Citation
Particulate phosphorus transport in the water conveyance systems of the Everglades Agricultural Area

Material Information

Title:
Particulate phosphorus transport in the water conveyance systems of the Everglades Agricultural Area
Creator:
Stuck, James Donald, 1940-
Publication Date:
Language:
English
Physical Description:
xxviii, 404 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Calibration ( jstor )
Canals ( jstor )
Hydraulics ( jstor )
Particulate materials ( jstor )
Phosphorus ( jstor )
Pumps ( jstor )
Sediments ( jstor )
Shear stress ( jstor )
Simulations ( jstor )
Velocity ( jstor )
Agricultural and Biological Engineering thesis, Ph. D ( lcsh )
Agriculture -- Environmental aspects -- Florida -- Everglades ( lcsh )
Dissertations, Academic -- Agricultural and Biological Engineering -- UF ( lcsh )
Water -- Phosphorus content -- Florida -- Everglades ( lcsh )
The Everglades ( local )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1996.
Bibliography:
Includes bibliographical references (leaves 392-403).
Additional Physical Form:
Also available online.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by James Donald Stuck.

Record Information

Source Institution:
University of Florida
Rights Management:
Copyright James Donald Stuck. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
36742012 ( OCLC )
026491902 ( ALEPH )

Downloads

This item has the following downloads:


Full Text










PARTICULATE PHOSPHORUS TRANSPORT IN THE
WATER CONVEYANCE SYSTEMS OF THE
EVERGLADES AGRICULTURAL AREA









By

JAMES DONALD STUCK










A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY


UNIVERSITY OF FLORIDA 1996
























Copyright 1996

by

James Donald Stuck

























This work is dedicated to the two people in my life who have helped me negotiate all the forks in the road, my wife Beverly, and my son Clark; and to the memory of Christopher.














ACKNOWLEDGMENTS


The author would like to express his appreciation to his major professor, Dr. K.L. Campbell of the Department of Agricultural and Biological Engineering. for his support throughout this effort. He would also like to acknowledge the other members of his committee: Dr. J. J. Delfino of Environmental Engineering Sciences, Dr. F. T. Izuno of Agricultural and Biological Engineering, Everglades Research and Education Center, Dr. A. J. Mehta of Coastal and Oceanographic Engineering, Dr. K. R. Reddy of Soil and Water Science, and Dr. A. B. Bottcher, President of Soil and Water Engineering Technology, special committee member.

Thanks are due to the Departments of Agricultural and Biological Engineering, Coastal and Oceanographic Engineering, and Soil and Water Science, and the Everglades Research and Education Center, all of which contributed laboratory space and equipment to this interdisciplinary study.

The author is indebted to many individuals who provided invaluable assistance during the course of this work: Ms. Yu Wang and Dr. Winston Davis for their analytical help, Ray Garcia and Curtis Miller for their patient advice and essential help in the field, Laurene Capone for helping get the resources and smoothing the path so often, Mauricio Abreu for his many long hours working on the erosion studies, and Dr. Nigel Pickering for his considerable computer skills.

Financial support is gratefully acknowledged from the U. S. Department of Agriculture, which provided the author with a Research Fellowship in Hydrologic Science while this research was in progress, and from the Everglades Agricultural Area Environmental Protection District, which provided the operating funds for this research as iv















a part of their "Best Management Practices Implementation" program for reducing phosphorus loading to the Everglades.

The author would particularly like to thank his loving, patient and supportive wife, Beverly, who contributed to the completion of this work in ways too numerous to detail.



































V















TABLE OF CONTENTS


ACKNOWLEDGMENTS .............................. ................................... iv

LIST O F TA BLES ........................................ ..... .................. xi

LIST O F FIG U R E S........................................................................................ ............ xiii

NOMENCLATURE............................ .. ............. xxii

A B STRA CT ....................................................... xxvii

CHAPTERS

1 INTRODUCTION, BACKGROUND, AND RESEARCH
M OTIVATION ...............................................................

Historical Summary ................................................................. I
Sources of Phosphorus in the EAA............................................. ..........5
Proposed Regulatory Remediation ............................................. .......... 7
Areas of Uncertainty in Remediation Plan ...................... ............
EAA M odeling Activity....................................................10

2 PRIOR DATA RELEVANT TO PARTICULATE PHOSPHORUS
TRANSPORT IN THE EVERGLADES REGION ..................................... 12

Introduction...................................... ......... 12
Farm-Field Scale Studies........................................ 13
CH2M Hill study............................ ................. 13
Izuno, et al study ....................................... ...................15
Canal Transport Studies.......................................16
Southeast Florida canals water quality studies ..................................... 16
Water Conservation Area water quality correlations ....................................17
EAA Canal Sediment Studies................................... .............. 17
Anderson and Hutcheon Engineers study (1992) ......................................18
Andreis/U.S. Sugar proposed sediment control BMPs (1993) .................. 19
Sediment control demonstration project ............................. 20
South Florida Water Management District Canal Data Set .............................22

3 GENERAL LITERATURE REFERENCES RELEVANT TO
PARTICULATE PHOSPHORUS TRANSPORT ...................................... 24


vi










Introduction........................24....... .. .. ..................24
Sources of Particulate Phosphorus.......................... .... .. .................24
Early D iagenesis ...................................... .............. .........27
Erosion and Transport of Organic Material ..................... .....28
General erosion and transport ..................................... .......... 28
Hydrologic approach to seston transport .................... ....37
A Brief Discussion of Sediment Transport Theory ..................................... 39
Non-Cohesive sediment transport.......................... .................39
Cohesive sediment transport..................................43
Particle entrainm ent simulator ..................... ........................... 48

4 INITIAL HYPOTHESES, OBJECTIVES, AND RESEARCH PLANS............51

-Problem Overview ........................................ ..................51
Summary of sources.....................................................51
Original conceptual model .................................. .......................53
Scope of This Research.....................................................55
Minimum criteria for experimental and modeling efforts .........................55
Application of minimum criteria and resulting conceptual model
simplification ....................... .. .......................56
Research goals for this study ....................... ........... .............58
General Research Plan ....................... .....................................60

5 SEDIMENT SURVEY AND PHYSICAL-CHEMICAL
CHARACTERIZATIONS OF SELECTED SEDIMENTS.............................62

Sediment Survey .............................................................62
Selection of representative farms....................................................62
Sediment survey measurement methods .................... ....64
Sediment survey results .......................... ..........65
Sediment survey discussion .................................. .....................66
Selection of Primary Target Farm......................................................69
Particle-Size Property Distribution Study..................... ......69
Sources for particle size fractionation study ..............................................70
Soil and sediment particle size fractionation ..................................... 71
Fraction analysis ................................................73
Particle size distribution ........................ ..................................74
Volatile matter content..... .................................77
Particle specific gravity ...................... .....................79
Total phosphorus content................... ..... ...................80
Phosphorus fractionation results ............................. .......................82
Adsorption-Desorption experimental methods.............................. ...84
Adsorption-Desorption data reduction.................... ......85
Potential environmental significance of adsorption-desorption ................91
Sedimentation parameters of B9B10 sediment................................95


vii









6 SURFICIAL SEDIMENT TRANSPORT STUDIES ......................................99

Introduction................. .............. .... ................ 99
Prototype Sedim ent ...................................... ............................................ 100
Laboratory Erosion Simulation Device- The Counter-Rotating Annular
Flum e ........................................ ................. .... .... .............. 101
Flum e description ....................................................................... ..........101
Flum e operation .................................................................................... 105
Organic sediment flume studies............................. .............107
Organic Sediment Particle Entrainment Simulator Studies...........................132
Particle Entrainment Simulator configuration ..................................... 132
Particle Entrainment Simulator operation preliminary clay tests............. 136
Particle Entrainment Simulator operation organic sediment tests............ 137
Calibration of PES against flume and extrapolation to field
parameters ............................................................. 141
Field Measurements using the calibrated particle entrainment
sim ulator .................................................. 148

7 FIELD EVENT STUDIES--TIME SERIES AND SYNOPTIC ....................155

Introduction........... ................. ........................155
Time Series Discharge Studies ............................................ ...........156
Data Monitoring and sample acquisition ........................................ 156
Detailed hydrography of the target study farm--UF9200......................159
Particulate phosphorus content determination........................ .. 161
Particulate phosphorus event studies--Farm UF9200 .................................163
Suspended solids transport--Farm UF9200 ....................................... 168
Particulate phosphorus content of suspended solids--Farm UF9200..........176
Particulate phosphorus event studies--Farm UF9206, North and
South stations............................ ..... ..... ..................... 181
Suspended solids transport Farm UF9206................................................89
Particulate phosphorus content of suspended solids Farm UF9206.........196
Intensive Synoptic Studies at UF9200 .......................... ..........................200
Sampling configuration.............................................200
Ditches vs. discharge ........................ ...............204
Influence of the North Canal ................................................. 207
Downstream variation of suspended solids concentration ..........................208
Order of Magnitude Transport Estimates ..................................... 211
Summary .... .................. ...................... .......................216

8 FIELD SPECIFIC STUDIES--FIELD SEDIMENTS, MACROPHYTE
AND DETRITUS STUDIES, LARGE COMPOSITE SAMPLES ...............217

Introduction.................... ... .... ......................................... 217
Sediment Phosphorus Content Synoptic Surveys............................. ....218
Macrophyte and Detritus Studies.......................... ........................224


viii









Macrophyte areal density and gross total phosphorus content studies ....... 226 Dislodgable detritus studies ................................................... 228
Large Composite Sample Studies .............................................. .........233
Adsorption/desorption tests ..................................... ... ........ 235
Phosphorus fractionation Tests on Large-Scale Composite Samples.........240
Sum m ary............................................ .... ......... ... ....... ...........243

9 MODEL DEVELOPMENT.....................................244

Introduction ............................ ............ ......................................................244
Exclusion of Overland Flow Erosion and Sorption from the Model ................246
Model Platform Selection and Hydraulic Calibration ...................................248
M odel review and selection ..................................... ... .............248
General characteristics of the DUFLOW program ..................................250
DUFLOW water quality module................................. .........252
Field drainage model.......................................................254
Hydraulic model calibration ..................................... ............256
The Evidence for Biological-Growth Controlled Particulate Phosphorus
Transport ............................................ ..............265
M odified M odel Development Plan ..................................... ......................267
Model Development ........................................................ 269
General model assumptions .................................. ........... ..........269
M odel derivation........................... ........ ......................270
Transport model calibration methodology ............................................. 273

10 TRANSPORT MODEL RESULTS AND DISCUSSION.............................278

Introduction............................ .... .....................278
Calibration Parameter Results .............................................. ...........278
Normal Wet Season Simulation Results................................................ 281
Suspended solids simulation................................................ .................281
Cumulative suspended solids load and particulate phosphorus load
simulation results ...............................................289
Small flows ...........................................296
Correlation of Calibration Values of CEM(O), Initial Erodable Mass ..............297
Comparison of Calibration Results for Erodable Mass with Literature
Values ....... ........................................301
Model Validation Prediction of the Suspended Solids and Particulate
Phosphorus Loads for the Late Season Storms....................304
Summary, Model Qualifications and Applications.......................... ....310
Summary ....................... .............................. 310
Model qualifications .................................................... ...................... 13
M odel applications..... ........... ......................................................314

11 MODEL APPLICATION, CONCLUSIONS AND
RECOMMENDATIONS ..................................... 318


ix










Introduction ............................................... .. ............ ....... ..................... 3 18
Evaluation of Potential Management Practices ............................................. 318
Reduction of transportable mass in the system................................ 319
Effect of sediment trap........................ .. ....................320
Effect of pumping modification.....................................322
Effect of wider discharge outlets ................................ ............ .......324
Hydraulic mining and recycle analysis ...................... ....................... 326
Specific effect of reduced velocity ............................ ......... 332
Evaluation of Suspended Solids and Particulate Phosphorus Export at
Farm UF9206.......................................................334
Possible dominant influence of channel water depth on velocity and
transport ..............................................................334
Correlation of export suspended solids with channel depth ....................337
Particulate phosphorus export at UF9206............................... ....341
Relation of Particulate Phosphorus Export to Soluble Phosphorus Export
at Both Farms........................ ....... ........................345
Conclusions and Critique............... ...................................346
Conclusions.......................... ....... ........... ............346
Critique of the model in its current form ....................... 350
Recommendations.................................................. .. ................. ....352
BMP recommendations for field implementation ...................................352
Model framework development................ .... .........353
M odel development ....................... ....................................354
BM P development ........................ .....................................355

APPENDICES

A PHYSICAL SAMPLING TECHNIQUES...........................356

B PHYSICAL ANALYSIS............................. ........................359

C CHEMICAL ANALYSIS ...................... ........................363

D APPROXIMATE MODEL FOR CRAF .............................................365

E RAINFALL AND WATER LEVELS AT UF9200 .................................. 369

F DIMENSIONS OF UF9200 USED IN DUFLOW FORMAT.......................373

G FARM LAYOUTS FOR UF9200 AND UF9206 ..................................... 386

H DUPROL EROSION PROGRAM.........................................389

REFERENCES ... .............................................................. ....................... 392

BIOGRAPHICAL SKETCH ..............................................404
















LIST OF TABLES

Table page

5.1 Sediment Survey Physical Results ..........................................................65

5.2 Sediment Survey Analytical Results............................. ....................65

5.3 Estimates of Total Phosphorus Mass in Target Main Canal Sediments ................68

5.4 Soil and Sediment Screen Fractionation Sequence.............................. ....72

5.5 Phosphorus Fractionation Results for WPBC and B9B10 Sediments................83

5.6 Sorption Parameters and Coefficient of Determination for WPBC and
B9B10 Sediments.......................................................90

5.7 Summary of Phosphorus Distribution in Demonstration Example 1 ..................93

5.8 Summary of Phosphorus Distribution in Demonstration Example 2..................94

5.9 Solids Content and Bulk Densities of WPBC and B9B 10 <38 Micrometer
F ractions........................................................................................................... 98

6.1 Constants for Fit of Equation 6.5 to Multiple Shear Erosion Tests ..................... 113

6.2 Constants for Fit of Equation 6.8 to Multiple Shear Deposition Tests with
Water Depth of 30.5 cm ....................................................... 117

6.3 Constants for Fit of Equation 6.8 to Multiple Shear Deposition Tests with
Water Depth of 15.2 cm .......................................121

6.4 Field PES Sample Point Descriptions........................ ................. 150

7.1 Average Conveyance Dimensions for UF9200.................................. 161

7.2 Summary of Pumping Events for Target Farm UF9200.................................. 165

7.3 Summary of Pumping Events for Station UF9206N ..................................... 183

7.4 Summary of Pumping Events for Station UF9206S ......................................... 184


xi











7.5 Comparison of Pumping Events Between Farms UF9200 and UF9206 for
Julian Dates 218 through 350 ............................................................185

7.6 Sampling Locations for Intensive Synoptic Studies at UF9200 .......................202

7.7 UF9200 Events 220-285 Expressed in Areal Loading Terms ..........................212

7.8 UF9200 Weighted Areal Loads Estimated From Equations 7.7 and 7.8.............214

8.1 Field Ditch Surficial Sediment Phosphorus Content UF9200 Synoptic
Survey of Julian Date 229........................................................219

8.2 Canal Surficial Sediment Phosphorus Content UF9200 Synoptic Survey
of Julian Date 229 ...........................................................220

8.3 Canal Surficial Sediment Phosphorus Content UF9200 Survey of Julian
Date 300......................................... ............223

8.4 Macrophyte Mass Density. Volatile Content, and Phosphorus Content ..............227

8.5 Water Lettuce Dislodgable Detritus Study .................................... 231

8.6 Phosphorus Fractionation Analysis of Large-Scale Composite Samples ............241

8.7 Comparison Between Large-Scale Composite Samples and Prototype
Surficial Sediment for Phosphorus Content Distribution .................................242

9.1 Results of Hydraulic Calibration for UF9200 .........................263

10.1 Erosion Parameters Determined in the Calibration Process .............................279

10.2 Placement of Example Nodes for Event UF9200-327 Simulation ...................286

10.3 Interevent Time and Values Of CEM(O) Used for Validation Simulation...............305

11.1 Load Weighted Distribution of Soluble and Particulate Phosphorus at
UF9200, UF9206N, and UF9206S ......... ....... ........................346

D. Example of Spreadsheet Output. .................................367

D.2 Spreadsheet Formulas ........................................368








xii
















LIST OF FIGURES

Figure pg

1.1 Original Everglades Watershed ................. : ....... ...................

4.1 Farm Scale Phosphorus Transport Original Conceptual Model ....................53

4.2 Farm Scale Phosphorus Transport Simplified Conceptual Model ................59

5.1 Particle Size Distribution of Dry and Hydrated B9B10 Soil ...........................75

5.2 Particle Size Distribution of Hydrated B9B10 Soil and B9B10 Sediment.........77

5.3 Particle Size Distribution of Hydrated B9B10 Soil, B9B10 Sediment,
and WPBC Sediment ..............................................78

5.4 Volatile Fraction (Organic Content) Distribution of Hydrated B9B10
Soil, B9B10 Sediment, and WPBC Sediment .......................................79

5.5 Particle Specific Gravity Distribution of Hydrated B9B10 Soil, B9BI0
Sediment, and WPBC Sediment ........ ........... .....................80

5.6 Correlation of Particle Specific Gravity with Organic Content for
Hydrated B9B10 Soil, B9B10 Sediment, and WPBC Sediment .....................81

5.7 Total Phosphorus (TP) Content Distribution of Hydrated B9B10 Soil,
B9B10 Sediment, and WPBC Sediment..... ..... ......................81

5.8 Phosphorus Fractions vs. Organic Fraction for WPBC and B9B10
Sedim ents ....................................... ............................................83

5.9 Adsorption Isotherm-WPBC Sediment 75-150 Micrometer Range ................88

5.10 Adsorption Isotherm-WPBC Sediment <38 Micrometer Range ...................89

5.11 Adsorption Isotherm-B9Bl0 Sediment 75-150 Micrometer Range ................89

5.12 Adsorption Isotherm-B9B1 0 Sediment <38 Micrometer Range .....................90




xlli










5.13 Sedimentation Velocity Distribution as Determined by The Bottom
Withdrawal Technique-B9B10 Sediment <38 Micrometer Particle Range........97

5.14 Mean Sedimentation Velocity vs. Test Duration ....................... ............. 97

6.1 Schematic Diagram of Counter-Rotating Annular Flume (CRAF) ...............102

6.2 Photograph of Counter-Rotating Annular Flume (CRAF) ............................ 103

6.3 CSME vs. Time Into Run for Placed Organic Sediment Bed with 15.2
cm. W ater Depth .................................................... 108

6.4 CSME as Function of Consolidation Time for Organic Sediment Bed
with 15.2 cm. Water Depth ........................................ 110

6.5 CSME as Function of Shear Time at Constant Calculated Shear Rate of
2.0 dynes/cm2 for Consolidated Organic Sediment Bed with 30.5 cm.
Water Depth............................................. 11

6.6 CSME as Function of Shear Time at Several Calculated Shear Rates for
Erosion of Consolidated Prototype Organic Sediment Bed with 30.5 cm.
Water Depth ............................................. 112

6.7 Initial Erosion Rates vs. Applied Shear Stress for Organic Sediment
Extended Run................................ ............................ 114

6.8 Asymptotic CSME vs. Applied Shear Stress for Organic Sediment
Extended Run Erosion at 30.5 cm. Water Depth.................... 115

6.9 CSME as Function of Shear Time at Several Calculated Shear Rates for
Deposition of Consolidated Prototype Organic Sediment Bed at 30.5 cm.
Water Depth .......... ....................... ............ 117

6.10 Asymptotic CSME vs. Applied Shear Stress for Organic Sediment
Extended Run Erosion and Deposition with 30.5 cm. Water Depth ..............118

6.11 CSME as Function of Shear Time at Several Calculated Shear Rates for
Erosion of Consolidated Prototype Organic Sediment Bed at 15.2 cm.
Water Depth .................. .................. ....... 122

6.12 Asymptotic CSME vs. Applied Shear Stress for Prototype Organic
Sediment Extended Run Erosion with 30.5 cm. and 15.2 cm. Water
Depth.. .............. .. ......... ....... ........... ........... 123

6.13 Asymptotic CSME vs. h213V2 for Prototype Organic Sediment Erosion
Extended Run with 30.5 cm. and 15.2 cm. Water Depth..................................131


xiv











6.14 Schematic Diagram of Particle Entrainment Simulator (PES) .........................133

6.15 Photograph of Particle Entrainment Simulator (PES) ................................... 134

6.16 PES-CRAF Calibration Results Using Cohesive Clays............................... 137

6.17 Typical PES Erosion Curve with Organic Sediment ....................................139

6.18 Results of PES Runs with Prototype Organic Sediment CSME vs. Grid
Oscillation Frequency ........................................141

6.19 Velocity Profile Measurements Reported by Mehta (1973) in CRAF with
22.9 cm W ater Depth ........................................................ 145

6.20 Schematic Representation of Superposition of Extended Logarithmic
Velocity Profile on CRAF Sigmoidal Velocity Profile.............................. 146

6.21 Final Calibration Curve of Estimated Equivalent Open Channel Mean
Velocity For Erosion as a Function of PES Oscillation Frequency ...............147

6.22 Example of Field PES Test Results ............................................... 151

6.23 Compiled Field PES Test Results for UF9200 ...............................153

6.24 Compiled Field PES Test Results for UF9206 ................... .........................154

7.1 Correlation of Event TSS Load with Event Hydraulic Load for Farm
UF9200 .............. ......... ..........................167

7.2 Correlation of Event PP Load with Event Hydraulic Load for Farm
UF9200 ..................... ............................167

7.3 Hydraulic and Total Suspended Solids Profiles for Event UF9200-154 ..........169

7.4 Hydraulic and Total Suspended Solids Profiles for Event UF9200-220 ..........169

7.5 Hydraulic and Total Suspended Solids Profiles for Event UF9200-237 ..........170

7.6 Hydraulic and Total Suspended Solids Profiles for Event UF9200-244 ..........170

7.7 Hydraulic and Total Suspended Solids Profiles for Event UF9200-252 ..........171

7.8 Hydraulic and Total Suspended Solids Profiles for Event UF9200-258 ..........171

7.9 Hydraulic and Total Suspended Solids Profiles for Event UF9200-262 ..........172

7.10 Hydraulic and Total Suspended Solids Profiles for Event UF9200-285 ..........172


xv











7.11 Hydraulic and Total Suspended Solids Profiles for Event UF9200-319 ..........173

7.12 Hydraulic and Total Suspended Solids Profiles for Event UF9200-336 ..........173

7.13 Hydraulic and Total Suspended Solids Profiles for Event UF9200-355 ..........174

7.14 UF9200 Phosphorus Content of TSS as a Function of TSS Concentration......177

7.15 UF9200 Phosphorus Content of TSS as a Function of TSS Concentration
(Expanded Scale) ...........................................................177

7.16 UF9200 Particulate Phosphorus Concentration in Discharge as a
Function of TSS Concentration (Expanded Scale) ..................................... 180

7.17 Correlation of Event TSS Load with Event Hydraulic Load for Station
U F9206N .................................................. ....... .....................187

7.18 Correlation of Event PP Load with Event Hydraulic Load for Station
UF9206N .................................... ................... 187

7.19 Correlation of Event TSS Load with Event Hydraulic Load for Station
UF9206S ..............................................................188

7.20 Correlation of Event PP Load with Event Hydraulic Load for Station
UF9206S ..............................................................188

7.21 Time Series-Flow and Total Suspended Solids Events UF9206N-209,
213................ ........... ...... .......................... 190

7.22 Time Series-Flow and Total Suspended Solids Events UF9206N-224,
234, 239................................................190

7.23 Time Series-Flow and Total Suspended Solids Events UF9206N-25 1,
256, 260............... .. ..................................191

7.24 Time Series-Flow and Total Suspended Solids Events UF9206N-268,
273,279..................... ....... ....... ................................ 191

7.25 Time Series-Flow and Total Suspended Solids Events UF9206N-283,
300... ...................................................... 192

7.26 Time Series-Flow and Total Suspended Solids Event UF9206N-318..............192

7.27 Time Series-Flow and Total Suspended Solids Event UF9206N-336.............. 193




xvi











7.28 Time Series-Flow and Total Suspended Solids Events UF9206S-209,
213,218,224, 234.....................................1......................193

7.29 Time Series-Flow and Total Suspended Solids Events UF9206S-260.
268...................................... ............ ....................... 194

7.30 Time Series-Flow and Total Suspended Solids Events UF9206S-280,
3 0 0 .....................................................................................................................19 4

7.31 Time Series-Flow and Total Suspended Solids Event UF9206S-317 ..............195

7.32 Time Series-Flow and Total Suspended Solids Event UF9206S-336 ..............195

7.33 UF9206N Phosphorus Content of TSS as a Function of TSS
Concentration............................ .......................197

7.34 UF9206S Phosphorus Content of TSS as a Function of TSS
Concentration................................. ............. 197

7.35 UF9206N Phosphorus Content of TSS as a Function of TSS
Concentration (Expanded Scale)............................... .....................198

7.36 UF9206S Phosphorus Content of TSS as a Function of TSS
Concentration (Expanded Scale)............................... .. ...................198

7.37 Layout of UF9200 with Synoptic Sampling Locations.......................... 203

7.38 UF9200 Intensive Study-Event 252 .................... .........205

7.39 UF9200 Intensive Study-Event 258 ............................. .... ..................205

7.40 UF9200 Intensive Study-Event 262 .............................. ............... .......206

7.41 UF9200 Intensive Study-Event 285 ................................. .... ...... 206

7.42 UF9200 Intensive Study-Event 258 South-East Canal Sample Locations,
First 10 Hours ............................................209

7.43 UF9200 Intensive Study-Event 262 South-East Canal Sample Locations,
First 12 Hours ..........................................................209

7.44 UF9200 Intensive Study-Event 285 South-East Canal Sample Locations,
First 32 Hours ............................................210

8.1 UF9200 Two-Point Adsorption/Desorption Curves (Adsorption Mode).........236



xvii











8.2 UF9200 Two-Point Adsorption/Desorption Curves (Desorption Mode) .........236

8.3 UF9206N Two-Point Adsorption/Desorption Curves (Adsorption Mode)......237

8.4 UF9206N: Two-Point Adsorption/Desorption Curves (Desorption
M ode)...................... .... .... ..................................237

8.5 UF9200 Desorption Data Eadie-Hofstee Plot................................ ...239

8.6 UF9206N Desorption Data Eadie-Hofstee Plot.......................................... 239

9.1 Simplified Schematic of UF9200 DUFLOW Nodal Structure .........................257

9.2 Calculated and Measured Flows, Event UF9200-237 ...................................261

9.3 Calculated and Measured Flows, Event UF9200-252 ..................................261

9.4 Calculated and Observed Levels, Event UF9200-220 ...................................262

10.1 Simulation Results-Total Suspended Solids for Event UF9200-220 ................282

10.2 Simulation Results-Total Suspended Solids for Event UF9200-237................282

10.3 Simulation Results-Total Suspended Solids for Event UF9200-252................283

10.4 Simulation Results-Total Suspended Solids for Event UF9200-258................283

10.5 Simulation Results-Total Suspended Solids for Event UF9200-262................284

10.6 Simulation Results-Total Suspended Solids for Event UF9200-285................284

10.7 Simulation of TSS at Selected Nodes for Event UF9200-237.................. 287

10.8 Simulation of EM at Selected Nodes for Event UF9200-237........................287

10.9 Simulation Results-Cumulative Suspended Solids Load for Event
UF9200-220 ..................... ...........................290

10.10 Simulation Results-Cumulative Particulate Phosphorus Load for Event
UF9200-220 .............................. ......................290

10.11 Simulation Results-Cumulative Suspended Solids Load for Event
UF9200-237.. ............. .......................................................... 291

10.12 Simulation Results-Cumulative Particulate Phosphorus Load for Event
UF9200-237 ....................... .........................291



xviii











10.13 Simulation Results-Cumulative Suspended Solids Load for Event
UF9200-252 .............................................................. 292

10.14 Simulation Results-Cumulative Particulate Phosphorus Load for Event
UF9200-252 ....................... .........................292

10.15 Simulation Results-Cumulative Suspended Solids Load for Event
UF9200-258 ....................................................293

10.16 Simulation Results-Cumulative Particulate Phosphorus Load for Event
UF9200-258 ............................................ ..... ...........293

10.17 Simulation Results-Cumulative Suspended Solids Load for Event
UF9200-262 .......................................... ............294

10.18 Simulation Results-Cumulative Particulate Phosphorus Load for Event
UF9200-262 ......................................... ... .............294

10.19 Simulation Results-Cumulative Suspended Solids Load for Event
UF9200-285 ................................................................295

10.20 Simulation Results-Cumulative Particulate Phosphorus Load for Event
UF9200-285 ............................................................295

10.21 Correlation of Initial Erodable Mass with Interevent Time ...........................297

10.22 Chronological Values of CEM(O) ......................... ............................................299

10.23 CEM(0) Correlated as an Exponential Growth Equation............................ 300

10.24 Suspended Solids Simulation Results for Event UF9200-319 ......................307

10.25 Suspended Solids Simulation Results for Event UF9200-355 ......................307

10.26 Cumulative Suspended Solids Load Simulation Results for Event
UF9200-319 ................................................308

10.27 Cumulative Suspended Solids Load Simulation Results for Event
UF9200-355 ....................................................308

10.28 Cumulative Particulate Phosphorus Load Simulation Results for Event
UF9200-319 ..................................................309

10.29 Cumulative Particulate Phosphorus Load Simulation Results for Event
UF9200-355 ....................................................309



xix











11.1 Effect of Transportable Mass on Exported Suspended Solids for Event
UF9200-220 .............................................319

11.2 Simulated Effects of Sediment Trap on Event UF9200-220 .........................320

11.3 Simulated Erodable Mass Concentrations at Selected Sections of SouthEast Canal for Trap Simulation........................... .................321

11.4 Effect of Pumping Modification (Simulation A) ..................................... 324

11.5 Effect of Pumping Modification and Flow Width Modification
(Simulation B)........ .......... .......... ....................325

11.6 Effect of Pumping Modification and Flow Width Modification
(Simulation C).......................................... ........... ...... 326

11.7 Solids Recycle Simulation RCL Short Duration, Short Distance..............328

11.8 Solids Recycle Simulation RCL2 Long Duration, Long Distance ..............329

11.9 Solids Recycle Simulation RCL3 Long Duration, Long Distance,
Higher Flow, No Backflow Prevention.......................... ............330

11.10 Simulation RCL4 Long Duration, Short Distance, Higher Flow,
Backflow Prevention........................... .. ..... ......................331

11.11 Specific Effect of Reduced Velocity via Constant Discharge............333

11.12 Water Elevation Trends for UF9206 Pump Stations ....................................336

11.13 Water Elevation Trends for UF9200 Pump Station..........................................336

11.14 Correlation of TSS Concentration with Canal Level at UF9206N ................338

11.15 Correlation of TSS Concentration with Corrected Canal Level at
UF9206S ..................... ... ................ .............339

11.16 Correlation of TSS Concentration with Canal Depth at UF9200 ..................339

11.17 PP Content vs. TSS Correlations for All Three Farms ..................................341

11.18 Cumulative Fractional PP Load vs. TSS for All Three Farms......................342

11.19 W ater Level Control Example for UF9206N.................................................344

11.20 Water Level Control Example for UF9206S ........................................344



xx











D.1 CRAF Approximate Model: Velocity-Depth Profile.....................................366

D.2 CRAF Approximate Model: Velocity-Depth Profile (Dimensionless in
Velocity)........................ ... .... ........... .................... 366

E.1 Rainfall at UF9200 Julian Days 150-356................. .........370

E.2 Selected Water Levels at UF9200 Julian Days 150-300.............................371

E.3 Selected Water Levels at UF9200 Julian Days 300-365 ...............................372

G.1 Layout of UF9200 ............................................386

G.2 Layout of UF9206 .............................................387

G.3 Recycle Simulations UF9200 ........ ........... .....................388




































xxi















NOMENCLATURE



a datum level elevation A area under water column, or crossectional area B water surface width C concentration in solution or suspension Co initial concentration at time zero Ca concentration at datum level a Cb concentration of suspended sediment CEM surface concentration of erodable mass CEM(O) concentration of erodable mass at time zero Cj concentration of suspended particles of class 'f' Css volumetric concentration of suspended solids CSME Cumulative Specific Mass Eroded CSMEo CSME at start of application of new shear stress CSME0o CSME at the low level of erosion CSMEhigh CSME at the high level of erosion d particle diameter D dispersion coefficient Deff effective drain depth EMj erodable mass of class '"' present at time t Fd deposition flux FE erosive phosphorus flux FN net phosphorus flux xxii









Fs sedimentation phosphorus flux gB weight rate of bed transport per unit width h water column height or actual canal water surface elevation JD Julian Date

k partition coefficient or surface roughness ks grain roughness Ks saturation constant for sorbate-substrate system Ksa, saturated hydraulic conductivity m water level difference between mid-field and ditch n Manning's coefficient of roughness P production of a constituent per unit section length Pdcs mass of phosphorus desorbed from sediment PC, phosphorus content of discharged TSS for the first four hours of an event PC, phosphorus content of discharged TSS for the remainder of an event PCN phosphorus content of discharged TSS for station UF9206N PCs phosphorus content of discharged TSS for station UF9206S PPI particulate phosphorus concentration in the discharge for the first four

hours of an event

PP2 particulate phosphorus concentration in the discharge for the remainder of

an event

q volume rate of water flow per unit width qs volume rate of suspended solids discharge per unit width Q flow rate
Re, wall Reynolds number Rh hydraulic radius of channel RPM grid oscillation frequency (revolutions/min) s ditch spacing


xxiii










S bed slope, or hydraulic gradient, or substrate concentration of sorbate. or

the quantity of a constituent passing a cross-section per unit time So native sorbate concentration originally present on substrate Sa substrate concentration of sorbate adsorbed from liquid phase Sm maximum sorbate concentration at complete saturation of sorption sites t time

TSS total suspended solids concentration u velocity U. shear velocity V volume of water column, or channel velocity VC critical channel mean velocity for erosion V, velocity which corresponds to Tcj Vmean mean stream velocity w specific weight of water W, settling velocity of sediment Wsed sedimentation velocity W4 Sedimentation velocity of particles in class 'f' x x-coordinate, or h2/3V2 y elevation, or distance from surface yo water surface elevation z elevation of sediment surface above datum plane




Greek Letters

a rate coefficient, or erosion constant

0 constant of proportionality y specific mass of water


xxiv










7,5 specific mass of sediment c erosion rate. or erosion coefficient E:r floc erosion rate EM erosion rate constant
Y
710


la reference level K von Karman's constant, or the collected constants epn2 v kinematic viscosity p water mass density T shear stress to surface shear stress Tb shear stress applied to bed Tc Critical shear stress for erosion Tcd critical shear stress for deposition Tcj Critical shear stress for erosion of class '". Td critical depositional shear stress Ts Bed shear strength AV CRAF ring-channel relative velocity




Abbreviations


BMP Best Management Practices CRAF Counter-Rotating Annular Flume CSFFCD Central and South Florida Flood Control District





xxv









CSME Cumulative Specific Mass Eroded EAA Everglades Agricultural Area EAAEPD Everglades Agricultural Area Environmental Protection District ENP Everglades National Park EPA Everglades Planning Area EPD Environmental Protection District EREC Everglades Research and Education Center FPOM Fine Particulate Organic Matter IFAS Institute of Food and Agricultural Sciences IP Insoluble Phosphorus OFCD Okeechobee Flood Control District PES Particle Entrainment Simulator PP Particulate Phosphorus SFWMD South Florida Water Management District SP Soluble Phosphorus STA Storm Treatment Area SWET Soil and Water Engineering Technology TDP Total Dissolved Phosphorus TP Total Phosphorus TSS Total Suspended Solids USACE United States Army Corps of Engineers WCA Water Conservation Area





xxvi














Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy


PARTICULATE PHOSPHORUS TRANSPORT IN THE WATER CONVEYANCE SYSTEMS OF THE EVERGLADES AGRICULTURAL AREA By

James Donald Stuck

December, 1996

Chairman: Kenneth L. Campbell Co-Chairman: Konda R. Reddy Major Department: Agricultural and Biological Engineering


The Everglades Agricultural Area (EAA) is a source of phosphorus nutrients to the downstream watershed, including Everglades National Park. This work considered export of particulate phosphorus (PP) from large EAA farms that are extensively drained by pumped canal systems.

The research objectives were to determine the major sources, locations, and mode of transport of PP, develop a model for PP transport and apply it to a representative farm, using an existing hydrodynamic model adapted to include subsurface drainage.

Representative sediments had organic matter and phosphorus contents of about 75% and 0.1%, respectively, with weak phosphorus adsorption characteristics. Typical farm canals contained sediment phosphorus equivalent to 15 to 30 years of discharged PP. Sediment transport properties, studied in an annular flume and a particle entrainment simulator, were used to develop a prototype model that included the concepts of a critical



xxvii









shear for erosion, supply limitation of erodable mass, and reversible erosionsedimentation.

Field studies showed the exported PP also had weak sorption characteristics but had higher, more labile phosphorus content than farm sediments. Additional studies led to the hypothesis that canal biological growth was the primary source of exported PP. The model, modified to incorporate this concept, was calibrated on the PP discharge data from the target farm during a normal wet season. It was validated on two tropical depression events, simulating PP export within about 10% of actual.

Farm management practices to reduce exported PP were simulated using the model. Macrophyt reduction and velocity control were shown to be critical factors. Sediment traps that do not reduce velocity appreciably were shown to be of limited value. In the simulations, canal modifications that reduced pump intake velocity reduced the first flush phenomenon, where a large fraction of the PP load is discharged at pump startup. Level control and elimination of pump cycling were shown to be cost-effective PP reduction mechanisms. The removal of untransported detritus from main canals was shown to be necessary for long-term control. Hydraulic transport schemes to achieve removal were recommended.

The model may be adapted to other farms to optimize water table control practices which minimize PP export.
















xxviii














CHAPTER 1I
INTRODUCTION, BACKGROUND, AND RESEARCH MOTIVATION



Historical Summary

The South Florida Everglades is a natural resource unique in the United States, if not in the world. The original Everglades watershed (See Fig. 1-1) was primarily a broad. freshwater marsh that extended from what is now the Kissimmee River basin, through Lake Okeechobee to the southern tip of the Florida peninsula, encompassing some 7500 square miles (-20,000 square km) of upland and wetland territory. The natural flow, arising from a shallow land slope of only a few cm/km, was at very low velocity from north to south along a riverbed which was often 80-100 km wide, with ultimate discharge into the waters surrounding the southern tip of Florida (Jones, 1948).

Flow was episodic and seasonal in nature, with Lake Okeechobee acting as a buffering reservoir. During the wet season (May through September) the level of Lake Okeechobee would rise above its shallow southern bank heights, providing a consistent flow of water through the wide riverbed. In the dry season the level of Lake Okeechobee would fall below the level of its southern banks, essentially shutting off the flow of water from the Kissimmee and Okeechobee basins to the land farther south. Under these conditions of reduced inflow and precipitation the marshlands south of Lake Okeechobee would slowly dry until, in the spring season, they would be subject to frequent and extensive fires. The onset of the wet season in the late spring would quench the fires and complete the annual wet-dry cycle (Douglas, 1988).














- ATLANTIC OCEAN






GULF
OF
MEXICO FLORIDA











SCALE IN MILES 0 40 80









Figure 1.1: Original Everglades Watershed (Adapted from Jones, 1948)






3


It is important to note for later reference that for thousands of years the hydrologic regime in the Everglades watershed was predominantly that of large volume discharge spread over shallow channel beds tens of kilometers wide flowing at velocities that were in the range of meters/day (Douglas, 1988).

The first serious human disturbance of the Evcrgladeswatershedcamej 188 when developer Hamilton Disston dug a canal between Lake Okeechobee and the Caloosahatchee River in an attempt to drain a 1.6 million ha (4 million acre) tract of wetlands. From that time through the present the planning and implementation of water management in the Everglades region has passed from the control of wealthy developers through a series of state trustee boards, commissions, and agencies. From 1906 to 1931 the Everglades Drainage District (EDD) followed by the Okeechobee Flood Control District (OFCD) installed a number of canals, levees, locks, and dams in the Everglades Region.

One of the objectives of the public works activity was to drain mucklands just south of Lake Okeechobee for agricultural use. This region, known as the Everglades Agricultural Area (EAA), consists of 700,000 acres (-283,000 ha) of rich organic soil that has supported a variety of agricultural activities since its creation. Currently. major winter vegetable and sugarcane industries are the predominant activities in this area. Because of the low elevation and flat terrain extant in the EAA and the swings in water supply from wet to dry season the crops grown there are highly dependent on artificial control of water table levels by off-farm pumping during wet periods and irrigation during dry periods. Much of the man-made hydraulic system installed in South Florida has had as one of its objectives the facilitation of water table control in the EAA.

In 1948 joint state-federal action created the Central and Southern Florida Flood Control District (CSFFCD), which acted in concert with the United States Army Corps of Engineers (USACE) to "develop and implement plans to provide flood protection, ensure adequate water supply, prevent saltwater intrusion along the Florida Lower East Coast,






4


enhance the region's fish, wildlife, and other environmental resources, and provide a water supply for the Everglades National Park" (Izuno and Bottcher. 1994). The CSFFCD has since been succeeded by the South Florida Water Management District (SFWMD), which assumed the responsibilities held by the CSFFCD in 1972. Cooperation with the USACE has proceeded uninterrupted since 1948 between first the CSFFD and then the SFWMD.

During the tenure of the CSFFCD major modifications were made in the hydraulic patterns and land use of thousands of square kilometers of Everglades wetlands. The EAA was completely canalized and diked. Fifteen hundred square miles (-3900 square kilometers) of wetlands to the south between the EAA and the Everglades National Park (ENP) were designated as Water Conservation Areas (WCAs) and dedicated to fish and wildlife conservation and surplus water storage. The storage capacity of Lake Okeechobee was increased and the capability to backpump farm stormwater runoff (or more appropriately farm stormwater "pumpoff') from the EAA collector canals upstream to the lake was implemented.

Unfortunately, during these many decades of major public works projects little attention was paid to the water quality ramifications of the profound hydraulic and land use modifications being imposed on the Everglades region. The ecological community in the Everglades originally evolved as an oligotrophic system, that is, one with low biological productivity, with phosphorus as the limiting nutrient. Studies conducted at the WCAs and ENP indicate that high nutrient concentrations, including phosphorus as a principal contributor, are causing significant shifts in the balance of the ecosystem (Whalen et al., 1992). According to the South Florida Water Management District, agricultural activities in the EAA, especially drainage and fertilization practices, have resulted in a major decline in the quality of water entering the north end of the Everglades Planning Area (EPA), the area encompassing the WCAs and ENP (Whalen et al., 1992).






5


The relationship of the EAA with Lake Okeechobee was recognized in the 1970s as being detrimental. A number of studies were conducted that linked the eutrophication of the lake to agricultural nutrient inputs, of which the EAA was a major contributor (Izuno and Bottcher, 1994). A number of remedial actions have been implemented to reduce the nutrient load on Lake Okeechobee, including eliminating backpumpingto _the lake from the EAA for all but few small private drainage districts. The effect of this policy is to force essentially all farm water discharge in the EAA to flow south to the WCAs, which, in the absence of any remediating action, would further increase the nutrient load vectored from the EAA toward the WCAs and the ENP.


Sources of Phosphorus in the EAA

Inputs to the potentially mobile phosphorus pool in the EAA can come from fertilizer application, rainfall, inflow water from Lake Okeechobee, and the natural mineralization process of the organic soil, where organic matter is converted to carbon dioxide and inorganic compounds. Fertilizer phosphorus input to the EAA has been estimated to be on the order of 18% of the total annual phosphorus input (Morris, 1975), however the fraction of this input that is available for transport to the drainage water is a strong function of fertilizer placement, crop uptake, soil microbial pool, and the iron and aluminum content of the soil (Sanchez and Porter, 1994). Rainfall is a poorly quantified but non-trivial source of phosphorus input. Studies on-going at the time of this writing indicate that the phosphorus content of rainfall in the EAA may be on the order of 0.050.08 mg/l (Izuno, 1995). At an EAA area of 280,000 ha with an average annual rainfall of 1.45 m the rainfall load might be in the range of 200,000-325,000 kg/year, or about

0.36-0.58 kg/ha-yr.

Precise estimates of the amount of phosphorus supplied to the EAA from Lake Okeechobee are difficult to generate because of the flow-through nature of the canal






6


system, the reuse of upstream farm discharge water by downstream farms, and the surprising lack of historic data on phosphorus concentrations in water directly at the lake outlets. A rough approximation of the maximum potential supply may be made using published average figures. Mean annual total phosphorus in Lake Okeechobee ranged from 0.05 to 0.1 mg/l for a 12 year period from 1973-1985 (Izuno et al., 1991). Annual flows from the EAA have been reported to be in the range of 1.2 x 109 m3 at the end of this period (Izuno, 1994). Assuming approximate equivalence between inflow and outflow, a potential maximum loading on the order of 50,000-100,000 kg/year may be estimated, or 0.18-0.36 kg/ha/yr.

The soils in the EAA are primarily Histosols, with depths ranging from 3 meters to 0.3 meters. The soil is underlain by marlad limestone. Approximately 200,000 ha are planted to sugarcane, about 40,000 ha are planted to vegetables, rice, and sod (Izuno et al., 1991). Phosphorus concentration in virgin EAA Histosols is estimated by Sanchez and Porter (1994) to be in the range of 0.08-0.35%. They note that in virgin soil about 24% of the phosphorus is inorganic, whereas in cultivated EAA soils up to 72% of the phosphorus is inorganic, implying significant mineralization resulting from cultivation. They further estimate that potential phosphorus mineralization is on the order of 20-150 kg/ha/yr for EAA soils. This estimate may be compared with Koch's (1991) estimate of mean total phosphorus storage in the upper 30 cm of EAA soil of 276 kg/ha. These quantities may be put into some perspective when it is calculated that 50.8 cm (20 in) of annual rainfall runoff from one hectare need only dissolve 0.508 kg of phosphorus to have a dissolved phosphorus concentration of 0.1 mg/l. The concentration of 0.1 mg/l is well above the natural level prevailing in the uncultivated interiors of the WCAs (see below) or the ENP.

Organic soils such as the EAA Histosols were formed under submerged, reducing conditions. When they are drained, cultivated, and aerated, they become net sources of phosphorus due to mineralization arising from increased aerobic microbial activity






7


(Sanchez and Porter, 1994). It has also been postulated that cropping increases the net rate of mineralizatio by further increasig the activity of the aerobic microbal population (Cosgrove, 1977). For these and other reasons related to crop and water management, the activities in the EAA have combined to increase the phosphorus loading outside the EAA.

Izuno et al. (1991) illustrate the increase with some analytical results which

indicate that the water pumped to Lake Okeechobee in the period 1983-1985 had a total phosphorus concentration in the range of 0.19-0.57 mg/l, compared with the lake's mean total phosphorus content of 0.05-0.10 mg/l. During the period 1978-1986 water pumped to the WCAs had total phosphorus concentrations in the range of 0.07-0.16 mg/l. The total phosphorus content in water in the interior of the WCAs, on the other hand, was in the range of 0.009-0.014 mg/I (Whalen et al., 1992). a factor of 6 to 7 lower than the incoming water. These figures serve to highlight the cause for concern regarding the modification of the oligotrophic status of the WCAs, and ultimately of the ENP.


Proposed Regulatory Remediation

The South Florida Water Management District, under the terms of the Florida House of Representatives' "Marjory Stoneman Douglas Everglades Protection Act of 1991" has the responsibility for ensuring mitigation of the phosphoruseffluxfrom-the EAA to the WCAs and the ENP. The District has proposed a remediation plan that requires implementation by EAA growers a series-ofBestManagelent Practices (BMPs), developed by Bottcher and Izuno (1993), and a massive treatment plan that involves the construction of four Stormwater Treatment Areas (STAs) within the EAA.

The BMPs, which are to affect 25% of the phosphorus load reduction, fall into two categories; fertility BMPs and water management BMPs.









Fertility BMPs include the use of calibrated soil testing to minimize overfertilization, banding of fertilizer to minimize placement of fertilizer in areas where it will not be accessible to the root zone, implementation of methods to prevent misplacement of fertilizer into drainage ditches, and split multiple applications of smaller loads of fertilizer to more closely match the crop uptake. Water management BMPs are aimed at maintaining the water table as close as possible to the maximum feasible level at all times to minimize oxidation and inraUzationfAhcsoiJjand reccling of runoff to retain phosphorus on-site, where possible. They include measures to minimize water table fluctuations, maintain uniform spatial and temporal distribution of the water table, and store runoff on fallow land or water insensitive cropland.

The water management BMPs require the grower to exercise much more

sophisticated control over runoff, irrigation, and ground water than has been practiced in the past. The grower is required to develop and track field water budgets, which incorporate irrigation, rainfall, and evapotranspiration accounting to optimize pumping and minimize excessive downward water table movement. Minimum canal-level pump shut-off controls are recommended to reduce excessive canal drawdown. Comparison of predicted rainfall with available field water storage capacity is recommended to reduce unnecessary prepumping in anticipation of storms. A number of hydraulic BMPs relate to achieving and maintaining hydraulic uniformity within the farm canal system to minimize unnecessary local drawdown of the water table. Water management BMPs are also detailed for retention of drainage water in on-farm reservoirs where possible, transfer of drainage water from more water-sensitive crops to fields with less water-sensitive crops, and the use of aquatic cover crops such as rice for uptake of phosphorus released elsewhere on-farm.

The STAs, which are to affect 75% of the phosphorus load reduction, are planned to be large artificial wetlands which will act as phosphorus removal buffers between the EAA and the WCA/ENP areas. These STAs may ultimately cover as much as 14,500 ha,






9


and are estimated to cost on the order of $320 million (Whalen et al., 1992). Cost of construction of these facilities will be borne, in part, by assessments on the EAA landowners. The current STA design goal is to achieve an average effluent total phosphorus concentration of less than 0.05 mg/I (Whalen et al., 1992). The design of each of the four STAs is based on treating a long-term average load from the sub-basin within the EAA served by the STA. This average load is determined from historical data and corrected for land removed from production for the STAs and for expected reduction in phosphorus loading resulting from grower implementation of BMPs.

Implementation of the combination of BMPs and construction and operation of the STAs was projected in 1992 to reduce phosphorus input to the WCAs by 80% within five years after initiation (Whalen et al., 1992).


Areas of Uncertainty in Remediation Plan

The possibility exists that the actual input to the STAs may be lowered below the long-term historical averages by more than 25% by implementation of the BMPs recommended by Bottcher and Izuno, by implementation of district-wide water management practices not yet specified, and by exploitation of biological and chemical processes in the irrigation ditches and drainage canals. The possibility also exists, with a somewhat lower probability, that release of phosphorus currently stored in the canal and drainage ditch sediments may tend to buffer the exit phosphorus concentrations from the various sub-basins and reduce the impact of the BMPs on the phosphorus load to the STAs. Significant changes of loading in either direction could have a substantial impact on the amount of land required for the STAs, or, alternatively, on the effluent phosphorus concentrations achieved by the STAs.

For any given time period, weather is always an uncertainty. Measurement of the effectiveness of any specific remediation strategy may be hampered by changes in the






10


precipitation and evapotranspiration patterns between time periods The long-term "endof-pipe" phosphorus loads to the ENP are set with some correction for antecedent flow at Shark River Slough (Whalen et al., 1992), which was based on statistical analysis of loading over an 11 year period, but statistical correlation over a multi-year period of multiple combinations of BMPs at the farm level is not an efficient method for evaluating and optimizing the effectiveness of BMPs. More precise correction methods are needed.

An appropriate alternative to long-term statistical evaluation is development and application of mathematical models which adequately describe the dominant hydrodynamic and chemodynamic processes controlling phosphorus transport and export at the field, farm, and basin scale. Such models could be used to forecast the impact of current or proposed BMPs under very site specific conditions, and to evaluate the actual performance of BMPs under field conditions which deviate from the long-term mean.


EAA Modeling Activity

The need for, and value of, models for phosphorus transport in the EAA has been recognized by both the growers and governmental agencies. One specific agency that combines aspects of both is the Everglades Agricultural Area Environmental Protection District (EPD), a state chartered organization with assessment powers that coordinates the environmental regulatory compliance activities of growers in the EAA. Among numerous research activities, the EPD has provided funds for BMP research.

The research is aimed at optimization of the original set of BMPs recommended by Bottcher and Izuno (1993) and development and application of new or modified BMPs for reduction of phosphorus export from the EAA. An integral part of this research activity is the development of field and farm scale models, specific to the EAA environment, for phosphorus sourcing and transport. This effort was originally initiated by the University of Florida Institute of Food and Agricultural Sciences (IFAS) and






11


continues at the time of this writing as a joint effort between IFAS and Soil and Water Engineering Technology (SWET), a private agro-environmental engineering firm in Gainesville. Florida. It is under this overall research charter that the work described in this dissertation was conducted.

The original water management BMPs focused on reducing phosphorus discharge by reducing the volume of water discharged from the farm and by minimizing water table fluctuations, that h~g an.aelerating effect on soil mineralization and phosphorus release. Thus the focus was on procedures that could reduce hydraulic load and minimize phosphorus solubilization. The contribution of particulate matter to phosphorus export was recognized (Izuno et al., 1991), however there were insufficient data available at the time of formulation of the original BMPs to quantitate the impact of management practices on particulate phosphorus transport and export. As a consequence none of the original water management BMPs spoke directly to control of particulate phosphorus transport. In addition, the main focus of the early modeling effort was on field scale transport, with the irrigation and drainage systems existing as boundary conditions. This resulted in an exclusion of the conveyance systems (field ditches, farm collector canals, main farm canals, pumping systems, and ultimately the district canals) from analysis and evaluation regarding phosphorus sourcing, assimilation, or transport. The impetus for the work reported herein was the need to fill that recognized gap by development of a first generation model which would give a reasonable approximation of the predominant phosphorus mobilization, transformation, and transport processes prevalent in the farm irrigation and drainage conveyance systems of the EAA.














CHAPTER 2
PRIOR DATA RELEVANT TO PARTICULATE PHOSPHORUS TRANSPORT IN THE EVERGLADES REGION


Introduction

The salient characteristics of the Everglades such as topography, soils and sediments, climate, and hydrology are not typical of those found in agricultural areas elsewhere in the United States. The area is extremely flat land at low elevation. frequently subject to intense local rainfall and periodically subject to intense regional rainfall, with drainage and irrigation under almost total human control.

The soils of the area are histosols, poorly drained muck and peatysoils low in

mineral content but rich in organic matter (in excess of 65% by weight), with low specific gravity and low hydraulic conductivity (Snyder, 1994), bearing more similarity to the soils and sediments of the Louisiana bayous than to upland mineral soils.

The climate is sub-tropical. The annual average solar radiation and wet season temperatures are more similar to those prevailing in the Caribbean and Central America rather than the United States southeast and western plains agricultural regions (Marsh, 1987, Trewartha and Horn, 1980).

The heavy reliance on pumping for water transport introduces flow-no-flow situations and hydraulic transients which are not typical of a more normal upland watershed. In many respects the hydrology of the EAA is more akin to that of the polder region of the Northern Netherlands where discharge hydrology is completely under human control (Snyder, 1994).

Given the individualistic character of the Everglades region in general and the EAA in particular it is not unreasonable to assume that particulate matter transport


12






13


standards which apply well to other locations in the United States might not translate completely to the Everglades environment. This assumption has underlain the planning and execution of the research work in this dissertation. With this in mind it may be appropriate to first examine the limited body of work available which relates directly to particulate matter and particulate phosphorus transport in the Everglades Region.



Farm-Field Scale Studies

There are two studies which have been done on the field or farm scale which are particularly pertinent to phosphorus transport in the EAA. They are those done by the engineering firm CH2M Hill (1978) on the farm scale and by Izuno et al. (1991) on the field scale.


CH2M Hill study

The CH2M Hill study evaluated a number of water quality parameters and was conducted over a period of 15 months in 1976-77 on multiple farm sites within the EAA. Intensive study was directed at one sugarcane field, one cattle ranch, and one vegetable farm which grew a variety of vegetables throughout the study period. In addition multiple secondary sites for each land use were studied less intensively and served as checkpoints for the representativeness of the intensive data. Key conclusions from the study regarding phosphorus follow.

The sugarcane farm showed study-period average effluent total

phosphorus of 0.110 mg/1, soluble phosphorus of 0.070 mg/1 (64% of

total), particulate phosphorus of 0.040 mg/I (36% of total) and an annual

effluent total phosphorus load of 0.65 kg/ha.






14


* The vegetable farm showed study average effluent total phosphorus of

0.460 mg/l, soluble phosphorus of 0.355 mg/I (77% of total), particulate

phosphorus of 0.105 mg/1 (23% of total) and an annual effluent total

phosphorus load of 2.10 kg/ha, roughly four times that of the sugarcane

farm.
* The cattle ranch showed study-period average effluent total phosphorus of

0.167 mg/l, soluble phosphorus of 0.122 mg/l (73% of total), particulate

phosphorus of 0.045 mg/1 (27% of total) and an annual effluent total phosphorus load of 0.55 kg/ha. Concentrations were higher than for

sugarcane but load was lower because less water was discharged.

* Soil water was monitored. Average soil water soluble phosphorus was

5.74 mg/l at the vegetable farm, 3.88 mg/l at the cattle ranch, and an

extremely low 0.04 mg/l at the sugarcane farm.
* Variability was high. Most phosphorus concentrations showed standard

deviations approaching or exceeding the mean.
* Seasonal variation of phosphorus concentrations was much higher for

vegetables than for sugarcane.
* Phosphorus applied as fertilizer per unit area was about 6.5 times as high

for vegetables as for sugarcane.
* A comparison of all studied sites and additional data obtained from

sampling of private drainage districts during the study indicated that the

cropland which had been under cultivation longer had less ability to retain

phosphorus.






15



Izuno et al. study

The work by Izuno et al. (1991) was a study carefully designed to allow

evaluation of phosphorus discharge from various crop and field conditions before and after implementation of BMPs. The baseline (before BMP) study was reported in the 1991 publication. Replicate fields of sugarcane, fallow, flooded fallow, and selected specific vegetables were evaluated over a 10-13 month period. Key conclusions follow.

The phosphorus concentrations in the sugarcane fields averaged 0.28 mg/l

total phosphorus, 0.17 mg/1l soluble phosphorus (61% of total) and 0.11

mg/l particulate phosphorus (39% of total). The cabbage fields averaged 0.56 mg/1l total phosphorus, 0.31 mg/I soluble phosphorus (55% of total)

and 0.25 mg/l particulate phosphorus (45% of total). The radish fields

averaged 0.25 mg/I total phosphorus, 0.20 mg/I soluble phosphorus (80% of total) and 0.05 mg/I particulate phosphorus (20% of total). It was noted

that radish has a lower fertilization requirement than cabbage.

The effects of hydraulics were evident in this study. The displacement

phenomenon was noted, wherein a portion of the initial contents of the

ditch and the surface runoff must pass through the pump before the effects

of the field groundwater on soluble phosphorus concentration are seen.

Main farm canal discharges averaged 0.16 mg/l total phosphorus, 0.08

mg/I soluble phosphorus (50% of total) and 0.08 mg/1 particulate

phosphorus (50% of total), which was consistently lower in phosphorus

concentration than the effluent from the fields, suggesting that either dilution or removal of phosphorus was taking place in the irrigation

conveyance systems.






16


The drained fallow fields, which were interspersed with the sugarcane

fields and drained under the same conditions, showed effluent phosphorus

concentrations averaging 0.43 mg/l total phosphorus, 0.28 mg/i soluble

phosphorus (65% of total) and 0.15 mg/l particulate phosphorus (35% of

total), which was significantly higher than the neighboring sugarcane

fields, suggesting a strong crop-effect on the availability of total

phosphorus for drainage. (Recall the very low pore-water phosphorus

found in the CH2M Hill sugarcane fields.)

The highest loads observed during the study came when flooded fallow

radish fields were drained, and immediately after fertilization of the

cabbage fields.

This study also showed high variability, with standard deviations of

phosphorus concentrations approaching the mean in most cases.


Canal Transport Studies

Several studies that contain data related to phosphorus transport in canals were performed in the Everglades region.


Southeast Florida canals water quality studies


Lutz (1977) studied the variation of water quality with time over an 18 month interval at various points within several Southeast Florida canals. One of the canals studied was the West Palm Beach Canal from its point of exit from the EAA to a point 16 km (10 miles) downstream. His experimental procedure entailed taking individual grab (non-flow weighted) samples at the various sample stations for each canal on the same day every two weeks. Lutz' data showed the same variability as was seen in the previously mentioned farm/field studies, that is, standard deviations approximately equal






17


to the mean. He concluded that spatial variation (upstream to downstream) was not statistically significant. Seasonal variation in the West Palm Beach Canal was also difficult to detect in his results because month to month variations were quite significant.

What is particularly interesting in Lutz' data, however, is what arises from visual inspection of the various time series. In about 15% of the days sampled there are appearances of large spikes at one individual sample point which were not reflected at the other sample points. This is true for soluble as well as total phosphorus. If the data plots are examined from a subjective phenomenological viewpoint one might possibly interpret the data as showing wave-like gradients in concentration down the canal on occasions of high concentration, however such an interpretation is very subjective, given the absence of close-interval time-series data. The main point from this study is that the data, taken at face value, indicate that there were times during the study when the canal was seeing short bursts of shock load with a more subjective indication of wave-like phosphorus transport under certain circumstances when phosphorus concentrations were high.


Water Conservation Area water quality correlations


Mattraw et al. (1987) attempted to develop statistical correlations of water quality variables to detect trends at 10 inflow or outflow locations around the periphery of Water Conservation Area 3A. Several of the structures were direct outflows from the EAA. They evaluated numerous model types relating to discharge and antecedent rainfall conditions and concluded that the only correlations of significance were orthophosphate and nitrate with 7 or 30 day antecedent rainfall conditions. No temporally significant trends were found over a five year period. The primary result of interest in the context of the present research is that for the one structure evaluated which corresponds to a major EAA discharge, S-8, the best model correlation had an R2 of only 0.36. This emphasizes the difficulty in correlation of phosphorus content using only hydrologic variables.






18


EAA Canal Sediment Studies

Several studies, which have been conducted in private and District canals within the EAA, have particular relevance to the area of particulate phosphorus transport. For the most part these activities have been private, non-peer-reviewed studies that have had limited report circulation. Nonetheless they have tended to form a basis for the perspective of the EAA grower community toward potential phosphorus reduction activities and require some significant attention at this point.


Anderson and Hutcheon Engineers study (1992)

Anderson et al. (1992), in a private study conducted for the Florida Sugar Cane League reported results of some sediment core sampling conducted in the EAA's four main and two connector canals, which are operated by the SFWMD. They also presented data from what was reported to be a sediment transport study conducted at coded private sites within the EAA.

The sediment core analyses showed a sediment phosphorus content in the range of 58-77 mg phosphorus/ kg dry sediment (58-77 ppm) for five of the six canals. Only the Miami Canal sediment, which averaged 252 mg P/kg sediment, exceeded 100 ppm.

The data reported in the sediment transport study include some velocity and total suspended solids (TSS) data in conjunction with the total phosphorus (TP), soluble or total dissolved phosphorus (TDP), and particulate or particulate phosphorus (PP) concentrations. Average concentrations for the total data set were TP = 0.153 mg/l, TDP = 0.042 mg/l (27% of total), PP = 0.111 (73% of total), and TSS = 187 mg/l.

On average the phosphorus content of the TSS was 594 mg P/kg TSS (ppm), but a more detailed examination of the data indicates that the sample set may have represented a bimodal distribution of samples, with one set having a phosphorus content of about 390 ppm and the other set having a phosphorus content of about 1700 ppm.






19


There was not sufficient detail of the channel configuration presented in the report to do a substantive evaluation of transport dynamics, but there appeared to be a step change in suspended solids concentration at a velocity somewhere between 0.06 m/sec and 0.18 m/sec, above 0.18 m/sec there was no clear response to increasing velocity.

Included in the report were data tables representing excerpts of data from a SFWMD sampling program which covered a nine year time span from 1983 to 1991. The data summary presented averages of TSS = 23 mg/l, ortho-P = 0.080 mg/l, total P = 0.161 mg/l, and particulate P = 0.081 mg/l. The implication from this summary was that the distribution of phosphorus in the SFWMD canals was split approximately 50%/50% between soluble and particulate forms during this period. This data set is discussed in more detail later in this section.


Andreis/U.S. Sugar Proposed sediment control BMPs (1993)

The Research and Development Department of U. S. Sugar Corporation, in an attempt to develop a particulate phosphorus control program for the EAA, informally proposed a number of sediment-control BMPs to the grower community and the SFWMD in 1993 (Andreis 1993). Although these practices were not proposed in formal publications, they have been given widespread distribution, and have received varying levels of acceptance within the grower community. They were discussed with H. J. Andreis, Senior Vice President of Research for U.S. Sugar, by this writer in 1994. At the time of discussion some field data were available on the effectiveness of the BMPs.


There were some sixteen BMPs recommended, which fall into five categories, as follows:

1. Surface soil erosion reduction,

2. Stabilization of ditch and canal banks and sidewalls,

3. Minimization of disturbance of in-canal sediments,






20


4. Installation of sediment trapping configurations, and

5. Canal cleaning programs.

Data available in 1994 at the time of discussion related primarily to comparative evaluations of test facilities that incorporated BMPs in the first three categories. Typical evaluation consisted of comparing some measure of particulate phosphorus (PP) concentration or load at a point immediately downstream of an installation before and after implementation of a specific BMP. In all cases discussed the implementation of a BMP resulted in significant reduction of the PP measure after implementation.

There were, however, no side-by-side controls run in any of the tests, and no data were available to relate the contribution of the problem being addressed to the overall phosphorus load from the farm of which the test plots were a part. Under these circumstances it must be judged that the BMPs seemed to move in the direction of goodness, but the levels of their contributions had not been quantified. Strong emphasis was placed on the last two categories of BMPs as "pipeline" and "end of pipeline" measures but no data were available at that time to evaluate the effectiveness of these measures because full scale tests were still in progress.


Sediment control demonstration project

A full-scale eighteen month duration test of canal dredging and sediment trapping procedures, funded by the EPD, was completed in May of 1995 and reported by Hutcheon Engineers (1995). At two separate locations canals were dredged for specific lengths to construct a canal segment, with increased depth relative to the immediate upstream and downstream sections, that would act as a sediment trap. Stated simply, the sediment trap operates by reducing the velocity of the carrying fluid which in turn reduces the carrying capacity of the fluid, causing some suspended sediment to precipitate. The upstream and downstream ends of the traps were equipped with sediment barriers constructed of rock






21


piled to the height of the original sediment to prevent sloughing of existing sediment into the study areas.

Effectiveness of the traps were measured several ways. Sediment accumulation pans were placed within the confines of the dredged sections at the start of the study and retrieved by a diver 18 months later at the end of the study. Bottom soundings were made in the dredged sections on a monthly schedule to estimate the increase in bottom elevation with time. Regular water sampling was conducted before and after dredging of the traps and upstream and downstream of the dredged sections to ascertain the degree of removal of phosphorus resulting from the presence of the traps.

The results were interesting. The sounding data indicated that the dredged bottom elevations increased at the rate of 1.4-1.8 ft/year (-0.43-0.55 m/yr). The sediment accumulation pans, which were 5 cm deep and had been placed on bare bottom at the start of the study, proved to be not especially useful because they were buried under 20-60 cm of overlying sediment at the time of retrieval. The water sample phosphorus analyses, however, gave no significant indication of phosphorus reduction arising from installation of the sediment traps. Pre-dredging and post-dredging farm effluent phosphorus contents were virtually identical at both locations. The average phosphorus analyses were 0.108 mg/I pre-dredging versus 0.107 mg/I post-dredging at one location and 0.086 mg/l both pre-dredging and post-dredging at the other. Upstream and downstream phosphorus measurements showed similar results. At one location upstream average phosphorus was

0.111 mg/l, downstream average phosphorus reduced to 0.096 mg/l. At the other location, upstream average phosphorus was 0.067 mg/l, downstream average phosphorus increased to 0.077 mg/1. In neither case was the difference statistically significant at the p=0.30 level.






22



South Florida Water Management District Canal Data Set

The data set from which Anderson et al. (1992) extracted their excerpts was

requested from the SFWMD for evaluation. What was provided (Mucinic 1994) was an extensive data set containing analytical data on 1,996 samples taken from 1984-1993 at various locations along the main EAA canals, along with logged flow data for the various pumping stations at the periphery of the EAA. The primary thrust of the sampling program was determination of total phosphorus and ortho-phosphorus, which were taken at each sample site roughly monthly. Samples that included analysis for total suspended solids were taken roughly quarterly.

Upon examination of this data set and previous interpretations of its analog for 1984-1991 it became evident that there were some opportunities for misinterpretation. First, the pumping records showed that there were numerous occasions when samples were taken that there was no flow in the sampled canals. Second, there was no independent analysis of particulate or particulate phosphorus. Prior evaluations of this data set had calculated particulate phosphorus as the difference between total phosphorus and soluble ortho-phosphorus. This may introduce non-trivial error if there is significant dissolved organic phosphorus present, which is not detected by the soluble orthophosphorus analytical technique.

The data set provided by the District did contain a small sub-set of samples, taken from several of the northern EAA pump stations which discharge into Lake Okeechobee, upon which both soluble ortho-phosphorus and total dissolved phosphorus analyses had been conducted. These data were used to develop an estimate of the proportions of soluble phosphorus which were organic (24.9%) and inorganic (75.1%). This distribution was assumed to hold for all samples and was used as a correction factor to estimate actual total dissolved phosphorous from soluble ortho-phosphorus data.






23


The data set was subjected to a screen for samples which met the criteria of:

1. Taken from one of the four major southern discharge stations

2. Taken during times of flow

3. Having analytical results for Total Suspended Solids, Total Phosphorus.

and Ortho-phosphorus.

This screen produced 55 samples from the entire data set which met all three

criteria. The above-mentioned soluble phosphorus correction was applied to the screened data set. The results showed an average Total Suspended Solids (TSS) of 14 mg/1, average total phosphorus of 0.140 mg/l, average soluble phosphorus of 0.099 mg/l (70% of total) and average particulate or particulate phosphorus of 0.042 mg/l (30% of total). Regression of particulate phosphorus on TSS indicated a phosphorus content of the suspended solids on the order of 2200 mg P/kg TSS (ppm).

The phosphorus content of 2200 mg/kg and the presumed upper mode of 1700 mg/kg of the Anderson and Hutcheon Engineers Study are both considerably higher than the phosphorus contents of the canal sediments reported in the Anderson and Hutcheon Engineers Study, 58-77 mg/kg for 5 of 6 canals and 252 mg/kg for the remaining canal. This major difference in phosphorus content poses an intriguing clue to the possible sourcing and transport mechanisms of particulate phosphorus in the EAA. The topic will be pursued in some depth in later chapters.














CHAPTER3
GENERAL LITERATURE REFERENCES RELEVANT TO PARTICULATE PHOSPHORUS TRANSPORT


Introduction


This chapter will focus on information in the open literature which may have particular influence on the direction of the research, which may provide insight for formulation of hypotheses presented later in this work, or which may be used to rationalize later conclusions. The chapter begins with a discussion of the sources of particulate phosphorus, continues with a discussion of diagenesis, or the fate of deposited particulate matter, and concludes with a brief review of the concepts of sediment transport, or the remobilization of deposited particulate matter.


Sources of Particulate Phosphorus

Particulate matter in ecosystems is typically classified as either allochthonous (of external origin) or autochthonous (of internal origin). Allochthonous materials which may be of importance in the EAA might include soil particles, livestock wastes, and particulate agricultural chemicals transported by wind or water erosion, stream bank eroded particles, leaf and ground cover litter transported by wind or water erosion, and deposited detritus from streamside trees and plants. Autochthonous materials would include aquatic macrophytes and aquatic microbes, both of which can fix soluble carbon and nutrients in their growth and reproduction processes, streambed parent inorganic particles, chemically precipitated inorganic materials, and invertebrate and vertebrate organisms. Given the minimal terrain slopes, sub-tropical climate, high insolation and



24






25


plentiful organic carbon supply it is intuitively expected that the regions channels are highly autochthonous.

Surficial agricultural residue will be a highly variable function of land use and management practices, but dairy farm soils in the area of interest may have typical phosphorus content of 1500-3000 mg/kg (Reddy et al., 1994). Tree leaves may contain 500-2500 mg P/kg dry weight, while woody litter may have a phosphorus content of 200300 mg P/kg (Reddy and DeBusk, 1987).

Macrophytes play multiple roles in the particulate phosphorus dynamics of subtropical streams. Because of their rapid turnover times they serve as both sinks during growth and sources during senescence for particulate and soluble phosphorus. For illustration, dense stands of water hyacinths may assimilate 100-300 mg P/square meterday by biological fixation (Reddy and DeBusk, 1987) and slough 5-15 mg P/square meter-day from normal detrital formation (DeBusk and Dierberg, 1989). Typical floating macrophytes may have tissue phosphorus concentrations of 1500-12,000 mg P/kg dry weight (Reddy and DeBusk, 1987) so, unless the macrophytes are removed from the system, the phosphorus stored in their tissue is available for release upon plant death and senescence.

By creating zones of hydraulic quiescence macrophytes may affect sediment deposition under normal flow conditions, but a portion of this sediment may be remobilized during times of increased hydraulic activity (Svendsen et al., 1995). In subtropical and tropical climates macrophytes, particularly floating macrophytes, may host substantial masses of loosely attached epiphytic microbial material, which effectively increases the in-stream concentration of organic phosphorus-containing material (Engle and Melack. 1990).

Microbial activity can play a major role in the particulate phosphorus dynamics of a water body. In eutrophic lakes, for example, phosphorus associated with deposited bacteria can equal or exceed that contributed by organic detritus (Gachter and Meyer,






26


1993). Microbial activity associated with litter decomposition can frequently cause an increase in the total phosphorus concentration as the higher-phosphorus-concentration microbes utilize the carbon source of the lower-phosphorus-concentration substrate to which they are attached (Elwood et al., 1988). Cyanobacteria have been shown to. on occasion, exhibit a reverse sedimentation effect where they detach from the sediment and migrate into the water column. The conditions under which they do this also favor a high bacterial phosphorus content so they may become an important source of suspended particulate phosphorus under conditions of high water temperature and favorable C/N ratios. Petersson et al. (1993) showed that particulate phosphorus flux into the water column from this source could be in excess of 2.5 mg P/m2-day.

Phosphorus content is high for planktonic materials, for example Behrendt (1990) reported values averaging 5100 mg P/kg biomass for diatoms, 7700 mg/kg for blue-green algae, and 13,700 mg/kg for zooplankton. The high phosphorus content can be a source of soluble phosphorus. Montigny and Prairie (1993) have shown that cell lysis of bacteria can produce high levels of soluble phosphorus even in the presence of iron which would be expected to precipitate the phosphorus.

Macrophytes and microbes typically dominate the mass of non-soluble material present in streams, but macroscopic organisms, particularly the invertebrates, can play a major role in modifying the physical character of the particulate matter resident in the stream. Webster (1983) describes the complex process where invertebrate shredders and collector-gatherers convert coarse organic matter to more transportable and accessible fine particulate organic matter and also create a bi-directional flux of the fine matter between the suspended and benthic compartments.

Many of the phosphorus content values noted above for autochthonous materials may be considerably higher than those typical of average EAA basin soils or stream sediments (See Chapter 5). This illustrates the important role of in-stream processes may play in modifying the physical and chemical nature of the basin phosphorus cycle. The






27


actual contributions, however, will be governed by local system dynamic parameters including carbon and nutrient availability, light source, intensity and extinction, competitive population dynamics and size of standing crops, air and water temperature. and hydraulic fluctuations.


Early Diagenesis

The processes of early diagenesis, or the first steps of conversion of sediment to inert material, are important in that the process pathways and kinetics can have a significant effect on the nutrient content of the resuspended sediments and on the amount of sediment-contained nutrients released back to the water column. The organic sediment constituents which have the highest phosphorus content will also tend to be the most bioavailable (See preceding discussion of sources). A brief discussion of the expectations of the fate of this material follows.

Lucotte (1993) has estimated that up to 50% of the P initially buried in estuarine sediments of the St. Lawrence River are released back to the water column because of bioturbation.

Montigny and Prairie (1993) have shown that bacteria in sediments will lyse

rapidly in the anaerobic regime, releasing high levels of soluble phosphorus. Contrary to conventional theory, however, the soluble phosphorus did not combine with ferric iron and precipitate when it diffused into the oxic zone. The reason for this was the chelation of the ferric iron with simultaneously released organic acids, which removed the iron from effective interaction with soluble phosphorus.

Menendez et al. (1993) found that the initial decomposition of macrophyte

detritus and release of phosphorus proceeded at an extremely rapid rate, about 20%/day, in the anaerobic zone of lagoon sediments.






28


Keizer et al. (1993) found that in peaty sediments of the Netherlands the

conversion of calcium carbonate to calcium phosphate complexes was inhibited to the point of insignificance.

These studies were not necessarily under conditions identical to those found in the EAA watersheds, but they do tend to indicate that, unless significant quantities of sorptive clays are present or calcium-phosphorus complexation is favored, high levels of sediment phosphorus recycle to the water column would be expected. This is a good place to note that Svendsen et al.(1995) observed that the sediments in their studies showed a long-term net retention of phosphorus of only one-half to one-third of their short-term gross retention rates.


Erosion and Transport of Organic Material



General erosion and transport

To state the obvious, transport of particulate phosphorus within a system is

governed by the overall transport of particulate matter (PM or seston) within that system. As noted above, sources of PM in a waterbody may include soil and ground cover erosion, deposition from external vegetation, and internal generation and transformation. Once deposited or generated the PM is constantly subjected to buoyancy, gravitational, kinetic, and chemical forces, the integrated interaction of which determine the fate of the particle.

Factors which affect the exchange of particles between their streambed or vegetation-attached locations and the water column have a large impact on particle transport by streams. Initiation of motion occurs when water velocity is sufficient to create local shear stress in excess of some critical value for detachment or incipient motion. Once moving, some particles may remain in contact with, or in close proximity






29


to, the stream bed. This material is conventionally referred to as bedload. Smaller or less dense particles may be carried into the water column when turbulent forces are considerably in excess of gravitational forces. This material represents the suspended load. When turbulent suspension forces fall below their critical threshold values suspended materials are re-deposited. Similarly when bed shear forces fall below their critical threshold values bedload transport ceases. A variety of physical factors affect particle transport including particle characteristics such as particle size, shape, density, fall velocity, and electrical charge, and stream characteristics such as width, depth, flow velocity, slope, roughness, water temperature, macrophyte population type, density and location, and flow-stage relationships (Webster et al., 1987).

The literature on erosion of organic soils is exceedingly sparse. Studies of soil erosion tend to focus on upland peat bogs or moors in north temperate or arctic climates. Representative of these is the work of Labadz et al. (1991), who studied sediment yield and delivery from blanket peat in Great Britain. They concluded that erosion from upland peatlands is highly variable spatially and temporally with a strong stochastic component attributable to the constant variation of surface morphology. In a similar vein, but in the context of different terrain, Benda (1990) concluded that organic debris flows were an important factor in determining the channel morphology of streams in the Pacific Northwest. Benda hypothesized that the stochastic nature of sediment supply from debris flows promotes cycling between channel aggradation and channel degradation, accentuating temporal and spatial variability of channel morphology. These and similar studies of upland erosion are of interest in that they may provide some qualitative framework to understand the variability of particulate organic transport, but they do not necessarily apply to the hydrology, hydrography, or surface morphology of the South Florida region.

Research with more relevance to the South Florida hydrologic regime is that which has included the study of production and movement of organic sediments in the






30


riverine, lacustrine, or estuarine environment. Cushing et al. (1993) studied the transport of carbon-14 labeled fine particulate organic matter (FPOM) of particle size less than 100 micrometers to estimate the transport distances and residence times in the riverine environment. They measured labeled suspended FPOM concentrations in the river as a function of time and distance, and labeled FPOM concentration in the river sediments as a function of time. Several interesting results were reported. The mean particle deposition velocity, calculated from concentration gradients, was 0.43 m/hr, about one order of magnitude less than the sedimentation velocity of particles from the same source measured in the lab under quiescent conditions. Calculated mean time to deposition under a mean velocity of 0.27 m/sec and average reach depth of 0.33 m was 51 minutes, with an average of 83% of the injected FPOM being deposited in the 1 km reach study area. However after 24 hours bottom sediment within the reach contained only 1% of the originally deposited material. The authors concluded that particles in surficial sediments exchange rapidly with the water column and migrate episodically throughout a riverine system, leading to strong longitudinal connection of sediment distribution. Unstated, but evident from the results, are the potential errors involved in attempting to apply laboratory sedimentation test results to field conditions.

Kronvang (1992), using a synoptic water sampling design augmented with

increased sampling frequency during storm events, investigated the export of particulate matter, with specific emphasis on phosphorus, from two Danish agricultural basins. He found significant phosphorus content enrichment in the transported sediments compared to typical soils in the watershed (sediment phosphorus concentrations of 7 to 14 times those found in the soils). The phosphorus content of the particulate organic matter was found to stay relatively constant at about 1%, regardless of the total particulate concentration, while the inorganic particulate phosphorus content decreased as particulate concentration increased, but approached the phosphorus content of the local soil only at extremely high particulate matter concentrations. The results of this study emphasize the






31


difference in sources of particulate inorganic and organic phosphorus and the apparent selectivity of transport which tends to favor particles which happen to have higher phosphorus content. The cumulative load chemographs for the various forms of phosphorus are illustrative of typical load distributions observed elsewhere, including the Everglades area. They show 50% of the annual particulate phosphorus load being transported during less than 5% of the total observation time, which may be interpreted as the bulk of the particulate phosphorus transport occurring during major flow events.

Findlay et al. (1991) conducted automated synoptic water sampling and stream gauging over a 150 km reach of the lower Hudson river for a three year period. Vertical as well as horizontal sample profiles were obtained. Profile analysis of particulate loadings led to the conclusion that during low to medium flows particle resuspension was as important as tributary contribution in determining riversediment loading. They found significant contributions from both autochthonous particulate organic matter and resuspended detrital material during the ice-free seasons and concluded that transport of particulate organic matter was controlled to a significant extent by processes occurring within the river and were not simply related to loadings from outside.

Godshalk and Wetzel (1984) used piston coring devices to sample sediment

transects of a small hardwater lake in Michigan. They segregated sediment particle size fractions by wet screening and then determined the various molecular weight fractions of humic and fulvic acids, as well as conducting total organic carbon, fluorescence, and UV absorbence analyses. Humic and fulvic acid proportions and the carbon content of each fraction were used to estimate the relative age of each sediment sample (Humic-older, Fulvic-younger, increasing carbon content indicating more refractory material). The results showed a succession of organic particulate matter sources in the lacustrine environment. The indigenous fine particulate organic matter (FPOM) was transported first, but there was regular production of FPOM from coarse particulate organic matter (CPOM) as time progressed. Thus the reservoir of FPOM was replenished on an irregular






32


periodic basis determined by transport events which affected resuspension of the nascent FPOM. The authors postulated a succession scenario wherein macrophyte colonization causes rapid detrital accrual near-shore. The detritus is reduced in size over time by microbial decay and the smaller particles are transported by periodic hydraulic excursions. Lability of smaller transportable particles depends on origin, for example planktonic versus detrital, and age. Chemical and biochemical interaction of the small particles with the environment is a complex function of space and time (depth, light availability, redox potential, temperature. etc.) and the history of the particle.

Calvo et al. (1991) used the Particle Entrainment Simulator (PES) to study the relative compositions of parent sediment and resuspended material in multiple locations in a shallow (I m average depth) marinelagoo4. Undisturbed core samples were taken directly in the PES cylinders by divers and then immediately subjected to typical shear stresses in the range of 6-9 dynes/cm2 in the PES onboard ship. Suspended particulate material was sampled and analyzed, along with parent sediment, for Total Organic Carbon, Total Kjeldahl Nitrogen. Exchangeable Ammonia Nitrogen, and Total Organic Nitrogen. The authors found that the carbon/nitrogen ratio was always eor4h suspended material than for the parent sediment, indicating a lower degree of mineralization for the suspended material, and that the C/N ratio increased as the suspending shear stress was increased. This latter finding suggests that the more readily transportable the material was, the younger and less decomposed it was. In addition, the largest deviations between suspended material and parent sediment composition occurred in the regions which were identified as having the largest standing crop of phytoplankton. No phosphorus analyses were done, but it follows from the hypothesis regarding age and degree of decomposition versus transportability that the same results would have been obtained for phosphorus. This study reinforces the concept of the selectivity of transport with respect to composition, expressed by the authors as selective resuspension of freshly deposited material.






33


Kemp et al. (1984) carried out a comprehensive study of sediment mass flux and chemical composition as influenced by suhbsracph ~ in a Chesapeake Bay tributary, using a variety of sampling and analytical techniques which included harvest of standing crops, core sampling, water sampling, respirometry studies, analyses for chlorophyll-a, stable carbon isotopes, total carbon, total nitrogen, nutrients, suspended material and wet sieve particle size analysis. They also included work from previous studies which included seasonal budgets for organic carbon, sediments, and nitrogen. They found that the vascular aquatic vegetation played a strong source-sink role in trapping POM and retarding its movement during the growing season but contributed about one-third of the total annual organic carbon budget during times of senescence and death. Phytoplankton content of trapped sediment ranged from 10%-40% as estimated from chlorophyll-a and stable carbon isotope analysis. Their annual budgets were particularly interesting. They showed that the sediment sink load within the macrophyte beds, expressed in kg/yr., was more than twice the total sediment source load on the study area from river input and shore erosion, indicating that processes internal to the study area contributed as greatly to sediment load as external inputs.

The nature of transportable organic sediment as a collection of agglomerates or aggregates of organic and inorganic matter has been addressed by several authors. Kranck (1984), studying transport in estuaries, found that the suspended particulate flocculated matter in three separate estuaries had an organic content in the range of 6575% by volume. She theorized that this organic content represented an optimum composition for floe formation under the estuarine conditions studied. Assuming specific gravities of 1.1 for organic matter and 2.7 for minerals would make Kranck's values 4555% by weight. Mirbagheri et al. (1988) studied a California agricultural runoff and irrigation watershed using floating single-stage and multi-depth autosamplers, Van Dorn plankton-sampling bottles, and Ponar bottom sediment samplers. They evaluated total suspended solids, suspended organic matter, suspended algae, chlorophyll-a and






34


suspended matter biochemical oxygen demand (BOD). They found that the suspended matter in the channelized riverine environment averaged about 40% organic matter by weight. The organic matter itself consisted of about 25% viable planktonic-type material (as estimated by chlorophyll-a analysis) and 75% nonliving biomass. This nonliving biomass was further categorized by BOD analysis as 35% (of total suspended organic matter) biodegradable and 40% "refractory" organic matter. Bokuniewicz and Arnold (1984), who carried out water sampling at 16 stations in freshwater reaches of the Lower Hudson River, found the average organic content of the tidally influenced freshwater suspended sediments to be 22% by weight.

Tipping et al. (1993) sampled suspended material in riverine environments and determined particle aggregate sizes by microscopy. They also collected freshly sedimented material using sediment traps placed in various hydraulic regimes within the reaches under study and determined particle size distributions using particle counters. Subsamples were analyzed for total mass, carbon, and nitrogen. They found that the average particle density decreased significantly as average aggregate diameter increased, with an increasing concentration of organic matter as aggregate size increased. They postulated that little, if any, agglomeration takes place in-stream, rather that the agglomerated particles either enter the stream as a result of original erosion of fields and banks, or they form on the sediment surface in relatively quiescent (dead) zones and are resuspended by periodic hydraulic excursions. They note that dead zones can play an important contributory role, causing either concentration spiking or tailing when they are disturbed. The magnitude of the disturbance and the exchange rate with the main stream govern the type of contribution. They also point out that dead zones with normal turnover times of days can be important contributors of phytoplankton.

The selective transport of constituents resulting from variations in particle size, density, origin, and composition gives rise to the phenomenon known as "focusing" which is the spatially inhomogeneous distribution of contaminants in sediments (Eadie






35


and Robbins, 1987). Degree of focusing, as would be expected intuitively, is a strong function of biology and hydrodynamics. Bloesch and Uehlinger (1986), using a lakewide distribution of sediment traps. found it to be relatively unimportant in a eutrophic lake with high productivity and modest turnover velocities. Murchie (1985) used 210-Pb dating to study the geologic history of a freshwater bay in Minnesota. He found that focusing decreased as basinwide productivity increased, but also that high-organiccontent sediment was more intensely focused than heavier siliceous or calcareous sediment. Kronvang and Christiansen (1986) used a combination of traps, cores, and water samples to develop a spatially distributed sediment budget for a hydrodynamically active estuary in Denmark. They found that focusing was strong in the upper estuary zone and recommended specific dredging locations and seasonal times which would allow for an optimization of the dredging effort to recovery ratio. This approach may have particular interest in regions such as South Florida where hydrodynamics are, or can be, partially controlled.

Specific quantitative erosion data have been generated in a few cases for the

organic fraction of sewage and stormwater flows. Kleijwegt et al. (1990) studied erosion of cohesive synthetic sewer solids in a laboratory flume. They determined that the upper limit of the critical shear stress for initiation of erosion of cohesive sewer sediments appears to be in the range of 5-7 Pascals. Ashley and Crabtree (1993) categorized sewer sediments into five classes: Class A coarse granular material, Class B agglutinated Class A deposits, Class C mobile fines, Class D organic biofilms, and Class E deposits found in tanks. In a related study Ashley et al. (1993) noted that definition of bedload was very difficult with Class C and D sediments because bedload for these materials was in the form of a dense cloud of sediment close to the surface of the transporting channel, as opposed to the traditional riverine definition of bedload as a saltating layer of individual particles. The thickness and density of the cloud is affected by hydrodynamic conditions as is the interchange of suspended particles with the bedload






36


cloud. Class C (mobile fine) solids were noted to be weakly resistant to erosion. Class D (biofilm) solids were reported to be poorly studied and ill-defined. More research was recommended in the area of biofilm solids transport because of their high organic matter content. The authors reported that flume studies of freshly deposited organic sewer solids exhibited a critical shear stress of about 1.8 Pa, and that 75% of the eroded solids had a particle size of less than 100 micrometers.

The only literature source found which deals directly with quantitative erosion

characteristics of natural organic sediments is the work of Hwang (1989). Hwang studied the erosion characteristics of Lake Okeechobee sediments which had organic fractions in the range of 40-45%. The experimental procedure included placing sediments harvested from Lake Okeechobee in an annular flume as a thick slurry, covering the placed bed with lake water, allowing the mixture to consolidate for several days, and then measuring the concentration in the overlying water as various shear stresses were applied to the bed. He developed a two-component model incorporating a "surface-fluff"' component, which had a critical shear stress for initiation of erosion of zero Pa (immediate erosion upon any disturbance), with a bed-surface-erosion component that had a critical shear stress of about 0.45 Pa (much lower than that reported for sewer sediment). Erosion rates as a function of shear stress were determined for both components. He further correlated the erosion rates with sediment bulk density and developed a map of regions of bulk density and shear stress where various types of erosion would occur. The experimental method utilized may have underestimated the impact of light flocculated organic matter which may have been intermixed with the base sediment at harvest, nevertheless this study reports the lowest critical shear stress found for organic sediments in this literature survey.






37


Hydrologic approach to seston transport

The application of traditional non-cohesive sediment transport theory to transport of seston has been shown to be inappropriate for several reasons (Webster et al., 1987). One major assumption in most non-cohesive sediment transport models is that transported particles are similar to particles in the streambed (for example, Einstein 1950). This is often not the case in stream transport of seston, where the transported matter may consist of small and/or low density matter with a high organic content while the streambed may be primarily sand, rock, and large organic particles. A second major assumption is that of unlimited supply of transportable solids (for example, Bagnold 1966). This is often not the case with seston transport (for example, Allen 1977) where depletion of supply is frequently evident, for example over sequential storm events of similar or increasing magnitude (for example, Svendsen and Kronvang, 1993).

These violations of traditional sediment transport assumptions may be

circumvented by using models derived from cohesive (Fine Grained) sediment transport theory which allow for variation of bed composition and recognize supply limitation. The theory of cohesive sediment transport formed the basis for much of the work planned in this project and is discussed in the accompanying section entitled "A Brief Discussion of Sediment Transport Theory".

An alternative approach to the quantification of seston transport has been to

develop phenomenological relationships between basin hydrologic parameters and seston transport. Typical are attempts to correlate particulate organic matter loads to stream discharge or stream order (Hawkes 1975). These approaches have met with limited success, mainly because they are implicitly based on the noncohesive sediment transport model assumptions. A more sophisticated approach was proposed by Sedell et al. (1978) where stream power, rather than discharge, was used to correlate seston discharge. This approach reduced data scatter somewhat, but suffered from the limitations that






38


correlations were very stream-type specific, and that for a given stream a hysteretic pattern was observed over a time series when transported load was correlated with stream power. This hysteresis was explained by Webster et al. (1987) as being due to the limitation of supply of seston, which reduces the potential transportable load as time progresses in an event or series of closely related events. The hysteretic effect may also arise from seasonal factors which affect the availability and transportability of seston even in the absence of significant hydraulic excursions. They proposed that the stream power correlations be modified into separate correlations for the rising and falling limbs of storm hydrographs, and that inter-event seston production and storage functions be developed for use in conjunction with the power correlations. This approach results in a higher level of predictability, but it also requires a fairly intimate knowledge of the biological productivity of a system in order to be useful in a predictive capacity.

Several recent studies in Denmark on the Brabrand Lake (agricultural) watershed (Kronvang 1992) and the Gjern A (lowland) basin (Svendsen and Kronvang, 1993, and Svendsen et al., 1995) have provided good illustrations of the complexity of sestonrelated phosphorus transport in the context of seasonal time scale basin hydrology. These studies showed the following results:

Significant differences in Particulate P/Total P ratios in adjacent

subcatchments

Decreasing P-content of inorganic particulates as flow increased
Relatively constant P-content of organic particulates as flow increased

Reduction in the P-load vs discharge correlation with sequential storm events

Seasonal changes in the P-load vs discharge correlation

High levels of seston removal by macrophyte beds during normal flow but

high levels of release during storm flow, with this effect accentuated when

storms followed extended low flow periods.

Episodic increases in particulate-P load arising from macrophyte cutting






39


The basin as a whole acting as a phosphorus sink during a dry year. but

becoming a net source the following year, which was wetter than average

Mass balances showing an increasing bias toward underestimation of

phosphorus export as frequency of sampling decreased


The authors rightly noted that current understanding of non-point P sources and the routing of P compounds within watersheds is very poor.


A Brief Discussion of Sediment Transport Theory



Non-Cohesive sediment transport

Traditional (non-cohesive) sediment transport places mobilized sediment in various categories, which have been noted earlier in this chapter. Grains making up a substantial part of the movable bed of a stream are called bed material. Bed material moving within a few grain diameters of the bed is known as bed load. Material in suspension which is not present in any quantity in the bed (usually very fine sediment) is known as wash load. Bed material swept up into the main flow stream by turbulence becomes part of what is referred to as the suspended load. The suspended load is the combined mass of this fugitive bed material and the wash load. The relative proportions of the bed material that move as bed load and suspended load depend on the characteristics of the bed material, such as size and density, and the flow conditions (Middleton and Southard, 1984).

The analytical approach taken in non-cohesive sediment dynamics treats bed load and suspended load as two separate entities. subject to two different sets of physical forces. Typical equations for bed load horizontal flux incorporate the difference between a critical shear stress for incipient motion and the actual bed shear stress as a driving






40


force and particle size and mass as a resistance. There are several such equations (Raudkivi 1976). Shield's equation is presented here for the purpose of example.


g, = 10 ( -)qS 3.1 1d


where go = weight rate of bed transport per unit width, kg/m

q = volume rate of water flow per unit width, m3/im

S = bed slope, m/m

d = particle diameter, m

T = bed shear stress, Newton/m2

To = critical shear stress for motion, Newton/m2

y = specific mass of water, kg/m3

Ys = specific mass of sediment, kg/m3

Suspended load transport may be treated within one or more of three conceptual frameworks, diffusional, energy, or statistical.

Diffusional approaches incorporate a turbulent eddy diffusivity term into the equations of motion and solve for the steady state concentration profile of mass in the vertical direction. An equation often proposed as a basis for analysis is one presented by Rouse (1937), that describes the concentration profile above a datum plane at which point it is assumed the concentration is known:



CF(,-y a 32
Ca y0- a) j



with z=
where C concentration at KUelevation

where C = concentration at elevation y, kg/ms






41


Ca = concentration at datum level a, kg/m3

y = elevation, m

yo = water surface elevation, m

a = datum level elevation, m

w = particle settling velocity, m/sec

13 = constant of proportionality
K = von Karman's constant

U* = shear velocity, m/sec.

Application of this equation gives sigmoidal concentration curves, the shape of which depends on the value of z.

Energy approaches analyze the suspended solids from the standpoint of an energy balance, where the momentum transfer to the suspended solids by the turbulent fluid must equal the excess weight of the solids in motion. Velikanov's gravitational theory (Velikanov 1954), for example, uses this approach to arrive at an equation that describes the concentration profile as follows:



C = w cy t, dn 3.3
e 3.3 C. U* yS (1 )1In



where 4


Tla = reference level, m k,
30y0

ks = grain roughness

Application of this approach gives concentration curves similar to those produced by the diffusion approach.






42


The statistical approach may be applied in several ways. First the diffusional and energy approaches, as presented above, apply to only one specific monodisperse class of particles. One statistical approach is to apply a statistical distribution to particle classes and then apply the equations to each class. A more detailed approach is to write the equations of motion as stochastic equations and then either solve them using simplifying assumptions and asymptotic solution methods (Hinze 1957, Hino 1963) or use them to run Monte Carlo computer simulations (Yalin 1972). Specifics will not be presented here.

The total suspended load is obtained by integrating the equation


q,= Cudy 3.4

where qs = volume rate of suspended solids discharge per unit width, m /m-sec

u = velocity at elevation y, m3/sec


The details of the integration of this equation depend entirely on the choices of concentration distribution, velocity distribution and reference datum plane. There are many options available (Raudkivi 1976, Vanoni 1975), some of which derive from basic principals and some of which contain empirically derived fitting parameters. The latter tend to give better results than the former in specific situations but cannot necessarily be extrapolated beyond their range of calibration.

There are several drawbacks in the application of non-cohesive sediment transport theory to the study of organic sediment transport. First most of the useful formulations assume steady state with respect to sediment suspension and deposition. Development of transient formulations would require reversion to the equations of motion and mass transport, where the dynamics of organic sediments have virtually no investigational






43


history. Second, and more important, the traditional non-cohesive approach makes no assumptions about supply limitation of transportable material, which is a major drawback in the evaluation and simulation of organic sediment transport.


Cohesive sediment transport

Cohesive sediments are those that exhibit particle-particle interactive forces

strong enough to cause them to not act as a collection of elementary particles. They are typically composed of fine-grained particles which, under the appropriate circumstances, assume an electrochemical surface charge that contributes to their interactive nature. The particle size, chemical charge, and applied shear can operate in close packed environments to impart theological properties to the sediments, which affects their erosion characteristics, and in dispersed environments to cause flocculation and disaggregation of suspended particles, which affect their sedimentation characteristics.

Partheniades (1977), using concepts developed and presented in earlier papers, proposed a unified theory that treated bed load, suspended load and wash load all as special cases of a general erosion-transport-deposition model. His conceptual framework includes a minimum critical shear stress for inception of erosion, a maximum critical shear stress for termination of deposition, addition of inter-particle forces to the force balance for detachment of a particle from the bed, incorporation of probabilities that a particle may suspend or deposit which are related to shear levels, and the concept of supply limitation and carrying capacity saturation for any given class of sediment. One of his important conclusions is that there are classes of sediment that can behave either as wash load or as bed load, depending on channel discharge characteristics.

Organic sediments do not generally possess electrochemical charges strong

enough to affect their compacted strength but their complex physical configurations may impart some degree of particle-particle interaction in the settled phase. There are






44


electrochemical charges present that may be sufficient to affect some degree of flocculation in the suspended phase. Given their source, there is definitely the potential for supply limitation, particularly in larger channels and flows. Their specific gravities may be considerably less than those of the inorganic materials typically found in sediments so the buoyancy forces acting on them may have a much greater impact than on inorganic sediments. For these reasons it may be appropriate to approach organic sediments as hybrids between cohesive and non-cohesive sediments and attempt to adopt appropriate techniques from the study of each of these two types of sediment.

Several reviews (Kranck 1984a, Mehta 1984, Mehta 1988, Mehta et al., 1981, Mehta et al., 1989, Parchure and Mehta, 1985) cover the practical aspects of cohesive sediment transport well. The following summary draws from these reviews.

Erosion Modeling of the instantaneous erosion rates of cohesive sediments can be approached from the perspective that the critical shear stress for erosion corresponds to the shear strength of the eroding sediment (Mehta et al., 1981). The key to developing analytical expressions for cohesive sediment erosion is the determination of the bed shear strength and the form of the rate of erosion as a function of the shear stress which is applied to the sediment bed in excess of the bed shear strength.

Beds that are artificially placed in erosion simulation devices for study (placed beds) usually have constant properties in the vertical direction, which makes their study relatively simple. The instantaneous erosion rates for these beds may be expressed in the normalized form


E = 3.5


where e = Erosion rate, kg/sec-m2

EM = Erosion rate constant, kg/sec-m2 T = Shear stress applied to bed, n/m2






45


Is = Bed shear strength, n/m2

although formulations have been proposed that make the erosion rate exponential with respect to the excess shear stress. Note that for a specific constant bed shear strength. EM and ts might be combined in Equation 3.5.

Beds that are deposited, under either flow or non-flow conditions, will normally have properties which vary with vertical position within the sediment. Parchure and Mehta (1985) have presented an analysis that correlates the depthwise variation of shear strength of specific sediments with the vertical increase in sediment density. They proposed an equation of the form


E = exp ca( -'t(z)) 3.6

where Ef = floc erosion rate, kg/sec-m2

c = rate coefficient, m/Nos

z = elevation of sediment surface above datum plane, m

The floc erosion rate represents a non-zero erosion rate when the bed shear stress equals the bed shear strength, a recognition that some erosion occurs at this point because of random turbulent fluctuations at the bed surface. Given the (known) relationships between sediment depth, density, and shear strength, and given a known applied shear stress pattern, Equation 3.6 may, in theory, be used to predict sediment suspension throughout the course of an erosion event.

Aalomeration and De-agglomeration Particle-particle collisions, caused by

Brownian motion, velocity gradients within the suspending fluid, and differential rates of settling among particles of various dimensions and densities may give rise to flocculation or agglomeration of suspended matter, increasing the dimensions of the particlecollection structure, which generally tends to increase the sedimentation rate of the particle-collection. Turbulence may increase the opportunity for agglomeration but it also






46


provides shear forces that act on the particle-collection, tending to break down the structure. The floce size distribution depends in a complex way on floce strength and turbulence structure but a qualitative interpretation is that low levels of turbulence tend to promote flocculation and produce light, diffuse floc structures, while high levels of turbulence tend to promote de-agglomeration and produce dense, compact floc structures.

Sedimentation Sedimentation of cohesive particles falls in one of three general categories. At concentrations of less than about 300 mg/l particles act as individuals, are not influenced by their neighbors, and maintain a constant settling velocity independent of particle concentration (Krone 1962). This region is known as the free-settling zone. At concentrations above the free settling zone, particle-particle interaction and differential settling cause the settling velocity to increase with increasing concentration as there become more and more opportunities for agglomeration as concentration increases. This region is known as the flocculation settling zone. Beyond a certain concentration, typically 5000-10,000 mg/l (Mehta et al., 1989) the flocculant structure of the sediment becomes so extensive that bridging begins to occur and the floc structure becomes partially self-supporting. At this point and beyond settling velocity tends to decrease with increasing concentration. This region is known as the hindered settling zone.

Deposition The concept of deposition is treated differently in cohesive sediment dynamics than in non-cohesive sediment dynamics. For a class of potentially cohesive particles there may exist a critical bed shear stress beyond which no deposition will take place at any reasonable concentrations. At bed shear stresses less than critical a fraction of the suspended material may deposit. The fraction of deposition is related to the departure of bed shear stress below critical. The simplest expression of this theory for a monodisperse particle collection is the Krone formulation (Krone 1962)



F, [l= wC for, < and F, = 0 for tb "Td 3.7 L td3.






47


where FD = deposition flux, kg/m2

Tb = bed shear stress, N/m2

rd = critical depositional shear stress, N/m2
ws = settling velocity of sediment, m/sec.

Cb = concentration of suspended sediment, kg/m3

This expression becomes more complex as additional classes of particles are added and as the effects of agglomeration and de-agglomeration are considered. One practical aspect of these relationships is that erosion and deposition of cohesive sediments may be studied independently of one another if the critical shear stress for deposition is less than the critical shear stress for erosion.

Consolidation Here there are similarities between the behavior of cohesive

sediments and the presumed behavior of organic sediments. Consolidation takes place because the forces exerted by the self weight of the settled material exceeds the strength of the interlocking sediment structure and its contained water. In cohesive sediments the interlocking sediment structure contains agglomerated structures as well as individual particles. The strength of these agglomerated structures is often the controlling element in the rate of consolidation. Organic sediments, with their potentially complex physical structures, might be expected to exhibit characteristics similar to cohesive sediments in the consolidation process, such as a rapid approach to an asymptotic density profile as supporting structural networks are broken down during consolidation.






48



Particle Entrainment Simulator

Quantitative prediction of the transport of cohesive sediments requires the

estimation of net sediment flux at the sediment-water interface as a function of sediment properties and the applied shear stress. The same holds true for sediments that behave cohesively due to physical particle-particle interaction.

Numerous studies have been conducted in laboratory flumes to correlate erosion rates with sediment properties and shear stress. A condensed listing includes Parthenaides (1965), Mehta and Parthenaides (1975), Fukuda and Lick (1980), Lee et al. (1981), Lick (1982), Mehta et al. (1982), Parchure and Mehta (1985), Ockenden (1993), and Verbeek et al. (1993)

Difficulty often arises because the state of sediments in the laboratory

environment differs significantly from that in the field. In addition, direct measurement of spatial variation of erosion characteristics in the field cannot be managed in any efficient way because the laboratory apparatus' are large and cumbersome and require relatively large quantities of sediment for a single test. Attempts have been made to circumvent this limitation by developing, for example, in-situ flumes (Young 1977), but such devices suffer from their own logistical and reliability problems (Lavelle and Davis, 1987).

An alternative to in-situ flumes has been the development of a small portable device that applies an oscillating-periodic but zero-mean velocity to the surface of a sample of bottom sediment. Appropriate sampling techniques can allow the sample to approximate an undisturbed sample of original sediment. A device with this capability was first used by Rouse (1938) and later adapted to the study of fine lacustrine sediment by Tsai and Lick (1986).






49


The device is of simple construction. consisting of a cylindrical chamber inside of which a horizontal grid oscillates vertically. The sediment to be studied is placed in the bottom of the cylinder and is overlain with water, preferably of composition similar to that of the original sediment environment. The grid oscillates in the water and creates turbulence that penetrates to the sediment-water interface, causing sediment resuspension. The turbulence, and thus the extent of resuspension, is a function of the grid oscillation frequency (Tsai and Lick, 1986)

The structure of the turbulence generated in the PES is different from that generated in a typical field shear flow. The mean velocity vector in the PES changes direction many times a second, whereas the mean shear velocity vector in field channel or tidal flows is relatively steady over short time periods. Never the less the statistical properties of the turbulence generated in the PES can be assumed to be similar to those in field shear flow (Lavelle and Davis, 1987). The statistical similarity assumption allows the PES to be calibrated to standard parameters.

The calibration procedure, developed by Tsai and Lick (1986), involves

determining the erosion extent of a specific sediment in the PES for various oscillation frequencies. The erosion extent of an identical sediment, prepared under similar conditions, is also determined in a shallow flume, for which the shear stress correlations are known with some degree of accuracy. The flume is also operated over a range of shear stresses. Under the assumption of statistically similar turbulence yielding similar erosion extent, it is possible to match a specific erosion extent at a known oscillation frequency in the PES with a corresponding equivalent erosion extent in the flume at a known shear stress. Ultimately this empirical matching procedure yields a correlation of PES oscillation frequency with applied shear stress in a uniform flow field. This is an important translation, because all useful erosion-hydrodynamic models require an expression of shear stress or equivalent shear velocity at the sediment-water interface for calculation of erosion rates.






50


One caveat must be noted. Recent work (Chapter 6) has shown that the PES

calibration procedure can be sensitive to large variations in sediment type. That is, PESflume calibrations developed for sedimented kaolinite-type clay beds, compacted bentonite-type clay beds, and sedimented organic-type beds were internally typeconsistent but did not compare well inter-type on a shear stress basis. This indicates the advisability of using calibration data for the PES from a sediment of similar type.

The portability of the PES has allowed its use on shipboard (Tsai and Lick, 1986, Lavelle and Davis, 1987, Ziegler et al., 1987, Sfrisco et al., 1991), as well as in the lab. Studies utilizing the PES have been conducted in freshwater (Tsai and Lick, 1986, Maclntyre et al., 1990, Davis and Abdelrhman, 1992, Mehta et al., 1994) as well as in the marine environment (Tsai and Lick, 1987, Lavelle and Davis, 1987, Ziegler et al., 1987, Sfrisco et al., 1991, Davis and Abdelrhman, 1992). Specialized studies utilizing the PES have included investigation of the effect of bioturbation on sediment erosion characteristics (Davis and Means, 1989, Davis 1993) and studies of the impact of catastrophic events on short term sediment transport (Mehta et al., 1994).














CHAPTER 4
INITIAL HYPOTHESES, OBJECTIVES, AND RESEARCH PLANS


Problem Overview

At this point a reiteration of the primary issues and objectives of the overall program is useful to place the objectives of the research reported in this document in proper context.


Summary of sources

The overall objective of the program funded by the EAA EPD is to develop methods of irrigation and drainage control for the farms of the EAA that minimize the quantities of phosphorus exported off-farm while not causing material detrimental effects to the crops grown on-farm. The primary on-farm sources of soluble phosphorus (SP) and particulate phosphorus (PP) are considered to be soil mineralization (SP), fertilizer application (SP and PP), and soil/litter/sediment mobilization (PP), where the term sediment is taken in the most general sense to designate any water-resident particulate material.

Organic soil subsidence can result from microbially mediated oxidation of the

organic matter in the soil. The oxidation destroys soil structure by converting soil organic mass to soluble organic and inorganic matter and CO2. The process also causes conversion of the nutrients incorporated in the organic matter to their inorganic form, for example conversion of organic phosphorus to soluble orthophosphorus. Soil subsidence may be retarded by reducing the oxidation-reduction potential of the soil to anoxic or




51






52


anaerobic conditions. Such retardation may be partially affected by control of soil water table levels.

The objective of fertilization is to provide growth factors (nutrients) to the crops when there is insufficient or growth limiting supply from the natural environment. Fertilization contributes directly to phosphorus export totheextentthatthefetilizer application is inefficient, that is, to the extent that applied fertilizer is not taken up by the target crop. Fertilization efficiency may be improved by correct placement and timing of fertilizer application. The most efficient placement and timing would maximize the fraction of applied nutrient that reaches the root zone and optimize the root zone nutrient concentration relative to plant uptake requirements.

Soil/litter/sediment mobilization is a complex function of biological and hydrodynamic conditions that relates not only to the physical character of the transportable material but also to the mode in which water is applied to and removed from the fields and water conveyance systems of the farm. In the soil/litter/sedimentwater system it is necessary, in addition, to consider the inter-conversion of soluble and particulate forms of phosphorus arising from physical adsorption/desorption and biological assimilation and decomposition.

The movement of water within the EAA is dominated by pump operation and characterized by much more rapid hydrodynamic transients than would be the case for a less hydraulically developed area of similar terrain, such as the Everglades National Park. In addition, the District Canal water levels are typically at a greater elevation than those found in the farm wate copveyance systems so there is usually a hydraulic gradient tending to drive water back onto the farm by groundwater flow or structure leak~e.

Other factors that may have an effect on phosphorus transport are the local water chemistry (high calcium content, high photosynthesis potential), phosphorus content of precipitation, and miscellaneous sources arising from wind erosion, vehicle movement, etc.






53

Original conceptual model

Within the context of the modeling portion of this project, fertilizer contribution to phosphorus export is being handled as a subset of soluble phosphorus export in the field groundwater, so for the sake of conciseness fertilization will be excluded from further discussion. With this caveat, the remaining factors, as hypothesized at the initiation of this program, are presented in conceptual form in Figure 4.1.

































Figure 4.1: Farm Scale Phosphorus Transport Original Conceptual Model L '- I (SP)






Lhr
Figre 1 Fam SalePhophous ranpor rignalConeptal ode






54




The figure shows a qualitative model with the following main features:

Soil water supplies or removes soluble phosphorus (SP) to or from the

canal water depending on whether irrigation or drainage is being practiced.

*, Soil water, moving through (perpendicular to) the canal wall, supplies or
removes particulate phosphorus (PP) to or from the canal water depending on drainage (perpendicular flow erosion) or irrigation (perpendicular flow filtration). PP is also supplied to the canal water from the bed by parallel flow erosion arising from longitudinal flow in the channel, and removed

from the canal water by sedimentation.

Precipitation supplies SP and PP directly to the canal water, and causes

surface runoff under some circumstances which also supplies SP and PP.
Biological activity (shown as, but not limited to, phytoplankton and
X
macrophytes) immobilizes SP, and increases or decreases pH, depending

on photosynthetic vs. respiration activity. Particulate detrital biological

material is deposited on the top layer of sediment, which is assumed to be

oxic.

Inorganic particulate matter suspended in the ditch and canal water also

interchanges with the oxic sediments via erosion and sedimentation.

Oxic sediments interchange with anoxic sediments, and both

compartments undergo mineralization of organic PP to SP. Anoxic

sediments may interchange with suspended particulates under the

appropriate hydrodynamic conditions. Anoxic sediments also interchange
with the high-calcium-containing substratum layer.

Soil water may infiltrate through the substratum, losing SP and gaining

Ca2+, that is carried into the bulk phase canal water for precipitation

reaction with SP under appropriately high pH (from photosynthesis). All






55


or part of this precipitate may dissolve under reduced pH conditions (from

respiration).

Canal water may also interact directly with exposed substratum in the bed

to adsorb or desorb phosphorus, depending on surface and bulk phase

conditions.

Hydrodynamic conditions affect interchange of sediments with the bulk

phase, interchange of soil water with the bulk phase, erosion and filtration,

infiltration, redox potential in the sediments, downstream transport of all

mobile constituents and, indirectly, soil/soil-water/crop interchanges.

The model is an unsteady-state one so all interactions must be assumed to

be dependent on antecedent conditions.


Scope of This Research



Minimum criteria for experimental and modeling efforts


Given the fact that virtually no channel transport experimental or modeling

activity had taken place on the EAA farm scale at the time of inception of this research, the decision was made to restrict this effort to fundamental approaches. The field-scale groundwater chemodynamic model for soluble phosphorus was being developed separately and was planned to be integrated with the farm-scale model at a later date. Basically the field scale model is to provide the soluble phosphorus dynamic boundary conditions at the periphery of all water conveyance channels. The model to be developed for the conveyance systems was required to have at least the following properties.

1. It should adequately represent the surface water, groundwater, and channel

flows of a network typical of the EAA farm scale over the time scale of an

entire pumping event, that may cover multiple days.






56


2. It should be capable of sufficiently detailed time scales to allow simulation-with reasonable accuracy of transients that exist at pump start-up and shutdown, however modeling of rapid transients, such as hydraulic jumps, is

not necessary because of the level terrain of the EAA.

3. It should model with reasonable accuracy particulate phosphorus
mobilization, transport, and deposition as a function of some readily

defined hydraulic parameter such as shear stress or average stream

velocity.

4. Phosphorus interchange between soluble and particulate forms by

adsorption-desorption should be represented in the model.

5. The dynamic impact of biological growth and senescence in the aqueous

system on phosphorus transport should be incorporated into the model.


Application of minimum criteria and resulting conceptual model simplification

Water chemistry reactions resulting from interflow through the substratum were deferred for later consideration. Interactions between oxic and anoxic sections of the base sediment can be very sensitive to redox potential profiles and tend to be important over long term (multi-annum) time scales. The field portion of this study was intended to last through only one wet season. It was decided to defer this portion of the study to a later program.

When these minimum criteria were applied and the program was evaluated in

light of prior knowledge, available resources, and time constraints, it became necessary to affect a truncation of the conceptual model to a simpler form.

Particulate filtration and resuspension arising from groundwater flow

perpendicular to ditch and canal banks may be a meaningful contributor to sediment flux but experimental determination of field-relevant parameters could require considerable






57


effort that, from a pragmatic standpoint, should not be expended until the nature of erosion/sedimentation associated with shear flow is reasonably well understood. It was decided the best allocation of resources for this project would be to attempt to understand the nature of shear flow erosion and defer perpendicular flow erosion studies if and until it became evident that they were necessary.

Current restrictions by the South Florida Water Management District limit the growers to a nominal farm pump discharge of one inch of farm runoff per day, therefore the time scale of pumping events in the EAA may usually be on the order of one to three days after a significant rainfall event. Typical time constants for biological assimilation and degradation processes may be on the order of weeks (USEPA 1985). For the first generation models to be developed in this study it was deemed appropriate to ignore biological process transients occurring during the pumping events, that is, to consider that all biological processes of significance occur during the inter-event quiescent periods. Given the accuracy with which biological processes in natural systems can be portrayed, this is a reasonable first approximation.

The conceptual model resulting from these simplifications, illustrated in Figure

4.2, is as follows

Soluble phosphorus is supplied from the fields as a boundary condition in

surface and ground water runoff to the farm drainage ditches.

Soluble phosphorus may also be added to the system with the release of

sediment interstitial pore water when sediment is resuspended.

Particulate phosphorus may be supplied to the system as:

eroded soil or litter carried by overland flow of surface runoff,

biological matter already suspended at event initiation,

sediment that is resuspended as a result of turbulent shear stresses arising from channel flow.

Particulate phosphorus may be removed from the system by sedimentation.






58


Phosphorus may be exchanged between the soluble and particulate phases

by adsorption/desorption

All biological processes are considered to be active only during the

quiescent inter-event periods


Research goals for this study

The simplified conceptual model now leads directly to the general research goals for this study that are, for a specific study farm, as follows:

1. Adapt an existing hydraulic network model for use in the subirrigation/drainage mode that prevails at the specific site.

2. Adapt or develop a water quality sub-module for the hydraulic model that

will allow expression of the water quality processes in a dynamic

hydraulic regime.

3. Determine the nature, frequency, and quantitative contribution of overland

flow erosion to particulate phosphorus in the water conveyance system.

4. Develop a description of the conveyance system sediment with respect to

quantities, locations, chemical/physical characteristics,

erosion/sedimentation characteristics, and temporal variations that is

adequate to allow modeling of sediment transport and pore water release

within the system.

5. Determine the adsorption/desorption characteristics of the system

sediment and incorporate this process into the model where appropriate.

6. Determine the biological processes in the system that contribute

significantly to phosphorus transport and develop a first approximation

lumped parameter model of these processes.

7. Develop, calibrate, and verify the model on the target study site.













sd~aRipms(~ ) SParfPatae SPad nupimrf fmi.n- msscd
mro as FphouA (PP) fmm dth nom as pratptpnm d Wl ai. th a d m an (a E.m4ss anm, wici wourat.

dch~ (Smd layat Ns) T.. I ..T
sdypvrtbi li(tchd r ad Esainguapcur Rchlocaumad
































AA--A
genurat F Sledoprun calivampaoiartfall helonalr M e wkpala&n aaando ds u nsta tre asrbic aninsefuagP










C nl o i l ( ( a e








RFtRnC.e R -.F-G-V
Sednitr F-maClCuk Dayn




F Sigu abr u Fn cu AAna












RqutF-mos 4.IBresDiad R~MWFn IW Ne
I


















Figure 4.2: Farm Scale Phosphorus Transport Simplified Concentual Model






60



General Research Plan

The preceding provides the framework for discussion of the original general research plan. It is appropriate to reiterate that at the time of development of this plan there was no information in the open literature relating specifically to transport of high organic content materials such as the soils and sediments of the EAA. Given that lack of information it was deemed appropriate to start with a basic evaluation of the target systems and then develop a fundamental data gathering program. Following is an outline of this approach.

1. Conduct a general survey of several representative farm conveyance

systems in the EAA, restricted initially to main farm canals, to determine:

a. General canal dimensions and pumping capacities, b. Volume and mass of sediment contained in canals,

c. Selected physical and chemical characteristics of representative sediments.

2. Select a target farm for detailed study.

3. Obtain sufficient hydrographic and topographic data to allow the farm

drainage network to be modeled.

4. Select a representative source of sediment from the target farm and

conduct physical and chemical characterization.

In the laboratory, determine erosion and sedimentation characteristics of

the representative sediment and use this as a prototype system for

modeling purposes.

6. Develop or adapt some small-scale device that would allow estimation of

erosion characteristics in the field and use this device to study geographic

and temporal variations of sediment erodability.






61


7. Set up regular discharge monitoring of particulate phosphorus at the target

farm.

8. Plan and execute synoptic studies to:

a. determine variation of sediment characteristics as a function of location within the conveyance system and b. evaluate suspended solids concentration variations within the conveyance system during transport events.

9. Determine, in the field, the requirements for biological data depending on

the outcome of the preceding characterizations and monitoring programs.

10. Adapt the chosen computer program, develop the water quality program

for phosphorus, and calibrate and verify for the target farm.



The subsequent four chapters discuss the specific activities directed toward sediment characterization and transport property determination.














CHAPTER 5
SEDIMENT SURVEY AND PHYSICAL-CHEMICAL CHARACTERIZATIONS OF SELECTED SEDIMENTS


Sediment Survey

The first field activity of the program was to develop an order of magnitude

estimate of quantities of sediment present in representative farm canals within the EAA.

Typical field ditch construction technique in the EAA involves digging or

dredging at least down to the limerock/marl substratum underlying the organic soils. In the case of main canals, the limerock is usually excavated to a depth of several feet below the marl surface to improve hydraulic flow conditions. Ensuing physical and biological processes combine to contribute a sediment layer that builds up in the conveyance system over time. Periodically the system components must be dredged to remove the sediment build-up. Dredging is a relatively costly process, especially for the small grower who must contract out the work, thus the extent of ditch and canal maintenance via dredging is variable among growers.

The excavation or dredging technique, which initially removes all soil from the channel, does simplify the measurement of sediment inventory, because all transportable particulates present in the channels may be assumed to be sediment.


Selection of representative farms

The research program "Implementation and Verification of Best Management

Practices for Reducing Phosphorus Loading in the EAA", conducted by the University of Florida's Institute of Food and Agricultural Sciences (IFAS), had a total of ten farms



62






63


within the EAA under intensive study. These farms had been chosen to represent a typical range of farm size, crop type, soil type, and geographical distribution within the four main sub-basins of the EAA.

The farms have, by mutual agreement with the appropriate regulatory authorities. been coded to preserve the anonymity of the growers who were participating in the study program on a voluntary basis. The code numbers were in a series of UF9200 through UF9209, where "UF" indicates University of Florida study, "92" indicates the start year of the BMP study program and "00" through "09" indicates the code number assigned to each participating farm. Further coding indicated the location of permanent sampling sites at a specific location. "A" refers to the main farm discharge; some farms have two major discharge locations so "B" is reserved for the second discharge location where appropriate. Sites "C" and beyond refer to internal farm sampling locations that are added as the program dictates. For example farm UF9206 has two pump discharge locations, so UF9206B refers to the permanent sampling location at the south pump station of that farm. That coding protocol will be modified later in this document where UF9206A and B will be referred to as UF9206N and S respectively.

Three sites were selected for the preliminary sediment inventory based on the recommendations of Dr. F. T. Izuno, the IFAS Program Principal Investigator. They were:

UF9200 A 1280 acre (518 ha) farm with two main canals and one pump station,

that planted only sugarcane. This farm was representative of medium to large

sugarcane-only fields with relatively deep soil that are actively managed.


UF9202 A 320 acre (130 ha) farm with one main canal and one pump station that planted only sugarcane. This farm was representative of small to medium sugarcane-only fields with relatively shallow soil that receive low to moderate

management.






64




UF9206 A 1750 acre (710 ha) farm with a complex drainage pattern that

included three main canals, several perimeter canals, two pump stations, and gated interconnects among the various sections of the farm. This was a multi-crop farm

that maintained stands of sod and sugar cane and rotated plantings of rice and

winter vegetables at various locations. This farm was representative of medium to

large farms with relatively deep soil that maintain multi-crop patterns and are

actively managed.


Sediment survey measurement methods

Water depth and sediment depth transects were taker(by bo~tfrom cordoned canal sections. At three foot (0.91 m) increments, the distance from the water surface to the sediment surface was determined by lowering a neutrally buoyant one ft2 (0.09 m2 ) pad attached to a calibrated rod until it rested on the sediment surface. At the same location the distance from the water surface to the canal bottom was determined by driving a 0.5 in. (1.3 cm) diameter calibrated steel penetrometer rod through the sediment until it met resistance at the marl surface. Sediment depth at each location was determined by difference between water-surface-to-sediment-surface depth and water-surface-to-marlsurface depth.

At each site water and sediment depth transects were taken at multiple locations depending on the site main canal configurations. Four transects were taken at UF9202, eight at UF9200, and twelve at UF9206. Two sediment core samples were taken at each site using the piston core sampler described in Appendix A. The cores were expunged on-site into plastic bags, sealed, refrigerated, and transported to the Gainesville, FL, laboratory for bulk analysis. Analyses were conducted according to the methods described in Appendices B and C.







65





Sediment survey results


The physical results of the sediment survey are shown in Table 5.1, the analytical results appear in Table 5.2.



Table 5.1: Sediment Survey Physical Results


Location Total Main Farm Area Average Estimated Total Estimated Unit Canal Length (ha) Sediment Depth Sediment Sediment
(m) (m) Volume Volume m) (m'/ha)

UF9200 6437 518 0.79 39277 75.8 UF9202 1609 130 0.65 5486 42.2 UF9206 11715 710 0.54 40368 56.85



Table 5.2: Sediment Survey Analytical Results



Location Core % Dry % Ash % Volatil Pore Wet Bulk Solids Pore Solids TP Length Solids (Dry) Solids Water pH Density Specific Water SP Content
(cm) (Dry) (gm/ml) Gravity (mg/l) mg/kg UF9200 60.7 14.77 52.5 47.5 7.21 1.088 1.596 0.35 787 #I

UF9200 52.8 18.09 46.1 53.9 7.16 1.082 1.453 0.20 868 #2
UF9202 55.1 18.61 53.4 46.6 7.12 1.094 1.505 0.03 613 #1
UF9202 41.7 18.44 43.1 56.9 7.16 1.058 1.315 0.56 572 #2-Top
UF9202 33.0 45.99 77.7 22.3 7.06 1.378 1.822 0.02 206 #2-Bott.
UF9206 371 25.16 47.9 52.1 7.02 1.114 1.453 0.07 415 #1
UF9206 59.4 22.40 42.7 57.3 7.05 1.100 1.446 0.17 445 #2






66


At locations UF9200 and UF9206 samples #1 and #2 were taken at mid-length of each of the two major farm canals. At UF 9202, which had only one main canal, samples #1 and #2 were taken at 25% and 75% of the upstream length, respectively. All samples except UF9202 #2 were homogeneous in appearance and were treated, for this order-ofmagnitude exercise, as a single bulk sample. UF9202 #2 exhibited two definite layers, expressed as a sudden color change from black to medium gray at the 41.7 cm depth level, so it was split into two subsamples, denoted as "top" and "bottom" in Table 5.2.


Sediment survey discussion

The sediment average depths at the three target farms ranged from 0.54 m (1.75 ft) to 0.79 m (2.6 ft). The estimated total sediment volumes varied greatly as would be expected because of the farm-to-farm variation in total canal length. An attempt at normalization is shown in the last column of Table 5.1, where the canal sediment volume is expressed on a per-unit -farm-area basis. On this basis the unit sediment volume ranged from 42.2 to 75.8 m3 canal sediment/hectare farm area.

The analytical data presented in Table 5.2 (excluding UF9202 #2 Bottom from the analysis) were used to draw some preliminary inferences. The dry solids content ranged from approximately 15 to 25%, indicating relatively well compacted sediments. Volatile contents ranging from about 47% to 57% and solids specific gravities of approximately

1.3-1.6 indicate a high organic content in the sediments. The volatiles content is somewhat lower than the typical 70-80% volatile content found in the organic soils of the EAA, and may represent selective transport or diagenesis or both. The specific gravities of 1.3-1.6 are of interest because they are substantially lower than the typical 2.1-2.2 values of inorganic sediments, and reinforce the position that direct application of traditional sediment transport concepts may be inappropriate.






67


The pore water pH values ranged from 7.02 to 7.21, only slightly on the alkaline side, indicating an environment not particularly conducive to precipitation of calcium or magnesium phosphorus complexes. The pore water soluble phosphorus content showed a significant range, from a low of 0.03 mg/I to a high of 0.56 mg/l1, with an average of 0.23 mg/l. In the absence of adsorption equilibrium data a full interpretation could not be placed on the significance of these values at the time of acquisition, but the average value of 0.23 mg/i was not out of line with aqueous phase concentrations reported elsewhere and discussed in Chapters 1 and 2 of this document. The approximate concurrence of the pore water phosphorus concentrations and typical aqueous phase phosphorus concentrations gave at least a preliminary indication that direct pore water contribution to bulk aqueous phosphorus was of secondary importance. A quantitative estimate of pore water contributions given later in this chapter substantiates this inference.

The sediment-column average phosphorus content for the three farms shows a reasonable level of agreement between same-farm samples, implying intra-farm geographical homogeneity (coefficients of variation of 5-7%) with a higher level of interfarm variation (coefficient of variation of 29%). The average sediment phosphorus content over the three farms was 617 mg/kg. It is interesting to note that this sediment value falls well within the range of 289-834 mg/kg and close to the average of 520 mg/kg for the phosphorus content of the top 10 cm of Pahokee Muck soil from eleven locations within the EAA reported by Fiskell and Nicholson (1986). It is tempting to infer the implication that the soil surface was the source of the sediment but, as will become evident in later sections, this inference may lead to erroneous conclusions.

The accumulated data may be used to calculate an estimated total mass of phosphorus contained in the main canal sediments of each farm. These estimates are shown in Table 5.3, where it is seen that the Unit Phosphorus mass estimates range from

5.0 to 11.2 kg of canal sediment total phosphorus per hectare of farm area. It is now instructive to calculate the supply potential of these masses of phosphorus.






68




Table 5.3: Estimates of Total Phosphorus Mass in Target Main Canal Sediments


Location Dry Solids Mass Total Phosphorus Mass Unit Phosphorus Mass
(kg) I (kg) (kg/ha) UF9200 7.00xl06 579.7 11.2 UF9202 1.09x10, 64.7 5.0 UF9206 10.63x10 457.0 6.4



In Chapters I and 2 various values of off-farm phosphorus export were reported. A reasonably typical set of values from the data of Izuno et al. (1991) and CH2M Hill (1978) would be average total discharge phosphorus concentration of 0.2 mg/I, with 35% or 0.07 mg/I present as particulate phosphorus. A conservative estimate of average annual off-farm pumping would be 50.8 cm (20 inches) of rainfall equivalent. On a perhectare basis 50.8 cm would equal 5080 m3/yr. A particulate concentration of 7xl0"6 kg/m3 (0.07 mg/1) would yield an annual particulate phosphorus export of 0.355 kg/ha.

Comparing the estimated typical particulate phosphorus annual export rate with the inventory range of 5.0-11.2 kg/ha it is evident that the main canal sediments have a particulate phosphorus yield potential equivalent to something on the order of 15-30 years worth of phosphorus supply. There is no intent in this exercise to imply that all the sediment particulate phosphorus is readily available or fully mobilizable, however this order-of-magnitude estimate does serve to emphasize the potential importance of the farm canal sediments as sources of phosphorus, both short term and long term. It should be reiterated here that this estimate was restricted to the main farm canals only and did not include sediment stored in field ditches (lateral channels that run perpendicular to and intersect with the main farm canals). A more detailed study of the sediment in the field ditches of the primary target farm is presented in Chapter 6.






69


Selection of Primary Target Farm

Availability of resources at the time of initiation of the field portion of this

research dictated that the intensive field activity be restricted to one target farm, with the understanding that the results from the target farm research would serve as a basis for determining further research work needed to generalize the model. The selection criteria were limited to a few important points, which were:

I. The target farm should be one of the ten farms participating in the IFAS

Best Management Practices Program.

2. It should have a relatively straightforward drainage layout and hydraulic

management program.

3. It should, as the prototype farm, be restricted to a sugarcane-only planting

on typical soil type and depth. At the time sugarcane represented

approximately 70% of the total agricultural acreage in the EAA.

4. It should be a medium-to-large size farm that practiced at least average and

preferably aggressive crop and water management policies.

5. The grower should be willing and able to provide access, assistance, and

information on a regular basis.

Farm UF9200 met these criteria and was chosen as the target farm for more intensive study.


Particle-Size Property Distribution Study

The results of the sediment survey represented on-farm particulate transport potential. The data set from the SFWMD (Mucinic 1994), discussed in Chapter 2, represented actual transported material sampled at EAA discharge points. A comparison of representative particulate phosphorus contents showed the on-farm sediment to average 617 mg P/kg solids, while the selected data set from the SFWMD implied that






70


particulate matter at the EAA discharge points might contain on the order of 2200 mg P/kg suspended solids. This three-to-four fold difference indicates the possibility of selective transport or enrichment processes being of some significance in the EAA water conveyance systems. One approach to evaluating the possibility of selective transport was to fractionate material from various sources by particle size and determine if there were appreciable differences in key properties of the several fractions.


Sources for particle size fractionation study

It was desired to evaluate farm soil, farm conveyance system sediment, and Water Management District Canal sediment. Sampling was conducted in mid-July, approximately 1.5 months into the wet season. The farm sediment was obtained from the midpoint of a field ditch in farm UF9200. The ditch chosen was one approximately halfway upstream and on the south side of the south canal of UF9200. It was well maintained and free of emergent growth. Samples of the ditch surficial sediment were obtained at ditch mid-length by Eckman dredge technique. Simultaneously composite samples were taken of the field soil at locations immediately adjacent to the sediment sample site and 5-10 meters in-field. The UF9200 grower identifies his fields by increasing number from east to west and by letter from north to south (See Appendix G). Field ditches are identified by the two adjacent fields. Field ditch B9B10 was the sample site and was the ditch between fields 9 and 10 on the south side of the central (South) canal. This sample will be referred to hereafter as "B9B10", the soil sample will be referred to as "Soil". Farm UF9200 discharges into the West Palm Beach Canal. The District Canal sample was taken from this canal approximately I km downstream of the UF9200 discharge point at a location intermediate between two downstream farm pump stations. This sample was also of surficial sediment, taken by Eckman dredge technique and is referred to as "WPBC". Approximately 20 liters of each type material were






71


collected, stored in sealed plastic buckets, and transported to the Gainesville laboratories for fractionation


Soil and sediment particle size fractionation

Several fractionation methods were evaluated and tested on surrogate sediments prior to sampling the target sediments. The methods included screening, hydraulic classification, and differential sedimentation. It was hoped to develop a method that would fractionate on the basis of sedimentation velocity, which would relate to a key physical parameter of field interest. Unfortunately all the hydraulic classification and differential sedimentation techniques evaluated either suffered from a lack of available equipment that could process 20 liters of material in any reasonable time span or subjected the test material to contact with large volumes of water. The latter condition was deemed to be inadvisable because of the potential for alteration of the physical or chemical nature of the materials over repeated dilutions with water. Ultimately the fractionation method of choice reduced to wet and dry screening. The three samples were fractionated in their entirety on US Standard screens in the sequence shown in Table 5.4. Note that "+" means "retained on", while "-" means "passed through".

Soil Fractionation The "Soil" sample was first subjected to coarse screening to remove clumps greater than I cm. The remaining material was dried at 700 C for 72 hours and then subsampled to produce a representative composite. The composite was split into two sub-segments of roughly 315 gm each. Each sub-segment was subjected to screening in a stack of USS sieves on a Ro-Tap shaker table for 30 minutes, then removed, weighed, and recombined. In order to place the evaluation of the soil particle fractions on the same basis as the sediment it was deemed appropriate to hydrate the soil to simulate the process of soil eroding into a conveyance system and becoming subjected to an aqueous environment.






72


Table 5.4: Soil and Sediment Screen Fractionation Sequence



Screen Fraction Particle Size Range -micrometers

"Coarse" >10,000

+8 USS 10,000-2360

-8 USS+16 USS 1180-2360 -16 USS+30 USS 600-1180 -30 USS+50 USS 300-600 -50 USS+100 USS 150-300 -100 USS+200 USS 75-150 -200 USS+400 USS 38-75

-400 USS <38



A small sub-sample of the recombined sample was taken for moisture analysis and the rest of the sample was gently combined with 7 liters of filtered native EAA water that had been collected at the farm side of the UF9200 pumphouse. This mixture was gently agitated on a twice-daily basis for eight days at room temperature and then subjected to wet screen analysis (See next section)

Sediment Fractionation The sediment wet creening fractionation process was much more labor intensive than the soil dry creening. When the sieves were stacked the cohesive, almost gelatinous, nature of the sediment samples caused the sieving process to stop almost immediately after the suspension completely wet the screens because of lack of back-flow of air from the lower sieves to the higher sieves. In order to overcome this problem the sieving process was done manually, one screen fraction at a time, starting with the coarse 1 cm screen. Wash for the wet screening process was obtained by recycling decanted supernate from settled suspension that had already passed through the




Full Text
60
General Research Plan
The preceding provides the framework for discussion of the original genera]
research plan. It is appropriate to reiterate that at the time of development of this plan
there was no information in the open literature relating specifically to transport of high
organic content materials such as the soils and sediments of the EAA. Given that lack of
information it was deemed appropriate to start with a basic evaluation of the target
systems and then develop a fundamental data gathering program. Following is an outline
of this approach.
1. Conduct a general survey of several representative farm conveyance
systems in the EAA, restricted initially to main farm canals, to determine:
a. General canal dimensions and pumping capacities,
b. Volume and mass of sediment contained in canals,
c. Selected physical and chemical characteristics of representative
sediments.
2. Select a target farm for detailed study.
3. Obtain sufficient hydrographic and topographic data to allow the farm
drainage network to be modeled.
4. Select a representative source of sediment from the target farm and
conduct physical and chemical characterization.
In the laboratory, determine erosion and sedimentation characteristics of
the representative sediment and use this as a prototype system for
modeling purposes.
6. Develop or adapt some small-scale device that would allow estimation of
erosion characteristics in the field and use this device to study geographic
and temporal variations of sediment erodability.
J-


Introduction..... 318
Evaluation of Potential Management Practices 318
Reduction of transportable mass in the system 319
Effect of sediment trap 320
Effect of pumping modification 322
Effect of wider discharge outlets 324
Hydraulic mining and recycle analysis 326
Specific effect of reduced velocity 1 332
Evaluation of Suspended Solids and Particulate Phosphorus Export at
Farm UF9206 334
Possible dominant influence of channel water depth on velocity and
transport 334
Correlation of export suspended solids with channel depth 337
Particulate phosphorus export at UF9206 341
Relation of Particulate Phosphorus Export to Soluble Phosphorus Export
at Both Farms 345
Conclusions and Critique 346
Conclusions 346
Critique of the model in its current form 350
Recommendations 352
BMP recommendations for field implementation 352
Model framework development 353
Model development 354
BMP development 355
APPENDICES
A PHYSICAL SAMPLING TECHNIQUES 356
B PHYSICAL ANALYSIS 359
C CHEMICAL ANALYSIS 363
D APPROXIMATE MODEL FOR CRAF 365
E RAINFALL AND WATER LEVELS AT UF9200 369
F DIMENSIONS OF UF9200 USED IN DUFLOW FORMAT 373
G FARM LAYOUTS FOR UF9200 AND UF9206 386
H DUPROL EROSION PROGRAM 389
REFERENCES 392
BIOGRAPHICAL SKETCH 404
x


161
30 field ditches (a very gross assumption) yields an estimate of initial field ditch velocity
of 0.033 m/sec.
Table 7.1: Average Conveyance Dimensions for UF9200
Average Elevation
North Section
South Section
Marl Layer
1.75 m
1.81 m
Sediment Surface
2.05 m
2.07 m
Field Surface
2.98 m
3.03 m
Average Field Ditch Width
2.97 m
3.03 m
Canal Measurements
Widths (Min/Max/Avg)
Bottom
Elevation(Min/Max/Avg)
South Canal
1.83/7.01/4.96 m
1.97/2.26/2.12 m
East Canal
6.40/7.62/7.11 m
1.79/1.93/1.85 m
North Canal
3.35/12.24/9.75 m
1.38/1.99/1.77 m
Particulate phosphorus content determination
An important aspect of the analytical effort in this program was the determination
of the concentration of suspended solids in discharge samples and the phosphorus content
of the suspended solids on a mass fraction basis. Early in the program a number of
options were evaluated for attempting to determine the origin of the discharged


10
precipitation and evapotranspiration patterns between time periods The long-tenn "end-
of-pipe" phosphorus loads to the ENP are set with some correction for antecedent flow at
Shark River Slough (Whalen et al., 1992), which was based on statistical analysis of
loading over an 11 year period, but statistical correlation over a multi-year period of
multiple combinations of BMPs at the farm level is not an efficient method for evaluating
and optimizing the effectiveness of BMPs. More precise correction methods are needed.
An appropriate alternative to long-term statistical evaluation is development and
application of mathematical models which adequately describe the dominant
hydrodynamic and chemodynamic processes controlling phosphorus transport and export
at the field, farm, and basin scale. Such models could be used to forecast the impact of
current or proposed BMPs under very site specific conditions, and to evaluate the actual
performance of BMPs under field conditions which deviate from the long-term mean.
EAA Modeling Activity
The need for, and value of, models for phosphorus transport in the EAA has been
recognized by both the growers and governmental agencies. One specific agency that
combines aspects of both is the Everglades Agricultural Area Environmental Protection
District (EPD), a state chartered organization with assessment powers that coordinates the
environmental regulatory compliance activities of growers in the EAA. Among
numerous research activities, the EPD has provided funds for BMP research.
The research is aimed at optimization of the original set of BMPs recommended
by Bottcher and Izuno (1993) and development and application of new or modified BMPs
for reduction of phosphorus export from the EAA. An integral part of this research
activity is the development of field and farm scale models, specific to the EAA
environment, for phosphorus sourcing and transport. This effort was originally initiated
by the University of Florida Institute of Food and Agricultural Sciences (IFAS) and


297
Correlation of Calibration Values of Cfmihi. Initial Erodable Mass
Recall from the calibration process, described in Chapter 9, that Cem(O) was
initially assumed to vary linearly with interevent time and then in the final stages of the
calibration process was adjusted independently for each event to achieve an optimum fit.
The final values of Cem, correlated with interevent time, are shown in Figure 10.21.
Here the interevent time is calculated based only on operation of the large pump. This
means that Event 244, the event where only the small pump was run intermittently, is
ignored in the calibration process.
Figure 10.21 indicates that there appeared to be a non-linear relationship between
interevent time and Cem(O), which was expressed empirically as a parabolic equation,
Qm(o) = 0.0013/2 0.3121/ + 337.96
where Cem(0) has units of gm/m2 and the interevent time, t, has units of hours.
10.1
Initial Erodable Mass Correlation with Interevent Time
Figure 10.21: Correlation of Initial Erodable Mass with Interevent Time


I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
Kenneth L. Campbell Chair
Professor of Agricultural and Biological
Engineering
I certify that 1 have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and quality-,
as a dissertation for the degree of Doctor of Philospph
jpbyr\
Konda R. Reddy, Co-Chair
Professor of Soil and Water Science
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
Adelbert B. Bottcher
Professor of Agricultural and Biological
Engineering
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
Joseph J. Delfino
Professor of Environmental Engineering
Sciences
1 certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
Forrest T. [zuno,
Professor of Agricultural and Biological
Engineering


43
history. Second, and more important, the traditional non-cohesive approach makes no
assumptions about supply limitation of transportable material, which is a major drawback
in the evaluation and simulation of organic sediment transport.
Cohesive sediment transport
Cohesive sediments are those that exhibit particle-particle interactive forces
strong enough to cause them to not act as a collection of elementary particles. They are
typically composed of fine-grained particles which, under the appropriate circumstances,
assume an electrochemical surface charge that contributes to their interactive nature. The
particle size, chemical charge, and applied shear can operate in close packed
environments to impart rheological properties to the sediments, which affects their
erosion characteristics, and in dispersed environments to cause flocculation and dis
aggregation of suspended particles, which affect their sedimentation characteristics.
Partheniades (1977), using concepts developed and presented in earlier papers,
proposed a unified theory that treated bed load, suspended load and wash load all as
special cases of a general erosion-transport-deposition model. His conceptual framework
includes a minimum critical shear stress for inception of erosion, a maximum critical
shear stress for termination of deposition, addition of inter-particle forces to the force
balance for detachment of a particle from the bed, incorporation of probabilities that a
particle may suspend or deposit which are related to shear levels, and the concept of
supply limitation and carrying capacity saturation for any given class of sediment. One of
his important conclusions is that there are classes of sediment that can behave either as
wash load or as bed load, depending on channel discharge characteristics.
Organic sediments do not generally possess electrochemical charges strong
enough to affect their compacted strength but their complex physical configurations may
impart some degree of particle-particle interaction in the settled phase. There are


APPENDIX E
RAINFALL AND WATER LEVELS AT UF9200
The rainfall records and water levels at selected locations at farm UF9200 for the
study period of June 1 -December 31,1994 (Julian Days 152-365) are shown in the
accompanying figures. Figure E.l shows rainfall for the entire period. Figures E.2 and
E.3 show the canal water level at the pump station and the field water table level in Field
A16, the field well most remote from the pump station. Figure E.2 approximately covers
the normal wet season, Julian Dates 150-300. Figure E.3 covers the remainder of the
year, Julian Dates 300-365. The onset of a series of tropical depressions began about
Julian Date 318.
369


190
Time Series-Flow and Total Suspended Solids
Events UF9206N-209. 213
Pumping Rate O -TSS Concentration
Figure 7.21: Time Series-Flow and Total Suspended Solids Events UF9206N-209,213
Time Series-Flow and Total Suspended Solids
Events UF9206N-224. 234, 239
Pumping Rate O TSS Concentration
Figure 7.22: Time Series-Flow and Total Suspended Solids Events UF9206N-224,234,
239


14
The vegetable farm showed study average effluent total phosphorus of
0.460 mg/1, soluble phosphorus of 0.355 mg/1 (77% of total), particulate
phosphorus of 0.105 mg/1 (23% of total) and an annual effluent total
phosphorus load of 2.10 kg/ha, roughly four times that of the sugarcane
farm.
The cattle ranch showed study-period average effluent total phosphorus of
0.167 mg/1, soluble phosphorus of 0.122 mg/1 (73% of total), particulate
phosphorus of 0.045 mg/1 (27% of total) and an annual effluent total
phosphorus load of 0.55 kg/ha. Concentrations were higher than for
sugarcane but load was lower because less water was discharged.
Soil water was monitored. Average soil water soluble phosphorus was
5.74 mg/1 at the vegetable farm, 3.88 mg/1 at the cattle ranch, and an
extremely low 0.04 mg/1 at the sugarcane farm.
Variability was high. Most phosphorus concentrations showed standard
deviations approaching or exceeding the mean.
Seasonal variation of phosphorus concentrations was much higher for
vegetables than for sugarcane.
Phosphorus applied as fertilizer per unit area was about 6.5 times as high
for vegetables as for sugarcane.
A comparison of all studied sites and additional data obtained from
sampling of private drainage districts during the study indicated that the
cropland which had been under cultivation longer had less ability to retain
phosphorus.


144
profile close to the bed (within 3.8 cm) was parabolic and was well represented by the
equation for turbulent flow in the rough regime (Equation 6.29).
It is interesting to note that a somewhat similar result may be obtained if the
hydraulic profile is estimated assuming that the ring is hydraulically smooth, obeying
Equation 6.30, the bed is hydraulically rough, obeying Equation 6.29, and the effects of
the walls are negligible. An approximate model that illustrates this is given in Appendix
D. Under these conditions there is no flattening of the profile but the velocity at mid
depth is on the order of 35% of the ring-channel AV for roughness and shear stresses
assumed to be representative of the organic sediment system.
In order to choose a reasonable characteristic velocity, let us consider the flow in
the CRAF as if it were occurring in an open channel with the same near-bed velocity
profile. Under this assumption the vertical velocity profile segment and the upper limb of
the sigmoidal curve would be replaced by an extension of the logarithmic velocity profile
from the near-bed region. The logarithmic velocity profile would represent the
transmission of the same shear stress to the bed in an open, wide channel as is being
transmitted in the CRAF. This is shown schematically in Figure 6.20.
Now it can be shown that for a fully developed logarithmic profile in an open,
wide, hydraulically rough channel the velocity at 0.368d (where d = total flow depth) is
equal to the channel mean velocity. What is required is a relationship between the ring-
channel AV and the velocity of the extended logarithmic profile at this elevation. Using
Mehtas fitted velocity equation of
6.31
it is possible to calculate an apparent value for k of 0.64 cm, which converts by
Sticklers approximation (Stickler 1923) to a Mannings n of 0.018.


7.5 Comparison of Pumping Events Between Farms UF9200 and UF9206 for
Julian Dates 218 through 350 185
7.6 Sampling Locations for Intensive Synoptic Studies at UF9200 202
7.7 UF9200 Events 220-285 Expressed in Areal Loading Terms 212
7.8 UF9200 Weighted Areal Loads Estimated From Equations 7.7 and 7.8 214
8.1 Field Ditch Surficial Sediment Phosphorus Content UF9200 Synoptic
Survey of Julian Date 229 219
8.2 Canal Surficial Sediment Phosphorus Content UF9200 Synoptic Survey
of Julian Date 229 220
8.3 Canal Surficial Sediment Phosphorus Content UF9200 Survey of Julian
Date 300 223
8.4 Macrophyte Mass Density, Volatile Content, and Phosphorus Content 227
8.5 Water Lettuce Dislodgable Detritus Study 231
8.6 Phosphorus Fractionation Analysis of Large-Scale Composite Samples 241
8.7 Comparison Between Large-Scale Composite Samples and Prototype
Surficial Sediment for Phosphorus Content Distribution 242
9.1 Results of Hydraulic Calibration for UF9200 263
10.1 Erosion Parameters Determined in the Calibration Process 279
10.2 Placement of Example Nodes for Event UF9200-327 Simulation 286
10.3 Interevent Time and Values of CEm(0) Used for Validation Simulation 305
11.1 Load Weighted Distribution of Soluble and Particulate Phosphorus at
UF9200, UF9206N, and UF9206S 346
D. 1 Example of Spreadsheet Output 367
D.2 Spreadsheet Formulas 368
xii


198
Station 9206N Phosphorus Content of Suspended Solids as a Function of Suspended
Solids Concentration (Expanded Scale)
Suspended Solids Concentration-mg/1
Figure 7.35: UF9206N Phosphorus Content of TSS as a Function of TSS Concentration
(Expanded Scale)
Figure 7.36: UF9206S Phosphorus Content of TSS as a Function of TSS Concentration
(Expanded Scale)


52
anaerobic conditions. Such retardation may be partially affected by control of soil water
table levels.
The objective of fertilization is to provide growth factors (nutrients) to the crops
when there is insufficient or growth limiting supply from the natural environment.
Fertilization contributes directly to phosphorus export to the extent that the fertilizer
application is inefficient, that is, to the extent that applied fertilizer is not taken up by the
target crop. Fertilization efficiency may be improved by correct placement and timing of
fertilizer application. The most efficient placement and timing would maximize the
fraction of applied nutrient that reaches the root zone and optimize the root zone nutrient
concentration relative to plant uptake requirements.
Soil/litter/sediment mobilization is a complex function of biological and
hydrodynamic conditions that relates not only to the physical character of the
transportable material but also to the mode in which water is applied to and removed
from the fields and water conveyance systems of the farm. In the soil/litter/sediment-
water system it is necessary, in addition, to consider the inter-conversion of soluble and
particulate forms of phosphorus arising from physical adsorption/desorption and
biological assimilation and decomposition.
The movement of water within the EAA is dominated by pump operation and
characterized by much more rapid hydrodynamic transients than would be the case for a
less hydraulically developed area of similar terrain, such as the Everglades National Park.
In addition, the District Canal water levels are typically at a greater elevation than those
found in the farm water conveyance systems so there is usually a hydraulic gradient
tending to drive water back onto the farm by groundwater flow or structure leakage.
Other factors that may have an effect on phosphorus transport are the local water
chemistry (high calcium content, high photosynthesis potential), phosphorus content of
precipitation, and miscellaneous sources arising from wind erosion, vehicle movement,
etc.


397
Kizer, P., Galas. J., Sinke, A., and de Joode, P. 1993. Phosphate Immobilization in
Peaty Sediment. Hvdrobioloeia. 253:374-375
Kleijwegt.R.A.. Veldkamp.R.G., and Nalluri.C. 1990. Sediment in Sewers: Initiation of
Transport, Water Sci. Technol.. 22:239-246
Koch.M. 1991. Soil and Surface Water Nutrients in the Everglades Nutrient Removal
Project, South Florida Water Management District Tech. Pub. No. 91-04. West
Palm Beach. FL
Konhya. K.D.. Skaggs, R.W., Gilliam. J.W., 1992 Effects of Drainage and Water
Management Practices on Hydrology, J. Irrig. Drain. Engng.. 118(5): 807-819
Kranck,K., 1984a. Settling Behavior of Cohesive Sediments, In Estuarine Cohesive
Sediment Dynamics, A.J. Mehta, Ed., Springer-Verlag, Berlin
Kranck,K., 1984b. Role of Flocculation in the Filtering of Particulate Matter in
Estuaries, In The Estuary as a Filter, Academic Press, Orlando
Krone R. B., 1962. Flume Studies of the Transport of Sediment in Estuarial Shoaling
Process. Final Report, Hydraulic Engineering Laboratory and Sanitary
Engineering Research Laboratory, University of California, Berkeley
Kronvang,B. 1992. The Export of Particulate Matter, Particulate Phosphorus, and
Dissolved Phosphorus from Two Agricultural Basins: Implications on Estimating
the Non-Point Phosphorus Load, Water Res.. 10:1347-1358
Kronvang.B., and Christiansen,C. 1986. Paths of the Suspended Particulate Inorganic
and Organic Matter in a Small Urban Estuary, Nordic Hydrology. 17:31-46
LabadzJ.C., Burt.T.P., and Potter,A.W.R. 1991. Sediment Yield and Delivery in the
Blanket Peat of the Moorlands of the Southern Pennines, Earth Surf. Proc.
Landforms. 16:255-271
Lavelle.J.W., and Davis, W.R. 1987. Measurements of Benthic Sediment Erodibility in
Puget Sound, Washington, NOAA Technical Memorandum ERL PMEL-72
Lee.D.Y., Lick.W., and Kang.S.W. 1981. The Entrainment and Deposition of Fine-
Grained Sediments in Lake Erie", J, Great Lakes Res.. 7:224-233
Lick.W. 1982. Entrainment, Deposition, and Transport of Fine-Grained Sediments in
Lakes, Hvdrobiologia. 91:31 -40
Lick, W, Lick, J., and Ziegler, C.K. 1992. Flocculation and its Effect on the Vertical
Transport of Fine-Grained Sediments, Hvdrobiologia. 235/236: 1-16


328
back to the canal discharges so the end result was essentially no reduction in net solids
export.
Figure 11.7: Solids Recycle Simulation RCL1 Short Duration, Short Distance
Simulation RCL2 Long Duration. Lone Distance In this simulation the recycle
location is moved 1600 m upstream in the North Canal to the canals midlength point, in
what is assumed to be an external conveyance element, either a pipe or a separate clean
channel. The recycle rate is held at 1.1 m3/sec but the time of recycle is increased from 4
hours to 72 hours. Figure 11.8 shows the results of this simulation. In this case the
reverse velocity did not exist at the lower end of the North Canal to move solids
upstream. In addition a portion of the solids delivered at the receiving point migrated
downstream with the recirculating currents to increase the start-up erodable mass
concentration in these sections. The result in the early part of the simulation was that
reduction of the initial solids transport in the East Canal was rapidly counterbalanced and
then slightly exceeded by increased transport in the North Canal.


329
This was a temporary situation, however, because later in the simulation, at about
60,000 m3 of net cumulative discharge, solids export fell substantially below that of the
original simulation and remained there for the duration. This was attributable to two
factors. First the long recycle period allowed export to the recycle receiving point of
upstream solids from the South-East Canal system. These solids were not available for
export from the South-East system later in the simulation. Second, a large fraction of
these solids that were exported to the recycle point were deposited during the recycle
phase in regions that subsequently had lower velocities during pump-down than the
regions where they were originally located. The end result of this process of moving
solids from areas of higher maximum velocity to areas of lower maximum velocity during
the recycle period was a net simulated reduction of solids export from the original 10,420
kg to 3960 kg for this case.
Figure 11.8: Solids Recycle Simulation RCL2 Long Duration, Long Distance


75
B9B10 Soil Panicle S12E Distnbution-Hydraied and Dry
Screen Opening Size-Micrometers
*Hydrated B9B10 Soil O
-Dry B9B10Soil
Figure 5.1: Particle Size Distribution of Dry and Hydrated B9B10 Soil
Drainage Ditch B9B10 Sediment The screening process for the drainage ditch
sediment showed a phenomenon which had not been in evidence in the soil screening
study. As the sediment was subjected to its initial rough screening through a 1 cm screen
it quickly became obvious that there were large masses of filamentous growths present in
the 1 cm screen retntate. This growth contributed to screen blinding and caused a great
deal of difficulty in the initial screening process. Samples of the material were identified
by the University of Florida Center for Aquatic Plants as an algae of the genus lyngbya a
hardy filamentous algae which thrives in neutral to alkaline pH conditions and is
ubiquitous throughout Florida.
Lyngbya had been observed to be present to some extent at all ten of the test farms
participating in the BMP studies, in some cases as floating mats, so, in retrospect, the
presence of the material in the sediment should have been expected. The screening
process was carried out with particular care to insure that washing of the rough screen
with decanted subnate was adequate to release smaller size matter held within the


106
After refurbishment was completed a number of experiments were run in the CRAF using
kaolinite, kaolinite/bentonite and kaolinite/attapulgite clay and clay blends, which were
chosen to represent a wide range of erodability. These experiments, which are not
detailed here, were run for the purposes of checking flume operation and developing
technique and obtaining standard data that could be used in the calibration of the field
device.
There are two basic bed types that may be studied in the flume, the placed bed and
the deposited bed. The placed bed is made of material that is blended external to the
flume, placed in the flume by pouring or scooping, and leveled mechanically. After
placement, overlying fluid is added carefully to not disturb the bed. The placed bed is, in
theory, homogeneous in the vertical dimension and should exhibit uniform erosion
properties over an extended test period. The deposited bed is made by allowing initially
suspended sediment to deposit to the bottom of the flume, either under applied shear or
under quiescent conditions. The deposition and consolidation processes which prevail
cause a deposited bed to have non-uniform density, credibility, and possibly composition,
in the vertical dimension. Further, these characteristics may be influenced by the shear
stresses applied during deposition and by the self-weight compaction forces which prevail
after deposition. The deposited bed. then, has characteristics that are a function of bed
material, nature of deposition, and time of compaction since deposition. Both placed and
deposited beds were studied during the preliminary clay-bed tests.
Once the bed has been positioned, either by placement or deposition, the
experimental procedure is to start the bed and ring rotation at the speeds coinciding to the
lowest shear stress to be tested and drawing timed 50 ml samples from one or more of the
seven taps located at various heights on the side of the flume. Sample tap purge is
recycled back into the top of the flume, along with make-up water equivalent to volume
lost due to sample withdrawal. For long runs make-up water must also be added
periodically to compensate for evaporative losses. After the desired time has passed the


90
Figure 5.12: Adsorption Isotherm-B9B10 Sediment <38 Micrometer Range
Table 5.6: Sorption Parameters and Coefficient of Determination for WPBC and
B9B10 Sediments
Sample
Organic
Content (%)
S
(mg/kg)
sm
(mg/kg)
Ks
(mg/1)
k
(1/kg)
r2
WPBC
75-150 p
6.4
1.1
110.8
2.0
56.3
0.544
WPBC
<38 p
42.5
12.0
517.4
2.6
197.9
0.877
B9B10
75-150 m
79.5
89.1
495.4
7.8
63.9
0.944
B9B10
<38p
80.2
113.0
519.8
6.3
83.1
0.935


176
mg/1 for event 319 to a low of 21 mg/1 for event 258. There was not necessarily a
repeating pattern to the shape of the U. On some occasions it was roughly symmetrical
but on others it was skewed to either the left or right.
Particulate phosphorus content of suspended solidsFarm UF9200
Plots of the particulate phosphorus concentration as a function of time give curves
similar to Figures 7.3-7.13 and will not be shown here. What is more informative is an
evaluation of phosphorus content as a function of the suspended solids concentration in
the discharge. Figure 7.14 shows this relationship for the population of all samples taken
between Julian Dates 220 and 291 (which excludes the tropical depression data). Figure
7.15 is the same data plotted on an expanded scale to emphasize TSS concentrations
below 85 mg/1. The first two samples (first 4 hours of pump time) of each event have
been placed in one population. All other samples were placed in a second population.
Several things are readily apparent in Figures 7.14-15. There is a clear decrease in
phosphorus content of the discharged suspended solids as the suspended solids
concentration increases. The phosphorus content appears to decrease asymptotically,
approaching a value of 1200-1300 mg/kg at the highest discharge TSS concentrations. At
the lower TSS concentrations the phosphorus content appears to increase exponentially
with substantial data scatter as TSS concentration decreases, approaching values on the
order of 7,000-10,000 mg/kg at TSS levels in the 5 mg/1 range. The population
representing the first four hours of each event also appears to follow the same pattern, but
the average phosphorus content of the discharged suspended solids appears to be higher
than the phosphorus content for the remaining population by ratios of from 1.5 at TSS of
3 mg/1 to 1.9 at TSS of 80 mg/1.


158
optimum sampling schedule was to take one 125 ml sample for every 30 minutes of
pumping time, compositing four of these samples in one bottle, and then moving to the
next bottle. Under these circumstances each bottle represented a four sample composite
of two hours of pump time and the entire sample caddie of 24 bottles could hold samples
representing up to 48 hours cumulative pumping time. This had been found to be a
reasonable compromise that sacrifices some of the resolution of discrete sampling in
exchange for extended temporal sampling capacity. On occasions of intensive continuous
pumping however, particularly over weekends, the 24 bottle sample caddie may fill
completely before pick-up, so some sampling opportunities may be lost. In order to
maintain consistency with the existing program the same sampling procedure was
adopted for the particulate phosphorus transport study.
The adaptations made to the existing systems to accommodate the samples for the
particulate phosphorus transport study consisted of installing another sample head in
close proximity to the existing heads at each of the three locations. The new head was
connected to a separate ISCO sampler system, but the separate ISCO system was
controlled by the same CR-10 datalogger as controlled the sampler used for total
phosphorus sampling, so all samples were taken on the same cycles. Particulate
phosphorus sample pick-up was incorporated into the regular routine at the three sites so
all samples were picked-up on the same schedule. There was one significant difference
between the total phosphorus and the particulate phosphorus samples. Total phosphorus
sampling protocol for regulatory monitoring requires that the samples be acidified with
strong acid to retard biological activity. This acidification was affected by depositing 2
ml of concentrated sulfuric acid in each sample container prior to its placement in the
field. This procedure was deemed to be unacceptable for the particulate phosphorus
monitoring process because of the potential impact of the low pH environment on
solubilization of inorganic phosphorus compounds and hydrolysis of organic particulates.


Table 8.6: Phosphorus Fractionation Analysis of Large-Scale Composite Samples
Sample ID
Sodium Bicarbonate Extractable
mg/kg
Hydrochloric Acid Extractable
mg/kg
Refractory
mg/kg
Total P
mg/kg
Discrete
Samples
Inorganic
Organic
Total
Inorganic
Organic
Total
Organic
Inorganic
Organic
Total
Total
UF9206N
JD236
297
431
728
612
132
746
1056
908
1619
2527
2084
UF9206N
JD256
253
460
713
661
135
796
1102
914
1697
2611
1917
Average
275
446
721
636
133
770
1079
911
1658
2569
Std. Dev.
31
21
10
35
2
37
32
4
55
59
Coeff. of
Variation
11.2
4.6
1.4
5.5
1.6
4.8
3.0
0.4
3.3
2.3
% of Total
P
10.70
17.35
28.05
24.76
5.19
29.95
41.99
35.47
64.53
100.0
UF9200
JD237
582
749
1331
1017
206
1223
1994
1599
2949
4548
3649
UF9200
JD241
470
844
1313
1001
188
1189
1488
1471
2520
3990
NA
UF9200
JD252
741
668
1410
993
242
1235
2371
1735
3281
5016
4598
UF9200
JD258
578
1181
1760
1363
215
1578
2100
1941
3497
5438
5766 I
Average
593
861
1453
1094
213
1306
1988
1686
3062
4748
Std. Dev.
112
226
209
180
22
182
370
201
426
622
Coeff. of
Variation
18.9
26.2
14.4
16.43
10.52
13.94
18.58
11.9
13.9
13.1
% of Total
P
12.48
18.13
30.61
23.03
4.48
27.51
41.88
35.52
64.48
100.0
1


175
reached the low-level cut-off point and the pump went through repeated on-off cycles for
the next 12 hours until the system was switched to the small pump and ran in this
configuration for an additional 6 hours. At the end of this period pumping was stopped
for 16 hours and then resumed on the small pump, which went through several on-off
cycles that were of lesser frequency because of its lower capacity and the reduction of
ground water levels at this point in the drainage cycle. On Julian Date 224 43.5 mm (1.7
inches) of rainfall was received. The grower elected to continue pumping with the small
pump into Julian Day 226, at which time he decided that his groundwater levels were not
dropping rapidly enough and switched over to the large pump that ran for 12 hours,
drawing the canal level down to the point where it went into an on-off cycle mode for the
next 18 hours. All pumping was terminated the afternoon of Julian Day 227. Note that
this was a case where the end of the event outlasted the sampler capacity so samples were
not obtained for the last 28 hours of the event.
Inspection of the curves indicates that the solids export tended to follow a
consistent pattern. Typically the first 2 hour composite sample would contain a relatively
high concentration of suspended solids but the concentration would reduce in subsequent
samples to a minimum and then increase as the cycle approached pump shut-down,
yielding a U shaped curve of suspended solids concentration with time. The suspended
solids concentration then would typically decrease with some random variation through
the pump cycling mode and then continue at a relatively low level for the duration of the
event. Cumulative load analysis showed that the portion of the event encompassed by the
U shaped curve and the following pump-cycling period contributed to from 75% to
95% of the total event suspended solids load, depending on the specific circumstances
surrounding the event. There were, of course, some departures from this general pattern
but these departures usually appeared later in the event. While the general pattern
repeated itself through each event, the specific concentrations associated with the pattern
were variable. The first-sample concentration, for example, ranged from a high of 155


153
entire conveyance width, the floating macrophytes are quite mobile and can move
considerable distances under the force of, for example, a weak but steady wind. If there
are floating macrophytes on a conveyance channel, their absence at a particular spot does
not preclude their presence at that spot in the immediate past or future. At the time this
tentative conclusion was a surprise. It had been anticipated that the lyngbya would play a
strong role in determining erosion characteristics, but the results seemed to indicate that
this was not necessarily the case and, if anything, the fresh lyngbya observed in sample
UF9200-3 had almost no erosive characteristics.
The observations made in these tests were the first indication that the presence of
floating macrophytes might have a major effect on the erosion characteristics of the
underlying sediments. The test results themselves provided information that would allow
order of magnitude estimates to be made of erosion contributions once the hydraulic
characteristics of the systems were determined.


188
Correlation of TSS Load with Hydraulic Load
Farm UF9206S
60000
50000
J? 40000
c 30000
£ 20000
10000
0
0 200.000 400.000 600.000 800.000 1.000.000 1.200.000 1.400.000
Event Hydraulic Load nr
O Events 209-281 Events 300-336
Figure 7.19: Correlation of Event TSS Load with Event Hydraulic Load for Station
UF9206S
Correlation of Particulate Phosphorus Load with Hydraulic Load
Farm UF9206S
200.000 400.000 600.000 800.000 1.000,000 1,200,000 1.400,000
Event Hydraulic Load mJ
O Events 209-281 Events 300-336
Figure 7.20: Correlation of Event PP Load with Event Hydraulic Load for Station
UF9206S


184
Table 7.4: Summary of Pumping Events for Station UF9206S
Event
Duration
(hrs)
Pump Time
(hrs)
Interevent
Time
(hrs)
Pumped
Volume
(m3)
Suspended
Solids Load
(kg)
Particulate
P Load
(kg)
209
31
31
Start
0.77x10s
418
2.62
213
18
18
71
0.39x10s
266
0.90
218
44
44
106
1.31x10s
532
4.80
224
68
68
94
2.04x10s
663
4.39
234
38
29
178
0.70x105
291
2.20
239
97
66
79
1.26x10s
1633
4.53
251
65
57
191
1.20x10s
3385
7.92
256
27
26
47
0.60x10s
3145
4.11
260
137
116
76
2.16x10s
3978
7.37
268
94
89
50
1.57x10s
1687
5.70
273
177
169
24
1.97x10s
6714
9.60
280
208
130
24
2.57x10s
44312
50.50
300
149
86
243
2.65x10s
11970
21.90
317
353
307
262
13.20x10s
48654
120.60
336
143
89
96
2.64x10s
10841
26.90


69
Selection of Primary Target Farm
Availability of resources at the time of initiation of the field portion of this
research dictated that the intensive field activity be restricted to one target farm, with the
understanding that the results from the target farm research would serve as a basis for
determining further research work needed to generalize the model. The selection criteria
were limited to a few important points, which were:
1. The target farm should be one of the ten farms participating in the IF AS
Best Management Practices Program.
2. It should have a relatively straightforward drainage layout and hydraulic
management program.
3. It should, as the prototype farm, be restricted to a sugarcane-only planting
on typical soil type and depth. At the time sugarcane represented
approximately 70% of the total agricultural acreage in the EAA.
4. It should be a medium-to-large size farm that practiced at least average and
preferably aggressive crop and water management policies.
5. The grower should be willing and able to provide access, assistance, and
information on a regular basis.
Farm UF9200 met these criteria and was chosen as the target farm for more
intensive study.
Particle-Size Property Distribution Study
The results of the sediment survey represented on-farm particulate transport
potential. The data set from the SFWMD (Mucinic 1994), discussed in Chapter 2,
represented actual transported material sampled at EAA discharge points. A comparison
of representative particulate phosphorus contents showed the on-farm sediment to
average 617 mg P/kg solids, while the selected data set from the SFWMD implied that


APPENDIX F
DIMENSIONS OF UF9200 USED IN DUFLOW FORMAT
Following is the dimensional data from farm UF9200 used in the DUFLOW
program. For convenience it is presented in the DUFLOW file format used to specify the
nodal layout (*.nod) and the channel and structure dimensions (*.net). The specific
example files shown have a k value of 17. where k is the reciprocal of Mannings
n.
STN17.NOD
The NOD file is in the format
Node No...X Coordinate (m)...Y Coordinate (m)...Area drained(m2)... RunoffFactor
* DUFLOW data file :220\STN 17.NOD
* Network data program version: 2.02
+F1
1
180.0
3017
1609
0E+00 1.00
2
3017
817 2428E+02 1.00
3
3017
805
OE+OO 1.00
4
3017
792
1214E+02 1.00
5
3017
12
1214E+02 1.00
6
3017
0
0E+00 1.00
7
2816
1609
0E+00 1.00
8
2816
817
1619E+02 1.00
9
2816
805
0E+00 1.00
10
2816
792
8090E+01 1.00
11
2816
12
8090E+01 1.00
12
2816
0
0E+00 1.00
13
2615
1609
0E+00 1.00
14
2615
817
1619E+02 1.00
15
2615
805
0E+00 1.00
373


204
Two of the four event studies had sampling difficulties that resulted in some
information truncation. Event 262 had been preceded for about a three week period by a
number of intense convective storms throughout the EAA. The backlog resulting from
the influx of samples from the ten study farms caused a sudden shortage of the specially
configured ISCO sample bottles and containers. The program analytical manager was
forced to ration sample containers so the sample program for Event 262 was reduced to
meet the reduced availability of equipment.
During Event 285 two samplers encountered mechanical difficulty. A large mass
of floating macrophytes collected on the vertical support arm of the sampler at location
#4, presenting a large crossectional area normal to the direction of flow. The moment
arm that resulted tore the sampler from its base mooring and caused it to float on the
surface loosing sampling capability after the 32nd hour of the event. The sampler at
location #3 suffered an internal structural failure that caused it to jam after the 24th hour
of the event. Given these qualifications on the data set, it is instructive to view the four
synoptic event studies for the trends they may exemplify.
Ditches vs. discharge
The first important observation is that the suspended solids concentration in the
discharge from the field ditches was much lower than the suspended solids concentration
in the discharge from the farm pump station. The field ditches were not monitored 100%
of the time during all synoptic event studies, but if we compare measures of central
tendency during periods when both ditch and farm discharge data are available we see
that the ditches TSS discharge concentration had an average of 4.0 mg/1 TSS (a = 4.3),
the farm TSS discharge concentration had an average of 20.0 mg/1 TSS (a = 24.9), a
factor of 5 difference between ditch and farm discharge.


172
Time Series-Flow and Total Suspended Solids
Event UF9200-262
Pumping Rate O TSS Concentration
Figure 7.9: Hydraulic and Total Suspended Solids Profiles for Event UF9200-262
Time Series-Flow and Total Suspended Solids
Event UF9200-285
2.0 160
Pumping Rate OTSS Concentration
Figure 7.10: Hydraulic and Total Suspended Solids Profiles for Event UF9200-285


364
Phosphorus Sorption/Desorption Measurements (Olila and Reddy, 1993)
Sediment samples of known solids content were treated with precalculated
standard phosphorus solutions of KH2PO4 with initial P concentrations of 0, 0.06, 0.2.
0.6, 1, 2, 6, 10, 20, and 60 mg/1. The solid:solution ratio (w/v) was maintained between
1:30 and 1:40. The treated sediment samples were equilibrated in an end-to-end shaker
for 24 hours, after which they were withdrawn and centrifuged. Supemate was decanted
off, filtered through a 0.45 pm membrane filter and then stored at 4C until it was
analyzed for orthophosphate by APHA Method 4500-P E. Centrifuged residues were
resuspended in phosphorus-free water, again agitated for 24 hours, and then centrifuged
and decanted again. Supemate was again decanted, membrane filtered, and analyzed for
orthophosphate, this time to study the desorption characteristics of the sediment solids.
Phosphorus Fractionation (Ivanoff. 1994, APHA, 1989)
Sediments were sequentially extracted with 0.5 M NaHC03 for 16 hrs.,
centrifuged and extracted with 1 M HC1 for 3 hrs.. then centrifuged for final separation.
The extracting liquids were split. One half was analyzed directly for inorganic P by
APHA Method 4500-P E., the other half was digested by the sulfuric acid-persulfate
method (Method 4500-P B.5. APHA, 1989), and analyzed for total P by APHA Method
4500-P E. Organic P was calculated by difference. The final residue was analyzed for
total P (assumed to be refractory organic P) by the ignition method of Anderson (1976)


189
appropriate now to examine the individual events for each station and their internal
correlations.
Suspended solids transport Farm UF9206
Figures 7.21 through 7.27 present the event time series of TSS and flow for
UF9206N, Figures 7.28 through 7.32 present the event time series of TSS and flow for
UF9206S. In the case of both of these stations multiple events are presented on each
chart for convenience because detailed event-by-event analysis will not be conducted for
farm UF9206. Several general points should be observed on these charts.
First of all note that there are dramatic changes in scale on the TSS axis among
events, which reflects the extremely large range of discharge suspended solids observed
over the course of the study. TSS at UF9206N ranged from a minimum of 3.7 mg/1 to a
maximum of 1894 mg/1, TSS at UF9206S ranged from a minimum of 0.7 mg/1 to a
maximum of 1474 mg/1. This is in comparison with a range of 0.3 mg/1 to 291 mg/1 for
UF9200.
The event-wise patterns for UF9206S showed the TSS concentration starting high
and decreasing in some, but not all, of the events. There was some evidence of a U
shaped pattern in some events but no clear repetitive pattern across events was evident.
The suspended solids load, unlike UF9200, did not show a consistent pattern of decline as
the events continued.
The event-wise patterns for UF9206N seldom showed the TSS starting off high
and reducing. In fact there typically tended to be an increase in TSS concentration for the
first part of an event. Any U shaped patterns appeared to be random in time. There
were approximately as many convex U patterns as there were concave which is a good
indication of randomness. Here also there was no indication of a consistent pattern of
suspended solids load decline as the events continued.


Copyright 1996
by
James Donald Stuck


1%
In general the temporal patterns of flow and suspended solids at both stations of
UF9206 showed less organization and more variation than was observed at UF9200.
It will be shown in Chapter 11 that some of this apparent random behavior was related to
the larger selection of pumping rates available from the variable speed pumps and to the
absence of level control at either station at UF9206. At this point, however, it is more
appropriate to pursue the relationships between phosphorus content and discharge
suspended solids at UF9206.
Particulate phosphorus content of suspended solids Farm UF9206
The phosphorus content data from both UF9206 stations showed considerably
more scatter than was observed at UF9200, when examined over approximately the same
time period, Julian Dates 209-291, which covered the normal wet season data and
excluded the tropical depressions. Figures 7.33 and 7.34 show the phosphorus content vs.
TSS data on the whole-data-set basis respectively for UF9206N and UF9206S. Figures
7.35 and 7.36 show the same relationships on an expanded-scale to emphasize the bulk of
the data that represents the lower TSS concentrations.
In spite of the data scatter the same general trend was obvious. Figures 7.33 and
7.34 clearly show the decreasing phosphorus content with increasing discharge TSS and
the asymptotic trend at higher concentrations for both stations. Note again the much
larger number of high TSS concentration occurrences at UF9206 than at UF9200 (Figure
7.14). Figures 7.35 and 7.36 illustrate the scatter in the data, especially for UF9206N, the
more hydraulically active of the two stations. Attempts to extract a significant
discrimination between the first several samples of an event and the rest of the event data
were unsuccessful for both stations, most likely due to the additional variables that were
introduced by the more complex operating conditions at UF9206. The data for both
stations were best correlated using an asymptotic power function model .


Soluble Phosphorus (SP)
SP and F^roculac
SP and FP fian surface
on soil water
Riccphoms (FP) fian ditch
runoff
as direct preapttaiicri. wind
wall sal. ditch sediment.
(Pen ode Excursions!
T
and decaying plant marta in
ditch
1
etc
(Secaidarv Contn buttons)
T
T
T
tisal type, water table,
If ditch dtmaistans aid
flsal type, crop cover.
degree of rmneializanai.
water level, development
culdvaiai practices. rainfall
redox potential, antecedent
and maintaiance history.
intensity, wata table
canil tier is and usage p.
conditions....)
sal type, rainfall intensity .
location, damage svstem
content in preapitation. sal
hydrodynames, sedunent
histay and transport
characteristics.)
configurations and
managematt practices, field
heterogenaty...)
moisture..)
QamisDitdiWM Drainage DitdiWM
Soluble Ftaphems Mnasluancnaxl panedate Phorrhoms
(SB g*11' m
L
Low- Riasphaus Dlution Waa fian
Ctamage Dtdi Wbter
damage Dtch Sedment
(see above)
Carbonate Rock Interflow
T
Farm Canal
V
Sedment Rae Wfiter
Soluble Rtasphcrus
FarmCaial Water
Efecaying
T
A
Mater
Soluble Riosphonis
1
Macrophytes
ahen ancr Adsorption
rpdon
Farm Canal
Will
Particulate RxsphctiLS

Rtytoplankton
T^fcuspenstan
Sedmaifa(gi
Farm Canal
Sediment
I
Repeal Farm Canal Processes
CtainageDstnct
Canal Wter
Rqoeat Fam Canal Ftaxesses
i
Repeal Pam Carat Recesses Mnagonncismci
Canal Waa
1
ExilEAA
Rqxat Fam Canal Processes
Figure 4.2: Farm Scale Phosphorus Transport Simplified Conceptual Model


95
Preliminary Conclusion Regarding Environmental Impact of Sorption The
preliminary conclusions drawn from the results of the adsorption/desorption studies run
on these selected sediments were as follows:
1. The sediments, particularly the smallest particle-size fractions that are high
in organic content, have exchangeable phosphorus contents on the order of
50-150 mg/kg, which typically represents well less than 10% of the total
sediment phosphorus.
2. The phosphorus partition coefficients exhibited by these sediments are on
the order of 75-200 1/kg, which indicates weak binding of phosphorus.
3. Under the conditions expected to prevail normally in sediment
resuspension, the contribution of pore water and desorbed phosphorus to
the system phosphorus would be on the order of 5% of the soluble
phosphorus and less than 4% of the total phosphorus.
4. Given conditions similar to those that these studies represent, the mass of
phosphorus sourced from pore water resuspension and desorption from
suspended particulates is negligible compared to the mass of soluble
phosphorus that is present in the groundwater and the mass of particulate
phosphorus which may be contributed by sediment resuspension.
5. Under these circumstances a first approximation environmental model,
which is what was to be developed in this work, may proceed by ignoring
the processes of adsorption and desorption, so long as the assumptions
made here are checked against subsequent measurements.
Sedimentation parameters of B9B10 sediment
Sedimentation Velocity At this point in the project there was virtually no
information available on the settling rates of highly organic particulate sediments such as


219
The phosphorus content analysis of the samples is shown in Table 8.1 for the field
ditch samples and Table 8.2 for the canal samples.
Table 8.1: Field Ditch Surficial Sediment Phosphorus Content UF9200
Synoptic Survey of Julian Date 229
I Field Ditch Designation
South End Sample
mg P/kg TSS
Midlength Sample
mg P/kg TSS
North End Sample
mg P/kg TSS
South Side
Bl-2
NA
NA
617
B3-4
NA
NA
492
B4-5
NA
NA
850
B6-7
NA
NA
406
B9-10
NA
NA
538
B12-13
NA
NA
638
B14-15
NA
NA
750
North Side
A1-2
517
427
324
A4-5
890
NA
NA
A6-7
383
323
471
A9-10
460
316
465
A12-13
471
NA
NA
A14-15
492
344
478


132
CSME =-2224 x 10-+4.21 Ox 10"-2.780 x 10'12x',,s 6.27
where
x = hV2
The correlation is reasonably good (r2 =0.987) and correlates the data much better
than its original presentation in Figure 6.12. Equation 6.27 applied to the 15.2 cm depth
case (where incipient erosion was assumed to be observed) gives a value of critical
erosion velocity of 29.8 cm/sec. Considering the assumptions made to develop the
derivation, the approximations involved in estimating asymptotic CSMEs, and the data
scatter observed in several of the runs, the resulting correlation rationalizes the observed
depth differences and provides a very acceptable tool for the analysis of the erosion
observed in the flume and its translation to the field device.
Organic Sediment Particle Entrainment Simulator Studies
Particle Entrainment Simulator configuration
The device chosen for the field studies was the Particle Entrainment Simulator, or
PES, which was discussed in Chapter 3. The PES, shown schematically in Figure 6.14
and pictorially in Figure 6.15, is of simple construction, consisting of a 30.5 cm (12 inch)
tall cylindrical chamber with inside diameter of 11.75 cm (4 5/8 inch) and outside
diameter of 12.7 cm (5 inches). Inside the cylinder an 11 cm (4.33 inch) diameter
horizontal grid is suspended by a linkage bar through the cylinder top to an external cam.
The grid oscillates vertically, producing homogeneous turbulence at the sediment-water
interface. The grid is a 0.6 cm thick disc of Plexiglas that is perforated with 1.2 cm
diameter holes spaced 1.5 cm center-to-center. The cam connection is offset by 1.27 cm


229
piston core sampling technique, which was designed to sample reasonably well
consolidated sediments. In other circumstances this kind of light, flocculant material
would be classed as wash load which is typically not considered to be part of the
sediment (Raudkivi 1976, Parthenaides 1977).
Under turbulent conditions the floating macrophyte mat and submerged algal
growth could present a significant flow resistance and be subject to shear forces that
would accelerate the detachment process, contributing additional biological material to
the suspended solids load. It seems reasonable to assume that the material dislodged
under quiescent conditions might differ in physical, chemical, and biological
characteristics from the material dislodged under the more forceful flow conditions. For
example, if the material released under quiescent conditions were more senescent it might
be expected to have a lower phosphorus content than more viable material released under
more vigorous hydraulic conditions. The age of the plant and the time of the season
would also be expected to affect dislodgability and phosphorus content.
The studies reported in this section were not designed to confront all the elements
of this hypothesis. Instead the primary objective was to determine plausibility of the
hypothesis by estimating the maximum amount of dislodgable material present in the
surface macrophyte matrix and evaluating the phosphorus content of the plant and its
detritus.
Water lettuce (Pistia stratiotes) was chosen as the representative macrophyte for
this study because, at UF9200, it was the predominant species at most locations. The
physiological configuration of water lettuce also lent itself best to the harvest technique
that was required for this study. Individual plants were harvested by carefully moving a
submerged plastic container under an individual plant and then slowly raising the
container, isolating the plant and its root structure, severing connections to adjacent plants
if necessary, and then capping the container for transportation. Samples were taken from
clusters and masses as well as from free-floating plants. It was found very quickly that


234
by the particulate material would have negligible effect on the total phosphorus content of
the farm discharge. This conclusion then led to the important corollary that sorption
dynamics could be ignored in the first approximation model to be developed in this work.
At the time the first sorption work was done, plans were made to check the sorption
characteristics of material from actual farm discharge. The hypothesis that a large
fraction of this material is of in-conveyance plant growth origin adds to the necessity of
verification of the original assumption.
The primary objective of the Large Composite Sample Studies was to obtain
enough representative material from several events to allow a credible estimation of
sorption characteristics. A secondary objective was to conduct some phosphorus
fractionation tests on the discharge suspended solids, the results of which might be useful
in determining both the source and fate of the exported particulate phosphorus.
The collection process was very straightforward. A pair of 100 liter
polypropylene tanks, equipped with bottom drain valves, were set up at both the UF9200
and the UF9206N pump stations. Each pair of tanks was coupled with a self-contained,
battery driven ISCO sampler, which was programmed to take a five liter sample evety
half-hour from a temporary sample head that had been placed adjacent to the sample
heads used for the discharge time-series sampling program. Both tanks were sealed
against intrusion of airborne particulates.
At the start of an event these samplers were started manually and proceeded to
sample on the same frequency as the time-series samplers, but offset by 10 minutes to
avoid the possibility of interference between the two sample heads competing for sample.
During the sampling process the sites were visited on a regular basis, the sample delivery
tube was switched to the off-line tank, and the collected volume in the previously on-line
tank was drained into a polypropylene carboy for transportation to the EREC lab, where it
was added to a polypropylene sedimentation vessel for separation at 4C. The collection
period was 24 hours or the duration of the pumping period, whichever was less.


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
PARTICULATE PHOSPHORUS TRANSPORT IN THE
WATER CONVEYANCE SYSTEMS OF THE
EVERGLADES AGRICULTURAL AREA
By
James Donald Stuck
December, 1996
Chairman: Kenneth L. Campbell
Co-Chairman: Konda R. Reddy
Major Department: Agricultural and Biological Engineering
The Everglades Agricultural Area (EAA) is a source of phosphorus nutrients to
the downstream watershed, including Everglades National Park. This work considered
export of particulate phosphorus (PP) from large EAA farms that are extensively drained
by pumped canal systems.
The research objectives were to determine the major sources, locations, and mode
of transport of PP, develop a model for PP transport and apply it to a representative farm,
using an existing hydrodynamic model adapted to include subsurface drainage.
Representative sediments had organic matter and phosphorus contents of about
75% and 0.1%, respectively, with weak phosphorus adsorption characteristics. Typical
farm canals contained sediment phosphorus equivalent to 15 to 30 years of discharged
PP. Sediment transport properties, studied in an annular flume and a particle entrainment
simulator, were used to develop a prototype model that included the concepts of a critical
xxvii


324
Eliminating the cycling of the large pump and replacing it with a period of
quiescence followed by operation of the small pump had a particularly noteworthy effect
on solids transport. This action caused the simulated velocity to be below critical for the
remainder of the simulation period, so the model predicted no additional solids transport.
The net effect of the two modifications of pumping practice was to reduce simulated net
solids export from 10,420 kg for the original simulation to 4370 kg for the subject
simulation.
Effect of Wider Discharge Outlets (Simulations B and Cf
The system calibration process involved configuring the sections closest to the
pump such that the section flow area narrowed to approximately the diameter of the
discharge culvert in the North Canal or the pump inlet in the East Canal. These


174
8 1.5
"e
Time Series-Flow and Total Suspended Solids
Event UF9200-355
90
80
70 -
60 ¥
30 c
355 356 357 358 359 360 361 362 363 364
Julian Date
Pumping Rate O TSS Concentration
Figure 7.13: Hydraulic and Total Suspended Solids Profiles for Event UF9200-355
The hydraulic curves presented in Figures 7.3-7.13 represent the result of this
operational philosophy. The typical event starts with the canals at a high level and the
high capacity pump running at a relatively low head differential close to maximum
capacity. As the canals are pumped down and level in the receiving sump increases
because of a rise in the West Palm Beach Canal (because of multi-farm pumping) the
head differential increases and the pump capacity decreases. Eventually the level at the
pump intake reaches the low level cut-off point and the pump shuts down. After this the
high capacity pump may go through several on-off cycles until the system is switched
over the lower capacity pump. If there is rainfall during the pumping event then the
profile shifts to reflect this increase in water supply.
Event 220 provides a somewhat extreme but informative illustration of this
process. Antecedent conditions were 56 mm (2.2 inches) of rainfall on Julian Date 219
(See Appendix E for rainfall and level data for the study period). The canal level at the
pump station started at 2.86 m when the large pump was started. After 14 hours the level


39
The basin as a whole acting as a phosphorus sink during a dry year, but
becoming a net source the following year, which was wetter than average
Mass balances showing an increasing bias toward underestimation of
phosphorus export as frequency of sampling decreased
The authors rightly noted that current understanding of non-point P sources and
the routing of P compounds within watersheds is very poor.
A Brief Discussion of Sediment Transport Theory
Non-Cohesive sediment transport
Traditional (non-cohesive) sediment transport places mobilized sediment in
various categories, which have been noted earlier in this chapter. Grains making up a
substantial part of the movable bed of a stream are called bed material. Bed material
moving within a few grain diameters of the bed is known as bed load. Material in
suspension which is not present in any quantity in the bed (usually very fine sediment) is
known as wash load. Bed material swept up into the main flow stream by turbulence
becomes part of what is referred to as the suspended load. The suspended load is the
combined mass of this fugitive bed material and the wash load. The relative proportions
of the bed material that move as bed load and suspended load depend on the
characteristics of the bed material, such as size and density, and the flow conditions
(Middleton and Southard, 1984).
The analytical approach taken in non-cohesive sediment dynamics treats bed load
and suspended load as two separate entities, subject to two different sets of physical
forces. Typical equations for bed load horizontal flux incorporate the difference between
a critical shear stress for incipient motion and the actual bed shear stress as a driving


268
surface. The potential mobility and temporal variability of the macrophyte and algal
population density and its detrital phosphorus content introduces uncertainty into the
source locations and quantities. The probability of multiple populations with distributed
erosion and sedimentation characteristics geometrically increases the amount of sampling
and testing required to develop an explicit distributed parameter particulate phosphorus
delivery model.
The obvious increase in complexity and the objective to adhere to a relatively
simple first approximation level model at this point in the program led to the decision to
develop a physically-based but lumped-parameter transport model. Instead of using
explicitly determined erosion and sedimentation rate functions, the erosion and
sedimentation parameters would be fit to the existing data as calibration parameters. In
essence the model extracts the key qualitative elements of the performance of the
prototype surficial sediment without incorporating the actual transport rates measured for
the prototype.
Basic Precepts in the Model Development Several precepts were followed which
set limits on the development of the final model. They are as follows:
From 75-90% of the total particulate phosphorus exported from the target
farm during an event occurred during the first major pump cycle. The
model development should concentrate on fitting this portion of the events
with lower priority given to later times in the events.
The potential complexity of the population suggests a multi-component
model but the lack of explicit data on the elements of the population
mediate against such a model at this time. A multi-component model,
which undoubtedly could be made to fit well, would be much more
speculative than a single-population model, therefore the model
development should be restricted to a single lumped biological population.


9
and are estimated to cost on the order of $320 million (Whalen et al., 1992). Cost of
construction of these facilities will be borne, in part, by assessments on the EAA
landowners. The current STA design goal is to achieve an average effluent total
phosphorus concentration of less than 0.05 mg/1 (Whalen et ah, 1992). The design of
each of the four STAs is based on treating a long-term average load from the sub-basin
within the EAA served by the STA. This average load is determined from historical data
and corrected for land removed from production for the STAs and for expected reduction
in phosphorus loading resulting from grower implementation of BMPs.
Implementation of the combination of BMPs and construction and operation of
the STAs was projected in 1992 to reduce phosphorus input to the WCAs by 80% within
five years after initiation (Whalen et ah, 1992).
Areas of Uncertainty in Remediation Plan
The possibility exists that the actual input to the STAs may be lowered below the
long-term historical averages by more than 25% by implementation of the BMPs
recommended by Bottcher and Izuno, by implementation of district-wide water
management practices not yet specified, and by exploitation of biological and chemical
processes in the irrigation ditches and drainage canals. The possibility also exists, with a
somewhat lower probability, that release of phosphorus currently stored in the canal and
drainage ditch sediments may tend to buffer the exit phosphorus concentrations from the
various sub-basins and reduce the impact of the BMPs on the phosphorus load to the
STAs. Significant changes of loading in either direction could have a substantial impact
on the amount of land required for the STAs, or, alternatively, on the effluent phosphorus
concentrations achieved by the STAs.
For any given time period, weather is always an uncertainty. Measurement of the
effectiveness of any specific remediation strategy may be hampered by changes in the


254
Field drainage model
The Field Drainage Model was developed by Dr. Nigel Pickering of Soil and
Water Engineering Technology to allow adaptation of DUFLOW to the subsurface
drainage mode practiced in the EAA. The model has been discussed at some length by
Bottcher and Pickering (1995). What follows is an adaptation of that discussion.
A simple, rapidly executable model was developed to simulate the drainage into
each farm ditch in response to the hydrological inputs of rainfall and evapotranspiration
and the water levels in the fields and the field ditches. The model was developed using a
quasi-steady state approach, allowing the use of a steady-state algorithm that incorporated
the Hooghoudt equation (Raadsma 1974). The Hooghoudt equation, expressed in
drainage flow form is
g_44,(2*zyW) 94
s2
where
q = Drainage flow, m3 /sec- m2 of subsurface flow area.
Ksat = Saturated hydraulic conductivity, m/sec
m = water level difference between mid-field and ditch, m
Deff = Effective drain depth, m
s = ditch spacing, m
Equation 9.4 is modified slightly for use in the model by defining the effective drain
depth as the water elevation in the ditch plus the effective drain depth. The effective
depth of the ditch, which corrects for convergence effects as groundwater flows approach
the drainage pipe or ditch, was calculated using published correlations (Van Schilfgaarde
1974). Because the ditch width was much greater than typical drainage tiles the effective
depth was almost identical to the depth to the impeding marl layer.


25
plentiful organic carbon supply it is intuitively expected that the regions channels are
highly autochthonous.
Surficial agricultural residue will be a highly variable function of land use and
management practices, but dairy farm soils in the area of interest may have typical
phosphorus content of 1500-3000 mg/kg (Reddy et al., 1994). Tree leaves may contain
500-2500 mg P/kg dry weight, while woody litter may have a phosphorus content of 200-
300 mg P/kg (Reddy and DeBusk, 1987).
Macrophytes play multiple roles in the particulate phosphorus dynamics of sub
tropical streams. Because of their rapid turnover times they serve as both sinks during
growth and sources during senescence for particulate and soluble phosphorus. For
illustration, dense stands of water hyacinths may assimilate 100-300 mg P/square meter-
day by biological fixation (Reddy and DeBusk, 1987) and slough 5-15 mg P/square
meter-day from normal detrital formation (DeBusk and Dierberg, 1989). Typical floating
macrophytes may have tissue phosphorus concentrations of 1500-12,000 mg P/kg dry
weight (Reddy and DeBusk, 1987) so, unless the macrophytes are removed from the
system, the phosphorus stored in their tissue is available for release upon plant death and
senescence.
By creating zones of hydraulic quiescence macrophytes may affect sediment
deposition under normal flow conditions, but a portion of this sediment may be
remobilized during times of increased hydraulic activity (Svendsen et al., 1995). In sub
tropical and tropical climates macrophytes, particularly floating macrophytes, may host
substantial masses of loosely attached epiphytic microbial material, which effectively
increases the in-stream concentration of organic phosphorus-containing material (Engle
and Melack, 1990).
Microbial activity can play a major role in the particulate phosphorus dynamics of
a water body. In eutrophic lakes, for example, phosphorus associated with deposited
bacteria can equal or exceed that contributed by organic detritus (Gachter and Meyer,


BMP development
1. There should be a program initiated to develop optimum pumping policies
that would minimize maximum channel velocity without significantly
sacrificing the reduction of field water table level. This program could
have concurrent modeling and field verification components to allow for
real time interaction between field data and model simulation. One might
envision microprocessor control of pumping rates with feedback from
field and channel level monitors and velocity probes.
2. Hydraulic mining and transport looks promising. More detailed
simulation and pilot testing should be undertaken to determine the
feasibility of implementing this practice with minimal additional
investment in structures and pumps.
3. Pilot programs should also be undertaken to test the feasibility of weirs or
floating pump intakes to reduce pump-region velocities and reduce or
eliminate the first flush of suspended solids at pump start-up.
4. Focus should be placed on inexpensive means of macrophyte control
and/or removal. Hydraulics may play an important role in prevention of
propagation of macrophytes and also in optimization of collection and
removal schemes. Plant growth regulation and integrated pest
management may be worthwhile approaches to pursue.
5. Pilot studies should be conducted on sediment trap designs that affect a
significant reduction in channel velocity.
6. Existing canal depth/phosphorus discharge data should be evaluated at
sites where such is available to determine if there are situations where
modest controls of level could affect meaningful reductions in exported
phosphorus.


127
in the water column, gm/cm
Wj Sedimentation velocity of particles in class
cm/sec
Now, because there is no significant sediment removal from the flume, the
concentration of particles of class / in the water column is expressed by a simple mass
balance that says that the suspended mass is equal to the original mass minus the
remaining unsuspended mass, or
(em,(0)-em\a
Cj = i -U- 6.14
V
where
A = area under water column, cm2
V = volume of water column, cm3
and EMj(O) and EMj are as defined before,
but A/V =h, so
Cj = J d 6.15
h
Now a differential equation may be written for the change in sediment
mass in class as follows:
d(EM:)
dt
W.
- E(T -
6.16
At steady state (attainment of asymptotic CSME), the left-hand side of
Equation 6.16 becomes zero. Also, by definition


320
with selected removal at key downstream points. The effect of this action may be
exemplified in the course of simulating the action of a sediment trap.
Effect of Sediment Trap
For the purposes of this simulation it is assumed that the sediment trap design is
such that the cross-sectional area of the channel is not altered enough to have a significant
effect on the channel velocity. The sediment trap discussed in the Hutcheon Engineers
(1995) report had a length of 1600 ft. (-488 m). For this evaluation that type of trap was
simulated by setting the concentration of Cem(o> equal to zero for 500 m upstream from
the pump station in both the North and East canals of UF9200. This is equivalent to
removing all transportable sediment and biological growth in the stated canal sections.
Figure 11.2 shows the results of this simulation compared to the simulation of the original
event, cast in the format of cumulative TSS load versus cumulative water discharge.
Figure 11.2: Simulated Effects of Sediment Trap on Event UF9200-220


199
The correlation curves, shown as solid lines in the figures, represent the
correlation equations
PCV =15653 x(TSS)152 +857 15
PC, = 23773 x (TSS)-''6 + 1 163 7.6
where PCN = Phosphorus content of discharged TSS for station
UF9206N, mg/kg
PC, = Phosphorus content of discharged TSS for station
UF9206S, mg/kg
TSS = Discharged suspended solids concentration, mg/1
and the constant term in each equation represents the best fit asymptotic phosphorus
content for each data set. The higher pre-multiplier and the higher constant term for
UF9206S over UF9206N are consistent with the higher event-correlated phosphorus
content calculated earlier.
It is evident in Figures 7.34 and 7.35 that there are a number of data points that
fall below the asymptotic value for each location. Part of this is simply a result of the
statistics of the curve-fitting process, but a detailed examination of the data showed that
the majority of the points that are low-phosphorus-content deviants for both locations
were concentrated in time. At both UF9206N and UF9206S there was a concentration
centered around Julian Date 280, at UF9206N there was an additional concentration
around Julian Date 240. During the approximate period Julian Dates 268-292 the rice
fields, which were bordered by both the North and South canals, were being drained and
harvested. Detailed records of internal pumping events were not kept by the grower, but
it is likely that this process contributed suspended solids that had been contained in the
rice fields and were part of a different population than the typical ditch and canal solids.


76
filamentous matrix. The retained greater than 1 cm material, including the filamentous
algae was dried as one sample, but after drying material that was obviously lyngbya was
separated manually by visual inspection under magnification and treated as a separate
sample. The algae rapidly lost its visual identity with decreasing particle size. Under
microscopic examination at particle sizes of less than 1180 micrometers (-16 USS) it was
difficult to discern coherent entities within the sediment matrix that could be identified as
lyngbya fragments. The remainder of the screening runs at sizes less than 1 cm
proceeded normally.
After final separation it was determined that the algae that had been retained on
the 1 cm screen constituted approximately 15% of the total dry mass of the drainage ditch
sediment. The soil and hydrated soil samples were prescreened to remove all material
greater than 10,000 micrometers. For the purposes of particle size comparison, the mass
distribution in the ditch sediment was calculated on the basis of the 10,000-micrometer-
and-less fractions only. Figure 5.2 shows the comparison of distributions between the
hydrated soil and the drainage ditch sediment. It is evident here that in terms of
potentially transportable matter, the particle size distributions of the two samples are
quite similar. The mass mean particle sizes of both samples are in the range of 240-250
micrometers and the distributions of the lower 50th percentile of particle sizes are almost
identical. In the upper 50th percentile of particle sizes the drainage ditch sediment is
more heavily weighted toward the larger particle sizes. The fraction of most readily
transportable material, the <38 micrometer size, was 10.4% of the total mass considered.
West Palm Beach Canal (WPBC) Sediment There was no evidence of
filamentous algae in the WPBC sediment; in fact the WPBC sediment coarse fraction (>
1 cm) was very different in visual appearance from the farm sediment. The major
constituent of this fraction appeared to be shells and shell fragments of fresh water
mollusks and some few limerock and marl fragments. The wet screening of the WPBC
sediment proceeded normally.


285
differences between observed and simulated suspended solids concentrations toward the
end of the first major pump cycle. This is the point where a solids discharge from the
North Canal was observed during the intensive synoptic studies. In general, however, the
other five events show good agreement throughout the major pump cycle and into the on-
off cycles, as long as the large pump was in operation.
The model does not simulate the response well when the small pump is in
operation. This is exemplified in Figures 10.1,10.4, and 10.5 where it may be seen that
there are times when the simulation predicts a negligible suspended solids concentration
while non-zero concentrations were actually observed. These times correspond to times
of operation of the small pump. The small pump has a capacity roughly one-fifth that of
the large pump. During simulation of its operation the critical velocity for erosion is
seldom, if ever, reached so the model predicts net transport of zero, when in fact some
suspended solids are still being transported. Reduction of the critical velocity to
accommodate the transport taking place in this flow mode caused an unacceptable
upward distortion of the trough of the U curves during the major pump cycle. In other
words, modification of the critical velocity to simulate low concentrations at low flows
caused over-prediction of low concentrations at high flows, which was obviously a non-
optimal trade-off. This is one of the limitations of simulating an entire population with
one set of transport parameters.
The qualitative predictions of the model for internal nodes are interesting to
observe. An illustrative example has been chosen from the first 30 hours of Event 237.
Figures 10.7 and 10.8 show the simulated values for Total Suspended Solids (TSS) and
Erodable Mass (EM) for a series of nodes in the South and East Canal system. The nodes
were chosen to illustrate the sequential movement of TSS and EM downstream as the
event progresses. The nodes start with Node 75 which is 1609 meters upstream of the
pump station in the South Canal and continue downstream in increasing numerical order


77
Figure 5.2: Particle Size Distribution of Hydrated B9B10 Soil and B9B10 Sediment
Figure 5.3 shows the distribution (again excluding the greater than 1 cm
fraction) for the WPBC sediment superimposed on that of the UF9200 soil and drainage
ditch sediment. The mass mean particle size of the WPBC sediment is about 230
micrometers, similar to that of the UF9200 soil and drainage ditch sediment, but this is
influenced by a large fraction in the 150-300 micrometer range. The distribution at lower
particle sizes is less than that exhibited by the hydrated farm soil and drainage ditch
sediment. The <38 micrometer fraction was 7.5% compared to 11.7% for the hydrated
soil sample and 10.4% for the drainage ditch sediment.
Volatile matter content
Organic content is approximated by the estimate of volatile matter content (total
mass minus the ash content) of each fraction. The results of this analysis is shown for the
various fractions of each of the three wet samples in Figure 5.4. The greater than 1 cm


215
Note that the weighted average for the Low level estimate, 0.00076 gm/cnr. is
less than the lowest value in Column 6 of Table 7.7, 0.0016 gin cm" The weighted
average for the High level estimate, 0.00186 gm/cm2, is only slightly above the lowest
value in Column 6, and considerably lower than the range of 0.0077-0.0193 gm/cm2 that
prevailed for the larger events. Numbers 220, 237, and 285. Recall also that the High"
erosion level estimate derived from a test that represented only one location of the eight
sampled at UF9200.
These estimates have some important implications.
First, they indicate that the field ditches may not be making significant
direct contributions to the farm suspended solids discharge load because of
low ditch velocities. This indication is supported by the observations
made during the intensive synoptic studies.
Second, they imply that the surficial sediments found in seven out of eight
locations, if they existed in the canals, did not have erosion characteristics
sufficient to supply even the minimum suspended solids load for any of
the events from 220 to 285.
Third, they imply that even if the canals were populated completely with
sediments with erosion characteristics that fell in the High category,
resuspension would supply sufficient suspended solids for only about half
of the events studied, and would fall far short of supplying the suspended
solid loads for the three larger events, 220,237, and 285. All of this is in
addition to the question of how the sediment supply would become
replenished after an erosion event.


Selected Water Levels-UF9200
Julian Date
Field AI6 Pump Station
Figure E.3: Selected Water Levels at UF9200Julian Days 300-365


294
Figure 10.17: Simulation Results-Cumulative Suspended Solids Load for Event
UF9200-262
Event UF9200-262


137
significant magnitude difference between the two curves, especially at the lower shear
stresses.
This difference could arise from different methods used to estimate the bed shear
stress, or from differences in the PES and flume performances on different materials, or
from both causes. The results tend to emphasize both the opportunity for extrapolation
error and the need for as direct a representation as possible of the parameters used to
calibrate devices that estimate erosion. This appears to be especially true at low shear
stresses where erosion of organic sediment would be expected to take place.
Figure 6.16: PES-CRAF Calibration Results Using Cohesive Clays
Particle Entrainment Simulator operation organic sediment tests
The organic sediment tests were run with sediment that was inserted into the PES
cylinder by a simulated coring technique, similar to what would be done in the field.
Samples of surficial sediment from UF9200 field ditch B9B10, which were contained in


243
Summary
In this chapter the results of synoptic sediment sampling were presented which
added additional support to the hypothesis that much of the exported particulate
phosphorus was sourced from origins other than surficial sediment.
Experiments on, and sample analysis of, surface macrophytes and their
dislodgable detritus indicated that they were capable of producing erodable material of
phosphorus content similar to that measured in the exported particulate material.
Areal density measurements were made on the predominant macrophytes at
UF9200. These measurements will be used later in interpretation and justification of the
model results.
Large-scale composite samples were taken of farm discharge particulate matter.
Sorption studies on this material confirmed the original hypothesis of a weak absorption-
desorption character for the particulate discharge and supported the decision to eliminate
sorption from the conveyance system modeling effort. Fractionation of the phosphorus in
these samples showed that the farm discharge material had a much higher fraction of
labile phosphorus than the surficial sediments, adding additional support to the hypothesis
of alternative sourcing of discharged particulate phosphorus.
These findings required that the approach to modeling of particulate phosphorus
transport that had originally been conceived for this project be modified to incorporate
alternative sources. The next chapter deals with those modifications and the development
of the transport model.


202
Table 7.6: Sampling Locations for Intensive Synoptic Studies at UF9200
Sample Point ID
No.
Canal System
Distance
Upstream from
Pump Station m
Description
1
S-E (Ditch)
2623
Field Ditch B9B10, sampled at upstream end of
discharge culvert
2.0
S-E
2413
Main Canal Culvert 8-9, sampled 3 m downstream
of culvert discharge
2.1
S-E
1407
Main Canal Culvert 3-4, sampled 3 m downstream
of culvert discharge
2.2
S-E (Ditch)
1005
Field Ditch B1B2, sampled at upstream end of
discharge culvert
3
S-E
804
South Canal discharge into East canal at Culvert
0-1, sampled 5 m downstream of culvert discharges
4
S-E
220
Footbridge crossing lower reach of East canal,
sampled at bridge center, downstream side
5
N
11
North Canal discharge, sampled 3m downstream
of dual culverts
Discharge
Both
NA
Normal farm discharge sample taken during this
event


187
Correlation of TSS Load with Hydraulic Load
Farm UF9206N
100000
90000
80000
70000
60000
50000
40000
30000
20000
10000
0
0 200,000 400,000 600,000 800.000 1,000.000
Event Hydraulic Load m3
OEvents209-283 Events 300-336
Figure 7.17: Correlation of Event TSS Load with Event Hydraulic Load for Station
UF9206N
Con-elation of Particulate Phosphorus Load with Hydraulic Load
Farm UF9206N
0
Event Hydraulic Load m3
O Events209-283 Events 300-336
Figure 7.18: Correlation of Event PP Load with Event Hydraulic Load for Station
UF9206N


49
The device is of simple construction, consisting of a cylindrical chamber inside of
which a horizontal grid oscillates vertically. The sediment to be studied is placed in the
bottom of the cylinder and is overlain with water, preferably of composition similar to
that of the original sediment environment. The grid oscillates in the water and creates
turbulence that penetrates to the sediment-water interface, causing sediment resuspension.
The turbulence, and thus the extent of resuspension, is a function of the grid oscillation
frequency (Tsai and Lick, 1986)
The structure of the turbulence generated in the PES is different from that
generated in a typical field shear flow. The mean velocity vector in the PES changes
direction many times a second, whereas the mean shear velocity vector in field channel or
tidal flows is relatively steady over short time periods. Never the less the statistical
properties of the turbulence generated in the PES can be assumed to be similar to those in
field shear flow (Lavelle and Davis, 1987). The statistical similarity assumption allows
the PES to be calibrated to standard parameters.
The calibration procedure, developed by Tsai and Lick (1986), involves
determining the erosion extent of a specific sediment in the PES for various oscillation
frequencies. The erosion extent of an identical sediment, prepared under similar
conditions, is also determined in a shallow flume, for which the shear stress correlations
are known with some degree of accuracy. The flume is also operated over a range of
shear stresses. Under the assumption of statistically similar turbulence yielding similar
erosion extent, it is possible to match a specific erosion extent at a known oscillation
frequency in the PES with a corresponding equivalent erosion extent in the flume at a
known shear stress. Ultimately this empirical matching procedure yields a correlation of
PES oscillation frequency with applied shear stress in a uniform flow field. This is an
important translation, because all useful erosion-hydrodynamic models require an
expression of shear stress or equivalent shear velocity at the sediment-water interface for
calculation of erosion rates.


224
The results of these two surficial sediment sample sets added strong support to the
hypothesis that what was being sampled as surficial sediment in both the ditches and
canals was not contributing significantly to much of the discharge particulate phosphorus
load. The results from the three samples that were taken earlier in regions where there
was a notable surface macrophyte population strongly suggested that the surface
macrophytes, and possibly other in-stream plant growth, might be playing a major role as
sources of particulate phosphorus. A sampling program was initiated to allow a
preliminary evaluation of this hypothesis.
Macrophyte and Detritus Studies
The population of floating macrophytes at UF9200 consisted predominantly of
water lettuce (Pislia stratiotes), water hyacinth (Eichhornia crassipes), and water
pennywort (Hydrocotyle umbellata). Visual observations made at various times in the
season gave an estimate that roughly 50-75% of the total areal coverage by floating
macrophytes was by Pislia stratiotes, 25-35% was Eichhornia crassipes, and 10-15% was
Hydrocotyle umbellata.
Areal coverage was extremely variable from location to location within the farm
and varied with time in both a random and a systematic way. The random variation arose
from several sources. The population dynamics contributed to variation in that a small
collection of macrophytes may propagate in a logistic pattern, exhibiting the growth
phases of lag, logarithmic, linear, and declining (Reddy and DeBusk, 1983). Throughout
the farm there were various collections of macrophytes in various stages of development.
The pumping and wind patterns also contributed to changes in location and density of the
macrophytes in a manner already described. The grower also made periodic unscheduled
attempts to control macrophytes by spot herbicide spraying and some mechanical removal
when the macrophyte mats appeared to interfere with culvert and pump station operation.


160
ultrasonic surveying instrument. Also at 100 meter increments along all field ditches
water depth from the surface and distance to the non-yielding conveyance bottom
(assumed to be the sediment bottom/marl layer top) was measured at the ditch centerline
using the neutrally buoyant footpad and the sediment penetrometer mentioned in Chapter
5. At these locations the field surface elevation relative to the water level was also
measured.
In the canals water depth from the surface to the top of the sediment was
measured at canal centerline and at locations one meter out from each bank, again on 100
meter increments. True dimensions and elevations, including sediment fills and partially
collapsed shapes, were taken for all culverts. The end result of this effort was a relatively
detailed survey of the elevations and dimensions of the conveyance system of farm
UF9200. The hydrographic survey is included in the hydraulic model development
discussed in Chapter 9, but some average values were used early in this part of the
program and are presented in Table 7.1.
These values may be used to make some approximations that will be used later for
estimation purposes. The surface area of the conveyance system represents
approximately 2.5% of the total farm surface area. A water elevation of 2.6 m MSL
represents a typical surface elevation at start of pumping for UF9200. The volume of
water contained in the conveyance system at this water surface elevation is approximately
66,400 m3 which is roughly equivalent to 12.7 mm (0.5 inches) of rainfall. A typical
initial pumping rate at UF9200 for the study period was 1.6 m3/sec (13 8,240 m3/day) so
the canal volume represents on the order of 12 hours of pumping volume, which may
offer significant buffering capacity between the fields and the discharge point. Using the
same initial pumping rate and assuming an equal split between the two main canal
systems yields an estimate of initial maximum velocity in the canals of from 0.15 to 0.21
m/sec when estimated contributions from ditches are matched with the corresponding
canal crossectional areas. Assuming that the initial flow is equally distributed among all


a part of their Best Management Practices Implementation program for reducing
phosphorus loading to the Everglades.
The author would particularly like to thank his loving, patient and supportive
wife, Beverly, who contributed to the completion of this work in ways too numerous to
detail.
v


30
riverine, lacustrine, or estuarine environment. Cushing et al. (1993) studied the transport
of carbon-14 labeled fine particulate organic matter (FPOM) of particle size less than 100
micrometers to estimate the transport distances and residence times in the riverine
environment. They measured labeled suspended FPOM concentrations in the river as a
function of time and distance, and labeled FPOM concentration in the river sediments as
a function of time. Several interesting results were reported. The mean particle
deposition velocity, calculated from concentration gradients, was 0.43 m/hr, about one
order of magnitude less than the sedimentation velocity of particles from the same source
measured in the lab under quiescent conditions. Calculated mean time to deposition
under a mean velocity of 0.27 m/sec and average reach depth of 0.33 m was 51 minutes,
with an average of 83% of the injected FPOM being deposited in the 1 km reach study
area. However after 24 hours bottom sediment within the reach contained only 1% of the
originally deposited material. The authors concluded that particles in surficial sediments
exchange rapidly with the water column and migrate episodically throughout a riverine
system, leading to strong longitudinal connection of sediment distribution. Unstated, but
evident from the results, are the potential errors involved in attempting to apply
laboratory sedimentation test results to field conditions.
Kronvang (1992), using a synoptic water sampling design augmented with
increased sampling frequency during storm events, investigated the export of particulate
matter, with specific emphasis on phosphorus, from two Danish agricultural basins. He
found significant phosphorus content enrichment in the transported sediments compared
to typical soils in the watershed (sediment phosphorus concentrations of 7 to 14 times
those found in the soils). The phosphorus content of the particulate organic matter was
found to stay relatively constant at about 1 %, regardless of the total particulate
concentration, while the inorganic particulate phosphorus content decreased as particulate
concentration increased, but approached the phosphorus content of the local soil only at
extremely high particulate matter concentrations. The results of this study emphasize the


70
particulate matter at the EAA discharge points might contain on the order of 2200 mg
P/kg suspended solids. This three-to-four fold difference indicates the possibility of
selective transport or enrichment processes being of some significance in the EAA water
conveyance systems. One approach to evaluating the possibility of selective transport
was to fractionate material from various sources by particle size and determine if there
were appreciable differences in key properties of the several fractions.
Sources for particle size fractionation study
It was desired to evaluate farm soil, farm conveyance system sediment, and Water
Management District Canal sediment. Sampling was conducted in mid-July,
approximately 1.5 months into the wet season. The farm sediment was obtained from the
midpoint of a field ditch in farm UF9200. The ditch chosen was one approximately
halfway upstream and on the south side of the south canal of UF9200. It was well
maintained and free of emergent growth. Samples of the ditch surficial sediment were
obtained at ditch mid-length by Eckman dredge technique. Simultaneously composite
samples were taken of the field soil at locations immediately adjacent to the sediment
sample site and 5-10 meters in-field. The UF9200 grower identifies his fields by
increasing number from east to west and by letter from north to south (See Appendix G).
Field ditches are identified by the two adjacent fields. Field ditch B9B10 was the sample
site and was the ditch between fields 9 and 10 on the south side of the central (South)
canal. This sample will be referred to hereafter as B9B10, the soil sample will be
referred to as Soil. Farm UF9200 discharges into the West Palm Beach Canal. The
District Canal sample was taken from this canal approximately 1 km downstream of the
UF9200 discharge point at a location intermediate between two downstream farm pump
stations. This sample was also of surficial sediment, taken by Eckman dredge technique
and is referred to as WPBC. Approximately 20 liters of each type material were


206
UF9200 Intensive Studs'
Event 262
25
20
0 10 20 30 40 50 60
Time fiom Start of Punping hr
O(Ditch) 0#2 0#2.1 -l-#2.2(Ditch) --#3 Discharge
Figure 7.40: UF9200 Intensive Study-Event 262
UF9200 Intensive Study
Event 285
0#l(Ditch) #2 0tt2.\ --#2.2(Ditch) -#-#3 -M-M #5(NorthCanal) Discharge
Figure 7.41: UF9200 Intensive Study-Event 285


390
{
v=Q/As;
Define Stream Average Velocity
w=As/Z;
Define Average Stream Width
TERM=(vA2-VCA2);
Define Excess V2
VSQCONV=7464960000;
Convert Velocity Terms to m/day
ksed=wsed/z;
Define Settling term for this cycle
if (TERM<=0.00000)
KZEDTSS=0.00000;
Avoid Negative Settling
}
else
KZEDTSS=EPSI*VSQCONV*TERM*w/As*EM;
\
Calculate Internal TSS Zero Order Rate Constant
i
if (TSS<=0.00000)
{
KONETSS=0.00000;
Avoid Negative Erosion
}
else
{
K.ONETSS=-ksed;
Calculate Internal TSS First Order Rate Constant
if (TSS<=0.00000)
{


200
There was no readily available explanation for the Julian Date 240 grouping but it may be
surmised that a similar internal event took place here also.
Although there are greater deviations than were observed at UF9200 it is clear that
both locations at UF9206 exhibited the same response of decreasing phosphorus content
with increasing discharge TSS concentration. The differences between the two UF9206
stations are consistent with the concept of decreasing phosphorus content with increasing
hydraulic activity. More details of this hypothesis will be discussed later, after additional
supporting data are presented.
Intensive Synoptic Studies at UF9200
Sampling configuration
Information on the variation of suspended solids concentration and phosphorus
content with time at various points within the farm conveyance system during an event
may offer some insight into potential sources of particulate phosphorus. When this
information is combined with hydraulic measurements or estimates additional insight
might be gained regarding transport mechanisms and the relative magnitude of the
contribution of the potential sources. Part of this program entailed attempts to acquire
such synoptic data at UF9200.
Five internal sampling stations were installed in the conveyance system of
UF9200 between Events 244 and 252. Following the test run on Event 252, two
additional sampling stations were installed, for a total of seven. These stations were
operated with varying degrees of success through Events 258,262, and 285. Following
the end of Event 285 the stations were removed in order to not interfere with the growers
cane harvest operations.


26
1993). Microbial activity associated with litter decomposition can frequently cause an
increase in the total phosphorus concentration as the higher-phosphorus-concentration
microbes utilize the carbon source of the lower-phosphorus-concentration substrate to
which they are attached (Elwood et al., 1988). Cyanobacteria have been shown to, on
occasion, exhibit a reverse sedimentation effect where they detach from the sediment and
migrate into the water column. The conditions under which they do this also favor a high
bacterial phosphorus content so they may become an important source of suspended
particulate phosphorus under conditions of high water temperature and favorable C/N
ratios. Petersson et al. (1993) showed that particulate phosphorus flux into the water
column from this source could be in excess of 2.5 mg P/m2-day.
Phosphorus content is high for planktonic materials, for example Behrendt (1990)
reported values averaging 5100 mg P/kg biomass for diatoms, 7700 mg/kg for blue-green
algae, and 13,700 mg/kg for zooplankton. The high phosphorus content can be a source
of soluble phosphorus. Montigny and Prairie (1993) have shown that cell lysis of bacteria
can produce high levels of soluble phosphorus even in the presence of iron which would
be expected to precipitate the phosphorus.
Macrophytes and microbes typically dominate the mass of non-soluble material
present in streams, but macroscopic organisms, particularly the invertebrates, can play a
major role in modifying the physical character of the particulate matter resident in the
stream. Webster (1983) describes the complex process where invertebrate shredders and
collector-gatherers convert coarse organic matter to more transportable and accessible
fine particulate organic matter and also create a bi-directional flux of the fine matter
between the suspended and benthic compartments.
Many of the phosphorus content values noted above for autochthonous materials
may be considerably higher than those typical of average EAA basin soils or stream
sediments (See Chapter 5). This illustrates the important role of in-stream processes may
play in modifying the physical and chemical nature of the basin phosphorus cycle. The


395
Findlay.S.. Pace.M.. and Lints,D. 1991. Variability and Transport of Suspended
Sediment. Particulate and Dissolved Organic Carbon in the Tidal Freshwater
Fludson River, Bioaeochemistry. 12:149-169
Fisher, M.M., Brenner, M, and Reddy, K.R. 1992. A Simple Inexpensive Piston Corer
for Collecting Undisturbed Sediment/Water Interface Profiles, L
Paleolimnoloev. 7:157-161
Fiskell,J.G.A and Nicholson,I.K. 1986. Organic Phosphorus Content of Pahokee Muck
and Spodosols in Florida, Soil and Crop Sci. Soc. Fla. Prbc.. 45:6-11
Fread, D.L., 1978. National Weather Service Operational Dynamic Wave Model,
National Weather Service. Silver Spring, Md
Fukuda,M,K, and Lick,W. 1980. The Entrainment of Cohesive Sediments in
Freshwater, J. Geophvs. Res.. 85J(C5):2813-2824
Gachter, R and Meyer, J.S. 1993. The Role of Microorganisms in Mobilization and
Fixation of Phosphorus in Sediments, Flvdrobioloeia. 253:103-121
Gichucki, F.N., Walker, W.R., and Merkley, G.P., 1990. Transient Hydraulic Model for
Simulating Canal-Network Operation. J. Irrig. Drain. Enana.. 116(1): 67-81
Godschalk.G.L., and Wetzel, R.G. 1984. Accumulation of Sediment Organic Matter in a
Hardwater Lake with Reference to Lake Ontogeny". Bull. Mar. Sci.. 35:576-586
Hakanson.L. 1981. Determination of Characteristic Values for Physical and Chemical
Lake Parameters. Water Resource Res.. 17:1625-1640
Hakanson.L. 1986. Modifications of the Eckman Sampler, Int. Rev. Ges. Hvdrobiol..
71:719
Hawkes, H.A., 1975. River Zonation and Classification, In River Ecology, B.A.
Whitton, Ed., U. Calif. Press, Berkeley
Hio, M., 1963. Turbulent Flow with Suspended Particles, Proceedings ASCE.
89(HY4):
Hinze, J.O., 1959. Turbulence, An Introduction to its Mechanism and Theory, McGraw-
Hill, New York
Husain. T Abderrahman, W.A., Kahn, H.U., Kahn, S.M., Equinaibi, B.S., 1988. Flow
Simulation Using Channel Network Model, J, Irrig. Drain. Engng.. 114(3): 424-
441


32
periodic basis determined by transport events which affected resuspension of the nascent
FPOM. The authors postulated a succession scenario wherein macrophyte colonization
causes rapid detrital accrual near-shore. The detritus is reduced in size over time by
microbial decay and the smaller particles are transported by periodic hydraulic
excursions. Lability of smaller transportable particles depends on origin, for example
planktonic versus detrital, and age. Chemical and biochemical interaction of the small
particles with the environment is a complex function of space and time (depth, light
availability, redox potential, temperature, etc.) and the history of the particle.
Calvo et al. (1991) used the Particle Entrainment Simulator (PES) to study the
relative compositions of parent sediment and resuspended material in multiple locations
in a shallow (1 m average depth) marine lagoot^. Undisturbed core samples were taken
directly in the PES cylinders by divers and then immediately subjected to typical shear
stresses in the range of 6-9 dynes/cm2 in the PES onboard ship. Suspended particulate
material was sampled and analyzed, along with parent sediment, for Total Organic
Carbon, Total Kjeldahl Nitrogen. Exchangeable Ammonia Nitrogen, and Total Organic
Nitrogen. The authors found that the carbon/nitrogen ratio was always lower for the
suspended material than for the parent sediment, indicating a lower degree of
mineralization for the suspended material, and that the C/N ratio increased as the
suspending shear stress was increased. This latter finding suggests that the more readily
transportable the material was, the younger and less decomposed it was. In addition, the
largest deviations between suspended material and parent sediment composition occurred
in the regions which were identified as having the largest standing crop of phytoplankton.
No phosphorus analyses were done, but it follows from the hypothesis regarding age and
degree of decomposition versus transportability that the same results would have been
obtained for phosphorus. This study reinforces the concept of the selectivity of transport
with respect to composition, expressed by the authors as selective resuspension of freshly
deposited material.


122
Prototype Organic Sediment-Consolidated Bed
Stepwise Shear Stresses
(Erosion 15.3 cm Water Depth)
00100
0.0090
0.0080
00070 Tau=0.65 Tau=092 Tart .30 Taif=1.70
E 00060 dynes/cm2 dynes/cm2 dynes/cm2 dynes/cm2
0.0050 -
U)
P 0.0040 -
Tau=2.Q2 'q
dvnes/cm2
1
o o
*
Tau=2.80
dynes/cnT
U 0.0030 -
00020 0 .
0.0010 dP
nonm '' to-O- O'
0 20 40 60 80
100 120
140 160 180 200
Time-hr
Figure 6.11: CSME as Function of Shear Time at Several Calculated Shear Rates for
Erosion of Consolidated Prototype Organic Sediment Bed at 15.2 cm Water Depth
First, there was minimal detectable erosion at the two lowest calculated shear
stresses of 0.65 and 0.92 dynes cnr The data at these two shear levels showed random
variations and did not conform to an asymptotic exponential model; only when the
calculated shear level was increased to 1.30 dynes/cm2 did the system begin to show the
asymptotic exponential type of response exhibited in previous runs. Second, and more
important, the CSME response as a function of calculated shear stress was very dissimilar
from that obtained at the 30.5 cm water depth. The 15.2 cm water depth run was
correlated (r2 = 0.995), excluding the two lowest shear stresses, by the equation
CSME = 0.005737t 0.00725 6.11
where CSME and t are defined the same as before. The ordinate intercept of this
equation is CSME = 0 at t = 1.26 dynes/cm2.


400
Parchure.T.M.. and Mehta,A.J. 1985. Erosion of Soft Cohesive Sediment Deposits', L
Hydraulic Engineering ASCE. 111:1302-1326
Parthenaides,E. 1965. Erosion and Deposition of Cohesive Soils, J. Hydraulics Div.
ASCE.9HHY11: 105-139
Parthenaides.E. 1977. Unified View of Wash Load and Bed Material Load. J.
Hydraulics Div. ASCE 103HY91: 1037-1057
Partheniades.E., Kennedy ,J.F., Etter.R.J., and Hover.R.I1. 1966. Investigations of the
Depositional Behavior of Fine Cohesive Sediments in an Annular Rotating
Channel, Report No. 96, Hydrodynamics Lab, MIT, Cambridge, MA
Pettersson, K., Herlitz, E and Istvanovics, V. 1993. The Role of Gloeotrichia echinulata
in the Transfer of Phosphorus from Sediments to Water in Lake Erken,
Hvdrobiologia. 253:123-129
Raadsma, S., 1974. Drainage Practices in Europe, In Drainage for Agriculture,
Agronomy Monograph No. 17, J. Van Schilfgaarde, Ed., American Society of
Agronomy, Inc, Madison, WI
Raudkivi, A.J. 1976. Loose Boundary Hydraulics, 2nd ed., Pergamon Press, Oxford
Reddy, K.R. 1988. Water Haycinth Biomass Cropping systems: I. Production, In
Methane from Biomass A Systems Approach, W.H. Smith and J.R. Frank, Eds.,
Elsevier Appl. Sci., London
Reddy, K.R. 1994. Lake Okeechobee Phosphorus Dynamics Study: Biogeochemical
Processes in the Sediments Final Report, Submitted to South Florida Water
Management District, Contract No. C91-2393
Reddy, K.R., and DeBusk, W.F. 1983. Growth Characteristics of Aquatic Macrophytes
Cultured in Nutrient Enriched Water: I. Water Hyacinth, Water Lettuce, and
Pennywort, Economic Botany. 38:229-239
Reddy, K.R., and DeBusk, W.F. 1987. Nutrient Storage Capabilities of Aquatic and
Wetland Plants. In Aquatic Plants for Water Treatment and Resource Recovery.
K.R. Reddy and W. H. Smith, Eds., Magnolia Publishing, Orlando
Reddy, K.R., and DeBusk, W.F. 1991. Decomposition of Water Hyacinth Detritus in
Eutrophic Lake Water, Hvdrobiologia. 211:101-109
Rouse, H. 1937. Modem Conceptions of the Mechanics of Fluid Turbulence,
Transactions ASCE. 102: 463-543


85
agitation, centrifugation, decantation, and acidification as before, this time to determine
sorption equilibrium in a desorption mode. The data gathered during this phase was
designated as Desorption data.
Adsorption-Desorption data reduction
The objective of the data reduction was to develop a simple mathematical
relationship that could accurately describe the sediment adsorption/desorption
equilibrium at the relatively low phosphorus concentrations which prevail in the EAA
waters. The mathematical construct that was chosen was the Langmuir adsorption model,
which predicts a linear relationship between liquid phase and sorbed phase concentrations
at low liquid phase concentrations and an approach to a saturated maximum sorbed phase
concentration at high liquid phase concentrations.
The Langmuir isotherm equation can be expressed in a number of forms, but the
form most convenient for these purposes is:
S. + S.
s c
Ks + C
(5.1)
where:
Sa
So
Sm
Ks
Substrate concentration of srbate that has been adsorbed
from liquid phase, mg P/kg substrate
Native substrate concentration of srbate that was originally
present on substrate, mg P/kg substrate
Maximum srbate concentration that prevails at complete
saturation of all sorption sites, mg/kg substrate
Saturation constant for sorbate-substrate system, which is
related to affinity of substrate for srbate, mg/1


149
had not undergone the same type program so the biological condition of the two farms
differed somewhat in that more ditches and canal segments had obvious evidence of
bottom-rooted weed and floating macrophyte growth at UF9206 than at UF9200.
The sampling procedure was as follows. A preliminary reconnaissance visit was
made to each site to evaluate the cropping and hydraulic situation and to ascertain the
condition of each ditch and canal on each farm. From the information gained ditches
were picked that were judged to be representative of a specific subsection of the farms. A
total of eight conveyances were chosen from each farm for sediment sampling. The
actual samples were taken from farm ditches only, at the ditch crossectional midpoint
approximately 50 meters upstream of the culvert through which the ditch discharged to its
farm canal. The sample was taken by entering the ditch about 5 meters downstream of
where the sample was to be taken and slowly wading upstream. At the sample location
the PES bottom was removed and the PES cylinder was inserted approximately 5 cm into
the sediment. The PES bottom was then slowly slid under the cylinder bottom and
secured in place with the O ring seal. The PES cylinder, which at this time was full of
ditch water, was then capped, removed from the water, inserted upright into a
transportation cradle, and allowed to settle for 10 minutes. Prior to moving to the next
sampling location the settled sample was inspected. If the sample was judged to be
questionable (usually because of a bed that was too thick, too thin or sloped), the
sampling procedure was repeated 5 meters upstream. At the end of the sampling process
for each site the PES cylinders were transported to the University of Florida Everglades
Research and Education Center (EREC) Agricultural Engineering Laboratories in Belle
Glade, Florida, where they were stored at 4C for at least 48 hours before being tested on
the PES device. The sampling procedure precluded taking any samples from the farm
canals, which had water depths of 1.2-1.8 meters. A brief description of each sample
location is given in Table 6.4.


118
where CSME and t are defined the same as before. The ordinate intercept of this
equation is CSME = 0 at r = 0.55 dynes/cm".
Figure 6.10: Asymptotic CSME vs. Applied Shear Stress for Organic Sediment Extended
Run Erosion and Deposition with 30.5 cm Water Depth
A qualitative comparison of the curves for erosion and deposition leads to some
interesting tentative conclusions. The two curves do differ in slope by about 22% and in
intercept by about 34%, but the differences are possibly explainable by the extra shear
time applied to the system before deposition was initiated and the agglomerative potential
observed in the bottom withdrawal sedimentation tests described in Chapter 5. The extra
shear time provided the opportunity for bed surface energy input and the resuspension of
material beyond the asymptotic CSME and also may have affected size reduction of the
suspended particle population. This could explain the higher deposition CSME values at
the higher shear levels. As the shear level was reduced, however, the shear-induced
flocculation of the suspended particles may have outweighed the floc-destroying action of
turbulent shear to the point where the average sedimentation velocity for the population


136
shaft length of the grid support rod is adjustable, so there is considerable flexibility in
elevations of each of the PES elements.
The sediment to be studied is placed in the bottom of the cylinder and is overlain
with water, preferably of composition similar to that of the original sediment
environment. The grid oscillates in the water and creates turbulence that penetrates to the
sediment-water interface, causing sediment resuspension. The turbulence, and thus the
extent of resuspension, is a function of the grid oscillation frequency (Tsai and Lick,
1986).
Particle Entrainment Simulator operation preliminary clay tests
The PES was given extensive preliminary operational testing using the same
kaolinite, kaolinite/bentonite and kaolinite/attapulgite clay and clay blends that had been
used for the CRAF preliminary testing. The PES was calibrated against the CRAF for
each sediment using the CSME approach wherein the oscillation speed in the PES that
produces a specific CSME is assumed to be equivalent to the shear force that produces
the same CSME in the CRAF. This produces a calibration of apparent applied shear
force as a function of PES oscillation speed.
The details of the calibration trials on clays will not be discussed here, but it is
informative to look at some representative results. Figure 6.16 shows the results of
calibration using a composite of all synthetic kaolinite and kaolinite/bentonite-blend clay
runs. Also plotted is the original calibration curve developed by Tsai and Lick (1986),
using natural sediments from Long Island Sound. The original calibration curve has been
used by many workers (Davis 1993, Davis and Abdelrhman, 1992, Davis and Means,
1989, Lavelle and Davis, 1987, MacIntyre et al., 1990, Sfrisco et al 1991, and Ziegler et
al., 1987) as the single calibration curve for the PES device. Note that there is a


207
The large standard deviations reflect variation from event to event in the ditch
data and the intra-event variation (U shaped curves) in the farm data. Given the data
variability, a comparison of medians may be more appropriate. The median ditch TSS
discharge concentration was 2.7 mg/1 compared to a median of 11.7 mg/1 for the farm
discharge, a factor of 4.3 difference between ditch and farm discharge. By either measure
it is clear that the ditches were contributing little in the way of suspended solids directly
to the farm discharge.
The ditch vs. farm discharge comparison is best illustrated in the uncomplicated
context of Figure 7.38, Event 252, where Location #1, Ditch B9B10, is observed to
maintain a relatively constant concentration around 5 mg/1 TSS, while the farm discharge
varied from a minimum of 14 to a maximum of 47 mg/1 TSS through the course of the
major pump-down cycle. Occasional short term increases in ditch discharge TSS
concentration could be observed. This is well exemplified in Figure 7.39, Event 258,
where the ditches show a peak in the time range of 16-22 hours into the pumping event.
Examination of the canal and field level records indicates that this time frame roughly
coincides with the time during which the ditch water levels may have approached or
fallen below the crown of their outlet culverts, possibly creating surface turbulence that
may have disturbed floating macrophytes which tend to collect at the outlet culvert.
Influence of the North Canal
The North canal appeared to have an erratic but steadily increasing input of
suspended solids as time progressed from the start of Event 252 through the end of Event
285 on Julian Date 289.5, a span of around 37 days. The data from Event 252 show the
suspended solids from the North Canal maintaining a relatively low and constant level of
8-11 mg/1 TSS compared to the final-sample farm discharge of 47 mg/1. During Event
258 the North Canal discharge was much more erratic, reaching a maximum of 56 mg/1,


345
previously stated, controlling to these levels would probably not have saved the full one-
third of the exported particulate phosphorus that occurred above 100 mg/1 TSS
concentration. The control policy, however, would have been invoked on only six out of
ninety days during the period of consideration, which makes its implementation very
attractive. From a more qualitative standpoint, this example illustrates the potentially
high payback at UF9206 for avoiding a few extreme canal level excursions, which could
be done with minimal control effort.
Relation of Particulate Phosphorus Export to Soluble Phosphorus Export at Both Farms
In-depth investigation of soluble phosphorus export was not within the scope of
this work but it is extremely important to report the relationships between soluble and
particulate phosphorus that were deduced from the studies done at the three farm pump
stations. Soluble phosphorus analysis was not done on the samples collected for
suspended solids and particulate phosphorus analysis, but total phosphorus analysis was
conducted by others on parallel samples collected from the same locations at
approximately the same times as part of the Best Management Practices program.
These samples were used to calculate soluble phosphorus by difference,
subtracting the particulate phosphorus concentration obtained in this study from the total
phosphorus concentration obtained in the BMP study for corresponding samples to obtain
an estimate of soluble phosphorus. The results, expressed as load-averages for the
normal wet season studied, are shown in Table 11.1.
The results, which show particulate phosphorus fractions in the range of 28-35%
of the total phosphorus for the three locations, are consistent with the results derived from
the Izuno, et al. work (1991). It is important to note here that there were no indications at
any point in the study of the three pump station locations of sustained particulate
phosphorus fractions much in excess of 35%. In fact the value of 34.8% calculated for


388
South Canal
North Canal
LEGEND:
Field ditch
Roadway
I l Canal
I Pump
Riser
Culvert
* all field ditches are
joined to the
main farm canal
and the perimeter
ditch by riser and
culvert connections
not shown.
Figure G.3: Recycle Simulations UF9200


209
UF9200 Intensive Study-Event 258
South-East Canal System Stations Only
40
35
#2 1 -#-*3 #4 Discharge
Figure 7.42: UF9200 Intensive Study-Event 258 South-East Canal Sample Locations.
First 10 Hours
Figure 7.43: UF9200 Intensive Study-Event 262 South-East Canal Sample Locations,
First 12 Hours


LD
1780
199,
sm
UNIVERSITY OF FLORIDA
3 1262 08554 9235


193
Time Series-Flow and Total Suspended Solids
Event UF9206N-336
Julian Date
Pumping Rate O -TSS Concentration
Figure 7.27: Time Series-Flow and Total Suspended Solids Event UF9206N-336
Time Series-Flow and Total Suspended Solids
Events UF9206S-209, 213, 218, 224, 234
14 o 16
Julian Date
Pumping Rate O TSS Concentration
Figure 7.28: Time Series-Flow and Total Suspended Solids Events UF9206S-209, 213,
218,224,234


CHAPTER 7
FIELD EVENT STUDIES-TIME SERIES AND SYNOPTIC
Introduction
Detailed farm discharge data is key to the development of a coherent model for
particulate phosphorus transport. The only prior data available within the EAA farm
systems was the CH2M Hill farm scale study (CH2M Hill 1978) and the Izuno. et al.
(Izuno, et al., 1991) field scale studies. Both studies presented data of good quality but
neither study had erosion as a specific objective, so the data necessary to develop
particulate phosphorus erosion models were not gathered. Specifically what is needed at
minimum is time series data on farm particulate discharge, phosphorus content of the
discharged particulates, accurate representation of the hydraulic conditions that prevailed
during the discharge, and reliable estimates of antecedent and inter-event conditions.
The monitoring portion of the research program Implementation and Verification
of Best Management Practices for Reducing Phosphorus Loading in the EAA. conducted
by the University of Floridas Institute of Food and Agricultural Sciences (IFAS), had
developed the basic data gathering infrastructure to allow the acquisition of the type of
data needed to develop a particulate phosphorus transport model. This basic
infrastructure was adapted for particulate phosphorus monitoring at the primary target
farm UF9200, which had one discharge station. The same adaptations were made on both
of the pump discharge stations at UF9206, which was to be used for back-up and
comparison data. The field event studies can be placed in three categories, the time series
discharge studies, intensive synoptic studies, and large sample composite studies. This
chapter will discuss the first two study types methodologies, results, and implications
155


7
(Sanchez and Porter. 1994). It has also been postulated that cropping increases the net
rate of mineralization by further increasing the acti vity of the aerobic microbial
population (Cosgrove, 1977). For these and other reasons related to crop and water
management, the activities in the EAA have combined to increase the phosphorus loading
outside the EAA.
Izuno et al. (1991) illustrate the increase with some analytical results which
indicate that the water pumped to Lake Okeechobee in the period 1983-1985 had a total
phosphorus concentration in the range of 0.19-0.57 mg/1, compared with the lake's mean
total phosphorus content of 0.05-0.10 mg/1. During the period 1978-1986 water pumped
to the WCAs had total phosphorus concentrations in the range of 0.07-0.16 mg/1. The
total phosphorus content in water in the interior of the WCAs, on the other hand, was in
the range of 0.009-0.014 mg/1 (Whalen et al., 1992), a factor of 6 to 7 lower than the
incoming water. These figures serve to highlight the cause for concern regarding the
modification of the oligotrophic status of the WCAs, and ultimately of the ENP.
Proposed Regulatory Remediation
The South Florida Water Management District, under the terms of the Florida
House of Representatives' "Marjory Stoneman Douglas Everglades Protection Act of
1991". has the responsibility for ensuring mitigation of the phosphorus efflux from-the
EAA to the WCAs and the ENP. The District has proposed a remediation plan that
requires implementation by EAA growers of a series of Best Management Practices
(BMPs), developed by Bottcher and Izuno (1993), and a massive treatment plan that
involves the construction of four Stormwater Treatment Areas (STAs) within the EAA.
The BMPs, which are to affect 25% of the phosphorus load reduction, fall into
two categories; fertility BMPs and water management BMPs.


10.13 Simulation Results-Cumulative Suspended Solids Load for Event
UF9200-252 292
10.14 Simulation Results-Cumulative Particulate Phosphorus Load for Event
UF9200-252 292
10.15 Simulation Results-Cumulative Suspended Solids Load for Event
UF9200-258 293
10.16 Simulation Results-Cumulative Particulate Phosphorus Load for Event
UF9200-258 293
10.17 Simulation Results-Cumulative Suspended Solids Load for Event
UF9200-262 294
10.18 Simulation Results-Cumulative Particulate Phosphorus Load for Event
UF9200-262 294
10.19 Simulation Results-Cumulative Suspended Solids Load for Event
UF9200-285 295
10.20 Simulation Results-Cumulative Particulate Phosphorus Load for Event
UF9200-285 295
10.21 Correlation of Initial Erodable Mass with Interevent Time 297
10.22 Chronological Values of Cem(0) 299
10.23 Cemio) Correlated as an Exponential Growth Equation 300
10.24 Suspended Solids Simulation Results for Event UF9200-319 307
10.25 Suspended Solids Simulation Results for Event UF9200-355 307
10.26 Cumulative Suspended Solids Load Simulation Results for Event
UF9200-319 308
10.27 Cumulative Suspended Solids Load Simulation Results for Event
UF9200-355 308
10.28 Cumulative Particulate Phosphorus Load Simulation Results for Event
UF9200-319 309
10.29 Cumulative Particulate Phosphorus Load Simulation Results for Event
UF9200-355 309
xix


205
UF9200 Intensive Study
Event 252
Figure 7.38: UF9200 Intensive Study-Event 252
UF9200 Intensive Study
Event 258
80
70
60
O#1 (Ditch) 0-#2.1 #2.2(Ditch) -#-#3 #4 #5(North Canal) Discharge
Figure 7.39: UF9200 Intensive Study-Event 258


105
z b = 0.0 1 0(AV )'27
6.2
For 30.48 cm (12 inch) depth:
t = 0.005(AV)'37 6.3
where tb has units of dynes/cm2 and AV has units of cm/sec. The inherent
assumption in this process is that the subject sediments have surface roughness equivalent
to that of the Plexiglas false bottom. It should be noted that the basic configuration and
calibrations of the CRAF have been used as standards by others in the field, see. for
example, Verbeek et al. (1993).
The flume is equipped with vertical Plexiglas view ports located every 90, seven
sample taps located at elevations ranging from 3.8 cm (1.5 inches) to 30.8 cm (12.125
inches), and a bottom drain tap. Adjustment of the ring height is done with winch and set
screws. Fine adjustment of the rings elevation and leveling can be done with four
adjustable support couplings located every 90 at the juncture of the ring spine and the
support struts. The channel and the ring each are driven by concentric shafts attached
independently by belt drive to high torque DC motors. Each motor is independently
controlled by a variable speed controller equipped with a feedback loop to maintain
constant speed.
Flume operation
The CRAF had been out of service for some time at the start of this phase of the
project so it was necessary to conduct some refurbishment on the flume and controllers.


This work is dedicated to the two people in my life who have helped me negotiate
all the forks in the road, my wife Beverly, and my son Clark; and to the memory of
Christopher.


159
The particulate phosphorus samples were not acidified in the field, but were stored at 4C
between their return from the field and their analysis.
The integration of the particulate phosphorus monitoring program within the
infrastructure of the Best Management Practices program allowed the critical data to be
collected that was required for the development of a particulate transport model. Samples
were available with a two hour temporal resolution for the analysis of particulate
discharge with time. Data were available with a one hour resolution for determination of
the rainfall and hydraulic conditions existing at the pump station and groundwater and
conveyance elevations were available at the same level of resolution for selected locations
within the farm. What remained to be added to the program was determination of the
detailed hydrography of the target farm and an appropriate method for determining the
phosphorus content of the exported particulate material.
Detailed hydrography of the target study farm-UF9200
The hydrography of the ditch and canal system of UF9200 was mapped over a
period of several months using the canal water surface as reference. This required
periods which were devoid of rain, pumping, or irrigation for at least five days prior to the
mapping. It also required a return trip at a time into the dry season when the water level
was lower than the minimum observed during the first rounds in order to map the lower
levels of the channel system and obtain direct measurement of low-lying culvert
dimensions.
The mapping process was relatively simple in concept. When a mapping session
was started, water level was checked at the pump station and at the internal wells to
insure the absence of hydraulic gradient. The water level, referenced to mean sea level,
then became the primary reference. Bank width measurements were made at regular
elevations at 100 meter increments along each field ditch and canal using a calibrated


Table 5.5: Phosphorus Fractionation Results for WPBC and B9B10 Sediments
(Figures in parenthesis refer to per cent of total phosphorus)
Sample
Labile P
(Bicarb. Extractable)
mg/kg
Acid Hydrolyzable P
(HCI Extractable)
mg/kg
Refractory P
(by Difference)
mg/kg
Total P
mg/kg
Organic
Content (%)
WPBC
75-150 p
4 (0.2%)
-2570 (>99%)
nil
2548
6.4
WPBC
<38 p
(5(1.3
783 (68.5%)
345 (30.2%)
1143
42.5
B9B10
75-150 p
71 (8.9%)
305 (38.0%)
426(53.1%)
802
79.5
B9B10
<38 p
61 (6.5%)
343 (36.3%)
540 (57.2%)
944
80.2
Phosphorus Fractions vs Organic Fraction
WPBC and B9B10 Sediments
1
b 0.9
0.8
OP
f= 0.7
i ,1 0.6
£ E
0.5
0.4
0.3
0.2
0.1
0

0.2 0.3 0.4 0.5 0.6 0.7 0.8
Organic Fraction-mg Organic Mass/mg Total Mass
Acid Hydrolyzable A Bicarbonate Extractable O Refractory
Figure 5.8: Phosphorus Fractions vs. Organic Fraction for WPBC and B9B10 Sediments


Recommendations
BMP recommendations for field implementation
The results of this project allow some recommendations to be made at this point
for management practices to reduce particulate phosphorus export.
Source Control This study has presented strong circumstantial evidence to
support the contention that the biological growth in the main farm canal systems
is a prime contributor of exported particulate phosphorus. Control of the
concentration of this material by growth inhibition or physical removal is the
single most effective practice that can be recommended. Efforts that focus on
field ditch control are expected to have limited direct impact on particulate
phosphorus export.
Herbicide Use The use of herbicide or plant growth regulators to inhibit
growth is recommended. The use of herbicide to Wll existing viable plants is not
recommended because the potential to transform nonerodable viable plant
material to erodable detritus is great.
Sediment Traps Sediment traps that do not appreciably alter the stream
velocity are not recommended. Sediment traps should be designed to reduce the
stream velocity to near the critical velocity for initiation of transport, which at this
stage of knowledge is estimated to be about 0.1 m/sec for the organic materials
studied.
Flow Area Increase Any effort to reduce the stream velocity in the region of the
pump intake is in the direction of goodness and may reduce or eliminate the first
leg of the U curve, or the first flush as it has been called. This must be done,
however, by an actual increase in flow area, not storage area. This would


222
Macrophytes tend to collect in this region and. at the time of sampling, had bridged across
the canal for a distance of about 50 m lengthwise. The commonality among these three
samples was that they were all taken from sediments that were immediately below a
collection of floating macrophytes, while all other samples in this set were taken from
sediments that were under relatively clear water.
Later in the season, on Julian Date 300, a smaller set of canal surficial sediment
samples were taken approximately 260 hours after the termination of Event 285. By this
time it had been determined from in-channel flow measurements that for the typical
pumping event about 50% of the flow in the South-East Canal system came from the
easternmost (downstream) 25% of the farm, so the highest velocities in the South Canal
would occur in the regions of Fields B1-B4. Sampling was skewed so that four samples
came from this region and four samples came from locations distributed at other points in
the North and South Canals. All samples were again of the top 2 cm. taken in open water
conditions (no macrophytes in the immediate vicinity) with the piston core sampler. This
time samples were run in duplicate to evaluate reproducibility.
The results are shown in Table 8.3. For the whole population there was
considerable variation in sediment solids content (Avg.-l 1.4%, Std. Dev-5.0%, CV-44%)
but much less variation in organic (volatile) fraction (Avg.-0.52, Std. Dev-0.08, CV-
15%). The population average phosphorus content of 841 mg/kg was higher than that
found in the equivalent set in the earlier sampling (540 mg/kg) but was still much lower
than the average of the last ten recorded hours of preceding Event 285, which was about
9350 mg/kg, and the average of the first ten hours of the subsequent event, Event 318,
which was about 6770 mg/kg. The samples taken from the lower end of the South Canal
had approximately the same statistics as. and did not appear to differ in any meaningful
way from, the whole population. The maximum phosphorus content observed was an
average of 1208 mg/kg from Location B3 in the South Canal. Even this value was


CHAPTER 6
SURFICIAL SEDIMENT TRANSPORT STUDIES
Introduction
Within the context of this study, surficial sediment is the sediment that, by normal
judgment standards, would be expected to contribute to solids resuspension under erosive
conditions and would also be the receptacle for suspended solids which settle to the
bottom under conditions of net sedimentation. Operationally it is considered to be the
sediment obtained by surficial dredge or mild hydraulic mining techniques or by selective
removal of the top section of a piston core sample. As an initial working definition in
this project the top two centimeters of the sediment column were considered to be the
upper limit of the range of the surficial sediment layer.
The specific objectives of the tests discussed in this chapter were to:
1. Choose a prototype sediment that was deemed to be representative of
surficial sediments expected to be found in EAA farm drainage
conveyance systems.
2. Evaluate the erosion characteristics of the chosen prototype sediment
under controlled laboratory conditions in a well-proven erosion
characterization device to provide a standard parameter set.
3. Choose a small-scale portable erosion characterization device for use in
the field and evaluate the characteristics of the prototype sediment in that
device.
4. Calibrate the prototype sediment erosion characteristics in the field device
back to the standard parameter set.
99


CHAPTER 10
TRANSPORT MODEL RESULTS AND DISCUSSION
Introduction
The preceding chapter discussed the model and its calibration methodology. In
this chapter the results of the model calibration on the six wet season events are presented
and the implications of the resulting calibration parameters are discussed. The model is
then extended to the late-season tropical-depression storms in an attempt at model
validation. The assumptions necessary to arrive at a logical application of the model give
important guidelines to the model limitations and set some priorities for future work.
Calibration Parameter Results
The calibration parameters that gave the best fit by the process described in
Chapter 9 are presented in Table 10.1. Note that the internal calculation algorithm of
DUFLOW requires that the time dimension of the quality model parameters be expressed
in terms of days. The erosion coefficient and the sedimentation velocity were input in
these terms, the critical velocity was input in the dimensions of meters/sec, the same
dimensions as other velocities are calculated in DUFLOW, and then converted to m/day
in the quality program. This is the reason for the mixed units in Table 10.1.
The erosion coefficient is a characteristic of the particular system that was
calibrated and the particular form of the model, which uses mean velocity rather than
shear stress, so there are no existing values against which to compare it at present. The
sedimentation velocity and critical velocity for erosion, however, may be compared with
results obtained earlier in the prototype sediment study.
278


CHAPTER 5
SEDIMENT SURVEY AND PHYSICAL-CHEMICAL CHARACTERIZATIONS OF
SELECTED SEDIMENTS
Sediment Survey
The first field activity of the program was to develop an order of magnitude
estimate of quantities of sediment present in representative farm canals within the EAA.
Typical field ditch construction technique in the EAA involves digging or
dredging at least down to the limerock/marl substratum underlying the organic soils. In
the case of main canals, the limerock is usually excavated to a depth of several feet below
the marl surface to improve hydraulic flow conditions. Ensuing physical and biological
processes combine to contribute a sediment layer that builds up in the conveyance system
over time. Periodically the system components must be dredged to remove the sediment
build-up. Dredging is a relatively costly process, especially for the small grower who
must contract out the work, thus the extent of ditch and canal maintenance via dredging is
variable among growers.
The excavation or dredging technique, which initially removes all soil from the
channel, does simplify the measurement of sediment inventory, because all transportable
particulates present in the channels may be assumed to be sediment.
Selection of representative farms
The research program Implementation and Verification of Best Management
Practices for Reducing Phosphorus Loading in the EAA, conducted by the University of
Floridas Institute of Food and Agricultural Sciences (IFAS), had a total of ten farms
62


255
The drainage model also includes a water balance for rainfall inputs and
evapotranspiration outputs, both of which are user specified. The water balance tracks
the air pore volume rather than the soil profile water volume. The water table depth is
tracked from cumulative air volume vs. depth curves, using separate curves for rising and
falling water tables.
Operationally the groundwater flow model inserts itself as field runoff, using field
sizes and drainage nodes that are specified in the regular DUFLOW input, but the runoff
is calculated by the field drainage module instead of the normal DUFLOW field runoff
routines. The groundwater flow module is invoked by specifying a particular negative
rainfall number in the rainfall boundary conditions. -1 to run in the drainage mode with
no rainfall. -99 to run in the drainage mode and read an external rainfall file.
The model was originally calibrated with selected data for a single field from
UF9200 using a saturated hydraulic conductivity of 12.2 m/day that represented the
average measured local hydraulic conductivity found at UF9200 using the auger-hole
method (Bouwer and Jackson, 1974). The calibration process involved holding hydraulic
conductivity constant and varying drain spacing. It was found that a drain spacing of 70
m, instead of the actual ditch spacing of 200 m, was necessary to match field conditions.
The difference was attributed to the necessity of compensating for the mole drains with a
lumped parameter system. This assumption was later modified in the farm-scale
calibration process. The air-volume versus water table depth curves were also adjusted
during the calibration process to match observed water table values. These values were
preserved during the farm-scale calibration process.
In its current form the model contains seven parameters that are input by the user
for the groundwater drainage computations. There is also a three parameter surface
runoff sub-routine in the model. The surface runoff model uses a form of the Manning
Equation and assumes hydraulic gradient is proportional to the mid-field surface water


256
depth. There is the capability to specify surface storage and the fraction of field area
contributing to runoff. The model parameters are as follows.
Elevations:
The elevation of the impermeable layer m
The elevation of the ditch bottom m
The elevation of the field surface m
The elevation of mid-field water level at start of simulation m
Field Ditch Parameters:
Effective ditch spacing m
Drain radius cm
Soil Parameters
Hydraulic conductivity, Ksat, m/day
Surface Runoff Parameters
Mannings n for surface flow on the field
Surface storage before runoff (initial abstraction) cm
Runoff factor, the fractional area of the field contributing to runoff
Hydraulic model calibration
General Data Regarding UF9200 Hydraulic Model The hydraulic model cast the
UF9200 field drainage gridiron in an assemblage of 104 nodes which defined the end
points of 68 canal and ditch sections and 47 conveyance structures (culverts). All
conveyance elevation and crossectional dimension data and structure dimensional data
were determined from the detailed hydrography field survey discussed in Chapter 7. The
details of the network dimensional data input are given in Appendix F. A simple
schematic of the gridiron is shown in Figure 9.1


381
SECT 68 68 68 74 201 1.93
W 270.0 0.0
H 0.0000.35400 1.0910 1.1220
BS1 3.4687 3.4687 13.590 13.654
BS2 5.6690 5.6690 14.523 14.883
SECT 69 69 69 70 792 1.58
W 180.0 0.0
H 0.0000 1.4690
BS 3.7580 3.7580
SECT 75 75 75 78 191 1.75
W 270.0 0.0
H 0.0000.53300 1.3010
BS1 5.6920 5.6920 9.0580
BS2 7.1870 7.1870 10.463
SECT 72 72 72 73 780 1.50
W 180.0 0.0
H 0.0000 1.5450
BS 3.0110 3.0110
SECT 74 74 74 81 202 1.99
W 270.0 0.0
H 0.0000.29300.79500 1.0610
BS1 5.6690 5.6690 14.523 14.883
BS2 6.7140 6.7140 14.489 14.636
SECT 76 76 76 77 792 1.91
W 180.0 0.0
H 0.0000 1.1400
BS 2.9960 2.9960
SECT 78 78 78 84 201 1.65
W 270.0 0.0
H 0.0000.63400 1.4020
BS1 7.1870 7.1870 10.463
BS2 7.5430 7.5430 11.301
SECT 79 79 79 80 780 1.60
W 180.0 0.0
H 0.0000 1.4450
BS 2.9960 2.9960
SECT 81 81 81 87 201 1.90
W 270.0 0.0
H 0.0000.37200 1.0550 1.1400
BS1 6.7140 6.7140 14.489 14.636
BS2 6.7100 6.7100 14.490 14.490
BB1 6.7100 6.7100 14.490 14.640
BB2 9.4880 9.4880 15.389 15.474
SECT 82 82 82 83 792 1.86
W 180.0 0.0
H 0.0000 1.1920
BS 3.0480 3.0480
SECT 84 84 84 90 201 1.77
W 270.0 0.0
H 0.0000 .50900 1.2770
1.99 17.00 17.00
1.58 17.00 17.00
1.65 17.00 17.00
1.50 17.00 17.00
1.90 17.00 17.00
1.91 17.00 17.00
1.77 17.00 17.00
1.60 17.00 17.00
1.90 17.00 17.00
1.86 17.00 17.00
1.76 17.00 17.00


341
Particulate phosphorus export at UF9206
The hypothesized effects of operating levels on erodable mass transport
supplement and clarify the concept of decreased phosphorus content of erodable mass
with increased hydraulic activity. Periodic situations where extreme velocities transport
large amounts of erodable mass and base sediment would be expected to reduce the
concentration of higher phosphorus-content, highly mobile young erodable mass.
These same episodes, by de-compacting and remobilizing diagenetic base sediment, may
serve to add a low phosphorus-content diluent to the mobile particulate phosphorus pool
when the mobilized base sediment settles out and co-mingles with the remaining erodable
mass.
The observed lower phosphorus content correlations of UF9206 may be in part a
manifestation of this process. Figure 11.17 reiterates the Particulate Phosphorus Content
versus Total Suspended Solids Correlations for all three farm pump locations.
Correlations of PP Content with TSS for All Three Farm Pump Stations
UF9200 UF9206N UF9206S
Figure 11.17: PP Content vs. TSS Correlations for All Three Farms


378
BS1 6.9090 6.9090 10.271
BS2 7.1620 7.1620 10.098
SECT 16 16 16 17 780
W 180.0 0.0
H 0.0000.98800
BS 3.2000 3.2000
SECT 18 18 18 24 201
W 270.0 0.0
H 0.0000.76500 1.5270 1.5330
BS1 9.0060 9.0060 13.191 13.200
BS2 9.4150 9.4150 12.560 12.786
SECT 19 19 19 20 792
W 180.0 0.0
H 0.0000.98800
BS 3.8620 3.8620
SECT 25 25 25 28 190
W 270.0 0.0
H 0.0000.43900 1.2070
BS1 7.1620 7.1620 10.098
BS2 5.1300 5.1300 8.9270
SECT 22 22 22 23 780
W 180.0 0.0
H 0.0000 1.0390
BS 3.7580 3.7580
SECT 24 24 24 31 201
W 270.0 0.0
H 0.0000.75300 1.3870 1.5210
BS1 9.4150 9.4150 12.560 12.786
BS2 8.0160 8.0160 12.649 12.993
SECT 26 26 26 27 792
W 180.0 0.0
H 0.0000 1.1160
BS 2.8440 2.8440
SECT 28 28 28 34 201
W 270.0 0.0
H 0.0000.50900 1.2770
BS1 5.1300 5.1300 8.9270
BS2 4.5200 4.5200 7.7200
SECT 29 29 29 30 780
W 180.0 0.0
H 0.0000 1.1920
BS 3.5050 3.5050
SECT 31 31 31 37 201
W 270.0 0.0
H 0.0000 .46300 1.0520 1.2310
BS1 8.0160 8.0160 12.649 12.993
BS2 9.7140 9.7140 13.432 13.971
SECT 32 32 32 33 792
W 180.0 0.0
H 0.0000 1.1920
2.06 2.06 17.00 17.00
1.51 1.53 17.00 17.00
2.06 2.06 17.00 17.00
1.84 1.77 17.00 17.00
2.01 2.01 17.00 17.00
1.53 1.82 17.00 17.00
1.93 1.93 17.00 17.00
1.77 1.91 17.00 17.00
1.86 1.86 17.00 17.00
1.82 1.72 17.00 17.00
1.86 1.86 17.00 17.00


I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
Professor of Coastal and Oceanographic
Engineering
This dissertation was submitted to the Graduate Faculty of the College of
Engineering and to the Graduate School and was accepted as partial fulfillment of the
requirements for the degree of Doctor of Philosophy.
December. 1996
Winfred M. Phillips
Dean, College of Engineering
Karen A. Holbrook
Dean. Graduate School


304
potential of this algae suggests it could contribute significantly to the erodable sediment
load in addition to the other sources that have been discussed in this section.
As has been the case several times earlier in this work, the objective here was not
to quantify specific data, but to illustrate the credibility of specific assumptions or results,
in this case the calibration values of Cem<0), the initial concentration of erodable mass for
each event. The order-of-magnitude estimates presented here coincide with the
calibration values at the lower end of the correlation range, and the estimated senescent
and detrital material production rates coincide approximately with the correlation time
constant of 0.04 days"1.
The reported studies of epiphyton production add the potential for a substantial
source of algal biomass deposition especially if the surface macrophytes are periodically
subjected to shear from wind driven motion, wind driven currents, and the direct impact
of rainfall. The high growth rates of the lyngbya filamentous algae add a potentially
substantial but unquantified source of erodable organic material. At this point this is the
best that can be offered. More accurate confirmation must await future work.
Model Validation Prediction of the Suspended Solids and Particulate Phosphorus Loads
for the Late Season Storms
All model calibration for hydraulics, suspended solids transport, and phosphorus
content was done with six major events occurring during the normal wet season. For
model validation the correlations developed from the normal season data were applied to
the pumping events arising from the late season tropical depression storms that occurred
in November and December, normally the beginning of the dry season and sugarcane
harvest time.
Recall from the Small Flows section that the operation of the small pump was
excluded from calculation of interevent time, which excluded Event 244 from
consideration. The same situation prevailed in the late season pumping events. Events


93
Solution of the simultaneous equilibrium and mass balance equations for
phosphorus gives a value of 0.00495 mg for PdeS, the mass of phosphorus desorbed to the
control volume. The 50 mg of sediment, which originally contained (50 mg x 1000 parts
per million) = 0.05 mg of total phosphorus, has transferred 0.00495 mg to the soluble
phase. These results are summarized in Table 5.7.
Table 5.7: Summary of Phosphorus Distribution in Demonstration Example 1
1 Phosphorus Origin
Concentration-mg/1
% of Soluble P
% of Total P
Original Soluble
0.10000
94.76
66.41
Pore Water Soluble
0.00057
0.54
0.38
[| Desorbed Soluble
0.00495
4.69
3.28
Total Soluble P
0.10552
100
70.07
Suspended Sediment
(Insoluble)
0.04505
NA
29.93
Total P
0.15057
NA
100
Note that for this example, which used upper observed limits for pore water
phosphorus and the partition coefficient, the combined contribution of soluble
phosphorus from pore water and desorption was about 5.2% of the system soluble
phosphorus and about 3.7% of the system total phosphorus.
Demonstration Example 2 As a second example, consider the B9B10 <38p
fraction which had a statistically estimated value of So of 113 mg/kg, and a total
phosphorus content of 944 mg/kg. The fractionation process solubilized about 1% of the
total phosphorus in the sediment, or about 10 mg/kg. Adding that fraction to So gives a
total original sorbed phosphorus concentration of 123 mg/kg and an original total
phosphorus concentration of 954 mg/kg. The previous calculations may be redone with


331
than they had in RCL2 and thus were available for transport in regions of higher velocity.
The end result was that these solids were exported later in the event to such an extent that
the final net export was almost identical with the original simulation.
Simulation RCL4 Long Duration. Short Distance. Higher Flow. Backflow
Prevention This simulation comes full circle back to the configuration of RCL1. where
recycle flow is discharged from the pump station to a point immediately upstream of the
discharge culverts of the North Canal with the (assumed) gates closed. The difference
between this simulation and RCL1 is that the recycle flow is now 1,65m3/sec and the
recycle duration is now 72 hours. Figure 11.10 shows the results.
Figure 11.10: Simulation RCL4 Long Duration, Short Distance, Higher Flow,
Backflow Prevention
This simulation may be viewed as the beginnings of a racetrack design where
erodable mass is moved at high velocities downstream in one canal and upstream in the
other to transport solids to regions of subsequent lower velocities during discharge


276
From this point on the values of Cem were relieved of the interevent time
ratio constraint and were adjusted as needed for the best fit for each event.
The calibration process then became iterative on all events until the best fit
was judged to have been found on all six events. At this point changes of
plus and minus 25% were made in each of the three transport parameters.
Vc, e and wsed one at a time, and the calibration process was repeated for
each change to determine if a better fit across all six events could be
achieved. The parameters presented in Chapter 10 are the end result of
multiple iterations of this calibration process.
The simulation of Event 220, which had the highest value of Cem<0) of the
six events, showed that at the end of the event the value of Cem at several
nodes closest to the pump station was considerably higher than the starting
value of the next event. Number 237. In this case the end values of Cem at
the two nodes closest to the pump station were carried forward as the
values of Cem(O) for event 237. This same situation prevailed for Event
252, but not for subsequent events.
The calibration process, which was done manually, was extremely labor intensive.
Given the complexity of the process and the calibration across six events, a concise
demonstration of sensitivity is difficult so some general observations will be reported
here. The shape of the U curve was different for each event. The ends of the curve
were fit by adjusting s, wsed and Cem(0) The values of e and wsed would determine the
relative heights of the two legs of the U and the value of Cem(0) would scale the
absolute magnitude of the curve height. Within limits for any single event various values
of e and wsed could be used as long as their ratio was held constant. That is, increasing
the erosion coefficient by 25% and increasing the sedimentation velocity by 25% would
give approximately the same results, all other variables being constant. However from


APPENDIX A
PHYSICAL SAMPLING TECHNIQUES
Piston Core Sampler (Livingston, 1955, Fisher, et al., 1992)
The piston core sampler consisted of a 3 inch (7.62 cm) outside diameter, 2.75
inch (6.99 cm) inside diameter polycarbonate tube 4 feet (122 cm) in length with a
cambered edge at its tip. The tube was fitted at its upper end with reducing fittings and
threaded extension handles of 1.5 inch, schedule 40 PVC pipe. The piston was made
from a #13 1/2 neoprene rubber stopper which had been turned to adjust its diameter
appropriately. The piston was fitted with a bolted screw-eye assembly and a small valved
vent tube. The screw-eye was attached a flexible cable which was threaded up the inside
of the core and emerged from a hole in the side of the first handle extension, to then run
to the upper end of the handle on its exterior.
Sampling was done from a floating platform which was stabilized to reduce
vertical oscillations. Prior to sampling, the vent valve was closed, the lower end of the
piston was positioned slightly above the end of the core tube, and the assembly was filled
with ambient water by submergence to eliminate buoyancy effects. The lower end of the
corer was then positioned at a predetermined position which was intended to place it
several centimeters above the sediment/water interface. The cable was then secured at the
platform to eliminate the possibility of downward movement of the piston. The handle of
the device was then pushed down until there was no additional vertical movement under
full body weight. Under these conditions the corer cylinder penetrated the sediment and
the stationary piston preserved the integrity of the sediment/water interface. The entire
assembly was then lifted to the surface. Wall friction of the consolidated sediment at the
356


125
component was sometimes assumed to be zero during extended erosion tests. The work,
for example, of Parchure (Parchure 1984, Parchure and Mehta. 1985) on the mechanisms
of erosion of cohesive clays had a basic implicit assumption that the erosion extent with
time observed in the CRAF was independent of any sedimentation contribution, or. stated
differently, that the critical shear stress for sedimentation of each population class was
less than the shear stress applied to suspend that class, so the probability of deposition of
all suspended classes was zero as long as the shear stress was not decreasing.
This conclusion posed a setback to the concept of relating the erosion properties
of the organic sediment to the fundamental property of bed shear stress, and also to the
objective of developing correlation parameters for the calibration of the field device. The
scope of this project did not allow for the considerable resource investment required to
develop new calibrations for the flume with asymmetric friction factor and stress
dissipation distributions, nor was the scope of this preliminary assessment of the
prototype sediment intended to include detailed evaluation of the distribution of
sedimentation properties. It was obviously necessary to develop an alternative correlation
scheme for the organic sediment that did not depend upon shear stress calculations and
which reconciled the depth differences observed in the two extended runs.
Alternative Erosion Parameter Correlation Scheme The development of the
alternative correlation scheme was done by considering the limiting case of a simplified
model of an eroding system with significant surface roughness and appreciable
sedimentation potential.
Consider the sediment system as being composed of classes of sediment,
where classy is defined as all sediment that will erode at shear stresses between Xj-i and Xj,
with to = zero. Consider a control volume of length 1 cm, width w cm, and depth h cm,
with surfaces parallel to the normal planes and flow parallel to the bed. Assume the
following mechanisms:


306
overprediction on the combination of both events. This is almost identical with the 3.4%
over prediction observed for the combined TSS loads for the six calibration events.
The ultimate particulate phosphorus loads were not quite as close. For the
simulation periods of the two events the combined particulate phosphorus load observed
was 110.87 kg compared to a simulation value of 98.31 kg, an 11.3% underprediction on
the combination of both events. This compares reasonably well to the underprediction of
7.8% in the combined total for the calibration runs.
The late season events transported from 1.4 to 2.3 times as much water as Event
285, the largest of the calibration events. Considering the dramatically expanded
hydraulics of the late season and the potential for seasonal effects to manifest themselves
late in the year the agreement between simulated and observed suspended solids loads is
remarkable.
Figures 10.28 and 10.29 seem to indicate that there was a trend for the model to
underpredict phosphorus content early in the event and then partially compensate by
overpredicting later in the event. This could possibly be the result of seasonal effects,
where lower water temperatures would lead to somewhat slower diagenesis, which in turn
would give the more readily transportable matter a higher phosphorus content.
The total mass of material transported in these events was greater than that
transported in the calibration events, so the extension of the suspended solids-phosphorus
content correlation to the end of the events may not have been totally valid because of an
exhaustion of material typical of the calibration events. Given these considerations the
agreement between simulated and observed particulate phosphorus is certainly
acceptable.


148
This curve was fit by the equation:
V_ = -4.6628 + 0.1115 x (RPM) 9.4263 x 10-5 x (RPM)2 6.32
where RPM =grid oscillation frequency, revolutions/min. Equation 6.32 is the
calibration equation that was used for all subsequent erosion studies on field sediments.
Field Measurements using the calibrated Particle Entrainment Simulator
At the time the samples were taken for field PES evaluation the decision was
made to expand the scope of the field portion of the study to include both farms UF9200
and UF 9206. UF 9200 was still designated as the primary target farm but UF9206 was
added for additional monitoring to provide a back-up site in case of loss of UF9200 for
any reason and also to provide comparison data to UF9200.
Recall from Chapter 5 that UF9200 was a 1280 acre (518 ha) farm with two main
canals and one pump station, planted only to sugarcane, which was representative of
medium to large actively managed sugarcane-only fields with relatively deep soil.
UF9206 was a 1750 acre (710 ha) farm with a complex drainage pattern that included
three main canals, several perimeter canals, two pump stations, and gated interconnects
among the various sections of the farm, with multi-crop plantings that included stands of
sod and sugar cane and rotated plantings of rice and winter vegetables that was
representative of medium to large actively managed multi-crop farms with relatively deep
soil. Because of its cropping patterns and hydraulic layout UF9206 typically undergoes
more active water management than UF9200, with periodic shifts of water inventory from
one on-farm location to another.
The time of sampling was early July, about one month into the rainy season. At
this time UF9200 was planted completely to sugarcane, UF9206 was planted to only two
crops, sugarcane (~75%) and sod (-25%). UF9200 had undergone an extensive ditch and
canal cleaning and maintenance program about two months prior to sampling. UF9206


44
electrochemical charges present that may be sufficient to affect some degree of
flocculation in the suspended phase. Given their source, there is definitely the potential
for supply limitation, particularly in larger channels and flows. Their specific gravities
may be considerably less than those of the inorganic materials typically found in
sediments so the buoyancy forces acting on them may have a much greater impact than on
inorganic sediments. For these reasons it may be appropriate to approach organic
sediments as hybrids between cohesive and non-cohesive sediments and attempt to adopt
appropriate techniques from the study of each of these two types of sediment.
Several reviews (Kranck 1984a, Mehta 1984, Mehta 1988, Mehta et al., 1981,
Mehta et al., 1989. Parchure and Mehta, 1985) cover the practical aspects of cohesive
sediment transport well. The following summary draws from these reviews.
Erosion Modeling of the instantaneous erosion rates of cohesive sediments can be
approached from the perspective that the critical shear stress for erosion corresponds to
the shear strength of the eroding sediment (Mehta et al., 1981). The key to developing
analytical expressions for cohesive sediment erosion is the determination of the bed shear
strength and the form of the rate of erosion as a function of the shear stress which is
applied to the sediment bed in excess of the bed shear strength.
Beds that are artificially placed in erosion simulation devices for study (placed
beds) usually have constant properties in the vertical direction, which makes their study
relatively simple. The instantaneous erosion rates for these beds may be expressed in the
normalized form
6
T,
where e
Tb
Erosion rate, kg/sec-m2
Erosion rate constant, kg/sec-m2
Shear stress applied to bed, n/m2
3.5


45
Ts = Bed shear strength, n/m2
although formulations have been proposed that make the erosion rate exponential with
respect to the excess shear stress. Note that for a specific constant bed shear strength. Em
and Ts might be combined in Equation 3.5.
Beds that are deposited, under either flow or non-flow conditions, will normally
have properties which vary with vertical position within the sediment. Parchure and
Mehta (1985) have presented an analysis that correlates the depthwise variation of shear
strength of specific sediments with the vertical increase in sediment density. They
proposed an equation of the form
3.6
where Ef = floe erosion rate, kg/sec-m2
a = rate coefficient, m/N5
elevation of sediment surface above datum plane, m
z
The floe erosion rate represents a non-zero erosion rate when the bed shear stress
equals the bed shear strength, a recognition that some erosion occurs at this point because
of random turbulent fluctuations at the bed surface. Given the (known) relationships
between sediment depth, density, and shear strength, and given a known applied shear
stress pattern, Equation 3.6 may, in theory, be used to predict sediment suspension
throughout the course of an erosion event.
Agglomeration and De-agglomeration Particle-particle collisions, caused by
Brownian motion, velocity gradients within the suspending fluid, and differential rates of
settling among particles of various dimensions and densities may give rise to flocculation
or agglomeration of suspended matter, increasing the dimensions of the particle-
collection structure, which generally tends to increase the sedimentation rate of the
particle-collection. Turbulence may increase the opportunity for agglomeration but it also


317
UF9200 with the practice in place. We will also take an overall look at the results from
UF9206 in light of the model results from UF9200 and attempt to extrapolate some
general guidelines for particulate phosphorus transport management.


298
The empirical curve suggests the existence of two zones of temporal influence.
First at interevent times less than around 300 hours there may have been a zone with a
relatively constant level of initial erodable mass independent of interevent times. Beyond
about 300 hours there may have existed a zone where the initial erodable mass begins to
increase with interevent time. This type of response may be rationalized as follows.
Both the field studies and the model indicated that the bulk of the exported
suspended solids was sourced from the lower ends of the canal system. For the purpose
of making a first approximation, consider that the exported suspended solids came from
the lower 25% of the North and South Canals and all of the East Canal. The dimensions
given in Table 7.1 may be used to estimate the combined area of this system as 17600 m2.
Now note that the longest interevent time and the highest value of Cem<0)
coincided with Event 220. the first event of the calibration series. The next event of the
series was Event 237. which had a Cem(0) of 350 gm/m2. The subsequent three events,
252,258 and 262, all had values of Cem(0) in the range of 300-350 gm/m2 and were within
four to six days of one another. The maximum mass of suspended solids exported during
any of these three events was 1315 kg. Considering the above area of 17600 m2 as the
source area we can estimate a maximum contribution of around 75 gm/m2 for each of
these three events, which is only 20-25% of the value of Cem calibration process.
What we seem to have here is a high starting value of Cem(o> arising from the long
interevent time preceding Event 220 reduced to a baseline value of Cem(0) of around 350
gm/m2 by the combined flushing effects of Events 220 and 237. The subsequent three
events (recall that we have eliminated Event 244 from the calibration process) do not
deplete the baseline to any great extent so there is a rapid return to the baseline value of
around 350 gm/m2. This rapid return may be similar to the quick regrowth observed in
the studies of epiphytic material on the roots of aquatic macrophytes in the Amazon flood
plain by Engle and Melack (1990). It may also arise from redistribution of the aquatic


235
At the end of the collection period both tanks were washed thoroughly with
membrane-filtered ambient water, which was added to the lab sedimentation vessel. The
underflow from the sedimentation vessel was further concentrated in an Imhoff cone and
then stored at 4C until analysis. The analyses were conducted according to the methods
detailed in Appendix C.
A total of six large-scale composite samples were collected. Two were taken at
UF9206N during Events 236 and 256. Four were taken at UF9200 on Julian Dates 237.
241,252, and 258. Retrospective examination of the UF9200 pumping patterns showed
that the Julian Date 241 sample was part of Event 237, so three events were sampled on a
large scale basis at UF9200.
Adsorption/desorption tests
In the case of the Large-Scale Composite Samples the sorption characteristics
were evaluated at only two concentrations because of limited sample mass. Under these
circumstances there were insufficient data points for each sample to use the non-linear
optimization procedure that was used on the earlier isotherm tests, reported in Chapter 5.
Instead the more traditional method of data analysis was used. Figures 8.1 through 8.4
show the equilibrium substrate phosphorus adsorption concentration for both the
adsorption and desorption modes (technique described in Chapter 5) for the several
samples.


212
results of the field erosion studies reported in Chapter 6. To do this we will make use of
some of the representative values presented in the Detailed Hydrography section of this
chapter. If we take the total surface area of the conveyance system to be 129,500 nr
(distributed as 72,400 m2 of ditches and 57,100 m2 of canals) we may calculate the
exported suspended solids loads of each event on an areal basis and recast the results in a
way that allows direct comparison with Figure 6.23. the erosion test results from UF9200.
The results, for Events 220-285, are listed in Table 7.7 where the event average
concentration (total suspended solids load divided by total hydraulic load) is shown and
the areal load is expressed both on the basis of total conveyance area and canal area only.
Table 7.7: UF9200 Events 220-285 Expressed in Areal Loading Terms
I Event
TSS Load
kg
Hydraulic Load
m3
Event Avg.
TSS Cone.
mg/1
Areal TSS Loading
(Total)
gm/cm2
Areal TSS Loading
(Canals Only))
gm/cm2
220
11031
1.99 x 10!
55
0.0085
0.0193
237
4368
3.18 x 10s
14
0.0034
0.0077
244
1132
0.39 x 105
29
0.0009
0.0020
252
1315
0.55 x 10s
24
0.0010
0.0023
258
1146
0.89 x 105
13
0.0009
0.0020
262
861
1.45 x 10s
6
0.0007
0.0016
285
6280
3.21 x 105
20
0.0048
0.0109
Referring back to Figure 6.23, recall that of the eight erosion tests run on surficial
sediment from UF9200 one did not show any evidence of erosion six fell in a category
of having roughly the same slopes and relatively low CSMEs, and one had a much


253
In this form the transport equation closely resembles the flow equations so similar
numerical techniques may be applied for its solution.
The key element in the DUFLOW water quality module is the sub-routine that
describes P, the production (and consumption) term. The DUFLOW computational
package provides a discrete external model file that is created by the user following a few
defined syntax rules. This ASCII model file is then compiled by an interpreter called
DUPROL into a form that allows its output to be read by the program as input data at
each time step. This format allows the user to modify the quality program as needed
without having to recompile the entire program.
The configuration of the quality module does have some limitations, however.
The specific format requires that all processes be described as rate equations that
ultimately are consolidated as zero order and first order rate constants that are passed out
to the computational module at each iteration. Casting the process as a collection of zero
and first order rate processes is not a problem, but this format, which deals only with
rates, does not allow the user to specify any constant outputs, or baseline concentrations.
The computational procedure requires that the state variables used to calculate the
rate processes be evaluated at the previous time step. This requirement makes the
evaluation of the process description explicit, which reduces the precision of the quality
calculation process and allows more opportunity for instabilities in the quality
calculations. This program limitation manifests itself in sensitivity of concentration to
time step size under some circumstances. Quality time step must be equal to or greater
than the hydraulic time steps so in cases where short quality time steps are desired the
hydraulic program time step must be reduced appropriately. In general, however, the
configuration of the quality module of DUFLOW provides a convenient and flexible
mode of integrating water quality processes with the hydraulic calculations.


BIOGRAPHICAL SKETCH
The author. James D. Stuck, was bom in Charleston, West Virginia, on April 20.
1940. He received a B.S. degree in chemical engineering from West Virginia University
in 1963, and worked for Olin Chemical Corporation for five years in Niagara Falls. New
York. He began graduate study in chemical engineering at the State University of New
York at Buffalo in 1968. receiving his M.S.ChE in 1970 and his PhD in 1973.
Upon completion of his first round of graduate education he went to work in the
Development Department of Linde Environmental Systems Division of Union Carbide
Corp. in Tonawanda, New York. He subsequently moved to the Process Engineering
Department where he held various managerial positions in pilot plant operation, technical
service, and process design. In 1980 he transferred to the Engineering Development
Department of Union Carbide Agricultural Products Company in Jacksonville, Florida,
working out of the Woodbine. Georgia, plant. For the subsequent ten years he held
various technical management positions in UCAPCo and in Rhone-Poulenc Corp., when
that company acquired the agricultural chemicals business of Union Carbide.
In 1990 he made the decision to leave the world of economic poisons and return
to the environmental field, enrolling in the Environmental Engineering Sciences
Department of the University of Florida and receiving his M.E. in 1991. He was then
fortunate to receive a research fellowship in hydrologic science from the USDA that
allowed him to pursue a second PhD degree in agricultural and biological engineering at
the University of Florida. At the time of completion of this work he resides in Atlantic
Beach. Florida, which he intends to make his permanent residence for the duration.
404


17
to the mean. He concluded that spatial variation (upstream to downstream) was not
statistically significant. Seasonal variation in the West Palm Beach Canal was also
difficult to detect in his results because month to month variations were quite significant.
What is particularly interesting in Lutz' data, however, is what arises from visual
inspection of the various time series. In about 15% of the days sampled there are
appearances of large spikes at one individual sample point which were not reflected at the
other sample points. This is true for soluble as well as total phosphorus. If the data plots
are examined from a subjective phenomenological viewpoint one might possibly interpret
the data as showing wave-like gradients in concentration down the canal on occasions of
high concentration, however such an interpretation is very subjective, given the absence
of close-interval time-series data. The main point from this study is that the data, taken at
face value, indicate that there were times during the study when the canal was seeing
short bursts of shock load with a more subjective indication of wave-like phosphorus
transport under certain circumstances when phosphorus concentrations were high.
Water Conservation Area water quality correlations
Mattraw et al. (1987) attempted to develop statistical correlations of water quality
variables to detect trends at 10 inflow or outflow locations around the periphery of Water
Conservation Area 3A. Several of the structures were direct outflows from the EAA.
They evaluated numerous model types relating to discharge and antecedent rainfall
conditions and concluded that the only correlations of significance were orthophosphate
and nitrate with 7 or 30 day antecedent rainfall conditions. No temporally significant
trends were found over a five year period. The primary result of interest in the context of
the present research is that for the one structure evaluated which corresponds to a major
EAA discharge, S-8, the best model correlation had an R2 of only 0.36. This emphasizes
the difficulty in correlation of phosphorus content using only hydrologic variables.


179
phosphorus content material is entrained in the water stream so the average phosphorus
content goes down.
The difference between the solids phosphorus content during the first four hours
of the events and the remainder of the data may be a crude representation of biological
differences among various elements of the transportable population, and also an
indication of important short term time dependencies, where time dependencies can be
both inter-event and intra-event. Clearly there is no implicit reason why there should be a
demarcation in phosphorus content at the end of four hours, and probably there is not.
Given the level of resolution inherent in this study the discrimination of differences
beyond four hours was difficult at best. It might be expected that a larger population with
more precise sampling techniques and more thorough statistical analysis might yield time
discrimination well beyond the four hour mark. At present the lack of time
discrimination beyond the four hour time limit probably contributes to the scatter evident
in the remainder of the event data.
In spite of the obvious data scatter, the trends are unmistakable. There is a
definite segeeation effect with respect to phosphorus content and particle transport. This
is the kind of trend that would be expected to arise from the processes variously known as
mineralization, humification, and diagenesis. This trend was alluded to in the discussions
of Chapters 2 and 3 and was sought but not found in the surficia! sediment particle size
distribution studies discussed in Chapter 5. The implications of this finding are important
in that they support the interaction between biology and phosphorus transport and also
provide one more tool for attempting to track the source of the exported phosphorus.
The correlations presented as Equations 7.1 and 7.2 may be recast in a form more
suitable for model application, as follows:
PPX = 0.01897 x (7XS)06574
7.3


246
Exclusion of Overland Flow Erosion and Sorption from the Model
The one thing Overland Flow and Sorption have in common in the context of this
work is that they were both excluded from the model. The two processes are discussed
together in this section from that standpoint.
Overland Flow Erosion The current thinking at the time of inception of this study
was that overland flow and soil and bank erosion would make a significant contribution
to the suspended solids load of the farm discharge in the EAA. This proposition was
investigated from several perspectives, specific to the target and back-up farms UF9200
and UF9206.
First, the specific soil contour and drainage system configurations were evaluated
for runoff potential. At both farms the primary mode of drainage is subsurface transport,
with hydraulic conductivity enhanced where possible by the construction of mole drains
that are tubular sub-surface channels cut into the soil by use of a bullet-shaped probe on
the end of a knife-edged plow inserted to near-marl depths and drawn by tractor
perpendicular to adjacent field ditches. The conveyance system construction and
maintenance procedures are such as to create an elevated bank at the edge of the field
ditch, which effectively functions as a dam. Field maintenance procedures are designed
to preserve that embankment. Where it is necessary to provide for field draining directly
to the field ditch, the drainage channels are introduced as far upstream of the ditch outlet
as possible. In general the farm fields are designed to prevent open field runoff.
Second, the frequency of potential for significant open field runoff was evaluated
at UF9200 by examining the field well records over the time period that covered the
major event span of the normal wet season, Events 220 through 285, to determine the
fraction of the time the field groundwater level rose to the surface. The author directly
observed the runoff conditions at UF9200 periodically over two seasons and found that at


321
The results look impressive. The solids load at 100.000 nr' of discharge was
reduced from -10,420 kg to -2360 kg, about a 77% reduction. Clearly the benefit of
selective removal of transportable mass is exhibited for one event. It is informative,
however, to look at the simulation of the movement of the upstream transportable mass
which was not removed. Figure 11.3 shows this simulation for the South-East Canal
system. The figure shows simulated concentrations of transportable mass in each section
as a function of time into the event. For conciseness the individual sections have not
been labeled, but the sections that have erodable mass concentrations starting at zero are
in the trap section, while the sections that have erodable mass concentrations starting at
1800 gm/m2 are the sections upstream of the South-East Canal trap.
Simulated Erodable Mass Concentrations at Selected Sections of South-East Canal
for Trap Simulation
Figure 11.3: Simulated Erodable Mass Concentrations at Selected Sections of
South-East Canal for Trap Simulation
Note the relatively steady progression of erodable mass downstream in a wave
like pattern. In the simulation this progression causes several of the sections in the trap.


263
The final calibration parameter set, shown in Table 9.1, had constant hydraulic
conductivity across all events but the flow resistance was allowed to vary for each event.
This is justified on the basis of the potential for the macrophyte population to change
from event to event and affect a change in channel resistance, and from the fact that the
system was being modeled without spatial variability of channel or soil properties, so
these variations would be taken up as variations in the parameters allowed to float.
Table 9.1: Results of Hydraulic Calibration for UF9200
Model Parameter
Final Value
Event
Calibration Value of
Mannings n
Impeding Layer Elevation
1.30 m
220
0.059
Ditch Bottom Elevation
1.85 m
237
0.052
Field Surface Elevation
3.10m
252
0.059
Drain Spacing
200 m
258
0.059
Ditch Radius
100 cm
262
0.056
Hydraulic Conductivity
55 m/day
285
0.050
Manning n-Surface Runoff
0.05
Avg
0.056
Runoff Factor
1.0
Std. Dev.
0.004
Initial Abstraction
1.0 cm
The average of 0.056 for the farm-wide Mannings n value is fairly high for a
non-meandering system that has relatively gradual changes in channel width throughout
most of its length. The South Florida Water Management District recommends a value of
0.035 for n for these types of channels. Chow (1959) in his examples of typical


142
calculated from these estimates may not reflect the actual local shear stress seen by the
sediment bed.
In the approach taken, by necessity, in this study the erosion parameters are
correlated with the square of a characteristic velocity. By doing this there is an inherent
assumption made that the roughness coefficient of the bed surface is constant, thus the
procedure is implicitly incorporating an average roughness coefficient for all correlations.
For the first approximation level, which is the objective of this work, it is probably better
to assume that the average implicit roughness coefficient developed in the course of this
work is representative of the average organic sediment bed than it is to attempt to
estimate the roughness from some external model. In addition the choice of some
characteristic velocity as the correlation parameter yields a lab and field parameter that
can be estimated with reasonable accuracy, provided the proper characteristic velocity is
chosen.
One step remains before the PES and CRAF data is extrapolated to field
conditions, and that is the choice of velocity for correlation. Hydraulic and hydrologic
parameters are frequently correlated using the mean channel velocity which is simply the
volumetric flow at a point in a channel divided by the channel crossectional area at that
point. For fully developed uniform flow in a rough surfaced channel the flow profile is
logarithmic and the mean velocity can be related to any other, including the maximum, by
standard formulas, such as
6.29
for hydraulically rough (Re > 70) flow and


286
to Node 100 which is 94 meters upstream of the pump station in the East Canal. The
placement of all six nodes is as in Table 10.2
Table 10.2: Placement of Example Nodes for Event UF9200-327 Simulation
Node
75
90
96
98
99
100
Distance Upstream from
1 Pump Station-meters
1609
1006
804
523
355
94
Note that Figure 10.7 illustrates the phenomena that was referred to earlier as the
suspended solids moving downstream in wave-like fashion. Each of the nodes goes
through a suspended solids concentration maximum that generally appears at a later time
and with a higher magnitude as the position proceeds downstream. This is attributable to
several factors. First, the concentration seen by a given packet of water moving
downstream increases with time in-stream as long as net erosion exceeds net
sedimentation. Secondly, in the North and South Canals velocity is increasing with
distance downstream because of inflow from field ditches, so the driving force for erosion
is increasing as position moves downstream. These two factors tend to cause the increase
in concentration.
Counteracting these processes is the loss of erodable mass in the channel, which
tends to reduce erosive suspended solids production. This process is clearly illustrated in
Figure 10.8, where it can be seen that the three most upstream nodes consistently lose
erodable mass because erosion is greater than sedimentation of material from farther
upstream. The three downstream nodes first increase in erodable mass because net
sedimentation of material from upstream exceeds erosion of material at that location, then
eventually start to decrease as their upstream supply of sedimentable material decreases.
*


64
UF9206 A 1750 acre (710 ha) farm with a complex drainage pattern that
included three main canals, several perimeter canals, two pump stations, and gated
interconnects among the various sections of the farm. This was a multi-crop farm
that maintained stands of sod and sugar cane and rotated plantings of rice and
winter vegetables at various locations. This farm was representative of medium to
large farms with relatively deep soil that maintain multi-crop patterns and are
actively managed.
Sediment survey measurement methods
Water depth and sediment depth transects were takeitby boat from cordoned canal
sections. At three foot (0.91 m) increments, the distance from the water surface to the
sediment surface was determined by lowering a neutrally buoyant one ft2 (0.09 m2) pad
attached to a calibrated rod until it rested on the sediment surface. At the same location
the distance from the water surface to the canal bottom was determined by driving a 0.5
in. (1.3 cm) diameter calibrated steel penetrometer rod through the sediment until it met
resistance at the marl surface. Sediment depth at each location was determined by
difference between water-surface-to-sediment-surface depth and water-surface-to-marl-
surface depth.
At each site water and sediment depth transects were taken at multiple locations
depending on the site main canal configurations. Four transects were taken at UF9202,
eight at UF9200, and twelve at UF9206. Two sediment core samples were taken at each
site using the piston core sampler described in Appendix A. The cores were expunged
on-site into plastic bags, sealed, refrigerated, and transported to the Gainesville, FL,
laboratory for bulk analysis. Analyses were conducted according to the methods
described in Appendices B and C.


shear for erosion, supply limitation of erodable mass, and reversible erosion-
sedimentation.
Field studies showed the exported PP also had weak sorption characteristics but
had higher, more labile phosphorus content than farm sediments. Additional studies led
to the hypothesis that canal biological growth was the primary source of exported PP.
The model, modified to incorporate this concept, was calibrated on the PP discharge data
from the target farm during a normal wet season. It was validated on two tropical
depression events, simulating PP export within about 10% of actual.
Farm management practices to reduce exported PP were simulated using the
model. Macrophytereduction and velocity control were shown to be critical factors.
Sediment traps that do not reduce velocity appreciably were shown to be of limited value.
In the simulations, canal modifications that reduced pump intake velocity reduced the
first flush phenomenon, where a large fraction of the PP load is discharged at pump start
up. Level control and elimination of pump cycling were shown to be cost-effective PP
reduction mechanisms. The removal of untransported detritus from main canals was
shown to be necessary for long-term control. Hydraulic transport schemes to achieve
removal were recommended.
The model may be adapted to other farms to optimize water table control practices
which minimize PP export.
xxviii


Selected Water Levels-UF9200
Julian Date
Field AI6 Pump Station
Figure E.2: Selected Water Levels at UF9200Julian Days 150-300



Figure 1.1: Original Everglades Watershed (Adapted from Jones, 1948)


119
was greater than it was during suspension, thus deposition was enhanced at the lower
shear rates. It may be appropriate to point out at this time that there was some indication
of agglomeration during the erosion runs. Close examination of Figure 6.6 shows that at
the three lowest shear levels there was an indication of overshoot of the CSME at the
knee of the erosion curve for each shear level. That is, the CSME was higher 8-12 hours
into each shear level run than it was later into the run.
The important observation to make is that, considering the above rationalization,
the two curves are quite similar in terms of mass transported at a given shear rate. The
implication here is that, unlike some fine-grained cohesive sediments, the prototype
sediment deposits about as readily as it erodes. If we adopt the rationalization and accept,
for the sake of illustration, the argument that the two curves are approximately congruent
then the results may be interpreted qualitatively as follows.
Recall the discussion of erosion and deposition of fine-grained cohesive clays
from Chapter 3 where, separately, the cumulative mass eroded was shown to be a
function of bed parameters, and the cumulative mass deposited was shown to be a
function of suspended solids parameters. Specifically the erosion mass depended on bed
depth and density profiles down to the depth where bed shear strength equaled the applied
shear stress (Mehta, et al, 1981), while deposition mass depended on initial
concentration, suspended solids particle size and settling velocity distributions, critical
deposition shear stresses, and applied shear stress.
With the erosion and deposition extents depending on entirely different properties
there is no reason to expect, a priori, that erosion and deposition would be reversible, that
is, curves such as Figure 6.10 would inherently be expected to exhibit hysteresis. If we
apply the simplification of assuming that the mass suspended during each shear level
interval was of uniform character and then use the integrated Krone expression for
deposition (Mehta 1988) which is


117
Prototype Organic Sediment-Consolidated Bed
Stepwise Shear Stresses
(Deposition 30.5 cm Water Depth)
0.018
0.016
0.014
A 0.012
I 0.01
8 0.008
T3
5 0.006
| 0.004
0.002
0
0 10
1.35 dyne/cm2
*0- .
20 30 40
1.65 dyne/cm2
-o- .
0.95 dyne/cm2
50 60 70 80
Time-hr
0.60 dyne/cm2
-O
90 100
Figure 6.9: CSME as Function of Shear Time at Several Calculated Shear Rates for
Deposition of Consolidated Prototype Organic Sediment Bed at 30.5 cm Water Depth
Table 6.2: Constants for Fit of Equation 6.8 to Multiple Shear Deposition Tests
with^WaterDegtlutfSthScn^
Tau -
dynes/cm2
a gm/cm2
k -hr*1
ak
gm/cm2-hr
CSMEo-
gm/cm2
Asymptotic
CSME -
gm/cm2
r2
1.65
0.003141
0.099
0.00031
0.01694
0.01380
0.975
1.35
0.00425
0.118
0.00050
0.01351
0.00983
0.999
0.95
0.00445
0.188
0.00083
0.00863
0.00418
0.989
0.60
0.00225
0.137
0.00064
0.00286
0.00077
0.993
The deposition curve can also be correlated well (r2 =0.996) by the linear equation
CSME = 0.0122571-0.00686
6.9


290
Event UF9200-220
CUM TSS Load
12000
....
Cum TSS Load-kg
//
0 ^
0 50,000
100,000 150.000 200.000
Cum Hydraulic Load-mA3
Observed Simulated
Figure 10.9: Simulation Results-Cumulative Suspended Solids Load for Event
UF9200-220
25
Event UF9200-220
CUM PP Load
20
OO
!5 /
i .o y
3 /
U
5
o y~~
0 20.000 40,000 60,000
80.000 100,000 120,000 140.000 160,000 180,000 200.000
Cum Hydraulic Load-mA3
Observed Simulated
Figure 10.10: Simulation Results-Cumulative Particulate Phosphorus Load for
Event UF9200-220


112
CSME = 0.0030 + 0.0048(1 e"2 m') + 0.0094(1 e1321) 6.4
Equation 6.4 suggests the presence of several subdivisions of sediment, each of
which has individual erosion characteristics. The constant term would represent material
that is almost instantaneously resuspended at the applied shear stress, while the two
exponential terms would represent two sediment components that have an increased
resistance to erosion. This is admittedly an extremely simplified characterization of the
system, but one that is descriptive and readily understandable.
Multiple Shear Extended Runs A more detailed evaluation of the erosion
characteristics of this sediment was obtained in the next extended run. The bed was
allowed to consolidate for 7 days, then shear stresses were applied in ascending steps of
0.96,1.30, 1.64, and 2.0 dynes/cm2 as calculated by Equation 6.3. At each level the shear
stress was held constant for 24 hours before the next step was implemented. The results,
presented in Figure 6.6, show a family of response curves similar to Figure 6.5.
Figure 6.6: CSME as Function of Shear Time at Several Calculated Shear Rates for
Erosion of Consolidated Prototype Organic Sediment Bed with 30.5 cm Water Depth


13
standards which apply well to other locations in the United States might not translate
completely to the Everglades environment. This assumption has underlain the planning
and execution of the research work in this dissertation. With this in mind it may be
appropriate to first examine the limited body of work available which relates directly to
particulate matter and particulate phosphorus transport in the Everglades Region.
Farm-Field Scale Studies
There are two studies which have been done on the field or farm scale which are
particularly pertinent to phosphorus transport in the EAA. They are those done by the
engineering firm CFI2M Hill (1978) on the farm scale and by Izuno et al. (1991) on the
field scale.
CH2M Hill studv
T~
The CH2M Hill study evaluated a number of water quality parameters and was
conducted over a period of 15 months in 1976-77 on multiple farm sites within the EAA.
Intensive study was directed at one sugarcane field, one cattle ranch, and one vegetable
farm which grew a variety of vegetables throughout the study period. In addition multiple
secondary sites for each land use were studied less intensively and served as checkpoints
for the representativeness of the intensive data. Key conclusions from the study regarding
phosphorus follow.
The sugarcane farm showed study-period average effluent total
phosphorus of 0.110 mg/1, soluble phosphorus of 0.070 mg/1 (64% of
total), particulate phosphorus of 0.040 mg/1 (36% of total) and an annual
effluent total phosphorus load of 0.65 kg/ha.


124
the relatively good reproducibility of CSME versus shear under repeated similar
experimental conditions at the same depths. The conclusion which had to be drawn from
these results was that the depth dependent relationships assumed to this point in the
analysis of sediment response to shear were invalid for the organic sediments under
current test. There are at least two potential reasons for this.
First, the shear stresses calculated using equations 6.1 and 6.3 may not in fact be
the shear stresses actually being applied to the organic sediment bed. The possible
explanation for this is that the original correlations with clays represented a system where
the roughness or friction factor of the Fiberglas walls of the flume was on the same order
of that of the clay bed, so that the walls dissipated an appreciable fraction of the energy
applied by the ring. Using Mehtas (Mehta 1973) correlations for shear applied by the
ring and shear recorded by the false bottom it may be shown that for a AV range of 20-70
cm/sec the shear transmitted to the bottom is a relatively constant 16.9-17.1 % of the shear
applied by the ring for the 15.2 cm depth and a more variable 7.4-9.9 % of the shear
applied by the ring for the 30.5 cm depth, so from about 83-93% of the applied shear was
being dissipated by the walls in these cases. In the case of the organic sediment, however,
it is obvious from visual observations that the bed roughness is considerably greater than
that of the Fiberglas walls. Under these circumstances the fractional contribution of the
walls could become negligible and the correlations of bed shear with depth would become
inaccurate.
Second, the prototype organic sediment showed appreciable sedimentation at all
shear levels (see previous discussion of Figure 6.10). Appreciable sedimentation is
defined in this case as having a critical shear stress for sedimentation that is greater than
the applied shear stress, so that probability of deposition of each class of particles is
greater than zero when the applied shear stress is less than or equal to the shear stress
originally applied to erode the particles in that class. This is different from much of the
prior erosion work done in the CRAF using clays. In the case of clays the sedimentation


239
Figure 8.6: UF9206N Desorption Data Eadie-Flofstee Plot


227
condition of the lyngbya no estimates of areal density were made. These samples were
processed in the same manner as the surface macrophyte samples and analyzed for
phosphorus content. This series of macrophyte and algae samples was taken over the
period Julian Dates 240-248. The results are shown in Table 8.4. The areal dry weight
Table 8.4: Macrophyte Mass Density, Volatile Content, and Phosphorus Content
Macrophyte
Areal Density
gm/m2
Volatile Fraction
Phosphorus Content
mg/kg
Water Lettuce
663
0.729
5794
u
606
0.715
5363
u
611
0.719
5898
Avg(SD)
627(32)
0.721(0.007)
5687(284)
Water Hyacinth
1472
0.826
4043
1356
0.830
4452

1776
0.806
3843
Avg(SD)
1535(217)
0.821(0.013)
4113(310)
Pennywort
976
0.758
5448
1075
0.786
4718
1296
0.760
4312
Avg (SD)
1116(164)
0.7687(0.016)
4826(576)
Filamentous Algae
NA
0.501
2418
NA
0.521
2135
-
NA
0.516
2221
Avg (SD)
NA
0.513(0.010)
2258(145)


probably require mechanical changes such as weir installation or floating pump
intakes that should not be made casually. Experimentation with temporary
structures would be in order to test the effectiveness of this concept.
Level Control Controlling to eliminate extreme downward excursions in
level was shown to be a potentially productive policy with minimal investment of
time or resources. In the longer term there may well be a site-specific optimum
lower level control point that could be found by evolutionary' operations at each
site.
Pump Cycling Substitution of lower pump speed or transfer of operations
to a smaller pump should be practiced in place of on-off cycling of large pumps to
avoid the high maximum velocities incurred during cycling.
Velocity Reduction Pumping policies that reduce the maximum stream velocity
have the potential to be extremely productive in reducing particulate phosphorus
export. They also have the potential for not remobilizing the resulting unexported
material during subsequent events.
Model framework development
1. The model response to major rainfall during pumping events should be
improved by, for example, incorporating spatial and water-content
dependent functionalities into the field hydraulic conductivities, or
modifying the shape of the elliptical water table curve assumed in the
drainage equations.
2. The model calibration process should be expedited by developing interface
programs to link DUFLOW to statistical evaluation or, ideally,
optimization programs.


sedimentation phosphorus flux
weight rate of bed transport per unit width
water column height or actual canal water surface elevation
Julian Date
partition coefficient or surface roughness
grain roughness
saturation constant for sorbate-substrate system
saturated hydraulic conductivity
water level difference between mid-field and ditch
Mannings coefficient of roughness
production of a constituent per unit section length
mass of phosphorus desorbed from sediment
phosphorus content of discharged TSS for the first four hours of an event
phosphorus content of discharged TSS for the remainder of an event
phosphorus content of discharged TSS for station UF9206N
phosphorus content of discharged TSS for station UF9206S
particulate phosphorus concentration in the discharge for the first four
hours of an event
particulate phosphorus concentration in the discharge for the remainder of
an event
volume rate of water flow per unit width
volume rate of suspended solids discharge per unit width
flow rate
wall Reynolds number
hydraulic radius of channel
grid oscillation frequency (revolutions/min)
ditch spacing


11.1 Effect of Transportable Mass on Exported Suspended Solids for Event
UF9200-220 319
11.2 Simulated Effects of Sediment Trap on Event UF9200-220 320
11.3 Simulated Erodable Mass Concentrations at Selected Sections of South-
East Canal for Trap Simulation 321
11.4 Effect of Pumping Modification (Simulation A) 324
11.5 Effect of Pumping Modification and Flow Width Modification
(Simulation B) 325
11.6 Effect of Pumping Modification and Flow Width Modification
(Simulation C) 326
11.7 Solids Recycle Simulation RCL1 Short Duration, Short Distance 328
11.8 Solids Recycle Simulation RCL2 Long Duration, Long Distance 329
11.9 Solids Recycle Simulation RCL3 Long Duration, Long Distance,
Higher Flow, No Backflow Prevention 330
11.10 Simulation RCL4 Long Duration, Short Distance, Higher Flow,
Backflow Prevention 331
11.11 Specific Effect of Reduced Velocity via Constant Discharge 333
11.12 Water Elevation Trends for UF9206 Pump Stations 336
11.13 Water Elevation Trends for UF9200 Pump Station 336
11.14 Correlation of TSS Concentration with Canal Level at UF9206N 338
11.15 Correlation of TSS Concentration with Corrected Canal Level at
UF9206S 339
11.16 Correlation of TSS Concentration with Canal Depth at UF9200 339
11.17 PP Content vs. TSS Correlations for All Three Farms 341
11.18 Cumulative Fractional PP Load vs. TSS for All Three Farms 342
11.19 Water Level Control Example for UF9206N 344
11.20 Water Level Control Example for UF9206S 344
xx


121
of this reduction was to obtain a consistency check on the system and to attempt some
runs at higher shear rates. After the end of the deposition run the bed was allowed to
settle overnight then water was drained out to reduce the water column height to 15.3 cm
(6 inches). The system was then subjected to a calculated shear stress of 2.0 dynes/cm'
for 24 hours and then allowed to consolidate for 7 days. After consolidation the system
was run at calculated shear stresses of 0.65, 0.92, 1.30, 1.70, 2.02, and 2.80 dynes/cm2.
The first 5 shear levels were sustained for 24 hours each, the last, 2.80 dynes/cm2, was
continued for 72 hours. The data, presented in Table 6.3 and shown graphically in Figure
6.11, provided some unexpected results.
Table 6.3: Constants for Fit of Equation 6.8 to Multiple Shear Deposition Tests with
Water Depth of 15.2 cm
1 Tau -
dynes/cm2
a gm/cm2
k -hr'1
ak
gm/cm2-hr
CSMEo -
gm/cm2
Asymptotic
CSME -
gm/cm2
r2
0.65
0.00000
0.000
0.00000
0.00042
0.00042
NA
0.92
0.00000
0.000
0.00000
0.00031
0.00031
NA
1.30
0.000184
0.754
0.00014
0.00031
0.00050
0.823
1.70
0.00141
0.090
0.00013
0.00079
0.00219
0.980
2.02
0.00195
0.074
0.00014
0.00229
0.00423
0.803
2.80
0.00462
0.131
0.00061
0.00434
0.00895
0.780


81
Figure 5.6: Correlation of Particle Specific Gravity with Organic Content for Hydrated
B9B10 Soil, B9B10 Sediment, and WPBC Sediment
Total Phosphorus Content-B9B10 Hydrated Soil, B9B10 Sediment, and WPBC Sediment
(Based on Total Solids)
Figure 5.7: Total Phosphorus (TP) Content Distribution of Hydrated B9B10 Soil, B9B10
Sediment, and WPBC Sediment


5
The relationship of the EAA with Lake Okeechobee was recognized in the 1970s
as being detrimental. A number of studies were conducted that linked the eutrophication
of the lake to agricultural nutrient inputs, of which the EAA was a major contributor
(Izuno and Bottcher, 1994). A number of remedial actions have been implemented to
reduce the nutrient load on Lake Okeechobee, including eliminating_backpumping to the
lake fromjhe EAA for all but a few small private drainage districts. The effect of this
policy is to force essentially all farm water discharge in the EAA to flow south to the
WCAs, which, in the absence of any remediating action, would further increase the
nutrient load vectored from the EAA toward the WCAs and the ENP.
Sources of Phosphorus in the EAA
Inputs to the potentially mobile phosphorus pool in the EAA can come from
fertilizer application, rainfall, inflow water from Lake Okeechobee, and the natural
mineralization process of the organic soil, where organic matter is converted to carbon
dioxide and inorganic compounds. Fertilizer phosphorus input to the EAA has been
estimated to be on the order of 18% of the total annual phosphorus input (Morris, 1975),
however the fraction of this input that is available for transport to the drainage water is a
strong function of fertilizer placement, crop uptake, soil microbial pool, and the iron and
aluminum content of the soil (Sanchez and Porter, 1994). Rainfall is a poorly quantified
but non-trivial source of phosphorus input. Studies on-going at the time of this writing
indicate that the phosphorus content of rainfall in the EAA may be on the order of 0.05-
0.08 mg/1 (Izuno, 1995). At an EAA area of 280,000 ha with an average annual rainfall
of 1.45 m the rainfall load might be in the range of 200,000-325,000 kg/year, or about
0.36-0.58 kg/ha-yr.
Precise estimates of the amount of phosphorus supplied to the EAA from Lake
Okeechobee are difficult to generate because of the flow-through nature of the canal


248
to form the original hypothesis. The original hypothesis that allowed exclusion of
adsorption/desorption and pore water release from the first stage of model development
was thus supported by field data and the exclusion was made.
Model Platform Selection and Hydraulic Calibration
The conveyance network model must have the capability of simulating unsteady-
state hydraulics in a complex branched network of pumped and controlled channels that
may be subject to distributed surface and subsurface inflow and outflow. Several
working finite difference models appeared to be adaptable to these requirements.
Model review and selection
Schaffranek (1987) presented a model used by the USGS for river and canal
network analysis which is based on the one-dimensional St. Venant equations, and is an
extension of work done by Fread (1974) at NOAA. Swain and Chin (1990) presented a
similar methodology that incorporates the effects of control structures and channel bed
leakage. Their computational algorithm is claimed to be faster than that of Schaffranek
for problems requiring long simulation times and more stringent convergence criteria.
Gichuki, et al. (1990) have presented a model that was developed specifically for
analysis of control schemes in surface irrigation networks. This algorithm is also based
on the one-dimensional St. Venant equations, but incorporates a computational scheme
that calculates submergence at transitions using the downstream depth of the previous
time step. This methodology is claimed to decouple each canal reach from the system,
thus reducing the number of computational steps in each iteration. Konyha, et al. (1992)
have coupled a field scale quantity model (DRAINMOD) for lateral inflow with an open-
channel stream-routing model modified to handle overbank flooding during extreme
events. Their routing model is based on the work of Amein (1972) at NOAA and appears


384
MU
1.000
1.000
0.830
0.830 25.6
STRU 47
347
46
47 11 2.26
0.48
2.87
MU
1.000
1.000
0.830
0.830 27.0
STRU 48
348
48
49 11 1.42
0.60
2.18
MU
1.000
1.000
0.830
0.830 0.0
STRU 52
352
52
53 11 1.86
0.48
2.47
MU
1.000
1.000
0.830
0.830 27.0
STRU 54
354
53
54 11 1.65
0.48
2.26
MU
1.000
1.000
0.830 0.830 27.0
STRU 55
355
55
56 11 2.52
0.32
2.92
MU
1.000
1.000
0.830 0.830 0.0
STRU 58
358
58
59 11 1.73
0.48
2.34
MU
1.000
1.000
0.830
0.830 27.0
STRU 60
360
59
60 11 1.69
0.48
2.30
MU
1.000
1.000
0.830
0.830 27.0
STRU 61
361
61
62 11 2.31
0.48
2.92
MU
1.000
1.000
0.830
0.830 27.0
STRU 64
364
64
65 11 2.04
0.48
2.65
MU
1.000
1.000
0.830
0.830 27.0
STRU 66
366
65
66 11 1.73
0.48
2.34
MU
1.000
1.000
0.830
0.830 27.0
STRU 67
367
67
68 11 2.06
0.48
2.67
MU
1.000
1.000
0.830
0.830 0.0
STRU 70
370
70
71 11 1.73
0.48
2.34
MU
1.000
1.000
0.830
0.830 27.0
STRU 71
371
71
75 11 1.49
1.68
2.56
MU
1.000
1.000
0.830
0.830 27.0
STRU 72
372
71
72 11 1.65
0.48
2.26
MU
1.000
1.000
0.830
0.830 27.0
STRU 73
373
73
74 11 1.76
0.48
2.37
MU
1.000
1.000
0.830 0.830 0.0
STRU 77
377
77
78 11 2.06
0.48
2.67
MU
1.000
1.000
0.830
0.830 27.0
STRU 79
379
78
79 11 1.76
0.48
2.37
MU
1.000
1.000
0.830
0.830 27.0
STRU 80
380
80
81 11 2.06
0.48
2.67
MU
1.000
1.000
0.830 0.830 0.0
STRU 83
383
83
84 11 2.01
0.48
2.62
MU
1.000
1.000
0.830
0.830 27.0
STRU 86
386
86
87 11 1.88
0.48
2.49
MU
i.000
1.000
0.830
0.830 27.0
STRU 89
389
89
90 11 1.70
0.48
2.31
MU
1.000
1.000
0.830
0.830 27.0
STRU 91
391
90
91 11 1.65
0.48
2.26
MU
1.000
1.000
0.830
0.830 27.0
STRU 94
394
94
96 11 1.07
1.91
2.29
MU
1.000
1.000
0.830
0.830 27.0
STRU 95
395
95
101 11 2.06
0.91
3.07
MU
1.000
1.000
0.830 0.830 30.0
STRU 46
346
46
50 11 2.08
1.20
3.05


115
CSME = 0.00954t 0.00352
6.7
where CSME has units of gm/cm2, and t has units of dynes/cm2.
The ordinate intercept of this equation is CSME = 0 at t = 0.36 dynes/cm2. The
value of 0.36 dynes/cm2 may be interpreted to represent the critical shear stress for
erosion of this sediment, a value below which no appreciable erosion will take place. The
functional relationship shown in Equation 6.7 forms, in a modified format, the basis for
the calibration of the field device discussed later.
Figure 6.8: Asymptotic CSME vs. Applied Shear Stress for Organic Sediment Extended
Run Erosion at 30.5 cm Water Depth
Deposition-Under-Shear Extended Runs Following the end of the multiple shear
extended runs the system was kept at the maximum shear stress of 2.0 dynes/cm2 for an
additional 24 hours and then shear stress on the bed was reduced in four steps to 1.65,
1.35,0.95, and 0.6 dynes/cm'. Again the shear stress was held constant for 24 hours at
each shear level. It should be emphasized at this point that the purpose of this series of


133
(0.5 inches) from center which produces a total top-to -bottom travel distance of the grid
of 2.54 cm (1 inch). The cam is driven by a variable speed 1/8 hp DC motor.
Figure 6.14: Schematic Diagram of Particle Entrainment Simulator (PES) (From Tsai
and Lick, 1986)


340
The preceding analysis may be used to draw some logical extensions that
potentially explain the observed differences between the solids and particulate
phosphorus export patterns at the two farms. The model applied to UF9200 illustrated
that the first leg of the U shaped curve arose from transport of solids in the high
velocity section of the channel near the pump station. These solids were hypothesized to
be from several sources, which consisted of an uneroded baseload, carry-over of the
unexported end of an erodable mass wave from the preceding event, and inter-event
generation. In the case of the low-level, high velocity situations presumed for UF9206 it
is possible that a large proportion of the erodable mass was exported along with an
appreciable amount of base sediment that would be expected to have lower erosion rates
and higher sedimentation velocities than the average" erodable material.
Under these circumstances a good fraction of the erodable mass in the region of
the pump was probably depleted and replaced with base sediment that had been
transported from upstream under high velocity conditions. Upon start-up in the
subsequent pump cycle there would not be an appreciable concentration of erodable mass
in the pump region so the first leg of the U curve would not be exhibited. As pump-
down progressed to a point where velocities start to become exponential with decreasing
depth, transport increased rapidly and the second leg of the U curve was manifested.
Because there was not always coordination of pumping rate with variables related to
water supply, the pump might be stopped for a short time and then restarted at the same
speed, to repeat the cycle, thus giving several second legs of the U without any first
legs. Conversely the pumps were also allowed to run with speed reduced on occasion.
This would account for the solids export curves that were convex with time, where the
peak of the curve should coincide roughly with the time at which pump speed was
reduced.


36
cloud. Class C (mobile fine) solids were noted to be weakly resistant to erosion. Class D
(biofilm) solids were reported to be poorly studied and ill-defined. More research was
recommended in the area of biofilm solids transport because of their high organic matter
content. The authors reported that flume studies of freshly deposited organic sewer solids
exhibited a critical shear stress of about 1.8 Pa. and that 75% of the eroded solids had a
particle size of less than 100 micrometers.
The only literature source found which deals directly with quantitative erosion
characteristics of natural organic sediments is the work of Hwang (1989). Hwang studied
the erosion characteristics of Lake Okeechobee sediments which had organic fractions in
the range of 40-45%. The experimental procedure included placing sediments harvested
from Lake Okeechobee in an annular flume as a thick slurry, covering the placed bed with
lake water, allowing the mixture to consolidate for several days, and then measuring the
concentration in the overlying water as various shear stresses were applied to the bed.
He developed a two-component model incorporating a "surface-fluff' component, which
had a critical shear stress for initiation of erosion of zero Pa (immediate erosion upon any
disturbance), with a bed-surface-erosion component that had a critical shear stress of
about 0.45 Pa (much lower than that reported for sewer sediment). Erosion rates as a
function of shear stress were determined for both components. He further correlated the
erosion rates with sediment bulk density and developed a map of regions of bulk density
and shear stress where various types of erosion would occur. The experimental method
utilized may have underestimated the impact of light flocculated organic matter which
may have been intermixed with the base sediment at harvest, nevertheless this study
reports the lowest critical shear stress found for organic sediments in this literature
survey.


58
Phosphorus may be exchanged between the soluble and particulate phases
by adsorption/desorption
All biological processes are considered to be active only during the
quiescent inter-event periods
Research goals for this study
The simplified conceptual model now leads directly to the general research goals
for this study that are, for a specific study farm, as follows:
1. Adapt an existing hydraulic network model for use in the sub
irrigation/drainage mode that prevails at the specific site.
2. Adapt or develop a water quality sub-module for the hydraulic model that
will allow expression of the water quality processes in a dynamic
hydraulic regime.
3. Determine the nature, frequency, and quantitative contribution of overland
flow erosion to particulate phosphorus in the water conveyance system.
4. Develop a description of the conveyance system sediment with respect to
quantities, locations, chemical/physical characteristics,
erosion/sedimentation characteristics, and temporal variations that is
adequate to allow modeling of sediment transport and pore water release
within the system.
5. Determine the adsorption/desorption characteristics of the system
sediment and incorporate this process into the model where appropriate.
6. Determine the biological processes in the system that contribute
significantly to phosphorus transport and develop a first approximation
lumped parameter model of these processes.
7. Develop, calibrate, and verify the model on the target study site.


31
difference in sources of particulate inorganic and organic phosphorus and the apparent
selectivity of transport which tends to favor particles which happen to have higher
phosphorus content. The cumulative load chemographs for the various forms of
phosphorus are illustrative of typical load distributions observed elsewhere, including the
Everglades area. They show 50% of the annual particulate phosphorus load being
transported during less than 5% of the total observation time, which may be interpreted as
the bulk of the particulate phosphorus transport occurring during major flow events.
Findlay et al. (1991) conducted automated synoptic water sampling and stream
gauging over a 150 km reach of the lower Hudson river for a three year period. Vertical
as well as horizontal sample profiles were obtained. Profile analysis of particulate
loadings led to the conclusion that during low to medium flows particle resuspension was
as important as tributary contribution in determining river sediment loading. They found
significant contributions from both autochthonous particulate organic matter and
resuspended detrital material during the ice-free seasons and concluded that transport of
particulate organic matter was controlled to a significant extent by processes occurring
within the river and were not simply related to loadings from outside.
Godshalk and Wetzel (1984) used piston coring devices to sample sediment
transects of a small hardwater lake in Michigan. They segregated sediment particle size
fractions by wet screening and then determined the various molecular weight fractions of
humic and fulvic acids, as well as conducting total organic carbon, fluorescence, and UV
absorbence analyses. Humic and fulvic acid proportions and the carbon content of each
fraction were used to estimate the relative age of each sediment sample (Humic-older,
Fulvic-younger, increasing carbon content indicating more refractory material). The
results showed a succession of organic particulate matter sources in the lacustrine
environment. The indigenous fine particulate organic matter (FPOM) was transported
first, but there was regular production of FPOM from coarse particulate organic matter
(CPOM) as time progressed. Thus the reservoir of FPOM was replenished on an irregular


288
It would be very gratifying if these internal simulation records matched the
concentrations seen in the intensive studies. Unfortunately they did not. The simulation
values for nodes that coincided with the sample points in the South-East Canal system
were consistently higher than the values observed during the tests, although the
qualitative trends were similar between simulated and observed.
To rationalize this discrepancy, recall that the system calibration parameters were
determined based on data from the discharge of the pump station. This discharge is a
combined stream from the North Canal and the South-East Canal system. These two
canal systems share some field ditches, but in fact there is little or no hydraulic
communication between them. Water flows either to one system or the other, not from
one system to the other. In general the North Canal had lower flows, lower velocities and
a greater surface area per unit length than the South-East Canal system.
What is presumed is that the biology and sedimentology of the two canal systems
developed independently. The North Canal with lower velocities and higher surface areas
may well have developed a biological system that had more erodable mass per unit area
than its counterpart locations in the South-East Canal system. In addition it is possible, in
fact probable, that the average erosion properties of the detritus and sediment in the North
Canal were different from those of the South-East Canal system. If the North Canal was
in fact contributing more suspended solids mass per unit of discharge than the South-East
Canal system and the model was calibrated on the basis of uniform properties across the
farm then the model could tend to overestimate the South-East Canal system and
underestimate the North Canal while still arriving at a good fit for the combined farm
discharge. This was in fact the case for the two events (258 and 285) where adequate data
were taken on the North Canal discharge to compare to the simulation. In both cases the
North Canal discharge TSS concentrations were higher than simulated. There was not
enough individual data available on the North Canal to develop a separate set of erosion
parameters.


326
is clear from the graph that for the bulk of the discharge (up to about 70,000 m3) the
cumulative solids load discharged in Simulation C was less than the corresponding point
in Simulation B. The important point to note from this simulation is that there might be
additional opportunity for solids export reduction by shutting the large pump off sooner
and starting the small pump sooner while still maintaining the same total time to
discharge a given volume of water.
Figure 11.6: Effect of Pumping Modification and Flow Width Modification
(Simulation C)
Hydraulic Mining and Recycle Analysis
The discussion of the trap analysis alluded to the fact that erodable mass not
exported may be available for transport in subsequent events. Alternatives to mechanical
dredging of channels might be hydraulic transport of erodable matter to regions of relative
quiescence. In these regions they could undergo consolidation over adequate time to


35
and Robbins, 1987). Degree of focusing, as would be expected intuitively, is a strong
function of biology and hydrodynamics. Bloesch and Uehlinger (1986), using a lake
wide distribution of sediment traps, found it to be relatively unimportant in a eutrophic
lake with high productivity and modest turnover velocities. Murchie (1985) used 210-Pb
dating to study the geologic history of a freshwater bay in Minnesota. He found that
focusing decreased as basinwide productivity increased, but also that high-organic-
content sediment was more intensely focused than heavier siliceous or calcareous
sediment. Kronvang and Christiansen (1986) used a combination of traps, cores, and
water samples to develop a spatially distributed sediment budget for a hydrodynamically
active estuary in Denmark. They found that focusing was strong in the upper estuary
zone and recommended specific dredging locations and seasonal times which would
allow for an optimization of the dredging effort to recovery ratio. This approach may
have particular interest in regions such as South Florida where hydrodynamics are, or can
be, partially controlled.
Specific quantitative erosion data have been generated in a few cases for the
organic fraction of sewage and stormwater flows. Kleijwegt et al. (1990) studied erosion
of cohesive synthetic sewer solids in a laboratory flume. They determined that the upper
limit of the critical shear stress for initiation of erosion of cohesive sewer sediments
appears to be in the range of 5-7 Pascals. Ashley and Crabtree (1993) categorized sewer
sediments into five classes: Class A coarse granular material. Class B agglutinated
Class A deposits, Class C mobile fines, Class D organic biofiims, and Class E -
deposits found in tanks. In a related study Ashley et al. (1993) noted that definition of
bedload was very difficult with Class C and D sediments because bedload for these
materials was in the form of a dense cloud of sediment close to the surface of the
transporting channel, as opposed to the traditional riverine definition of bedload as a
saltating layer of individual particles. The thickness and density of the cloud is affected
by hydrodynamic conditions as is the interchange of suspended particles with the bedload


38
correlations were very stream-type specific, and that for a given stream a hysteretic
pattern was observed over a time series when transported load was correlated with stream
power. This hysteresis was explained by Webster et al. (1987) as being due to the
limitation of supply of seston, which reduces the potential transportable load as time
progresses in an event or series of closely related events. The hysteretic effect may also
arise from seasonal factors which affect the availability and transportability of seston
even in the absence of significant hydraulic excursions. They proposed that the stream
power correlations be modified into separate correlations for the rising and falling limbs
of storm hydrographs, and that inter-event seston production and storage functions be
developed for use in conjunction with the power correlations. This approach results in a
higher level of predictability, but it also requires a fairly intimate knowledge of the
biological productivity of a system in order to be useful in a predictive capacity.
Several recent studies in Denmark on the Brabrand Lake (agricultural) watershed
(Kronvang 1992) and the Gjem A (lowland) basin (Svendsen and Kronvang, 1993, and
Svendsen et al., 1995) have provided good illustrations of the complexity of seston-
related phosphorus transport in the context of seasonal time scale basin hydrology. These
studies showed the following results:
Significant differences in Particulate P/Total P ratios in adjacent
subcatchments
Decreasing P-content of inorganic particulates as flow increased
Relatively constant P-content of organic particulates as flow increased
Reduction in the P-load vs discharge correlation with sequential storm events
Seasonal changes in the P-load vs discharge correlation
High levels of seston removal by macrophyte beds during normal flow but
high levels of release during storm flow, with this effect accentuated when
storms followed extended low flow periods.
Episodic increases in particulate-P load arising from macrophyte cutting


347
contribute directly to phosphorus export by remobilization of sediments or
indirectly by solubilization and uptake by aquatic macrophytes, which in turn may-
produce transportable phosphorus-containing particulate matter.
Studies of farm-sourced base sediment showed it to be relatively homogeneous,
with -80% organic content, phosphorus content on the order of 1000 mg/kg. with
a low labile phosphorus content, and containing significant amounts of aquatic
plant remains. Samples of base sediment from an EAA main canal showed
similar properties with the exception that plant remains were not readily
identifiable.
The prototype base sediment exhibited very weak phosphorus sorption
characteristics and highly flocculant settling properties.
The Counter Rotating Annular Flume and Particle Entrainment Simulator
experimental devices were adapted for use in measuring the erosion
characteristics of organic sediments in the laboratory and field respectively. The
key characteristics of the prototype organic sediment studied were:
a) It exhibited a critical shear or critical stream velocity for inception of
erosion. Thereafter erosion extent was proportional to an apparent shear
level.
b) The extent of erosion at a given shear was sediment-supply limited and
could be approximated by an asymptotic exponential decay model that
was consistent with a first-order erosion supply model.
c) Erosion and deposition were approximately reversible, implying that the
critical shear level for deposition was greater than the critical shear for
erosion for all elements of the sediment population.
Field erosion studies showed that the sediments tended to fall in two categories of
erodability and that there was a correlation between sediment erodability and the
presence of macrophytes in the channel from which the sediment was sourced.


163
was still a detectable residual. Filter blanks were run with each sample set. the residual
averaged 2.05 micrograms of phosphorus per filter over all blanks run. This value
represents only about 4% of the phosphorus that was found in a typical sample containing
30 mg/1 TSS, but it may represent on the order of 40% of the phosphorus that was found
in a typical sample containing 1 mg/1 TSS. This explains some of the variability that will
be seen in the low concentration samples to be presented later in this chapter.
A second complicating factor introduced by analyzing the filter contents directly
was the effect of the presence of the filter on the physical character of the digestion
process. The filter, under digestion conditions, tended to nucleate boiling similar to a
boiling stone at points where it came into contact with the digestion vessel walls. This
caused bumping, which could be particularly severe within a confined digestion tube or
flask. On several occasions early in the program when a confined digestion tube method
was being tested by a contract lab the samples were lost because the filter pads were
expelled from the digestion tubes with considerable force. Ultimately all filter pad
analyses were done by the author using the open beaker HC1 digestion method described
in Appendix C. Careful placement of the filter pads and close observation of the
digestion process was necessary to prevent sample loss even with the open beaker
digestion procedure but the proper technique was developed fairly quickly and
incorporated into the digestion routine.
Particulate phosphorus event studiesFarm UF92Q0
The event studies from target farm UF9200 were to form the basis for the model
development so they will be treated in some detail here. Sampling started with the June
3, 1994, event (Julian Day 154) and continued through December 29, 1994, which was
the last day of the event that started on December 21, 1994, (Julian Day 355). All events


6
system, the reuse of upstream farm discharge water by downstream farms, and the
surprising lack of historic data on phosphorus concentrations in water directly at the lake
outlets. A rough approximation of the maximum potential supply may be made using
published average figures. Mean annual total phosphorus in Lake Okeechobee ranged
from 0.05 to 0.1 mg/1 for a 12 year period from 1973-1985 (Izuno et al 1991). Annual
flows from the EAA have been reported to be in the range of 1.2 x 109 m3 at the end of
this period (Izuno, 1994). Assuming approximate equivalence between inflow and
outflow, a potential maximum loading on the order of 50,000-100.000 kg/year may be
estimated, or 0.18-0.36 kg/ha/yr.
The soils in the EAA are primarily Histosols. with depths ranging from 3 meters
to 0.3 meters. The soil is underlain by marl and limestone. Approximately 200,000 ha
are planted to sugarcane, about 40,000 ha are planted to vegetables, rice, and sod (Izuno
et al., 1991). Phosphorus concentration in virgin EAA Histosols is estimated by Sanchez
and Porter (1994) to be in the range of 0.08-0.35%. They note that in virgin soil about
24% of the phosphorus is inorganic, whereas in cultivated EAA soils up to 72% of the
phosphorus is inorganic, implying significant mineralization resulting from cultivation.
They further estimate that potential phosphorus mineralization is on the order of 20-150
kg/ha/yr for EAA soils. This estimate may be compared with Koch's (1991) estimate of
mean total phosphorus storage in the upper 30 cm of EAA soil of 276 kg/ha. These
quantities may be put into some perspective when it is calculated that 50.8 cm (20 in) of
annual rainfall runoff from one hectare need only dissolve 0.508 kg of phosphorus to
have a dissolved phosphorus concentration of 0.1 mg/1. The concentration of 0.1 mg/1 is
well above the natural level prevailing in the uncultivated interiors of the WCAs (see
below) or the ENP.
Organic soils such as the EAA Histosols were formed under submerged, reducing
conditions. When they are drained, cultivated, and aerated, they become net sources of
phosphorus due to mineralization arising from increased aerobic microbial activity


343
An example of the interpretation of these curves is that at UF9200, 50% of the
particulate phosphorus exported during the normal wet season was discharged at
suspended solids concentrations of 22 mg/1 or less. The same interpretation for UF9206N
and UF9206S would give values of 60.6 mg/1 and 51.2 mg/1 respectively. Relative to the
operation at UF9206, it appears from the curves that about one-third of the particulate
phosphorus exported was in discharges that had TSS concentrations in excess of 100
mg/1. The correlation equations for TSS vs. Canal Depth (uncorrected in this case for
UF9206S) from Figures 11.14 and 11.15 were:
TSS = 51.87/T28' 11.1
for UF9206N. and
TSS = 7254(/i + 0.43) 309 112
for UF9206S. where
h = Actual canal water surface elevation, m MSL
TSS = Expected concentration of exported suspended solids, mg/1
If these equations are used to calculate the level corresponding to 100 mg/1
discharge concentration they yield actual canal levels of 0.80 m and 0.47 m for UF9206N
and UF9206S respectively. Figures 11.19 and 11.20 show the wet season canal level time
series for each location with these levels superimposed as proposed control points for
minimum canal level. The figures show that controlling to these minimum levels would
have involved clipping the lower excursions from only four pumping cycles at UF9206N
and only two cycles at UF9206S. It is fully recognized that, for all the reasons


311
Model Type The model is a mixed type distributed parameter hydraulic
model, lumped parameter suspended solids erosion and sedimentation
model, with an empirical correlation relating phosphorus content to
suspended solids concentration. Erosion is first order in concentration of
erodable mass, second order in the difference between mean channel
velocity and the critical velocity for initiation of erosion, and quasi-first
order in sedimentation. Single parameters, uniformly distributed spatially
and temporally, are used to describe the erosion and sedimentation
characteristics of the erodable mass, which in fact has a distributed
population of particle sizes and types.
Key Assumptions in Model Calibration There were several important
assumptions necessary to achieve a credible calibration and validation of
the transport model.
1. The flow areas in the sections of the North and East Canals
adjacent to the pump station were restricted to approximately
those of the discharge culverts or the pump intake. This was
necessary to simulate the first leg of the U shaped suspended
solids transport curve.
2. Transport was assumed to be negligible below the critical
velocity for erosion, which was represented as a single velocity
across all events.
3. During times when the farms small pump was in operation the
channel velocities seldom exceeded the critical velocity.
Assumption #2 led to the assumption that these portions of the
events would be treated as non-pumping times when
calculating interevent times. This led to the exclusion of Event


167
Correlation of TSS Load with Hydraulic Load
Farm UF9200
16000
14000
12000
-f 10000
I 8000
g 6000
4000
2000
0
0 100,000 200,000 300.000 400.000 500.000 600.000 700.000 800.000
Event Hydraulic Load m3
O
Slope=20.45 mg/1
oo O o
O Events 220-285 O Events 319-355
Figure 7.1: Correlation of Event TSS Load with Event Hydraulic Load for Farm UF9200
Correlation of Particulate Phosphorus Load with Hydraulic Load
Farm UF9200
O Events 220-285 Events 319-355
Figure 7.2: Correlation of Event PP Load with Event Hydraulic Load for Farm UF9200


CHAPTER 9
MODEL DEVELOPMENT
Introduction
At this point it is appropriate to recall from Chapter 4 the basic criteria and goals
for the particulate phosphorus transport model to be developed as the end product of this
work. The minimum criteria in abbreviated form are as follows.
Adequate representation of the surface water, groundwater, and channel flows
of a network typical of the EAA farm scale over the time scale of an entire
pumping event, with sufficiently detailed time scales to allow simulation with
reasonable accuracy of transients that exist at pump start-up and shut-down.
Modeling with reasonable accuracy particulate phosphorus mobilization,
transport, and deposition as a function of some readily defined hydraulic
parameter such as shear stress or average stream velocity.
Representation of phosphorus interchange between soluble and insoluble
forms by adsorption-desorption.
Incorporation of the dynamic impact on phosphorus transport of biological
growth and senescence in the aqueous system.
The goals of this study to achieve those minimum criteria were set as follows.
1. Adapt an existing hydraulic network model for use in the sub
irrigation/drainage mode that prevails at the target site.
2. Adapt or develop a water quality sub-module for the hydraulic model that
will allow expression of the water quality processes in a dynamic
hydraulic regime.
244


126
1. Each class has, at time zero, an associated specific erodable mass.
EMj(O), with units of gm/cm2.
2. Specific erosion flux, gm/cm2-sec, of the sediment in class / is
proportional to the product of the excess shear stress, x xc,, and the
amount of specific erodable mass present, EM¡, according to the equation
Ft = -e(x xc,)EM, 6.12
where
Fc = Flux due to erosion, gm/cm2-sec
s = Erosion coefficient, cm2/dynes-sec, assumed to be
constant for all classes
x = Shear stress at time t, dynes/cm2
xCj = Critical shear stress for erosion of class
dynes/cm2
EMj = Erodable mass of class / present at time t,
gm/cm2
3. All particles in class have the same settling velocity, which is
independent of concentration. The deposition flux of particles in class f\
gm/cm2-sec, is equal to the product of the concentration and the velocity,
as given by:
Fd = ^C, 6.13
where
Fd = Deposition flux, gm/cm2-sec
Q = Concentration of suspended particles of class /


94
these values and the experimentally determined partition coefficient for the B9B10 <38p
fraction of 83.1 1/kg.
In this case the pore water concentration must first be calculated using the same
type of equations expressed in Eq. 5.4-5.6. The calculated pore water in this case was
1.20 mg/1, higher than the prior example because of a higher original sorbed phosphorus
concentration and a lower partition coefficient. The results of this example are shown in
Table 5.8.
Table 5.8: Summary of Phosphorus Distribution in Demonstration Example 2
Phosphorus Origin
Concentration-mg/1
% of Soluble P
% of Total P
Original Soluble
0.10000
94.61
67.70
Pore Water Soluble
0.00114
1.07
0.77
Desorbed Soluble
0.00456
4.32
3.09
Total Soluble P
0.10570
100
71.56
Suspended Sediment
0.04200
NA
28.44
(Insoluble)
Total P
0.14770
NA
100
Here the contribution to soluble phosphorus from pore water and desorption is
calculated to be about 5.4% of the total soluble and about 3.9% of the total phosphorus.
In both cases the results were very similar, desorption was calculated to contribute on the
order of 5% of the soluble phosphorus for the demonstration system assumptions.
Contribution is roughly linear with suspended solids concentration, so as suspended
solids vary the contribution varies roughly proportionally. Note that the value of 50 mg/1
for suspended solids is significantly higher than the average value of 14 mg/1 reported in
Chapter 2 from the SFWMD canal data (Mucinic 1994).


218
Sediment Phosphorus Content Synoptic Surveys
On Julian Date 229 a synoptic field ditch and canal sediment sampling program
was conducted at UF9200. Sampling was done with a piston core sampler and only the
top 2 cm of sediment (the operationally defined surficial sediment) were retained for
analysis. The original intent of this study was to evaluate for the presence of spatial
gradients in phosphorus content along the ditches and canals after a pumping event. In
the preceding event. Number 220, pumping was terminated at Julian Date 227.73, so
about 38 to 46 hours had transpired from the time the pumping had stopped and the
samples were taken. The subsequent event, Number 237, started at Julian Date 237.7 so
about 192-200 hours transpired before the next event.
Field ditch samples were divided into two sets. One set of 7 samples was taken
50 m south (upstream) of the discharge culvert of every second ditch on the south side of
UF9200. All of these ditches drain to the South Canal. On the north side, which drains
to both the North and South Canals, four sets of three samples each were taken 50 m in
from the south culvert, at the midlength point, and 50 m in from the north culvert in four
ditches roughly evenly spaced along the length of the farm. Two additional samples were
taken at two more south culverts on the north side in order to have six sets of samples that
were in ditches directly opposite one another along the South Canal.
Canal samples were also taken in two groups. The first group consisted of
samples taken 500 m downstream of the upstream end and 500 m upstream of the
downstream end of the North and South Canals and at the midlength of the East Canal.
The second group consisted of samples taken at the pump suction and at points 100, 200,
and 300 m upstream in the East Canal and 150, 250. and 350 m upstream in the North
Canal.


282
400
Suspended Solids Simulation Results
Event UF9200-220
Time from start of Pumping-hr
oooooooooo
50 60 70
O Observed Simulated
Figure 10.1: Simulation Results-Total Suspended Solids for Event UF9200-220
Suspended Solids Simulation Results
Event UF9200-237
160
O Observed Simulated
Figure 10.2: Simulation Results-Total Suspended Solids for Event UF9200-237


63
within the EAA under intensive study. These farms had been chosen to represent a
typical range of farm size, crop type, soil type, and geographical distribution within the
four main sub-basins of the EAA.
The farms have, by mutual agreement with the appropriate regulatory authorities,
been coded to preserve the anonymity of the growers who were participating in the study
program on a voluntary basis. The code numbers were in a series of UF9200 through
UF9209, where UF indicates University of Florida study, 92 indicates the start year
of the BMP study program and 00 through 09 indicates the code number assigned to
each participating farm. Further coding indicated the location of permanent sampling
sites at a specific location. A refers to the main farm discharge; some farms have two
major discharge locations so B is reserved for the second discharge location where
appropriate. Sites C and beyond refer to internal farm sampling locations that are
added as the program dictates. For example farm UF9206 has two pump discharge
locations, so UF9206B refers to the permanent sampling location at the south pump
station of that farm. That coding protocol will be modified later in this document where
UF9206A and B will be referred to as UF9206N and S respectively.
Three sites were selected for the preliminary sediment inventory based on the
recommendations of Dr. F. T. Izuno, the IFAS Program Principal Investigator. They
were:
UF9200 A 1280 acre (518 ha) farm with two main canals and one pump station,
that planted only sugarcane. This farm was representative of medium to large
sugarcane-only fields with relatively deep soil that are actively managed.
UF9202 A 320 acre (130 ha) farm with one main canal and one pump station
that planted only sugarcane. This farm was representative of small to medium
sugarcane-only fields with relatively shallow soil that receive low to moderate
management.


338
the depth to the fourth power. A third power dependency may imply somewhere between
a rectangular and a triangular channel, such as might be exhibited by an acute trapezoid.
The exponents of approximately 2 at UF9200 and approximately 3 at the UF9206
stations might be interpreted to imply that the relatively narrow pumping level variation
at UF9200 kept the canal operation within an approximately rectangular crossectional
configuration, while the broader variation of levels observed at the UF9206 stations
Correlation of Discharge TSS Concentration with Canal Level for UF9206N
Figure 11.14: Correlation of TSS Concentration with Canal Level at UF9206N
frequently brought the levels into a range where there was channel constriction with
decreased depth. This channel constriction would tend to accentuate the velocity increase
with decreasing level and contribute to a disproportionate increase in mass exported.
This is in fact the result observed at both UF9206 pump stations where reduction of levels
below about 1 meter resulted in export suspended solids concentrations in the range of
100-1000 mg/1.


34
suspended matter biochemical oxygen demand (BOD). They found that the suspended
matter in the channelized riverine environment averaged about 40% organic matter by
weight. The organic matter itself consisted of about 25% viable planktonic-type material
(as estimated by chlorophyll-a analysis) and 75% nonliving biomass. This nonliving
biomass was further categorized by BOD analysis as 35% (of total suspended organic
matter) biodegradable and 40% "refractory" organic matter. Bokuniewicz and Arnold
(1984), who carried out water sampling at 16 stations in freshwater reaches of the Lower
Hudson River, found the average organic content of the tidally influenced freshwater
suspended sediments to be 22% by weight.
Tipping et al. (1993) sampled suspended material in riverine environments and
determined particle aggregate sizes by microscopy. They also collected freshly
sedimented material using sediment traps placed in various hydraulic regimes within the
reaches under study and determined particle size distributions using particle counters.
Subsamples were analyzed for total mass, carbon, and nitrogen. They found that the
average particle density decreased significantly as average aggregate diameter increased,
with an increasing concentration of organic matter as aggregate size increased. They
postulated that little, if any, agglomeration takes place in-stream, rather that the
agglomerated particles either enter the stream as a result of original erosion of fields and
banks, or they form on the sediment surface in relatively quiescent (dead) zones and are
resuspended by periodic hydraulic excursions. They note that dead zones can play an
important contributory role, causing either concentration spiking or tailing when they are
disturbed. The magnitude of the disturbance and the exchange rate with the main stream
govern the type of contribution. They also point out that dead zones with normal
turnover times of days can be important contributors of phytoplankton.
The selective transport of constituents resulting from variations in particle size,
density, origin, and composition gives rise to the phenomenon known as "focusing"
which is the spatially inhomogeneous distribution of contaminants in sediments (Eadie


87
however, it was decided to use a standard non-linear optimization routine to develop a
statistical fit of the data by minimizing some form of the variance of calculated versus
experimentally determined values of S, the total substrate concentration of srbate at a
given liquid phase concentration. The variance measure used was a form of weighted
residuals, where the square deviation of each calculated versus experimental data pair was
divided by the experimental value. The use of this statistic normalizes the contribution of
each data pair to the variance analysis and thus reduces the dominance of the data points
at high concentrations (Draper and Smith, 1981).
The value of using the optimization approach is that it allowed all data points,
both Adsorption and Desorption, to be considered in the same analysis. This was
done by determining, by difference, the amount of sorbed mass remaining on the substrate
for each desorption sample, which requires an estimate of So- The optimization approach
eliminates determination of So by extrapolation, and instead determines the best fit of all
three equation parameters simultaneously. Figures 5.9 through 5.12 show the isotherms
generated by this method. Table 5.6 summarizes the relevant parameter sets.
Correlations for three of the four sediments, both B9B10 samples and the WPBC
<38|i sample, had reasonably good coefficients of determination (r2) of 0.877-0.944. The
weighted residuals method caused the analysis of the WPBC 75-150(1 sample to reduce
the contribution to the correlation of the highest concentration sample in both the
adsorption and desorption run so there was significant deviation at the highest
2
concentrations, resulting in an r of 0.544. When the contribution of the highest
adsorption-desorption data pair is removed the coefficient of determination for the
WPBC 75-150(1 sample increases to 0.735.
It should be recalled that the WPBC 75-150(i fraction was the sample that had a
relatively low organic content and a relatively high phosphorus content. With this in
mind several observations on Table 5.5 are in order. S0 ,the native adsorbed phosphorus,
was much lower for the WPBC samples (1.1 and 12.0 mg/kg) than for the B9B10


96
found in the EAA. The original intent of this work was to investigate the sedimentation
characteristics of the test sediments, develop a simple sedimentation model for these
organic sediments, and then use that model as a prototype to study spatial and temporal
variation in the field. The mechanism chosen for the first approximation was the
Bottom Withdrawal technique. This technique, which is described in Appendix B is a
process of initially suspending sediment in a 1 meter column of water contained in a large
burette-like tube, withdrawing equal height increments from the tube over a period of
time and analyzing the water for suspended solids. Through the use of graphical
transformations the cumulative density function of fraction of mass with settling
velocities less than a given value may be estimated.
The B9B10 <38p fraction was chosen for the initial testing. The tests were run at
five different test durations. A 90 minute test means that the entire contents of the
burette were withdrawn, incrementally, over 90 minutes. The durations ranged from 90
minutes to 1440 minutes. Figure 5.13 which shows the calculated settling velocity
distributions, clearly indicates that the observed settling velocity distribution was a
function of the duration of the test. Figure 5.14 shows the mean velocity, V50 as a
function of test time, where V50 increases from 0.53 m/day at 90 minutes duration to 2.36
m/day at 1440 minutes duration. The data are fit well (r2 =0.998) with a linear
relationship. These test results strongly suggest a linear increase in sedimentation
velocity with time in suspension that may extend well beyond the maximum experimental
time of 1440 minutes or one day.
These results may be explained by assuming that the sediment is capable of
undergoing autogeneous flocculation. Under these circumstances, the particle diameter
distribution may increase steadily with time in suspension. If there is not an off-setting
decrease in apparent particle density there will be a steady increase in particle settling
velocity. Flocculation kinetics can be highly nonlinear (van Leussen 1988, Lick et al.,


348
Field time-series and synoptic studies showed that there was a much higher
phosphorus content in the exported solids than existed in the base sediments, that
the solids export profile was concave with maxima at the beginning and end of
major pumping cycles, and that the suspended solids tended to move downstream
in wave-like patterns. The studies also indicated that direct contributions from
field ditches to farm suspended solids export was minimal.
The erodabiiity of the majority of the studied sediments was not adequate to
supply the observed farm export of suspended solids. The phosphorus content of
base sediments was far lower than the average phosphorus content of the exported
material, but the phosphorus content of sediments in the region of macrophytes
approached the average phosphorus content of the exported material, so the basic
hypothesis of the research was modified to incorporate major contributions of
erodable mass from channel macrophytes, detachable detritus, and planktonic
growth. Studies of aquatic macrophytes showed that they were capable of
providing adequate mass at sufficiently high phosphorus content to satisfy the
phosphorus export requirements observed in the time-series studies.
Large scale composite studies verified that the actual farm export suspended
solids had weak sorption characteristics, and that the labile fraction was much
higher than had been observed in the base sediment studied earlier.
The hydraulic model DUFLOW, developed by a Dutch technical consortium and
adapted for groundwater flow by Soil and Water Engineering Technology, Inc.,
was programmed with a simple erosion-deposition model that had characteristics
derived from the prototype sediment:
a) First order in erodable mass, adopting the decaying exponential erosion
exhibited by the prototype.
b) Exhibiting a critical stream velocity for erosion and second order in the
excess of stream velocity over critical.


APPENDIX D
APPROXIMATE MODEL FOR CRAF
The CRAF was modeled in an approximate manner by ignoring wall resistance
and assuming that the velocity profile for the upper ring was expressed by the equation
for turbulent hydraulically smooth flow:
D.l
and the velocity profile for the channel was expressed by the equation for turbulent
hydraulically rough flow:
D. 2
where terms are defined in Chapter 6.
If wall resistance is ignored then To is equal for both equations. The model is a
simple spreadsheet superposition of the velocity profile calculated for the upper half of
the flume by Equation D. 1 on the velocity profile calculated for the lower half of the
flume by Equation D.2 where y for Equation D. 1 is calculated from the water surface
down and y for Equation D.2 is calculated from the channel bottom up. The
spreadsheet calculates the ratio of mid-depth velocity to total velocity and the ratio of
velocity at a depth of 0.368 x channel depth to total velocity, where all velocities are
relative to the channel bottom. The spreadsheet also displays the resulting velocity
profile. Example calculations are shown in Figures D. 1 (Dimensional) and D.2 (Non-
dimensionalized with respect to velocity).
365


238
Figures 8.1 and 8.3 show clearly that there was a significant variation in initial
sorbed phosphorus content, So. as represented by the negative of the ordinate intercept of
each line, for each sample. The desorption curves for both locations, however, showed
good coincidence, implying that although the samples had variable amounts of
phosphorus sorbed initially, their sorption characteristics were quite similar at a given
location. This information suggested that the traditional Eadie-Hofstee single reciprocal
plot of the desorption data grouped by location should give a good estimate of the
location sample sorption parameters. Figures 8.5 and 8.6 show these plots for UF9200
and UF9206N respectively.
An advantage of the single reciprocal plot in this case is that the ordinate intercept
is equal to the limiting value of S/C as S approaches zero (see Chapter 5). S/C at low
concentrations is, by definition, equal to the partition coefficient k, thus the intercept of
the Eadie-Flofstee plot is equal to the primary value of interest, the partition coefficient.
Recall from Chapter 5 that the value of the partition coefficient used for the
demonstration examples which showed sorption to have negligible impact was 2001/kg.
The correlation of UF9200 data using the single reciprocal method on the grouped
desorption data had an r2 value of 0.891. The intercept of this curve was 119.61/kg. The
correlation for UF9206N. with one low data point eliminated, had an r2 value of 0.918.
The intercept for this curve was 119.0 1/kg.
The calculated k values for both locations were essentially identical, and were
well below the 200 1/kg used in the demonstration example, so it was concluded that the
assumption of negligible impact by sorption on total discharge phosphorus concentration
was supported by the adsorption/desorption test results on the large-scale composite
samples taken at both study locations.


208
which was considerably higher than the farm discharge maximum of 29 mg/1. The
equipment shortage precluded North Canal sampling for Event 262. By the time Event
285 was completed, the North Canal had an enormous discharge maximum of 713 mg/1
that dominated the farm discharge for several hours before and after this peak.
Recall that hydraulic measurements showed that the fractional volumetric
contribution from the North Canal reduces rapidly as canal levels are reduced, so the
large differences observed between the high North Canal concentrations and the farm
discharge concentrations are partially attributable to the dilution effect of the lower TSS
concentration water from the East Canal. None the less the contribution of the suspended
solids load from the North Canal seemed to show a definite increase with time from
Event 252 through Event 285.
Downstream variation of suspended solids concentration
The sampling locations 2,2.1. 3,4, and Discharge form a linear spatial sequence
moving downstream in the South-East Canal system. Isolation on these sampling
locations gives some insight into the patterns of suspended movement down the canal
with time. Recall that the bulk of the suspended solids load for an event was found to
occur from the time the large pump started until it stopped for the first time on low-level
shut down. Figures 7.42 through 7.44 show the locations 2,2.1, 3,4, (where data were
taken) and Discharge only, for the first pump-down time sequence for Events 258, 262,
and 285. The curves have been smoothed in these graphs to emphasized a pattern. Event
252 may be viewed in Figure 7.38.
Starting with Figure 7.38, note that all four event graphs suggest a spatial and
temporal sequence of suspended solids that could be viewed as a wave-like pattern of
suspended solids moving longitudinally down the canal. The pattern appears to show


293
UF9200-258
Event UF9200-258


284
Suspended Solids Simulation Results
Event UF9200-285
Figure 10.6: Simulation Results-Total Suspended Solids for Event UF9200-285


157
controlled the schedule of the ISCO sampler peristaltic sampling pump. There were also
stand-alone strip-chart level recorders at various locations within each farm which
continuously recorded levels of groundwater wells and selected conveyance channels.
The strip-chart data was digitized periodically and integrated with the CR-10 data report.
Discharge flow was calculated from directly measured pump head and RPM data
and previously determined calibration correlations. Optical transducers, mounted below
each pumps pulley flywheel, sensed the passage of reflective tape attached to one leg of
the pulley. The CR-10 polled the transducers on a five minute cycle and recorded the
cumulative pulley revolutions if the pump was running. Upstream and downstream heads
were also recorded on a five minute cycle. Calibration correlations had been developed
for each pump relating measured discharge flows at specific pump speeds to head
differential across the pump. Interpolation and extrapolation formulae were developed
for each pump for heads and speeds not measured as part of the calibration process.
The CR-10 kept a cumulative account of the pump run-time and sent a signal to
the ISCO sample pump after a predetermined time-of-pumping had passed. The ISCO
sampler was programmed to extract a predetermined volume of sample from a sample
head located approximately 0.5 meters above the bottom surface of the pumphouse
conveyance channel. At UF9200 the pump discharged into a 10 x 10 x 3 meter sump
prior to discharging into the West Palm Beach Canal. The sample point was downstream
of the pump, approximately in the middle of the sump. At UF9206 North and South the
pumps discharged directly into a private drainage district canal that was shared with a
number of other farms. Here the sampling points were on the farm side of the pump
approximately 3 meters upstream of the pump intake at both the North and South
location.
The ISCO sampler held twenty-four 500 ml bottles and could be programmed to
fill these bottles in a number of ways. The samples and data from the ten farms in the
IFAS program were collected on a twice-weekly basis, and experience had shown that the


316
Another example might be shunting the start-up discharge to a distribution
system that would portion it to upstream field ditches at very low
velocities, where the initial solids load of the U curve could be settled in
channels that would never see a velocity close to critical.
5. Velocity Control The presumed existence of the critical velocity and the
second order relationship between velocity and erosion suggest that there
may be significant advantage to reducing maximum velocities in the
conveyance system. This could be done with programmed operation of
pumps to increase pumping rates when channel depths are greatest and
reduce pumping when water depths are reduced. This occurs naturally now
because of head effects but velocities still increase as water level recedes.
This practice may have only temporary effects, however, if the erodable
mass builds up in the channel as a result of the reduced velocities and does
not consolidate rapidly.
6. Configuration Control In this context configuration control refers to
eliminating or reducing configurations that can give rise to locally high
velocities, such as those presumed for the East and North Canal discharges
in the calibration process. Configuration control might mean introducing a
broad floating or submerged inlet weir at the pump inlet or enlarging
culverts to reduce the effects of streamline convergence or re-channelizing
a canal reach to eliminate constrictions such as that which exists in the
midsection of the North Canal at UF9200.
Some of these management actions can be tested by the model. In the final
chapter several of these actions will be exemplified by simulation of specific events at


280
micrometer fraction of the B9B10 prototype sediment was in the range of 0.5-2.5 m/day.
depending on the amount of flocculation time the suspended sediment had been exposed
to in the test process (See Figure 5.12). The calibration value of 125 m/day is 50 times
the upper value observed in those tests so there was no agreement between the calibration
value and the smallest particle size fraction of the prototype sediment.
In Chapter 5 a relationship between organic content and particle specific gravity
was established for the prototype sediment. If we assume an organic content of 70%, the
particle specific gravity would be around 1.7. With this specific gravity we can use
Stokes Law to make some approximations of particle sizes that would give rise to these
various settling velocities. The velocity range of 0.5-2.5 m/day observed in the bottom
withdrawal tests would correspond to particle diameters of 33-8.7 micrometers, while the
calibration velocity of 125 m/day would correspond to a particle diameter of 61.4
micrometers. The value of 33-8.7 micrometers was derived from material that had been
specifically segregated by screening through a 38 micrometer opening sieve, and then
subjected to floe destroying agitation of multiple column inversions in the suspension
process of the bottom withdrawal test.
Contrast that process with the hypothesized process in the canals, where detrital
material is produced by plant senescence and microbial action and the process that
dislodges or resuspends the particulate material is a result of the shearing action of the
turbulent flow, which has relatively modest mean velocities in the range of 0.1-03 m/sec.
The particle size of 33-8.7 micrometers spans the range described as very fine silt to
coarse clay, while the 61.4 micrometer diameter is at the upper range of the category
coarse silt (Vanoni 1975).
Although the calibration sedimentation velocity and the resulting calculated
particle diameter are considerably greater than the measured values for the <38
micrometer fraction of the prototype sediment, they are still in the range of very fine
grained sediments, defined by Raudkivi (1976) as less than 100 micrometers. Although


55
or part of this precipitate may dissolve under reduced pH conditions (from
respiration).
Canal water may also interact directly with exposed substratum in the bed
to adsorb or desorb phosphorus, depending on surface and bulk phase
conditions.
Hydrodynamic conditions affect interchange of sediments with the bulk
phase, interchange of soil water with the bulk phase, erosion and filtration,
infiltration, redox potential in the sediments, downstream transport of all
mobile constituents and, indirectly, soil/soil-water/crop interchanges.
The model is an unsteady-state one so all interactions must be assumed to
be dependent on antecedent conditions.
Scope of This Research
Minimum criteria for experimental and modeling efforts
Given the fact that virtually no channel transport experimental or modeling
activity had taken place on the EAA farm scale at the time of inception of this research,
the decision was made to restrict this effort to fundamental approaches. The field-scale
groundwater chemodynamic model for soluble phosphorus was being developed
separately and was planned to be integrated with the farm-scale model at a later date.
Basically the field scale model is to provide the soluble phosphorus dynamic boundary
conditions at the periphery of all water conveyance channels. The model to be developed
for the conveyance systems was required to have at least the following properties.
1. It should adequately represent the surface water, groundwater, and channel
flows of a network typical of the EAA farm scale over the time scale of an
entire pumping event, that may cover multiple days.


120
for
for
* <
6.10a
6.10ft
where C = concentration in the water column at time t
Co = initial concentration at time zero
Tb = shear stress applied to bed
ted = critical shear stress for deposition
Ws = suspended sediment settling velocity
h = water column height
t = time from start of deposition
all in consistent units, we see that for each collection of particles suspended at a specific
shear level the critical shear stress for deposition is always greater than the shear stress
required to suspend that family of particles. If this were otherwise then, in the simplified
case assumed here, there would be no deposition. In the more realistic case presented by
Mehta (1988) there would be some deposition but there would not necessarily be a return
to the starting concentration.
The implication of the assumption of critical deposition shear always being
greater than the suspending shear is important from a modeling standpoint, because it
may allow a simplification of the deposition model. It is also important from an
environmental standpoint because it implies that in-stream shear reduction for an
adequate time period should allow deposition of all material eroded at greater shear
levels.
Series 3 Extended Erosion Tests at Reduced Water Depths All the previously
discussed runs in Series 2 were done at water depths of 30.5 cm (12 inches). For the runs
in Series 3 the water depth in the flume was reduced to 15.3 cm (6 inches). The purpose


349
c) Sedimentation rate determined by sedimentation velocity and water depth.
Parameters were lumped for the whole population to affect single constants for the
erosion coefficient, critical velocity, and sedimentation velocity. A statistically
derived correlation was used to forecast particulate phosphorus content as a
function of suspended solids concentration, but no assumptions were made that
limited the source of the suspended solids to the sediment.
The model parameters were calibrated by fitting to six separate events occurring
during the normal wet season. In the calibration process the parameter that
represented initial concentration of erodable mass gave the best fit when it was
positively correlated with interevent time, implying generation of erodable mass
between events.
When the physical representation of the farm conveyance system in the model was
modified to reflect flow constrictions at the pump station outlet end of the farm
canals, the model was able to reproduce the concave suspended solids profiles
observed in the time-series studies. The final version of the calibrated model gave
good fits across six diverse rainfall and pumping events, and exhibited the
wavelike movement of suspended solids that was observed in the synoptic studies.
The model was validated on two late season tropical depression storms with
approximately the same degree of accuracy as was observed in the calibration
process.
Example simulations modeling the application of different management practices
to the target farm illustrated:
a) the linear relationship between reducing erodable mass and reducing
exported mass,
b) the potential for rapid loss of effectiveness of sediment traps that do not
appreciably reduce the stream velocity,
c) the potential benefit of reducing stream velocities in the vicinity of the


308
Figure 10.26: Cumulative Suspended Solids Load Simulation Results for Event
UF9200-319
UF9200-355


40
force and particle size and mass as a resistance. There are several such equations
(Raudkivi 1976). Shields equation is presented here for the purpose of example.
(h)-4?S
3.1
where gB = weight rate of bed transport per unit width, kg/m
q = volume rate of water flow per unit width, m3/m
S = bed slope, m/m
d = particle diameter, m
x = bed shear stress, Newton/m2
to = critical shear stress for motion, Newton/m2
y = specific mass of water, kg/m3
ys = specific mass of sediment, kg/nf
Suspended load transport may be treated within one or more of three conceptual
frameworks, diffusional, energy, or statistical.
Diffusional approaches incorporate a turbulent eddy diffusivity term into the
equations of motion and solve for the steady state concentration profile of mass in the
vertical direction. An equation often proposed as a basis for analysis is one presented by
Rouse (1937), that describes the concentration profile above a datum plane at which point
it is assumed the concentration is known:
3.2
with
w
z =
P*c£/*
where C = concentration at elevation y, kg/m3


262
The Pump Station calculated and observed values track quite well through the
period of the large pump continuous operation. The levels are reported on an hourly basis
and thus do not reflect many of the level swings arising from the on-off pump operation.
The levels at mid-farm in the South Canal also track well but with a slight under
prediction. The under-prediction is more pronounced for the upper reach of the North
Canal. The North Canal is less regularly formed than the South and East, resembling a
creekbed more than a canal at some locations, and suffering some constrictions at around
midlength that often served as a collection point for floating macrophytes. These factors
probably contributed to the observed levels being higher than calculated.
Figure 9.4: Calculated and Observed Levels, Event UF9200-220


394
Cosgrove.D.J. 1977. "Microbial Transformations in the Phosphorus Cycle In Advances
in Microbial Ecology, T. Alexander, Ed.. Plenum Press. New York
Cunge, J.A. 1989. Review of Recent Developments in River Modeling. In Hydraulic
and Environmental Modeling of Coastal. Estuarine, and River Waters. R.A.
Falconer, P. Goodwin, and R.G.S. Matthew, Eds., Gower Technical, Aldershot.
U.K.
Cunge, J.A.. Holly, F.M., Verwey.A. 1980. Practical Aspects of Computational River
Hydraulics, Iowa Inst, of Hydr. Res., University of Iowa. Iowa City, Iowa
Cushing,C.E., Minshall.G. W., and Newbold,J.D. 1993. Transport Dynamics of Fine
Particulate Organic Matter in two Idaho Streams". Limnol. Oceanoer., 38:1101-
1115
Davis.W.R. 1993. The role of Bioturbation in Sediment Resuspension and its Interaction
with Physical Shearing. J. Exner. Mar. Biol. Ecol.. 171:187-200
Davis.W.R.. and Abdelrhman, M.A. 1992. Geophysical Transport Assessment of
Contaminated Sediment: Experimental and Mathematical Procedures for the
Particle Entrainment Simulator (PES). Report to EPA Region II (New York)
Davis.W.R., and Means.J.C. 1989. A Developing Model of Benthic-Water Contaminant
Transport in Bioturbated Sediment. In Proceedings of the 21st European Marine
Biology Symposium, Gdansk, Poland, PL ISSN 0078-3234
DeBusk, T.A., and Dierberg, F.E. 1989 Effects of Nutrient Availability on Water
Hyacinth Standing Crop and Detritus Deposition, Hvdrobiologia, 174:151-159
Douglas, M.S. 1988 The Everglades-River of Grass, Pineapple Press, Sarasota, FL
Draper.N.R., and Smith.H. 1981. Applied Regression Analysis, John Wiley and Sons,
New York
Eadie.B.J.. and Robbins,J.A. 1987. Role of Particulate Matter in the Movement of
Contaminants in the Great Lakes, in Sources and Fates of Aquatic Pollutants,
Eisenreich,S.J., Ed., American Chemical Society, Washington, DC
Elwood, J.W., Mulholland, P.J., and Newbold, J.D. 1988. Microbial Activity and
Phosphorus Uptake on Decomposing Leaf detritus in a Heterotrophic Stream,
Vehr. Intemat. Verein. Limnol.. 23(2):1198-1208
Engle, D.L., and Melack, J.M. 1990. Floating Meadow Epiphyton: Biological and
Chemical Features of Epiphytic Material in an Amazon Floodplain Lake,
Freshwater Biology, 23:479-494


336
UF9206, on the other hand, there was no specific effort to control level. Also at UF9206
the cane fields were used as water storage areas early in the season but their water table
Water Elevation Trends for UF9206 Pump Stations
Julian Date
UF9206N- UF9206S Trend-UF9206N Trend-UF9206S
Figure 11.12: Water Elevation Trends for UF9206 Pump Stations
Water Elevation Trends for UF9200 Pump Station
Figure 11.13: Water Elevation Trends for UF9200 Pump Station


50
One caveat must be noted. Recent work (Chapter 6) has shown that the PES
calibration procedure can be sensitive to large variations in sediment type. That is, PES-
flume calibrations developed for sedimented kaolinite-type clay beds, compacted
bentonite-type clay beds, and sedimented organic-type beds were internally type-
consistent but did not compare well inter-type on a shear stress basis. This indicates the
advisability of using calibration data for the PES from a sediment of similar type.
The portability of the PES has allowed its use on shipboard (Tsai and Lick, 1986,
Lavelle and Davis, 1987, Ziegler et al., 1987, Sffisco et al., 1991), as well as in the lab.
Studies utilizing the PES have been conducted in freshwater (Tsai and Lick, 1986,
MacIntyre et al., 1990, Davis and Abdelrhman. 1992, Mehta et al., 1994) as well as in the
marine environment (Tsai and Lick, 1987, Lavelle and Davis, 1987, Ziegler et al., 1987,
Sffisco et al., 1991, Davis and Abdelrhman, 1992). Specialized studies utilizing the PES
have included investigation of the effect of bioturbation on sediment erosion
characteristics (Davis and Means, 1989, Davis 1993) and studies of the impact of
catastrophic events on short term sediment transport (Mehta et al., 1994).


379
BS 3.4040 3.4040
SECT 34 34 34 40 202 1.91 1.98 17.00 17.00
W 270.0 0.0
H 0.0000.37200 1.1400
BS1 4.5200 4.5200 7.7200
BS2 4.6720 4.6720 8.5430
SECT 35 35 35 36 780 2.31 2.31 17.00 17.00
W 180.0 0.0
H 0.0000.73500
BS 3.2520 3.2520
SECT 37 37 37 43 202 1.72 1.68 17.00 17.00
W 270.0 0.0
H 0.0000.55800 1.0520 1.3260
BS1 9.7140 9.7140 13.432 13.971
BS2 9.2290 9.2290 13.429 13.929
SECT 38 38 38 39 792 1.58 1.58 17.00 17.00
W 180.0 0.0
H 0.0000 1.4690
BS 3.2000 3.2000
SECT 40 40 40 46 201 1.98 1.98 17.00 17.00
W 270.0 0.0
H 0.0000.30500 1.0730
BS1 4.6720 4.6720 8.5430
BS2 3.9620 3.9620 7.8970
SECT 41 41 41 42 780 2.26 1.81 17.00 17.00
W 180.0 0.0
H 0.0000.78650
BS 3.2000 3.2000
SECT 43 43 43 49 201 1.68 1.41 17.00 17.00
W 270.0 0.0
H 0.0000.59700 1.1400 1.3650
BS1 9.2290 9.2290 13.429 13.929
BS2 11.701 11.701 16.894 17.010
SECT 44 44 44 45 792 1.73 1.73 17.00 17.00
W 180.0 0.0
H 0.0000 1.3170
BS 6.0960 6.0960
SECT 50 50 50 53 191 1.98 1.94 17.00 17.00
W 270.0 0.0
H 0.0000 .30500 1.0730
BS1 3.9620 3.9620 7.8970
BS2 3.6570 3.6570 6.7140
SECT 47 47 47 48 780 2.11 1.26 17.00 17.00
W 180.0 0.0
H 0.0000.93600
BS 4.5720 4.5720
SECT 49 49 49 56 201 1.41 1.91 17.00 17.00
W 270.0 0.0
H 0.0000 .86900 1.5670 1.6370
BS1 11.701 11.701 16.894 17.010


272
where
e = ap, combined erosion coefficient, time/area
Vc = Critical channel mean velocity for erosion,
length/time
The sedimentation flux is defined by the equation
Fs ~wstdCss 95
where
wsed = sedimentation velocity, length/time
Css = Volumetric concentration of suspended solids.
mass/length3
Now the change in suspended solids concentration in the control volume with
time (ignoring dispersion) is the flux times the area divided by the volume or
Qv F b&x _Fn
dt v zMx z
9.6
and the change in surface concentration of erodable material is the negative of the net flux
or
dt
= -Fu
9.7
Substituting Equations 9.4 and 9.5 into Equation 9.1 and 9.1 into 9.6 and 9.7 gives
the final set of coupled equations for the simplified transport model
^2-K2)C,-hwQv
dt z
9.8


97
1992) so it was somewhat surprising to see the very linear increase in mean sedimentation
Figure 5.13: Sedimentation Velocity Distribution as Determined by The Bottom
Withdrawal Technique-B9B10 Sediment <38 Micrometer Particle Range


APPENDIX C
CHEMICAL ANALYSES
Soluble Orthophosphate (APHA, Method 4500-P E)
The method used was the Ascorbic acid method wherein ammonium molybdate
and potassium antimonyl tartrate react in acidic medium with orthophosphate to form
phosphomolybdic acid that is reduced to intensely colored molybdenum blue by ascorbic
acid, the color intensity of which is measured by the absorbence at 660 nm. The analyses
were conducted using an Alpkem Autoanalyzer following the procedure detailed in the
referenced method, A series of calibration samples were analyzed at the beginning and
end of each autoanalyzer run and reference samples were analyzed at every tenth sample.
Particulate Phosphorus (Anderson, 1976)
Particulate phosphorus samples came from several sources including soils,
sediments, plant matter, detritus, and particulate matter retained on glass microfiber
filters. Non-filter samples were dried at 70C for three days then milled or ground with
mortar and pestle, redried at 103-105C for 1 hour and reweighed. Filter samples were
subjected to the TSS analysis which included drying at 103-105C for 1 hour.
Subsequent to drying and weighing samples were placed in 50 ml high-silica beakers and
ignited in a muffle furnace at 550C for 1 hour. Following cooling the samples were
digested using a hydrochloric acid form of the strong acid digestion procedure described
by Anderson (1976). Samples were digested to dryness on hotplates from an initial
charge of 25 ml of 6N HC1 per sample. The sample was then reconstituted with 25 ml of
hot 0.3N HC1 and immediately filtered to remove particulate matter. The filtrate was
analyzed for orthophosphate using APHA Method 4500-P E.
363


270
5. Above the critical shear stress erosion rate, mass/unit time/unit area, is
proportional to the difference between the existing shear stress and the
critical shear stress.
6. Erosion rate is also proportional to the surface concentration of erodable
material remaining, and thus is a first order process with respect to
erodable material.
7. Sedimentation rate is represented by a single characteristic sedimentation
velocity that is assumed to apply to all particles.
8. At any point in the conveyance network the crossection of the channel can
be represented as a rectangle without significant loss of accuracy.
9. Axial dispersion may be ignored in the derivation of the model because it
is accounted for in the numerical solution computed by DUFLOW.
Model derivation
With these assumptions in place the derivation of the simplest possible model that
fulfills the assumptions follows.
Consider a control volume rectangular in shape with depth = z, width = b, length
= Ax, which contains erodable material on the bottom at a surface concentration of
Cem (mass/unit area), and suspended solids at a volumetric concentration of Css
(mass/unit volume). The net flux of erodable material from the bottom is given
by
Fr-R
9.1
where


33
Kemp et al. (1984) carried out a comprehensive study of sediment mass flux and
chemical composition as influenced hv submersed macrophytes in a Chesapeake Bay
tributary, using a variety of sampling and analytical techniques which included harvest of
standing crops, core sampling, water sampling, respirometry studies, analyses for
chlorophyll-a, stable carbon isotopes, total carbon, total nitrogen, nutrients, suspended
material and wet sieve particle size analysis. They also included work from previous
studies which included seasonal budgets for organic carbon, sediments, and nitrogen.
They found that the vascular aquatic vegetation played a strong source-sink role in
trapping POM and retarding its movement during the growing season but contributed
about one-third of the total annual organic carbon budget during times of senescence and
death. Phytoplankton content of trapped sediment ranged from 10%-40% as estimated
from chlorophyll-a and stable carbon isotope analysis. Their annual budgets were
particularly interesting. They showed that the sediment sink load within the macrophyte
beds, expressed in kg/yr., was more than twice the total sediment source load on the study
area from river input and shore erosion, indicating that processes internal to the study area
contributed as greatly to sediment load as external inputs.
The nature of transportable organic sediment as a collection of agglomerates or
aggregates of organic and inorganic matter has been addressed by several authors.
Kranck (1984), studying transport in estuaries, found that the suspended particulate
flocculated matter in three separate estuaries had an organic content in the range of 65-
75% by volume. She theorized that this organic content represented an optimum
composition for floe formation under the estuarine conditions studied. Assuming specific
gravities of 1.1 for organic matter and 2.7 for minerals would make Kranck's values 45-
55% by weight. Mirbagheri et al. (1988) studied a California agricultural runoff and
irrigation watershed using floating single-stage and multi-depth autosamplers, Van Dorn
plankton-sampling bottles, and Ponar bottom sediment samplers. They evaluated total
suspended solids, suspended organic matter, suspended algae, chlorophyll-a and


281
there was a major difference in the absolute values of sedimentation velocities determined
by the measurement techniques of Chapter 5 and the calibration techniques of Chapter 9.
the calibration value corresponds to an estimated particle size that is reasonably
consistent with the concept of the suspended solids being sourced from dislodged and
resuspended detrital material.
Normal Wet Season Simulation Results
Suspended solids simulation
The suspended solids simulation results for the six wet season events used in the
calibration process are shown in Figures 10.1-10.6. These figures will differ somewhat
from those presented in Chapter 7. They tend to emphasize the first major pumping
cycle, where the bulk of the solids transport occurs and where the calibration was
focused, and so will be of shorter duration than the complete curves of Chapter 7. The
Observed data points on these plots, which represent a composite of two hours of
pumping time, are located at the centerpoint of their respective time periods, rather than
at the endpoint as they were in Chapter 7. It is important to keep in mind when viewing
these curves that the Observed points represent composite sample concentrations
averaged over two hours of pumping time while the Simulated lines represent the
simulated instantaneous concentrations, so even in an ideal simulation the Observed
points would not always coincide with the Simulation line.
Qualitatively the model did a good job of simulating the U shaped transport
curve in its various configurations and magnitudes among the six events. It also appears
to simulate the transport occurring during pump on-off cycling reasonably well, as seen
particularly in Events 220,237, and 252 (Figures 10.1-10.3). Event 285 has some notable


88
samples (89.1 and 113.0 mg/kg) which had a substantially higher organic content. The
maximum srbate concentration. Sm was very similar (495-520 mg/kg) for the three
higher organic content samples and much lower (110.8 mg/kg) for the low organic
content WPBC 75-150p sample. The affinity of the substrate for the srbate is indicated
by the value of Ks, with affinity increasing as Ks decreases. Ks was lower for the WPBC
samples (2.0-2.6 mg/1) than for the B9B10 samples (6.3-7.8 mg/1), indicating a somewhat
higher affinity in the WPBC samples.
Adsorption Isotherm WPBC Sediment 75-150 Micrometer Particle
Range
1000
Figure 5.9: Adsorption Isotherm-WPBC Sediment 75-150 Micrometer Range
The parameter of particularly important interest in this analysis was the linearized
partition coefficient, k, which may be used to estimate equilibrium at low liquid phase
phosphorus concentrations. The values of k ranged from -56-198 1/kg. The lowest value
of 56.3 1/kg was exhibited by the WPBC 75-150p sample, which, it should be recalled,
had the highest contained phosphorus content and lowest organic content of the four


74
In addition two specific particle size cuts from the B9B10 and WPBC sediments
were given more intensive evaluation. The smallest particle size fraction, the -400 USS (-
38 micron) cuts, and a larger particle size fraction, the -100 USS +200 USS (75-150
micron) cuts, were selected fqr evaluation of phosphorus forms present and also for
development of phosphorus adsorption/desorption data.
The phosphorus forms evaluated (See Appendix C for methods) were
1. Bicarbonate extractable, which represents the labile fraction of phosphorus
2.
that is readily available for biological uptake,
jyfcL ,, ...
Hydrochloric acid extractable, which represents the nominal maximum
amount of phosphorus present that is exchangeable with aqueous solutions
under environmental conditions and, by difference,
3. Y Residual phosphorus, which is relatively refractory, usually organically
bound phosphorus, that requires biochemical breakdown of the containing
organic matrix in order to become biologically available.
The adsorption/desorption data were developed using the methods described later
in this chapter.
Particle size distribution
Soil Dry and Hydrated Figure 5.1 shows the results of the dry screening before
and the wet screening after the eight day hydration. There appeared to be substantial de
agglomeration resulting from the hydration, as indicated by the reduction of the mass
1 I
mean particle size from approximately 800 micrometers in the dry soil down to
approximately 250 micrometers in the hydrated soil. This de-agglomeration appeared to
produce a large population of particles in the most-readily transportable range of <38
micron diameter (-400 USS) as evidenced by the increase of this mass fraction from 0.4%
dry to 11.7% hydrated.


337
was gradually drawn down as the season progressed. This may offer partial explanation
for the greater decreasing trends at UF9206.
Correlation of export suspended solids with channel depth
All these observations may be interpreted to imply that the conveyance system at
UF9206 was subjected to more extreme variations in level and, by inference, velocity,
than was the conveyance system at UF9200. An evaluation of suspended solids
correlation with canal level at the pump stations is summarized in Figure 11.14 for
UF9206N and Figure 11.15 for UF9206S. Several negative levels were recorded for
UF9206S. In order to allow a power-law type correlation for this location a correction
value of 0.43 meters was arbitrarily added to all levels so that the minimum correlation
level for both canals was about 0.3 m. For comparison the same type correlation was
developed for UF9200. At UF9200 the bottom elevation of the canal was known so the
correlation may be presented in the form of canal depth. This correlation is presented in
Figure 11.16.
It is clear from visual observation of the figures that there was a relatively strong
correlation between canal level and discharge TSS concentration at both UF9206 pump
stations and a relatively weak correlation at UF9200. This is borne out by the r2 values of
0.66 for UF9206N, 0.49 for UF9206S. and only 0.12 for UF9200. The canal levels at
UF9206 are very approximately equal to the actual depth so the exponent of the level may
have some physical meaning. The exponent in both cases at UF9206 is approximately 3,
while the (poorly correlated) exponent at UF9200 is approximately 2. Now at any
constant pump rate the velocity in a rectangular channel is proportional to the depth so the
velocity squared is proportional to the depth squared. Similarly, in a triangular channel
the velocity is proportional to the depth squared so the velocity squared is proportional to


279
Table 10.1: Erosion Parameters Determined in the Calibration Process
Calibration Parameter
Value
Erosion Coefficient, 8
1.5 x 1 O'7 days/m2
Sedimentation Velocity, wse(i
125 m/day
Critical Velocity for erosion, Vc
0.095 m/sec
Initial Erodable Mass/Unit Area, Cemioi
Value
Event 220
1800 gm/m2
Event 237
350 gm/m2
Event 252
350 gm/m2
Event 258
300 gm/m2
Event 262
325 gm/m2
Event 285
550 gm/m2
The correlations presented in Chapter 6 allow the calculation of a critical velocity
for erosion by setting CSME equal to zero in Equation 6.27 to determine the critical
oscillation frequency of the PES (141 RPM) and then substituting that value in Equation
6.31 to get a critical mean channel velocity of 0.091 m/sec, which is surprisingly close to
the value of 0.095 m/sec that was determined completely independently in the calibration
process. Note also that the calculated critical velocity for the PES field samples from
UF9200 and UF9206 which showed high erosion rates were in the range of 0.10-0.11
m/sec (See Figures 6.23 and 6.24), so the calibration value of critical velocity appears to
be fairly close to the critical velocity determined for several other organic sediments in
this study.
The sedimentation velocity, however, does not show the same relationship.
Recall from Chapter 5 that the mean sedimentation velocity determined for the <38


152
(UF9200) and 6.24 (UF9206). In these graphs, CSME is plotted as a function of (Vmean)2,
the square of the equivalent open channel velocity estimated using Equation 6.31.
For the UF9200 results six of the seven eroding samples showed similar slopes of
CSME with (Vmean)2 (Average = 0.060 gm-sec2 /cm4 ct = 0.012). For these six samples
the highest CSMEs (at 275 RPM) were in the range of 0.0012 to 0.0021 gm/cm2,
(average 0.0018 gm/cm2) The remaining sample showed a significant departure from the
others. Its slope was 0.342 gm-sec2 /cm4, almost six times the average of the other
samples, its highest CSME (at 275 RPM) was 0.0074 gm/cm2, which was four times that
of the average of the other samples.
For the UF9206 there was a bit more variability but note that UF9206-1 had some
weed stalks protruding through the surface of the sediment. This condition has been
associated with accelerated erosion in the lab studies. If the results from UF9206-1 are
removed from consideration then the UF9206 PES results show a pattern similar to the
UF9200 results. Four samples had an average slope of 0.104 gm-sec2 /cm4, while the
remaining two samples had an average slope of 0.621 gm-sec2 /cm4, over six times
greater. The same four samples had an average maximum CSME (at 275 RPM) of
0.0028 gm/cm2 compared to an average of 0.0153 gm/cm2 for the two remaining samples,
a five-and-one-half-fold difference. While the absolute values of erosion extent were
greater for the UF9206 samples than for the UF9200 samples, the relationships between
the high and low clusters in the UF9206 data set were similar to the relationship between
the one high sample and the remaining six samples from UF9200.
These results showed the opportunity for considerable variation in the erodability
of the farm sediments studied but at the same time indicated a possible pattern for that
variability. Study of Table 6.4 indicates that the one parameter that the three high eroding
samples had in common and the remaining less erodable samples did not was the
presence of floating macrophytes either at or in the vicinity of the sample point. At this
time it is appropriate to reiterate that, unless they are so prevalent that they cover the


LD
1780
199,
sm
UNIVERSITY OF FLORIDA
3 1262 08554 9235


54
The figure shows a qualitative model with the following main features:
Soil water supplies or removes soluble phosphorus (SP) to or from the
canal water depending on whether irrigation or drainage is being practiced.
Soil water, moving through (perpendicular to) the canal wall, supplies or
removes particulate phosphorus (PP) to or from the canal water depending
on drainage (perpendicular flow erosion) or irrigation (perpendicular flow
filtration). PP is also supplied to the canal water from the bed by parallel
flow erosion arising from longitudinal flow in the channel, and removed
from the canal water by sedimentation.
Precipitation supplies SP and PP directly to the canal water, and causes
surface runoff under some circumstances which also supplies SP and PP.
Biological activity (shown as, but not limited to, phytoplankton and
P\
macrophytes) immobilizes SP, and increases or decreases pH, depending
on photosynthetic vs. respiration activity. Particulate detrital biological
material is deposited on the top layer of sediment, which is assumed to be
oxic.
Inorganic particulate matter suspended in the ditch and canal water also
interchanges with the oxic sediments via erosion and sedimentation.
Oxic sediments interchange with anoxic sediments, and both
compartments undergo mineralization of organic PP to SP. Anoxic
sediments may interchange with suspended particulates under the
appropriate hydrodynamic conditions. Anoxic sediments also interchange
with the high-calcium-containing substratum layer.
Soil water may infiltrate through the substratum, losing SP and gaining
Ca that is earned into the bulk phase canal water for precipitation
reaction with SP under appropriately high pH (from photosynthesis). All


277
event to event there was a narrow range of absolute values of e and wsed that could be
used and still give a reasonable fit across all six events.
The value of Vc had almost no effect on the ends of the curve but strongly
impacted the lower values in the concave portion, or trough of the curve. As mentioned
above, Cemio) would scale the absolute magnitude of the curve height, but if it was too
low there would not be enough erodable material available under any circumstances and
the U curve could not be reproduced. If it was too high there would be too much
erodable material available the U curve would flatten by increasing the trough, and
much higher values of Vc would be necessary to fit the field data.
It is possible that other combinations of parameters could have fit the data equally
well or better and this must be kept in mind when the simulation results are interpreted,
but it is felt that the method of calibration gave parameters that are reasonably close to
optimum. The results are discussed in the next chapter.


250
was supported by a staff in The Netherlands who could provide technical service. Most
important from the perspective of this project, it was the only program in the group that
already contained an integrated water quality subroutine.
The advantages of the DUFLOW program were significantly greater than any of
the others reviewed, but there were some disadvantages. First, neither DUFLOW nor any
of the other existing programs had the capability to represent field runoff as subsurface
flow moving with the hydraulic gradient between the groundwater and the receiving
channel, the primary mode of transport in the EAA conveyance systems. Second, the
source code was not available for release to users. These two disadvantages were
circumvented. Partial access to the source code for the purpose of extending the
capabilities of the model was negotiated between Soil and Water Engineering Technology
and Informtica Centrum voor Infrastructura en Milieu (ICIM), the group responsible for
maintaining and administering the program. The groundwater flow deficiency was
corrected by the development, by Dr. Pickering, of a groundwater-canal stage algorithm
as an adjunct sub-routine of the program. This addition gave the program all the platform
elements necessary to model particulate transport in the EAA conveyance systems.
General characteristics of the DUFLOW program
The DUFLOW program was reviewed thoroughly by Clemmens, et al. (1993).
The following description derives from that review and the DUFLOW operating manual
(ICIM 1992). The computational methodology utilizes a four point implicit Preissmann
scheme (Cunge. et al. 1980) to solve the complete St. Venant equations of continuity and
motion using a Gaussian elimination technique, which allows any network or loop
configuration to be included. There is a program option that allows the user to simplify
the calculations by eliminating or damping the so-called Froude Term. -Q-. in the
At dx
advection term of the momentum equation. This term has small impact on large, slow


273
and
at
with the initial conditions
Css = zero at time zero
and
Cem = Cem(O) at time zero, where Cemioi is the uniform initial distribution
of erodable mass for each event
The model now consists of two coupled linear differential equations that share the
three basic erosion/sedimentation parameters of the erosion coefficient, e, the critical
velocity for erosion, vc and the sedimentation velocity, wseij, and have an initial
condition that has yet to have its form specified. The DUFLOW sub-routine that
expresses these equations in DUPROL format is shown in Appendix FI.
Transport model calibration methodology
The methodology of the transport model calibration is presented here, the results
are presented and discussed in the next chapter. The top priority for the calibration
process was to duplicate the primary pump cycle U shaped suspended solids
concentration curve in its various forms for each event of the wet season from Event 220
through Event 285. This represents calibration over 75-90% of the total suspended solids
load for each event. Once calibrated for suspended solids concentration, the particulate
phosphorus concentration would be calculated using the phosphorus content-suspended
solids concentration correlations developed in Chapter 7. The criteria for optimization
was visual evaluation of suspended solids concentration fit during the first major pump
cycle to minimize the sum of the absolute deviations between calculated and observed
suspended solids. The difficulty was compounded somewhat by the fact that the field


382
BS1 7.5430 7.5430 11.301
BS2 6.5770 6.5770 9.9540
SECT 85 85 85 86 780
W 180.0 0.0
H 0.0000 1.4960
BS 2.8440 2.8440
SECT 87 87 87 93 201
W 270.0 0.0
H 0.0000 .3 7800 1.0970 1.1460
BS1 6.7100 6.7100 14.490 14.640
BS2 6.7100 6.7100 14.490 14.640
BB1 9.4880 9.4880 15.389 15.474
BB2 11.789 11.789 17.787 17.787
SECT 88 88 88 89 792
W 180.0 0.0
H 0.0000 1.4960
BS 3.2520 3.2520
SECT 90 90 90 94 190
W 270.0 0.0
H 0.0000.51500 1.2830
BS1 6.5770 6.5770 9.9540
BS2 5.4100 5.4100 7.9210
SECT 91 91 91 92 780
W 180.0 0.0
H 0.0000 1.5450
BS 2.9960 2.9960
SECT 93 93 93 193 95
W 270.0 0.0
H 0.0000.33200 1.1000
BS 6.7100 6.7100 14.490
BB1 11.789 11.789 17.787
BB2 9.3100 9.3100 13.300
SECT 96 96 96 97 121
W 180.0 0.0
H 0.0000.47500 1.2440
BS1 6.7300 6.7300 8.8140
BS2 5.3580 5.3800 8.5860
SECT 97 97 97 98 162
W 180.0 0.0
H 0.0000.63400 1.4020
BS1 5.3580 5.3580 8.5860
BS2 6.5280 6.5280 8.3690
SECT 98 98 98 99 168
W 180.0 0.0
H 0.0000 .68600 1.4540
BS1 6.5280 6.5280 8.3690
BS2 7.0860 7.0860 10.405
SECT 99 99 99 199
W 180.0 0.0
H 0.0000.78000 1.5480
1.55 1.55 17.00 17.00
1.90 1.95 17.00 17.00
1.55 1.55 17.00 17.00
1.76 1.51 17.00 17.00
1.50 1.50 17.00 17.00
1.95 1.83 17.00 17.00
1.80 1.65 17.00 17.00
1.65 1.59 17.00 17.00
1.59 1.50 17.00 17.00
167 1.50 1.45 17.00 17.00


203
South Canal
North Canal
X
.-XT)
-X2'1) X
: V
/s
o II
A
iTV^*
LEGEND:
Field ditch
HI Roadway
Canal
1 Pump
Riser
Culvert
West Palm Beach Canal
Figure 7.37: Layout of UF9200 with Synoptic Sampling Locations


313
Validation Results When Assumption #3 was adhered to in the validation
process the model simulated the ultimate suspended solids and particulate
phosphorus loads with approximately the same accuracy as was achieved
in the calibration period. The interim values of cumulative phosphorus
load tended to be underpredicted early in the simulation and overpredicted
in the last stages. Considering the seasonal and hydrodynamic differences
between the calibration and validation periods the validation process was
remarkably successful.
Model qualifications
The assumptions and calibration results noted above impose some important
qualifications and limitations on the model.
First, because of the way in which the model was calibrated, its parameters must
be considered to be site specific. The biota throughout the EAA are similar but densities,
locations, and population distributions would be expected to vary from farm to farm in
response to changes in cropping practices, hydraulic configurations, water management
policies, and conveyance maintenance procedures. The pump configuration and pump
operation techniques at UF9200 are not necessarily representative of all EAA farms.
Differences in pump configuration and operation at other locations could bring factors
into play that were not considered at UF9200, as we shall see in the next chapter.
Second, the lack of adequate simulation of transport during small flows might
tend to give an over-optimistic prediction of the results of remedial actions predicated on
the models simulations. This was not particularly troublesome for the calibration period
but exclusion of Event 336 from the validation period simulation resulted in the loss of


240
Phosphorus fractionation tests on large-scale composite samples
Phosphorus fractionation tests, similar to those run on the prototype sediments
reported in Chapter 5, were run on the six large-scale composite samples. In this case the
inorganic and organic acid and base extractable fractions were analyzed separately and
the refractory residue was taken to be completely organic. The results appear in Table
8.6, which also includes the average phosphorus content of the discrete time-series
samples taken at the same time the large-scale composite samples were taken.
The Large-Scale Composite samples showed a trend toward higher phosphorus
content than the parallel discrete time-series samples. The Large-Scale Composite
samples were brought back from the field on 8-12 hour schedules and were placed under
refrigeration sooner than the typical discrete samples. This may be an indication that
there was some solubilization of labile particulate phosphorus in the discrete samplers
during their stay in the field.
The two UF9206N samples were very similar in phosphorus content and in the
way the phosphorus content was distributed. The UF9200 samples were somewhat more
variable, but showed coefficients of variation usually in the range of 10-20%, which is
reasonably low. In general there was good internal consistency in the phosphorus fraction
analyses at each site.
The samples from UF9206N were lower in phosphorus content than those from
UF9200 which mirrors the general trend seen throughout the season, but the interesting
comparison is the marked similarity between the two sites of the average percentage
distribution of the various fractions. Each fraction from each site averaged within one or
two percentage points of the corresponding fraction at the other site, suggesting strong
similarity between the phosphorus sources at each location, with the absolute difference
in phosphorus content possibly being due to dilution with non-phosphorus-containing
material.


APPENDIX G
FARM LAYOUTS FOR UF9200 AND UF9206
Figure G. 1: Layout of UF9200
386


71
collected, stored in sealed plastic buckets, and transported to the Gainesville laboratories
for fractionation
Soil and sediment particle size fractionation
Several fractionation methods were evaluated and tested on surrogate sediments
prior to sampling the target sediments. The methods included screening, hydraulic
classification, and differential sedimentation. It was hoped to develop a method that
would fractionate on the basis of sedimentation velocity, which would relate to a key
physical parameter of field interest. Unfortunately all the hydraulic classification and
differential sedimentation techniques evaluated either suffered from a lack of available
equipment that could process 20 liters of material in any reasonable time span or
subjected the test material to contact with large volumes of water. The latter condition
was deemed to be inadvisable because of the potential for alteration of the physical or
chemical nature of the materials over repeated dilutions with water. Ultimately the
fractionation method of choice reduced to wet and dry screening. The three samples were
fractionated in their entirety on US Standard screens in the sequence shown in Table 5.4.
Note that + means retained on, while means passed through.
Soil Fractionation The Soil sample was first subjected to coarse screening to
remove clumps greater than 1 cm. The remaining material was dried at 70 C for 72
hours and then subsampled to produce a representative composite. The composite was
split into two sub-segments of roughly 315 gm each. Each sub-segment was subjected to
screening in a stack of USS sieves on a Ro-Tap shaker table for 30 minutes, then
removed, weighed, and recombined. In order to place the evaluation of the soil particle
fractions on the same basis as the sediment it was deemed appropriate to hydrate the soil
to simulate the process of soil eroding into a conveyance system and becoming subjected
to an aqueous environment.


16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
374
2615
792
8090E+01 1.00
2615
12
8090E+01 1.00
2615
0
OE+OO 1.00
2414
1609
OE+OO 1.00
2414
817
1619E+02 1.00
2414
805
0E+00 1.00
2414
792
8090E+01 1.00
2414
12
8090E+01 1.00
2414
0
OE+OO 1.00
2403
805
OE+OO 1.00
2213
1609
OE+OO 1.00
2213
817
1619E+02 1.00
2213
805
OE+OO 1.00
2213
792
8090E+01 1.00
2213
12
8090E+01 1.00
2213
0
OE+OO 1.00
2012
1609
OE+OO 1.00
2012
817
1619E+02 1.00
2012
805
OE+OO 1.00
2012
792
8090E+01 1.00
2012
12
8090E+01 1.00
2012
0
OE+OO 1.00
1810
1609
OE+OO 1.00
1810
817
1619E+02 1.00
1810
805
OE+OO 1.00
1810
792
8090E+01 1.00
1810
12
8090E+01 1.00
1810
0
OE+OO 1.00
1609
1609
OE+OO 1.00
1609
817
1619E+02 1.00
1609
805
OE+OO 1.00
1609
792
8090E+01 1.00
1609
12
8090E+01 1.00
1609
0
OE+OO 1.00
1599
805
OE+OO 1.00
1408
1609
OE+OO 1.00
1408
817
1616E+02 1.00
1408
805
OE+OO 1.00
1408
792
8090E+01 1.00
1408
12
8090E+01 1.00
1408
0
OE+OO 1.00
1207
1609
OE+OO 1.00
1207
817
1619E+02 1.00
1207
805
OE+OO 1.00
1207
792
8090E+01 1.00
1207
12
8090E+01 1.00
1207
0
OE+OO 1.00
1006
1609
OE+OO 1.00
1006
817
1619E+02 1.00
1006
805
OE+OO 1.00


269
Spatial variability undoubtedly exists but there are insufficient data to
define the specifics. Spatial variability of erosion and sedimentation
parameters should be excluded from the model development.
Some temporal variation of parameters may be discernible from the field
data. Temporal variation of selected parameters should be allowed if it
may be justified logically.
The model should exhibit the simplest possible combination of functions
to describe the observed results.
Model Development
General model assumptions
The basic assumptions for the model development are as follows.
1. At the start of each event there is a specific amount of erodable material
uniformly distributed across the entire conveyance system. For
convenience the material is assumed to be on the bottom of the channel
but that is not necessary.
2. The distribution of the erodable material is expressed in terms of a surface
concentration, mass/unit area, where, for convenience, the area is based on
the channel bottom.
3. Shear forces acting on the erodable material are proportional to the square
of the channel mean velocity. The effects of changes in hydraulic radius
are ignored.
4. The erosion of this material exhibits a threshold or critical shear stress
below which erosion is negligible.


6SURFICIAL SEDIMENT TRANSPORT STUDIES
.99
Introduction 99
Prototype Sediment 100
Laboratory Erosion Simulation Device- The Counter-Rotating Annular
Flume 101
Flume description 101
Flume operation 105
Organic sediment flume studies 107
Organic Sediment Particle Entrainment Simulator Studies 132
Particle Entrainment Simulator configuration 132
Particle Entrainment Simulator operation preliminary clay tests 136
Particle Entrainment Simulator operation organic sediment tests 137
Calibration of PES against flume and extrapolation to field
parameters 141
Field Measurements using the calibrated particle entrainment
simulator 148
7 FIELD EVENT STUDIES-TIME SERIES AND SYNOPTIC 155
Introduction 155
Time Series Discharge Studies 156
Data Monitoring and sample acquisition 156
Detailed hydrography of the target study farm~UF9200 159
Particulate phosphorus content determination 161
Particulate phosphorus event studiesFarm UF9200 163
Suspended solids transport-Farm UF9200 168
Particulate phosphorus content of suspended solidsFarm UF9200 176
Particulate phosphorus event studiesFarm UF9206, North and
South stations 181
Suspended solids transport Farm UF9206 189
Particulate phosphorus content of suspended solids Farm UF9206 196
Intensive Synoptic Studies at UF9200 200
Sampling configuration 200
Ditches vs. discharge 204
Influence of the North Canal 207
Downstream variation of suspended solids concentration 208
Order of Magnitude Transport Estimates 211
Summary 216
8 FIELD SPECIFIC STUDIES-FIELD SEDIMENTS, MACROPHYTE
AND DETRITUS STUDIES, LARGE COMPOSITE SAMPLES 217
Introduction 217
Sediment Phosphorus Content Synoptic Surveys 218
Macrophyte and Detritus Studies 224
viii


366
CRAF Approximate Model: Velocity-Depth Pro fi les
n = 0 025 n = 0 .0 I 8
Figure D.l: CRAF Approximate Model: Velocity-Depth Profile
CRAF Approximate Model: V e lo c ity-D e p th Profiles
(N on-D im ensionalized in Velocity)
n = 0 .02 5 n = 0.018
Figure D.2: CRAF Approximate Model: Velocity-Depth Profile (Dimensionless
in Velocity)


8
Fertility BMPs include the use of calibrated soil testing to minimize over
fertilization, banding of fertilizer to minimize placement of fertilizer in areas where it will
not be accessible to the root zone, implementation of methods to prevent misplacement of
fertilizer into drainage ditches, and split multiple applications of smaller loads of fertilizer
to more closely match the crop uptake. Water management BMPs are aimed at
maintaining the water table as close as possible to the maximum feasible level at all times
to minimize oxidation andmineralization of the soil and recycling of runoff to retain
phosphorus on-site, where possible. They include measures to minimize water table
fluctuations, maintain uniform spatial and temporal distribution of the water table, and
store runoff on fallow land or water insensitive cropland.
The water management BMPs require the grower to exercise much more
sophisticated control[over runoff, irrigation, and ground water than has been practiced in
the past. The grower is required to develop and track field water budgets, which
incorporate irrigation, rainfall, and evapotranspiration accounting to optimize pumping
and minimize excessive downward water table movement. Minimum canal-level pump
shut-off controls are recommended to reduce excessive canal drawdown. Comparison of
predicted rainfall with available field water storage capacity is recommended to reduce
unnecessary prepumping in anticipation of storms. A number of hydraulic BMPs relate to
achieving and maintaining hydraulic uniformity within the farm canal system to minimize
unnecessary local drawdown of the water table. Water management BMPs are also
detailed for retention of drainage water in on-farm reservoirs where possible, transfer of
drainage water from more water-sensitive crops to fields with less water-sensitive crops,
and the use of aquatic cover crops such as rice for uptake of phosphorus released
elsewhere on-farm.
The STAs, which are to affect 75% of the phosphorus load reduction, are planned
to be large artificial wetlands which will act as phosphorus removal buffers between the
EAA and the WCA/ENP areas. These STAs may ultimately cover as much as 14,500 ha,


73
screen. This way contact of sediment with fresh water was eliminated. Each screen size
fractionation was judged to be complete when visual examination of the screen subnate
indicated that relatively clear liquid was being passed.
After completion of each screening run the screen retntate was scraped into a
plastic bag and the screen was backwashed with a small amount of clear subnate. The
backwash was added to the retntate bag and the bag was refrigerated until analysis.
Screen subnate, that contained all particle sizes less than the current screen run, was
stored under refrigerated conditions between screen runs. The final cut, the <38
micrometer material, was separated from the screen subnate water by centrifugation and
stored in plastic bags until analysis.
The final subnate water was analyzed for soluble phosphorus to determine if a
substantial fraction of the sediment phosphorus had been released. In the case of B9B10
sediment the total mass of phosphorus in the final subnate water was about 1% of the
total mass of phosphorus remaining in the sediment fractions, and was about 0.4% for the
hydrated soil samples. For the WPBC sediment the subnate soluble phosphorus was less
than 0.1 % of the remaining sediment phosphorus. Based on these results it was .
reasonable to assume that the wet sieving process had
sediment particulate phosphorus content.
Fraction analysis
The various fractions were analyzed for:
1. Total dry mass
2. Volatile matter/Ash content
3. Particle specific gravity
Total phosphorus content
4.


169
Time Series-Flow and Total Suspended Solids
Event UF9200-154
Figure 7.3: Hydraulic and Total Suspended Solids Profiles for Event UF9200-154
Time Series-Flow and Total Suspended Solids
Event UF9200-220
Pumping Rate 0 TSS Concentration
Figure 7.4: Hydraulic and Total Suspended Solids Profiles for Event UF9200-220


226
What was deemed important at the time the field data were being gathered was
some justification that the macrophytes were in fact capable of contributing significant
amounts of particulate matter that were sufficiently high in phosphorus content to
coincide with the observed high phosphorus contents of the discharged suspended solids.
Several studies that were set up to provide this justification are described in the following
sections.
Macrophyte areal density and gross total phosphorus content studies
Samples of each of the three predominant species were taken in such a manner
that their areal mass density could be estimated. These samples were processed in-total
and then sub-sampled for phosphorus content analysis in a way that maximized the
probability that the sub-sample was representative of the average of the total sample.
Specifically the process was as follows.
A one meter square floating grid was cast into a section of canal where the
macrophyte population was predominantly of the species of interest. All plant growth
within this grid was harvested, drained and transported to the EREC laboratory at Belle
Glade, FL, where the growth was sorted by species. The non-target species were counted
then removed from the main target mass. The main target mass was then air dried for at
least three days to reduce volume and mass. After air drying the plant mass was
transferred to loose-weave bags and dried further at 70C for at least two weeks.
Following the second drying the plant mass was milled to less than 100 micrometers in a
screened hammer mill and then blended by shaking in a closed bucket to insure
homogeneity. Subsamples were withdrawn from the blend, dried at 103C for one hour,
and stored in a dessicator for further analysis.
Samples were also collected of floating submerged lyngbya filamentous algae,
which had not been incorporated into the sediment, but because of the submerged


22
South Florida Water Management District Canal Data Set
The data set from which Anderson et al. (1992) extracted their excerpts was
requested from the SFWMD for evaluation. What was provided (Mucinic 1994) was an
extensive data set containing analytical data on 1,996 samples taken from 1984-1993 at
various locations along the main EAA canals, along with logged flow data for the various
pumping stations at the periphery of the EAA. The primary thrust of the sampling
program was determination of total phosphorus and ortho-phosphorus, which were taken
at each sample site roughly monthly. Samples that included analysis for total suspended
solids were taken roughly quarterly.
Upon examination of this data set and previous interpretations of its analog for
1984-1991 it became evident that there were some opportunities for misinterpretation.
First, the pumping records showed that there were numerous occasions when samples
were taken that there was no flow in the sampled canals. Second, there was no
independent analysis of particulate or particulate phosphorus. Prior evaluations of this
data set had calculated particulate phosphorus as the difference between total phosphorus
and soluble ortho-phosphorus. This may introduce non-trivial error if there is significant
dissolved organic phosphorus present, which is not detected by the soluble ortho
phosphorus analytical technique.
The data set provided by the District did contain a small sub-set of samples, taken
from several of the northern EAA pump stations which discharge into Lake Okeechobee,
upon which both soluble ortho-phosphorus and total dissolved phosphorus analyses had
been conducted. These data were used to develop an estimate of the proportions of
soluble phosphorus which were organic (24.9%) and inorganic (75.1%). This distribution
was assumed to hold for all samples and was used as a correction factor to estimate actual
total dissolved phosphorous from soluble ortho-phosphorus data.


216
Summary
This chapter has discussed the results of extensive particulate phosphorus export
studies for three pumping stations on large farms within the EAA and synoptic studies of
particulate transport for four events at the target farm. The inferences presented by the
data are unmistakable.
In most circumstances the phosphorus content of the exported suspended solids
was significantly higher than that of the surficial canal and ditch sediments that had been
studied earlier in the program. The phosphorus content appeared to be correlated with
hydraulic activity. That correlation was an inverse one, where the phosphorus content of
exported particulate matter decreased with increasing hydraulic activity.
The synoptic studies indicated that the sediments from the field ditches were most
likely not contributing in any meaningful way to the exported particulate phosphorus
load, and that most of the load was probably originating in the main farm canals. The
erosion studies and the transport estimates made therefrom indicated that the surficial
sediments studied were not capable of providing sufficient transportable matter to satisfy
the measured suspended solids export loads for most studied events.
The suspended solids concentrations measured in the synoptic studies indicated
that the suspended solids tended to move in wavelike patterns with peaks of
concentrations moving downstream over time.
The findings with respect to surficial sediment transport raised particularly
important questions regarding the source of the exported particulate phosphorus.
Additional studies in this direction were pursued and will be discussed in the following
chapter.


104
complicating factor of an external recycle loop that would be necessary in a raceway
system. The closed loop also allows the study of sedimentation under dynamic
conditions.
The major drawback of the annular flume is that the centrifugal and coriolis forces
generated by its angular velocity set up secondary currents that tend to rotate about axes
normal to the radius of the flume. The presence of these secondary currents introduce
uncertainty into the evaluation of the actual stress applied to the flume bed. Mehta (1973)
showed that these secondary currents may be almost completely suppressed if the ring
and flume rotate in opposite directions, hence the counter-rotation aspect of the annular
flume operation. Counter-rotation alone is not sufficient to suppress the secondary
currents, the absolute speeds of the ring and annulus must be set at prescribed levels for
each value of relative speed difference between the two. These prescribed levels were
determined by Mehta by visual observation of the motion of neutrally buoyant beads in
the flume under various operating conditions.
The flume was originally calibrated by Mehta using strain gages attached to the
ring and to a similar ring attached to the bottom of the channel which functioned as a
false bottom. Shear stresses applied to the false bottom were calculated from strain
gage and dimensional and rotational data. These measured shear stresses were correlated
with the ring-channel relative speed, AV, for water depths of 15.24, 22.86, and 30.48 cm
(6, 9. and 12 inches). The strain gages were removed prior to actual sediment testing and
the correlations have been used to calculate bed shear stress on all subsequent test
sediments. The equations are:
For 15.24 cm (6 inch) depth:
Tt = 0.022(AV)"5 6.1
For 22.86 cm (9 inch) depth:


PARTICULATE PHOSPHORUS TRANSPORT IN THE
WATER CONVEYANCE SYSTEMS OF THE
EVERGLADES AGRICULTURAL AREA
By
JAMES DONALD STUCK
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1996


180
PP._ = 0.01377 x (TSS)0
7.4
where PP¡ = Particulate phosphorus concentration in the discharge for
the first four hours of an event, mg/1
PP, = Particulate phosphorus concentration in the discharge for
the remainder of an event, mg/1
TSS = Discharged suspended solids concentration, mg/1
Figure 7.16 is the analog of Figure 7.15, displaying the data cast in the form of Equations
7.3 and 7.4 in the expanded-scale mode. For the purposes of illustration, and to prepare
Figure 7.16: UF9200 Particulate Phosphorus Concentration in Discharge as a Function of
TSS Concentration (Expanded Scale)
for discussion later in this chapter, the curve that would correspond to a constant
discharge phosphorus content of 1000 mg/kg has been plotted as a hypothetical case in
Figure 7.16. This is the case that would prevail if, for example, all exported particulate


6.14 Schematic Diagram of Particle Entrainment Simulator (PES) 133
6.15 Photograph of Particle Entrainment Simulator (PES) 134
6.16 PES-CRAF Calibration Results Using Cohesive Clays 137
6.17 Typical PES Erosion Curve with Organic Sediment 139
6.18 Results of PES Runs with Prototype Organic Sediment CSME vs. Grid
Oscillation Frequency 141
6.19 Velocity Profile Measurements Reported by Mehta (1973) in CRAF with
22.9cm Water Depth 145
6.20 Schematic Representation of Superposition of Extended Logarithmic
Velocity Profile on CRAF Sigmoidal Velocity Profile 146
6.21 Final Calibration Curve of Estimated Equivalent Open Channel Mean
Velocity For Erosion as a Function of PES Oscillation Frequency 147
6.22 Example of Field PES Test Results 151
6.23 Compiled Field PES Test Results for UF9200 153
6.24 Compiled Field PES Test Results for UF9206 154
7.1 Correlation of Event TSS Load with Event Hydraulic Load for Farm
UF9200 167
7.2 Correlation of Event PP Load with Event Hydraulic Load for Farm
UF9200 167
7.3 Hydraulic and Total Suspended Solids Profiles for Event UF9200-154 169
7.4 Hydraulic and Total Suspended Solids Profiles for Event UF9200-220 169
7.5 Hydraulic and Total Suspended Solids Profiles for Event UF9200-237 170
7.6 Hydraulic and Total Suspended Solids Profiles for Event UF9200-244 170
7.7 Hydraulic and Total Suspended Solids Profiles for Event UF9200-252 171
7.8 Hydraulic and Total Suspended Solids Profiles for Event UF9200-258 171
7.9 Hydraulic and Total Suspended Solids Profiles for Event UF9200-262 172
7.10 Hydraulic and Total Suspended Solids Profiles for Event UF9200-285 172
xv


265
earlier than actual. This problem was probably also a result of simulating the entire farm
with one set of field elevations, so that runoff becomes an all-or-nothing phenomenon.
Again there was minimal impact on the objective of this work because minimal
suspended solids transport took place during these times.
The Evidence for Biological-Growth Controlled Particulate Phosphorus Transport
At this point a crucial juncture is reached in the approach to developing a
particulate phosphorus transport model. The original concept at the initiation of this
work was that the erosion characteristics of the surficial sediment could be measured and
estimated and used in an erosion/transport model to predict the export of particulate
matter, which could be related to the export of particulate phosphorus by the phosphorus
content of the sediment. A substantial amount of work went into estimating the transport
characteristics of prototype sediment, developing field techniques, and measuring field
surficial sediment erosion characteristics. In the course of the work, however, there was a
considerable body of evidence developed that cast doubt on the sampled surficial
sediment as the major source of exported particulate phosphorus.
The major points are reiterated here.
EAA soils from the top 10 cm of ten farms averaged 510 mg/kg
phosphorus content in the study by Fiskell and Nicholson (1986).
Average phosphorus content of the canal sediments from the original
three-farm survey in this work was 617 mg/kg (Chap. 5).
Phosphorus content of the prototype sediment from UF9200 field ditch
B9B10 was in the range of 800-900 mg/kg depending on size fraction.
The soil from the adjacent field also had a phosphorus content in this
range (Chap. 5).


128
CSME, =(EM,(0)-EM,) 6.17
Setting Equation 6.16 equal to zero, substituting Equation 6.17,
eliminating EMj and collecting terms gives
(csme)
Wy- J-L
' h
= e(t t(j ^EM, (0) e(t t )CSM£(
6.18
Solving for CSME, gives
CSME] =
e(t-t,)£M,(0)
1V,
- + E|
(T-Tv)
6.19
Now to eliminate shear stress from Equation 6.19 recall that
r = PgRhS
where
6.20
t = Channel unit tractive force (Shear Stress,
dynes/cm2)
p = density of water, gms/cm3
g = gravitational acceleration, cm/sec2
Rh = Hydraulic Radius of channel, cm
S Hydraulic gradient, cm/cm
Expressing S in terms of velocity may be done by using the Manning Equation
V =
-R^V'
62\
which, when solved for S gives


18
EAA Canal Sediment Studies
Several studies, which have been conducted in private and District canals within
the EAA, have particular relevance to the area of particulate phosphorus transport. For
the most part these activities have been private, non-peer-reviewed studies that have had
limited report circulation. Nonetheless they have tended to form a basis for the
perspective of the EAA grower community toward potential phosphorus reduction
activities and require some significant attention at this point.
Anderson and Hutcheon Engineers study (1992)
Anderson et al. (1992), in a private study conducted for the Florida Sugar Cane
League reported results of some sediment core sampling conducted in the EAA's four
main and two connector canals, which are operated by the SFWMD. They also presented
data from what was reported to be a sediment transport study conducted at coded private
sites within the EAA.
The sediment core analyses showed a sediment phosphorus content in the range of
58-77 mg phosphorus/ kg dry sediment (58-77 ppm) for five of the six canals. Only the
Miami Canal sediment, which averaged 252 mg P/kg sediment, exceeded 100 ppm.
The data reported in the sediment transport study include some velocity and total
suspended solids (TSS) data in conjunction with the total phosphorus (TP), soluble or
total dissolved phosphorus (TDP), and particulate or particulate phosphorus (PP)
concentrations. Average concentrations for the total data set were TP = 0.153 mg/1, TDP
= 0.042 mg/1 (27% of total), PP = 0.111 (73% of total), and TSS = 187 mg/1.
On average the phosphorus content of the TSS was 594 mg P/kg TSS (ppm), but a
more detailed examination of the data indicates that the sample set may have represented
a bimodal distribution of samples, with one set having a phosphorus content of about 390
ppm and the other set having a phosphorus content of about 1700 ppm.


194
Time Series-Flow and Total Suspended Solids
Events UF9206S-260. 268
0.9 O
70
60
50 1
40 6
30 J
20 £
10
0
Pumping Rate O -TSS Concentration
Figure 7.29: Time Series-Flow and Total Suspended Solids Events UF9206S-260,268
Time Series-Flow and Total Suspended Solids
Events UF9206S-280. 300
Pumping Rate O TSS Concentration
Figure 7.30: Time Series-Flow and Total Suspended Solids Events UF9206S-280, 300


186
Table 7.5 illustrates the difference between the water management policies of the
two farms. UF9206 received about 33% more rain volume than UF9200 but discharged
slightly over 3 times the volume of UF9200. The attempt to keep the sod fields dry is
clearly illustrated by the 64.3% pumpoff ratio at UF9206N compared with 31.9% at
UF9206S (cane and rice) and only 18.2% at UF9200 (cane only). The increased
hydraulic load at UF9206 produced an increase in total suspended solids and particulate
phosphorus loads. The particulate phosphorus load ratio (3.28) was roughly equivalent to
the hydraulic load ratio (3.01), but the total suspended solids ratio (9.30) was extremely
elevated compared to the hydraulic load ratio. Note that the hydraulically active sod
section contributed about 67% of the TSS load and 57% of the PP load even though it
represented only about 30% of the farm area. Given what was theorized from the results
at UF9200 regarding the relationship between hydraulic activity and phosphorus content
it would be expected that the phosphorus content at UF9206 would be lower.
Repetition of the event correlation process yields Figures 7.17 and 7.18, TSS and
PP correlation with hydraulic load for UF9206N, and Figures 7.19 and 7.20, TSS and PP
correlation with hydraulic load for UF9206S. The values of interest, the slopes and ratio
of the slopes, are respectively 104.15 mg/1 TSS, 138.9 pg/1 PP, and 1320 mg/kg PP/TSS
for UF9206N, and 39.36 mg/1 TSS, 89.3 pg/1 PP. and 2268 mg/kg PP/TSS for UF9206S.
Note that the event correlated particulate phosphorus contents of 1320 mg/kg for
UF9206N and 2268 mg/kg for UF9206S are considerably lower than the value of 4562
mg/kg calculated for UF9200, where the event correlated TSS value was only 20.45 mg/1.
The value of 1320 mg/kg for UF9206N is approaching the range of 900-1150 mg/kg
which was observed in the surficial sediment studies reported in Chapter 5. The general
trends for UF9206, where the North station had a lower event correlated phosphorus
content than the South station, which in turn was lower than UF9200, support the
hypothesis of decreased phosphorus content with increased hydraulic activity. It is




15
Izuno et al. study
The work by Izuno et al. (1991) was a study carefully designed to allow
evaluation of phosphorus discharge from various crop and field conditions before and
after implementation of BMPs. The baseline (before BMP) study was reported in the
1991 publication. Replicate fields of sugarcane, fallow, flooded fallow, and selected
specific vegetables were evaluated over a 10-13 month period. Key conclusions follow.
The phosphorus concentrations in the sugarcane fields averaged 0.28 mg/1
total phosphorus, 0.17 mg/1 soluble phosphorus (61% of total) and 0.11
mg/1 particulate phosphorus (39% of total). The cabbage fields averaged
0.56 mg/1 total phosphorus, 0.31 mg/1 soluble phosphorus (55% of total)
and 0.25 mg/1 particulate phosphorus (45% of total). The radish fields
averaged 0.25 mg/1 total phosphorus, 0.20 mg/1 soluble phosphorus (80%
of total) and 0.05 mg/1 particulate phosphorus (20% of total). It was noted
that radish has a lower fertilization requirement than cabbage.
The effects of hydraulics were evident in this study. The displacement
phenomenon was noted, wherein a portion of the initial contents of the
ditch and the surface runoff must pass through the pump before the effects
of the field groundwater on soluble phosphorus concentration are seen.
Main farm canal discharges averaged 0.16 mg/1 total phosphorus, 0.08
mg/1 soluble phosphorus (50% of total) and 0.08 mg/1 particulate
phosphorus (50% of total), which was consistently lower in phosphorus
concentration than the effluent from the fields, suggesting that either
dilution or removal of phosphorus was taking place in the irrigation
conveyance systems.


D.l CRAF Approximate Model: Velocity-Depth Profile 366
D.2 CRAF Approximate Model: Velocity-Depth Profile (Dimensionless in
Velocity) 366
E.l Rainfall at UF9200 Julian Days 150-356 370
E.2 Selected Water Levels at UF9200 Julian Days 150-300 371
E.3 Selected Water Levels at UF9200 Julian Days 300-365.. 372
G. 1 Layout of UF9200 386
G.2 Layout of UF9206 387
G.3 Recycle Simulations UF9200 388
xxi


47
n2
H = bed shear stress, N/m2
id = critical depositional shear stress. N/m2
ws = settling velocity of sediment, m/sec.
Cb = concentration of suspended sediment, kg/m3
This expression becomes more complex as additional classes of particles are added and as
the effects of agglomeration and de-agglomeration are considered. One practical aspect
of these relationships is that erosion and deposition of cohesive sediments may be studied
independently of one another if the critical shear stress for deposition is less than the
critical shear stress for erosion.
Consolidation Here there are similarities between the behavior of cohesive
sediments and the presumed behavior of organic sediments. Consolidation takes place
because the forces exerted by the self weight of the settled material exceeds the strength
of the interlocking sediment structure and its contained water. In cohesive sediments the
interlocking sediment structure contains agglomerated structures as well as individual
particles. The strength of these agglomerated structures is often the controlling element
in the rate of consolidation. Organic sediments, with their potentially complex physical
structures, might be expected to exhibit characteristics similar to cohesive sediments in
the consolidation process, such as a rapid approach to an asymptotic density profile as
supporting structural networks are broken down during consolidation.


401
Rouse.H. 1938. Experiments of the Mechanics of Sediment Suspension, In
Proceedings, Fifth Congress for Applied Mechanics, J. Wiley, New York
Sanchez,C.A., and Porter.P.S. 1994 Chapter 5-Phosphorus in the Organic Soils of the
EAA In Everglades Agricultural Area (EAA)-Water, Soil, Crop, and
Environmental Management, A.B. Bottcher and F.T. lzuno, Eds., University Press
of Florida, Gainesville, FL
Saunders, W.M.H., and Williams, E.G. 1955. Observations on the Determination of
Total Organic Phosphorus in Soil, J. Soil Sci.. 6:254-267
Schaffrenek. R.W., 1987. Flow Model for Open Channel Reach or Network, Professional
Paper No. 1348, U.S. Geological Survey, Washington, D.C.
Sedell, J.R., Naiman, R.J., Cummins, K.W., Minshall, G.W., and Vannote, R.L. 1978.
Transport of Particulate Organic Matter in Streams as a Function of Physical
Processes. Vehr. Intemat. Verein. Limnol.. 20:1366-1375
Sfrisco.A., Donazzolo.R., Calvo,C., and Orio,A.A. 1991. Field Resuspension of
Sediments in the Venice Lagoon, Env. Technol. Letters. 12:371-379
Snyder.G.H. 1994. Chapter 3-Soils of the EAA In Everglades Agricultural Area
(EAA)-Water, Soil, Crop, and Environmental Management, A.B. Bottcher and
F.T. Izuno, Eds., University Press of Florida, Gainesville, FL
Svendsen, L.M., and Kronvang, B. 1993. Retention of Nitrogen and Phosphorus in a
Danish Lowland: Implications for the Export from the Watershed,
Hvdrobiologia. 251:123-135
Svendsen, L.M., Kronvang, B., Kristensen, P., and Graesbol, P. 1995. Dynamics of
Phosphorus Compounds in a Lowland River System: Importance of Retention and
Non-Point Sources, Hydrological Processes. 9:119-142
Swain, E.D., Chin, D.A., 1990. Model of Flow in Regulated Open-Channel Networks,
J. Irrig. Drain. Engng.. 116(4): 537-554
Tipping,E., Woof,C and Clarke,K. 1993. Deposition and Resuspension of Fine
Particles in a Riverine Dead Zone, Hvdrol, Processes. 7:263-277
Trewartha, G.T., and Horn, L.H. 1980. An Introduction to Climate, 5th ed., McGraw-Hill
Tsai.C.H., and Lick,W. 1986. A Portable Device for Measuring Sediment
Resuspension, J. Great Lakes Res.. 12:314-321


16
The drained fallow fields, which were interspersed with the sugarcane
fields and drained under the same conditions, showed effluent phosphorus
concentrations averaging 0.43 mg/1 total phosphorus, 0.2B mg/1 soluble
phosphorus (65% of total) and 0.15 mg/1 particulate phosphorus (35% of
total), which was significantly higher than the neighboring sugarcane
fields, suggesting a strong crop-effect on the availability of total
phosphorus for drainage. (Recall the very low pore-water phosphorus
found in the CH2M Hill sugarcane fields.)
The highest loads observed during the study came when flooded fallow
radish fields were drained, and immediately after fertilization of the
cabbage fields.
This study also showed high variability, with standard deviations of
phosphorus concentrations approaching the mean in most cases.
Canal Transport Studies
Several studies that contain data related to phosphorus transport in canals were
performed in the Everglades region.
Southeast Florida canals water quality studies
Lutz (1977) studied the variation of water quality with time over an 18 month
interval at various points within several Southeast Florida canals. One of the canals
studied was the West Palm Beach Canal from its point of exit from the EAA to a point 16
km (10 miles) downstream. His experimental procedure entailed taking individual grab
(non-flow weighted) samples at the various sample stations for each canal on the same
day every two weeks. Lutz data showed the same variability as was seen in the
previously mentioned farm/field studies, that is, standard deviations approximately equal


261
Calculated and Observed Flow Rates al
[vfajor Structures tor Event UF9200-237
1.00
Time firm Start of Punping-hr
Bridge Culvert 12-13 0
CUlvat 8-9 CUKert 4-5
S to E Canal Discharge North Canal Discharge 0
Figure 9.2: Calculated and Measured Flows, Event UF9200-237
Calculated and Observed Flow Rates at
Major Structures for Event UF9200-252
1.00
Time from Start of Pumping-hr
Brid8e Culvert 12-13 o
Culvert 8-9 Culvert4.5 .
S lo E Canal Discharge North Canal Discharge o
Figure 9.3: Calculated and Measured Flows, Event UF9200-252


100
5. Use the portable device to determine operational parameters in the field
and hopefully determine spatial and temporal variations.
The ultimate objective, of course, was to be able to translate measured or
calculated hydrodynamic parameters in the conveyance systems, specifically local shear
rates or velocities, to the rates of sediment suspension and transport in the conveyance
system.
Prototype Sediment
For the prototype sediment it was decided to continue to use field ditch B9B10 of
UF9200 as the standard source. To reiterate, this ditch was a well maintained
conveyance, relatively free of macrophytes, located approximately in the center of a
typical medium-to-large sugarcane-only farm. In anticipation of these tests a large sister
sample had been taken at the same time as the sample was taken for the studies reported
in Chapter 5. Approximately 100 kg of surficial sediment was collected using a hand
held clamshell-type dredge. This sample was taken over a distance of about 50 meters at
a location just upstream of the drainage ditch B9B10 mid-length. It was transported to
the Gainesville laboratories along with the samples used for fractionation and stored in 20
liter sealed containers under anoxic conditions at laboratory temperature. It was felt that
refrigeration was not required because the samples were essentially being kept in the
same state under which they had existed in the field, chemical analysis of these samples
was of little interest, and the samples were to serve as a prototype for calibration between
the laboratory and field erosion-simulation devices.


177
Figure 7.14: UF9200 Phosphorus Content of TSS as a Function of TSS Concentration
Farm 9200 Phosphorus Content of Suspended Solids as a Function of Suspended
Solids Concentration (Expanded Scale)
0 -l t 1 1 1 t-
0 10 20 30 40 50 60 70 80
Suspended Solids Concentration-mg/1
O First 4 Hours of Events A Beyond 4 Hours
Figure 7.15: UF9200 Phosphorus Content of TSS as a Function of TSS Concentration
(Expanded Scale)


287
It is this depletion of erodable mass that causes the maximum and the subsequent decline
of TSS concentration at each node.
Simulation of Event UF9200-237
TSS at Selected Nodes
Simulation Time-hr
Figure 10.7: Simulation of TSS at Selected Nodes for Event UF9200-237
Simulation of Event UF9200-237
CEm at Selected Nodes
Simulation Time-hr
Figure 10.8: Simulation of EM at Selected Nodes for Event UF9200-237


42
The statistical approach may be applied in several ways. First the diffusional and
energy approaches, as presented above, apply to only one specific monodisperse class of
particles. One statistical approach is to apply a statistical distribution to particle classes
and then apply the equations to each class. A more detailed approach is to write the
equations of motion as stochastic equations and then either solve them using simplifying
assumptions and asymptotic solution methods (Hinze 1957, Hino 1963) or use them to
run Monte Carlo computer simulations (Yalin 1972). Specifics will not be presented
here.
The total suspended load is obtained by integrating the equation
3.4
where qs = volume rate of suspended solids discharge per unit width,
m3/m-sec
u = velocity at elevation y, m3/sec
The details of the integration of this equation depend entirely on the choices of
concentration distribution, velocity distribution and reference datum plane. There are
many options available (Raudkivi 1976, Vanoni 1975), some of which derive from basic
principals and some of which contain empirically derived fitting parameters. The latter
tend to give better results than the former in specific situations but cannot necessarily be
extrapolated beyond their range of calibration.
There are several drawbacks in the application of non-cohesive sediment transport
theory to the study of organic sediment transport. First most of the useful formulations
assume steady state with respect to sediment suspension and deposition. Development of
transient formulations would require reversion to the equations of motion and mass
transport, where the dynamics of organic sediments have virtually no investigational


72
Table 5.4: Soil and Sediment Screen Fractionation Sequence
Screen Fraction
Particle Size Range -micrometers
Coarse
>10,000
+8 USS
10.000-2360
-8 USS+16 USS
1180-2360
-16 USS+30 USS
600-1180
-30 USS+50 USS
300-600
-50 USS+100 USS
150-300
-100 USS+200 USS
75-150
-200 USS+400 USS
38-75
-400 USS
<38
A small sub-sample of the recombined sample was taken for moisture analysis
and the rest of the sample was gently combined with 7 liters of filtered native EAA water
that had been collected at the farm side of the UF9200 pumphouse. This mixture was
gently agitated on a twice-daily basis for eight days at room temperature and then
subjected to wet screen analysis (See next section).
Sediment Fractionation The sediment wet Screening fractionation process was
much more labor intensive than the soil dry ^creening. When the sieves were stacked the
cohesive, almost gelatinous, nature of the sediment samples caused the sieving process to
stop almost immediately after the suspension completely wet the screens because of lack
of back-flow of air from the lower sieves to the higher sieves. In order to overcome this
problem the sieving process was done manually, one screen fraction at a time, starting
with the coarse 1 cm screen. Wash for the wet screening process was obtained by
recycling decanted supemate from settled suspension that had already passed through the


327
become incorporated in the base sediment, which has been shown to be much less
transportable than the surficial erodable sediment. This type of operation might be
viewed as hydraulic mining of the erodable mass.
In order for hydraulic mining to be practical it would have to move large masses
of erodable matter to the desired locations over reasonable time periods. The following
simulations provide several examples that allow qualitative judgments to be made on the
practicality of hydraulic mining and solids transport via recycle to upstream locations.
They use only the existing pumps and structures. In some cases gates are assumed on the
discharge culverts of the North Canal to prevent back flow during recycle.
In all of these simulations the original pumping pattern is assumed, including the
cycling of the large pump. No attempts are made in this series to super-impose
alternative pumping schemes other than the recycle simulation, which occurs prior to the
discharge pumping. See Appendix G, Fig. G.3, for the various recycle routings.
Simulation RCL1 Short Duration. Short Distance In this simulation the recycle
pumping is assumed to flow from the discharge of the Pump Station to a point
immediately upstream of the discharge culverts of the North Canal at 1.1 m3/sec for four
hours with the (assumed) gates closed on the North Canal discharge culverts. After four
hours the normal pumpdown cycle of Event UF9200-220 was simulated. This is the
simplest possible recycle, with recycle water being pumped only about 20 meters from
the pump discharge across the farm entrance road to the lower end of the North Canal.
Figure 11.7 shows that this recycle scheme did virtually nothing to affect ultimate mass
transport.
The solids that were transported from the East Canal upstream into the North
Canal, along with North Canal solids that were transported upstream by the reverse flow
were not available for initial transport when the discharge cycle was started so there was a
temporary reduction in solids export. The transported masses, however, were eventually
exported when the normal flow reversed their migration and ultimately brought them


98
It may be surmised that the flocculation kinetics of the observed systems were in
early stages of reaction extent so that the dominant processes exhibited pseudo-first
order kinetics, but this is purely speculative at this point. The apparently strong
flocculation mechanisms exhibited by the sediments introduced complicating factors into
the plan to develop a simple sedimentation prototype model. At this point it was
determined that, if modeling conditions demanded, a dynamic sedimentation model that
included flocculation processes might be necessary. No further studies were conducted
with the bottom withdrawal system on the sediments collected during this phase, however
the studies conducted on the B9B10 sediment gave benchmark estimates of V50 for the
<38 micrometer particle size fraction in the range of 0.50-2.50 m/day.
Sediment Solids Content and Bulk Density The solids content and bulk densities
of the <38 micrometer fractions of the WPBC and B9B10 sediments under conditions of
minimal self-weight (< 2 cm depth) were measured using the methods noted in Appendix
B. The measurements were done in five replicates. The measured and calculated
properties are shown in Table 5.9.
Table 5.9: Solids Content and Bulk Densities of WPBC and B9B10 <38 Micrometer
Fractions
WPBC <38 Micrometer
B9B10 <38 Micrometer
Fractions
Fractions
Particle Specific Gravity
2.32
1.710
Sediment % Solids
12.85%
7.65%
Content
Bulk Density gm/ml
1.079
1.033


LIST OF FIGURES
Figure Page
1.1 Original Everglades Watershed : 2
4.1 Farm Scale Phosphorus Transport Original Conceptual Model 53
4.2 Farm Scale Phosphorus Transport Simplified Conceptual Model 59
5.1 Particle Size Distribution of Dry and Flydrated B9B10 Soil 75
5.2 Particle Size Distribution of Flydrated B9B10 Soil and B9B10 Sediment 77
5.3 Particle Size Distribution of Hydrated B9B10 Soil, B9B10 Sediment,
and WPBC Sediment 78
5.4 Volatile Fraction (Organic Content) Distribution of Hydrated B9B10
Soil. B9B10 Sediment, and WPBC Sediment 79
5.5 Particle Specific Gravity Distribution of Hydrated B9B10 Soil, B9B10
Sediment, and WPBC Sediment 80
5.6 Correlation of Particle Specific Gravity with Organic Content for
Hydrated B9B10 Soil, B9B10 Sediment, and WPBC Sediment 81
5.7 Total Phosphorus (TP) Content Distribution of Hydrated B9B10 Soil,
B9B10 Sediment, and WPBC Sediment 81
5.8 Phosphorus Fractions vs. Organic Fraction for WPBC and B9B10
Sediments 83
5.9 Adsorption Isotherm-WPBC Sediment 75-150 Micrometer Range 88
5.10 Adsorption Isotherm-WPBC Sediment <38 Micrometer Range 89
5.11 Adsorption Isotherm-B9B10 Sediment 75-150 Micrometer Range 89
5.12 Adsorption Isotherm-B9B 10 Sediment <38 Micrometer Range 90
xiii


251
moving irrigation canals, and elimination of the term can speed up calculation time with
minimal sacrifice in precision. The Preissmann technique is well suited for use when
boundary and internal conditions such as weirs and gates exist.
All initial flows and levels must be specified by the user. Boundary conditions
may be specified to include fixed water levels or discharges, levels or discharges that vary
according to a schedule input by the user, discharges as a user defined function of stage,
rainfall/runoff relationships, and wind velocity and direction. Structures, primarily weirs
and culverts, are treated as internal boundary conditions with zero storage that are
governed by standard equations that relate flow to upstream and downstream head and
hydraulic characteristics of the structure.
There is no major limitation on the system configuration that can be analyzed by
DUFLOW, but the number of nodes is limited to a maximum of 448, sections to a
maximum of 300, and structures to a maximum of 148. Looping is automatically
accounted for, and multiple structures at a single location pose no problem.
Cross-sections are defined by the user at each node in terms of top width of flow
at given depths and the program interpolates between the given dimensions. A number of
depths, up to a maximum of 15, may be specified to define an irregular channel
configuration. The user may also, by separate dimension specification, define storage
volumes that do not participate in flow. These storage volumes are incorporated in the
equations of continuity but not in the equations of flow rate, velocity, or momentum.
Dimensions and hydraulic characteristics of structures must be defined by the user.
Frictional resistances for structures are expressed in terms of the Chezy equation,
frictional resistances for the channels may be expressed in terms of either the Chezy
equation or the Manning-Strickler equation (where the Strickler k, the inverse of
Mannings n, is the input parameter).


185
Similar to the UF9200 record, the events covered a broad range of discharge, from
0.59x10s m3 to 9.47x10s m3 for UF9206N and from 0.39x10s m3 to 13.20x10s m3 for
UF9206S. Suspended solids loads ranged from 798 kg to 99,494 kg at UF9206N and
from 266 kg to 48.654 kg at UF9206S. Particulate phosphorus loads ranged from 4.27 kg
to 132.42 kg at UF9206N and from 0.90 kg to 120.60 kg at UF9206S. As was the case at
UF9200, the late season tropical depressions had a dominant impact on the statistics for
the period. Both farms had similar rainfall during the period of comparison, UF9200 had
0.922 m. UF9206 had 0.896 m.
A comparison of discharge statistics for comparable periods among the three
stations (Events 218-336 for UF9206N and S and Events 220-336 for UF9200) is shown
in Table 7.5, where the totals for the two stations at UF9206 are compared to the results
from UF9200.
Table 7.5: Comparison of Pumping Events Between Farms UF9200 and UF9206 for
Julian Dates 218 through 350
UF9206N
UF9206S
Total UF9206
UF9200
Ratio
9206/9200
Area-ha
-525
-1225
1750
1280
1.37
Rain Depthm
0.896
0.896
0.896
0.922
0.97
Rain Vol.-m3
47.0x10s
109.8x10s
156.8x10s
118.0x10s
1.33
Discharge-m3
29.8x10s
35.0x10s
64.8x10s
21.5x10s
3.01
% Pump-off
63.4%
31.9%
41.3%
18.2%
2.26
TSS Loadkg
277,071
138,489
415,560
44.651
9.30
Partic. P loadkg
370.7
274.0
644.7
196.31
3.28


354
Model development
1. The current version of the model should be evaluated at more sites, in
conjunction with the improved model calibration interfaces, to develop a
statistical base for parameter ranges and to correlate the parameter ranges
with basic farm hydrologic and biological characteristics such as percent
pump-off, level variations, areal coverage of macrophytes, etc.
2. In a parallel effort studies should be implemented to expand the scope of
the model by better characterizing the transported material. This might
entail a major collection effort of exported solids to obtain sufficient
material to run extensive flocculation/sedimentation studies and erosion
studies to develop a distribution of parameters for the transported material
instead of the lumped parameters used in the current model.
3. The acquisition of parameter distribution estimates would then allow the
development of a multi-population model that would be more responsive
to the actual transport at lower flows. This is important for the
development of several of the BMP recommendations.
4. There should be in-depth studies of the interevent consolidation and
immobilization processes through erosion testing of freshly deposited
materials through various stages of consolidation. This will require
development of better sampling techniques than those used in this study in
order to insure that the light flocculant sediments are sampled in the proper
proportion and not selectively excluded. Hydraulic mining has promise in
this area but the techniques need to be developed that insure representative
sampling.


342
At hydraulic conditions that give rise to low TSS concentrations, where it is
presumed that only highly mobile fresh organic matter is transported, all three locations
have approximately the same phosphorus content of exported solids. Similarly at
hydraulic conditions that give rise to the very high concentrations of suspended solids,
which are presumed to be principally base sediment, the phosphorus content of exported
solids is again similar at all three locations. In the intermediate range, where most of the
total phosphorus load is generated, the two UF9206 locations are quite similar, but are
lower in phosphorus content than UF9200. This pattern is consistent with the hypothesis
of dilution of transportable matter by periodically mobilized base sediment at UF9206.
At this point it is also instructive to look at the impact of the operating extremes at
UF9206, because level control is such an easy practice to implement. Figure 11.18 shows
a summary, for the normal wet season, of the cumulative fractional particulate
phosphorus load of sampled periods as a function of the TSS concentration of the
sampled period, where a unit period is the two hours of pumping time programmed for
each sample.
Figure 11.18: Cumulative Fractional PP Load vs. TSS for All Three Farms


140
of this linear estimate extended back to time zero. The process is illustrated in Figure
6.17.
A total of 20 organic sediment samples were run. The first four runs were used to
develop technique and were not included in the data analysis. Of the remaining 16. three
were eliminated from consideration because of aberrant behavior. Two samples had large
amounts of structured debris protruding through the surface and eroded much more
rapidly and to a greater extent than their counterparts and one sample simply would not
erode under any conditions within the chosen test range. The 13 remaining samples were
used to develop a correlation of CSME versus cam RPM that was deemed to be a more
convenient variable than its inverse, grid oscillation period.
Figure 6. IB shows the results of the PES Organic Sediments runs. The data
were correlated by the linear regression equation
CSME = 4.709 x 1 O'5 x RPM 0.006667 628
where RPM = grid oscillation frequency, revolutions/min
There was considerable data scatter, illustrated by the r2 value of 0.454 for
Equation 6.28. This scatter, however, is not surprising considering the visual
observations made of the variable condition of the bed surface. Given this data scatter it
was deemed inappropriate to attempt to correlate the PES test results with any higher
order functions of RPM even though one would expect the dependency on grid speed to
be higher than first order.
The results from the PES-Organic Sediment tests and the CRAF-Organic
Sediment tests now give enough information to allow calibration of the PES to
parameters measured in the flume and extrapolation to parameters which may be obtained
in the field.


274
data represented a two hour composite per data point while the simulation data output
was essentially continuous. This introduced the most difficulty when a pump cycle
ended, because the corresponding sample period usually did not end simultaneously.
The calibration process for the transport model was more complex than for the
hydraulic model because there were essentially four adjustable parameters to fit across all
events. The approach taken was to jointly adjust the erosion coefficient, E, the critical
velocity for erosion, Vc, and the sedimentation velocity, wsed,across the events and adjust
Cem(O) the initial surface concentration individually for each event to obtain an optimum
fit. The values of Cemd thus developed were then examined for possible relationships
that could lead to a more general correlation.
The general calibration process proceeded as follows.
Preliminary data evaluation showed that there seemed to be a trend of
increasing exported solids with increasing interevent time so an
assumption made to start calibration was a linear relationship between
interevent time and Cem(O) the initial surface concentration of erodable
solids.
The calibration process was roughed out with the three major events of the
wet season, 220, 237, and 285. After the parameter range had been
narrowed to a reasonable level the three smaller events, 252, 258, and 262,
were added to the process. Event 244, a small event during which only the
small pump was run intermittently, was excluded ffom the process.
To start, Vc was assumed equal to zero, a trial value was assumed for
Cemio) for Event 220, the event with the longest preceding interevent time,
and the values of s and wsc end values of the U curve. Once trial values of s and wSCd were
established for Event 220, they were used in the simulation of Events 237
and 285 along with values of CEm(0) that were ratioed down from the Event


312
244 from the calibration process and Event 336 from the
validation process.
4. End-of-simulation values of concentration of erodable mass,
Cem, at the lower extremes of the canals which exceeded the
start values of the subsequent event were carried forward as the
start value, Cemioi. for the simulation of Events 237 and 252.
Calibration Results The key results of the model calibration process were:
1. The model represented the U shaped curve of exported
suspended solids concentration quite well for five of six of the
calibration events. It qualitatively expressed but did not closely
track the large suspended solids discharge from the North
Canal observed during Event 285.
2. The model did not simulate suspended solids discharge well
during the operation of the small pump, which represented a
small fraction of the total solids discharge for most events.
3. The cumulative suspended solids loads were overpredicted by
slightly more than 3%, the cumulative particulate phosphorus
loads were underpredicted by slightly less than 8% for the
combined output of the simulation periods over the six
calibration events.
4. When applied to the separate canals, the model tended to
overpredict the transport in the South-East Canal system and
underpredict the transport in the North Canal.


27
actual contributions, however, will be governed by local system dynamic parameters
including carbon and nutrient availability, light source, intensity and extinction,
competitive population dynamics and size of standing crops, air and water temperature,
and hydraulic fluctuations.
Early Diagenesis
The processes of early diagenesis, or the first steps of conversion of sediment to
inert material, are important in that the process pathways and kinetics can have a
significant effect on the nutrient content of the resuspended sediments and on the amount
of sediment-contained nutrients released back to the water column. The organic sediment
constituents which have the highest phosphorus content will also tend to be the most
bioavailable (See preceding discussion of sources). A brief discussion of the expectations
of the fate of this material follows.
Lucotte (1993) has estimated that up to 50% of the P initially buried in estuarine
sediments of the St. Lawrence River are released back to the water column because of
bioturbation.
Montigny and Prairie (1993) have shown that bacteria in sediments will lyse
rapidly in the anaerobic regime, releasing high levels of soluble phosphorus. Contrary to
conventional theory, however, the soluble phosphorus did not combine with ferric iron
and precipitate when it diffused into the oxic zone. The reason for this was the chelation
of the ferric iron with simultaneously released organic acids, which removed the iron
from effective interaction with soluble phosphorus.
Menendez et al. (1993) found that the initial decomposition of macrophyte
detritus and release of phosphorus proceeded at an extremely rapid rate, about 20%/day,
in the anaerobic zone of lagoon sediments.


303
suggesting that macrophytes in periodic motion might act as detrital phytoplankton
generators, dislodging and replacing algal masses quite rapidly.
Reddy and DeBusk (1983) provide data that allows the calculation of specific
growth rates of water lettuce and water hyacinth as a function of plant density. Again we
will assume a surface occupied 75% by water lettuce at a density of 627 gm/m2 and 25%
by water hyacinth at a density of 1535 gm/m2.
The specific growth rates associated with each of these species at the given
density is 0.06 days'1 for water lettuce and 0.025 days'1 for water hyacinth. This
translates to a weighted biomass increase rate of 37.8 gm/m2/day. Now Reddy (1988)
states that biomass grown in eutrophic conditions of Lake Apopka in Central Florida
contained approximately 65% of the plant as live tissue and 35% as senescent plant
matter and detached detritus. Applying that 35% value to the production figure of 37.8
gm/m2/day gives a potential detrital production of 13.2 gm/m2/day, close to the 14
gm/m2/day of erodable material mentioned earlier in this section.
The above arguments deal only with the macrophyte mat. Other than the 14%
sediment content, found in the prototype sediment particle fractionation tests reported in
Chapter 5, no data were gathered on concentrations or densities of the lyngbya
filamentous algae, but it is reasonable to assume that this ubiquitous plankton makes a
meaningful contribution to the erodable sediment load.
Tubea, et al. (1981) reported that the filamentous algae lyngbya exhibited
doubling times of 0.8-2.0 days, depending on nitrogen concentration, but the growth rate
was relatively independent of phosphorus concentration down to 0.003 mg/1 P. These
doubling times correspond to specific growth rates of 0.87 days'1 and 0.35 days'1
respectively, growth rates that are much higher than those discussed for the floating
macrophytes.
No attempt will be made to estimate the lyngbya contribution to erodable
sediment because of lack of data on densities and senescence rates but the high growth


214
erosion would start. An approach to this water level was never seen in the water level
records. These observations imply that erosion of surficial sediment in the field ditches
would not make any significant contribution to the suspended solids loading of most
events, if the sediments tested were representative. This holds true for both the Low
and High erosion levels.
If we accept the preceding observation at face value it implies that the bulk of the
event suspended solids load may have been coming from the canals. Consider the last
column of Table 7.7, where the areal loading is calculated on the basis of the canals. It is
possible to make a very rough estimate of the weighted areal loading that would arise
from sediments that obeyed Equations 7.7 or 7.8 by the simple approximation of canal
velocity decreasing proportionally as the number of upstream fields decreases. If we
assume for simplicity that the maximum velocity does not change in each canal and that
the velocity stays constant in the East Canal (which receives no drainage ditches) and
decreases over 15 equal increments moving up the North and South Canals we can
calculate a farm-wide average areal load weighted by canal length for both the Low and
High" erosion levels. This calculation is summarized in Table 7.8.
Table 7.8: UF9200 Weighted Areal Loads Estimated From Equations 7.7 and 7.8
Canal
Avg. CSME Calculated using
Avg. CSME Calculated using
Low Level Erosion Eqn.
High Level Erosion Eqn.
South
0.00088 gm/cm2
0.00251 gm/cm2
East
0.00162 gm/cm2
0.00341 gm/cm2
North
0.00052 gm/cm2
0.00090 gm/cm2
Weighted Average
0.00076 gm/cm2
0.00186 gm/cm2


346
Table 11.1: Load Weighted Distribution of Soluble and Particulate Phosphorus at
UF9200, UF9206N. and UF9206S
Location
Phosphorus Particulate
Phosphorus Soluble
Fraction
Fraction
UF9200
32.1%
67.9%
UF9206N
28.6%
71.4%
UF9206S
34.8%
65.2%
UF9206S tends to be skewed upward by the several extremely low canal levels incurred
at the end of the normal wet season. There was no evidence gathered during these studies
to support the contention (noted in Chapter 2) that 50-75% of the phosphorus exported
from EAA farms is in the particulate form.
Conclusions and Critique
The conclusions drawn from the experimental results of this study and the model
adaptation and application will be summarized concisely in this section. Following that
an auto-critique will summarize qualifications to the model. This will then lead directly
to a series of recommendations for control and for future work in the area of particulate
phosphorus transport in the EAA.
Conclusions
The preliminary multi-farm canal survey showed that the phosphorus contained in
farm canal sediments in the EAA is equivalent to the particulate phosphorus
exported in pumping events for 15 or more years. This phosphorus content might


Table D.2: Spreadsheet Formulas


292
UF9200-252
Event UF9200-252


103
Figure 6.2: Photograph of Counter-Rotating Annular Flume (CRAF)


332
pumping. In this simulation the cumulative solids exported at 100,000 m of discharge
was -3650 kg, the lowest of the four simulations. This compares favorably with RCL2,
which had a net export of 3960 kg and it represents a simpler operating scheme, with no
external piping or new channel as was assumed in RCL2.
Specific Effect of Reduced Velocity Throughout the Benchmark Simulation Period
In each simulation changes of velocity have been noted as having a significant
direct or indirect effect on the ultimate cumulative solids export load at the benchmark
hydraulic load of 100,000 m3. The original event delivered 100.000 m3 after
approximately 40 hours total elapsed time. This is an average flow of approximately 0.7
m3/sec.
The final simulation in this series assumes a constant discharge of 0.7 m3/sec for
40 hours, in place of large pump operation, followed by large pump cycling, followed by
small pump operation. This achieves the identical discharge but reduces maximum
velocities at all points in the system. The original channel configuration is assumed, no
modifications have been assumed to increase flow area at the canal discharges. Figure
11.11 shows the results, labeled as Constant Discharge", along with Simulation B and
RCL4, the best results of the other series.
The maintenance of a constant average discharge over the pumping period of
interest produced a simulated cumulative solids load of only about 840 kg by the
benchmark discharge point of 100,000 m3, which is less than half of the cumulative
simulated load produced by Simulation B, the best of the preceding simulations. These
results are attributable to the fact that lower maximum velocities were achieved
throughout the system.
Because there were lower maximum velocities the overall driving force for
erosion, which is proportional to the square of the excess velocity over critical, was


275
220 value of Cem(O) by the ratio of the interevent time. This process was
repeated until there was a reasonable agreement between simulated and
observed suspended solids concentrations at the extremes of the U curve
for all three events.
It quickly became evident in the calibration process that the hydraulic
system as then configured could not replicate the first leg of the U curve.
An evaluation was made of logical system configuration modifications that
might produce this leg of the curve. The original configuration specified
the channel width for flow at the ends of the North and East canals as the
actual canal width measured just upstream of the North Canal discharge
culverts, and just upstream of the pumphouse, the end of the East Canal.
Experimentation with the model showed that specification of the flow
width at the ends of the North and East Canals at 1.8 and 2.0 meters
respectively allowed the system to give reasonable approximations of both
ends of the U curve. These values may be viewed as approximations of
the culvert and pump inlet dimensions. This substitution is certainly
logical from the hydrodynamic standpoint in that it forces the model to
simulate the increase in velocity that prevails as the streamlines converge
at the culvert and the pump inlet. The hydraulic model was recalibrated
with this new configuration and the transport model calibration process
was restarted.
Once the first trial set of values for Cem(0), s and wse the three major events with the modified hydraulic configuration the
concave portion of the U curve was fit by assuming non-zero values for
Vc for all three events. Completion of this process then gave a full set of
starting parameters for the six event simulation.


130
bed resistance would be zero wall resistance, with all shear force transmitted to the bed.
Under these circumstances the walls would become hydraulically transparent, which is
hydraulically equivalent to an infinitely wide channel. For an infinitely wide channel the
hydraulic radius. Rh, equals the depth, h. Substituting this relationship into Equation
6.25, collecting terms, and rearranging gives
CSME =
where
K
I
EMj( 0)
W
hmV2K11--^-
the collected constants epgn2
6.26
The equation suggests a dependency on h2/3V2. Equation 6.26 may now be used
to make some semi-quantitative observations. First, note that as velocity increases the
first term in the denominator approaches zero for each class below the currently eroding
class, so the CSME curve approaches the sum of the erodable masses as velocity
increases. This curve would not necessarily be expected to be linear. Second, at higher
velocities the V2 term reduces the contribution of the first term in the denominator very
rapidly, so the impact of depth, h, on the correlation should be reduced as velocity
increases. Third, if a threshold velocity exists for the first (most easily eroded) class of
sediment the equation says that the velocity observed at incipient erosion is the threshold
velocity, regardless of depth. This can be shown by setting j= 1 and taking the limit of
Equation 6.26 as V approaches Vc, The result is V= VCJ when CSME approaches zero.


181
phosphorus were sourced from the surficial sediment used for the studies reported in
Chapter 5.
The front-end phosphorus loading and the U shaped transport curves of the first
pumping cycle for UF9200 are key elements that provide guidance and also must be
matched in the model development. The correlation forms given by Equations 7.3 and
7.4 will be used as the empirical bridge between transported suspended solids and
discharge particulate phosphorus concentrations in the model developed for the target
farm UF9200. It is instructive now, however, to look at the parallel findings for the back
up stations at UF9206N and UF9206S to determine if they bear any semblance to the
UF9200 results.
Particulate phosphorus event studies-Farm UF9206, North and South stations
It has been mentioned previously that the owner of farm UF9206 took an
aggressive approach to water management. This multi-crop farm had numerous internal
pumps and structures to aid in water management and made extensive use of both cane
and rice fields as equalization basins for water storage. The discharge stations each had
two diesel-driven, adjustable speed pumps, which gave the water manager considerable
latitude in discharge control. At the time of this study the farm was planted
approximately 30% to turf, 60% to cane, and 10% to rice. Toward the end of the year the
rice was harvested and the fields that were planted to rice, plus some early-harvest cane
fields, were planted to vegetables. Throughout the time of this study there was a
concerted effort to maintain relatively dry conditions in the sod fields, which had shown
themselves to be less tolerant of upward water table fluctuations than sugarcane. This
control was achieved by installing an internal booster pump at the downstream end of the
sod field drainage system and essentially dedicating the North pump station to pumpage


183
Table 7.3: Summary of Pumping Events for Station UF9206N
Event
Duration
(hrs)
Pump Time
(hrs)
Interevent
Time
(hrs)
Pumped
Volume
(m3)
Suspended
Solids Load
(kg)
Particulate
P Load
(kg)
209
85
69
Start
1.68x10s
Incomplete
Incomplete
213
233
208
27
3.28xl05
5012
14.51
224
71
71
34
1.31x10s
798
4.27
234
65
58
179
1.13x10s
4799
9.37
239
97
66
52
1.24x10s
3453
7.56
251
64
56
191
0.90x10s
6133
10.45
256
57
40
47
0.59x10s
6744
9.11
260
138
121
45
1.80x10s
7616
13.85
268
95
88
50
1.24x10s
6535
12.35
273
120
94
24
1.34x10s
18737
20.80
279
70
40
26
0.47x10s
6592
7.51
283
107
89
36
1.06x10s
14312
16.75
300
149
66
292
1.47x10s
17245
15.62
318
318
260
279
9.47x10s
99494
132.42
336
240
127
114
2.78x10s
79601
96.09


LIST OF TABLES
Table E2ge
5.1 Sediment Survey Physical Results 65
5.2 Sediment Survey Analytical Results 65
5.3 Estimates of Total Phosphorus Mass in Target Main Canal Sediments 68
5.4 Soil and Sediment Screen Fractionation Sequence 72
5.5 Phosphorus Fractionation Results for WPBC and B9B10 Sediments 83
5.6 Sorption Parameters and Coefficient of Determination for WPBC and
B9B10 Sediments 90
5.7 Summary of Phosphorus Distribution in Demonstration Example 1 93
5.8 Summary of Phosphorus Distribution in Demonstration Example 2 94
5.9 Solids Content and Bulk Densities of WPBC and B9B10 <38 Micrometer
Fractions 98
6.1 Constants for Fit of Equation 6.5 to Multiple Shear Erosion Tests 113
6.2 Constants for Fit of Equation 6.8 to Multiple Shear Deposition Tests with
Water Depth of 30.5 cm 117
6.3 Constants for Fit of Equation 6.8 to Multiple Shear Deposition Tests with
Water Depth of 15.2 cm 121
6.4 Field PES Sample Point Descriptions 150
7.1 Average Conveyance Dimensions for UF9200 161
7.2 Summary of Pumping Events for Target Farm UF9200 165
7.3 Summary of Pumping Events for Station UF9206N 183
7.4 Summary of Pumping Events for Station UF9206S 184
xi


398
Livingston. D.A., 1955. A Lightweight Piston Sampler for Lake Deposits, Ecology.
36(1): 137-139
Lucotte, M. 1993. Early Diagenesis of Phosphorus in Recent Sediments of the Estuary
and Gulf of St. Lawrence, Hvdrobiologia. 253:100-101
Lutz.J. 1977. Water Quality Characteristics of Several Southeast Florida Canals, South
Florida Water Management District Tech. Pub. No. 77-4, West Palm Beach. FL
Maclntyre.S., Lick.W., and Tsai.C.H. 1990. Variability of Entrainment of Cohesive
Sediments in Freshwater. Bioeeochemistrv. 9:187-209
Marsh, W.M. 1987. Earthscape: A Physical Geography, New York, Wiley
Mattraw.H.C., Scheidt,D.J., and Federico,A.C. 1987 Analysis of Trends in Water-Quality
Data for Water Conservation Area 3A, The Everglades, Florida, USGS WR1
Report 87-4142
Mehta,A.J. 1973. Depositional Behavior of Cohesive Sediments, PhD Dissertation,
University of Florida, Gainesville, FL
Mehta, A.J., 1984. Characterization of Cohesive Sediment Properties and Transport
Processes in Estuaries, In Estuarine Cohesive Sediment Dynamics, A.J. Mehta,
Ed., Springer-Verlag, Berlin
Mehta,A.J. 1988. Laboratory Studies on Cohesive Sediment Deposition and Erosion,
In Physical Processes in Estuaries, J. Dronkers and W. van Leussen, Eds.,
Springer-Verlag, Berlin
Mehta,A.J., Hayter,E.J., Parker,R.W., Krone,R.B., and Teeter,A.M., 1989. Cohesive
Sediment Transport. 1: Process Description, J. Flvdr. Enana. ASCE. 115(8):
1076-1093
Mehta,A.J., Lee,S.C., Li,Y., Vinzon.S.B., and Abreu,M.G., 1994. Analysis of Some
Sedimentary Properties and Erodibility Characteristics of Bottom Sediments from
the Rodman Reservoir, Florida. University of Florida Coastal Engineering Report
UFL/COEL/MP-94/03, University of Florida, Gainesville
Mehta,A.J., Parchure,T.M., Dixit,J.G., and Ariathuria,R., 1981. Resuspension Potential
of Deposited Cohesive Sediment Beds, In Estuarine Comparisons V.S.
Kennedy, Ed., Academic Press, New York
Mehta,A. J., and Parthenaides.E. 1975. An Investigation of the Depositional Properties of
Flocculated Sediments. J. Hydraulic Res.. 13:361-381


305
319 and 355 entailed prolonged operation of the large pump, and in the case of Event
319, operation of both large pumps for some periods.
By contrast Event 336 consisted of over 20 on-off cycles of the small pump,
interspersed with one short (7 hour) run of the large pump. Following the convention set
in the calibration process, and considering the magnitude of operation of the other two
events, Event 336 was excluded from calculation of interevent time for correlation
purposes.
The interevent times used for the validation simulations and the corresponding
correlation value of Cem calculated from Equation 10.2 are shown in Table 10.3.
Table 10.3: Interevent Time and Values of Cem Used for Validation Simulation
Event
Interevent Time-days
CEMiorgmhn
319
32.00
872
355
30.42
820
The results of the simulations of the two late season events are shown in Figures
10.24-10.29. The qualitative agreement of the suspended solids profile with time was
good in both cases. The U shaped curve was reproduced in approximately the correct
proportions although there was some overshoot in Event 319 and some time lead in Event
355.
The ultimate suspended solids loads were reproduced quite well even though the
aforementioned shifts caused some deviation between the interim observed and simulated
cumulative values. For the simulation periods of the two events the combined suspended
solids load observed was 24,119 kg compared to a simulation value of 24,921 kg, a 3.3%


242
It may be informative to compare the distributions of these samples with the same
fractions analyzed in the prototype sediment taken from UF9200 Field Ditch B9B10 and
reported in Chapter 5, Table 5.5. This is summarized in Table 8.7, where it may be seen
that the percentage distribution of the Large-Scale-Composite samples was much lower in
refractory phosphorus and acid hydrolyzable phosphorus and much higher in labile
bicarbonate extractable phosphorus.
Table 8.7: Comparison Between Large-Scale Composite Samples and Prototype Surficial
Sediment for Phosphorus Content Distribution
I! Source
Labile Bicarbonate
Extractable Phosphorus
% of Total
Hydrochloric Acid
Extractable Phosphorus
% of Total
Refractory Phosphorus
% of Total
L.S.Comp.-UF9206N
28.05
29.95
41.99
L.S.Comp.-UF9200
30.61
27.51
41.88
B9B10, <38p
6.46
36.33
57.21
B9B10, 75n-150u
8.85
38.03
53.12
Recall that the prototype surficial sediment from Field Ditch B9B10 lost less than
3% of its total phosphorus content during processing prior to analysis so it was well
representative of surficial sediment in its original field condition. The large difference
between the labile fraction content of the Large-Scale Composite samples and the
prototype surficial sediment is one more strong indication of contribution to discharge
particulate phosphorus from sources other than the surficial sediment.


131
Thus the qualitative predictions are that correlation of CSME with h2,3V2 should
give a family of curves that have an X axis intercept which varies by h233 and all
approach the same limiting curve asymptotically as V increases.
Figure 6.13 shows the data plotted in the form of CSME vs. h2/3V2. For this
representation the lowest two data points in the 15.2 cm depth run were assumed to be
below the critical threshold and whatever small erosion was observed was assumed to be
a result of the debris generated by the cumulative shear stresses applied to the sediment
bed over the two month test period.
Figure 6.13: Asymptotic CSME vs.. h2/3V2 for Prototype Organic Sediment Erosion
Extended Run with 30.5 cm and 15.2 cm Water Depth
The solid line in Figure 6.13 is an empirical correlation fitting all the represented
data points with one curve, which is


245
3. Determine the nature, frequency, and quantitative contribution of overland
flow erosion to particulate phosphorus in the water conveyance system.
4. Develop a description of the conveyance system sediment with respect to
quantities, locations, chemical/physical characteristics,
erosion/sedimentation characteristics, and temporal variations that is
adequate to allow modeling of sediment transport and pore water release
within the system.
5. Determine the adsorption/desorption characteristics of the system
sediment and incorporate this process into the model where appropriate.
6. Determine the biological processes in the system that contribute
significantly to phosphorus transport and develop a first approximation
lumped parameter model of these processes.
7. Develop, calibrate, and verify the model on the target study site.
This chapter is organized into four sections. First, overland flow and
sorption (Items 3 and 5) are discussed briefly with the objective of presenting the
rationale for their exclusion. Second, the adapted hydraulic model platform and
its quality module (Items 1 and 2) are discussed, along with the hydraulic
calibration of the model. Third, the impact of the findings suggesting that
biological growth strongly affects particulate phosphorus transport is discussed
(Items 4 and 6) and the decision to alter the method of particulate model
development is justified. Finally, the transport model format and the revised
calibration methodology is presented (Item 7). The results of the event simulation
for UF9200 are presented and discussed in Chapter 10.


28
Keizer et al. (1993) found that in peaty sediments of the Netherlands the
conversion of calcium carbonate to calcium phosphate complexes was inhibited to the
point of insignificance.
These studies were not necessarily under conditions identical to those found in the
EAA watersheds, but they do tend to indicate that, unless significant quantities of
sorptive clays are present or calcium-phosphorus complexation is favored, high levels of
sediment phosphorus recycle to the water column would be expected. This is a good
place to note that Svendsen et al.( 1995) observed that the sediments in their studies
showed a long-term net retention of phosphorus of only one-half to one-third of their
short-term gross retention rates.
Erosion and Transport of Organic Material
General erosion and transport
To state the obvious, transport of particulate phosphorus within a system is
governed by the overall transport of particulate matter (PM or seston) within that system.
As noted above, sources of PM in a waterbody may include soil and ground cover
erosion, deposition from external vegetation, and internal generation and transformation.
Once deposited or generated the PM is constantly subjected to buoyancy, gravitational,
kinetic, and chemical forces, the integrated interaction of which determine the fate of the
particle.
Factors which affect the exchange of particles between their streambed or
vegetation-attached locations and the water column have a large impact on particle
transport by streams. Initiation of motion occurs when water velocity is sufficient to
create local shear stress in excess of some critical value for detachment or incipient
motion. Once moving, some particles may remain in contact with, or in close proximity


230
small disturbances could cause the release of what appeared by visual judgment to be
significant quantities of floe. Extreme care was necessary when harvesting individual
plants from clusters and a number of samples were discarded because the judgment was
that they had received excessive agitation in the sampling process. All samples were
taken from a region spanning the lower 400 meters of the South Canal and the upper 400
meters of the East Canal.
Once the samples were harvested they were transported to the EREC Station
laboratory, where they were each subjected to agitation by shaking the plant vigorously
ten times in its native water in its own container. After the agitation process the plant and
its water were separated. The plant was drained, dried, weighed, and milled in a manner
similar to that used for the Areal Density Studies. The container water was settled
overnight at 4C and decanted off. The bottom 3 cm of water and settled solids were
washed into an Imhoff cone and again concentrated overnight at 4C. The concentrated
solids were removed from the cone bottom, dried to constant weight at 103C, and milled
in a mortar and pestle.
The conditions of the test were certainly more stressful than the typical shear
stresses that would be expected to be applied during pumping, but they served to help
define an upper limit of dislodgable material. The sampling process itself was not
necessarily random so the representativeness of the results may be somewhat
questionable but the qualitative trends were clear.
The results, shown in Table 8.5, indicate a considerable amount of the total plant
mass was dislodgable under the conditions of the test. The samples taken from clusters
had an average of almost 38% of their total dry mass as dislodgable detritus under the test
conditions, while the free-floating samples averaged almost 30% of their total dry mass as
dislodgable detritus under the test conditions. The average size of the plants sampled
from the clusters was greater than the average size of the free-floating samples. This may
have been due to sampling bias, in that small plants were difficult to sample from the


191
Time Series-Flow and Total Suspended Solids
Events F9206N-251. 256, 260
0.9
0.8
0.7
0.6
0.3
0.2
600
500
400
300
200
100
0
-Pumping Rate
-TSS Concentration
Figure 7.23: Time Series-Flow and Total Suspended Solids Events UF9206N-251, 256,
260
Time Series-Flow and Total Suspended Solids
Events UF9206N-268. 273. 279
Figure 7.24: Time Series-Flow and Total Suspended Solids Events UF9206N-268,273,
279


344
Water Elevation Control Example for UF9206N
210 220 230 240 250 260 270 280 290 300
Julian Date
" 1 UF9206N Proposed Control Level
Figure 11.19: Water Level Control Example for UF9206N
Water Elevation Control Example for UF9206S
Julian Date
" 1 UF9206 S Proposed Control Level
Figure 11.20: Water Level Control Example for UF9206S


210
Figure 7.44: UF9200 Intensive Study-Event 285 South-East Canal Sample Locations,
First 32 Hours
internal amplification and/or attenuation in some cases, but the continuity of the pattern is
very recognizable. In Event 252, Figure 7.38, Location 2 maintained a low concentration
throughout the event, while Locations 3 and 4 steadily increased in similar patterns, with
Location 3 showing a more rapid increase than Location 4.
It takes a good bit of imagination to read a wave-like pattern into the data set from Event
252, but Event 258. Figure 7.42, illustrates this pattern much more clearly. Here the
samples from Location 2 were lost because of analytical error but Location 2.1 shows a
pattern similar to location 2 in the preceding data set, starting out low and never rising
above about 10 mg/1. There is considerable amplification between location 2.1 and
Location 3. which is 200 meters downstream, followed by attenuation at Location 4
which is another 585 meters downstream. Recall from Figure 7.39 that the North Canal
was contributing a high level of suspended solids, which undoubtedly affects the
discharge concentration, so it is difficult to extend the trend at Location 4 to the


146
I
Schematic of Superposition of Sigmoidal Velocity Profile with Extended Logarithmic Profile
Figure 6.20: Schematic Representation of Superposition of Extended Logarithmic
Velocity Profile on CRAF Sigmoidal Velocity Profile
Equation 6.31 may also be used to calculate the ratio of the pseudo-mean velocity
at depth 0.368d to the ring-channel AV. If the values of to measured by Mehta using the
false bottom are used the ratios for the two profiles presented by Mehta are 0.43 and 0.37,
for an average of 0.40.
Now if the roughness of 0.64 cm is used in the approximate model in Appendix D
the values of the ratio are 0.390 and 0.392 for the shear stresses that corresponded to 0.43
and 0.37 respectively in the calculation using Mehtas data and equations. The average of
0.391 compared to the average of 0.40 for the ratios calculated by the two separate
models provides some support for their correspondence. Now the approximate model can
be used to estimate the response of increasing the surface roughness from that of a clay
surface to what might be expected of an organic sediment surface. A representative
roughness was chosen (k=4.65 cm) which corresponded to a Mannings n of 0.025.
This is the value of n recommended as the minimum for earth channels with grass and


232
Here the detrital phosphorus content was still much higher than what had consistently
been observed for the clear water surficial sediment, and was in line with the phosphorus
content of the canal surficial sediment samples taken in regions of high macrophyte
density, shown in Table 8.2.
The average total plant phosphorus content, calculated by weighting the average
phosphorus contents of the two fractions, for the cluster samples was 4690 mg/kg, which
is lower than the average of 5687 mg/kg for the water lettuce samples (see Table 8.4). It
is interesting to note, however, that the phosphorus content of the remaining plant
material in the cluster samples of the dislodgable detritus studies, 5882 mg/kg, compared
well with the average of 5687 mg/kg for the water lettuce areal studies, possibly
indicating that harvesting the plants dislodged most of the lower phosphorus-content
detritus. The harvesting operation was much less vigorous than the detrital detachment
agitation, consisting simply of lifting the plants from the water into plastic bags. The
implication here might be that most of the detritus could be dislodged under conditions of
much less stress than occurred in the detachment tests.
Later in the season an abbreviated test was done on some additional samples.
During the Julian Date 300 canal sampling six samples of water lettuce were taken from
diverse locations around the farm, rather than the 800 meter reach of the South-East
Canal where the first samples were collected. These samples were deliberately skewed in
that only free-floating water lettuce samples were collected. The samples were subjected
to the same dislodgment process as was practiced previously, but this time the only
analysis done was for phosphorus content of the detritus.
The results were an average detrital phosphorus content of 4227 mg/kg, maximum
of 6517 mg/kg, minimum of 2864 mg/kg, with a standard deviation of 1421 mg/kg (CV
of 33.6%). The average of 4227 mg/kg was markedly higher than the previous average
for free floating samples of 2640 mg/kg. In fact the minimum of 2864 mg/kg was higher
than the maximum of 2639 mg/kg observed in the previous free-floating set. These


221
The field ditch sample results came as a surprise at the time they were obtained.
The population as a whole had an average phosphorus content of only 507 mg/kg. with a
standard deviation of 162 mg/kg (Coefficient of Variation of 32%). With such a high CV
any apparent trends were statistically insignificant, but more important the low average
value of phosphorus content of 507 mg/kg and the low maximum of 890 mg/kg in the
surficial sediment was surprising considering that the last ten hours of discharge samples
taken toward the end of the preceding pumping event averaged 8620 mg/kg phosphorus
content and the first ten hours of samples of the subsequent event averaged 4250 mg/kg.
The canal samples show the same general results with some interesting
exceptions. The in-stream samples showed no particular directional trend, although the
sample from the West end of the North Canal had a phosphorus content much higher than
the remainder of the in-stream samples (Significant at the 99% confidence level). The
spaced samples near the pump stations also showed no clear indication of a gradient but
the sample set contained two samples, one at the pump suction and one 300 m upstream
in the East Canal, that were considerably higher than the remainder. The two high
samples had an average phosphorus content of 2063 mg/kg compared to an average of
540 mg/kg for the remaining samples. These two populations were also statistically
different at the 99% confidence level.
The three samples in this canal sediment set had an important characteristic in
common. The West-End North Canal sample was taken in a region at the upper end of
the canal that had become populated with emergent weeds and some floating macrophytes
early in the season. The flow toward the pump station tends to carry floating
macrophytes along with it. When pumping stops the macrophytes can be driven in the
opposite direction by wind currents, but if the macrophytes bridge across the canal they
tend to resist wind-driven locomotion. The region around the pump intake was
completely bridged with macrophytes at the time of sampling. At a distance of around
280 m upstream from the pump station in the East Canal there is a slight eastward jog.


322
which started empty, to reach erodable mass concentrations in the range of 1000-1200
gm/m2. This simulated performance provides some insight into the possibly limited value
of sediment traps in dealing with vertically mobile sediments. The simulation says that
initially the trap performs its intended function because the material that was removed to
create the trap is not available for export. As the event proceeds, however, the trap
sections receive material from upstream, which is their function, but because the section
velocity is not appreciably altered this material is immediately available for transport.
The amount of transportable mass discharged from the farm is a small fraction of
the mass available for transport from the entire farm conveyance system. Because of this
there is an ample supply of upstream mass to replace what was removed to create the trap,
so the trap rapidly becomes a net donor of erodable material. Within the context of the
model the key to this performance is the channel mean velocity. If the velocity is not
altered appreciably then the same erosive forces will act on the material in the trap region
as act in the rest of the conveyance system and the material transported into the trap
region will tend to move out once the concentration of erodable mass in the trap region
becomes sufficient to make the trap a net donor. In the example simulation this appeared
to happen within the time frame of the first major pumping cycle, about 14 hours. It is
interesting to note that the final concentrations of erodable mass in the trap sections do
not differ dramatically from the final concentrations of the same sections in the original
simulation (not shown), which may imply that a trap of the size simulated may be
effective for only one large-size event.
Effect of Pumping Modification (Simulation A1
The second order effects of velocity in the model imply that pumping practices
that reduce maximum velocities while maintaining the same overall pumping rates should
reduce overall erodable mass export. These practices are simulated in two ways in this


89
fractions. The highest value of 197.9 1/kg was exhibited by the WPBC <38n fraction,
Figure 5.10: Adsorption Isotherm-WPBC Sediment <38 Micrometer Range
Figure 5.11: Adsorption Isotherm-B9B 10 Sediment 75-150 Micrometer Range


Introduction 24
Sources of Particulate Phosphorus 24
Early Diagenesis 27
Erosion and Transport of Organic Material 28
General erosion and transport 28
Hydrologic approach to seston transport 37
A Brief Discussion of Sediment Transport Theory 39
Non-Cohesive sediment transport 39
Cohesive sediment transport 43
Particle entrainment simulator 48
4 INITIAL HYPOTHESES, OBJECTIVES, AND RESEARCH PLANS 51
Problem Overview 51
_ Summary of sources 51
Original conceptual model 53
Scope of This Research 55
Minimum criteria for experimental and modeling efforts 55
Application of minimum criteria and resulting conceptual model
simplification 56
Research goals for this study 58
General Research Plan 60
5 SEDIMENT SURVEY AND PHYSICAL-CHEMICAL
CHARACTERIZATIONS OF SELECTED SEDIMENTS 62
Sediment Survey 62
Selection of representative farms 62
Sediment survey measurement methods 64
Sediment survey results 65
Sediment survey discussion 66
Selection of Primary Target Farm 69
Particle-Size Property Distribution Study 69
Sources for particle size fractionation study 70
Soil and sediment particle size fractionation 71
Fraction analysis 73
Particle size distribution 74
Volatile matter content 77
Particle specific gravity 79
Total phosphorus content 80
Phosphorus fractionation results 82
Adsorption-Desorption experimental methods 84
Adsorption-Desorption data reduction 85
Potential environmental significance of adsorption-desorption 91
Sedimentation parameters of B9B10 sediment 95
vii


173
Time Series-Flow and Total Suspended Solids
Event UF9200-3I9
Figure 7.11: Hydraulic and Total Suspended Solids Profiles for Event UF9200-319
Time Series-Flow and Total Suspended Solids
Event UF9200-336
Julian Date
100
90
70
60
50
40
30
20
10
0
352
Pumping Rate O TSS Concentration
Figure 7.12: Hydraulic and Total Suspended Solids Profiles for Event UF9200-336


113
Each response curve may be correlated well with its individual decaying
exponential equation of the form
CSME = a(l-e'fa) + CSME0 6.5
where
t is time from start of application of current shear stress, hr
CSMEo is CSME at the start of application of current shear stress, gm/cm2
and a and k are constants, listed in Table 6.1
Table 6.1: Constants for Fit of Equation 6.5 to Multiple Shear Erosion Tests
Tau -
dynes/cm2
a gm/cm2
k -hr'1
ak
gm/cm2-hr
CSMEo -
gm/cm2
Asymptotic
CSME -
gm/cm2
r2
0.96
0.00424
0.266
0.00113
0.00143
0.00567
0.968
1.30
0.00363
0.806
0.00292
0.00567
0.00930
0.946
1.64
0.00265
0.541
0.00143
0.00930
0.01195
0.950
2.00
0.00381
0.207
0.00079
0.01195
0.01576
0.982
Two characteristics of the erosion curves that are of particular interest are the
initial erosion rate when the new shear stress is first applied, which represents the
maximum erosion rate attainable at any given shear stress, and the asymptotic CSME,
which represents the ultimate amount of material transportable at the given shear stress,
or the steady-state balance between erosion and deposition at that shear stress.
Differentiating Equation 6.5 with respect to time, 1, and evaluating the derivative
at 1=0 gives the equation for the initial rate which is simply


65
Sediment survey results
The physical results of the sediment survey are shown in Table 5.1, the analytical
results appear in Table 5.2.
Table 5.1: Sediment Survey Physical Results
Location
Total Main
Canal Length
(m)
Farm Area
(ha)
Average
Sediment Depth
(m)
Estimated Total
Sediment
Volume
(m1)
Estimated Unit
Sediment
Volume
(m3/ha)
UF9200
6437
518
0.79
39277
75.8
UF9202
1609
130
0.65
5486
42.2
UF9206
11715
710
0.54
40368
56.85
Table 5.2: Sediment Survey Analytical Results
Location
Core
Length
(cm)
% Dry
Solids
% Ash
(Dry)
% Volatile
Solids
(Dry)
Pore
Water pH
Wet Bulk
Density
(gm/ml)
Solids
Specific
Gravity
Pore
Water SP
(mg/1)
Solids TP
Content
mg/kg
UF9200
#1
60.7
14.77
52.5
47.5
7.21
1.088
1.596
0.35
787
UF9200
#2
52.8
18.09
46.1
53.9
7.16
1.082
1.453
0.20
868
UF9202
#1
55.1
18.61
53.4
46.6
7.12
1.094
1.505
0.03
613
UF9202
#2-Top
41.7
18.44
43.1
56.9
7.16
1.058
1.315
0.56
572
UF9202
#2-Bott.
33.0
45.99
77.7
22.3
7.06
1.378
1.822
0.02
206
UF9206
#1
37.1
25.16
47.9
52.1
7.02
1.114
1.453
0.07
415
UF9206
#2
59.4
22.40
42.7
57.3
7.05
1.100
1.446
0.17
445


362
At time zero a solids suspension is introduced into the tube up to the 1 meter level and
uniformly dispersed by multiple inversion. The tube is then anchored in a vertical
position and samples are removed at timed intervals. The sample removal is conducted
so as to reduce the level of the water surface by 100 cm with each sample. The samples
are analyzed for total suspended solids.
The calculation procedure (See referenced method or Vanoni, 1975) involves
developing a table which ultimately yields an estimate of the time required to settle 1000
cm for each sample fraction. This information is plotted in terms of time to settle versus
percent of original material still in suspension. This is known as an Oden Curve. The
intersection of a tangent drawn to this curve at any particular X value of time with the
Y axis gives the percent of original material present at time zero which had a settling
velocity less than the chosen time. A complete derivation of this procedure is presented
in Vanoni (1975). This procedure can be used to develop a frequency distribution of
settling velocities for a collection of particles.


237
Mode)
Mode)


143
for hydraulically smooth (Rew <3.3) flow
6.30
where
Re =
Wall Reynolds number =
y = distance from surface, cm
V = Velocity at distance y from surface
to = surface shear stress, dynes/cm2,
p = water mass density, gm/cm3
k = surface roughness, cm
v = kinematic viscosity, cm2/sec
In the CRAF, however, the velocity profile is not parabolic. Because of the
counter-rotation aspect of the flumes operation and the shear dissipation of the flume
walls, the velocity profile with respect to the bed was shown by Mehta (1973) to be
similar to a vertically elongated sigmoid, with a large region of homogeneous turbulence
covering about 70-75% of the flow depth where the velocity profile is almost flat. Figure
6.19 shows an example of the velocity profiles measured by Mehta at a water depth of
22.9 cm for several ring-channel AVs. There was no sediment bed present when these
measurements were made.
The data presented by Mehta suggests that the velocity of this flat profile relative
to the bed is on the order of 34-36% of the ring-channel AV. Mehta mentioned that this is
different from the 50% value that would be expected in turbulent Couette flows and
attributed the difference to the effect of the walls. He also showed that the velocity


380
BS2 4.1120 4.1120 12.435 13.048
SECT 51 51 51 52 792
W 180.0 0.0
H 0.0000 1.3440
BS 3.8620 3.8620
SECT 53 53 53 59 201
W 270.0 0.0
H 0.0000.33800 1.1060
BS1 3.6570 3.6570 6.7140
BS2 5.4100 5.4100 9.8960
SECT 54 54 54 55 780
W 180.0 0.0
H 0.0000 1.5450
BS 2.7740 2.7740
SECT 56 56 56 62 201
W 270.0 0.0
H 0.0000 .36600 .72500 1.1340
BS1 4.1120 4.1120 12.435 13.048
BS2 4.2670 4.2670 13.032 13.548
SECT 57 57 57 58 792
W 180.0 0.0
H 0.0000 1.4690
BS 3.5570 3.5570
SECT 59 59 59 65 201
W 270.0 0.0
H 0.0000.41500 1.1380
BS1 5.4100 5.4100 9.8960
BS2 5.5870 5.5870 9.1190
SECT 60 60 60 61 780
W 180.0 0.0
H 0.0000 1.5090
BS 2.8958 2.8958
SECT 62 62 62 68 201
W 270.0 0.0
H 0.0000 .36000 .73500 1.1280
BS1 4.2670 4.2670 13.023 13.548
BS2 3.4680 3.4680 13.590 13.654
SECT 63 63 63 64 792
W 180.0 0.0
H 0.0000 1.1640
BS 2.9960 2.9960
SECT 65 65 65 71 201
W 270.0 0.0
H 0.0000.56700 1.3350
BS1 5.5870 5.5870 9.1190
BS2 5.6900 5.6900 9.0580
SECT 66 66 66 67
W 180.0 0.0
H 0.0000 1.4720
BS 3.0230 3.0230
1.70 1.70 17.00 17.00
1.94 1.87 17.00 17.00
1.50 1.50 17.00 17.00
1.91 1.92 17.00 17.00
1.58 1.58 17.00 17.00
1.87 1.71 17.00 17.00
1.54 1.54 17.00 17.00
1.92 1.93 17.00 17.00
1.88 1.88 17.00 17.00
1.71 1.75 17.00 17.00
780 1.58 1.58 17.00 17.00


134
Figure 6.15: Photograph of Particle Entrainment Simulator (PES)


351
primary pump-down cycle. This problem was circumvented by artificial
adjustment of rainfall timing in the calibration process, but that is not an
alternative if the model is to be used in a predictive mode.
3. The model does not yet specifically identify the erodable and transportable
population so it is not yet tied to measurable erosion and sedimentation
parameters that may be determined in the laboratory or the field.
4. The model parameters are determined by calibration so until there is
evidence to the contrary they must be considered to be site specific. The
particulate phosphorus-total suspended solids correlations were shown to
be site specific, or at least hydrologic-activity specific, between the target
farm and the back-up farm.
5. The site specificity may scale down to sub-system specificity when there is
a significant difference in hydrologic activity between the subsystems.
This was shown by the system-wide models tendency to overpredict the
South-East Canal system suspended solids and underpredict the North
Canal suspended solids at the target farm.
6. The model does not handle small flows well, underpredicting solids export
at low flow conditions. This was not a problem in the calibration process
but it eliminated one extended pumping event from analysis in the
validation process. In the study of the impact of significant velocity
reductions on particulate export the model as currently constituted would
probably give optimistic predictions.
7. The interevent solids generation and consolidation processes are currently
represented by an empirical growth function that is undoubtedly system
dependent and probably season dependent. This limits the ability to
forecast what happens to non-transported solids that may arise from
modified management practices.


150
Table 6.4: Field PES Sample Point Descriptions
Sample ID
Description and Comments
UF9200-1
Floating macrophytes and lyngbya present at sample point, some weeds
near bank
UF9200-2
Clear surface, no weeds, no floating macrophytes, some lyngbya present
UF9200-3
Clear surface, no weeds, no floating macrophytes, substantial lyngbya
present. Sample appeared to consist almost completely of Ivngbva.
UF9200-4
Clear surface, no weeds, no floating macrophytes, some lyngbya present
UF9200-5
Clear surface, no weeds, no floating macrophytes, no lyngbya
UF9200-6
Clear surface, some bank weeds, no floating macrophytes, some lyngbya
present
UF9200-7
Clear surface, no weeds, no floating macrophytes, some lyngbya present
UF9200-8
Clear surface, some bank weeds, no floating macrophytes, no lyngbya
UF9206-1
Cane field-heavv rooted weeds across channel, no Ivngbva. no floating
macrophytes. Some weed stalk protrusions from sample surface
UF9206-2
Sod fleld-Clear surface, no weeds, no floating macrophytes, no Ivngbva
UF9206-3
Cane field-FIoatins macrophvtes about 10 meters downstream, no weeds
no lyngbya
UF9206-4
Cane field-weeds about 2/3 of width into channel, no floating macrophvtes,
no lyngbya
UF9206-5
Cane field-floating macrophytes at sample point, no weeds, no lyngbya
UF9206-6
Sod field-Clear surface, no weeds, no floating macrophytes, no lyngbya
UF9206-7
Cane field-weeds at bank, no floating macrophytes, trace of lyngbya
UF9206-8
Cane field-weeds at bank, no floating macrophytes, no lyngbya


361
water so to avoid the drying and rehydration step which could have introduced some error
the procedure was modified to incorporate the native water with a correction introduced
for the dissolved solids content of the native water.
Sediment samples were diluted with native water to a total solids content of 3-4%
and well mixed. A subsample was taken and filtered (Whatman 934AH filter) to remove
suspended solids. Remaining (dissolved) solids content was determined on the filtrate by
APHA Method 2540B. The remainder of the sample was used to fill a pre-weighed 100
ml volumetric flask to the calibration mark. The filled flask was weighed, then the
contents were removed and three subsamples were taken. Total solids were determined
by APHA Method 2540B and the three results were averaged. In the calculations it was
assumed that the dissolved solids did not appreciably alter the volume of 1 gm. of water,
which was assumed to be 1 ml.
Calculation:
(D-C)f4l + B)-Bl
Specific Gravity = B.6
lOO-(D-CXl-rl)
where:
A = total solids content of sediment, gm TS/gm sediment
B = dissolved solids content of water, expressed as gm TS/gm
water
C = mass container, gm
D = mass container + sample, gm
Particle Sedimentation Velocity Distribution (USBR, 1989, Method 5345-89)
This method uses the bottom withdrawal sedimentation technique described in the
referenced Method. The bottom withdrawal sedimentation tube is constructed of 25.4
mm inside diameter glass tubing tapered at the lower end in a 60 constriction to a 7 mm
inside diameter nozzle, which is fitted with rubber tubing and a pinch clamp. The tube is
calibrated by height up to a height of I meter, with a 200 mm freeboard above that level.


383
BS1 7.0860 7.0860 10.400
BS2 7.0900 7.0900 10.400
BB1 7.0900 7.0900 10.400
BB2 9.6300 9.6300 11.020
SECT 100 100 100 101 94 1.39 1.33 17.00 17.00
W 180.0 0.0
H 0.0000.89000 1.5970
BS1 7.0900 7.0900 10.400
BS2 2.0000 2.0000 10.400
BB1 10.190 10.190 13.080
BB2 10.740 10.740 15.140
STRU
2
302
2
3 11
2.26
0.48 :
2.87
MU
1.000
1.000
0.830
0.830
27.0
STRU
4
304
3
4 11
2.06
0.48 :
2.67
MU
1.000
1.000
0.830
0.830
27.0
STRU
5
305
5
6 11
2.01
0.36 :
2.47
MU
1.000
1.000
0.830
0.830
0.0
STRU
8
308
8
9 11
1.81
0.48 :
2.42
MU
1.000
1.000
0.830
0.830
27.0
STRU
10
310
9
10 1
1 1.86
0.48
2.47
MU
1.000
1.000
0.830
0.830
27.0
STRU
11
311
11
12
11 2.11
0.36
2.57
MU
1.000
1.000
0.830
0.830
27.0
STRU
14
314
14
15
11 2.06
0.48
2.67
MU
1.000
1.000
0.830
0.830
27.0
STRU
16
316
15
16
11 2.21
0.48
2.82
MU
1.000
1.000
0.830
0.830
27.0
STRU
20
320
20
21
11 2.21
0.48
2.82
MU
1.000
1.000
0.830
0.830
27.0
STRU
21
321
21
25
11 2.01
0.92
2.82
MU
1.000
1.000
0.830
0.830
27.0
STRU
22
322
21
22
11 2.16
0.48
2.77
MU
1.000
1.000
0.830
0.830
27.0
STRU
23
323
23
24
11 2.06
0.36
2.52
MU
1.000
1.000
0.830
0.830
0.0
STRU
27
327
27
28
11 2.08
0.48
2.69
MU
1.000
1.000
0.830
0.830
27.0
STRU
29
329
28
29
11 2.01
0.48
2.62
MU
1.000
1.000
0.830
0.830
27.0
STRU
33
333
33
34
II 2.01
0.48
2.62
MU
1.000
1.000
0.830
0.830
27.0
STRU
35
335
34
35
11 2.47
0.48
3.08
MU
1.000
1.000
0.830
0.830
27.0
STRU
39
339
39
40
11 1.73
0.48
2.34
MU
1.000
1.000
0.830
0.830
27.0
STRU
41
341
40
41 1
11 2.41
0.48
3.02
MU
1.000
1.000
0.830
0.830
27.0
STRU
42
342
42
43 !
11 1.96
0.48
2.57
MU
1.000
1.000
0.830
0.830
27.0
STRU
45
345
45
46 1
11 1.88
0.48
2.49


315
2. Location of Control Actions The field studies and the model show that the
bulk of the exported material came from the lower ends of the canal
systems. This should allow control actions to be prioritized for specific
areas of the farms to maximize effectiveness at reasonable costs of. for
example, mechanical removal. It is also suggested, subject to the model
limitations noted above, that actions at the field ditch level may be less
productive than has been presumed to date.
3. Sediment Traps The critical velocity for erosion and interevent build-up
and the wave-like progression of erodable mass downstream during
events, all of which were successfully incorporated in the model, suggest
that sediment traps may not be particularly effective, or at best may be
effective for only a few events. Unless the sediment trap includes a
crossectional area change that reduces stream velocity significantly it will,
according to the model, come to equilibrium with the rest of the channel
bed over a matter of several days. Thus, even though the sediment trap
fills with sediment, it will not be preventing this sediment from
resuspending unless its design includes a decrease in stream velocity to
near critical.
4. Sedimentation Control As opposed to sedimentation traps, sedimentation
control implies using the presumed sedimentation and critical velocity
characteristics of the erodable mass to positive advantage. For example,
solids could be transported by internal hydraulic manipulation from areas
of high drainage velocity to areas of low drainage velocity, where they
could remain relatively quiescent until, through diagenesis, they
consolidate and become part of the less transportable base sediment.


252
DUFLOW water quality module
The DUFLOW water quality module is based on the one-dimensional transport
equation
3(BC)_ejec)+1(ADec)+p
ar
ck
clx
dx
where
C
=
Constituent concentration, gm/m3
Q
=
Flow, m3/sec
A
=
Crossectional area, m2
D
=
Dispersion coefficient, m2/sec
B
=
Crossectional area, m2
X
=
x-coordinate, m
t
=
time, sec
P
=
Production of the constituent per unit section
length, gm/m-sec
The numerical solution technique for Equation 9.1 involves introducing the
transport, S, which is the quantity of the constituent passing a cross-section per unit time
(gm/sec) and is defined as
S = QC-AD 92
Etc
and then rewriting Equation 9.1 as
3S 8(BC)
dx dt
- P = 0
9.3


29
to, the stream bed. This material is conventionally referred to as bedload. Smaller or less
dense particles may be carried into the water column when turbulent forces are
considerably in excess of gravitational forces. This material represents the suspended
load. When turbulent suspension forces fall below their critical threshold values
suspended materials are re-deposited. Similarly when bed shear forces fall below their
critical threshold values bedload transport ceases. A variety of physical factors affect
particle transport including particle characteristics such as particle size, shape, density,
fall velocity, and electrical charge, and stream characteristics such as width, depth, flow
velocity, slope, roughness, water temperature, macrophyte population type, density and
location, and flow-stage relationships (Webster et al., 1987).
The literature on erosion of organic soils is exceedingly sparse. Studies of soil
erosion tend to focus on upland peat bogs or moors in north temperate or arctic climates.
Representative of these is the work of Labadz et al. (1991), who studied sediment yield
and delivery from blanket peat in Great Britain. They concluded that erosion from upland
peatlands is highly variable spatially and temporally with a strong stochastic component
attributable to the constant variation of surface morphology. In a similar vein, but in the
context of different terrain, Benda (1990) concluded that organic debris flows were an
important factor in determining the channel morphology of streams in the Pacific
Northwest. Benda hypothesized that the stochastic nature of sediment supply from debris
flows promotes cycling between channel aggradation and channel degradation,
accentuating temporal and spatial variability of channel morphology. These and similar
studies of upland erosion are of interest in that they may provide some qualitative
framework to understand the variability of particulate organic transport, but they do not
necessarily apply to the hydrology, hydrography, or surface morphology of the South
Florida region.
Research with more relevance to the South Florida hydrologic regime is that
which has included the study of production and movement of organic sediments in the


86
C = Srbate concentration in liquid phase, mg/1
Equation 5.1 may be rewritten as
S =
(5.2)
where:
S = Total substrate concentration of srbate, mg P/kg substrate
As liquid phase srbate concentration, C, is decreased to zero, Equation 5.2
approaches, in the limit, the simple linear relationship
S = ^C = iC (5.3)
Ks
where:
k = Sm/Ks = Constant of proportionality, 1/kg
So at low liquid phase concentrations Equation 5.3 may be used to estimate
sorption equilibrium. The parameter k is often referred to as the partition coefficient'.
The variables to be determined for each data set under the assumption of the
Langmuir isotherm form are So, the native adsorbed phosphorus concentration, Sm, the
maximum substrate concentration at saturation, and Ks, the saturation constant. There are
several approaches to determining these constants which include statistical extrapolation
of sorption data to zero liquid phase concentration to estimate So (Olila and Reddy, 1993),
and the use of single reciprocal (Eadie-Hoffstee) plots or double reciprocal (Lineweaver-
Burke) plots to linearize Equation 5.1 (Atkinson 1991). For the purposes of this work,


264
channels with different values of n describes several channels with n values in the
range of 0.05-0.06 as containing growths of weeds and grass or large size growth of trees.
These hydraulic results are further indication of the impact of the in-channel growth on
the drainage system operation.
Note that the calibration value of hydraulic conductivity was 55 m/day, compared
to the value of 12.2 m/day for the average of the local conductivities noted earlier. This
difference represents the farm-wide average effect of the mole drains in increasing the
effective hydraulic conductivity.
In general the hydraulic calibration process produced reasonably good fits similar
to those shown in Figures 9.2-9.4 for the major pumping cycle of each event. There were
some issues encountered in the application of the model to UF9200 that should be
discussed here. The official pump flow calibration for the small pump, when applied
with the parameter set developed in this calibration process, would not match the
observed levels with any degree of accuracy. There were no independent flow
measurements made as part of this study during the times the small pump was running, so
no alternative calibration curves could be developed. To circumvent this problem the
system was simulated using head boundary conditions instead of flow boundary
conditions for the periods when the small pump was running. This is a less desirable
simulation mode and reduces the accuracy of those simulation periods. This has a
minimal effect on the objective of this work, which was to develop a first approximation
model, because the bulk of the suspended solids transport took place while the large
pump was running.
The hydraulic model did not respond well to the few situations where there was a
large rainfall while the major pumping cycle was in progress. This is probably the result
of the quasi-steady state assumption made in the development of the drainage model
being violated by intense rainfall distorting the field hydraulic gradient. These few
situations were handled by adjusting the timing of the rainfall in the simulation to occur


110
CSVE as ilGmsolidanai Ture) with sequentially Increasing Shear Stress (t)
Prototype Cyanic Sediment -15.2 cm \tter Depth
0.008
Figure 6.4: CSME as Function of Consolidation Time for Organic Sediment Bed with
15.2 cm Water Depth
Figure 6.4 shows clearly the decrease in erosion rates as the consolidation time
increases from 2.6 through 6.7 days. The similarity of the erosion extent curves for the
6.7 day and 8.7 day consolidation times indicates that the effect of consolidation on
erosion appeared to be essentially complete by 6.7 days. It was concluded from this
series of test runs that a seven day consolidation time should be adequate to minimize the
effects of consolidation on erosion test results.
Series 2 Extended Erosion Tests Series 1 was run with a channel water depth of
15.2 cm (6 inches). During the runs it was noted that there was some difficulty
maintaining constant speed control at the lower shear (lower speed) levels. There was
also some concern that at the high shear levels suspended solids concentrations of 700-
800 mg/1 were making sampling accuracy somewhat problematic. In an attempt to


11
continues at the time of this writing as a joint effort between IF AS and Soil and Water
Engineering Technology (SWET), a private agro-environmental engineering firm in
Gainesville, Florida. It is under this overall research charter that the work described in
this dissertation was conducted.
The original water management BMPs focused on reducing phosphorus discharge
by reducing the volume of water discharged from the farm and by minimizing water table
fluctuations, that have an accelerating effect on soil mineralization and phosphorus
release. Thus the focus was on procedures that could reduce hydraulic load and minimize
phosphorus solubilization. The contribution of particulate matter to phosphorus export
was recognized (Izuno et al., 1991). however there were insufficient data available at the
time of formulation of the original BMPs to quantitate the impact of management
practices on particulate phosphorus transport and export. As a consequence none of the
original water management BMPs spoke directly to control of particulate phosphorus
transport. In addition, the main focus of the early modeling effort was on field scale
transport, with the irrigation and drainage systems existing as boundary conditions. This
resulted in an exclusion of the conveyance systems (field ditches, farm collector canals,
main farm canals, pumping systems, and ultimately the district canals) from analysis and
evaluation regarding phosphorus sourcing, assimilation, or transport. The impetus for the
work reported herein was the need to fill that recognized gap by development of a first
generation model which would give a reasonable approximation of the predominant
phosphorus mobilization, transformation, and transport processes prevalent in the farm
irrigation and drainage conveyance systems of the EAA.


220
Table 8.2: Canal Surficial Sediment Phosphorus Content UF9200 Synoptic
Survey of Julian Date 229
Sample Type
Location
Phosphorus Content
mg P/kg TS
In-stream Canal Samples
South Canal-East End
718
South Canal-West End
314
North Canal-East End
520
North Canal-West End
1203
East Canal-Middle
437
Spaced Samples Near Pump Station
Pump Suction
1828
East Canal-100 m Upstream
533
East Canal-200 m Upstream
530
East Canal-300 m Upstream
2297
North Canal-150 m Upstream
419
North Canal-250 m Upstream
691
North Canal-350 m Upstream
527


192
Time Series-Flow and Total Suspended Solids
Events UF9206N-283. 300
1000
800
600
400
200
0
282 284 286 288 290 292 294 296 298 300 302 304 306 308
E
o
U
t/5
1/5
Julian Date
Pumping Rate O TSS Concentration
Figure 7.25: Time Series-Flow and Total Suspended Solids Events UF9206N-283, 300
Time Series-Flow and Total Suspended Solids
Event UF9206N-318
Figure 7.26: Time Series-Flow and Total Suspended Solids Event UF9206N-318


7.28 Time Series-Flow and Total Suspended Solids Events UF9206S-209.
213,218,224, 234 193
7.29 Time Series-Flow and Total Suspended Solids Events UF9206S-260.
268 194
7.30 Time Series-Flow and Total Suspended Solids Events UF9206S-280,
300 194
7.31 Time Series-Flow and Total Suspended Solids Event UF9206S-317 195
7.32 Time Series-Flow and Total Suspended Solids Event UF9206S-336 195
7.33 UF9206N Phosphorus Content of TSS as a Function of TSS
Concentration 197
7.34 UF9206S Phosphorus Content of TSS as a Function of TSS
Concentration 197
7.35 UF9206N Phosphorus Content of TSS as a Function of TSS
Concentration (Expanded Scale) 198
7.36 UF9206S Phosphorus Content of TSS as a Function of TSS
Concentration (Expanded Scale) 198
7.37 Layout of UF9200 with Synoptic Sampling Locations 203
7.38 UF9200 Intensive Study-Event 252 205
7.39 UF9200 Intensive Study-Event 258 205
7.40 UF9200 Intensive Study-Event 262 206
7.41 UF9200 Intensive Study-Event 285 206
7.42 UF9200 Intensive Study-Event 258 South-East Canal Sample Locations,
First 10 Hours 209
7.43 UF9200 Intensive Study-Event 262 South-East Canal Sample Locations,
First 12 Hours 209
7.44 UF9200 Intensive Study-Event 285 South-East Canal Sample Locations,
First 32 Hours 210
8.1 UF9200 Two-Point Adsorption/Desorption Curves (Adsorption Mode) 236
xvii


46
provides shear forces that act on the particle-collection, tending to break down the
structure. The floe size distribution depends in a complex way on floe strength and
turbulence structure but a qualitative interpretation is that low levels of turbulence tend to
promote flocculation and produce light, diffuse floe structures, while high levels of
turbulence tend to promote de-agglomeration and produce dense, compact floe structures.
Sedimentation Sedimentation of cohesive particles falls in one of three general
categories. At concentrations of less than about 300 mg/1 particles act as individuals, are
not influenced by their neighbors, and maintain a constant settling velocity' independent
of particle concentration (Krone 1962). This region is known as the free-settling zone.
At concentrations above the free settling zone, particle-particle interaction and differential
settling cause the settling velocity to increase with increasing concentration as there
become more and more opportunities for agglomeration as concentration increases. This
region is known as the flocculation settling zone. Beyond a certain concentration,
typically 5000-10,000 mg/1 (Mehta et al., 1989) the flocculant structure of the sediment
becomes so extensive that bridging begins to occur and the floe structure becomes
partially self-supporting. At this point and beyond settling velocity tends to decrease with
increasing concentration. This region is known as the hindered settling zone.
Deposition The concept of deposition is treated differently in cohesive sediment
dynamics than in non-cohesive sediment dynamics. For a class of potentially cohesive
particles there may exist a critical bed shear stress beyond which no deposition will take
place at any reasonable concentrations. At bed shear stresses less than critical a fraction
of the suspended material may deposit. The fraction of deposition is related to the
departure of bed shear stress below critical. The simplest expression of this theory for a
monodisperse particle collection is the Krone formulation (Krone 1962)
for Tb < Td and Fa = 0 for tb > td
3.7


154
These field erosion tests were the prelude to a number of additional field
observations that directed attention toward the macrophyte population as a potential
major contributor of exportable particulate phosphorus. These observations are detailed
in the subsequent two chapters.


249
to be derived from the same basic structure as that of Schaffranek (1987) and Swain and
Chin (1990).
Husain, et al. (1988) have developed a modified version of the dynamic wave
operational model, based on the DWOPER model developed by Fread (1978) at NOAA.
The modifications were made to incorporate junctions, submerged gates, bottom falls,
and dynamic inflow from springs and reservoirs in the Al-Hassa irrigation system in
Saudi Arabia. The model is stated to have interactive features that enhance its
adaptability to general open-channel network systems
In general it appeared that there were several operational models available that
may be adapted to the specific configurations of the EAA. Most of the models seemed to
derive from the work of Amein and Fread at NOAA in the 1970's. A review article by
Cunge (1989) refers to advanced, menu-driven "fourth-generation" PC based models
developed in Europe for modeling of complex riverine-estuary systems.
An ASCE Task Committee reviewed several of these models in addition to others
developed in the U.S. as part of a comprehensive study of irrigation and drainage models
(ASCE Task Committee 1993). These models, along with the models of Schaffranek
(1987) and Swain and Chin (1990), were evaluated for project applicability by the
computer programmer on this project, Dr. Nigel Pickering of Soil and Water Engineering
Technology, Gainesville, Florida.
The model chosen was in fact one of the European models, DUFLOW, which is
an abbreviation for DU(tch) FLOW. The hydraulic program was originally developed
jointly by three institutions in the Netherlands; The International Institute for Hydraulic
and Environmental Engineering (IIHEE), Delft; Rijkswaterstaat (Tidal Waters Division),
The Hague; and Delft University of Technology. A later modification was made to
extend the utility of the program by the addition of a water quality module, which was
developed by the Agricultural University of Wageningen. The program was chosen
because it met the criteria listed previously, was user friendly and well documented, and


228
density of the macrophytes varied considerably among species, averaging 627 gm/nr for
water lettuce. 1116 gm/m2 for pennywort, and 1535 gm/m2 for water hyacinth. Volatile
content was in the range of 72-82% for the macrophytes and 51.3% for the filamentous
algae. The phosphorus content of all three floating macrophytes was well above that of
the surficial sediment and well into the range found in the discharge samples. Water
lettuce was highest, averaging 5687 mg/kg, pennywort was next averaging 4828 mg/kg.
and water hyacinth was lowest, averaging 4113 mg/kg. The filamentous algae was
somewhat lower, averaging 2258 mg/kg, which was still considerably higher than the
surficial sediment samples and also more than twice as high as the phosphorus content of
the lyngbya remains that had been separated from the field ditch B9B10 surficial
sediment samples reported in Chapter 5.
These results were encouraging in that they provided several possible sources that
could match the phosphorus content often seen in the discharge samples. It was not
immediately obvious, however, how the particulate phosphorus in the whole plant would
find its way directly to the farm pump discharge.
Dislodgable detritus studies
What was hypothesized to be the most probable source mechanism for
macrophyte contribution of particulate phosphorus was that detritus resulting from plant
senescence and attached microbial growth in the macrophyte root structure could detach
under both quiescent and turbulent conditions. Under quiescent conditions the detritus
could detach as a result of natural growth and senescence processes, aquatic vertebrate
and invertebrate interaction, and movement due to wind action. The detached material,
which could be a complex combination of biological matter, might be in several states of
buoyancy and flocculation. The portion of the material that was near neutral in buoyancy
and highly flocculant might be present in significant amounts but not be nicked up bv the


156
regarding particulate phosphorus transport. Large sample composite studies are included
in Chapter 8.
Time Series Discharge Studies
Time series discharge data were obtained over a seven month period from June
1994 through December 1994, from the single pump station at UF9200 and from both
pump stations at UF9206. An event was defined as any series of off-farm pumping
occurrences that were separated from an antecedent or subsequent pumping series by 24
hours or more, but which did not contain any pump down-time of more than 24 hours.
This definition proved to be very functional because there were numerous instances at
both farms during discharge from specific rainfall events when the pumps were down for
several hours, but the 24 hour or more cut-off almost always coincided with the discharge
pumping which could be associated with a particular storm or storm series. During the
period of measure there were 14 events monitored at UF9200, 17 events were monitored
at the North pump station of UF9206 (denoted as UF9206N) and 17 events were
monitored at the South pump station of UF9206 (denoted as UF9206S). Although there
were an equal number of events at UF9206N and UF 9206S, the events did not always
coincide.
Data Monitoring and sample acquisition
The basic data gathering infrastructure at each farm centered around a Campbell
Scientific CR-10 datalogger operated in conjunction with peripheral instruments for
measuring pump speeds, water depths and heads, temperatures, rainfall, and solar
radiation. Data were recorded at the pumphouse station on a five minute cycle,
downloaded twice weekly, and processed by computer to report hourly averages, or
hourly cumulative values where appropriate. In addition to logging data, the CR-10 also


91
An objective of this study was to determine if there were significant changes in
basic physical and chemical properties with change in particle size. In general the native
sorbed phosphorus, the adsorption affinities, and the partition coefficients were higher in
the <38p particle size fractions than in the 75-150p fractions, but the data did not show
dramatic increases for the smallest particle size fraction.
Potential environmental significance of adsorption-desorption
At this point in the program there was adequate information to develop a first
approximation of the potential contribution of sorption processes in thfe organic sediments
to soluble phosphorus in the liquid phase upon resuspension. The original evaluation
consisted of a matrix of results that covered a range options for the parameters
considered. For the purposes of illustration here several examples that cover reasonable
parameter values will be presented. The representative nature of the parameter values not
yet noted will be justified in later chapters.
Demonstration Example 1 Consider for the sake of illustration a system with the
following characteristics.
1. Sediment solids phosphorus content is 1000 mg/kg (Typical of values
reported in this chapter)
2. Pore water phosphorus is 0.6 mg/1 (The maximum value found in the
original sediment survey)
3. Phosphorus partition coefficient is 200 1/kg (The maximum found in the
particle size fractionation study)
4. Sediment solids content is 5%w, with solids specific gravity of 1.57
5. Initial liquid phase soluble phosphorus concentration is 0.1 mg/1
6. Sediment is resuspended in the liquid phase to a concentration of 50 mg/1


182
of water from the sod fields. The internal water storage that took place was restricted to
the cane-rice section of the farm, that was in turn drained by the South pump station.
The grower chose to run his pumps to meet sod drainage, cane water table control,
rice irrigation-drainage, and emergency needs on a daily decision basis. He also did not
practice any form of canal level control. The result of this control policy was that
pumping for similar rainfall events was often more frequent than was observed at UF9200
and that level and pumping rate fluctuations were more extreme than those at UF9200.
Sampling at UF9206 went through a shake-down process because of some
equipment issues but began in earnest on July 28,1994 (Julian Date 209) at both stations
and continued through December 9, 1994 (Julian Date 343) at UF9206N and December
12, 1994 (Julian Date 346) at UF9206S. Because of the more frequent pumping at
UF9206 the samplers filled completely between pick-ups more often at UF9206 than at
UF9200, thus there were more gaps in sample acquisition and a less complete data record
regarding particulate phosphorus transport. None the less the data record was still
extensive and adequate for basic statistical evaluation.
At UF9206 the definition of an event was modified somewhat on occasion for
convenience. No events had less than a 24 hour separation time but some events included
pumping series that were separated by more than 24 hours when this was felt to be
reasonable based on antecedent rainfall and pump run-time patterns. This was justified
because most of the data examination at UF9206 was based on individual samples rather
than the events. With this qualification in place let us take a brief look at the event data
in order to draw some parallels with UF9200.
There were a total of 15 nominal pumping events recorded at each station at
UF9206, although the dates for each event did not always coincide for the two stations.
Tables 7.3 and 7.4 present the event summaries for UF9206N and UF9206S respectively.


21
piled to the height of the original sediment to prevent sloughing of existing sediment into
the study areas.
Effectiveness of the traps were measured several ways. Sediment accumulation
pans were placed within the confines of the dredged sections at the start of the study and
retrieved by a diver 18 months later at the end of the study. Bottom soundings were made
in the dredged sections on a monthly schedule to estimate the increase in bottom
elevation with time. Regular water sampling was conducted before and after dredging of
the traps and upstream and downstream of the dredged sections to ascertain the degree of
removal of phosphorus resulting from the presence of the traps.
The results were interesting. The sounding data indicated that the dredged bottom
elevations increased at the rate of 1.4-1.8 ft/year (-0.43-0.55 m/yr). The sediment
accumulation pans, which were 5 cm deep and had been placed on bare bottom at the start
of the study, proved to be not especially useful because they were buried under 20-60 cm
of overlying sediment at the time of retrieval. The water sample phosphorus analyses,
however, gave no significant indication of phosphorus reduction arising from installation
of the sediment traps. Pre-dredging and post-dredging farm effluent phosphorus contents
were virtually identical at both locations. The average phosphorus analyses were 0.108
mg/1 pre-dredging versus 0.107 mg/1 post-dredging at one location and 0.086 mg/1 both
pre-dredging and post-dredging at the other. Upstream and downstream phosphorus
measurements showed similar results. At one location upstream average phosphorus was
0.111 mg/1, downstream average phosphorus reduced to 0.096 mg/1. At the other
location, upstream average phosphorus was 0.067 mg/1, downstream average phosphorus
increased to 0.077 mg/1. In neither case was the difference statistically significant at the
p=0.30 level.


APPENDIX H
DUPROL EROSION PROGRAM
I* Erosion Model EROS7 DUFLOW v2.0 mod */
/* /
/* First Version(EROS5) JDStuck Feb. 7,1996 */
/* Modified to EROS6 on Feb. 15,1996 Eliminated EMI from equations */
/* Modified to EROS7 on Mar.5,1996 Replaced ksed with settling velocity */
Variable
Type
Name
Default
Value
Units
Description */
*/
Water
TSS
[ 0.00]
g-TSS/mA3
;Suspended solids
concentration in water
column
Bottom
EM
[700.00]
g/mA2
;Erodible mass per unit area
parm
EPSI
[0.0000001]
day/mA2
¡Erosion coefficient
parm
VC
[0.0000]
m/sec
¡Critical velocity for erosion
parm
wsed
[120.00]
m/day
¡Sedimentation velocity
FLOW
U
[0.0200]
m/sec
¡Water Velocity
FLOW
Q
[1.0000]
mA3/sec
¡Flow
FLOW
As
[7.2000]
mA2
¡Flow area
FLOW
Z
[1.0000]
m
¡Depth of water
389


295
Event UF9200-285
CUM TSS Load
7000
Cum Hydraulic Load-mA3
Observed Simulated
Figure 10.19: Simulation Results-Cumulative Suspended Solids Load for Event
UF9200-285
Figure 10.20: Simulation Results-Cumulative Particulate Phosphorus Load for
Event UF9200-285


80
Figure 5.5: Particle Specific Gravity Distribution of Hydrated B9B10 Soil, B9B10
Sediment, and WPBC Sediment
Linear correlation of particle specific gravity with organic matter content is shown
in Figure 5.6, where the expected trend of decreasing specific gravity with increasing
organic matter content is evident. The extrapolated linear correlation value at zero
organic matter is 2.73, which is in the mid-range of typical values of 2.85 for carbonate
particles and 2.65 for sands (Vanoni 1975). The extrapolated value at 100% organic
matter content is 1.28.
Total phosphorus content
The total phosphorus content of the various fractions is shown in Figure 5.7,
where phosphorus content is based on total solids mass. The WPBC sediment sample
showed an excursion of increased phosphorus content in the particle size range of 75-150
micrometers.


57
effort that, from a pragmatic standpoint, should not be expended until the nature of
erosion/sedimentation associated with shear flow is reasonably well understood. It was
decided the best allocation of resources for this project would be to attempt to understand
the nature of shear flow erosion and defer perpendicular flow erosion studies if and until
it became evident that they were necessary.
Current restrictions by the South Florida Water Management District limit the
growers to a nominal farm pump discharge of one inch of farm runoff per day, therefore
the time scale of pumping events in the EAA may usually be on the order of one to three
days after a significant rainfall event. Typical time constants for biological assimilation
and degradation processes may be on the order of weeks (USEPA 1985). For the first
generation models to be developed in this study it was deemed appropriate to ignore
biological process transients occurring during the pumping events, that is, to consider that
all biological processes of significance occur during the inter-event quiescent periods.
Given the accuracy with which biological processes in natural systems can be portrayed,
this is a reasonable first approximation.
The conceptual model resulting from these simplifications, illustrated in Figure
4.2, is as follows
\J Soluble phosphorus is supplied from the fields as a boundary condition in
surface and ground water runoff to the farm drainage ditches.
Soluble phosphorus may also be added to the system with the release of
sediment interstitial pore water when sediment is resuspended.
Particulate phosphorus may be supplied to the system as:
eroded soil or litter carried by overland flow of surface runoff,
* biological matter already suspended at event initiation,
sediment that is resuspended as a result of turbulent shear
stresses arising from channel flow.
Particulate phosphorus may be removed from the system by sedimentation.


225
The spot herbicide spraying appeared to have a very temporary impact in that the surface
layer of vegetation would die but not sink, so within a few days new macrophytes were
propagating on the floating mat left from the preceding generation.
The systematic change in areal coverage was a seasonal one. It was mentioned
earlier that the grower had implemented an extensive ditch and canal maintenance
program early in the spring. At the start of the wet season there was almost no aquatic
macrophyte growth visible at any point in the conveyance system. As the season
progressed there was clearly an increase in macrophyte coverage with bridging occurring
at many locations. The pump station region became completely covered with floating
macrophytes early in the season and remained so for the duration of the year. The
previously mentioned bridges also formed in the East Canal early in the season. East
Canal coverage grew steadily during the long pumping hiatus between Events 169 and
220 so that by Julian Date 220 the East Canal had roughly 40-50% of its surface area
substantially occupied by macrophytes. By Julian Date 240 the East Canal was 80-90%
covered by macrophytes. This situation prevailed for the remainder of the wet season.
The North and South Canals, which were each four times the length of the East Canal, did
not suffer the same percentage coverage as the East Canal but they had numerous and
extensive sections where macrophyte coverage increased, bridged, and expanded as the
season progressed.
There was not a scheduled program to estimate areal macrophyte coverage on a
regular basis at either UF9200 or UF9206. The above description was culled from field
logs made at various times for other purposes. In hindsight this was an unfortunate
omission because that information could have been interesting and useful, but as we shall
see the model development assumes lumped parameters that are calibrated to the field
data, so for the first level of model development that is attempted in this work the
omission was not critical.


178
Correlations were done with power functions and with asymptotic power
functions. Even though the data appears to exhibit asymptotic behavior, the simple
power functions gave the best correlations, expressed by the following equations.
PC; =18970 x(TSS)-34,3 7.1
PC, = 13770x (7XS)-0 42056 72
where PC, = Phosphorus content of discharged TSS for the first
four hours of an event, mg/kg
PC, = Phosphorus content of discharged TSS for the
remainder of an event, mg/kg
TSS = Discharged suspended solids concentration, mg/1
Equation 7.1 correlated the data with an r value of 0.581 and Equation 7.1 correlated
with an r2 value of 0.381, indicating the data scatter that is obvious in Figures 7.14 and
7.15.
It must be made clear at this point that correlating phosphorus content as a
function of total suspended solids concentration is not in any wav intended to imply that
phosphorus content is the dependent variable and total suspended solids concentration is
the independent variable. In fact the hypothesis arising from these observations is that
both parameters are variables that are dependent in a complex way on time, biology, and
hydrodynamics.
If we put aside the biological dependency for the time being, the tentative
hydrodynamic explanation for the correlations exhibited in Figures 7.14 and 7.15 is that
there is a direct relationship between phosphorus content and particle transportability, that
is, as phosphorus content decreases particle transportability also decreases. In this
situation as hydraulic energy input increases more of the less transportable, lower


260
match the upstream water levels in the upper reach of the North Canal.
Well 09.
Later in the quality modeling process it was deemed appropriate to define
the flow area at the pump station and the North Canal discharge as the
areas of the pump intake pipe and the discharge culverts respectively. The
process was repeated with this modification, and represents the final
hydraulic calibration.
This process was extremely iterative in nature. Initially there was considerable
difficulty encountered matching the pump station level records at the low levels
encountered at the end of the first major pump cycle, where a large portion of the particle
resuspension appeared to be taking place. In fact for several events no combination of
parameters could be found that would match this portion of the level data. An alternative
pump calibration curve was developed using a combination of the data points from the
official calibration curves and the five transects mentioned above. It was found that this
revised curve, with one outlying point removed, yielded a pump-flow calibration curve
that gave much more tractable results. This revised pump curve was used for all
subsequent calibration operations.
The calculated flows at the major structures and the measured values for these
structures are shown in Figures 9.2 and 9.3 for events 237 and 252. Note that the
calculated and observed values agree reasonably well for all locations, but the internal
structures at Culverts 4-5 and 8-9 deviate more than the others. This is again most likely
due to using uniform farm-wide values for channel resistance and hydraulic conductivity.
An example of the simulated and observed water levels at the calibration locations
is shown in Figure 9.4, in this case for the first 50 hours of event UF9200-220. The
pattern shown here is typical of all the events, and reflects the priorities of the calibration
process.


92
Now consider for calculation purposes one liter of liquid phase, ignoring the small
volume and mass changes of this control volume with sediment resuspension. With the
given pore water phosphorus and partition coefficient the adsorbed sediment phosphorus
would be 120 mg/kg. Upon resuspension the 50 mg of suspended solids would carry (50
mg x 120 x 1 O'6 mg/kg) = 0.006 mg sorbed phosphorus along with 950 mg (0.95 mi) of
pore water. The pore water would contribute (0.6 mg/1 x 0.95 x 1031) = 0.00057 mg of
soluble phosphorus to the liter control volume. The suspended 50 mg of sediment would
reach sorption equilibrium with the control volume.
The simple equations describing the equilibrium and mass balances for the one
liter control volume and the suspended sediment and pore water are:
Mass balance for phosphorus in the liquid
010057/^+7^
1liter
=C
54
Mass balance for phosphorus originally sorbed on the sediment
o-6
(50xl0"ufe)S+7V^ = 0.006mg
55
Equilibrium relationship
Where:
S = 200C
5.6
C = Soluble phosphorus concentration in liquid phase, mg/1
Pdes = Mass of phosphorus desorbed from the 50 mg of sediment, mg
S = Srbate concentration on suspended sediment, mg/kg


376
STN17.NET
Sections
For sections the NET file is in the format
First Line
SECT Label..Section No...Program Code No. for section...Start node...End Node
...Length...Start Bottom Elevation...End Bottom Elevation...Start k... End k
Subsequent Lines
W...Direction of Section, degrees... Direction of Wind, degrees
H...Height above bottom for first width... Height above bottom for second width...etc.
BS1...Start flow width at first height... Start flow width at second height...etc.
BS2...End flow width at first height...End flow width at second height...etc.
BB1...Start storage width at first height... Start storage width at second height...etc.
BB2...End storage width at first height...End storage width at second height...etc.
Where there are is no change in channel width with elevation DUFLOW abbreviates the
width dimensioning array to one line labeled BS.
Structures
For structures (all culverts in this case) the NET file is in the format
First Line
STRU Label...Structure No...Program Code No. for structure...Start node...End Node
...Length...Sill Elevation...Width...Gate Elevation
Second Line
MU Label...Loss coefficient (emerged) in the positive direction...Loss coefficient
(emerged) in the negative direction... Loss coefficient (submerged) in the positive
direction... Loss coefficient (submerged) in the negative direction...Chezy Coefficient


197
Station 9206N Phosphorus Content of Suspended Solids as a Function of Suspended
Solids Concentration
Suspended Solids Concentradon-mg/1
Figure 7.33: UF9206N Phosphorus Content of TSS as a Function of TSS Concentration
Station 9206S Phosphorus Content of Suspended Solids as a Function of Suspended Solids
Concentration
Suspended Solids Concentration-mg/I
Figure 7.34: UF9206S Phosphorus Content of TSS as a Function of TSS Concentration


53
Original conceptual model
Within the context of the modeling portion of this project, fertilizer contribution
to phosphorus export is being handled as a subset of soluble phosphorus export in the
field groundwater, so for the sake of conciseness fertilization will be excluded from
further discussion. With this caveat, the remaining factors, as hypothesized at the
initiation of this program, are presented in conceptual form in Figure 4.1.
Figure 4.1: Farm Scale Phosphorus Transport Original Conceptual Model


367
The examples are for a channel depth of 30 cm and a value of To of 1.2
dynes/cm2. Two values of Manning's n, converted to Nikuradses k are used. The
n value of 0.018 represents the value calculated from Mehta's (1973) data. The value
of 0.025 is the roughness chosen to represent the organic sediment surface. The
spreadsheet formulas are shown in Table D.l. an example of the spreadsheet output for
the n=0.025 condition is shown in Table D.2.
Table D. 1: Example of Spreadsheet Output
tauzero =
1 .200
k =
4 650
n =
0.025
V f=
1.095
m id V/Total v
38v/Total v
0.30712162
0.2889454
y h
top down v
Bottom up v
Total v
v/vt
0.0010143 29
9989857
0.00760081
14.3949604
40.6900407
1
0.01
29.99
6.27460221
14.39414
34 4222188
0.84596177
0.1
29.9
12.5804912
14.385909
28 1080989
0.69078572
0.5
29.5
16.9881 184
14.3490248
23.6635875
0.58155723
1
29
18.8863802
14.3022098
21.7185107
0.53375495
3
27
21.8950538
14.1065113
18.5141386
0.45500418
5
25
23.2940074
13.8957449
16.9044185
0.41544364
7
23
24.2154746
13.6673949
15.7546014
0.38718569
9
21
24.9037275
13.4182584
14.817212
0.36414837
10
20
25.1922691
13.284641 1
14.395053
0.35377337
1 1
19
25.4532868
13.1441686
13.9935628
0.34390634
12
18
25.6915773
12.9960994
13 6072032
0.33441 1 15
13
17
25.9107833
12.8395646
13.2314624
0.32517693
14
16
26 1137363
12.6735373
12.862482
0.31610885
15
15
26.302681 1
12 4967913
12.4967913
0.30712162
16
14
26.4794271
12.3078465
12.3078465
0.3024781 1
17
13
26.6454545
12.1048935
12.1048935
0.29749033
18
12
26.8019892
11.8856875
11.8856875
0.2921031 1
18.55
1 1.45
26.8844161
1 1.7572002
1 1 .7572002
0.2889454
20
10
27.0905309
1 1 .3863793
1 1.3863793
0.2798321
21
9
27.2241482
1 1 .0978377
1 1 .0978377
0.27274088
23
7
27,4732847
10 4095848
10.4095848
0.25582635
25
5
27.7016347
9.4881 176
9.4881 176
0.23318034
26
4
27.809045
8.87701382
8.87701382
0.21816183
27
3
27.9124012
8.08916402
8.08916402
0.19879961
28
2
28.01 1998
6.97875209
6.97875209
0.17151008
29
1
28.1080996
5.08049035
5.08049035
0.12485833
29.2
0.8
28.1269217
4 46938657
4 46938657
0.10983982
29.54585
0.45415
28.1591678
2.91882732
2.91 882732
0.07173321
29.845
0.155
28 1867566
-0.02518807
-0.02518807
-0.00061902


84
Adsorption-Desorption experimental methods
Two particle size ranges were chosen from the WPBC sediment and B9B10
sediment fractions for detailed evaluation of phosphorus adsorption/desorption
equilibrium data. The most mobile less than 38 micrometer particle size fraction of
both sediments was the logical first choice. The 75-150 micrometer fraction of the
WPBC sample showed a significant increase in phosphorus content and decrease in
organic matter content. This fraction of both sediments was chosen as the second particle
size range for sorption data.
The solids for each sample used in the sorption tests were concentrated by
centrifugation at 5000g for 30 min. The concentrate was analyzed for total solids, and
this analysis was used to convert betweeiywet and dry masses. Each sample, because of
its origin, was presumed to contain some initial amount of sorbed phosphorus, the
quantity of which was unknown. This unknown amount of initial phosphorus was
estimated statistically as a part of data reduction.
Standard phosphorus solutions were made up using potassium dihydrogen
phosphate and 0.1 M potassium chloride solution. The 0.1 M KC1 solution had been
shown to have approximately the same conductivity as ambient water samples from the
EAA (900-1000 micromhos/cm) and was assumed, as a first approximation, to also have
the same ionic strength. Solution strengths were 0.0, 0.06, 0.2,0.6, 1.0,2.0, 6.0, 10.0,
20.0, and 60.0 mg/1 phosphorus. The solids were mixed with the various solutions in a
weight ratio of 1 dry solid unit to 100 solution units, agitated in an end-over-end shaker
for 24 hrs, and then centrifuged at 3620g for 15 min. The liquid supemate was decanted,
acidified to pH 2.0 with 6N sulfuric acid, and stored at 4 C until analysis for total
phosphorus. The data gathered during this phase was designated as Adsorption data.
Following the decantation of supemate, the concentrated solids in the centrifuged
subnate were resuspended in phosphorus-free 0.1 M KC1 solution, and again subjected to


3
It is important to note for later reference that for thousands of years the hydrologic
regime in the Everglades watershed was predominantly that of large volume discharge
spread over shallow channel beds tens of kilometers wide flowing at velocities that were
in the range of meters/day (Douglas, 1988).
The first serious human disturbance of the Everglades watershed.camein 1883
when developer Hamilton Disston dug a canal between Lake Okeechobee and the
Caloosahatchee River in an attempt to drain a 1.6 million ha (4 million acre) tract of
wetlands. From that time through the present the planning and implementation of water
management in the Everglades region has passed from the control of wealthy developers
through a series of state trustee boards, commissions, and agencies. From 1906 to 1931
the Everglades Drainage District (EDD) followed by the Okeechobee Flood Control
District (OFCD) installed a number of canals, levees, locks, and dams in the Everglades
Region.
One of the objectives of the public works activity was to drain mucklands just
south of Lake Okeechobee for agricultural use. This region, known as the Everglades
Agricultural Area (EAA), consists of 700,000 acres (~283,000 ha) of rich organic soil that
has supported a variety of agricultural activities since its creation. Currently, major
winter vegetable and sugarcane industries are the predominant activities in this area.
Because of the low elevation and flat terrain extant in the EAA and the swings in water
supply from wet to dry season the crops grown there are highly dependent on artificial
control of water table levels by off-farm pumping during wet periods and irrigation
during dry periods. Much of the man-made hydraulic system installed in South Florida
has had as one of its objectives the facilitation of water table control in the EAA.
In 1948 joint state-federal action created the Central and Southern Florida Flood
Control District (CSFFCD), which acted in concert with the United States Army Corps of
Engineers (USACE) to "develop and implement plans to provide flood protection, ensure
adequate water supply, prevent saltwater intrusion along the Florida Lower East Coast,


23
The data set was subjected to a screen for samples which met the criteria of:
1. Taken from one of the four major southern discharge stations
2. Taken during times of flow
3. Having analytical results for Total Suspended Solids, Total Phosphorus,
and Ortho-phosphorus.
This screen produced 55 samples from the entire data set which met all three
criteria. The above-mentioned soluble phosphorus correction was applied to the screened
data set. The results showed an average Total Suspended Solids (TSS) of 14 mg/1,
average total phosphorus of 0.140 mg/1, average soluble phosphorus of 0.099 mg/1 (70%
of total) and average particulate or particulate phosphorus of 0.042 mg/1 (30% of total).
Regression of particulate phosphorus on TSS indicated a phosphorus content of the
suspended solids on the order of 2200 mg P/kg TSS (ppm).
The phosphorus content of 2200 mg/kg and the presumed upper mode of 1700
mg/kg of the Anderson and Hutcheon Engineers Study are both considerably higher than
the phosphorus contents of the canal sediments reported in the Anderson and Hutcheon
Engineers Study, 58-77 mg/kg for 5 of 6 canals and 252 mg/kg for the remaining canal.
This major difference in phosphorus content poses an intriguing clue to the possible
sourcing and transport mechanisms of particulate phosphorus in the EAA. The topic will
be pursued in some depth in later chapters.


258
Eighteen integrated velocity transects taken during three events. These
transects were taken at each major canal culvert and at a foot bridge over
the East Canal immediately upstream of the pump station. Two complete
sets were taken during Events 237 and 252, additional transects of the
North Canal Culverts and the Bridge were taken during Events 252 and
258. These transects allowed calculation of total mass flow at each major
culvert and also provided five pairs of coincident flow measurements in
the two canals feeding the pump station just upstream of the pumps. All
transects were run when the large pump was in operation.
Well records of mid-field water table at two locations, Well 01, located in
field A1 adjacent to the pump station, and Well 08, located in Field A12,
approximately 2300 meters West of the pump station.
Canal water elevation records for Well 02, located at the North culvert of
Field A1 (flowing to the North Canal), Well 05, located at the center of the
farm on the South Canal, and Well 09, located at the North culvert of Field
A12 (flowing to the North Canal).
Canal water elevation records for the pump station inlet.
Pump flow records and flow rates as calculated from the official BMP
project calibration curves.
Hydraulic Model Calibration Methodology and Results The calibration
methodology entailed a joint calibration over six events (220,237,252,258, 262, and
285) during the normal wet season, conducted as follows:
The calibrations were carried out over the first major pump cycle of each
event when the large pump ran until first shut-down on low level. For this
reason Event 244 was excluded because it was a small event where only
the small pump was used.


393
Bagnold. R.A.. 1966. "An Approach to the Sediment Transport Problem of General
Physics". USGS Professional Paper 422-1
Behrendt.H. 1990. The Chemical Composition of Phytoplankton and Zooplankton in a
Eutrophic Shallow Lake", Arch. Hvdrobiol.. 118(2): 129-145
Benda,L. 1990. Influence of Debris Flows on Channels and Valley Floors in the Oregon
Coast Range. U.S.A., Earth Surface Processes and Landforms. 15:457-466
Bloesch.J., and Uehlinger.U. 1986. Horizontal Sedimentation Differences in a Eutrophic
Swiss Lake, Limnol. Oceanogr.. 31:1094-1109
Bokuniewicz,H and Amold.C.L. 1984. Characteristics of Suspended Sediment
Transport in the Lower Hudson River, Northeastern Environmental Science.
3(3/4): 184-189
Bottcher.A.B.. and Izuno.F.T. 1993. Procedural Guide for the Development of Farm-
Level Best Management Practice Plans for Phosphorus Control in the Everglades
Agricultural Area, Florida Cooperative Extension Service, University of Florida.
Gainesville, FL
Bottcher,A.B., and Pickereing, N. 1995. Modeling Status for Phase III Final Report, in
Implementation and Verification of Best Management Practices for Reducing
Phosphorus Loading in the EAA: Phase III Final Report to EAA Environmental
Protection District, F.T. Izuno, Ed., Institute of Food and Agricultural Sciences
Everglades Research and Education Center, University of Florida, Belle Glade, FL
Bouwer, H and Jackson, R.D., 1974. Determining Soil Properties, In Drainage for
Agriculture, Agronomy Monograph No. 17, J. Van Schilfgaarde, Ed., American
Society of Agronomy, Inc. Madison. WI
Calvo.C., Grasso.M., and Gardenghi.G. 1991. Organic Carbon and Nitrogen in
Sediments and Resuspended Sediments of Venice Lagoon: Relationships with
PCB Contamination, Mar. Poll. Bull.. 22:543-547
Chow.V.T., 1959. Open Channel Hydraulics, McGraw-Hill, New York
CH2M Hill Consulting Engineers. 1978. Water Quality Studies in the Everglades
Agricultural Area of Florida, Report submitted to The Florida Sugar Cane
League
Clemmens, A.J., Holly, F.M., Jr., and Schurrmans, W. 1993. Description and
Evaluation of Program DUFLOW, J. Irrig. Drain. Enene.. 119(4): 724-734


291
UF9200-237
Event UF9200-237


5.13 Sedimentation Velocity Distribution as Determined by The Bottom
Withdrawal Technique-B9B10 Sediment <38 Micrometer Particle Range 97
5.14 Mean Sedimentation Velocity vs. Test Duration 97
6.1 Schematic Diagram of Counter-Rotating Annular Flume (CRAF) 102
6.2 Photograph of Counter-Rotating Annular Flume (CRAF) 103
6.3 CSME vs. Time Into Run for Placed Organic Sediment Bed with 15.2
cm. Water Depth 108
6.4 CSME as Function of Consolidation Time for Organic Sediment Bed
with 15.2 cm. Water Depth 110
6.5 CSME as Function of Shear Time at Constant Calculated Shear Rate of
2.0 dynes/cm2 for Consolidated Organic Sediment Bed with 30.5 cm.
Water Depth 111
6.6 CSME as Function of Shear Time at Several Calculated Shear Rates for
Erosion of Consolidated Prototype Organic Sediment Bed with 30.5 cm.
Water Depth 112
6.7 Initial Erosion Rates vs. Applied Shear Stress for Organic Sediment
Extended Run 114
6.8 Asymptotic CSME vs. Applied Shear Stress for Organic Sediment
Extended Run Erosion at 30.5 cm. Water Depth 115
6.9 CSME as Function of Shear Time at Several Calculated Shear Rates for
Deposition of Consolidated Prototype Organic Sediment Bed at 30.5 cm.
Water Depth 117
6.10 Asymptotic CSME vs. Applied Shear Stress for Organic Sediment
Extended Run Erosion and Deposition with 30.5 cm. Water Depth 118
6.11 CSME as Function of Shear Time at Several Calculated Shear Rates for
Erosion of Consolidated Prototype Organic Sediment Bed at 15.2 cm.
Water Depth 122
6.12 Asymptotic CSME vs. Applied Shear Stress for Prototype Organic
Sediment Extended Run Erosion with 30.5 cm. and 15.2 cm. Water
Depth 123
6.13 Asymptotic CSME vs. h2/3V2 for Prototype Organic Sediment Erosion
Extended Run with 30.5 cm. and 15.2 cm. Water Depth 131
xiv


335
lower the total particulate phosphorus load from UF9206 was over three times that of
UF9200. This difference in performance between the two farms may be evaluated by an
indirect analysis of velocity differences. Each pump station at UF9206 had two high
capacity variable speed pumps. These pumps were used aggressively to reduce canal
water depths, particularly in the North Canal system which drained the water-sensitive
sod fields.
No detailed canal topographic measurements were made at UF9206 so calculation
of specific velocities is not possible in this presentation. We can. however, use canal
water elevation as a partial surrogate for velocity, recognizing the uncertainty involved in
crossectional areas and pump flow-head relationships.
First it is useful to look at the wet season water elevation trends for both UF9206
stations. Figure 11.12 shows this trend for Julian Dates 210 through 300. Forthe
purpose of comparison, Figure 11.13 shows the same type data for the UF9200 pump
station. Note that there is a trend toward lower canal elevation as the season progresses at
all three locations, but the trend is modest at UF 9200, 1.2 mm/day, compared with 5.0
mm/day at UF9206N and 9.2 mm/day at UF9206S.
Second the extremes in elevations are much greater at the UF9206 pump stations
than they were at UF9200. At UF9200 the extremes are a high of 2.95 m MSL and a low
of 1.95 m MSL, a maximum variation over the season of 1 m. At UF9206 both stations
had maxima over 2.0 m MSL: and minima close to 0.0 m MSL, for a maximum seasonal
variation of about 2m, twice the variation of UF9200. If we consider each collection of
pump cycling episodes at UF9200 as single approaches to a minimum then the pump
stations at UF9206 have about twice as many sustained pumped approaches to a
minimum level as UF9200.
These differences are a direct result of the control policies practiced at each farm.
At UF9200 level was controlled by a pump shut-off switch that precluded canal levels
from falling below 1.95 meters. This set point was not altered during the season. At


CSME
Cumulative Specific Mass Eroded
EAA
Everglades Agricultural Area
EAAEPD
Everglades Agricultural Area Environmental Protection District
ENP
Everglades National Park
EPA
Everglades Planning Area
EPD
Environmental Protection District
EREC
Everglades Research and Education Center
FPOM
Fine Particulate Organic Matter
IFAS
Institute of Food and Agricultural Sciences
IP
Insoluble Phosphorus
OFCD
Okeechobee Flood Control District
PES
Particle Entrainment Simulator
PP
Particulate Phosphorus
SFWMD
South Florida Water Management District
SP
Soluble Phosphorus
STA
Storm Treatment Area
SWET
Soil and Water Engineering Technology
TDP
Total Dissolved Phosphorus
TP
Total Phosphorus
TSS
Total Suspended Solids
USACE
United States Army Corps of Engineers
WCA
Water Conservation Area
XXVI


403
Ziegler.C.K.. Tsai.C.H., and Lick.W. 1987. The Resuspension, Deposition, and Transpon
of Sediments in the Venice Lagoon. UCSB Report ME-87-3, University of
California. Santa Barbara


233
results may suggest several things. The Coefficient of Variation of the second detrital
phosphorus content set, 33.6%, was much higher than the CV for the first set of free-
floating samples, 4.9%. This may reflect the fact that the first sample set, taken in a
limited geographic location, may have represented a more narrow spectrum of the
population than the second, more geographically diverse set. The higher average
phosphorus content for the second detrital set may also reflect a seasonal difference,
however the data are insufficient to confirm this as a general trend.
The thrust of these detrital dislodgment studies was to evaluate the plausibility of
the macrophyte-supply hypothesis in providing a partial explanation for the gross
discrepancies between the surficial sediment phosphorus content and the phosphorus
content of the suspended solids in the farm discharge. The macrophytes were shown to
have a very high percentage of their total mass that could be detached by agitation and the
detached materia! in the first sample set was shown to have considerably higher
phosphorus content than the typical surficial sediment samples, but it was still lower than
many of the discharge samples. There was some evidence with the second set of samples
that higher detrital phosphorus content might be found in a wider geographical sampling
and that there might be a temporal effect on detrital phosphorus content. The studies
achieved what they were intended to accomplish in that they demonstrated the plausibility
of the in-conveyance plant growth as a potential source of relatively high phosphorus-
content suspended solids.
Large Composite Sample Studies
The results reported in Chapters 7 and 8 require that the original
adsorption/desorption assumptions, made in Chapter 5, be questioned and re-tested.
Recall that the tests run on samples-of surficial sediment, which showed weak sorption
isotherms, had led to the preliminary conclusion that phosphorus adsorption/desorption


7.11 Hydraulic and Total Suspended Solids Profiles for Event UF9200-319 173
7.12 Hydraulic and Total Suspended Solids Profiles for Event UF9200-336 173
7.13 Hydraulic and Total Suspended Solids Profiles for Event UF9200-355 174
7.14 UF9200 Phosphorus Content of TSS as a Function of TSS Concentration 177
7.15 UF9200 Phosphorus Content of TSS as a Function of TSS Concentration
(Expanded Scale) 177
7.16 UF9200 Particulate Phosphorus Concentration in Discharge as a
Function of TSS Concentration (Expanded Scale) 180
7.17 Correlation of Event TSS Load with Event Hydraulic Load for Station
UF9206N 187
7.18 Correlation of Event PP Load with Event Hydraulic Load for Station
UF9206N 187
7.19 Correlation of Event TSS Load with Event Hydraulic Load for Station
UF9206S 188
7.20 Correlation of Event PP Load with Event Hydraulic Load for Station
UF9206S 188
7.21 Time Series-Flow and Total Suspended Solids Events UF9206N-209,
213 190
7.22 Time Series-Flow and Total Suspended Solids Events UF9206N-224,
234,239 190
7.23 Time Series-Flow and Total Suspended Solids Events UF9206N-251,
256, 260 191
7.24 Time Series-Flow and Total Suspended Solids Events UF9206N-268,
273,279 191
7.25 Time Series-Flow and Total Suspended Solids Events UF9206N-283,
300 192
7.26 Time Series-Flow and Total Suspended Solids Event UF9206N-318 192
7.27 Time Series-Flow and Total Suspended Solids Event UF9206N-336 193
xvi


266
Farm surficial sediment samples averaged 507 mg/kg phosphorus content
in a synoptic survey taken early in the wet season, 841 mg/kg 2 months
later in the season (Chap. 8).
Surficial sediment samples taken near or beneath floating macrophytes had
phosphorus contents averaging 2063 mg/kg (Chap. 8).
Phosphorus content of farm discharge suspended solids for the study
period showed a characteristic average of greater than 4500 mg/kg at
UF9200, and greater than 2200 mg/kg at UF9206S, the sugarcane drainage
portion of the UF9206 system (Chap. 7).
The phosphorus content of discharged suspended solids decreased as total
suspended solids increased and asymptotically approached values that
were in the range of 900-1200 mg/kg only at very high discharge
suspended solids concentrations (Chap. 7).
Phosphorus contents of representative floating macrophytes from UF9200
ranged from 4100-5700 mg/kg. Algae samples averaged around 2200
mg/kg (Chap. 8).
Dislodgable detritus phosphorus samples from water lettuce averaged
around 2600 mg/kg phosphorus content early in the study period, 4200
mg/kg 2 months later (Chap. 8).
The large scale discharge composite samples verified the high phosphorus
content observed in the discrete samples and in fact were typically higher
in phosphorus content than the corresponding discrete samples (Chap. 8).
The labile fraction of phosphorus analyzed in the large scale composite
farm discharge samples averaged about 30.6% at UF9200, compared with
only 6-9% for the prototype surficial sediment samples from the same
farm (Chap. 8 and Chap. 5).


166
The second lapse reflects the time period between the last off-farm pumpage
(Event 285) before the nominal end of the wet season (End of October) and a series of
dry season tropical depressions that moved across South Florida during the last half of
November and all of December.
The two situations which caused the lapses were extremely fortuitous. The
passage of the tropical depressions provided the opportunity to include several extreme
weather events in the study, and the lapse caused by the grower's water storage BMP
gave a unique opportunity to evaluate the impact of cessation of off-farm pumping for an
extended period. This latter opportunity in fact gave concrete support to a major
modification in the approach to the particulate export model, which will be discussed
later in this chapter. For this portion of the discussion, however, it is appropriate to
consider some very basic parameters which may be derived from Table 7.2.
Consider Figures 7.1 and 7.2, which plot event TSS load and event particulate
phosphorus (PP) load respectively against event hydraulic load. Linear correlation
equations which pass through the origin for these data points have slopes which may be
considered to represent some average concentration which characterized the farm
discharge over the duration of the study. The slopes of the respective curves are 20.45
mg/1 for total suspended solids (TSS) and 93.3 pg/1 (or 0.093 mg/1) for particulate
phosphorus (PP). The ratio of these two slopes is 0.004562, which may be considered to
represent a characteristic average phosphorus content of the discharged suspended solids.
Stated in terms which have already been used in this work, the ratio of the slopes is
equivalent to a phosphorus content of 4562 mg/kg. For future reference it should be
noted here that this value is considerably higher than the values reported (900-1150
mg/kg) from the preliminary analyses of sediment fractions, discussed in Chapter 5.


307
Suspended Solids Simulation Results
Event UF9200-355
180
0 Observed Simulated
Figure 10.25: Suspended Solids Simulation Results for Event UF9200-355


170
Time Series-Flow and Total Suspended Solids
Event UF9200-237
Flow Rate
TSS Concentration
Figure 7.5: Hydraulic and Total Suspended Solids Profiles for Event UF9200-237
Time Series-Flow and Total Suspended Solids
Event UF9200-244
Julian Date
35
- 30
25 ^
E
20
- 15 5
C/5
7 10 h
5
0
246
Pumping Rate O TSS Concentration
Figure 7.6: Hydraulic and Total Suspended Solids Profiles for Event UF9200-244


141
Figure 6.18: Results of PES Runs with Prototype Organic Sediment CSME vs. Grid
Oscillation Frequency
Calibration of PES against flume and extrapolation to field parameters
In reality the difficulty encountered in correlating erosion of the prototype organic
sediment with the fundamental parameter of shear stress did not pose a major setback.
The observation of the diverse character of the sediment bed surface and the variation of
the PES results imply that the roughness coefficients and friction factors for freshwater
organic sediment beds existing in a dynamic sub-tropical ecosystem would be expected to
be much more variable than those of predominantly inorganic clay or sorted sand found in
marine environments.
Methods for measurement of the roughness coefficients or friction factors of
individual beds typically require substantial amounts of bed material in a linear flume.
Methods are available for estimation of overall bed roughness in channels (see, for
example, Chow 1959) however they incorporate many channel and hydraulic
characteristics in addition to the local roughness of the bed surface, so local shear stress


address both of these issues, the water depth in the flume was increased to 30.5 cm (12
inches) for the extended erosion tests. The increase in water depth caused the speeds
calculated for a given shear stress to increase (See Equations 6.1 and 6.3). and the
increased volume in the flume reduces the concentration of suspended solids for any
given cumulative mass eroded.
Single Shear Extended Run Series 2 consisted of three extended runs. In the first
run the bed was allowed to consolidate for 7 days, then shear was applied at a constant
calculated value of 2.0 dynes/cm2 for a period of 24 hours. The results, presented in
Figure 6.5, show an exponentially decreasing erosion extent with time that approaches an
asymptotic maximum CSME of about 0.0172 gm/cm2.
Organic Sediment-Extended Run
30.5 cm. Water Depth, Single Shear Stress of 2.0 dyne/cm2
Mid-depth Sample
jf "
"¡OOP

(j
0.000
0 5 10 1
Time hr
20 25
Figure 6.5: CSME as Function of Shear Time at Constant Calculated Shear Rate of 2.0
dynes/cm2 for Consolidated Organic Sediment Bed with 30.5 cm Water Depth
The data are correlated well (r2=0.99) with an equation that contains a constant
term and two decaying exponential terms, as follows:


385
11 2.06
30.0
10 1.33
95 95 1.83 1.70 17.00 17.00
MU 1.000 1.000 0.830 0.830 30.1
STRU 92 392 92 93 11 1.93 0.48 2.54
MU 1.000 1.000 0.830 0.830 27.0
STRU 102 402 95 101
MU 1.000 1.000 0.830 0.830
SECT 101 101 101 102
W 180.0 0.0
H 0.0000 .83000 1.6000
BS 10.740 10.740 15.140
SECT 193 193 193
W 270.0 0.0
H 0.0000.45000 1.3500
BS1 6.7100 6.7100 11.630
BS2 1.8100 1.8100 8.7600
BB1 9.3100 9.3100 13.300
BB2 6.8300 6.8300 8.7600
SECT 199 199 199 100
W 180.0 0.0
H 0.0000.83000 1.6000
BS 7.0900 7.0900 10.400
BB1 9.6300 9.6300 11.020
BB2 10.190 10.190 13.080
0.91 3.07
1.33 500.00 500.00
93 1.45 1.39 17.00 17.00


145
Figure 6.19: Velocity Profile Measurements Reported by Mehta (1973) in CRAF with
22.9 cm Water Depth


116
tests was simply to evaluate, at the reconnaissance level, the deposition performance of
the prototype sediment, not to obtain detailed deposition modeling data.
The CSME of the system increased from 0.0158 gm/cm2 to 0.0174 gm/cm2
during the extra 24 hour period. This is judged to be attributable to at least two factors.
First, continued turbulent shear stress on the bed after most of the erodable material is
resuspended may act to modify the interparticle interactions at the surface of the
sediment, allowing more material to slowly become available for resuspension. The
simple asymptotic model used in Equation 6.5 does not consider this effect. Second, if
the suspended solids mass loading is a balance between erosion and deposition, any
decrease in deposition rate may serve to increase the overall system mass loading. It is
quite likely that the imposition of the 2.0 dynes/cm2 shear stress for an additional 24
hours may have caused some suspended particle attrition, that would effectively reduce
the average settling rate and reduce the overall deposition rate. This modification of the
system would be expected to introduce some bias into the test results.
With the above mentioned caveat in mind let us examine the deposition
performance, shown in Figure 6.9.
The same type of curve fit was done for the deposition runs as was done for the
erosion runs using the equation
CSME = CSME0-a(l6.8
where a, k, t and CSMEo have the same definitions in Equation 6.5. Table 6.2
lists the constants for each shear level.
The prime point of interest with this analysis is the asymptotic deposition CSME
achieved at each shear level as a function of shear level. This is shown in Figure 6.10
along with the corresponding curve for asymptotic CSME for erosion.


319
Reduction of Transportable Mass in the System
This is the most obvious remedial action, but one which is not necessarily easy to
achieve economically. The effect of reduction of transportable mass is most simply-
simulated by reducing the value of Cem, the initial concentration of transportable mass.
Figure 11.1 shows the effect of this reduction of the cumulative mass exported at 100.000
m3 of discharge, where the points are simulation results and the line is a linear best-fit
through the simulation points.
Suspended Solids Export Simulation for Event UF9200-220 at Various
Assumed Values of Cemccji
12.000
oo
ft
10,000
3
-§ 8.000 0
C/3 O
& 6,000
£
g 4,000 -
t u
| 2,000 -O
ui
0 o
0 200 400 600 800 1000 1200 1400 1600 1800
Cem(o) gm/m"
Figure 11.1: Effect of Transportable Mass on Exported Suspended Solids for
Event UF9200-220
Note that the response to mass reduction is approximately linear in the range
evaluated, implying that a system-wide reduction of transportable mass by a given
fraction should result in that same reduction in exported solids. It should be emphasized
that these simulations were done with a uniform reduction across the entire system, not


CHAPTER 1
INTRODUCTION, BACKGROUND, AND RESEARCH MOTIVATION
Historical Summary
The South Florida Everglades is a natural resource unique in the United States, if
not in the world. The original Everglades watershed (See Fig. 1-1) was primarily a broad,
freshwater marsh that extended from what is now the Kissimmee River basin, through
Lake Okeechobee to the southern tip of the Florida peninsula, encompassing some 7500
square miles (-20,000 square km) of upland and wetland territory. The natural flow,
arising from a shallow land slope of only a few cm/km, was at very low velocity from
north to south along a riverbed which was often 80-100 km wide, with ultimate discharge
into the waters surrounding the southern tip of Florida (Jones, 1948).
Flow was episodic and seasonal in nature, with Lake Okeechobee acting as a
buffering reservoir. During the wet season (May through September) the level of Lake
Okeechobee would rise above its shallow southern bank heights, providing a consistent
flow of water through the wide riverbed. In the dry season the level of Lake Okeechobee
would fall below the level of its southern banks, essentially shutting off the flow of water
from the Kissimmee and Okeechobee basins to the land farther south. Under these
conditions of reduced inflow and precipitation the marshlands south of Lake Okeechobee
would slowly dry until, in the spring season, they would be subject to frequent and
extensive fires. The onset of the wet season in the late spring would quench the fires and
complete the annual wet-dry cycle (Douglas, 1988).
1


129
V2rr
Rkw
622
where
V = Average channel velocity, cm/sec
n = Mannings coefficient of roughness
Substituting Equation 6.22 into Equation 6.20 gives the result
. VWpg
622,
Substituting Equation 6.23 into Equation 6.19 gives
CSME,
h { V2
624
where
Velocity that corresponds to iCJ, cm/sec
Note that Equation 6.24 refers only to the CSME of the material in class j. The
measured asymptotic CSME at the /th shear stress would include the sum of all erodable
classes up to and including class j, or
CSME =
£pgn2V2
R,'12
f V2
1 --v
V2
W, epn2V2{,
h
C l
F2)
6.25
At this point let us consider the case of difference of roughness between the wall
and the bed. The limiting case where the wall resistance is negligible compared to the


396
Hutcheon Engineers. 1995. Sediment Control Demonstration Project Summary Report".
Report submitted to the Everglades Agricultural Area Environmental Protection
District, West Palm Beach, FL
Hwang,K. 1989. Erodability of Fine Sediment in Wave-dominated Environments. M.S.
Thesis, University of Florida, Department of Coastal and Oceanographic
Engineering, Gainesville. Florida
ICIM (Informtica Centrum voor Infrastructura en Milieu), 1992. DUFLOW-A Micro-
Computer Package for the Simulation of One-Dimensional Unsteady Flow and
Water Quality in Open Channel Systems, Bureau Icim, Rijswijk, The Netherlands
Ivanoff, D.B., 1994. Mineralization of Organic Phosphorus in Everglades Histosols
Under Aerobic and Anaerobic Conditions M.S. Thesis, University of Florida.
Gainesville, FI.
Izuno, F.T. 1994 Chapter 6-Physical Components and Water management in the Lake
Okeechobee-EAA-WCA-ENP System, In Everglades Agricultural Area (EAA)-
Water. Soil, Crop, and Environmental Management. A.B. Bottcher and F.T.
Izuno. Eds., University Press of Florida, Gainesville, FL
Izuno, F.T. 1995. Implementation and Verification of BMPs for Reducing P Loading in
the EAA, Phase III Final Report to EAA Environmental Protection District,
Institute of Food and Agricultural Sciences. University of Florida Everglades
Research and Education Station, Belle Glade, Florida
Izuno, F.T., and Bottcher, A.B. 1994 Chapter 2-The History of Water Management in
South Florida In Everglades Agricultural Area (EAA)-Water, Soil, Crop, and
Environmental Management. A.B. Bottcher and F.T. Izuno, Eds., University Press
of Florida, Gainesville, FL
Izuno.F.T., Sanchez.C.A., Coale.F.J., Bottcher.A.B., and Jones.D.B. 1991 Phosphorus
Concentrations in Drainage Water in the Everglades Agricultural Area, J. Env.
Oual.. 20:608-619
Jones,L.A.. 1948. Soils, Geology and Water Control in the Everglades Region,
Agricultural Experiment Station Bulletin No. 442, University of Florida,
Gainesville, FL
Kemp.W.M., Boynton,W.R., Twilley.R.R.. Stevenson,J.C., and Ward.L.G. 1984.
Influences of Submerged Vascular Plants on Ecological Processes in Upper
Chesapeake Bay. In The Estuary as a Filter, Academic Press. Orlando


82
The WPBC 75-150 micrometer range is also the particle size range that had the
significant reduction in organic content so it might be surmised that this particle size
range contained a concentration of high phosphorus-containing inorganic matter, perhaps
shells and shell fragments. With the exception of the 75-150 micrometer WPBC fraction,
the remainder of the fractions ranged in phosphorus content from 550 to 1150 mg/kg.
The less-than-38-micrometer fractions of all three samples were very similar in
phosphorus content, 1145 mg/kg for the WPBC sediment, 945 mg/kg for the B9B10
drainage ditch sediment, and 915 mg/kg for the hydrated B9B10 soil.
Phosphorus fractionation results
The results of the phosphorus fractionation analyses are shown in Table 5.5. The
WPBC 75-150p sample phosphorus content was found be almost entirely acid
hydrolyzable phosphorus not readily available to bicarbonate extraction. Only 0.2% of
the total phosphorus was present in the bicarbonate extractable form.
The refractory organic, fraction-was so small that it was lost in the experimental
error between the sum-ofthe bicarbonate and HC1 extractable forms and the total
phosphorus analysis. The WPBC <38p fraction showed a slight increase in the
bicarbonate extractable form to 1.3% of total phosphorus.('38.0% of this fractions total
phosphorus was acid hydrolyzable. The B9B10 fractions were similar in distribution with
6.5-8.9% bicarbonate extractable form, and 36.3-38.0% acid hydrolyzable form.
The fractions of phosphorus present in a particular form were related to the
organic matter content of the sample, as indicated in Figure 5.8. Recognizing the
minimum number of data points, it is still reasonable to infer that the refractory and acid
extractable forms were a linear function of organic matfei; content, while the labile
(bicarbonate extractable) form appeared to be a mildly exponential function of organic
matter content.


56
2. It should be capable of sufficiently detailed time scales to allow simulation
with reasonable accuracy of transients that exist at pump start-up and shut
down, however modeling of rapid transients, such as hydraulic jumps, is
not necessary because of the level terrain of the EAA.
3. It should model with reasonable accuracy particulate phosphorus
mobilization, transport, and deposition as a function of some readily
defined hydraulic parameter such as shear stress or average stream
velocity.
¡4. Phosphorus interchange between soluble and particulate forms by
adsorption-desorption should be represented in the model.
5. The dynamic impact of biological growth and senescence in the aqueous
system on phosphorus transport should be incorporated into the model.
Application of minimum criteria and resulting conceptual model simplification
Water chemistry reactions resulting from interflow through the substratum were
deferred for later consideration. Interactions between oxic and anoxic sections of the
base sediment can be very sensitive to redox potential profiles and tend to be important
over long term (multi-annum) time scales. The field portion of this study was intended to
last through only one wet season. It was decided to defer this portion of the study to a
later program.
When these minimum criteria were applied and the program was evaluated in
light of prior knowledge, available resources, and time constraints, it became necessary to
affect a truncation of the conceptual model to a simpler form.
Particulate filtration and resuspension arising from groundwater flow
perpendicular to ditch and canal banks may be a meaningful contributor to sediment flux
but experimental determination of field-relevant parameters could require considerable


339
Correlation of Discharge TSS Concentration with Canal Level for UF9206S
Figure 11.15: Correlation of TSS Concentration with Corrected Canal Level at
UF9206S
Correlation of Discharge TSS Concentration with Canal Depth for UF9200
Figure 11.16: Correlation of TSS Concentration with Canal Depth at UF9200


TABLE OF CONTENTS
ACKNOWLEDGMENTS iv
LIST OF TABLES xi
LIST OF FIGURES xiii
NOMENCLATURE xxii
ABSTRACT xxvii
CHAPTERS
1 INTRODUCTION. BACKGROUND. AND RESEARCH
MOTIVATION 1
Historical Summary 1
Sources of Phosphorus in the EAA 5
Proposed Regulatory Remediation 7
Areas of Uncertainty in Remediation Plan 9
EAA Modeling Activity 10
2 PRIOR DATA RELEVANT TO PARTICULATE PHOSPHORUS
TRANSPORT IN THE EVERGLADES REGION 12
Introduction 12
Farm-Field Scale Studies 13
CH2M Hill study 13
Izuno, et al study 15
Canal Transport Studies 16
Southeast Florida canals water quality studies 16
Water Conservation Area water quality correlations 17
EAA Canal Sediment Studies 17
Anderson and Hutcheon Engineers study (1992) 18
Andreis/U.S. Sugar proposed sediment control BMPs (1993) 19
Sediment control demonstration project 20
South Florida Water Management District Canal Data Set 22
3 GENERAL LITERATURE REFERENCES RELEVANT TO
PARTICULATE PHOSPHORUS TRANSPORT 24
vi


325
simulations relax that constraint and restore the section flow width to its actual physical
width. This might be accomplished with, for example, weir arrangements upstream of the
structures. Figure 11.5 shows the results of Simulation B. which incorporated the
pumping modifications included in Simulation A.
Figure 11.5: Effect of Pumping Modification and Flow Width Modification
(Simulation B)
Flere the simulated solids export at 100,000 m3 discharge was reduced to 1935 kg
from the original 10,420 kg. The principal contribution to the additional reduction
beyond that achieved in Simulation A was the reduction of the high solids discharge at
pump start-up by increasing the flow crossectional area and thus decreasing the start-up
velocities.
Simulation C maintains the flow width increase and the substitution of the small
pump operation for the large pump cycling, but eliminates the incorporation of a steeper
pumping curve on the large pump. Figure 11.6 shows the effects of this modification.
The end result is a higher cumulative solids discharge, 2395 kg, than Simulation B, but it


Event UF9200-319
CUM PP Load
Figure 10.28: Cumulative Particulate Phosphorus Load Simulation Results for
Event UF9200-319
Event UF9200-355


231
clusters, but it probably also reflects the condition that plants in the clusters tended to
look larger and older than the free floating plants. This may help explain the increased
level of dislodgable detritus in the cluster samples.
Table 8.5: Water Lettuce Dislodgable Detritus Study
Sample Source
Remaining Plant
Dry Mass-gm
Detritus Dry
Mass-gm
Fraction of Total
Mass Dislodged
Remaining Plant
Phosphorus
Content-mg/kg
Detritus
Phosphorus
Content-mg/kg
Cluster
20.18
10.46
0.342
5387
2526
it
21.45
7.84
0.268
5513
2725
17.59
14.30
0.449
6117
2858
24.72
18.27
0.425
6581
2804
28.79
15.08
0.344
5919
2564

21.08
17.07
0.447
5775
3004
Avg (SD)
22.3(3.9)
13.83(4.0)
0.379(0.073)
5882(395)
2747(181)
Free-Floating
12.7
5.51
0.303
6319
2639

15.51
5.93
0.277
6131
2436
7.56
2.57
0.254
7282
2496

11.42
6.17
0.351
6351
2350
Avg (SD)
11.80(3.3)
5.05(1.7)
0.296(0.042)
6521(517)
2480(122)
The phosphorus content analyses showed interesting results. The detritus had a
significantly lower phosphorus content than the remaining plant. The population average
plant phosphorus content remaining after agitation was 6137 mg/kg, the population
average detritus phosphorus content was 2640 mg/kg, about 43% of the plant content.


APPENDIX B
PHYSICAL ANALYSES
Total Suspended Solids (APHA, 1995 Method 2540D)
Weigh ISCO sample containers and then filter contents under vacuum through
pre-dried, pre-weighed glass microfiber filter (Whatman 934AH) which had received acid
wash and rinse described in Chapter 7. Rinse remaining solids from container onto filter
with de-ionized water. Dry filter pad at 103-105 C for 1 hour. Cool in dessicator and
weigh dried filter and residue.
Calculation:
(A-B)x 1,000
mg Total Suspended Solids / L = 4-- B. 1
(C-D)
where:
A = mass filter + dried residue, mg
B = mass filter, mg
C = mass sample + container, gm
D = mass container, gm
Dry Solids (APHA, 1995 Method 2540B)
Weigh a wet or moist sample onto a pre-weighed glass or aluminum evaporating
pan and dry to constant weight at 103-105 C.
Calculation:
% Dry Solids = i£~S)x W0 g 2
(A-B)
where:
A = mass original sample + pan, mg
359


CHAPTER 4
INITIAL HYPOTHESES, OBJECTIVES, AND RESEARCH PLANS
Problem Overview
At this point a reiteration of the primary issues and objectives of the overall
program is useful to place the objectives of the research reported in this document in
proper context.
Summary of sources
The overall objective of the program funded by the EAA EPD is to develop
methods of irrigation and drainage control for the farms of the EAA that minimize the
quantities of phosphorus exported off-farm while not causing material detrimental effects
to the crops grown on-farm. The primary on-farm sources of soluble phosphorus (SP)
and particulate phosphorus (PP) are considered to be soil mineralization (SP), fertilizer
application (SP and PP), and soil/litter/sediment mobilization (PP), where the term
sediment is taken in the most general sense to designate any water-resident particulate
material.
Organic soil subsidence can result from microbially mediated oxidation of the
organic matter in the soil. The oxidation destroys soil structure by converting soil organic
mass to soluble organic and inorganic matter and CO2. The process also causes
conversion of the nutrients incorporated in the organic matter to their inorganic form, for
example conversion of organic phosphorus to soluble orthophosphorus. Soil subsidence
may be retarded by reducing the oxidation-reduction potential of the soil to anoxic or
51


147
some weeds (Chow 1959) that was felt to best describe the prototype organic sediment
bed. For this value of n the approximate model gave a value of 0.29 for the ratio of the
pseudo-mean velocity to the ring-channel AV. This value was chosen as the multiplier of
the ring-channel AV to obtain a characteristic velocity that was assumed to approximate
the CRAFs equivalent open channel mean velocity when operating with organic
sediment.
To summarize now, the calibration procedure is as follows:
1. Pick a PES RPM and calculate the CSME produced using Equation
6.28.
2. Using Equation 6.27 calculate an equivalent CRAF ring-channel
AV for the flume, using the PES depth of 12.7 cm as the value for
h.
3.Multiply the calculated AV by 0.29 to obtain the estimated
equivalent open channel mean velocity for erosion.
Figure 6.21: Final Calibration Curve of Estimated Equivalent Open Channel Mean
Velocity For Erosion as a Function of PES Oscillation Frequency


123
Figure 6.12, which shows the respective responses at the two water depths, clearly
illustrates the significant differences between the CSME versus calculated shear stress
relationships. The calculated shear stresses applied to the bed at the 15.2 cm water depth
produced significantly less asymptotic cumulative specific mass eroded than the
equivalent calculated shear stresses applied to the same bed at 30.5 cm water depth. The
two lowest calculated shear stresses applied to the bed at 15.2 cm water depth appeared to
produce essentially no cumulative erosion. The small amount of suspended sediment
detected at the low shear stresses was most likely due to a minor amount of non-
flocculant debris that had accumulated over the seven weeks of testing with this prototype
organic sediment bed.
Figure 6.12: Asymptotic CSME vs.. Applied Shear Stress for Prototype Organic
Sediment Extended Run Erosion with 30.5 cm and 15.2 cm Water Depth
These results were in opposition to the observations made during previous runs.
The results of the 6.7 and 8.7 day consolidation tests (See Figure 6.6), the single shear
level of 2.0 dynes/cm2 and the same shear level in the multiple shear level tests (See
Figures 6.5 and 6.6) and the erosion-deposition comparison (See Figure 6.10) all showed


195
Time Series-Flow and Total Suspended Solids
Event UF9206S-317
Julian Date
Pumping Rate O TSS Concentration
Figure 7.31: Time Series-Flow and Total Suspended Solids Event UF9206S-317
Time Series-Flow and Total Suspended Solids
Event UF9206S-336
Figure 7.32: Time Series-Flow and Total Suspended Solids Event UF9206S-336


CHAPTER 3
GENERAL LITERATURE REFERENCES RELEVANT TO PARTICULATE
PHOSPHORUS TRANSPORT
Introduction
This chapter will focus on information in the open literature which may have
particular influence on the direction of the research, which may provide insight for
formulation of hypotheses presented later in this work, or which may be used to
rationalize later conclusions. The chapter begins with a discussion of the sources of
particulate phosphorus, continues with a discussion of diagenesis, or the fate of deposited
particulate matter, and concludes with a brief review of the concepts of sediment
transport, or the remobilization of deposited particulate matter.
Sources of Particulate Phosphorus
Particulate matter in ecosystems is typically classified as either allochthonous (of
external origin) or autochthonous (of internal origin). Allochthonous materials which
may be of importance in the EAA might include soil particles, livestock wastes, and
particulate agricultural chemicals transported by wind or water erosion, stream bank
eroded particles, leaf and ground cover litter transported by wind or water erosion, and
deposited detritus from streamside trees and plants. Autochthonous materials would
include aquatic macrophytes and aquatic microbes, both of which can fix soluble carbon
and nutrients in their growth and reproduction processes, streambed parent inorganic
particles, chemically precipitated inorganic materials, and invertebrate and vertebrate
organisms. Given the minimal terrain slopes, sub-tropical climate, high insolation and
24


165
of the first lapse. By Julian Day 220 the pond was full and the grower resumed off-farm
pumping.
Table 7.2: Summary of Pumping Events for Target Farm UF9200
Event
Duration
(hrs)
Pump Time
(hrs)
Interevent
Time
(hrs)
Pumped
Volume
(m3)
Suspended
Solids Load
(kg)
Particulate
P Load
(kg)
154
68
56
Start
3.56 x 105
2262
8.4
158
47
46
25
0.93 x 105
NA
NA
161
52
40
25
0.74 x 105
NA
NA
169
35
34
150
0.76 x 105
295
NA
220
172
134
1180
1.99 x 105
11031
23.9
237
114
100
243
3.18 x 105
4368
19.5
244
22
16
53
0.39 x 10!
1132
3.9
252
28
19
165
0.55 x 10s
1315
5.8
258
39
34
120
0.89 x 10s
1146
6.2
262
56
56
53
1.45 x I05
861
6.7
285
95
90
498
3.21 x 10s
6280
26.5
319
286
179
722
7.34 x 105
14262
77.4
336
323
142
126
2.45 x 105
4256
26.3
355
222
142
123
4.56 x 105
10344
35.9


289
Cumulative suspended solids load and narticulate phosphorus load simulation results
The quantitative performance of the model may be observed by examining the
simulated and observed cumulative total suspended solids and particulate phosphorus
loads through the course of each event. These cumulative loads are shown, paired for
each event, in Figures 10.9-10.20. Presented in this mode it is evident that the model
calibration has calculated the delivered suspended solids and particulate phosphorus
reasonably well. For the periods of calculation, the observed total delivered suspended
solids load was 24,762 kg, the simulated total load over the same periods was 25,700 kg
or 103.4% of the observed total load. The observed total particulate phosphorus load was
86.88 kg versus a simulated total load of 80.14 kg, or 92.2% of the observed total load.
The calibration based on fitting the suspended solids concentrations overpredicted
the observed suspended solids loads by 3.4% and underpredicted the particulate
phosphorus loads by 7.8%. Extension of the TSS-PP correlation developed for the first
four hours of pumping time to the first six hours would have brought the particulate
phosphorus simulation close to 100% of the observed value, but such an extension was
not really justifiable with the data in hand. Examination of the cumulative particulate
phosphorus load curves indicates that the preponderance of the deviation seems to occur
at the end of Events 220, 262, and 285 during times when the large pump was going
through on-off cycles. It is possible that the acceleration/deceleration that occurs during
these times may add a hydrodynamic component that causes an enrichment of the
particulate phosphorus content above the correlation values, for example by dislodging
detritus that is in an earlier stage of diagenesis. In any event the simulation particulate
phosphorus results were quite acceptable for the first approximation level model.


107
shear stress is increased to the next level by appropriate adjustment of ring and channel
speed and the sampling process is repeated. Samples are analyzed for total suspended
solids (TSS) by the method noted in Appendix B.
If dynamic deposition data are desired the process is run in reverse, where an
initially high shear stress is applied to suspend sediment, and then the shear is
progressively reduced and sampling is conducted to determine deposition parameters.
The raw data are in the form of concentration time series. Comparison among
devices, and ultimately application to the field, requires that the data be normalized. This
is done by multiplying the concentration at any time by the water volume in the flume to
get a cumulative mass of sediment eroded up to that time, then dividing by the sediment
bed area to get a specific mass per unit area. The resulting normalized value will be
referred to in this text as the Cumulative Specific Mass Eroded, or CSME. and has the
units of gm/cm2.
Organic sediment flume studies
The organic sediment flume studies were conducted in three series. Series 1 was a series
of range-finding tests run with placed and deposited beds at a water depth of 15.2 cm (6
inches), Series 2 was a series of extended runs at shear stresses chosen in consideration of
the results of the first series and anticipated field conditions run at a water depth of 30.5
cm (12 inches), and Series 3 was an extended series run at the same apparent shear
stresses as Series 2, but at a depth of 15.2 cm (6 inches) to investigate depth effects...
Series 1 Ranee-Finding Tests The first test of Series 1 was with a placed
organic bed. This bed was ultimately to be the source of all the ensuing deposited bed
organic sediment tests but it was initially run as a placed bed in an attempt to simulate a
base line that represented a deposited bed with extremely long consolidation time. The
bed was placed in the flume at a depth of 5.1 cm (2 inches) and leveled with a T-shaped


66
67
68
64
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
193
199
375
1006 792 8090E+01 1.00
1006 12 8090E+01 1.00
1006 0 OE+OO 1.00
805 1609 OE+OO 1.00
805 817 1619E+02 1.00
805 805 OE+OO 1.00
805 792 8090E+01 1.00
805 12 8090E+01 1.00
805 0 OE+OO 1.00
794 805 OE+OO 1.00
603 1609 OE+OO 1.00
603 817 1619E+02 1.00
603 805 OE+OO 1.00
603 792 8090E+01 1.00
603 12 8090E+01 1.00
603 0 OE+OO 1.00
402 1609 OE+OO 1.00
402 817 1619E+02 1.00
402 805 OE+OO 1.00
402 792 8090E+01 1.00
402 12 8090E+01 1.00
402 0 OE+OO 1.00
201 1609 OE+OO 1.00
201 817 2428E+02 1.00
201 805 OE+OO 1.00
201 792 1214E+02 1.00
201 12 1214E+02 1.00
201 0 OE+OO 1.00
11 805 OE+OO 1.00
11 0 OE+OO 1.00
0 805 OE+OO 1.00
0 684 OE+OO 1.00
0 522 OE+OO 1.00
0 354 OE+OO 1.00
0 94 OE+OO 1.00
0 0 OE+OO 1.00
0 -10 OE+OO 1.00
106 0 OE+OO 1.00
0 187 OE+OO 1.00


ACKNOWLEDGMENTS
The author would like to express his appreciation to his major professor. Dr. K.L.
Campbell of the Department of Agricultural and Biological Engineering, for his support
throughout this effort. He would also like to acknowledge the other members of his
committee: Dr. J. J. Delfino of Environmental Engineering Sciences, Dr. F. T. Izuno of
Agricultural and Biological Engineering, Everglades Research and Education Center. Dr.
A. J. Mehta of Coastal and Oceanographic Engineering, Dr. K. R. Reddy of Soil and
Water Science, and Dr. A. B. Bottcher, President of Soil and Water Engineering
Technology, special committee member.
Thanks are due to the Departments of Agricultural and Biological Engineering,
Coastal and Oceanographic Engineering, and Soil and Water Science, and the Everglades
Research and Education Center, all of which contributed laboratory space and equipment
to this interdisciplinary study.
The author is indebted to many individuals who provided invaluable assistance
during the course of this work: Ms. Yu Wang and Dr. Winston Davis for their analytical
help, Ray Garcia and Curtis Miller for their patient advice and essential help in the field,
Laurene Capone for helping get the resources and smoothing the path so often, Mauricio
Abreu for his many long hours working on the erosion studies, and Dr. Nigel Pickering
for his considerable computer skills.
Financial support is gratefully acknowledged from the U. S. Department of
Agriculture, which provided the author with a Research Fellowship in Hydrologic
Science while this research was in progress, and from the Everglades Agricultural Area
Environmental Protection District, which provided the operating funds for this research as
iv


377
* DUFLOW data file :220\STN 17.NET
* Network data program version: 2.02
*
SECT 1 1 1 2 792 2.11 2.11 17.00 17.00
W 180.0 0.0
H 0.0000.93600
BS 3.1480 3.1480
SECT 3 3 3 9 201 1.57 1.48 17.00 17.00
W 270.0 0.0
H 0.0000.71000 1.4780
BS1 7.1380 7.1380 9.8110
BS2 7.3670 7.3670 10.390
SECT 4 4 4 5 780 1.91 1.91 17.00 17.00
W 180.0 0.0
H 0.0000 1.1400
BS 3.3530 3.3530
SECT 6 6 6 12 201 1.66 1.71 17.00 17.00
W 270.0 0.0
H 0.0000.61900 1.2890 1.3870
BS1 9.9480 9.9480 13.347 13.539
BS2 7.2420 7.2420 11.472 11.725
SECT 7 7 7 8 792 1.65 1.65 17.00 17.00
W 180.0 0.0
H 0.0000 1.3930
BS 3.9100 3.9100
SECT 9 9 9 15 201 1.48 1.89 17.00 17.00
W 270.0 0.0
H 0.0000.79500 1.5640
BS1 7.3670 7.3670 10.390
BS2 6.9090 6.9100 10.271
SECT 10 10 10 11 780 1.70 1.70 17.00 17.00
W 180.0 0.0
H 0.0000 1.3440
BS 3.6570 3.6570
SECT 12 12 12 18 201 1.71 1.51 17.00 17.00
W 270.0 0.0
H 0.0000.56700 1.1800 1.3354
BS1 7.2420 7.2420 11.472 11.725
BS2 9.0060 9.0060 13.191 13.200
SECT 13 13 13 14 792 1.91 1.91 17.00 17.00
W 180.0 0.0
H 0.0000 1.1400
BS 3.1487 3.1487
SECT 15 15 15 21 201 1.87 1.84 17.00 17.00
W 270.0 0.0
H 0.0000.40500 1.1730


114
'd( CSME)
dt
Figure 6.7 shows the relationship between initial erosion rate and shear stress. It
is risky to infer a pattern from four data points, but it appears that, for this particular
sediment, there is a maximum in the initial erosion rate around a shear stress of 1.3
dynes/cm2, followed by a decrease in initial erosion rate as shear is increased. This
suggests a nonlinear distribution of interparticle forces through the sediment particle
population, which is not surprising given the morphology of the sediment that arises from
organic material in various stages of diagenesis.
Initial Erosion Rates vs Shear Stress for Organic Sediment Extended Run
0.0030
0.0025
"e
| 0.0020
5b
£ 0.0015
c
.o
£ 0.0010
0.0005
0.0000
0 0.5
1 1.5 2
Shear Stress-dynes/cm2
2.5
Figure 6.7: Initial Erosion Rates vs. Applied Shear Stress for Organic Sediment
Extended Run
The evaluation of asymptotic CSME as a function of shear stress, shown in Figure
6.8, presents a more clear-cut correlation. The relationship between asymptotic CSME
and shear stress is correlated well (r2 = 0.998) by the equation


267
Erosion studies found that seven of eight sampled locations at UF9200 had
far less capability to deliver suspended solids under typical flow
conditions than was required to account for the measured suspended solids
loads. A similar distribution of erosion characteristics was found at
UF9206 (Chap. 6).
The intensive synoptic suspended solids studies indicated that the field
ditches were contributing little to the particulate solids loading of the farm
discharge. This observation was supported by the results of the erosion
studies that showed that velocities in the ditches were too low to cause
erosion of the surficial sediment until late in the events, long after the bulk
of the suspended solids load had been discharged (Chap.7 and Chap. 6).
This body of evidence led to the conclusion that the original hypotheses must be
modified. The surficial sediment, as sampled, would not be considered as the primary
particulate phosphorus source. The primary source of exported particulate phosphorus
for the target farm, UF9200, would be considered to be the biological growth in the
canals. The sources are hypothesized to be a combination of detrital deposit and
dislodgable detritus from floating macrophytes, suspended and bottom growing
macrophytes and algae, and settled detritus.
Modified Model Development Plan
The modification in the basic hypothesis required a modification in the model
development plan. The diversity of the hypothesized sources adds considerable
complexity to the tasks of quantifying erosion characteristics. The hypothesis of
hydraulically dislodgable detritus contributing to particulate phosphorus export adds a
new dimension to the concept of erosion, which now may take place at the waters


360
B = mass pan, mg
C = mass dry sample + pan. mg
Fixed and Volatile Solids Ignited at 550 C (APHA, 1995 Method 2540F.)
Weigh sample which has been previously dried by Method 2540B into pre
weighed porcelain crucible or high silica glass beaker. Ignite in muffle furnace for one
hour at 550C (plus heat-up and cool-down time). Remove, cool in dessicator, and weigh
container with ignited residue.
Calculation:
(C-B)x 100
% Fixed Matter (Ash) = ----
(A-B)
% Volatile Matter (Organic) = 100
where:
, (C-B)
. (-B)
B3
BA
A = mass original sample + container, mg
B = mass container, mg
C = mass ignited sample + container, mg
Bulk Density (ASTM, 1993 Designation D4531-86)
Sediment is weighed into pre-weighed 50 ml graduated cylinder to the 50 ml
graduation mark. Combined sample and cylinder are weighed.
Calculation:
Bulk Density, gm / ml = ^ ^ B5
where:
A = mass original sample + cylinder, gm
B = mass cylinder, gm
Specific Gravity (ASTM, 1993 Designation D854-92 [Modified])
ASTM 854-92 calls for addition of distilled water to a known mass of dried solids
in a pycnometer. Most of the materials of interest were already in contact with native


271
Fn = Net phosphorus flux, mass/unit area
Fe Erosive phosphorus flux, mass/unit area
Fs = Sedimentation phosphorus flux, mass/unit area
The erosive flux is defined by the equation
FE=a(x-xc)CEU 92
where
x = Shear stress applied to eroding area, force/area
xc = Critical shear stress for erosion, force/area
Cem = Surface concentration of erodable material,
mass/area
a = Erosion constant, area/force-time
Now by assumption, shear stress is proportional to mean velocity squared or
t = PV2 93
where
V = Channel mean velocity, length/time
P = Proportionality constant, force-time2/length4
Substituting Equation 9.3 into 9.2 gives
FE=e(V2-Vc1)CEU
9.4


296
Small flows
At this point it is necessary to discuss the model's limitation with respect to small
flows because assumptions made to accommodate this limitation have important effects
on the remainder of the calibration process and on the validation process. It was
mentioned previously that the model does not predict transport well at the low flows and
velocities prevalent when the small pump was in operation. Attempts to incorporate this
flow regime by reducing the critical velocity for transport resulted in unacceptable
increases in the predicted transport during periods of lower velocity when the large pump
was running.
It was also noted that this is an inherent limitation in a model that is simulating a
mixed population with a single set of erosion parameters. There is a sub-set of the
erodable material population that has a much lower critical velocity for transport than the
average material for which the model was calibrated. This material is transported at
velocities that prevailed when the small pump was running, but was not simulated
because of the all-or-nothing effect of falling below the critical velocity in the simulation.
At present there is insufficient information from the studies done in this work to
estimate the amount of this material present in the system at any specific time, but
cumulative event analyses showed that the amount of suspended solid mass transported
during the small-pump portions of the wet season events was less than 10% for the four
largest events and less than 20% for the two smallest events. For the purposes of this
modeling effort the assumption is made that this material is negligible and that the
operation of the small pump has no impact on movement of erodable mass. It is
recognized that this is a very strong assumption but from a pragmatic standpoint it is
justifiable at this point to ignore phenomena that contribute small fractions to the
transport load. One direct application of this assumption is that the operation of the small
pump is excluded from the calculation of interevent time from here on.


78
fraction is shown for the drainage ditch sediment to include the effect of the lyngbya
which constituted approximately 50% of this fraction.
Figure 5.3: Particle Size Distribution of Hydrated B9B10 Soil, B9B10 Sediment, and
WPBC Sediment
The organic content of the drainage ditch sediment was quite constant across all
particle size fractions, varying between 76-80%, with an average of 79.1%. The organic
content of the hydrated field soil was lower than the drainage ditch sediment, averaging
60.6% and varying from 40% to 69%. Note in Figure 5.4 the minimum at particle size
fraction 75-150 micrometers. The organic content of the WPBC sediment was much
lower, averaging 21.6% and ranging from 7% to 43%. Note the drop in organic content
of the 75-150 and 150-300 micrometer fractions, similar to that observed in the hydrated
soil.


19
There was not sufficient detail of the channel configuration presented in the report
to do a substantive evaluation of transport dynamics, but there appeared to be a step
change in suspended solids concentration at a velocity somewhere between 0.06 m/sec
and 0.1B m/sec, above 0.18 m/sec there was no clear response to increasing velocity.
Included in the report were data tables representing excerpts of data from a
SFWMD sampling program which covered a nine year time span from 1983 to 1991.
The data summary presented averages of TSS = 23 mg/1, ortho-P = 0.080 mg/1, total P =
0.161 mg/1, and particulate P = 0.081 mg/1. The implication from this summary was that
the distribution of phosphorus in the SFWMD canals was split approximately 50%/50%
between soluble and particulate forms during this period. This data set is discussed in
more detail later in this section.
Andreis/U.S. Sugar Proposed sediment control BMPs (19931
The Research and Development Department of U. S. Sugar Corporation, in an
attempt to develop a particulate phosphorus control program for the EAA, informally
proposed a number of sediment-control BMPs to the grower community and the SFWMD
in 1993 (Andreis 1993). Although these practices were not proposed in formal
publications, they have been given widespread distribution, and have received varying
levels of acceptance within the grower community. They were discussed with H. J.
Andreis, Senior Vice President of Research for U.S. Sugar, by this writer in 1994. At the
time of discussion some field data were available on the effectiveness of the BMPs.
There were some sixteen BMPs recommended, which fall into five categories, as
follows:
1. Surface soil erosion reduction,
2. Stabilization of ditch and canal banks and sidewalls,
3. Minimization of disturbance of in-canal sediments,


164
are referenced by the Julian Day on which they start, and will be referred to as such from
here on.
There were a total of fourteen events recorded during the monitoring period. The
first event, UF9200-154, was in fact the first pumping event of the wet season at UF9200.
It was completely recorded and sampled. The next three events suffered from various
operational difficulties and yielded incomplete data. There was a malfunction of the
controller during events UF9200-158 and UF9200-161 that was not discovered until the
sample pick-up Julian Date 161 so no samples were taken for UF9200-158 and the first 9
hours of UF9200-161. Samples were taken properly thereafter, however the phosphorus
samples for UF9200-169 were the ones that were blown out of the digestion tubes at the
contract analytical lab, so there were no phosphorus analyses for UF9200-169. Pumped
volume reported in all cases refers to pumped volume that was sampled. In cases where
the sampler filled before the event was completed, the unsampled volume has not been
considered, that is, no attempt was made to estimate the composition of unsampled
discharge. With these qualifications in place, a summary of the events is presented in
Table 7.2.
The events covered a broad range of discharge, ranging from 0.39 x 105 m3 to
7.34 x 105 m3. Suspended solids load ranged from 295 to 14,262 kg, and particulate
phosphorus load ranged from 3.9 to 77.4 kg. Inspection of the inter-event time column
shows that there were two instances when there was a large time period between events.
Slightly over 49 days elapsed between the end of Event 169 and the start of Event 220,
slightly over 30 days elapsed between the end of Event 285 and the start of Event 319.
These lapses arise from two different causes. Early into the wet season the grower
at UF9200 made the decision to attempt a water storage BMP. For the time period from
Julian Day 170 up to Julian Day 220, whatever pumpage was necessary was directed to a
77 ha storage pond that had been created by diking two fallow cane fields in the most
remote comer of the farm, diagonally opposite from the pump station. This was the cause


247
only two times was any field runoff detected. Both of these occasions corresponded to
times when the field groundwater level was at field elevation. For the time period
evaluated the field groundwater levels were at field elevation for less than 10% of the
total time and less than 14% of the time during pumping periods. In addition, with the
exception of the times when the high field water level coincided with pump start-up, the
farm discharge suspended solids concentrations tended to be well below event average
when field water level was high.
Third, samples were taken during an event that generated field runoff. These
samples were taken in the 1993 wet season immediately at the end of a 16.5 cm (6.5 inch)
rainfall at UF9200 that had caused a breach in several of the field ditch walls. Samples
were taken in the field ditches immediately downstream of the field runoff, in the main
canals where the sampled ditches discharged, and at the pump discharge. In all cases the
total suspended solids in the samples was less than 4 mg/1, which is less than the average
observed in the ditches during the more typical flow of the intensive synoptic tests. These
sample results coincided with the previously mentioned observations of low TSS
concentration at high field water levels.
These observations have led to the conclusion that overland flow erosion was an
insignificant direct contributor to discharge suspended solids and consequently to
discharge particulate phosphorus load.
Adsorption/Desorption The sorption characteristics of both the prototype surficial
sediment and the discharge suspended solids have been determined and discussed in
Chapters 5 and 8. The hypothesis was posed in Chapter 5 that the weak sorption
characteristics of the organic sediments would cause the combination of desorption and
pore water release to contribute less than 4% of the total discharge phosphorus. This
hypothesis was based on trial calculations using typical parameters determined for the
prototype sediment. The sorption characteristics of the actual farm discharge Large-Scale
Composite samples, presented in Chapter 8, fell within the range of the parameters used


283
70
Suspended Solids Simulation Results
Event UF9200-252
Time from start of Pumping-hr
O Observed Simulated
Figure 10.3: Simulation Results-Total Suspended Solids for Event UF9200-252
Suspended Solids Simulation Results
Event UF9200-258
O Observed Simulated
Figure 10.4: Simulation Results-Total Suspended Solids for Event UF9200-258


61
7. Set up regular discharge monitoring of particulate phosphorus at the target
farm.
8. Plan and execute synoptic studies to:
a. determine variation of sediment characteristics as a function of
location within the conveyance system and
b. evaluate suspended solids concentration variations within the
conveyance system during transport events.
9. Determine, in the field, the requirements for biological data depending on
the outcome of the preceding characterizations and monitoring programs.
10. Adapt the chosen computer program, develop the water quality program
for phosphorus, and calibrate and verify for the target farm.
The subsequent four chapters discuss the specific activities directed toward
sediment characterization and transport property determination.


37
Hvdrologic approach to seston transport
The application of traditional non-cohesive sediment transport theory to transport
of seston has been shown to be inappropriate for several reasons (Webster et al., 1987).
One major assumption in most non-cohesive sediment transport models is that
transported particles are similar to particles in the streambed (for example, Einstein
1950). This is often not the case in stream transport of seston, where the transported
matter may consist of small and/or low density matter with a high organic content while
the streambed may be primarily sand, rock, and large organic particles. A second major
assumption is that of unlimited supply of transportable solids (for example, Bagnold
1966). This is often not the case with seston transport (for example. Allen 1977) where
depletion of supply is frequently evident, for example over sequential storm events of
similar or increasing magnitude (for example, Svendsen and Kronvang, 1993).
These violations of traditional sediment transport assumptions may be
circumvented by using models derived from cohesive (Fine Grained) sediment transport
theory which allow for variation of bed composition and recognize supply limitation.
The theory of cohesive sediment transport formed the basis for much of the work planned
in this project and is discussed in the accompanying section entitled A Brief Discussion
of Sediment Transport Theory.
An alternative approach to the quantification of seston transport has been to
develop phenomenological relationships between basin hydrologic parameters and seston
transport. Typical are attempts to correlate particulate organic matter loads to stream
discharge or stream order (Hawkes 1975). These approaches have met with limited
success, mainly because they are implicitly based on the noncohesive sediment transport
model assumptions. A more sophisticated approach was proposed by Sedel! et al. (1978)
where stream power, rather than discharge, was used to correlate seston discharge. This
approach reduced data scatter somewhat, but suffered from the limitations that


108
trowel that was designed to slide along the upper flanges of the flume and extend to
within 2 inches of the flume bottom. Water containing 0.1 M K.C1 was carefully added to
a water depth of 15.2 cm (6 inches). The flume then stood in a quiescent state for 48
hours.
The first run was conducted at three arbitrary shear levels. Ring and channel
speeds were selected that gave shear stresses, calculated by equation 6.1, of 1.0, 2.0. and
3.0 dynes/cm2. The system was run at the first shear level for 3 hours, and at the second
and third shear levels for 2 hours each. The results are shown in Figure 6.3.
Figure 6.3: CSME vs. Time Into Run for Placed Organic Sediment Bed with 15.2 cm
Water Depth
Samples were taken at two locations in the channel, the near-bed sample tap was
approximately 1 cm above the bed surface, while the mid-depth sample tap was
approximately 8 cm above the bed surface. The two sample points exhibit the same


162
suspended solids. These included radio-tracer techniques, microscopic techniques,
elemental analysis, and functional group analysis. It was judged that none of these
techniques offered a reasonable chance for identity determination given the nature of the
potential sources and the resources available for the project. Determination of the
phosphorus content of the solids was felt to offer an opportunity for some discrimination
of possible source by comparison. In effect the phosphorus mass fraction of the
suspended solids would serve as a tag, albeit a low resolution tag.
Typical methods for determination of particulate phosphorus operate by process of
elimination, that is, a complete sample is split, one portion is analyzed for total
phosphorus, the other portion is filtered and analyzed for total phosphorus, and the
particulate content is assumed to be equal to the difference between the two. This was
not an adequate methodology to achieve the needs of this project, so an alternative
analytical procedure was developed, which involved direct digestion and analysis of the
solids that had been collected on the glass microfiber filter pad for the total suspended
solids analysis.
This procedure introduced several complicating factors into the process. The
standard glass microfiber filters specified for total suspended solids (TSS) analysis, such
as Whatman GF filters, are not phosphorus free. There are specialty phosphorus-free
filters available, but unfortunately they are all membrane type filters that blind almost
immediately when filtering organic sediments. Most also have an inherent water content
so they give a false weight loss upon drying and are not suitable for TSS analysis. It was
decided after some experimentation with the membrane filters to revert to the glass
microfiber filters and attempt to reduce their phosphorus content. This was done by acid
wash and de-ionized water rinse. Each filter used in this program received a minimum of
48 hour soak in IN Hydrochloric acid, followed by triple rinsing with de-ionized water.
The filters were then dried to constant weight at 105C and stored in a desiccator until
used. The process reduced the phosphorus content of the filters by about 80%. but there


350
pump station inlet,
d) the potentially dramatic benefit of overall transport reduction by altering
pumping strategies to significantly reduce maximum velocities.
e) the possibility that remedial actions may simply delay the export of
phosphorus if the unexported material is not removed or moved to regions
of non-transport, and
f) the potential for hydraulic mining to move phosphorus-containing solids
from areas of transport to areas of non-transport. The simulations
showed that this might be achievable on the target farm with existing
equipment.
Application of the general principle of velocity reduction to analysis of data from
the back-up farm led to the conclusion that major benefit might be had by
implementation of simple level control and velocity reduction policies related
only to pump operation.
Critique of the model in its current form
The model was critiqued in Model Qualifications and Applications of Chapter
10. Those qualifications, along with others mentioned elsewhere in the text are
summarized here.
1. The calibration of the model, as it is currently configured, to farm-scale
conveyance systems is extremely tedious and labor-intensive. This could
be rectified by the development of programs that translate the ASCII
output files to files suitable for analysis by existing statistical programs
and control the transfer of such files to the target statistical program.
2. The level changes simulated by the hydraulic model do not respond
quickly enough in simulating events that receive major rainfall during the


4
enhance the regions fish, wildlife, and other environmental resources, and provide a
water supply for the Everglades National Park" (Izuno and Bottcher. 1994). The
CSFFCD has since been succeeded by the South Florida Water Management District
(SFWMD), which assumed the responsibilities held by the CSFFCD in 1972.
Cooperation with the USACE has proceeded uninterrupted since 1948 between first the
CSFFD and then the SFWMD.
During the tenure of the CSFFCD major modifications were made in the hydraulic
patterns and land use of thousands of square kilometers of Everglades wetlands. The
EAA was completely canalized and diked. Fifteen hundred square miles (-3900 square
kilometers) of wetlands to the south between the EAA and the Everglades National Park
(ENP) were designated as Water Conservation Areas (WCAs) and dedicated to fish and
wildlife conservation and surplus water storage. The storage capacity of Lake
Okeechobee was increased and the capability to backpump farm stormwater runoff (or
more appropriately farm stormwater "pumpoff') from the EAA collector canals upstream
to the lake was implemented.
Unfortunately, during these many decades of major public works projects little
attention was paid to the water quality ramifications of the profound hydraulic and land
use modifications being imposed on the Everglades region. The ecological community in
the Everglades originally evolved as an oligotrophic system, that is, one with low
biological productivity, with phosphorus as the limiting nutrient. Studies conducted at
the WCAs and ENP indicate that high nutrient concentrations, including phosphorus as a
principal contributor, are causing significant shifts in the balance of the ecosystem
(Whalen et al., 1992). According to the South Florida Water Management District,
agricultural activities in the EAA, especially drainage and fertilization practices, have
resulted in a major decline in the quality of water entering the north end of the Everglades
Planning Area (EPA), the area encompassing the WCAs and ENP (Whalen et al., 1992).


41
Ca = concentration at datum level a, kg/m3
y = elevation, m
y0 = water surface elevation, m
a = datum level elevation, m
w = particle settling velocity, m/sec
P = constant of proportionality
k = von Karmans constant
U* = shear velocity, m/sec.
Application of this equation gives sigmoidal concentration curves, the shape of which
depends on the value of z.
Energy approaches analyze the suspended solids from the standpoint of an energy
balance, where the momentum transfer to the suspended solids by the turbulent fluid must
equal the excess weight of the solids in motion. Velikanovs gravitational theory
(Velikanov 1954), for example, uses this approach to arrive at an equation that describes
the concentration profile as follows:
3.3
a
y
where q
yo
qa = reference level, m
k
a =
30y0
ks = grain roughness
Application of this approach gives concentration curves similar to those produced by the
diffusion approach.


8.2 UF9200 Two-Point Adsorption/Desorption Curves (Desorption Mode) 236
8.3 UF9206N Two-Point Adsorption/Desorption Curves (Adsorption Mode) 237
8.4 UF9206N: Two-Point Adsorption/Desorption Curves (Desorption
Mode) 237
8.5 UF9200 Desorption Data Eadie-Hofstee Plot 239
8.6 UF9206N Desorption Data Eadie-Hofstee Plot..... 239
9.1 Simplified Schematic of UF9200 DUFLOW Nodal Structure 257
9.2 Calculated and Measured Flows, Event UF9200-237 261
9.3 Calculated and Measured Flows, Event UF9200-252 261
9.4 Calculated and Observed Levels, Event UF9200-220 262
10.1 Simulation Results-Total Suspended Solids for Event UF9200-220 282
10.2 Simulation Results-Total Suspended Solids for Event UF9200-237 282
10.3 Simulation Results-Total Suspended Solids for Event UF9200-252 283
10.4 Simulation Results-Total Suspended Solids for Event UF9200-258 283
10.5 Simulation Results-Total Suspended Solids for Event UF9200-262 284
10.6 Simulation Results-Total Suspended Solids for Event UF9200-285 284
10.7 Simulation of TSS at Selected Nodes for Event UF9200-237 287
10.8 Simulation of EM at Selected Nodes for Event UF9200-237 287
10.9 Simulation Results-Cumulative Suspended Solids Load for Event
UF9200-220 290
10.10 Simulation Results-Cumulative Particulate Phosphorus Load for Event
UF9200-220 290
10.11 Simulation Results-Cumulative Suspended Solids Load for Event
UF9200-237 291
10.12 Simulation Results-Cumulative Particulate Phosphorus Load for Event
UF9200-237 291
xviii


323
example. Recall from Figure 7.4 that Event UF9200-220 started with operation of the
large pump until the shut-down level was reached. After the shut-down level placed the
large pump in its first off mode the large pump continued to cycle on and off for a number
of cycles.
This simulation makes two separate modifications in the pump operation. First,
while maintaining the same overall discharge volume for the first major pump cycle the
early pumping rates are increased and the later rates are decreased to create a steeper
decline in pumping rates as time proceeds. Second, in place of the large pump cycling in
an on-off mode the simulation had the large pump shut off and stay off and the small
pump come on earlier than it had during the actual event. The timing of the small pump
start-up was set so that the same time was required to reach 100,000 m3 discharged as
was observed in the actual event.
The combined effects of the two modifications achieved several ends. Early
velocities, which were relatively low because of maximum depth, were increased but later
velocities, which became high because of decreased depth, were reduced. Shutting down
the pumps completely for a period instead of allowing the large pump to cycle allowed
the level in the East Canal to increase. Starting the small pump earlier at the higher level
resulted in velocities that were below critical, so the model predicted zero mass transport.
The results of these combined changes are shown in Figure 11.4. Several management
practices will be progressively exemplified on this chart so this simulation is referred to
as Simulation A.
Note that the initial manifestation of the change is to increase net transport over
that presented in the original simulation because of the higher initial velocities. This is
more than compensated for, however, in the latter stages of the first major pump cycle
when the reduced flow at the lower depths cuts the velocity substantially, so that by the
time the large pump is shut off, at -72,000 m3 the net export is 4370 kg, compared to
6850 kg at the same point in the original simulation.


387
N
Mill Road Canal
Figure G.2: Layout of UF9206


CHAPTER 2
PRIOR DATA RELEVANT TO PARTICULATE PHOSPHORUS TRANSPORT IN
THE EVERGLADES REGION
Introduction
The salient characteristics of the Everglades such as topography, soils and
sediments, climate, and hydrology are not typical of those found in agricultural areas
elsewhere in the United States. The area is extremely flat land at low elevation,
frequently subject to intense local rainfall and periodically subject to intense regional
rainfall, with drainage and irrigation under almost total human control.
The soils of the area are histosols, poorly drained muck and peaty soils low in
mineral content but rich in organic matter (in excess of 65% by weight), with low specific
gravity and low hydraulic conductivity (Snyder, 1994), bearing more similarity to the
soils and sediments of the Louisiana bayous than to upland mineral soils.
The climate is sub-tropical. The annual average solar radiation and wet season
temperatures are more similar to those prevailing in the Caribbean and Central America
rather than the United States southeast and western plains agricultural regions (Marsh,
1987. Trewartha and Horn, 1980).
The heavy reliance on pumping for water transport introduces flow-no-flow
situations and hydraulic transients which are not typical of a more normal upland
watershed. In many respects the hydrology of the EAA is more akin to that of the polder
region of the Northern Netherlands where discharge hydrology is completely under
human control (Snyder, 1994).
Given the individualistic character of the Everglades region in general and the
EAA in particular it is not unreasonable to assume that particulate matter transport
12


391
KZEDEM=0.00000;
}
else
Avoid Negative Settling
{
KZEDEM=ksed*As/w*TSS;
Calculate Internal EM Zero Order Rate Constant
}
if (TERM<=0.00000)
KONEEM=0.00000; Avoid Negative Erosion
)
else
{
KONEEM=-EPSI*VSQCONV*TERM;
Calculate Internal EM First Order Rate Constant
}
KO(TSS)=KZEDTSS; Convert Internal Variables to the Required
K1(TSS)=K0NETSS; DUPROL Zero and First Order Output
KO(EM)=KZEDEM; Variable Names
K1 (EM)=KONEEM;
}


334
particulate phosphorus export, by simple manipulation of pumping rates to minimize
maximum velocities. Two prime caveats must again be mentioned. First, the model
predicts zero transport below the critical velocity. The field data presented in Chapter 7
showed that there is transport of (presumably very light) material at velocities below
critical under some circumstances. The model approximation will tend to underestimate
transport under these circumstances making the predicted export reduction somewhat
optimistic. Second, all or a portion of the material not transported may be available for
transport in the next event so improvements may not be totally cumulative unless
erodable solids removal is practiced between events.
Even with these qualifications it is reasonable to state that the simulations exhibit
the dominating role of channel velocity on particulate phosphorus export for a given
distribution of erodable mass. This concept may be generalized to other hydraulic
systems and used to evaluate the results obtained at the back-up location, farm UF9206.
Evaluation of Suspended Solids and Particulate Phosphorus Export at Farm UF9206
Possible Dominant Influence of Channel Water Depth on Velocity and Transport
Recall from the data presented in Chapter 7 that the suspended solids and
particulate phosphorus export curves observed at the two pump stations at UF9206 did
not always exhibit the characteristic U shaped curve that was observed at UF9200.
Both stations at UF9206 were more hydraulically active, on a water-volume-pumped-per-
unit-farm-area basis, than UF9200. This activity was manifested in a lower average
phosphorus content of exported solids at both stations than was observed at UF9200.
The combined suspended solids loads from the two UF9206 pump stations were
almost ten times that of UF9200 so even though the average phosphorus content was


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID ECCJVYMCZ_FKI2AF INGEST_TIME 2015-03-25T19:47:14Z PACKAGE AA00029751_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


299
macrophyte mat and replenishment of the lower canal reaches from farther upstream, both
resulting from wind driven currents and surface movement during interevent times.
Following Event 262 there was an interevent time of about 23 days where significant in-
stream growth and macrophyte population increase could have taken place, giving rise to
the increased value of Cemioi of 550 gm/m2 for Event 285. This entire process may be
viewed with a simple plot of Cem(O) as a function of chronological event start date. Figure
10.22 illustrates this.
Cem(o) vs Event Julian Start Date
Event Julian Start Date
Figure 10.22: Chronological Values of Cem(0)
The Cem(0) correlation may be viewed in a different perspective by plotting the
data as an exponential growth equation, in semi-logarithmic coordinates. Figure 10.23
shows this plot, which suggests a slightly different but related rationalization. The
exponential correlation equation, fit to all six events, is
Q(0) = 247.94c0 0393' 102


S bed slope, or hydraulic gradient, or substrate concentration of srbate, or
the quantity of a constituent passing a cross-section per unit time
So native srbate concentration originally present on substrate
Sa substrate concentration of srbate adsorbed from liquid phase
Sm maximum srbate concentration at complete saturation of sorption sites
t time
TSS total suspended solids concentration
u velocity
U shear velocity
V volume of water column, or channel velocity
Vc critical channel mean velocity for erosion
VCJ velocity which corresponds to tcj
Vmean mean stream velocity
w specific weight of water
ws settling velocity of sediment
wSed sedimentation velocity
Wj Sedimentation velocity of particles in class /
x x-coordinate, or hMV2
y elevation, or distance from surface
yo water surface elevation
z elevation of sediment surface above datum plane
Greek Letters
a rate coefficient, or erosion constant
P constant of proportionality
y specific mass of water
XXIV


257
The DUFLOW input allows only rectangular culverts so the circular UF9200
culverts were converted to rectangles with sill and crown elevations and cross-sectional
area equal to the actual culverts. Culvert frictional resistance was specified by assuming
a Mannings n of 0.025 for all culverts and then calculating the appropriate Chezy
coefficient for each culvert using its full flow hydraulic radius. Channel flow resistance
was a calibration variable and so was varied with each run. Wind effects were ignored
for all runs. The program was always run with the extra iteration and damped Froude
term options activated to increase precision. These selections added little to the overall
run time.
UF9200-Duflow Nodal Chan
1
N
E
2:















806
O



a




o



-a
















c/5
I

2


T
















on

-o-

-O-
o
-o
-D
a
o
-o-
-H3-
-d-
-Q

-O
^0
201
402
603
804
1005
1206
1407
1608
1809
2010
2211
2412
2613
2814
3015 3216
Pump
Meters West of Pump Inlet
Station
Figure 9.1: Simplified Schematic of UF9200 DUFLOW Nodal Structure
There were no data available to justify specific spatial variation of hydraulic input
parameters so the calibrations were carried out assuming uniformity across the farm.
Data Sources for Calibration The data used for hydraulic calibration were as
follows:


301
erodable material as a first order process. This might in turn be tied to a first order
growth process for the biological material presumed to be producing the erodable
material.
Comparison of Calibration Results for Erodable Mass with Literature Values
Direct evaluation of the credibility of the values obtained for Cem(O) by
comparison with literature values is difficult because of the lack of literature data on
dislodeable detritus quantities. As an alternative we will look at the limited data
available on detrital production and then attempt to relate this to plant growth rates under
conditions similar to those prevailing in the EAA.
Consider first as a representative example the value of 350 gm/m2 for Cem(O)
exhibited in several of the events of the calibration set. As a trial calculation let us assume
a surface occupied 75% by water lettuce at a density of 627 gm/m2 and 25% by water
hyacinth at a density of 1535 gm/m2 (See Table 8.4). This would translate to a weighted
average surface mass density of 854 gm/m2. If it is assumed that 38% of the mass is
dislodgable detritus (See Table 8.5) the dislodgable detritus mass density would be about
325 gm/m2, close to the 350 gm/m2 representative value.
Now Equation 10.2 has a time constant of 0.0393 days'1 which, at a density of 350
gm/m2, gives a production of erodable material on the order of 14 gm/m2/day. Debusk
and Dierberg (1989) reported average rates of detritus deposition of water hyacinth in the
range of 1.2-3.0 gm/m2/day, but this was in a passive system with minimal disturbance.
They noted that detritus deposition is a poor measure of detritus production because of
the highly retentive nature of the hyacinth root system. Moorehead, et al. (1988) reported
detritus accumulation of water hyacinth in the range of 4.7-6.7 gm/m2/day in a fertilized
system and 3.5-3.7 gm/m2/day in an unfertilized system. Their definition of detritus was
that which was retained on a 1 mm surrounding screen plus hand separated roots and


358
conditions a technique known as sheltered dredging was used, wherein the Eckman
dredge was brought slowly to a location just below the surface, at which point it was
inserted into a container which already contained ambient water, displacing its own
volume of water. The container was then removed from the water and the sediment
transfer from the dredge was completed. Suspended sediment was allowed to settle and
then excess water was decanted off if necessary.


357
bottom of the core tube was sufficient to prevent the sediment from falling during
retrieval.
Following retrieval the handles and upper corer flange were removed and a
calibrated rod, tipped with a sister piston, was inserted into the bottom of the corer. The
vent valve on the internal piston was opened and the internal piston was removed. The
sediment sample was then expunged in the desired length increments by securing the
corer tube, positioning a movable stop on the lower piston rod to the desired setting, and
slowly driving the rod upward until the stop was reached. The expunged sediment was
collected in a coated polystyrene collar channel which fit around the top of the corer tube
and allowed the sample to be directed to polyethylene sample bags.
Eckman Dredge (Mudroch and MacKnight, 1991, Hakanson, 1981, Hakanson, 1986)
The Eckman sampler is a stainless steel box with a pair of spring-loaded jaws for
closure and free-moving, hinged flaps on top for sample protection. An externally
mounted trigger device which causes the jaws to close is activated by a hollow cylindrical
weight which is dropped down the hoisting cable. The Eckman sampler is especially
effective in sampling light, fine grained sediments and provides larger samples and easy
access to the surficial sediment if sub-sampling is desired. The dredge used in these
studies had a cubic box size of approximately 150 cm on a side, with an effective volume
on the order of 3 liters.
The sampling procedure was to set the hinged jaws in the open position, lower the
dredge until it rested on the sediment as indicated by slackening of the hoisting rope,
remove the rope slack and drop the weight, which triggered the catch mechanism, closing
the jaws. The dredge and its contents were then hoisted to the surface and emptied into a
sample bucket.
In cases where light surficial sediment is expected there is a chance that removing
the Eckman dredge from the water will cause some loss of light sediment. Under these


139
accounted for all material which had been removed by sampling in addition to the matter
suspended in the water column at any specific time.
The calibration runs with organic sediment lasted 4 hours each. Figure 6.17
shows a typical test result, where CSME increased rapidly for the first half hour, then
showed a very modest rate of increase over the remaining 3.5 hours. The contents of the
PES were being diluted by sampling and replacement with clear water so after about 1
hour in a typical run the actual concentration in the PES would start to decrease.
OS8 178 RPM
CSME vs Time
0.0020
Least Squares Fit
0.0018
0.0016
0.0014
O
E
0.0012
0.0010

O o O-rs
o O O o
1
o
U.u
0
c
0
c
o
*'
2
0.0008
u
0.0006
:
Moving Average
0.0004
0.0002
0.0000

0
0.5
1 1.5 2 2.5
3 3.5 4
4.5
Time hrs
Figure 6.17: Typical PES Erosion Curve with Organic Sediment
Referring to sedimentation arguments made earlier in this chapter, it could be
surmised that the slight increase in CSME calculated after the first hour may not have
been due as much to continued low-rate erosion as it was to reduced sedimentation
because of a slowly declining solids concentration arising from the sampling procedure.
To correct for this possibility, a linear least squares curve fit was made of the last three
hours data for each run. The asymptotic CSME for each run was defined as the intercept


211
discharge. The point to note about Event 258 is the definite increase in suspended solids
concentration from Location 2.1 to Location 3 followed by the decrease at Location 4.
Viewed another way, there appears to be only a weak relationship between Locations 2.1
and 3, but Location 4 appears to present a similar response as Location 3 with about a 1
hour lag.
Event 262 suffered from a reduced sampling schedule so there is no Location 4
data, but the trends of the locations monitored are similar to prior runs. There was very
little increase in TSS concentration between Location 2 and location 2.1 which is 1408
meters downstream from Location 2. but there was a considerable increase in
concentration between Location 2.1 and Location 3 which is 200 meters farther
downstream. The discharge concentration appeared to show an attenuated wave pattern
similar to Location 3 with about a 2 hour lag from peak to peak.
The graph of Event 285 has been truncated at 32 hours which is short of the end
of the first major pump cycle (42 hours) to avoid most of the effects of the major TSS
influx from the North Canal that started around hour 26 and exceeded 100 mg/1 by hour
32. Location 3 data were lost after hour 22 during this event but the relationship between
Locations 3 and 4 again appears to be Location 4 approximately repeating the Location 3
pattern with a time lag. In this case the time lag appeared to be considerably longer, on
the order of 8-10 hours. During event 285 Location 2 started off with a TSS
concentration that peaked at about 10 mg/1 and stayed above 5 mg/1 through hour 14, but
this peak was not transferred to the next location, 2.1, that stayed consistently below 4
mg/1 throughout this portion of the event.
Order of Magnitude Transport Estimates
At this point it is instructive to compare the overall suspended solids transport
calculated for the discrete events of the UF9200 time series discharge studies with the


102
Figure 6.1: Schematic Diagram of Counter-Rotating Annular Flume (CRAF) (From
Mehta, 1973)


223
considerably lower than the above noted ten hour averages for the preceding and
subsequent events.
Table 8.3: Canal Surficial Sediment Phosphorus Content UF9200 Survey of
Julian Date 300
Location
% Solids Content
Solids Volatile Fraction
Phosphorus Content-mg/kg
South Canal-at Field B1
16.6
0.49
645
South Canal-at Field B1
15.0
0.46
705
South Canal-at Field B2
15.5
0.53
650
South Canal-at Field B2
15.3
0.52
657
South Canal-at Field B3
5.1
0.50
1156
South Canal-at Field B3
6.0
0.50
1260
South Canal-at Field B4
6.0
0.44
790
South Canal-at Field B4
7.3
0.42
802
South Canal-at Field B7
5.1
0.59
986
South Canal-at Field B7
6.2
0.58
958
South Canal-at Field B12
10.1
0.49
679
South Canal-at Field B12
11.1
0.51
860
North Canal-at Field A1
12.5
0.50
852
North Canal-at Field A1
13.0
0.49
777
North Canal-at Field A5
20.0
0.63
876
North Canal-at Field A5
18.3
0.74
802


300
where t is now in the dimensions of days, with an r2 value of 0.984, representing a
favorable correlation.
Cem(0) vs Interevent Time
Exponential Format
Interevent Time, t days
O Calibration Values Correlation Equation
Figure 10.23: Cem<0) Correlated as an Exponential Growth Equation
A physical interpretation of this equation may be that the production of erodable
material proceeded approximately as a first order process, with a time constant of 0.0393
days1. This time constant translates to an increase of erodable material of about 4% per
day, or a doubling time of around 17-18 days. The value of Cem(0) at time zero, 247.94
gm/nr could be considered as the true baseline concentration that might exist after a
number of average pumping events, and could represent the rapid regrowth of epiphyton
in the macrophyte mats, replenishment of subsurface erodable sediment by residual and
wind driven currents, the rapid growth of suspended and resuspendable phytoplankton in
the channels, and the average transportable mass not discharged during a typical event.
This approach is more appealing from an analytical standpoint because it allows
the correlation of Cem(0) to be linked to a physically sensible description of production of


PARTICULATE PHOSPHORUS TRANSPORT IN THE
WATER CONVEYANCE SYSTEMS OF THE
EVERGLADES AGRICULTURAL AREA
By
JAMES DONALD STUCK
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1996

Copyright 1996
by
James Donald Stuck

This work is dedicated to the two people in my life who have helped me negotiate
all the forks in the road, my wife Beverly, and my son Clark; and to the memory of
Christopher.

ACKNOWLEDGMENTS
The author would like to express his appreciation to his major professor. Dr. K.L.
Campbell of the Department of Agricultural and Biological Engineering, for his support
throughout this effort. He would also like to acknowledge the other members of his
committee: Dr. J. J. Delfino of Environmental Engineering Sciences, Dr. F. T. Izuno of
Agricultural and Biological Engineering, Everglades Research and Education Center. Dr.
A. J. Mehta of Coastal and Oceanographic Engineering, Dr. K. R. Reddy of Soil and
Water Science, and Dr. A. B. Bottcher, President of Soil and Water Engineering
Technology, special committee member.
Thanks are due to the Departments of Agricultural and Biological Engineering,
Coastal and Oceanographic Engineering, and Soil and Water Science, and the Everglades
Research and Education Center, all of which contributed laboratory space and equipment
to this interdisciplinary study.
The author is indebted to many individuals who provided invaluable assistance
during the course of this work: Ms. Yu Wang and Dr. Winston Davis for their analytical
help, Ray Garcia and Curtis Miller for their patient advice and essential help in the field,
Laurene Capone for helping get the resources and smoothing the path so often, Mauricio
Abreu for his many long hours working on the erosion studies, and Dr. Nigel Pickering
for his considerable computer skills.
Financial support is gratefully acknowledged from the U. S. Department of
Agriculture, which provided the author with a Research Fellowship in Hydrologic
Science while this research was in progress, and from the Everglades Agricultural Area
Environmental Protection District, which provided the operating funds for this research as
iv

a part of their Best Management Practices Implementation program for reducing
phosphorus loading to the Everglades.
The author would particularly like to thank his loving, patient and supportive
wife, Beverly, who contributed to the completion of this work in ways too numerous to
detail.
v

TABLE OF CONTENTS
ACKNOWLEDGMENTS iv
LIST OF TABLES xi
LIST OF FIGURES xiii
NOMENCLATURE xxii
ABSTRACT xxvii
CHAPTERS
1 INTRODUCTION. BACKGROUND. AND RESEARCH
MOTIVATION 1
Historical Summary 1
Sources of Phosphorus in the EAA 5
Proposed Regulatory Remediation 7
Areas of Uncertainty in Remediation Plan 9
EAA Modeling Activity 10
2 PRIOR DATA RELEVANT TO PARTICULATE PHOSPHORUS
TRANSPORT IN THE EVERGLADES REGION 12
Introduction 12
Farm-Field Scale Studies 13
CH2M Hill study 13
Izuno, et al study 15
Canal Transport Studies 16
Southeast Florida canals water quality studies 16
Water Conservation Area water quality correlations 17
EAA Canal Sediment Studies 17
Anderson and Hutcheon Engineers study (1992) 18
Andreis/U.S. Sugar proposed sediment control BMPs (1993) 19
Sediment control demonstration project 20
South Florida Water Management District Canal Data Set 22
3 GENERAL LITERATURE REFERENCES RELEVANT TO
PARTICULATE PHOSPHORUS TRANSPORT 24
vi

Introduction 24
Sources of Particulate Phosphorus 24
Early Diagenesis 27
Erosion and Transport of Organic Material 28
General erosion and transport 28
Hydrologic approach to seston transport 37
A Brief Discussion of Sediment Transport Theory 39
Non-Cohesive sediment transport 39
Cohesive sediment transport 43
Particle entrainment simulator 48
4 INITIAL HYPOTHESES, OBJECTIVES, AND RESEARCH PLANS 51
Problem Overview 51
_ Summary of sources 51
Original conceptual model 53
Scope of This Research 55
Minimum criteria for experimental and modeling efforts 55
Application of minimum criteria and resulting conceptual model
simplification 56
Research goals for this study 58
General Research Plan 60
5 SEDIMENT SURVEY AND PHYSICAL-CHEMICAL
CHARACTERIZATIONS OF SELECTED SEDIMENTS 62
Sediment Survey 62
Selection of representative farms 62
Sediment survey measurement methods 64
Sediment survey results 65
Sediment survey discussion 66
Selection of Primary Target Farm 69
Particle-Size Property Distribution Study 69
Sources for particle size fractionation study 70
Soil and sediment particle size fractionation 71
Fraction analysis 73
Particle size distribution 74
Volatile matter content 77
Particle specific gravity 79
Total phosphorus content 80
Phosphorus fractionation results 82
Adsorption-Desorption experimental methods 84
Adsorption-Desorption data reduction 85
Potential environmental significance of adsorption-desorption 91
Sedimentation parameters of B9B10 sediment 95
vii

6SURFICIAL SEDIMENT TRANSPORT STUDIES
.99
Introduction 99
Prototype Sediment 100
Laboratory Erosion Simulation Device- The Counter-Rotating Annular
Flume 101
Flume description 101
Flume operation 105
Organic sediment flume studies 107
Organic Sediment Particle Entrainment Simulator Studies 132
Particle Entrainment Simulator configuration 132
Particle Entrainment Simulator operation preliminary clay tests 136
Particle Entrainment Simulator operation organic sediment tests 137
Calibration of PES against flume and extrapolation to field
parameters 141
Field Measurements using the calibrated particle entrainment
simulator 148
7 FIELD EVENT STUDIES-TIME SERIES AND SYNOPTIC 155
Introduction 155
Time Series Discharge Studies 156
Data Monitoring and sample acquisition 156
Detailed hydrography of the target study farm~UF9200 159
Particulate phosphorus content determination 161
Particulate phosphorus event studiesFarm UF9200 163
Suspended solids transport-Farm UF9200 168
Particulate phosphorus content of suspended solidsFarm UF9200 176
Particulate phosphorus event studiesFarm UF9206, North and
South stations 181
Suspended solids transport Farm UF9206 189
Particulate phosphorus content of suspended solids Farm UF9206 196
Intensive Synoptic Studies at UF9200 200
Sampling configuration 200
Ditches vs. discharge 204
Influence of the North Canal 207
Downstream variation of suspended solids concentration 208
Order of Magnitude Transport Estimates 211
Summary 216
8 FIELD SPECIFIC STUDIES-FIELD SEDIMENTS, MACROPHYTE
AND DETRITUS STUDIES, LARGE COMPOSITE SAMPLES 217
Introduction 217
Sediment Phosphorus Content Synoptic Surveys 218
Macrophyte and Detritus Studies 224
viii

Macrophyte areal density and gross total phosphorus content studies 226
Dislodgable detritus studies 228
Large Composite Sample Studies 233
Adsorption/desorption tests 235
Phosphorus fractionation Tests on Large-Scale Composite Samples 240
Summary 243
9 MODEL DEVELOPMENT 244
Introduction 244
Exclusion of Overland Flow Erosion and Sorption from the Model 246
Model Platform Selection and Hydraulic Calibration 248
Model review and selection 248
General characteristics of the DUFLOW program 250
DUFLOW water quality module 252
Field drainage model 254
Hydraulic model calibration 256
The Evidence for Biological-Growth Controlled Particulate Phosphorus
Transport 265
Modified Model Development Plan 267
Model Development 269
General model assumptions 269
Model derivation 270
Transport model calibration methodology 273
10 TRANSPORT MODEL RESULTS AND DISCUSSION 278
Introduction 278
Calibration Parameter Results 278
Normal Wet Season Simulation Results 281
Suspended solids simulation 281
Cumulative suspended solids load and particulate phosphorus load
simulation results 289
Small flows 296
Correlation of Calibration Values of Cem(0), Initial Erodable Mass 297
Comparison of Calibration Results for Erodable Mass with Literature
Values 301
Model Validation Prediction of the Suspended Solids and Particulate
Phosphorus Loads for the Late Season Storms 304
Summary, Model Qualifications and Applications 310
Summary 310
Model qualifications 313
Model applications 314
11 MODEL APPLICATION, CONCLUSIONS AND
RECOMMENDATIONS 318
ix

Introduction..... 318
Evaluation of Potential Management Practices 318
Reduction of transportable mass in the system 319
Effect of sediment trap 320
Effect of pumping modification 322
Effect of wider discharge outlets 324
Hydraulic mining and recycle analysis 326
Specific effect of reduced velocity 1 332
Evaluation of Suspended Solids and Particulate Phosphorus Export at
Farm UF9206 334
Possible dominant influence of channel water depth on velocity and
transport 334
Correlation of export suspended solids with channel depth 337
Particulate phosphorus export at UF9206 341
Relation of Particulate Phosphorus Export to Soluble Phosphorus Export
at Both Farms 345
Conclusions and Critique 346
Conclusions 346
Critique of the model in its current form 350
Recommendations 352
BMP recommendations for field implementation 352
Model framework development 353
Model development 354
BMP development 355
APPENDICES
A PHYSICAL SAMPLING TECHNIQUES 356
B PHYSICAL ANALYSIS 359
C CHEMICAL ANALYSIS 363
D APPROXIMATE MODEL FOR CRAF 365
E RAINFALL AND WATER LEVELS AT UF9200 369
F DIMENSIONS OF UF9200 USED IN DUFLOW FORMAT 373
G FARM LAYOUTS FOR UF9200 AND UF9206 386
H DUPROL EROSION PROGRAM 389
REFERENCES 392
BIOGRAPHICAL SKETCH 404
x

LIST OF TABLES
Table E2ge
5.1 Sediment Survey Physical Results 65
5.2 Sediment Survey Analytical Results 65
5.3 Estimates of Total Phosphorus Mass in Target Main Canal Sediments 68
5.4 Soil and Sediment Screen Fractionation Sequence 72
5.5 Phosphorus Fractionation Results for WPBC and B9B10 Sediments 83
5.6 Sorption Parameters and Coefficient of Determination for WPBC and
B9B10 Sediments 90
5.7 Summary of Phosphorus Distribution in Demonstration Example 1 93
5.8 Summary of Phosphorus Distribution in Demonstration Example 2 94
5.9 Solids Content and Bulk Densities of WPBC and B9B10 <38 Micrometer
Fractions 98
6.1 Constants for Fit of Equation 6.5 to Multiple Shear Erosion Tests 113
6.2 Constants for Fit of Equation 6.8 to Multiple Shear Deposition Tests with
Water Depth of 30.5 cm 117
6.3 Constants for Fit of Equation 6.8 to Multiple Shear Deposition Tests with
Water Depth of 15.2 cm 121
6.4 Field PES Sample Point Descriptions 150
7.1 Average Conveyance Dimensions for UF9200 161
7.2 Summary of Pumping Events for Target Farm UF9200 165
7.3 Summary of Pumping Events for Station UF9206N 183
7.4 Summary of Pumping Events for Station UF9206S 184
xi

7.5 Comparison of Pumping Events Between Farms UF9200 and UF9206 for
Julian Dates 218 through 350 185
7.6 Sampling Locations for Intensive Synoptic Studies at UF9200 202
7.7 UF9200 Events 220-285 Expressed in Areal Loading Terms 212
7.8 UF9200 Weighted Areal Loads Estimated From Equations 7.7 and 7.8 214
8.1 Field Ditch Surficial Sediment Phosphorus Content UF9200 Synoptic
Survey of Julian Date 229 219
8.2 Canal Surficial Sediment Phosphorus Content UF9200 Synoptic Survey
of Julian Date 229 220
8.3 Canal Surficial Sediment Phosphorus Content UF9200 Survey of Julian
Date 300 223
8.4 Macrophyte Mass Density, Volatile Content, and Phosphorus Content 227
8.5 Water Lettuce Dislodgable Detritus Study 231
8.6 Phosphorus Fractionation Analysis of Large-Scale Composite Samples 241
8.7 Comparison Between Large-Scale Composite Samples and Prototype
Surficial Sediment for Phosphorus Content Distribution 242
9.1 Results of Hydraulic Calibration for UF9200 263
10.1 Erosion Parameters Determined in the Calibration Process 279
10.2 Placement of Example Nodes for Event UF9200-327 Simulation 286
10.3 Interevent Time and Values of CEm(0) Used for Validation Simulation 305
11.1 Load Weighted Distribution of Soluble and Particulate Phosphorus at
UF9200, UF9206N, and UF9206S 346
D. 1 Example of Spreadsheet Output 367
D.2 Spreadsheet Formulas 368
xii

LIST OF FIGURES
Figure Page
1.1 Original Everglades Watershed : 2
4.1 Farm Scale Phosphorus Transport Original Conceptual Model 53
4.2 Farm Scale Phosphorus Transport Simplified Conceptual Model 59
5.1 Particle Size Distribution of Dry and Flydrated B9B10 Soil 75
5.2 Particle Size Distribution of Flydrated B9B10 Soil and B9B10 Sediment 77
5.3 Particle Size Distribution of Hydrated B9B10 Soil, B9B10 Sediment,
and WPBC Sediment 78
5.4 Volatile Fraction (Organic Content) Distribution of Hydrated B9B10
Soil. B9B10 Sediment, and WPBC Sediment 79
5.5 Particle Specific Gravity Distribution of Hydrated B9B10 Soil, B9B10
Sediment, and WPBC Sediment 80
5.6 Correlation of Particle Specific Gravity with Organic Content for
Hydrated B9B10 Soil, B9B10 Sediment, and WPBC Sediment 81
5.7 Total Phosphorus (TP) Content Distribution of Hydrated B9B10 Soil,
B9B10 Sediment, and WPBC Sediment 81
5.8 Phosphorus Fractions vs. Organic Fraction for WPBC and B9B10
Sediments 83
5.9 Adsorption Isotherm-WPBC Sediment 75-150 Micrometer Range 88
5.10 Adsorption Isotherm-WPBC Sediment <38 Micrometer Range 89
5.11 Adsorption Isotherm-B9B10 Sediment 75-150 Micrometer Range 89
5.12 Adsorption Isotherm-B9B 10 Sediment <38 Micrometer Range 90
xiii

5.13 Sedimentation Velocity Distribution as Determined by The Bottom
Withdrawal Technique-B9B10 Sediment <38 Micrometer Particle Range 97
5.14 Mean Sedimentation Velocity vs. Test Duration 97
6.1 Schematic Diagram of Counter-Rotating Annular Flume (CRAF) 102
6.2 Photograph of Counter-Rotating Annular Flume (CRAF) 103
6.3 CSME vs. Time Into Run for Placed Organic Sediment Bed with 15.2
cm. Water Depth 108
6.4 CSME as Function of Consolidation Time for Organic Sediment Bed
with 15.2 cm. Water Depth 110
6.5 CSME as Function of Shear Time at Constant Calculated Shear Rate of
2.0 dynes/cm2 for Consolidated Organic Sediment Bed with 30.5 cm.
Water Depth 111
6.6 CSME as Function of Shear Time at Several Calculated Shear Rates for
Erosion of Consolidated Prototype Organic Sediment Bed with 30.5 cm.
Water Depth 112
6.7 Initial Erosion Rates vs. Applied Shear Stress for Organic Sediment
Extended Run 114
6.8 Asymptotic CSME vs. Applied Shear Stress for Organic Sediment
Extended Run Erosion at 30.5 cm. Water Depth 115
6.9 CSME as Function of Shear Time at Several Calculated Shear Rates for
Deposition of Consolidated Prototype Organic Sediment Bed at 30.5 cm.
Water Depth 117
6.10 Asymptotic CSME vs. Applied Shear Stress for Organic Sediment
Extended Run Erosion and Deposition with 30.5 cm. Water Depth 118
6.11 CSME as Function of Shear Time at Several Calculated Shear Rates for
Erosion of Consolidated Prototype Organic Sediment Bed at 15.2 cm.
Water Depth 122
6.12 Asymptotic CSME vs. Applied Shear Stress for Prototype Organic
Sediment Extended Run Erosion with 30.5 cm. and 15.2 cm. Water
Depth 123
6.13 Asymptotic CSME vs. h2/3V2 for Prototype Organic Sediment Erosion
Extended Run with 30.5 cm. and 15.2 cm. Water Depth 131
xiv

6.14 Schematic Diagram of Particle Entrainment Simulator (PES) 133
6.15 Photograph of Particle Entrainment Simulator (PES) 134
6.16 PES-CRAF Calibration Results Using Cohesive Clays 137
6.17 Typical PES Erosion Curve with Organic Sediment 139
6.18 Results of PES Runs with Prototype Organic Sediment CSME vs. Grid
Oscillation Frequency 141
6.19 Velocity Profile Measurements Reported by Mehta (1973) in CRAF with
22.9cm Water Depth 145
6.20 Schematic Representation of Superposition of Extended Logarithmic
Velocity Profile on CRAF Sigmoidal Velocity Profile 146
6.21 Final Calibration Curve of Estimated Equivalent Open Channel Mean
Velocity For Erosion as a Function of PES Oscillation Frequency 147
6.22 Example of Field PES Test Results 151
6.23 Compiled Field PES Test Results for UF9200 153
6.24 Compiled Field PES Test Results for UF9206 154
7.1 Correlation of Event TSS Load with Event Hydraulic Load for Farm
UF9200 167
7.2 Correlation of Event PP Load with Event Hydraulic Load for Farm
UF9200 167
7.3 Hydraulic and Total Suspended Solids Profiles for Event UF9200-154 169
7.4 Hydraulic and Total Suspended Solids Profiles for Event UF9200-220 169
7.5 Hydraulic and Total Suspended Solids Profiles for Event UF9200-237 170
7.6 Hydraulic and Total Suspended Solids Profiles for Event UF9200-244 170
7.7 Hydraulic and Total Suspended Solids Profiles for Event UF9200-252 171
7.8 Hydraulic and Total Suspended Solids Profiles for Event UF9200-258 171
7.9 Hydraulic and Total Suspended Solids Profiles for Event UF9200-262 172
7.10 Hydraulic and Total Suspended Solids Profiles for Event UF9200-285 172
xv

7.11 Hydraulic and Total Suspended Solids Profiles for Event UF9200-319 173
7.12 Hydraulic and Total Suspended Solids Profiles for Event UF9200-336 173
7.13 Hydraulic and Total Suspended Solids Profiles for Event UF9200-355 174
7.14 UF9200 Phosphorus Content of TSS as a Function of TSS Concentration 177
7.15 UF9200 Phosphorus Content of TSS as a Function of TSS Concentration
(Expanded Scale) 177
7.16 UF9200 Particulate Phosphorus Concentration in Discharge as a
Function of TSS Concentration (Expanded Scale) 180
7.17 Correlation of Event TSS Load with Event Hydraulic Load for Station
UF9206N 187
7.18 Correlation of Event PP Load with Event Hydraulic Load for Station
UF9206N 187
7.19 Correlation of Event TSS Load with Event Hydraulic Load for Station
UF9206S 188
7.20 Correlation of Event PP Load with Event Hydraulic Load for Station
UF9206S 188
7.21 Time Series-Flow and Total Suspended Solids Events UF9206N-209,
213 190
7.22 Time Series-Flow and Total Suspended Solids Events UF9206N-224,
234,239 190
7.23 Time Series-Flow and Total Suspended Solids Events UF9206N-251,
256, 260 191
7.24 Time Series-Flow and Total Suspended Solids Events UF9206N-268,
273,279 191
7.25 Time Series-Flow and Total Suspended Solids Events UF9206N-283,
300 192
7.26 Time Series-Flow and Total Suspended Solids Event UF9206N-318 192
7.27 Time Series-Flow and Total Suspended Solids Event UF9206N-336 193
xvi

7.28 Time Series-Flow and Total Suspended Solids Events UF9206S-209.
213,218,224, 234 193
7.29 Time Series-Flow and Total Suspended Solids Events UF9206S-260.
268 194
7.30 Time Series-Flow and Total Suspended Solids Events UF9206S-280,
300 194
7.31 Time Series-Flow and Total Suspended Solids Event UF9206S-317 195
7.32 Time Series-Flow and Total Suspended Solids Event UF9206S-336 195
7.33 UF9206N Phosphorus Content of TSS as a Function of TSS
Concentration 197
7.34 UF9206S Phosphorus Content of TSS as a Function of TSS
Concentration 197
7.35 UF9206N Phosphorus Content of TSS as a Function of TSS
Concentration (Expanded Scale) 198
7.36 UF9206S Phosphorus Content of TSS as a Function of TSS
Concentration (Expanded Scale) 198
7.37 Layout of UF9200 with Synoptic Sampling Locations 203
7.38 UF9200 Intensive Study-Event 252 205
7.39 UF9200 Intensive Study-Event 258 205
7.40 UF9200 Intensive Study-Event 262 206
7.41 UF9200 Intensive Study-Event 285 206
7.42 UF9200 Intensive Study-Event 258 South-East Canal Sample Locations,
First 10 Hours 209
7.43 UF9200 Intensive Study-Event 262 South-East Canal Sample Locations,
First 12 Hours 209
7.44 UF9200 Intensive Study-Event 285 South-East Canal Sample Locations,
First 32 Hours 210
8.1 UF9200 Two-Point Adsorption/Desorption Curves (Adsorption Mode) 236
xvii

8.2 UF9200 Two-Point Adsorption/Desorption Curves (Desorption Mode) 236
8.3 UF9206N Two-Point Adsorption/Desorption Curves (Adsorption Mode) 237
8.4 UF9206N: Two-Point Adsorption/Desorption Curves (Desorption
Mode) 237
8.5 UF9200 Desorption Data Eadie-Hofstee Plot 239
8.6 UF9206N Desorption Data Eadie-Hofstee Plot..... 239
9.1 Simplified Schematic of UF9200 DUFLOW Nodal Structure 257
9.2 Calculated and Measured Flows, Event UF9200-237 261
9.3 Calculated and Measured Flows, Event UF9200-252 261
9.4 Calculated and Observed Levels, Event UF9200-220 262
10.1 Simulation Results-Total Suspended Solids for Event UF9200-220 282
10.2 Simulation Results-Total Suspended Solids for Event UF9200-237 282
10.3 Simulation Results-Total Suspended Solids for Event UF9200-252 283
10.4 Simulation Results-Total Suspended Solids for Event UF9200-258 283
10.5 Simulation Results-Total Suspended Solids for Event UF9200-262 284
10.6 Simulation Results-Total Suspended Solids for Event UF9200-285 284
10.7 Simulation of TSS at Selected Nodes for Event UF9200-237 287
10.8 Simulation of EM at Selected Nodes for Event UF9200-237 287
10.9 Simulation Results-Cumulative Suspended Solids Load for Event
UF9200-220 290
10.10 Simulation Results-Cumulative Particulate Phosphorus Load for Event
UF9200-220 290
10.11 Simulation Results-Cumulative Suspended Solids Load for Event
UF9200-237 291
10.12 Simulation Results-Cumulative Particulate Phosphorus Load for Event
UF9200-237 291
xviii

10.13 Simulation Results-Cumulative Suspended Solids Load for Event
UF9200-252 292
10.14 Simulation Results-Cumulative Particulate Phosphorus Load for Event
UF9200-252 292
10.15 Simulation Results-Cumulative Suspended Solids Load for Event
UF9200-258 293
10.16 Simulation Results-Cumulative Particulate Phosphorus Load for Event
UF9200-258 293
10.17 Simulation Results-Cumulative Suspended Solids Load for Event
UF9200-262 294
10.18 Simulation Results-Cumulative Particulate Phosphorus Load for Event
UF9200-262 294
10.19 Simulation Results-Cumulative Suspended Solids Load for Event
UF9200-285 295
10.20 Simulation Results-Cumulative Particulate Phosphorus Load for Event
UF9200-285 295
10.21 Correlation of Initial Erodable Mass with Interevent Time 297
10.22 Chronological Values of Cem(0) 299
10.23 Cemio) Correlated as an Exponential Growth Equation 300
10.24 Suspended Solids Simulation Results for Event UF9200-319 307
10.25 Suspended Solids Simulation Results for Event UF9200-355 307
10.26 Cumulative Suspended Solids Load Simulation Results for Event
UF9200-319 308
10.27 Cumulative Suspended Solids Load Simulation Results for Event
UF9200-355 308
10.28 Cumulative Particulate Phosphorus Load Simulation Results for Event
UF9200-319 309
10.29 Cumulative Particulate Phosphorus Load Simulation Results for Event
UF9200-355 309
xix

11.1 Effect of Transportable Mass on Exported Suspended Solids for Event
UF9200-220 319
11.2 Simulated Effects of Sediment Trap on Event UF9200-220 320
11.3 Simulated Erodable Mass Concentrations at Selected Sections of South-
East Canal for Trap Simulation 321
11.4 Effect of Pumping Modification (Simulation A) 324
11.5 Effect of Pumping Modification and Flow Width Modification
(Simulation B) 325
11.6 Effect of Pumping Modification and Flow Width Modification
(Simulation C) 326
11.7 Solids Recycle Simulation RCL1 Short Duration, Short Distance 328
11.8 Solids Recycle Simulation RCL2 Long Duration, Long Distance 329
11.9 Solids Recycle Simulation RCL3 Long Duration, Long Distance,
Higher Flow, No Backflow Prevention 330
11.10 Simulation RCL4 Long Duration, Short Distance, Higher Flow,
Backflow Prevention 331
11.11 Specific Effect of Reduced Velocity via Constant Discharge 333
11.12 Water Elevation Trends for UF9206 Pump Stations 336
11.13 Water Elevation Trends for UF9200 Pump Station 336
11.14 Correlation of TSS Concentration with Canal Level at UF9206N 338
11.15 Correlation of TSS Concentration with Corrected Canal Level at
UF9206S 339
11.16 Correlation of TSS Concentration with Canal Depth at UF9200 339
11.17 PP Content vs. TSS Correlations for All Three Farms 341
11.18 Cumulative Fractional PP Load vs. TSS for All Three Farms 342
11.19 Water Level Control Example for UF9206N 344
11.20 Water Level Control Example for UF9206S 344
xx

D.l CRAF Approximate Model: Velocity-Depth Profile 366
D.2 CRAF Approximate Model: Velocity-Depth Profile (Dimensionless in
Velocity) 366
E.l Rainfall at UF9200 Julian Days 150-356 370
E.2 Selected Water Levels at UF9200 Julian Days 150-300 371
E.3 Selected Water Levels at UF9200 Julian Days 300-365.. 372
G. 1 Layout of UF9200 386
G.2 Layout of UF9206 387
G.3 Recycle Simulations UF9200 388
xxi

NOMENCLATURE
a
A
B
C
Co
C,
C
Cem
Cem(O)
Cj
Css
CSME
CSME
CSMEiou
CSMEh,gh
d
D
Dcff
EMj
Fa
Fe
Fn
datum level elevation
area under water column, or crossectional area
water surface width
concentration in solution or suspension
initial concentration at time zero
concentration at datum level a
concentration of suspended sediment
surface concentration of erodable mass
concentration of erodable mass at time zero
concentration of suspended particles of class
volumetric concentration of suspended solids
Cumulative Specific Mass Eroded
CSME at start of application of new shear stress
CSME at the low level of erosion
CSME at the high level of erosion
particle diameter
dispersion coefficient
effective drain depth
erodable mass of class / present at time t
deposition flux
erosive phosphorus flux
net phosphorus flux
XXII

sedimentation phosphorus flux
weight rate of bed transport per unit width
water column height or actual canal water surface elevation
Julian Date
partition coefficient or surface roughness
grain roughness
saturation constant for sorbate-substrate system
saturated hydraulic conductivity
water level difference between mid-field and ditch
Mannings coefficient of roughness
production of a constituent per unit section length
mass of phosphorus desorbed from sediment
phosphorus content of discharged TSS for the first four hours of an event
phosphorus content of discharged TSS for the remainder of an event
phosphorus content of discharged TSS for station UF9206N
phosphorus content of discharged TSS for station UF9206S
particulate phosphorus concentration in the discharge for the first four
hours of an event
particulate phosphorus concentration in the discharge for the remainder of
an event
volume rate of water flow per unit width
volume rate of suspended solids discharge per unit width
flow rate
wall Reynolds number
hydraulic radius of channel
grid oscillation frequency (revolutions/min)
ditch spacing

S bed slope, or hydraulic gradient, or substrate concentration of srbate, or
the quantity of a constituent passing a cross-section per unit time
So native srbate concentration originally present on substrate
Sa substrate concentration of srbate adsorbed from liquid phase
Sm maximum srbate concentration at complete saturation of sorption sites
t time
TSS total suspended solids concentration
u velocity
U shear velocity
V volume of water column, or channel velocity
Vc critical channel mean velocity for erosion
VCJ velocity which corresponds to tcj
Vmean mean stream velocity
w specific weight of water
ws settling velocity of sediment
wSed sedimentation velocity
Wj Sedimentation velocity of particles in class /
x x-coordinate, or hMV2
y elevation, or distance from surface
yo water surface elevation
z elevation of sediment surface above datum plane
Greek Letters
a rate coefficient, or erosion constant
P constant of proportionality
y specific mass of water
XXIV

ys specific mass of sediment
E erosion rate, or erosion coefficient
Ef floe erosion rate
Em erosion rate constant
1 ^
y<¡
T|a reference level
k von (Carman's constant, or the collected constants spn2
v kinematic viscosity
p water mass density
r shear stress
to surface shear stress
Tb shear stress applied to bed
tc Critical shear stress for erosion
Ted critical shear stress for deposition
Tcj Critical shear stress for erosion of class j"
Td critical depositional shear stress
ts Bed shear strength
AV CRAF ring-channel relative velocity
Abbreviations
BMP Best Management Practices
CRAF Counter-Rotating Annular Flume
CSFFCD Central and South Florida Flood Control District
xxv

CSME
Cumulative Specific Mass Eroded
EAA
Everglades Agricultural Area
EAAEPD
Everglades Agricultural Area Environmental Protection District
ENP
Everglades National Park
EPA
Everglades Planning Area
EPD
Environmental Protection District
EREC
Everglades Research and Education Center
FPOM
Fine Particulate Organic Matter
IFAS
Institute of Food and Agricultural Sciences
IP
Insoluble Phosphorus
OFCD
Okeechobee Flood Control District
PES
Particle Entrainment Simulator
PP
Particulate Phosphorus
SFWMD
South Florida Water Management District
SP
Soluble Phosphorus
STA
Storm Treatment Area
SWET
Soil and Water Engineering Technology
TDP
Total Dissolved Phosphorus
TP
Total Phosphorus
TSS
Total Suspended Solids
USACE
United States Army Corps of Engineers
WCA
Water Conservation Area
XXVI

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
PARTICULATE PHOSPHORUS TRANSPORT IN THE
WATER CONVEYANCE SYSTEMS OF THE
EVERGLADES AGRICULTURAL AREA
By
James Donald Stuck
December, 1996
Chairman: Kenneth L. Campbell
Co-Chairman: Konda R. Reddy
Major Department: Agricultural and Biological Engineering
The Everglades Agricultural Area (EAA) is a source of phosphorus nutrients to
the downstream watershed, including Everglades National Park. This work considered
export of particulate phosphorus (PP) from large EAA farms that are extensively drained
by pumped canal systems.
The research objectives were to determine the major sources, locations, and mode
of transport of PP, develop a model for PP transport and apply it to a representative farm,
using an existing hydrodynamic model adapted to include subsurface drainage.
Representative sediments had organic matter and phosphorus contents of about
75% and 0.1%, respectively, with weak phosphorus adsorption characteristics. Typical
farm canals contained sediment phosphorus equivalent to 15 to 30 years of discharged
PP. Sediment transport properties, studied in an annular flume and a particle entrainment
simulator, were used to develop a prototype model that included the concepts of a critical
xxvii

shear for erosion, supply limitation of erodable mass, and reversible erosion-
sedimentation.
Field studies showed the exported PP also had weak sorption characteristics but
had higher, more labile phosphorus content than farm sediments. Additional studies led
to the hypothesis that canal biological growth was the primary source of exported PP.
The model, modified to incorporate this concept, was calibrated on the PP discharge data
from the target farm during a normal wet season. It was validated on two tropical
depression events, simulating PP export within about 10% of actual.
Farm management practices to reduce exported PP were simulated using the
model. Macrophytereduction and velocity control were shown to be critical factors.
Sediment traps that do not reduce velocity appreciably were shown to be of limited value.
In the simulations, canal modifications that reduced pump intake velocity reduced the
first flush phenomenon, where a large fraction of the PP load is discharged at pump start
up. Level control and elimination of pump cycling were shown to be cost-effective PP
reduction mechanisms. The removal of untransported detritus from main canals was
shown to be necessary for long-term control. Hydraulic transport schemes to achieve
removal were recommended.
The model may be adapted to other farms to optimize water table control practices
which minimize PP export.
xxviii

CHAPTER 1
INTRODUCTION, BACKGROUND, AND RESEARCH MOTIVATION
Historical Summary
The South Florida Everglades is a natural resource unique in the United States, if
not in the world. The original Everglades watershed (See Fig. 1-1) was primarily a broad,
freshwater marsh that extended from what is now the Kissimmee River basin, through
Lake Okeechobee to the southern tip of the Florida peninsula, encompassing some 7500
square miles (-20,000 square km) of upland and wetland territory. The natural flow,
arising from a shallow land slope of only a few cm/km, was at very low velocity from
north to south along a riverbed which was often 80-100 km wide, with ultimate discharge
into the waters surrounding the southern tip of Florida (Jones, 1948).
Flow was episodic and seasonal in nature, with Lake Okeechobee acting as a
buffering reservoir. During the wet season (May through September) the level of Lake
Okeechobee would rise above its shallow southern bank heights, providing a consistent
flow of water through the wide riverbed. In the dry season the level of Lake Okeechobee
would fall below the level of its southern banks, essentially shutting off the flow of water
from the Kissimmee and Okeechobee basins to the land farther south. Under these
conditions of reduced inflow and precipitation the marshlands south of Lake Okeechobee
would slowly dry until, in the spring season, they would be subject to frequent and
extensive fires. The onset of the wet season in the late spring would quench the fires and
complete the annual wet-dry cycle (Douglas, 1988).
1


Figure 1.1: Original Everglades Watershed (Adapted from Jones, 1948)

3
It is important to note for later reference that for thousands of years the hydrologic
regime in the Everglades watershed was predominantly that of large volume discharge
spread over shallow channel beds tens of kilometers wide flowing at velocities that were
in the range of meters/day (Douglas, 1988).
The first serious human disturbance of the Everglades watershed.camein 1883
when developer Hamilton Disston dug a canal between Lake Okeechobee and the
Caloosahatchee River in an attempt to drain a 1.6 million ha (4 million acre) tract of
wetlands. From that time through the present the planning and implementation of water
management in the Everglades region has passed from the control of wealthy developers
through a series of state trustee boards, commissions, and agencies. From 1906 to 1931
the Everglades Drainage District (EDD) followed by the Okeechobee Flood Control
District (OFCD) installed a number of canals, levees, locks, and dams in the Everglades
Region.
One of the objectives of the public works activity was to drain mucklands just
south of Lake Okeechobee for agricultural use. This region, known as the Everglades
Agricultural Area (EAA), consists of 700,000 acres (~283,000 ha) of rich organic soil that
has supported a variety of agricultural activities since its creation. Currently, major
winter vegetable and sugarcane industries are the predominant activities in this area.
Because of the low elevation and flat terrain extant in the EAA and the swings in water
supply from wet to dry season the crops grown there are highly dependent on artificial
control of water table levels by off-farm pumping during wet periods and irrigation
during dry periods. Much of the man-made hydraulic system installed in South Florida
has had as one of its objectives the facilitation of water table control in the EAA.
In 1948 joint state-federal action created the Central and Southern Florida Flood
Control District (CSFFCD), which acted in concert with the United States Army Corps of
Engineers (USACE) to "develop and implement plans to provide flood protection, ensure
adequate water supply, prevent saltwater intrusion along the Florida Lower East Coast,

4
enhance the regions fish, wildlife, and other environmental resources, and provide a
water supply for the Everglades National Park" (Izuno and Bottcher. 1994). The
CSFFCD has since been succeeded by the South Florida Water Management District
(SFWMD), which assumed the responsibilities held by the CSFFCD in 1972.
Cooperation with the USACE has proceeded uninterrupted since 1948 between first the
CSFFD and then the SFWMD.
During the tenure of the CSFFCD major modifications were made in the hydraulic
patterns and land use of thousands of square kilometers of Everglades wetlands. The
EAA was completely canalized and diked. Fifteen hundred square miles (-3900 square
kilometers) of wetlands to the south between the EAA and the Everglades National Park
(ENP) were designated as Water Conservation Areas (WCAs) and dedicated to fish and
wildlife conservation and surplus water storage. The storage capacity of Lake
Okeechobee was increased and the capability to backpump farm stormwater runoff (or
more appropriately farm stormwater "pumpoff') from the EAA collector canals upstream
to the lake was implemented.
Unfortunately, during these many decades of major public works projects little
attention was paid to the water quality ramifications of the profound hydraulic and land
use modifications being imposed on the Everglades region. The ecological community in
the Everglades originally evolved as an oligotrophic system, that is, one with low
biological productivity, with phosphorus as the limiting nutrient. Studies conducted at
the WCAs and ENP indicate that high nutrient concentrations, including phosphorus as a
principal contributor, are causing significant shifts in the balance of the ecosystem
(Whalen et al., 1992). According to the South Florida Water Management District,
agricultural activities in the EAA, especially drainage and fertilization practices, have
resulted in a major decline in the quality of water entering the north end of the Everglades
Planning Area (EPA), the area encompassing the WCAs and ENP (Whalen et al., 1992).

5
The relationship of the EAA with Lake Okeechobee was recognized in the 1970s
as being detrimental. A number of studies were conducted that linked the eutrophication
of the lake to agricultural nutrient inputs, of which the EAA was a major contributor
(Izuno and Bottcher, 1994). A number of remedial actions have been implemented to
reduce the nutrient load on Lake Okeechobee, including eliminating_backpumping to the
lake fromjhe EAA for all but a few small private drainage districts. The effect of this
policy is to force essentially all farm water discharge in the EAA to flow south to the
WCAs, which, in the absence of any remediating action, would further increase the
nutrient load vectored from the EAA toward the WCAs and the ENP.
Sources of Phosphorus in the EAA
Inputs to the potentially mobile phosphorus pool in the EAA can come from
fertilizer application, rainfall, inflow water from Lake Okeechobee, and the natural
mineralization process of the organic soil, where organic matter is converted to carbon
dioxide and inorganic compounds. Fertilizer phosphorus input to the EAA has been
estimated to be on the order of 18% of the total annual phosphorus input (Morris, 1975),
however the fraction of this input that is available for transport to the drainage water is a
strong function of fertilizer placement, crop uptake, soil microbial pool, and the iron and
aluminum content of the soil (Sanchez and Porter, 1994). Rainfall is a poorly quantified
but non-trivial source of phosphorus input. Studies on-going at the time of this writing
indicate that the phosphorus content of rainfall in the EAA may be on the order of 0.05-
0.08 mg/1 (Izuno, 1995). At an EAA area of 280,000 ha with an average annual rainfall
of 1.45 m the rainfall load might be in the range of 200,000-325,000 kg/year, or about
0.36-0.58 kg/ha-yr.
Precise estimates of the amount of phosphorus supplied to the EAA from Lake
Okeechobee are difficult to generate because of the flow-through nature of the canal

6
system, the reuse of upstream farm discharge water by downstream farms, and the
surprising lack of historic data on phosphorus concentrations in water directly at the lake
outlets. A rough approximation of the maximum potential supply may be made using
published average figures. Mean annual total phosphorus in Lake Okeechobee ranged
from 0.05 to 0.1 mg/1 for a 12 year period from 1973-1985 (Izuno et al 1991). Annual
flows from the EAA have been reported to be in the range of 1.2 x 109 m3 at the end of
this period (Izuno, 1994). Assuming approximate equivalence between inflow and
outflow, a potential maximum loading on the order of 50,000-100.000 kg/year may be
estimated, or 0.18-0.36 kg/ha/yr.
The soils in the EAA are primarily Histosols. with depths ranging from 3 meters
to 0.3 meters. The soil is underlain by marl and limestone. Approximately 200,000 ha
are planted to sugarcane, about 40,000 ha are planted to vegetables, rice, and sod (Izuno
et al., 1991). Phosphorus concentration in virgin EAA Histosols is estimated by Sanchez
and Porter (1994) to be in the range of 0.08-0.35%. They note that in virgin soil about
24% of the phosphorus is inorganic, whereas in cultivated EAA soils up to 72% of the
phosphorus is inorganic, implying significant mineralization resulting from cultivation.
They further estimate that potential phosphorus mineralization is on the order of 20-150
kg/ha/yr for EAA soils. This estimate may be compared with Koch's (1991) estimate of
mean total phosphorus storage in the upper 30 cm of EAA soil of 276 kg/ha. These
quantities may be put into some perspective when it is calculated that 50.8 cm (20 in) of
annual rainfall runoff from one hectare need only dissolve 0.508 kg of phosphorus to
have a dissolved phosphorus concentration of 0.1 mg/1. The concentration of 0.1 mg/1 is
well above the natural level prevailing in the uncultivated interiors of the WCAs (see
below) or the ENP.
Organic soils such as the EAA Histosols were formed under submerged, reducing
conditions. When they are drained, cultivated, and aerated, they become net sources of
phosphorus due to mineralization arising from increased aerobic microbial activity

7
(Sanchez and Porter. 1994). It has also been postulated that cropping increases the net
rate of mineralization by further increasing the acti vity of the aerobic microbial
population (Cosgrove, 1977). For these and other reasons related to crop and water
management, the activities in the EAA have combined to increase the phosphorus loading
outside the EAA.
Izuno et al. (1991) illustrate the increase with some analytical results which
indicate that the water pumped to Lake Okeechobee in the period 1983-1985 had a total
phosphorus concentration in the range of 0.19-0.57 mg/1, compared with the lake's mean
total phosphorus content of 0.05-0.10 mg/1. During the period 1978-1986 water pumped
to the WCAs had total phosphorus concentrations in the range of 0.07-0.16 mg/1. The
total phosphorus content in water in the interior of the WCAs, on the other hand, was in
the range of 0.009-0.014 mg/1 (Whalen et al., 1992), a factor of 6 to 7 lower than the
incoming water. These figures serve to highlight the cause for concern regarding the
modification of the oligotrophic status of the WCAs, and ultimately of the ENP.
Proposed Regulatory Remediation
The South Florida Water Management District, under the terms of the Florida
House of Representatives' "Marjory Stoneman Douglas Everglades Protection Act of
1991". has the responsibility for ensuring mitigation of the phosphorus efflux from-the
EAA to the WCAs and the ENP. The District has proposed a remediation plan that
requires implementation by EAA growers of a series of Best Management Practices
(BMPs), developed by Bottcher and Izuno (1993), and a massive treatment plan that
involves the construction of four Stormwater Treatment Areas (STAs) within the EAA.
The BMPs, which are to affect 25% of the phosphorus load reduction, fall into
two categories; fertility BMPs and water management BMPs.

8
Fertility BMPs include the use of calibrated soil testing to minimize over
fertilization, banding of fertilizer to minimize placement of fertilizer in areas where it will
not be accessible to the root zone, implementation of methods to prevent misplacement of
fertilizer into drainage ditches, and split multiple applications of smaller loads of fertilizer
to more closely match the crop uptake. Water management BMPs are aimed at
maintaining the water table as close as possible to the maximum feasible level at all times
to minimize oxidation andmineralization of the soil and recycling of runoff to retain
phosphorus on-site, where possible. They include measures to minimize water table
fluctuations, maintain uniform spatial and temporal distribution of the water table, and
store runoff on fallow land or water insensitive cropland.
The water management BMPs require the grower to exercise much more
sophisticated control[over runoff, irrigation, and ground water than has been practiced in
the past. The grower is required to develop and track field water budgets, which
incorporate irrigation, rainfall, and evapotranspiration accounting to optimize pumping
and minimize excessive downward water table movement. Minimum canal-level pump
shut-off controls are recommended to reduce excessive canal drawdown. Comparison of
predicted rainfall with available field water storage capacity is recommended to reduce
unnecessary prepumping in anticipation of storms. A number of hydraulic BMPs relate to
achieving and maintaining hydraulic uniformity within the farm canal system to minimize
unnecessary local drawdown of the water table. Water management BMPs are also
detailed for retention of drainage water in on-farm reservoirs where possible, transfer of
drainage water from more water-sensitive crops to fields with less water-sensitive crops,
and the use of aquatic cover crops such as rice for uptake of phosphorus released
elsewhere on-farm.
The STAs, which are to affect 75% of the phosphorus load reduction, are planned
to be large artificial wetlands which will act as phosphorus removal buffers between the
EAA and the WCA/ENP areas. These STAs may ultimately cover as much as 14,500 ha,

9
and are estimated to cost on the order of $320 million (Whalen et al., 1992). Cost of
construction of these facilities will be borne, in part, by assessments on the EAA
landowners. The current STA design goal is to achieve an average effluent total
phosphorus concentration of less than 0.05 mg/1 (Whalen et ah, 1992). The design of
each of the four STAs is based on treating a long-term average load from the sub-basin
within the EAA served by the STA. This average load is determined from historical data
and corrected for land removed from production for the STAs and for expected reduction
in phosphorus loading resulting from grower implementation of BMPs.
Implementation of the combination of BMPs and construction and operation of
the STAs was projected in 1992 to reduce phosphorus input to the WCAs by 80% within
five years after initiation (Whalen et ah, 1992).
Areas of Uncertainty in Remediation Plan
The possibility exists that the actual input to the STAs may be lowered below the
long-term historical averages by more than 25% by implementation of the BMPs
recommended by Bottcher and Izuno, by implementation of district-wide water
management practices not yet specified, and by exploitation of biological and chemical
processes in the irrigation ditches and drainage canals. The possibility also exists, with a
somewhat lower probability, that release of phosphorus currently stored in the canal and
drainage ditch sediments may tend to buffer the exit phosphorus concentrations from the
various sub-basins and reduce the impact of the BMPs on the phosphorus load to the
STAs. Significant changes of loading in either direction could have a substantial impact
on the amount of land required for the STAs, or, alternatively, on the effluent phosphorus
concentrations achieved by the STAs.
For any given time period, weather is always an uncertainty. Measurement of the
effectiveness of any specific remediation strategy may be hampered by changes in the

10
precipitation and evapotranspiration patterns between time periods The long-tenn "end-
of-pipe" phosphorus loads to the ENP are set with some correction for antecedent flow at
Shark River Slough (Whalen et al., 1992), which was based on statistical analysis of
loading over an 11 year period, but statistical correlation over a multi-year period of
multiple combinations of BMPs at the farm level is not an efficient method for evaluating
and optimizing the effectiveness of BMPs. More precise correction methods are needed.
An appropriate alternative to long-term statistical evaluation is development and
application of mathematical models which adequately describe the dominant
hydrodynamic and chemodynamic processes controlling phosphorus transport and export
at the field, farm, and basin scale. Such models could be used to forecast the impact of
current or proposed BMPs under very site specific conditions, and to evaluate the actual
performance of BMPs under field conditions which deviate from the long-term mean.
EAA Modeling Activity
The need for, and value of, models for phosphorus transport in the EAA has been
recognized by both the growers and governmental agencies. One specific agency that
combines aspects of both is the Everglades Agricultural Area Environmental Protection
District (EPD), a state chartered organization with assessment powers that coordinates the
environmental regulatory compliance activities of growers in the EAA. Among
numerous research activities, the EPD has provided funds for BMP research.
The research is aimed at optimization of the original set of BMPs recommended
by Bottcher and Izuno (1993) and development and application of new or modified BMPs
for reduction of phosphorus export from the EAA. An integral part of this research
activity is the development of field and farm scale models, specific to the EAA
environment, for phosphorus sourcing and transport. This effort was originally initiated
by the University of Florida Institute of Food and Agricultural Sciences (IFAS) and

11
continues at the time of this writing as a joint effort between IF AS and Soil and Water
Engineering Technology (SWET), a private agro-environmental engineering firm in
Gainesville, Florida. It is under this overall research charter that the work described in
this dissertation was conducted.
The original water management BMPs focused on reducing phosphorus discharge
by reducing the volume of water discharged from the farm and by minimizing water table
fluctuations, that have an accelerating effect on soil mineralization and phosphorus
release. Thus the focus was on procedures that could reduce hydraulic load and minimize
phosphorus solubilization. The contribution of particulate matter to phosphorus export
was recognized (Izuno et al., 1991). however there were insufficient data available at the
time of formulation of the original BMPs to quantitate the impact of management
practices on particulate phosphorus transport and export. As a consequence none of the
original water management BMPs spoke directly to control of particulate phosphorus
transport. In addition, the main focus of the early modeling effort was on field scale
transport, with the irrigation and drainage systems existing as boundary conditions. This
resulted in an exclusion of the conveyance systems (field ditches, farm collector canals,
main farm canals, pumping systems, and ultimately the district canals) from analysis and
evaluation regarding phosphorus sourcing, assimilation, or transport. The impetus for the
work reported herein was the need to fill that recognized gap by development of a first
generation model which would give a reasonable approximation of the predominant
phosphorus mobilization, transformation, and transport processes prevalent in the farm
irrigation and drainage conveyance systems of the EAA.

CHAPTER 2
PRIOR DATA RELEVANT TO PARTICULATE PHOSPHORUS TRANSPORT IN
THE EVERGLADES REGION
Introduction
The salient characteristics of the Everglades such as topography, soils and
sediments, climate, and hydrology are not typical of those found in agricultural areas
elsewhere in the United States. The area is extremely flat land at low elevation,
frequently subject to intense local rainfall and periodically subject to intense regional
rainfall, with drainage and irrigation under almost total human control.
The soils of the area are histosols, poorly drained muck and peaty soils low in
mineral content but rich in organic matter (in excess of 65% by weight), with low specific
gravity and low hydraulic conductivity (Snyder, 1994), bearing more similarity to the
soils and sediments of the Louisiana bayous than to upland mineral soils.
The climate is sub-tropical. The annual average solar radiation and wet season
temperatures are more similar to those prevailing in the Caribbean and Central America
rather than the United States southeast and western plains agricultural regions (Marsh,
1987. Trewartha and Horn, 1980).
The heavy reliance on pumping for water transport introduces flow-no-flow
situations and hydraulic transients which are not typical of a more normal upland
watershed. In many respects the hydrology of the EAA is more akin to that of the polder
region of the Northern Netherlands where discharge hydrology is completely under
human control (Snyder, 1994).
Given the individualistic character of the Everglades region in general and the
EAA in particular it is not unreasonable to assume that particulate matter transport
12

13
standards which apply well to other locations in the United States might not translate
completely to the Everglades environment. This assumption has underlain the planning
and execution of the research work in this dissertation. With this in mind it may be
appropriate to first examine the limited body of work available which relates directly to
particulate matter and particulate phosphorus transport in the Everglades Region.
Farm-Field Scale Studies
There are two studies which have been done on the field or farm scale which are
particularly pertinent to phosphorus transport in the EAA. They are those done by the
engineering firm CFI2M Hill (1978) on the farm scale and by Izuno et al. (1991) on the
field scale.
CH2M Hill studv
T~
The CH2M Hill study evaluated a number of water quality parameters and was
conducted over a period of 15 months in 1976-77 on multiple farm sites within the EAA.
Intensive study was directed at one sugarcane field, one cattle ranch, and one vegetable
farm which grew a variety of vegetables throughout the study period. In addition multiple
secondary sites for each land use were studied less intensively and served as checkpoints
for the representativeness of the intensive data. Key conclusions from the study regarding
phosphorus follow.
The sugarcane farm showed study-period average effluent total
phosphorus of 0.110 mg/1, soluble phosphorus of 0.070 mg/1 (64% of
total), particulate phosphorus of 0.040 mg/1 (36% of total) and an annual
effluent total phosphorus load of 0.65 kg/ha.

14
The vegetable farm showed study average effluent total phosphorus of
0.460 mg/1, soluble phosphorus of 0.355 mg/1 (77% of total), particulate
phosphorus of 0.105 mg/1 (23% of total) and an annual effluent total
phosphorus load of 2.10 kg/ha, roughly four times that of the sugarcane
farm.
The cattle ranch showed study-period average effluent total phosphorus of
0.167 mg/1, soluble phosphorus of 0.122 mg/1 (73% of total), particulate
phosphorus of 0.045 mg/1 (27% of total) and an annual effluent total
phosphorus load of 0.55 kg/ha. Concentrations were higher than for
sugarcane but load was lower because less water was discharged.
Soil water was monitored. Average soil water soluble phosphorus was
5.74 mg/1 at the vegetable farm, 3.88 mg/1 at the cattle ranch, and an
extremely low 0.04 mg/1 at the sugarcane farm.
Variability was high. Most phosphorus concentrations showed standard
deviations approaching or exceeding the mean.
Seasonal variation of phosphorus concentrations was much higher for
vegetables than for sugarcane.
Phosphorus applied as fertilizer per unit area was about 6.5 times as high
for vegetables as for sugarcane.
A comparison of all studied sites and additional data obtained from
sampling of private drainage districts during the study indicated that the
cropland which had been under cultivation longer had less ability to retain
phosphorus.

15
Izuno et al. study
The work by Izuno et al. (1991) was a study carefully designed to allow
evaluation of phosphorus discharge from various crop and field conditions before and
after implementation of BMPs. The baseline (before BMP) study was reported in the
1991 publication. Replicate fields of sugarcane, fallow, flooded fallow, and selected
specific vegetables were evaluated over a 10-13 month period. Key conclusions follow.
The phosphorus concentrations in the sugarcane fields averaged 0.28 mg/1
total phosphorus, 0.17 mg/1 soluble phosphorus (61% of total) and 0.11
mg/1 particulate phosphorus (39% of total). The cabbage fields averaged
0.56 mg/1 total phosphorus, 0.31 mg/1 soluble phosphorus (55% of total)
and 0.25 mg/1 particulate phosphorus (45% of total). The radish fields
averaged 0.25 mg/1 total phosphorus, 0.20 mg/1 soluble phosphorus (80%
of total) and 0.05 mg/1 particulate phosphorus (20% of total). It was noted
that radish has a lower fertilization requirement than cabbage.
The effects of hydraulics were evident in this study. The displacement
phenomenon was noted, wherein a portion of the initial contents of the
ditch and the surface runoff must pass through the pump before the effects
of the field groundwater on soluble phosphorus concentration are seen.
Main farm canal discharges averaged 0.16 mg/1 total phosphorus, 0.08
mg/1 soluble phosphorus (50% of total) and 0.08 mg/1 particulate
phosphorus (50% of total), which was consistently lower in phosphorus
concentration than the effluent from the fields, suggesting that either
dilution or removal of phosphorus was taking place in the irrigation
conveyance systems.

16
The drained fallow fields, which were interspersed with the sugarcane
fields and drained under the same conditions, showed effluent phosphorus
concentrations averaging 0.43 mg/1 total phosphorus, 0.2B mg/1 soluble
phosphorus (65% of total) and 0.15 mg/1 particulate phosphorus (35% of
total), which was significantly higher than the neighboring sugarcane
fields, suggesting a strong crop-effect on the availability of total
phosphorus for drainage. (Recall the very low pore-water phosphorus
found in the CH2M Hill sugarcane fields.)
The highest loads observed during the study came when flooded fallow
radish fields were drained, and immediately after fertilization of the
cabbage fields.
This study also showed high variability, with standard deviations of
phosphorus concentrations approaching the mean in most cases.
Canal Transport Studies
Several studies that contain data related to phosphorus transport in canals were
performed in the Everglades region.
Southeast Florida canals water quality studies
Lutz (1977) studied the variation of water quality with time over an 18 month
interval at various points within several Southeast Florida canals. One of the canals
studied was the West Palm Beach Canal from its point of exit from the EAA to a point 16
km (10 miles) downstream. His experimental procedure entailed taking individual grab
(non-flow weighted) samples at the various sample stations for each canal on the same
day every two weeks. Lutz data showed the same variability as was seen in the
previously mentioned farm/field studies, that is, standard deviations approximately equal

17
to the mean. He concluded that spatial variation (upstream to downstream) was not
statistically significant. Seasonal variation in the West Palm Beach Canal was also
difficult to detect in his results because month to month variations were quite significant.
What is particularly interesting in Lutz' data, however, is what arises from visual
inspection of the various time series. In about 15% of the days sampled there are
appearances of large spikes at one individual sample point which were not reflected at the
other sample points. This is true for soluble as well as total phosphorus. If the data plots
are examined from a subjective phenomenological viewpoint one might possibly interpret
the data as showing wave-like gradients in concentration down the canal on occasions of
high concentration, however such an interpretation is very subjective, given the absence
of close-interval time-series data. The main point from this study is that the data, taken at
face value, indicate that there were times during the study when the canal was seeing
short bursts of shock load with a more subjective indication of wave-like phosphorus
transport under certain circumstances when phosphorus concentrations were high.
Water Conservation Area water quality correlations
Mattraw et al. (1987) attempted to develop statistical correlations of water quality
variables to detect trends at 10 inflow or outflow locations around the periphery of Water
Conservation Area 3A. Several of the structures were direct outflows from the EAA.
They evaluated numerous model types relating to discharge and antecedent rainfall
conditions and concluded that the only correlations of significance were orthophosphate
and nitrate with 7 or 30 day antecedent rainfall conditions. No temporally significant
trends were found over a five year period. The primary result of interest in the context of
the present research is that for the one structure evaluated which corresponds to a major
EAA discharge, S-8, the best model correlation had an R2 of only 0.36. This emphasizes
the difficulty in correlation of phosphorus content using only hydrologic variables.

18
EAA Canal Sediment Studies
Several studies, which have been conducted in private and District canals within
the EAA, have particular relevance to the area of particulate phosphorus transport. For
the most part these activities have been private, non-peer-reviewed studies that have had
limited report circulation. Nonetheless they have tended to form a basis for the
perspective of the EAA grower community toward potential phosphorus reduction
activities and require some significant attention at this point.
Anderson and Hutcheon Engineers study (1992)
Anderson et al. (1992), in a private study conducted for the Florida Sugar Cane
League reported results of some sediment core sampling conducted in the EAA's four
main and two connector canals, which are operated by the SFWMD. They also presented
data from what was reported to be a sediment transport study conducted at coded private
sites within the EAA.
The sediment core analyses showed a sediment phosphorus content in the range of
58-77 mg phosphorus/ kg dry sediment (58-77 ppm) for five of the six canals. Only the
Miami Canal sediment, which averaged 252 mg P/kg sediment, exceeded 100 ppm.
The data reported in the sediment transport study include some velocity and total
suspended solids (TSS) data in conjunction with the total phosphorus (TP), soluble or
total dissolved phosphorus (TDP), and particulate or particulate phosphorus (PP)
concentrations. Average concentrations for the total data set were TP = 0.153 mg/1, TDP
= 0.042 mg/1 (27% of total), PP = 0.111 (73% of total), and TSS = 187 mg/1.
On average the phosphorus content of the TSS was 594 mg P/kg TSS (ppm), but a
more detailed examination of the data indicates that the sample set may have represented
a bimodal distribution of samples, with one set having a phosphorus content of about 390
ppm and the other set having a phosphorus content of about 1700 ppm.

19
There was not sufficient detail of the channel configuration presented in the report
to do a substantive evaluation of transport dynamics, but there appeared to be a step
change in suspended solids concentration at a velocity somewhere between 0.06 m/sec
and 0.1B m/sec, above 0.18 m/sec there was no clear response to increasing velocity.
Included in the report were data tables representing excerpts of data from a
SFWMD sampling program which covered a nine year time span from 1983 to 1991.
The data summary presented averages of TSS = 23 mg/1, ortho-P = 0.080 mg/1, total P =
0.161 mg/1, and particulate P = 0.081 mg/1. The implication from this summary was that
the distribution of phosphorus in the SFWMD canals was split approximately 50%/50%
between soluble and particulate forms during this period. This data set is discussed in
more detail later in this section.
Andreis/U.S. Sugar Proposed sediment control BMPs (19931
The Research and Development Department of U. S. Sugar Corporation, in an
attempt to develop a particulate phosphorus control program for the EAA, informally
proposed a number of sediment-control BMPs to the grower community and the SFWMD
in 1993 (Andreis 1993). Although these practices were not proposed in formal
publications, they have been given widespread distribution, and have received varying
levels of acceptance within the grower community. They were discussed with H. J.
Andreis, Senior Vice President of Research for U.S. Sugar, by this writer in 1994. At the
time of discussion some field data were available on the effectiveness of the BMPs.
There were some sixteen BMPs recommended, which fall into five categories, as
follows:
1. Surface soil erosion reduction,
2. Stabilization of ditch and canal banks and sidewalls,
3. Minimization of disturbance of in-canal sediments,

20
4. Installation of sediment trapping configurations, and
5. Canal cleaning programs.
Data available in 1994 at the time of discussion related primarily to comparative
evaluations of test facilities that incorporated BMPs in the first three categories. Typical
evaluation consisted of comparing some measure of particulate phosphorus (PP)
concentration or load at a point immediately downstream of an installation before and
after implementation of a specific BMP. In all cases discussed the implementation of a
BMP resulted in significant reduction of the PP measure after implementation.
There were, however, no side-by-side controls run in any of the tests, and no data
were available to relate the contribution of the problem being addressed to the overall
phosphorus load from the farm of which the test plots were a part. Under these
circumstances it must be judged that the BMPs seemed to move in the direction of
goodness, but the levels of their contributions had not been quantified. Strong emphasis
was placed on the last two categories of BMPs as "pipeline and "end of pipeline"
measures but no data were available at that time to evaluate the effectiveness of these
measures because full scale tests were still in progress.
Sediment control demonstration project
A full-scale eighteen month duration test of canal dredging and sediment trapping
procedures, funded by the EPD, was completed in May of 1995 and reported by Hutcheon
Engineers (1995). At two separate locations canals were dredged for specific lengths to
construct a canal segment, with increased depth relative to the immediate upstream and
downstream sections, that would act as a sediment trap. Stated simply, the sediment trap
operates by reducing the velocity of the carrying fluid which in turn reduces the carrying
capacity of the fluid, causing some suspended sediment to precipitate. The upstream and
downstream ends of the traps were equipped with sediment barriers constructed of rock

21
piled to the height of the original sediment to prevent sloughing of existing sediment into
the study areas.
Effectiveness of the traps were measured several ways. Sediment accumulation
pans were placed within the confines of the dredged sections at the start of the study and
retrieved by a diver 18 months later at the end of the study. Bottom soundings were made
in the dredged sections on a monthly schedule to estimate the increase in bottom
elevation with time. Regular water sampling was conducted before and after dredging of
the traps and upstream and downstream of the dredged sections to ascertain the degree of
removal of phosphorus resulting from the presence of the traps.
The results were interesting. The sounding data indicated that the dredged bottom
elevations increased at the rate of 1.4-1.8 ft/year (-0.43-0.55 m/yr). The sediment
accumulation pans, which were 5 cm deep and had been placed on bare bottom at the start
of the study, proved to be not especially useful because they were buried under 20-60 cm
of overlying sediment at the time of retrieval. The water sample phosphorus analyses,
however, gave no significant indication of phosphorus reduction arising from installation
of the sediment traps. Pre-dredging and post-dredging farm effluent phosphorus contents
were virtually identical at both locations. The average phosphorus analyses were 0.108
mg/1 pre-dredging versus 0.107 mg/1 post-dredging at one location and 0.086 mg/1 both
pre-dredging and post-dredging at the other. Upstream and downstream phosphorus
measurements showed similar results. At one location upstream average phosphorus was
0.111 mg/1, downstream average phosphorus reduced to 0.096 mg/1. At the other
location, upstream average phosphorus was 0.067 mg/1, downstream average phosphorus
increased to 0.077 mg/1. In neither case was the difference statistically significant at the
p=0.30 level.

22
South Florida Water Management District Canal Data Set
The data set from which Anderson et al. (1992) extracted their excerpts was
requested from the SFWMD for evaluation. What was provided (Mucinic 1994) was an
extensive data set containing analytical data on 1,996 samples taken from 1984-1993 at
various locations along the main EAA canals, along with logged flow data for the various
pumping stations at the periphery of the EAA. The primary thrust of the sampling
program was determination of total phosphorus and ortho-phosphorus, which were taken
at each sample site roughly monthly. Samples that included analysis for total suspended
solids were taken roughly quarterly.
Upon examination of this data set and previous interpretations of its analog for
1984-1991 it became evident that there were some opportunities for misinterpretation.
First, the pumping records showed that there were numerous occasions when samples
were taken that there was no flow in the sampled canals. Second, there was no
independent analysis of particulate or particulate phosphorus. Prior evaluations of this
data set had calculated particulate phosphorus as the difference between total phosphorus
and soluble ortho-phosphorus. This may introduce non-trivial error if there is significant
dissolved organic phosphorus present, which is not detected by the soluble ortho
phosphorus analytical technique.
The data set provided by the District did contain a small sub-set of samples, taken
from several of the northern EAA pump stations which discharge into Lake Okeechobee,
upon which both soluble ortho-phosphorus and total dissolved phosphorus analyses had
been conducted. These data were used to develop an estimate of the proportions of
soluble phosphorus which were organic (24.9%) and inorganic (75.1%). This distribution
was assumed to hold for all samples and was used as a correction factor to estimate actual
total dissolved phosphorous from soluble ortho-phosphorus data.

23
The data set was subjected to a screen for samples which met the criteria of:
1. Taken from one of the four major southern discharge stations
2. Taken during times of flow
3. Having analytical results for Total Suspended Solids, Total Phosphorus,
and Ortho-phosphorus.
This screen produced 55 samples from the entire data set which met all three
criteria. The above-mentioned soluble phosphorus correction was applied to the screened
data set. The results showed an average Total Suspended Solids (TSS) of 14 mg/1,
average total phosphorus of 0.140 mg/1, average soluble phosphorus of 0.099 mg/1 (70%
of total) and average particulate or particulate phosphorus of 0.042 mg/1 (30% of total).
Regression of particulate phosphorus on TSS indicated a phosphorus content of the
suspended solids on the order of 2200 mg P/kg TSS (ppm).
The phosphorus content of 2200 mg/kg and the presumed upper mode of 1700
mg/kg of the Anderson and Hutcheon Engineers Study are both considerably higher than
the phosphorus contents of the canal sediments reported in the Anderson and Hutcheon
Engineers Study, 58-77 mg/kg for 5 of 6 canals and 252 mg/kg for the remaining canal.
This major difference in phosphorus content poses an intriguing clue to the possible
sourcing and transport mechanisms of particulate phosphorus in the EAA. The topic will
be pursued in some depth in later chapters.

CHAPTER 3
GENERAL LITERATURE REFERENCES RELEVANT TO PARTICULATE
PHOSPHORUS TRANSPORT
Introduction
This chapter will focus on information in the open literature which may have
particular influence on the direction of the research, which may provide insight for
formulation of hypotheses presented later in this work, or which may be used to
rationalize later conclusions. The chapter begins with a discussion of the sources of
particulate phosphorus, continues with a discussion of diagenesis, or the fate of deposited
particulate matter, and concludes with a brief review of the concepts of sediment
transport, or the remobilization of deposited particulate matter.
Sources of Particulate Phosphorus
Particulate matter in ecosystems is typically classified as either allochthonous (of
external origin) or autochthonous (of internal origin). Allochthonous materials which
may be of importance in the EAA might include soil particles, livestock wastes, and
particulate agricultural chemicals transported by wind or water erosion, stream bank
eroded particles, leaf and ground cover litter transported by wind or water erosion, and
deposited detritus from streamside trees and plants. Autochthonous materials would
include aquatic macrophytes and aquatic microbes, both of which can fix soluble carbon
and nutrients in their growth and reproduction processes, streambed parent inorganic
particles, chemically precipitated inorganic materials, and invertebrate and vertebrate
organisms. Given the minimal terrain slopes, sub-tropical climate, high insolation and
24

25
plentiful organic carbon supply it is intuitively expected that the regions channels are
highly autochthonous.
Surficial agricultural residue will be a highly variable function of land use and
management practices, but dairy farm soils in the area of interest may have typical
phosphorus content of 1500-3000 mg/kg (Reddy et al., 1994). Tree leaves may contain
500-2500 mg P/kg dry weight, while woody litter may have a phosphorus content of 200-
300 mg P/kg (Reddy and DeBusk, 1987).
Macrophytes play multiple roles in the particulate phosphorus dynamics of sub
tropical streams. Because of their rapid turnover times they serve as both sinks during
growth and sources during senescence for particulate and soluble phosphorus. For
illustration, dense stands of water hyacinths may assimilate 100-300 mg P/square meter-
day by biological fixation (Reddy and DeBusk, 1987) and slough 5-15 mg P/square
meter-day from normal detrital formation (DeBusk and Dierberg, 1989). Typical floating
macrophytes may have tissue phosphorus concentrations of 1500-12,000 mg P/kg dry
weight (Reddy and DeBusk, 1987) so, unless the macrophytes are removed from the
system, the phosphorus stored in their tissue is available for release upon plant death and
senescence.
By creating zones of hydraulic quiescence macrophytes may affect sediment
deposition under normal flow conditions, but a portion of this sediment may be
remobilized during times of increased hydraulic activity (Svendsen et al., 1995). In sub
tropical and tropical climates macrophytes, particularly floating macrophytes, may host
substantial masses of loosely attached epiphytic microbial material, which effectively
increases the in-stream concentration of organic phosphorus-containing material (Engle
and Melack, 1990).
Microbial activity can play a major role in the particulate phosphorus dynamics of
a water body. In eutrophic lakes, for example, phosphorus associated with deposited
bacteria can equal or exceed that contributed by organic detritus (Gachter and Meyer,

26
1993). Microbial activity associated with litter decomposition can frequently cause an
increase in the total phosphorus concentration as the higher-phosphorus-concentration
microbes utilize the carbon source of the lower-phosphorus-concentration substrate to
which they are attached (Elwood et al., 1988). Cyanobacteria have been shown to, on
occasion, exhibit a reverse sedimentation effect where they detach from the sediment and
migrate into the water column. The conditions under which they do this also favor a high
bacterial phosphorus content so they may become an important source of suspended
particulate phosphorus under conditions of high water temperature and favorable C/N
ratios. Petersson et al. (1993) showed that particulate phosphorus flux into the water
column from this source could be in excess of 2.5 mg P/m2-day.
Phosphorus content is high for planktonic materials, for example Behrendt (1990)
reported values averaging 5100 mg P/kg biomass for diatoms, 7700 mg/kg for blue-green
algae, and 13,700 mg/kg for zooplankton. The high phosphorus content can be a source
of soluble phosphorus. Montigny and Prairie (1993) have shown that cell lysis of bacteria
can produce high levels of soluble phosphorus even in the presence of iron which would
be expected to precipitate the phosphorus.
Macrophytes and microbes typically dominate the mass of non-soluble material
present in streams, but macroscopic organisms, particularly the invertebrates, can play a
major role in modifying the physical character of the particulate matter resident in the
stream. Webster (1983) describes the complex process where invertebrate shredders and
collector-gatherers convert coarse organic matter to more transportable and accessible
fine particulate organic matter and also create a bi-directional flux of the fine matter
between the suspended and benthic compartments.
Many of the phosphorus content values noted above for autochthonous materials
may be considerably higher than those typical of average EAA basin soils or stream
sediments (See Chapter 5). This illustrates the important role of in-stream processes may
play in modifying the physical and chemical nature of the basin phosphorus cycle. The

27
actual contributions, however, will be governed by local system dynamic parameters
including carbon and nutrient availability, light source, intensity and extinction,
competitive population dynamics and size of standing crops, air and water temperature,
and hydraulic fluctuations.
Early Diagenesis
The processes of early diagenesis, or the first steps of conversion of sediment to
inert material, are important in that the process pathways and kinetics can have a
significant effect on the nutrient content of the resuspended sediments and on the amount
of sediment-contained nutrients released back to the water column. The organic sediment
constituents which have the highest phosphorus content will also tend to be the most
bioavailable (See preceding discussion of sources). A brief discussion of the expectations
of the fate of this material follows.
Lucotte (1993) has estimated that up to 50% of the P initially buried in estuarine
sediments of the St. Lawrence River are released back to the water column because of
bioturbation.
Montigny and Prairie (1993) have shown that bacteria in sediments will lyse
rapidly in the anaerobic regime, releasing high levels of soluble phosphorus. Contrary to
conventional theory, however, the soluble phosphorus did not combine with ferric iron
and precipitate when it diffused into the oxic zone. The reason for this was the chelation
of the ferric iron with simultaneously released organic acids, which removed the iron
from effective interaction with soluble phosphorus.
Menendez et al. (1993) found that the initial decomposition of macrophyte
detritus and release of phosphorus proceeded at an extremely rapid rate, about 20%/day,
in the anaerobic zone of lagoon sediments.

28
Keizer et al. (1993) found that in peaty sediments of the Netherlands the
conversion of calcium carbonate to calcium phosphate complexes was inhibited to the
point of insignificance.
These studies were not necessarily under conditions identical to those found in the
EAA watersheds, but they do tend to indicate that, unless significant quantities of
sorptive clays are present or calcium-phosphorus complexation is favored, high levels of
sediment phosphorus recycle to the water column would be expected. This is a good
place to note that Svendsen et al.( 1995) observed that the sediments in their studies
showed a long-term net retention of phosphorus of only one-half to one-third of their
short-term gross retention rates.
Erosion and Transport of Organic Material
General erosion and transport
To state the obvious, transport of particulate phosphorus within a system is
governed by the overall transport of particulate matter (PM or seston) within that system.
As noted above, sources of PM in a waterbody may include soil and ground cover
erosion, deposition from external vegetation, and internal generation and transformation.
Once deposited or generated the PM is constantly subjected to buoyancy, gravitational,
kinetic, and chemical forces, the integrated interaction of which determine the fate of the
particle.
Factors which affect the exchange of particles between their streambed or
vegetation-attached locations and the water column have a large impact on particle
transport by streams. Initiation of motion occurs when water velocity is sufficient to
create local shear stress in excess of some critical value for detachment or incipient
motion. Once moving, some particles may remain in contact with, or in close proximity

29
to, the stream bed. This material is conventionally referred to as bedload. Smaller or less
dense particles may be carried into the water column when turbulent forces are
considerably in excess of gravitational forces. This material represents the suspended
load. When turbulent suspension forces fall below their critical threshold values
suspended materials are re-deposited. Similarly when bed shear forces fall below their
critical threshold values bedload transport ceases. A variety of physical factors affect
particle transport including particle characteristics such as particle size, shape, density,
fall velocity, and electrical charge, and stream characteristics such as width, depth, flow
velocity, slope, roughness, water temperature, macrophyte population type, density and
location, and flow-stage relationships (Webster et al., 1987).
The literature on erosion of organic soils is exceedingly sparse. Studies of soil
erosion tend to focus on upland peat bogs or moors in north temperate or arctic climates.
Representative of these is the work of Labadz et al. (1991), who studied sediment yield
and delivery from blanket peat in Great Britain. They concluded that erosion from upland
peatlands is highly variable spatially and temporally with a strong stochastic component
attributable to the constant variation of surface morphology. In a similar vein, but in the
context of different terrain, Benda (1990) concluded that organic debris flows were an
important factor in determining the channel morphology of streams in the Pacific
Northwest. Benda hypothesized that the stochastic nature of sediment supply from debris
flows promotes cycling between channel aggradation and channel degradation,
accentuating temporal and spatial variability of channel morphology. These and similar
studies of upland erosion are of interest in that they may provide some qualitative
framework to understand the variability of particulate organic transport, but they do not
necessarily apply to the hydrology, hydrography, or surface morphology of the South
Florida region.
Research with more relevance to the South Florida hydrologic regime is that
which has included the study of production and movement of organic sediments in the

30
riverine, lacustrine, or estuarine environment. Cushing et al. (1993) studied the transport
of carbon-14 labeled fine particulate organic matter (FPOM) of particle size less than 100
micrometers to estimate the transport distances and residence times in the riverine
environment. They measured labeled suspended FPOM concentrations in the river as a
function of time and distance, and labeled FPOM concentration in the river sediments as
a function of time. Several interesting results were reported. The mean particle
deposition velocity, calculated from concentration gradients, was 0.43 m/hr, about one
order of magnitude less than the sedimentation velocity of particles from the same source
measured in the lab under quiescent conditions. Calculated mean time to deposition
under a mean velocity of 0.27 m/sec and average reach depth of 0.33 m was 51 minutes,
with an average of 83% of the injected FPOM being deposited in the 1 km reach study
area. However after 24 hours bottom sediment within the reach contained only 1% of the
originally deposited material. The authors concluded that particles in surficial sediments
exchange rapidly with the water column and migrate episodically throughout a riverine
system, leading to strong longitudinal connection of sediment distribution. Unstated, but
evident from the results, are the potential errors involved in attempting to apply
laboratory sedimentation test results to field conditions.
Kronvang (1992), using a synoptic water sampling design augmented with
increased sampling frequency during storm events, investigated the export of particulate
matter, with specific emphasis on phosphorus, from two Danish agricultural basins. He
found significant phosphorus content enrichment in the transported sediments compared
to typical soils in the watershed (sediment phosphorus concentrations of 7 to 14 times
those found in the soils). The phosphorus content of the particulate organic matter was
found to stay relatively constant at about 1 %, regardless of the total particulate
concentration, while the inorganic particulate phosphorus content decreased as particulate
concentration increased, but approached the phosphorus content of the local soil only at
extremely high particulate matter concentrations. The results of this study emphasize the

31
difference in sources of particulate inorganic and organic phosphorus and the apparent
selectivity of transport which tends to favor particles which happen to have higher
phosphorus content. The cumulative load chemographs for the various forms of
phosphorus are illustrative of typical load distributions observed elsewhere, including the
Everglades area. They show 50% of the annual particulate phosphorus load being
transported during less than 5% of the total observation time, which may be interpreted as
the bulk of the particulate phosphorus transport occurring during major flow events.
Findlay et al. (1991) conducted automated synoptic water sampling and stream
gauging over a 150 km reach of the lower Hudson river for a three year period. Vertical
as well as horizontal sample profiles were obtained. Profile analysis of particulate
loadings led to the conclusion that during low to medium flows particle resuspension was
as important as tributary contribution in determining river sediment loading. They found
significant contributions from both autochthonous particulate organic matter and
resuspended detrital material during the ice-free seasons and concluded that transport of
particulate organic matter was controlled to a significant extent by processes occurring
within the river and were not simply related to loadings from outside.
Godshalk and Wetzel (1984) used piston coring devices to sample sediment
transects of a small hardwater lake in Michigan. They segregated sediment particle size
fractions by wet screening and then determined the various molecular weight fractions of
humic and fulvic acids, as well as conducting total organic carbon, fluorescence, and UV
absorbence analyses. Humic and fulvic acid proportions and the carbon content of each
fraction were used to estimate the relative age of each sediment sample (Humic-older,
Fulvic-younger, increasing carbon content indicating more refractory material). The
results showed a succession of organic particulate matter sources in the lacustrine
environment. The indigenous fine particulate organic matter (FPOM) was transported
first, but there was regular production of FPOM from coarse particulate organic matter
(CPOM) as time progressed. Thus the reservoir of FPOM was replenished on an irregular

32
periodic basis determined by transport events which affected resuspension of the nascent
FPOM. The authors postulated a succession scenario wherein macrophyte colonization
causes rapid detrital accrual near-shore. The detritus is reduced in size over time by
microbial decay and the smaller particles are transported by periodic hydraulic
excursions. Lability of smaller transportable particles depends on origin, for example
planktonic versus detrital, and age. Chemical and biochemical interaction of the small
particles with the environment is a complex function of space and time (depth, light
availability, redox potential, temperature, etc.) and the history of the particle.
Calvo et al. (1991) used the Particle Entrainment Simulator (PES) to study the
relative compositions of parent sediment and resuspended material in multiple locations
in a shallow (1 m average depth) marine lagoot^. Undisturbed core samples were taken
directly in the PES cylinders by divers and then immediately subjected to typical shear
stresses in the range of 6-9 dynes/cm2 in the PES onboard ship. Suspended particulate
material was sampled and analyzed, along with parent sediment, for Total Organic
Carbon, Total Kjeldahl Nitrogen. Exchangeable Ammonia Nitrogen, and Total Organic
Nitrogen. The authors found that the carbon/nitrogen ratio was always lower for the
suspended material than for the parent sediment, indicating a lower degree of
mineralization for the suspended material, and that the C/N ratio increased as the
suspending shear stress was increased. This latter finding suggests that the more readily
transportable the material was, the younger and less decomposed it was. In addition, the
largest deviations between suspended material and parent sediment composition occurred
in the regions which were identified as having the largest standing crop of phytoplankton.
No phosphorus analyses were done, but it follows from the hypothesis regarding age and
degree of decomposition versus transportability that the same results would have been
obtained for phosphorus. This study reinforces the concept of the selectivity of transport
with respect to composition, expressed by the authors as selective resuspension of freshly
deposited material.

33
Kemp et al. (1984) carried out a comprehensive study of sediment mass flux and
chemical composition as influenced hv submersed macrophytes in a Chesapeake Bay
tributary, using a variety of sampling and analytical techniques which included harvest of
standing crops, core sampling, water sampling, respirometry studies, analyses for
chlorophyll-a, stable carbon isotopes, total carbon, total nitrogen, nutrients, suspended
material and wet sieve particle size analysis. They also included work from previous
studies which included seasonal budgets for organic carbon, sediments, and nitrogen.
They found that the vascular aquatic vegetation played a strong source-sink role in
trapping POM and retarding its movement during the growing season but contributed
about one-third of the total annual organic carbon budget during times of senescence and
death. Phytoplankton content of trapped sediment ranged from 10%-40% as estimated
from chlorophyll-a and stable carbon isotope analysis. Their annual budgets were
particularly interesting. They showed that the sediment sink load within the macrophyte
beds, expressed in kg/yr., was more than twice the total sediment source load on the study
area from river input and shore erosion, indicating that processes internal to the study area
contributed as greatly to sediment load as external inputs.
The nature of transportable organic sediment as a collection of agglomerates or
aggregates of organic and inorganic matter has been addressed by several authors.
Kranck (1984), studying transport in estuaries, found that the suspended particulate
flocculated matter in three separate estuaries had an organic content in the range of 65-
75% by volume. She theorized that this organic content represented an optimum
composition for floe formation under the estuarine conditions studied. Assuming specific
gravities of 1.1 for organic matter and 2.7 for minerals would make Kranck's values 45-
55% by weight. Mirbagheri et al. (1988) studied a California agricultural runoff and
irrigation watershed using floating single-stage and multi-depth autosamplers, Van Dorn
plankton-sampling bottles, and Ponar bottom sediment samplers. They evaluated total
suspended solids, suspended organic matter, suspended algae, chlorophyll-a and

34
suspended matter biochemical oxygen demand (BOD). They found that the suspended
matter in the channelized riverine environment averaged about 40% organic matter by
weight. The organic matter itself consisted of about 25% viable planktonic-type material
(as estimated by chlorophyll-a analysis) and 75% nonliving biomass. This nonliving
biomass was further categorized by BOD analysis as 35% (of total suspended organic
matter) biodegradable and 40% "refractory" organic matter. Bokuniewicz and Arnold
(1984), who carried out water sampling at 16 stations in freshwater reaches of the Lower
Hudson River, found the average organic content of the tidally influenced freshwater
suspended sediments to be 22% by weight.
Tipping et al. (1993) sampled suspended material in riverine environments and
determined particle aggregate sizes by microscopy. They also collected freshly
sedimented material using sediment traps placed in various hydraulic regimes within the
reaches under study and determined particle size distributions using particle counters.
Subsamples were analyzed for total mass, carbon, and nitrogen. They found that the
average particle density decreased significantly as average aggregate diameter increased,
with an increasing concentration of organic matter as aggregate size increased. They
postulated that little, if any, agglomeration takes place in-stream, rather that the
agglomerated particles either enter the stream as a result of original erosion of fields and
banks, or they form on the sediment surface in relatively quiescent (dead) zones and are
resuspended by periodic hydraulic excursions. They note that dead zones can play an
important contributory role, causing either concentration spiking or tailing when they are
disturbed. The magnitude of the disturbance and the exchange rate with the main stream
govern the type of contribution. They also point out that dead zones with normal
turnover times of days can be important contributors of phytoplankton.
The selective transport of constituents resulting from variations in particle size,
density, origin, and composition gives rise to the phenomenon known as "focusing"
which is the spatially inhomogeneous distribution of contaminants in sediments (Eadie

35
and Robbins, 1987). Degree of focusing, as would be expected intuitively, is a strong
function of biology and hydrodynamics. Bloesch and Uehlinger (1986), using a lake
wide distribution of sediment traps, found it to be relatively unimportant in a eutrophic
lake with high productivity and modest turnover velocities. Murchie (1985) used 210-Pb
dating to study the geologic history of a freshwater bay in Minnesota. He found that
focusing decreased as basinwide productivity increased, but also that high-organic-
content sediment was more intensely focused than heavier siliceous or calcareous
sediment. Kronvang and Christiansen (1986) used a combination of traps, cores, and
water samples to develop a spatially distributed sediment budget for a hydrodynamically
active estuary in Denmark. They found that focusing was strong in the upper estuary
zone and recommended specific dredging locations and seasonal times which would
allow for an optimization of the dredging effort to recovery ratio. This approach may
have particular interest in regions such as South Florida where hydrodynamics are, or can
be, partially controlled.
Specific quantitative erosion data have been generated in a few cases for the
organic fraction of sewage and stormwater flows. Kleijwegt et al. (1990) studied erosion
of cohesive synthetic sewer solids in a laboratory flume. They determined that the upper
limit of the critical shear stress for initiation of erosion of cohesive sewer sediments
appears to be in the range of 5-7 Pascals. Ashley and Crabtree (1993) categorized sewer
sediments into five classes: Class A coarse granular material. Class B agglutinated
Class A deposits, Class C mobile fines, Class D organic biofiims, and Class E -
deposits found in tanks. In a related study Ashley et al. (1993) noted that definition of
bedload was very difficult with Class C and D sediments because bedload for these
materials was in the form of a dense cloud of sediment close to the surface of the
transporting channel, as opposed to the traditional riverine definition of bedload as a
saltating layer of individual particles. The thickness and density of the cloud is affected
by hydrodynamic conditions as is the interchange of suspended particles with the bedload

36
cloud. Class C (mobile fine) solids were noted to be weakly resistant to erosion. Class D
(biofilm) solids were reported to be poorly studied and ill-defined. More research was
recommended in the area of biofilm solids transport because of their high organic matter
content. The authors reported that flume studies of freshly deposited organic sewer solids
exhibited a critical shear stress of about 1.8 Pa. and that 75% of the eroded solids had a
particle size of less than 100 micrometers.
The only literature source found which deals directly with quantitative erosion
characteristics of natural organic sediments is the work of Hwang (1989). Hwang studied
the erosion characteristics of Lake Okeechobee sediments which had organic fractions in
the range of 40-45%. The experimental procedure included placing sediments harvested
from Lake Okeechobee in an annular flume as a thick slurry, covering the placed bed with
lake water, allowing the mixture to consolidate for several days, and then measuring the
concentration in the overlying water as various shear stresses were applied to the bed.
He developed a two-component model incorporating a "surface-fluff' component, which
had a critical shear stress for initiation of erosion of zero Pa (immediate erosion upon any
disturbance), with a bed-surface-erosion component that had a critical shear stress of
about 0.45 Pa (much lower than that reported for sewer sediment). Erosion rates as a
function of shear stress were determined for both components. He further correlated the
erosion rates with sediment bulk density and developed a map of regions of bulk density
and shear stress where various types of erosion would occur. The experimental method
utilized may have underestimated the impact of light flocculated organic matter which
may have been intermixed with the base sediment at harvest, nevertheless this study
reports the lowest critical shear stress found for organic sediments in this literature
survey.

37
Hvdrologic approach to seston transport
The application of traditional non-cohesive sediment transport theory to transport
of seston has been shown to be inappropriate for several reasons (Webster et al., 1987).
One major assumption in most non-cohesive sediment transport models is that
transported particles are similar to particles in the streambed (for example, Einstein
1950). This is often not the case in stream transport of seston, where the transported
matter may consist of small and/or low density matter with a high organic content while
the streambed may be primarily sand, rock, and large organic particles. A second major
assumption is that of unlimited supply of transportable solids (for example, Bagnold
1966). This is often not the case with seston transport (for example. Allen 1977) where
depletion of supply is frequently evident, for example over sequential storm events of
similar or increasing magnitude (for example, Svendsen and Kronvang, 1993).
These violations of traditional sediment transport assumptions may be
circumvented by using models derived from cohesive (Fine Grained) sediment transport
theory which allow for variation of bed composition and recognize supply limitation.
The theory of cohesive sediment transport formed the basis for much of the work planned
in this project and is discussed in the accompanying section entitled A Brief Discussion
of Sediment Transport Theory.
An alternative approach to the quantification of seston transport has been to
develop phenomenological relationships between basin hydrologic parameters and seston
transport. Typical are attempts to correlate particulate organic matter loads to stream
discharge or stream order (Hawkes 1975). These approaches have met with limited
success, mainly because they are implicitly based on the noncohesive sediment transport
model assumptions. A more sophisticated approach was proposed by Sedel! et al. (1978)
where stream power, rather than discharge, was used to correlate seston discharge. This
approach reduced data scatter somewhat, but suffered from the limitations that

38
correlations were very stream-type specific, and that for a given stream a hysteretic
pattern was observed over a time series when transported load was correlated with stream
power. This hysteresis was explained by Webster et al. (1987) as being due to the
limitation of supply of seston, which reduces the potential transportable load as time
progresses in an event or series of closely related events. The hysteretic effect may also
arise from seasonal factors which affect the availability and transportability of seston
even in the absence of significant hydraulic excursions. They proposed that the stream
power correlations be modified into separate correlations for the rising and falling limbs
of storm hydrographs, and that inter-event seston production and storage functions be
developed for use in conjunction with the power correlations. This approach results in a
higher level of predictability, but it also requires a fairly intimate knowledge of the
biological productivity of a system in order to be useful in a predictive capacity.
Several recent studies in Denmark on the Brabrand Lake (agricultural) watershed
(Kronvang 1992) and the Gjem A (lowland) basin (Svendsen and Kronvang, 1993, and
Svendsen et al., 1995) have provided good illustrations of the complexity of seston-
related phosphorus transport in the context of seasonal time scale basin hydrology. These
studies showed the following results:
Significant differences in Particulate P/Total P ratios in adjacent
subcatchments
Decreasing P-content of inorganic particulates as flow increased
Relatively constant P-content of organic particulates as flow increased
Reduction in the P-load vs discharge correlation with sequential storm events
Seasonal changes in the P-load vs discharge correlation
High levels of seston removal by macrophyte beds during normal flow but
high levels of release during storm flow, with this effect accentuated when
storms followed extended low flow periods.
Episodic increases in particulate-P load arising from macrophyte cutting

39
The basin as a whole acting as a phosphorus sink during a dry year, but
becoming a net source the following year, which was wetter than average
Mass balances showing an increasing bias toward underestimation of
phosphorus export as frequency of sampling decreased
The authors rightly noted that current understanding of non-point P sources and
the routing of P compounds within watersheds is very poor.
A Brief Discussion of Sediment Transport Theory
Non-Cohesive sediment transport
Traditional (non-cohesive) sediment transport places mobilized sediment in
various categories, which have been noted earlier in this chapter. Grains making up a
substantial part of the movable bed of a stream are called bed material. Bed material
moving within a few grain diameters of the bed is known as bed load. Material in
suspension which is not present in any quantity in the bed (usually very fine sediment) is
known as wash load. Bed material swept up into the main flow stream by turbulence
becomes part of what is referred to as the suspended load. The suspended load is the
combined mass of this fugitive bed material and the wash load. The relative proportions
of the bed material that move as bed load and suspended load depend on the
characteristics of the bed material, such as size and density, and the flow conditions
(Middleton and Southard, 1984).
The analytical approach taken in non-cohesive sediment dynamics treats bed load
and suspended load as two separate entities, subject to two different sets of physical
forces. Typical equations for bed load horizontal flux incorporate the difference between
a critical shear stress for incipient motion and the actual bed shear stress as a driving

40
force and particle size and mass as a resistance. There are several such equations
(Raudkivi 1976). Shields equation is presented here for the purpose of example.
(h)-4?S
3.1
where gB = weight rate of bed transport per unit width, kg/m
q = volume rate of water flow per unit width, m3/m
S = bed slope, m/m
d = particle diameter, m
x = bed shear stress, Newton/m2
to = critical shear stress for motion, Newton/m2
y = specific mass of water, kg/m3
ys = specific mass of sediment, kg/nf
Suspended load transport may be treated within one or more of three conceptual
frameworks, diffusional, energy, or statistical.
Diffusional approaches incorporate a turbulent eddy diffusivity term into the
equations of motion and solve for the steady state concentration profile of mass in the
vertical direction. An equation often proposed as a basis for analysis is one presented by
Rouse (1937), that describes the concentration profile above a datum plane at which point
it is assumed the concentration is known:
3.2
with
w
z =
P*c£/*
where C = concentration at elevation y, kg/m3

41
Ca = concentration at datum level a, kg/m3
y = elevation, m
y0 = water surface elevation, m
a = datum level elevation, m
w = particle settling velocity, m/sec
P = constant of proportionality
k = von Karmans constant
U* = shear velocity, m/sec.
Application of this equation gives sigmoidal concentration curves, the shape of which
depends on the value of z.
Energy approaches analyze the suspended solids from the standpoint of an energy
balance, where the momentum transfer to the suspended solids by the turbulent fluid must
equal the excess weight of the solids in motion. Velikanovs gravitational theory
(Velikanov 1954), for example, uses this approach to arrive at an equation that describes
the concentration profile as follows:
3.3
a
y
where q
yo
qa = reference level, m
k
a =
30y0
ks = grain roughness
Application of this approach gives concentration curves similar to those produced by the
diffusion approach.

42
The statistical approach may be applied in several ways. First the diffusional and
energy approaches, as presented above, apply to only one specific monodisperse class of
particles. One statistical approach is to apply a statistical distribution to particle classes
and then apply the equations to each class. A more detailed approach is to write the
equations of motion as stochastic equations and then either solve them using simplifying
assumptions and asymptotic solution methods (Hinze 1957, Hino 1963) or use them to
run Monte Carlo computer simulations (Yalin 1972). Specifics will not be presented
here.
The total suspended load is obtained by integrating the equation
3.4
where qs = volume rate of suspended solids discharge per unit width,
m3/m-sec
u = velocity at elevation y, m3/sec
The details of the integration of this equation depend entirely on the choices of
concentration distribution, velocity distribution and reference datum plane. There are
many options available (Raudkivi 1976, Vanoni 1975), some of which derive from basic
principals and some of which contain empirically derived fitting parameters. The latter
tend to give better results than the former in specific situations but cannot necessarily be
extrapolated beyond their range of calibration.
There are several drawbacks in the application of non-cohesive sediment transport
theory to the study of organic sediment transport. First most of the useful formulations
assume steady state with respect to sediment suspension and deposition. Development of
transient formulations would require reversion to the equations of motion and mass
transport, where the dynamics of organic sediments have virtually no investigational

43
history. Second, and more important, the traditional non-cohesive approach makes no
assumptions about supply limitation of transportable material, which is a major drawback
in the evaluation and simulation of organic sediment transport.
Cohesive sediment transport
Cohesive sediments are those that exhibit particle-particle interactive forces
strong enough to cause them to not act as a collection of elementary particles. They are
typically composed of fine-grained particles which, under the appropriate circumstances,
assume an electrochemical surface charge that contributes to their interactive nature. The
particle size, chemical charge, and applied shear can operate in close packed
environments to impart rheological properties to the sediments, which affects their
erosion characteristics, and in dispersed environments to cause flocculation and dis
aggregation of suspended particles, which affect their sedimentation characteristics.
Partheniades (1977), using concepts developed and presented in earlier papers,
proposed a unified theory that treated bed load, suspended load and wash load all as
special cases of a general erosion-transport-deposition model. His conceptual framework
includes a minimum critical shear stress for inception of erosion, a maximum critical
shear stress for termination of deposition, addition of inter-particle forces to the force
balance for detachment of a particle from the bed, incorporation of probabilities that a
particle may suspend or deposit which are related to shear levels, and the concept of
supply limitation and carrying capacity saturation for any given class of sediment. One of
his important conclusions is that there are classes of sediment that can behave either as
wash load or as bed load, depending on channel discharge characteristics.
Organic sediments do not generally possess electrochemical charges strong
enough to affect their compacted strength but their complex physical configurations may
impart some degree of particle-particle interaction in the settled phase. There are

44
electrochemical charges present that may be sufficient to affect some degree of
flocculation in the suspended phase. Given their source, there is definitely the potential
for supply limitation, particularly in larger channels and flows. Their specific gravities
may be considerably less than those of the inorganic materials typically found in
sediments so the buoyancy forces acting on them may have a much greater impact than on
inorganic sediments. For these reasons it may be appropriate to approach organic
sediments as hybrids between cohesive and non-cohesive sediments and attempt to adopt
appropriate techniques from the study of each of these two types of sediment.
Several reviews (Kranck 1984a, Mehta 1984, Mehta 1988, Mehta et al., 1981,
Mehta et al., 1989. Parchure and Mehta, 1985) cover the practical aspects of cohesive
sediment transport well. The following summary draws from these reviews.
Erosion Modeling of the instantaneous erosion rates of cohesive sediments can be
approached from the perspective that the critical shear stress for erosion corresponds to
the shear strength of the eroding sediment (Mehta et al., 1981). The key to developing
analytical expressions for cohesive sediment erosion is the determination of the bed shear
strength and the form of the rate of erosion as a function of the shear stress which is
applied to the sediment bed in excess of the bed shear strength.
Beds that are artificially placed in erosion simulation devices for study (placed
beds) usually have constant properties in the vertical direction, which makes their study
relatively simple. The instantaneous erosion rates for these beds may be expressed in the
normalized form
6
T,
where e
Tb
Erosion rate, kg/sec-m2
Erosion rate constant, kg/sec-m2
Shear stress applied to bed, n/m2
3.5

45
Ts = Bed shear strength, n/m2
although formulations have been proposed that make the erosion rate exponential with
respect to the excess shear stress. Note that for a specific constant bed shear strength. Em
and Ts might be combined in Equation 3.5.
Beds that are deposited, under either flow or non-flow conditions, will normally
have properties which vary with vertical position within the sediment. Parchure and
Mehta (1985) have presented an analysis that correlates the depthwise variation of shear
strength of specific sediments with the vertical increase in sediment density. They
proposed an equation of the form
3.6
where Ef = floe erosion rate, kg/sec-m2
a = rate coefficient, m/N5
elevation of sediment surface above datum plane, m
z
The floe erosion rate represents a non-zero erosion rate when the bed shear stress
equals the bed shear strength, a recognition that some erosion occurs at this point because
of random turbulent fluctuations at the bed surface. Given the (known) relationships
between sediment depth, density, and shear strength, and given a known applied shear
stress pattern, Equation 3.6 may, in theory, be used to predict sediment suspension
throughout the course of an erosion event.
Agglomeration and De-agglomeration Particle-particle collisions, caused by
Brownian motion, velocity gradients within the suspending fluid, and differential rates of
settling among particles of various dimensions and densities may give rise to flocculation
or agglomeration of suspended matter, increasing the dimensions of the particle-
collection structure, which generally tends to increase the sedimentation rate of the
particle-collection. Turbulence may increase the opportunity for agglomeration but it also

46
provides shear forces that act on the particle-collection, tending to break down the
structure. The floe size distribution depends in a complex way on floe strength and
turbulence structure but a qualitative interpretation is that low levels of turbulence tend to
promote flocculation and produce light, diffuse floe structures, while high levels of
turbulence tend to promote de-agglomeration and produce dense, compact floe structures.
Sedimentation Sedimentation of cohesive particles falls in one of three general
categories. At concentrations of less than about 300 mg/1 particles act as individuals, are
not influenced by their neighbors, and maintain a constant settling velocity' independent
of particle concentration (Krone 1962). This region is known as the free-settling zone.
At concentrations above the free settling zone, particle-particle interaction and differential
settling cause the settling velocity to increase with increasing concentration as there
become more and more opportunities for agglomeration as concentration increases. This
region is known as the flocculation settling zone. Beyond a certain concentration,
typically 5000-10,000 mg/1 (Mehta et al., 1989) the flocculant structure of the sediment
becomes so extensive that bridging begins to occur and the floe structure becomes
partially self-supporting. At this point and beyond settling velocity tends to decrease with
increasing concentration. This region is known as the hindered settling zone.
Deposition The concept of deposition is treated differently in cohesive sediment
dynamics than in non-cohesive sediment dynamics. For a class of potentially cohesive
particles there may exist a critical bed shear stress beyond which no deposition will take
place at any reasonable concentrations. At bed shear stresses less than critical a fraction
of the suspended material may deposit. The fraction of deposition is related to the
departure of bed shear stress below critical. The simplest expression of this theory for a
monodisperse particle collection is the Krone formulation (Krone 1962)
for Tb < Td and Fa = 0 for tb > td
3.7

47
n2
H = bed shear stress, N/m2
id = critical depositional shear stress. N/m2
ws = settling velocity of sediment, m/sec.
Cb = concentration of suspended sediment, kg/m3
This expression becomes more complex as additional classes of particles are added and as
the effects of agglomeration and de-agglomeration are considered. One practical aspect
of these relationships is that erosion and deposition of cohesive sediments may be studied
independently of one another if the critical shear stress for deposition is less than the
critical shear stress for erosion.
Consolidation Here there are similarities between the behavior of cohesive
sediments and the presumed behavior of organic sediments. Consolidation takes place
because the forces exerted by the self weight of the settled material exceeds the strength
of the interlocking sediment structure and its contained water. In cohesive sediments the
interlocking sediment structure contains agglomerated structures as well as individual
particles. The strength of these agglomerated structures is often the controlling element
in the rate of consolidation. Organic sediments, with their potentially complex physical
structures, might be expected to exhibit characteristics similar to cohesive sediments in
the consolidation process, such as a rapid approach to an asymptotic density profile as
supporting structural networks are broken down during consolidation.

48
Particle Entrainment Simulator
Quantitative prediction of the transport of cohesive sediments requires the
estimation of net sediment flux at the sediment-water interface as a function of sediment
properties and the applied shear stress. The same holds true for sediments that behave
cohesively due to physical particle-particle interaction.
Numerous studies have been conducted in laboratory flumes to correlate erosion
rates with sediment properties and shear stress. A condensed listing includes
Parthenaides (1965), Mehta and Parthenaides (1975), Fukuda and Lick (1980), Lee et al.
(1981), Lick (1982), Mehta et al. (1982). Parchure and Mehta (1985), Ockenden (1993),
and Verbeek et al. (1993)
Difficulty often arises because the state of sediments in the laboratory
environment differs significantly from that in the field. In addition, direct measurement
of spatial variation of erosion characteristics in the field cannot be managed in any
efficient way because the laboratory apparatus are large and cumbersome and require
relatively large quantities of sediment for a single test. Attempts have been made to
circumvent this limitation by developing, for example, in-situ flumes (Young 1977), but
such devices suffer from their own logistical and reliability problems (Lavelle and Davis,
1987).
An alternative to in-situ flumes has been the development of a small portable
device that applies an oscillating-periodic but zero-mean velocity to the surface of a
sample of bottom sediment. Appropriate sampling techniques can allow the sample to
approximate an undisturbed sample of original sediment. A device with this capability
was first used by Rouse (1938) and later adapted to the study of fine lacustrine sediment
by Tsai and Lick (1986).

49
The device is of simple construction, consisting of a cylindrical chamber inside of
which a horizontal grid oscillates vertically. The sediment to be studied is placed in the
bottom of the cylinder and is overlain with water, preferably of composition similar to
that of the original sediment environment. The grid oscillates in the water and creates
turbulence that penetrates to the sediment-water interface, causing sediment resuspension.
The turbulence, and thus the extent of resuspension, is a function of the grid oscillation
frequency (Tsai and Lick, 1986)
The structure of the turbulence generated in the PES is different from that
generated in a typical field shear flow. The mean velocity vector in the PES changes
direction many times a second, whereas the mean shear velocity vector in field channel or
tidal flows is relatively steady over short time periods. Never the less the statistical
properties of the turbulence generated in the PES can be assumed to be similar to those in
field shear flow (Lavelle and Davis, 1987). The statistical similarity assumption allows
the PES to be calibrated to standard parameters.
The calibration procedure, developed by Tsai and Lick (1986), involves
determining the erosion extent of a specific sediment in the PES for various oscillation
frequencies. The erosion extent of an identical sediment, prepared under similar
conditions, is also determined in a shallow flume, for which the shear stress correlations
are known with some degree of accuracy. The flume is also operated over a range of
shear stresses. Under the assumption of statistically similar turbulence yielding similar
erosion extent, it is possible to match a specific erosion extent at a known oscillation
frequency in the PES with a corresponding equivalent erosion extent in the flume at a
known shear stress. Ultimately this empirical matching procedure yields a correlation of
PES oscillation frequency with applied shear stress in a uniform flow field. This is an
important translation, because all useful erosion-hydrodynamic models require an
expression of shear stress or equivalent shear velocity at the sediment-water interface for
calculation of erosion rates.

50
One caveat must be noted. Recent work (Chapter 6) has shown that the PES
calibration procedure can be sensitive to large variations in sediment type. That is, PES-
flume calibrations developed for sedimented kaolinite-type clay beds, compacted
bentonite-type clay beds, and sedimented organic-type beds were internally type-
consistent but did not compare well inter-type on a shear stress basis. This indicates the
advisability of using calibration data for the PES from a sediment of similar type.
The portability of the PES has allowed its use on shipboard (Tsai and Lick, 1986,
Lavelle and Davis, 1987, Ziegler et al., 1987, Sffisco et al., 1991), as well as in the lab.
Studies utilizing the PES have been conducted in freshwater (Tsai and Lick, 1986,
MacIntyre et al., 1990, Davis and Abdelrhman. 1992, Mehta et al., 1994) as well as in the
marine environment (Tsai and Lick, 1987, Lavelle and Davis, 1987, Ziegler et al., 1987,
Sffisco et al., 1991, Davis and Abdelrhman, 1992). Specialized studies utilizing the PES
have included investigation of the effect of bioturbation on sediment erosion
characteristics (Davis and Means, 1989, Davis 1993) and studies of the impact of
catastrophic events on short term sediment transport (Mehta et al., 1994).

CHAPTER 4
INITIAL HYPOTHESES, OBJECTIVES, AND RESEARCH PLANS
Problem Overview
At this point a reiteration of the primary issues and objectives of the overall
program is useful to place the objectives of the research reported in this document in
proper context.
Summary of sources
The overall objective of the program funded by the EAA EPD is to develop
methods of irrigation and drainage control for the farms of the EAA that minimize the
quantities of phosphorus exported off-farm while not causing material detrimental effects
to the crops grown on-farm. The primary on-farm sources of soluble phosphorus (SP)
and particulate phosphorus (PP) are considered to be soil mineralization (SP), fertilizer
application (SP and PP), and soil/litter/sediment mobilization (PP), where the term
sediment is taken in the most general sense to designate any water-resident particulate
material.
Organic soil subsidence can result from microbially mediated oxidation of the
organic matter in the soil. The oxidation destroys soil structure by converting soil organic
mass to soluble organic and inorganic matter and CO2. The process also causes
conversion of the nutrients incorporated in the organic matter to their inorganic form, for
example conversion of organic phosphorus to soluble orthophosphorus. Soil subsidence
may be retarded by reducing the oxidation-reduction potential of the soil to anoxic or
51

52
anaerobic conditions. Such retardation may be partially affected by control of soil water
table levels.
The objective of fertilization is to provide growth factors (nutrients) to the crops
when there is insufficient or growth limiting supply from the natural environment.
Fertilization contributes directly to phosphorus export to the extent that the fertilizer
application is inefficient, that is, to the extent that applied fertilizer is not taken up by the
target crop. Fertilization efficiency may be improved by correct placement and timing of
fertilizer application. The most efficient placement and timing would maximize the
fraction of applied nutrient that reaches the root zone and optimize the root zone nutrient
concentration relative to plant uptake requirements.
Soil/litter/sediment mobilization is a complex function of biological and
hydrodynamic conditions that relates not only to the physical character of the
transportable material but also to the mode in which water is applied to and removed
from the fields and water conveyance systems of the farm. In the soil/litter/sediment-
water system it is necessary, in addition, to consider the inter-conversion of soluble and
particulate forms of phosphorus arising from physical adsorption/desorption and
biological assimilation and decomposition.
The movement of water within the EAA is dominated by pump operation and
characterized by much more rapid hydrodynamic transients than would be the case for a
less hydraulically developed area of similar terrain, such as the Everglades National Park.
In addition, the District Canal water levels are typically at a greater elevation than those
found in the farm water conveyance systems so there is usually a hydraulic gradient
tending to drive water back onto the farm by groundwater flow or structure leakage.
Other factors that may have an effect on phosphorus transport are the local water
chemistry (high calcium content, high photosynthesis potential), phosphorus content of
precipitation, and miscellaneous sources arising from wind erosion, vehicle movement,
etc.

53
Original conceptual model
Within the context of the modeling portion of this project, fertilizer contribution
to phosphorus export is being handled as a subset of soluble phosphorus export in the
field groundwater, so for the sake of conciseness fertilization will be excluded from
further discussion. With this caveat, the remaining factors, as hypothesized at the
initiation of this program, are presented in conceptual form in Figure 4.1.
Figure 4.1: Farm Scale Phosphorus Transport Original Conceptual Model

54
The figure shows a qualitative model with the following main features:
Soil water supplies or removes soluble phosphorus (SP) to or from the
canal water depending on whether irrigation or drainage is being practiced.
Soil water, moving through (perpendicular to) the canal wall, supplies or
removes particulate phosphorus (PP) to or from the canal water depending
on drainage (perpendicular flow erosion) or irrigation (perpendicular flow
filtration). PP is also supplied to the canal water from the bed by parallel
flow erosion arising from longitudinal flow in the channel, and removed
from the canal water by sedimentation.
Precipitation supplies SP and PP directly to the canal water, and causes
surface runoff under some circumstances which also supplies SP and PP.
Biological activity (shown as, but not limited to, phytoplankton and
P\
macrophytes) immobilizes SP, and increases or decreases pH, depending
on photosynthetic vs. respiration activity. Particulate detrital biological
material is deposited on the top layer of sediment, which is assumed to be
oxic.
Inorganic particulate matter suspended in the ditch and canal water also
interchanges with the oxic sediments via erosion and sedimentation.
Oxic sediments interchange with anoxic sediments, and both
compartments undergo mineralization of organic PP to SP. Anoxic
sediments may interchange with suspended particulates under the
appropriate hydrodynamic conditions. Anoxic sediments also interchange
with the high-calcium-containing substratum layer.
Soil water may infiltrate through the substratum, losing SP and gaining
Ca that is earned into the bulk phase canal water for precipitation
reaction with SP under appropriately high pH (from photosynthesis). All

55
or part of this precipitate may dissolve under reduced pH conditions (from
respiration).
Canal water may also interact directly with exposed substratum in the bed
to adsorb or desorb phosphorus, depending on surface and bulk phase
conditions.
Hydrodynamic conditions affect interchange of sediments with the bulk
phase, interchange of soil water with the bulk phase, erosion and filtration,
infiltration, redox potential in the sediments, downstream transport of all
mobile constituents and, indirectly, soil/soil-water/crop interchanges.
The model is an unsteady-state one so all interactions must be assumed to
be dependent on antecedent conditions.
Scope of This Research
Minimum criteria for experimental and modeling efforts
Given the fact that virtually no channel transport experimental or modeling
activity had taken place on the EAA farm scale at the time of inception of this research,
the decision was made to restrict this effort to fundamental approaches. The field-scale
groundwater chemodynamic model for soluble phosphorus was being developed
separately and was planned to be integrated with the farm-scale model at a later date.
Basically the field scale model is to provide the soluble phosphorus dynamic boundary
conditions at the periphery of all water conveyance channels. The model to be developed
for the conveyance systems was required to have at least the following properties.
1. It should adequately represent the surface water, groundwater, and channel
flows of a network typical of the EAA farm scale over the time scale of an
entire pumping event, that may cover multiple days.

56
2. It should be capable of sufficiently detailed time scales to allow simulation
with reasonable accuracy of transients that exist at pump start-up and shut
down, however modeling of rapid transients, such as hydraulic jumps, is
not necessary because of the level terrain of the EAA.
3. It should model with reasonable accuracy particulate phosphorus
mobilization, transport, and deposition as a function of some readily
defined hydraulic parameter such as shear stress or average stream
velocity.
¡4. Phosphorus interchange between soluble and particulate forms by
adsorption-desorption should be represented in the model.
5. The dynamic impact of biological growth and senescence in the aqueous
system on phosphorus transport should be incorporated into the model.
Application of minimum criteria and resulting conceptual model simplification
Water chemistry reactions resulting from interflow through the substratum were
deferred for later consideration. Interactions between oxic and anoxic sections of the
base sediment can be very sensitive to redox potential profiles and tend to be important
over long term (multi-annum) time scales. The field portion of this study was intended to
last through only one wet season. It was decided to defer this portion of the study to a
later program.
When these minimum criteria were applied and the program was evaluated in
light of prior knowledge, available resources, and time constraints, it became necessary to
affect a truncation of the conceptual model to a simpler form.
Particulate filtration and resuspension arising from groundwater flow
perpendicular to ditch and canal banks may be a meaningful contributor to sediment flux
but experimental determination of field-relevant parameters could require considerable

57
effort that, from a pragmatic standpoint, should not be expended until the nature of
erosion/sedimentation associated with shear flow is reasonably well understood. It was
decided the best allocation of resources for this project would be to attempt to understand
the nature of shear flow erosion and defer perpendicular flow erosion studies if and until
it became evident that they were necessary.
Current restrictions by the South Florida Water Management District limit the
growers to a nominal farm pump discharge of one inch of farm runoff per day, therefore
the time scale of pumping events in the EAA may usually be on the order of one to three
days after a significant rainfall event. Typical time constants for biological assimilation
and degradation processes may be on the order of weeks (USEPA 1985). For the first
generation models to be developed in this study it was deemed appropriate to ignore
biological process transients occurring during the pumping events, that is, to consider that
all biological processes of significance occur during the inter-event quiescent periods.
Given the accuracy with which biological processes in natural systems can be portrayed,
this is a reasonable first approximation.
The conceptual model resulting from these simplifications, illustrated in Figure
4.2, is as follows
\J Soluble phosphorus is supplied from the fields as a boundary condition in
surface and ground water runoff to the farm drainage ditches.
Soluble phosphorus may also be added to the system with the release of
sediment interstitial pore water when sediment is resuspended.
Particulate phosphorus may be supplied to the system as:
eroded soil or litter carried by overland flow of surface runoff,
* biological matter already suspended at event initiation,
sediment that is resuspended as a result of turbulent shear
stresses arising from channel flow.
Particulate phosphorus may be removed from the system by sedimentation.

58
Phosphorus may be exchanged between the soluble and particulate phases
by adsorption/desorption
All biological processes are considered to be active only during the
quiescent inter-event periods
Research goals for this study
The simplified conceptual model now leads directly to the general research goals
for this study that are, for a specific study farm, as follows:
1. Adapt an existing hydraulic network model for use in the sub
irrigation/drainage mode that prevails at the specific site.
2. Adapt or develop a water quality sub-module for the hydraulic model that
will allow expression of the water quality processes in a dynamic
hydraulic regime.
3. Determine the nature, frequency, and quantitative contribution of overland
flow erosion to particulate phosphorus in the water conveyance system.
4. Develop a description of the conveyance system sediment with respect to
quantities, locations, chemical/physical characteristics,
erosion/sedimentation characteristics, and temporal variations that is
adequate to allow modeling of sediment transport and pore water release
within the system.
5. Determine the adsorption/desorption characteristics of the system
sediment and incorporate this process into the model where appropriate.
6. Determine the biological processes in the system that contribute
significantly to phosphorus transport and develop a first approximation
lumped parameter model of these processes.
7. Develop, calibrate, and verify the model on the target study site.

Soluble Phosphorus (SP)
SP and F^roculac
SP and FP fian surface
on soil water
Riccphoms (FP) fian ditch
runoff
as direct preapttaiicri. wind
wall sal. ditch sediment.
(Pen ode Excursions!
T
and decaying plant marta in
ditch
1
etc
(Secaidarv Contn buttons)
T
T
T
tisal type, water table,
If ditch dtmaistans aid
flsal type, crop cover.
degree of rmneializanai.
water level, development
culdvaiai practices. rainfall
redox potential, antecedent
and maintaiance history.
intensity, wata table
canil tier is and usage p.
conditions....)
sal type, rainfall intensity .
location, damage svstem
content in preapitation. sal
hydrodynames, sedunent
histay and transport
characteristics.)
configurations and
managematt practices, field
heterogenaty...)
moisture..)
QamisDitdiWM Drainage DitdiWM
Soluble Ftaphems Mnasluancnaxl panedate Phorrhoms
(SB g*11' m
L
Low- Riasphaus Dlution Waa fian
Ctamage Dtdi Wbter
damage Dtch Sedment
(see above)
Carbonate Rock Interflow
T
Farm Canal
V
Sedment Rae Wfiter
Soluble Rtasphcrus
FarmCaial Water
Efecaying
T
A
Mater
Soluble Riosphonis
1
Macrophytes
ahen ancr Adsorption
rpdon
Farm Canal
Will
Particulate RxsphctiLS

Rtytoplankton
T^fcuspenstan
Sedmaifa(gi
Farm Canal
Sediment
I
Repeal Farm Canal Processes
CtainageDstnct
Canal Wter
Rqoeat Fam Canal Ftaxesses
i
Repeal Pam Carat Recesses Mnagonncismci
Canal Waa
1
ExilEAA
Rqxat Fam Canal Processes
Figure 4.2: Farm Scale Phosphorus Transport Simplified Conceptual Model

60
General Research Plan
The preceding provides the framework for discussion of the original genera]
research plan. It is appropriate to reiterate that at the time of development of this plan
there was no information in the open literature relating specifically to transport of high
organic content materials such as the soils and sediments of the EAA. Given that lack of
information it was deemed appropriate to start with a basic evaluation of the target
systems and then develop a fundamental data gathering program. Following is an outline
of this approach.
1. Conduct a general survey of several representative farm conveyance
systems in the EAA, restricted initially to main farm canals, to determine:
a. General canal dimensions and pumping capacities,
b. Volume and mass of sediment contained in canals,
c. Selected physical and chemical characteristics of representative
sediments.
2. Select a target farm for detailed study.
3. Obtain sufficient hydrographic and topographic data to allow the farm
drainage network to be modeled.
4. Select a representative source of sediment from the target farm and
conduct physical and chemical characterization.
In the laboratory, determine erosion and sedimentation characteristics of
the representative sediment and use this as a prototype system for
modeling purposes.
6. Develop or adapt some small-scale device that would allow estimation of
erosion characteristics in the field and use this device to study geographic
and temporal variations of sediment erodability.
J-

61
7. Set up regular discharge monitoring of particulate phosphorus at the target
farm.
8. Plan and execute synoptic studies to:
a. determine variation of sediment characteristics as a function of
location within the conveyance system and
b. evaluate suspended solids concentration variations within the
conveyance system during transport events.
9. Determine, in the field, the requirements for biological data depending on
the outcome of the preceding characterizations and monitoring programs.
10. Adapt the chosen computer program, develop the water quality program
for phosphorus, and calibrate and verify for the target farm.
The subsequent four chapters discuss the specific activities directed toward
sediment characterization and transport property determination.

CHAPTER 5
SEDIMENT SURVEY AND PHYSICAL-CHEMICAL CHARACTERIZATIONS OF
SELECTED SEDIMENTS
Sediment Survey
The first field activity of the program was to develop an order of magnitude
estimate of quantities of sediment present in representative farm canals within the EAA.
Typical field ditch construction technique in the EAA involves digging or
dredging at least down to the limerock/marl substratum underlying the organic soils. In
the case of main canals, the limerock is usually excavated to a depth of several feet below
the marl surface to improve hydraulic flow conditions. Ensuing physical and biological
processes combine to contribute a sediment layer that builds up in the conveyance system
over time. Periodically the system components must be dredged to remove the sediment
build-up. Dredging is a relatively costly process, especially for the small grower who
must contract out the work, thus the extent of ditch and canal maintenance via dredging is
variable among growers.
The excavation or dredging technique, which initially removes all soil from the
channel, does simplify the measurement of sediment inventory, because all transportable
particulates present in the channels may be assumed to be sediment.
Selection of representative farms
The research program Implementation and Verification of Best Management
Practices for Reducing Phosphorus Loading in the EAA, conducted by the University of
Floridas Institute of Food and Agricultural Sciences (IFAS), had a total of ten farms
62

63
within the EAA under intensive study. These farms had been chosen to represent a
typical range of farm size, crop type, soil type, and geographical distribution within the
four main sub-basins of the EAA.
The farms have, by mutual agreement with the appropriate regulatory authorities,
been coded to preserve the anonymity of the growers who were participating in the study
program on a voluntary basis. The code numbers were in a series of UF9200 through
UF9209, where UF indicates University of Florida study, 92 indicates the start year
of the BMP study program and 00 through 09 indicates the code number assigned to
each participating farm. Further coding indicated the location of permanent sampling
sites at a specific location. A refers to the main farm discharge; some farms have two
major discharge locations so B is reserved for the second discharge location where
appropriate. Sites C and beyond refer to internal farm sampling locations that are
added as the program dictates. For example farm UF9206 has two pump discharge
locations, so UF9206B refers to the permanent sampling location at the south pump
station of that farm. That coding protocol will be modified later in this document where
UF9206A and B will be referred to as UF9206N and S respectively.
Three sites were selected for the preliminary sediment inventory based on the
recommendations of Dr. F. T. Izuno, the IFAS Program Principal Investigator. They
were:
UF9200 A 1280 acre (518 ha) farm with two main canals and one pump station,
that planted only sugarcane. This farm was representative of medium to large
sugarcane-only fields with relatively deep soil that are actively managed.
UF9202 A 320 acre (130 ha) farm with one main canal and one pump station
that planted only sugarcane. This farm was representative of small to medium
sugarcane-only fields with relatively shallow soil that receive low to moderate
management.

64
UF9206 A 1750 acre (710 ha) farm with a complex drainage pattern that
included three main canals, several perimeter canals, two pump stations, and gated
interconnects among the various sections of the farm. This was a multi-crop farm
that maintained stands of sod and sugar cane and rotated plantings of rice and
winter vegetables at various locations. This farm was representative of medium to
large farms with relatively deep soil that maintain multi-crop patterns and are
actively managed.
Sediment survey measurement methods
Water depth and sediment depth transects were takeitby boat from cordoned canal
sections. At three foot (0.91 m) increments, the distance from the water surface to the
sediment surface was determined by lowering a neutrally buoyant one ft2 (0.09 m2) pad
attached to a calibrated rod until it rested on the sediment surface. At the same location
the distance from the water surface to the canal bottom was determined by driving a 0.5
in. (1.3 cm) diameter calibrated steel penetrometer rod through the sediment until it met
resistance at the marl surface. Sediment depth at each location was determined by
difference between water-surface-to-sediment-surface depth and water-surface-to-marl-
surface depth.
At each site water and sediment depth transects were taken at multiple locations
depending on the site main canal configurations. Four transects were taken at UF9202,
eight at UF9200, and twelve at UF9206. Two sediment core samples were taken at each
site using the piston core sampler described in Appendix A. The cores were expunged
on-site into plastic bags, sealed, refrigerated, and transported to the Gainesville, FL,
laboratory for bulk analysis. Analyses were conducted according to the methods
described in Appendices B and C.

65
Sediment survey results
The physical results of the sediment survey are shown in Table 5.1, the analytical
results appear in Table 5.2.
Table 5.1: Sediment Survey Physical Results
Location
Total Main
Canal Length
(m)
Farm Area
(ha)
Average
Sediment Depth
(m)
Estimated Total
Sediment
Volume
(m1)
Estimated Unit
Sediment
Volume
(m3/ha)
UF9200
6437
518
0.79
39277
75.8
UF9202
1609
130
0.65
5486
42.2
UF9206
11715
710
0.54
40368
56.85
Table 5.2: Sediment Survey Analytical Results
Location
Core
Length
(cm)
% Dry
Solids
% Ash
(Dry)
% Volatile
Solids
(Dry)
Pore
Water pH
Wet Bulk
Density
(gm/ml)
Solids
Specific
Gravity
Pore
Water SP
(mg/1)
Solids TP
Content
mg/kg
UF9200
#1
60.7
14.77
52.5
47.5
7.21
1.088
1.596
0.35
787
UF9200
#2
52.8
18.09
46.1
53.9
7.16
1.082
1.453
0.20
868
UF9202
#1
55.1
18.61
53.4
46.6
7.12
1.094
1.505
0.03
613
UF9202
#2-Top
41.7
18.44
43.1
56.9
7.16
1.058
1.315
0.56
572
UF9202
#2-Bott.
33.0
45.99
77.7
22.3
7.06
1.378
1.822
0.02
206
UF9206
#1
37.1
25.16
47.9
52.1
7.02
1.114
1.453
0.07
415
UF9206
#2
59.4
22.40
42.7
57.3
7.05
1.100
1.446
0.17
445

66
At locations UF9200 and UF9206 samples #1 and #2 were taken at mid-length of
each of the two major farm canals. At UF 9202, which had only one main canal, samples
#1 and #2 were taken at 25% and 75% of the upstream length, respectively. All samples
except UF9202 #2 were homogeneous in appearance and were treated, for this order-of-
magnitude exercise, as a single bulk sample. UF9202 #2 exhibited two definite layers,
expressed as a sudden color change from black to medium gray at the 41.7 cm depth
level, so it was split into two subsamples, denoted as top and bottom" in Table 5.2.
Sediment survey discussion
The sediment average depths at the three target farms ranged from 0.54 m (1.75 ft)
to 0.79 m (2.6 ft). The estimated total sediment volumes varied greatly as would be
expected because of the farm-to-farm variation in total canal length. An attempt at
normalization is shown in the last column of Table 5.1, where the canal sediment volume
is expressed on a per-unit -farm-area basis. On this basis the unit sediment volume
ranged from 42.2 to 75.8 m3 canal sediment/hectare farm area.
The analytical data presented in Table 5.2 (excluding UF9202 #2 Bottom from the
analysis) were used to draw some preliminary inferences. The dry solids content ranged
from approximately 15 to 25%, indicating relatively well compacted sediments. Volatile
contents ranging from about 47% to 57% and solids specific gravities of approximately
1.3-1.6 indicate a high organic content in the sediments. The volatiles content is
somewhat lower than the typical 70-80% volatile content found in the organic soils of the
EAA, and may represent selective transport or diagenesis or both. The specific gravities
of 1.3-1.6 are of interest because they are substantially lower than the typical 2.1-2.2
values of inorganic sediments, and reinforce the position that direct application of
traditional sediment transport concepts may be inappropriate.

67
The pore water pH values ranged from 7.02 to 7.21, only slightly on the alkaline
side, indicating an environment not particularly conducive to precipitation of calcium or
magnesium phosphorus complexes. The pore water soluble phosphorus content showed a
significant range, from a low of 0.03 mg/1 to a high of 0.56 mg/1, with an average of 0.23
mg/1. In the absence of adsorption equilibrium data a full interpretation could not be
placed on the significance of these values at the time of acquisition, but the average value
of 0.23 mg/1 was not out of line with aqueous phase concentrations reported elsewhere
and discussed in Chapters 1 and 2 of this document. The approximate concurrence of the
pore water phosphorus concentrations and typical aqueous phase phosphorus
concentrations gave at least a preliminary indication that direct pore water contribution to
bulk aqueous phosphorus was of secondary importance. A quantitative estimate of pore
water contributions given later in this chapter substantiates this inference.
The sediment-column average phosphorus content for the three farms shows a
reasonable level of agreement between same-farm samples, implying intra-farm
geographical homogeneity (coefficients of variation of 5-7%) with a higher level of inter
farm variation (coefficient of variation of 29%). The average sediment phosphorus
content over the three farms was 617 mg/kg. It is interesting to note that this sediment
value falls well within the range of 289-834 mg/kg and close to the average of 520 mg/kg
for the phosphorus content of the top 10 cm of Pahokee Muck soil from eleven locations
within the EAA reported by Fiskell and Nicholson (1986). It is tempting to infer the
implication that the soil surface was the source of the sediment but, as will become
evident in later sections, this inference may lead to erroneous conclusions.
The accumulated data may be used to calculate an estimated total mass of
phosphorus contained in the main canal sediments of each farm. These estimates are
shown in Table 5.3, where it is seen that the Unit Phosphorus mass estimates range from
5.0 to 11.2 kg of canal sediment total phosphorus per hectare of farm area. It is now
instructive to calculate the supply potential of these masses of phosphorus.

68
Table 5.3: Estimates of Total Phosphorus Mass in Target Main Canal Sediments
Location
Dry Solids Mass
(kg)
Total Phosphorus Mass
(kg)
Unit Phosphorus Mass
(kg/ha)
UF9200
7.00x10
579.7
11.2
UF9202
1.09x10
64.7
5.0
UF9206
10.63x10
457.0
6.4
In Chapters 1 and 2 various values of off-farm phosphorus export were reported.
A reasonably typical set of values from the data of Izuno et al. (1991) and CH2M Hill
(1978) would be average total discharge phosphorus concentration of 0.2 mg/1, with 35%
or 0.07 mg/1 present as particulate phosphorus. A conservative estimate of average
annual off-farm pumping would be 50.8 cm (20 inches) of rainfall equivalent. On a per-
hectare basis 50.8 cm would equal 5080 m3/yr. A particulate concentration of 7x1 O'4
kg/m3 (0.07 mg/1) would yield an annual particulate phosphorus export of 0.355 kg/ha.
Comparing the estimated typical particulate phosphorus annual export rate with
the inventory range of 5.0-11.2 kg/ha it is evident that the main canal sediments have a
particulate phosphorus yield potential equivalent to something on the order of 15-30 years
worth of phosphorus supply. There is no intent in this exercise to imply that all the
sediment particulate phosphorus is readily available or fully mobilizable, however this
order-of-magnitude estimate does serve to emphasize the potential importance of the farm
canal sediments as sources of phosphorus, both short term and long term. It should be
reiterated here that this estimate was restricted to the main farm canals only and did not
include sediment stored in field ditches (lateral channels that run perpendicular to and
intersect with the main farm canals). A more detailed study of the sediment in the field
ditches of the primary target farm is presented in Chapter 6.

69
Selection of Primary Target Farm
Availability of resources at the time of initiation of the field portion of this
research dictated that the intensive field activity be restricted to one target farm, with the
understanding that the results from the target farm research would serve as a basis for
determining further research work needed to generalize the model. The selection criteria
were limited to a few important points, which were:
1. The target farm should be one of the ten farms participating in the IF AS
Best Management Practices Program.
2. It should have a relatively straightforward drainage layout and hydraulic
management program.
3. It should, as the prototype farm, be restricted to a sugarcane-only planting
on typical soil type and depth. At the time sugarcane represented
approximately 70% of the total agricultural acreage in the EAA.
4. It should be a medium-to-large size farm that practiced at least average and
preferably aggressive crop and water management policies.
5. The grower should be willing and able to provide access, assistance, and
information on a regular basis.
Farm UF9200 met these criteria and was chosen as the target farm for more
intensive study.
Particle-Size Property Distribution Study
The results of the sediment survey represented on-farm particulate transport
potential. The data set from the SFWMD (Mucinic 1994), discussed in Chapter 2,
represented actual transported material sampled at EAA discharge points. A comparison
of representative particulate phosphorus contents showed the on-farm sediment to
average 617 mg P/kg solids, while the selected data set from the SFWMD implied that

70
particulate matter at the EAA discharge points might contain on the order of 2200 mg
P/kg suspended solids. This three-to-four fold difference indicates the possibility of
selective transport or enrichment processes being of some significance in the EAA water
conveyance systems. One approach to evaluating the possibility of selective transport
was to fractionate material from various sources by particle size and determine if there
were appreciable differences in key properties of the several fractions.
Sources for particle size fractionation study
It was desired to evaluate farm soil, farm conveyance system sediment, and Water
Management District Canal sediment. Sampling was conducted in mid-July,
approximately 1.5 months into the wet season. The farm sediment was obtained from the
midpoint of a field ditch in farm UF9200. The ditch chosen was one approximately
halfway upstream and on the south side of the south canal of UF9200. It was well
maintained and free of emergent growth. Samples of the ditch surficial sediment were
obtained at ditch mid-length by Eckman dredge technique. Simultaneously composite
samples were taken of the field soil at locations immediately adjacent to the sediment
sample site and 5-10 meters in-field. The UF9200 grower identifies his fields by
increasing number from east to west and by letter from north to south (See Appendix G).
Field ditches are identified by the two adjacent fields. Field ditch B9B10 was the sample
site and was the ditch between fields 9 and 10 on the south side of the central (South)
canal. This sample will be referred to hereafter as B9B10, the soil sample will be
referred to as Soil. Farm UF9200 discharges into the West Palm Beach Canal. The
District Canal sample was taken from this canal approximately 1 km downstream of the
UF9200 discharge point at a location intermediate between two downstream farm pump
stations. This sample was also of surficial sediment, taken by Eckman dredge technique
and is referred to as WPBC. Approximately 20 liters of each type material were

71
collected, stored in sealed plastic buckets, and transported to the Gainesville laboratories
for fractionation
Soil and sediment particle size fractionation
Several fractionation methods were evaluated and tested on surrogate sediments
prior to sampling the target sediments. The methods included screening, hydraulic
classification, and differential sedimentation. It was hoped to develop a method that
would fractionate on the basis of sedimentation velocity, which would relate to a key
physical parameter of field interest. Unfortunately all the hydraulic classification and
differential sedimentation techniques evaluated either suffered from a lack of available
equipment that could process 20 liters of material in any reasonable time span or
subjected the test material to contact with large volumes of water. The latter condition
was deemed to be inadvisable because of the potential for alteration of the physical or
chemical nature of the materials over repeated dilutions with water. Ultimately the
fractionation method of choice reduced to wet and dry screening. The three samples were
fractionated in their entirety on US Standard screens in the sequence shown in Table 5.4.
Note that + means retained on, while means passed through.
Soil Fractionation The Soil sample was first subjected to coarse screening to
remove clumps greater than 1 cm. The remaining material was dried at 70 C for 72
hours and then subsampled to produce a representative composite. The composite was
split into two sub-segments of roughly 315 gm each. Each sub-segment was subjected to
screening in a stack of USS sieves on a Ro-Tap shaker table for 30 minutes, then
removed, weighed, and recombined. In order to place the evaluation of the soil particle
fractions on the same basis as the sediment it was deemed appropriate to hydrate the soil
to simulate the process of soil eroding into a conveyance system and becoming subjected
to an aqueous environment.

72
Table 5.4: Soil and Sediment Screen Fractionation Sequence
Screen Fraction
Particle Size Range -micrometers
Coarse
>10,000
+8 USS
10.000-2360
-8 USS+16 USS
1180-2360
-16 USS+30 USS
600-1180
-30 USS+50 USS
300-600
-50 USS+100 USS
150-300
-100 USS+200 USS
75-150
-200 USS+400 USS
38-75
-400 USS
<38
A small sub-sample of the recombined sample was taken for moisture analysis
and the rest of the sample was gently combined with 7 liters of filtered native EAA water
that had been collected at the farm side of the UF9200 pumphouse. This mixture was
gently agitated on a twice-daily basis for eight days at room temperature and then
subjected to wet screen analysis (See next section).
Sediment Fractionation The sediment wet Screening fractionation process was
much more labor intensive than the soil dry ^creening. When the sieves were stacked the
cohesive, almost gelatinous, nature of the sediment samples caused the sieving process to
stop almost immediately after the suspension completely wet the screens because of lack
of back-flow of air from the lower sieves to the higher sieves. In order to overcome this
problem the sieving process was done manually, one screen fraction at a time, starting
with the coarse 1 cm screen. Wash for the wet screening process was obtained by
recycling decanted supemate from settled suspension that had already passed through the

73
screen. This way contact of sediment with fresh water was eliminated. Each screen size
fractionation was judged to be complete when visual examination of the screen subnate
indicated that relatively clear liquid was being passed.
After completion of each screening run the screen retntate was scraped into a
plastic bag and the screen was backwashed with a small amount of clear subnate. The
backwash was added to the retntate bag and the bag was refrigerated until analysis.
Screen subnate, that contained all particle sizes less than the current screen run, was
stored under refrigerated conditions between screen runs. The final cut, the <38
micrometer material, was separated from the screen subnate water by centrifugation and
stored in plastic bags until analysis.
The final subnate water was analyzed for soluble phosphorus to determine if a
substantial fraction of the sediment phosphorus had been released. In the case of B9B10
sediment the total mass of phosphorus in the final subnate water was about 1% of the
total mass of phosphorus remaining in the sediment fractions, and was about 0.4% for the
hydrated soil samples. For the WPBC sediment the subnate soluble phosphorus was less
than 0.1 % of the remaining sediment phosphorus. Based on these results it was .
reasonable to assume that the wet sieving process had
sediment particulate phosphorus content.
Fraction analysis
The various fractions were analyzed for:
1. Total dry mass
2. Volatile matter/Ash content
3. Particle specific gravity
Total phosphorus content
4.

74
In addition two specific particle size cuts from the B9B10 and WPBC sediments
were given more intensive evaluation. The smallest particle size fraction, the -400 USS (-
38 micron) cuts, and a larger particle size fraction, the -100 USS +200 USS (75-150
micron) cuts, were selected fqr evaluation of phosphorus forms present and also for
development of phosphorus adsorption/desorption data.
The phosphorus forms evaluated (See Appendix C for methods) were
1. Bicarbonate extractable, which represents the labile fraction of phosphorus
2.
that is readily available for biological uptake,
jyfcL ,, ...
Hydrochloric acid extractable, which represents the nominal maximum
amount of phosphorus present that is exchangeable with aqueous solutions
under environmental conditions and, by difference,
3. Y Residual phosphorus, which is relatively refractory, usually organically
bound phosphorus, that requires biochemical breakdown of the containing
organic matrix in order to become biologically available.
The adsorption/desorption data were developed using the methods described later
in this chapter.
Particle size distribution
Soil Dry and Hydrated Figure 5.1 shows the results of the dry screening before
and the wet screening after the eight day hydration. There appeared to be substantial de
agglomeration resulting from the hydration, as indicated by the reduction of the mass
1 I
mean particle size from approximately 800 micrometers in the dry soil down to
approximately 250 micrometers in the hydrated soil. This de-agglomeration appeared to
produce a large population of particles in the most-readily transportable range of <38
micron diameter (-400 USS) as evidenced by the increase of this mass fraction from 0.4%
dry to 11.7% hydrated.

75
B9B10 Soil Panicle S12E Distnbution-Hydraied and Dry
Screen Opening Size-Micrometers
*Hydrated B9B10 Soil O
-Dry B9B10Soil
Figure 5.1: Particle Size Distribution of Dry and Hydrated B9B10 Soil
Drainage Ditch B9B10 Sediment The screening process for the drainage ditch
sediment showed a phenomenon which had not been in evidence in the soil screening
study. As the sediment was subjected to its initial rough screening through a 1 cm screen
it quickly became obvious that there were large masses of filamentous growths present in
the 1 cm screen retntate. This growth contributed to screen blinding and caused a great
deal of difficulty in the initial screening process. Samples of the material were identified
by the University of Florida Center for Aquatic Plants as an algae of the genus lyngbya a
hardy filamentous algae which thrives in neutral to alkaline pH conditions and is
ubiquitous throughout Florida.
Lyngbya had been observed to be present to some extent at all ten of the test farms
participating in the BMP studies, in some cases as floating mats, so, in retrospect, the
presence of the material in the sediment should have been expected. The screening
process was carried out with particular care to insure that washing of the rough screen
with decanted subnate was adequate to release smaller size matter held within the

76
filamentous matrix. The retained greater than 1 cm material, including the filamentous
algae was dried as one sample, but after drying material that was obviously lyngbya was
separated manually by visual inspection under magnification and treated as a separate
sample. The algae rapidly lost its visual identity with decreasing particle size. Under
microscopic examination at particle sizes of less than 1180 micrometers (-16 USS) it was
difficult to discern coherent entities within the sediment matrix that could be identified as
lyngbya fragments. The remainder of the screening runs at sizes less than 1 cm
proceeded normally.
After final separation it was determined that the algae that had been retained on
the 1 cm screen constituted approximately 15% of the total dry mass of the drainage ditch
sediment. The soil and hydrated soil samples were prescreened to remove all material
greater than 10,000 micrometers. For the purposes of particle size comparison, the mass
distribution in the ditch sediment was calculated on the basis of the 10,000-micrometer-
and-less fractions only. Figure 5.2 shows the comparison of distributions between the
hydrated soil and the drainage ditch sediment. It is evident here that in terms of
potentially transportable matter, the particle size distributions of the two samples are
quite similar. The mass mean particle sizes of both samples are in the range of 240-250
micrometers and the distributions of the lower 50th percentile of particle sizes are almost
identical. In the upper 50th percentile of particle sizes the drainage ditch sediment is
more heavily weighted toward the larger particle sizes. The fraction of most readily
transportable material, the <38 micrometer size, was 10.4% of the total mass considered.
West Palm Beach Canal (WPBC) Sediment There was no evidence of
filamentous algae in the WPBC sediment; in fact the WPBC sediment coarse fraction (>
1 cm) was very different in visual appearance from the farm sediment. The major
constituent of this fraction appeared to be shells and shell fragments of fresh water
mollusks and some few limerock and marl fragments. The wet screening of the WPBC
sediment proceeded normally.

77
Figure 5.2: Particle Size Distribution of Hydrated B9B10 Soil and B9B10 Sediment
Figure 5.3 shows the distribution (again excluding the greater than 1 cm
fraction) for the WPBC sediment superimposed on that of the UF9200 soil and drainage
ditch sediment. The mass mean particle size of the WPBC sediment is about 230
micrometers, similar to that