• TABLE OF CONTENTS
HIDE
 Title Page
 Dedication
 Acknowledgement
 Table of Contents
 List of Tables
 List of Figures
 Abstract
 Introduction to the Everglades...
 Posing the questions
 Modeling the "River of Grass"
 A cross scale exploration of the...
 An end and a beginning
 Bibliography
 Biographical sketch














Title: Spatial and temporal dynamics in the Everglades ecosystem with implications for water deliveries to Everglades National Park /
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00097382/00001
 Material Information
Title: Spatial and temporal dynamics in the Everglades ecosystem with implications for water deliveries to Everglades National Park /
Physical Description: xvii, 239 leaves : ill. ; 29 cm.
Language: English
Creator: Gunderson, Lance H., 1952-
Publication Date: 1992
Copyright Date: 1992
 Subjects
Subject: Water-supply -- Florida -- Everglades   ( lcsh )
Hydrology -- Florida -- Everglades   ( lcsh )
Everglades National Park (Fla.)   ( lcsh )
Environmental Engineering Sciences thesis Ph. D
Dissertations, Academic -- Environmental Engineering Sciences -- UF
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
 Notes
Thesis: Thesis (Ph. D)--University of Florida, 1992.
Bibliography: Includes bibliographical references (leaves 226-238).
Additional Physical Form: Also available on World Wide Web
General Note: Typescript.
General Note: Vita.
Statement of Responsibility: by Lance H. Gunderson.
 Record Information
Bibliographic ID: UF00097382
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 001790405
oclc - 29233125
notis - AJL4064

Downloads

This item has the following downloads:

PDF ( 9 MBs ) ( PDF )


Table of Contents
    Title Page
        Page i
        Page i-a
        Page ii
    Dedication
        Page iii
    Acknowledgement
        Page iv
    Table of Contents
        Page v
        Page vi
    List of Tables
        Page vii
    List of Figures
        Page viii
        Page ix
        Page x
        Page xi
        Page xii
        Page xiii
        Page xiv
        Page xv
    Abstract
        Page xvi
        Page xvii
    Introduction to the Everglades ecosystem
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
    Posing the questions
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
    Modeling the "River of Grass"
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
    A cross scale exploration of the Everglades landscape
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
        Page 99
        Page 100
        Page 101
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
        Page 115
        Page 116
        Page 117
        Page 118
        Page 119
        Page 120
        Page 121
        Page 122
        Page 123
        Page 124
        Page 125
        Page 126
        Page 127
        Page 128
        Page 129
        Page 130
        Page 131
        Page 132
        Page 133
        Page 134
        Page 135
        Page 136
        Page 137
        Page 138
        Page 139
        Page 140
        Page 141
        Page 142
        Page 143
        Page 144
        Page 145
        Page 146
        Page 147
        Page 148
        Page 149
        Page 150
        Page 151
        Page 152
        Page 153
        Page 154
        Page 155
        Page 156
        Page 157
        Page 158
        Page 159
        Page 160
        Page 161
        Page 162
        Page 163
        Page 164
        Page 165
        Page 166
        Page 167
        Page 168
        Page 169
        Page 170
        Page 171
        Page 172
        Page 173
        Page 174
        Page 175
        Page 176
        Page 177
        Page 178
        Page 179
        Page 180
        Page 181
        Page 182
        Page 183
        Page 184
        Page 185
        Page 186
        Page 187
        Page 188
        Page 189
        Page 190
        Page 191
        Page 192
        Page 193
        Page 194
        Page 195
        Page 196
        Page 197
        Page 198
        Page 199
        Page 200
        Page 201
        Page 202
        Page 203
        Page 204
        Page 205
        Page 206
        Page 207
        Page 208
        Page 209
        Page 210
        Page 211
        Page 212
        Page 213
        Page 214
        Page 215
        Page 216
    An end and a beginning
        Page 217
        Page 218
        Page 219
        Page 220
        Page 221
        Page 222
        Page 223
        Page 224
        Page 225
    Bibliography
        Page 226
        Page 227
        Page 228
        Page 229
        Page 230
        Page 231
        Page 232
        Page 233
        Page 234
        Page 235
        Page 236
        Page 237
        Page 238
    Biographical sketch
        Page 239
        Page 240
        Page 241
        Page 242
Full Text








SPATIAL AND TEMPORAL DYNAMICS IN THE EVERGLADES ECOSYSTEM
WITH IMPLICATIONS FOR WATER DELIVERIES TO EVERGLADES
NATIONAL PARK











By

LANCE H. GUNDERSON


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


1992


























Copyright 1992

by

Lance H. Gunderson
























Dedicated to Gene, Dorothy and Sara Gunderson

My immediate links with the past and the future











ACKNOWLEDGMENTS

The author is indebted to numerous individuals and agencies, without

whom this work would not have been completed. First and foremost, my
deepest thanks go to Buzz Holling, for his support and guidance in the midst of
shifting paradigms, but most of all for teaching me that it is important to have

fun in the endeavor of science. I am very grateful to J.J. Delfino, who has been

extremely supportive as chairman of the committee. Committee members

Warren Viessman, Ronnie Best and Dan Spangler all provided expert comments

and assistance. Clay Montague and H.T. Odum are acknowledged for their
input in the early stages of the work.
Friends, family and colleagues all assisted with various tasks. John

Stenberg is gratefully acknowledged for his help on tasks too numerous to list.

Dave Sikkema, George Schardt, and Bob Johnson provided data from Everglades
National Park. John Stenberg and Alan Herndon were integral parts of the

evapotranspiration studies. Steve Davis, John Richardson, and Jennifer Silviera

all provided maps used in the cross scale analyses of the vegetation. Steve Light
helped enlighten (groan) me on the theories and practice of public policy and

water resources. No thanks would be too much for Carl Walters. Candy Lane,

Toni Carter, Keiley and Kenny Pilotto all helped with tasks along the way.
Finally, I am grateful to Bev, for taking care of all the details that enabled me to

tackle this project, and everything else.
This work was supported by grants from the South Florida Water

Management District, the Division of Sponsored Research at the University of

Florida and funds from the Arthur R. Marshall, Jr. Chair in Ecological Sciences.












TABLE OF CONTENTS

pae

A CKN O W LED G M EN TS ............................................................................................ iv

LIST OF TABLES .......................................................................................................... vii

LIST O F FIG U R ES ...................................................................................................... viii

A BST R A C T ............................................................................................................. xvi

CHAPTER I. INTRODUCTION TO THE EVERGLADES ECOSYSTEM ......... 1
The Everglades Ecosystem........................................ ........................... 3
H ydrology ......................................... ................. ........................... 7
History of Human Use ........................................ .................................... 10
W after M anagem ent.................................................................................... 12
Water Deliveries to Everglades National Park .......................................... 18
Managing Ecological Systems ............................................................. 23
Sum m ary ..................................................................................................... 25

CHAPTER 2. POSING THE QUESTIONS................................................... 26
Views of Ecosystem Structure and Function............................................ 27
H ypotheses .................................................................................................. 31
Water Deliveries to Everglades National Park The First
H ypothesis Set ................................................ ........................ 32
Cross-Scale Patterns In The Everglades Ecosystem The
Second Hypothesis Set ............................................ ........... ... 33
O objectives ................................................................................... . ............... 34

CHAPTER 3. MODELING THE "RIVER OF GRASS"......................................... 36
Background ................................................................................................. 38
Evapotranspiration in Wetlands ................................... ............ 38
Measurements of Evapotranspiration in southern Florida............ 40
Transpiration from Three Everglades Plant Communities............ 43
Measurements of community evapotranspiration...................... 46
Flow in W wetlands ...................................... .................................. 51
Ecological Models and Scale............................................................... 54
Model Description and Development................................................... 56
Hydrologic Components ................................................................. 56
Vegetation Com ponents ...................................................................... 59








Changes in Landscape Vegetation Types...................... ............ 62
Development of Flow Coefficients for Landscape Units ............ 65
Development of Evapotranspiration Coefficients for Landscape
U n its ...................................................................................... . ............... 67
R esu lts .......................................................................... . ............ .................. 70
Sensitivity Analysis- Flow and Evapotranspiration ..................... 71
Agreement with Historic Data .................................... ........... .. 73
Linkages between hydrology and vegetation................................ 77
Flow and upstream area.................................................................... 84
Summary ..................................................................................................... 89

CHAPTER 4. A CROSS SCALE EXPLORATION OF THE EVERGLADES
LANDSCAPE. .................................................................................................... 93
Methodology........................................................................ .. ........................ 94
Methods to Detect Discontinuities........ ...................................... 100
Methods to Analyze Patterns of Self-Similarity............................. 103
Techniques of estimating fractal dimensions................................. 107
Fourier analysis .................................................................................... 112
Summary of Methodology................................................................ 113
Data Sets Used In The Cross-Scale Analyses ............................................... 116
Fire data ........................................................... .............................. 122
Hydrologic data sets .......................................................................... 127
Evaporation Data ................................................................................. 127
Sea Level Data................................................ .................................... 128
Results Of Cross-Scale Analyses .................................................................. 129
Topographic data ................................................................................. 132
Vegetation Patterns ............................................................................ 138
F ire s ........................................................................................................ 144
R ain fall ................................................................................................... 153
W ater Levels ......................................................................................... 173
W ater Flow ............................................................................................ 184
F ires ........................................................................................................ 18 7
Sea Level .......................................................................................... 187
Temperature and Pan Evaporation ................................................. 192
Discussion of Results ...................................................................................... 203
Su m m ary ........................................................................................................... 214

CHAPTER 5. AN END AND A BEGINNING...................................................... 217
A Su m m ary ....................................................................................................... 217
Understanding Ecosystem Dynamics through Alternative
P arad igm s.......................................................................... ............................. 221
Prognosis for System Restoration ................................................................ 223

LITERATURE CITED .................................................................................................. 226

BIOGRAPHICAL SKETCH .................................................................................. 239












LIST OF TABLES

Table 1. History of major water management structures in the
Everglades ecosystem that influenced water deliveries to
Everglades N national Park. ..................................... ............. 14

Table 2. History of major water management policies that influenced
water deliveries to southern Everglades and Everglades
N national Park ............................................... ......................... 15

Table 3. Daily transpiration rates for three vegetation types in the
Everglades ....................................... .................... .................. 45

Table 4. Description of vegetation categories (landscape units) used
in Everglades m odel. ............................................ ........... .... 61

Table 5. Vegetation density and related flow coefficients as a function
of depth for sawgrass, tree island, wet and marl prairie
vegetation types..................................................................... ..... 64

Table 6. Spatially weighted flow coefficients for each landscape unit
used in the Everglades M odel. .................................. ............ 67

Table 7. Average daily and annual evapotranspiration for plant
communities used to develop evapotranspiration
coefficients for Everglades model.............................................. 68

Table 8. Summary of data sets used in cross-scale analyses of
Everglades ecosystem .................................................................. 130

Table 9. Summary of cross-scale analyses of spatial data sets,
indicating break points detected by fractal and gap
an alyses. .......................................................................................... 131

Table 10. Summary of cross-scale analyses of temporal data sets,
indicating dominant frequencies found in each set................ 131












LIST OF FIGURES


Figure 1. Location of the Kissimmee River, Lake Okeechobee and
Everglades Drainage Basin in Southern Florida.................... 4

Figure 2. Current broad scale land-use designations in the historic
freshwater Everglades drainage basin...................................... 6

Figure 3. Mean daily transpiration rates from sawgrass, tree island
and marl prairie plant communities.......................................... 47

Figure 4. Time course of mean daily evapotranspiration (cm/day) at
study sites P33 and P37 from February 1985 through
Septem ber 1986. ............................................. ....................... 49

Figure 5. Model grid used to depict hydrologic and vegetation
dynamics in Everglades ecosystem. .......................................... 57

Figure 6. Location of sample rain and stage gauges, flow sections and
pan evaporation sites within the Everglades region................. 60

Figure 7. Percent cover of sawgrass, wet prairie, and tree island in
various w indow sizes. .......................................... ............ ... 63

Figure 8. Effects of doubling and halving flow coefficients on
simulated flow through Tamiami flow section. ...................... 72

Figure 9. Effects of varying base evapotranspiration rates on
simulated flow through Tamiami flow section. ......................74

Figure 10. Time series of simulated (solid lines) and actual (dashed
lines) stages at gauge P33 from 1960 to 1988........................ 75

Figure 11. Time series of simulated (solid lines) and actual (dashed
lines) stages at gauge P35 from 1960 to 1988............................ 76

Figure 12. Map codes for initial landscape vegetation types in the
Everglades model. Dominant code in the Everglades
proper is type 1. Cover types developed from Davis
(1943). Explanation of codes is found in Table 4 .................... 79







Figure 13. Map codes of landscape vegetation types in grid cells of
Everglades model at end of 28 year simulation run. Note
presence of type 3 codes in right -central portion of array.
See Table 4 for explanation of codes....................................... 81


Figure 14.


Figure 15.





Figure 16.


Figure 17.



Figure 18.





Figure 19.


Results of changing vegetation patterns on simulated flow
through three flow sections in the southern Everglades.......... 83

Difference in predicted stage between models with and
without vegetation linkages of succession,
evapotranspiration and flow. The model without
linkages tended to predict higher stages (less than 0.5 ft),
under both the natural and managed scenario........................ 85

Three dimension plot showing increase in annual flow (z) as
upstream area is increased (y) over a 28 simulated year. ........ 87

Log-log plot of simulated flow versus upstream area. Scatter
due to differences in annual rainfall. Gray area
represents bounds of actual flow and equivalent area............. 90

Rank order plot, difference indices 1 and 2 for mock data set
designed to simulate continuous normal distribution.
Distribution has mean of 50. For smooth curves such as
this, no value of DI1 is significant. No sequential values
of D I2 are significant ............................................. ........... .... 99

Rank order plot, difference indices 1 and 2 for mock data set
derived from uniform random distribution. Values of
both DI1 and DI2 are significant, yet values between gaps
exhibit no pattern ......................................................................... 100


Figure 20. Rank Order Plot, Difference Indices 1 and 2 for mock data
set designed to show gaps. Gaps are values of difference
indices (1 and 2) outside a 95% confidence interval.
Clumps are not evident, but indicated by a pattern of
uniformity small groups of difference indices. ....................... 101


Figure 21.


Figure 22.


Sample Koch curves to show patterns of increasing fractal
dimension (1.5 1.7), from Mandelbrot (1983)......................... 108

Example of first four levels or divisions for rendering data
into a quadtree storage form at. .................................................. 111







Figure 23. Log-log plots of tile size versus number of tiles for three
Koch Curves of known fractal dimension. Fractal
dimension is estimated by the negative slope of curves.
Actual fractal dimensions (D) of curves are A) D=1.5, B)
D=1.63, C) D=1.71, as shown in Figure 21. .............................. 113

Figure 24. Fourier analysis of random mock time-series data, showing
time-series plot (top) and spectral plot (bottom). This
data set was created by taking 64 samples from a uniform
random number distribution. No peaks are significant in
the spectral plot. ........................................................................... 116

Figure 25. Fourier analysis of structured mock time series data,
showing time-series plot (top) and spectral plot (bottom).
This data set was created by combination of two sine
w av es ................................................................... ......................... 117

Figure 26. Top transect m aps................................................................................ 120

Figure 28. Sample vegetation map, showing patterns of sawgrass
w within a 160 m w indow ............................................................... 125

Figure 29. Sample vegetation map, showing pattern of sawgrass within
a 1600 m sam ple w window ............................................................ 126

Figure 30. Sample vegetation map, showing pattern of sawgrass within
a 16 km sample window. ............................................... 127

Figure 31. Location and size of sample window within which recent
fire histories were analyzed for spatial and temporal
patterns. ....................................................................................... 128

Figure 32. Topographic surveys from transects 1 through 4 ( see Figure
4 for locations) showing elevational variation with east
w est distance. .................................................................................. 135

Figure 33. Topographic surveys from transects 5 through 8 ( see Figure
4 for locations) showing elevational variation with east
w est distance. .................................................................................. 136

Figure 34 Topographic surveys from transects 9 through 11 ( see
Figure 4 for locations) showing elevational variation with
east w est distance. .......................................................................... 137







Figure 35.





Figure 36.






Figure 37.






Figure 38.


Log-log plot of transect length versus step length (top plot)
used to estimate breaks in fractal dimension. Bottom plot
shows results of rolling regressions. Break in top plot is
indicated at point of maximum regression coefficient
(bottom plot). ................................................................................. 138

Log-log plots of box size versus box count used to estimate
fractal dimensions of sawgrass vegetation within three
sample windows. Estimates of fractal dimension are
given by slope of regression. Data from three windows
are combined in lower right plot, and suggest a break in
the fractal dim ension. ................................................................. 141

Log-log plots of box size versus box count used to estimate
fractal dimensions of wet prairie vegetation within three
sample windows. Estimates of fractal dimension are
given by slope of regression. Data from three windows
are combined in lower right plot, and suggest a break in
the fractal dim ension. ................................................................. 142

Log-log plots of box size versus box count used to estimate
fractal dimensions of sawgrass vegetation within three
sample windows. Estimates of fractal dimension are
given by slope of regression. Data from three windows
are combined in lower right plot, and suggest a break in
the fractal dim ension. ................................................................. 143


Figure 39. Rank order plot (top plot) and difference indices (bottom
plot) for sawgrass patches sampled from 1600 m
w window s. ........................................................................................ 145


Figure 40.



Figure 41.



Figure 42.


Rank order plot (top plot) and difference indices (bottom
plot) for wet prairie patches sampled from 1609 m
w window s. ............................................................................ 147

Rank order plot (top plot) and difference indices (bottom
plot) for tree island patches sampled from 1609 m
w window s. .................... ........................... ....... .... 148

Rank order plot (top plot) and difference indices (bottom
plot) for sawgrass patches sampled from 16 km
w indow s. ............................. ................. .... .... 149


Figure 43. Rank order plot (top plot) and difference indices (bottom
plot) for wet prairie patches sampled from 16 km
windows...... ............................... ..........150







Figure 44.



Figure 45.


Figure 46.


Figure 47.


Figure 48.


Figure 49.



Figure 50.



Figure 51.


Figure 52.



Figure 53.


Figure 54.



Figure 55.



Figure 56.


Rank order plot (top plot) and difference indices (bottom
plot) for tree island patches sampled from 16 km
w indow s. ......................................................................................... 151

Rank order plot of log fire sizes from Shark River Slough,
1958-1979. ........................................................................................ 153

Difference Indices 1 and 2 used to determine gaps in fire
sizes, Shark River Slough, 1958-1979 ......................................... 154

Time series plot of daily rainfall data from Tamiami Ranger
Station, 1949 through 1977. ......................................................... 156

Time series plot of daily rainfall data from Royal Palm
Station, 1949 through 1977. ......................................................... 157

Spectral plots from Fourier analysis of daily rainfall from
Tamiami Ranger Station, indicating dominant annual and
m monthly cycles. ............................................................................. 158

Spectral plots from Fourier analysis of daily rainfall from
Royal Palm Station, indicating dominant annual and
m monthly cycles. ............................................................................. 159

Rank order plots of daily rainfall data, Tamiami and Royal
Palm Stations. ................................................................................ 160

Difference index 2 versus rank order for daily rainfall data.
Bars represent value of index, gray areas represent mean
+ 95% C .I........................................................................................ 161

Time series plot of monthly rainfall data from Royal Palm
and Tamiami Ranger Stations, 1949 through 1977.................. 163

Spectral plots from Fourier analysis of monthly rainfall from
Tamiami Station, indicating dominant annual and
m monthly cycles. ............................................................................. 164

Spectral plots from Fourier analysis of monthly rainfall from
Royal Palm Station, indicating dominant annual and
m monthly cycles. ............................................................................. 165

Rank order plot and difference index 2 for monthly rainfall
data from Tamiami Ranger Station. Arrows locate
significant gaps. ...................................... 166







Figure 57.



Figure 58.


Figure 59.


Figure 60.



Figure 61.


Figure 62.



Figure 63.


Figure 64.


Figure 65.


Figure 66.


Figure 67.


Figure 68.


Figure 69.


Figure 70.


Figure 71.


Rank order plot and difference index 2 for monthly rainfall
data from Royal Palm Station. Arrows locate significant
g ap s. ........................................................................................ . 167

Time series of total annual rainfall at Tamiami and Royal
Palm Stations. ............................................................................... 168

Spectral plots from Fourier analysis of annual rainfall data
from Royal Palm and Tamiami Stations................................. 170

Rank order plot and difference index 2 for monthly rainfall
data from Royal Palm Station. Arrows locate significant
g ap s. ........................................................................................ . 171

Hierarchical Cluster Tree for annual rainfall at Royal Palm
Station ............................................................................................. 172

Rank order plot and difference index 2 for monthly rainfall
data from Tamiami Ranger Station. Arrows locate
significant gaps. ............................................................................. 173

Hierarchical Cluster Tree for annual rainfall at Tamiami
R anger Station................................................... ........................... 174

Time series plots of monthly water levels at gauging stations
P33 and P38, from 1958-1979. ..................................................... 176

Time series plots of monthly water levels at gauging stations
P35 and P37, from 1958-1979. ..................................................... 177

Spectral plots from Fourier analysis of monthly water level
data from four P stations............................................................. 178

Spectral plots from Fourier analysis of daily water level data
from four P stations. .................................................................... 179

Rank order plots of daily water level data from stations P33
and P 35. .......................................................................................... 180

Rank order plots of daily water level data from stations P37
and P 38. .......................................................................................... 181

Rank order plot and difference index 1 for monthly water
level data at station P33. .............................................................. 182

Rank order plot and difference index 1 for monthly water










Figure 72.


Figure 73.


Figure 74.


Figure 75.


Figure 76.


Figure 77.


Figure 78.



Figure 79.



Figure 80.



Figure 81.


Figure 82.


Figure 83.



Figure 84.


Figure 85.


level data at station P35. .............................................................. 183

Rank order plot and difference index 1 for monthly water
level data at station P37. .............................................................. 184

Rank order plot and difference index 1 for monthly water
level data at station P38............................................................... 185

Time series plot of monthly water flow across Tamiami Trail
flow section 1940-1982. ................................................................ 187

Spectral plots from Fourier analysis of monthly flow across
Tamiami Trail flow section. ........................................................ 188

Rank order plot of monthly water flow across Tamiami flow
section ............................................................................................. 190

Difference indices 1 (top plot) and 2 (bottom plot) for
monthly flow data, Tamiami flow section................................ 191

Time series plot of log fire sizes (top plot) from 1958-1979,
and spectral analysis (bottom plot) indicating dominant
cycles in fire data. ........................................................................... 192

Time series plots of mean monthly sea level elevation at Key
West, Miami, and Naples for the time period 1910
through 1990. ................................................................................. 193

Spectral plots from Fourier analysis of detrended sea level
data from Key West. Dominant cycle is the annual
period .............................................................................................. 195

Spectral plots from Fourier analysis of detrended sea level
data from Miami. Dominant cycle is the annual period.......... 196

Spectral plots from Fourier analysis of detrended sea level
data from Naples. Dominant cycle is the annual period......... 197

Time series plot of mean monthly minimum and maximum
air temperatures from Belle Glade and Tamiami Ranger
S tatio n s............................................................................................. 198

Spectral plots from Fourier analyses of maximum and
minimum monthly temperature data from Belle Glade.......... 199

Spectral plots from Fourier analyses of maximum and










Figure 8(


Figure 8



Figure 8



Figure 8


Figure


Figure



Figure 9



Figure



Figure 9



Figure 95


minimum monthly temperature data from Tamiami
R anger Station............................................................................... 200

6. Time series plots of monthly pan evaporation from Belle
Glade and Tamiami Ranger Stations, 1965 through 1991........ 201

.7. Spectral plots from Fourier analyses of monthly pan
evaporation from Belle Glade and Tamiami Ranger
Stations, 1965 through 1991. ....................................................... 202

8. Time series plots of daily evapotranspiration and pan
evaporation at site P33, January 1985 through October
1986 ......................................................................................... . 203

9. Time series plots of daily evapotranspiration and pan
evaporation at site P37, January 1985 through October
1986 ................................................................................................. 204

90. Spectral plots from Fourier analyses of daily
evapotranspiration at sites P33 and P37. Bars represent
magnitude of cycle, gray 95% C.I................................................ 206

)1. Spectral plots from Fourier analyses of daily pan
evaporation at sites P33 and P37. Bars represent
m agnitude of cycle, gray 95% C.I. ............................................... 207

2. Results of mini-model frequency of rain input (A), ET
outflow (B), and stage fluctuations (C). Note difference
in time range, stage plot covers 40 simulated years................ 211

)3. Hydrologic hierarchies in the Everglades ecosystem,
showing scales of dominant frequencies in surface water
and atm ospheric variation. ......................................................... 213

,4. Topographic hierarchies in the Everglades ecosystem,
indicating breaks between microtopographic and
macrotopographic features and processes. .............................. 215

. Vegetation hierarchies in the Everglades ecosystem, showing
scales of plant species, communities, landscape units and
the Everglades ecosystem as defined by breaks in the
fractal dim ension............................................................................ 217












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


SPATIAL AND TEMPORAL DYNAMICS IN THE EVERGLADES ECOSYSTEM
WITH IMPLICATIONS FOR WATER DELIVERIES TO EVERGLADES
NATIONAL PARK

Lance H. Gunderson

December 1992


Chairman: Joseph J. Delfino
Cochairman: C.S. Holling
Major Department: Environmental Engineering Sciences


The Everglades is a unique wetland ecosystem. During this century, the

ecosystem has been partitioned for disparate uses of human habitation,

agriculture, water conservation and ecosystem conservation in a national park.

The sustainability of Everglades National Park is dependent upon upstream

water sources. Water management in the Everglades and water deliveries to the

Park are linked to human perceptions of ecosystem dynamics.

One line of inquiry used expansion of a state-of-the-art computer model to

examine the upstream area that once contributed water to the Park. Linkages

between vegetation and hydrology were added as vegetation mediation of

evapotranspiration and flow and hydrologically induced vegetation changes, but

neither addition appreciably improve understanding of hydrodynamics in the

Everglades system at the scale of the model. Prior to management, the entire

system, south of Lake Okeechobee, contributed flow to Everglades Park except







during dry years. Since the onset of intensive water management, an equivalent

area of only about one-third of the historic drainage basin has supplied water

into the Park. But these conclusions are dependent upon the assumptions made

to represent the system at a specific spatial-temporal scale in a model. At other

scales the conclusions could well be different. That led to the second major topic

of this thesis; that of cross-scale structure and dynamics.

A cross-scale mode of inquiry suggests that ecosystems exhibit

discontinuities in spatial structures and temporal patterns across time and space

due to the interaction of key processes operating over different scale ranges.

Spatial patterns in the topography, vegetation and fire data sets exhibited scale

regions of self-similarity separated by distinct breaks. Temporal patterns of

rainfall, stage, flow, evaporation and sea-level exhibited multiple cycles. These

analyses support the theory that ecosystems are structured around a few

keystone variables of mixed spatial and temporal dimensions. Dramatic

discontinuities appear in patterns as a result of the interactions of processes

operating at different space and time domains. This emerging viewpoint of

ecosystem structure and dynamics will hopefully provide a basis for new

understanding and hence improved management of this unique ecosystem.


xvii












CHAPTER 1.
INTRODUCTION TO THE EVERGLADES ECOSYSTEM


There are no other Everglades in the world.
-Marjory Stoneman Douglas

The Everglades is a wetland ecosystem unlike any other on Earth.

Situated in the subtropics of southern Florida, the unique combination of
physiography and biota blend into a landscape whose name is internationally

recognized. Undoubtedly some of the values and distinctions that the area now

holds are due to attributes of the natural system. During the twentieth century,
the human population in and around the Everglades ecosystem has increased

dramatically, resulting in a myriad of demands on and uses of a unique

ecosystem. Many of the current management problems are associated with the

historical spatial partitioning of resources within a once contiguous ecosystem.

The pattern unfolding throughout the past century is one of a transition of land

uses, from a pristine wetland with negligible human use to one dominated by a
variety of human uses each with characteristic spatial and temporal domain.

These varied land uses range from intensive agriculture in the northern

Everglades to Everglades National Park in the south.
The primary purpose of this work is to improve understanding of the

critical processes and factors in the Everglades ecosystem that influence water

deliveries to Everglades National Park. The issues surrounding water deliveries

to the park cannot be described from a single moment in history nor from a
spatial perspective of the current border of the park. Indeed, the water problems







of the park are woven throughout a rich tapestry extending back thousands of

years and covering the southern half of peninsular Florida. In order to make the

problem tractable, the dissertation is divided into five chapters. The introductory

chapter contains descriptions of the natural and human histories in the system

that lead to a conclusion that water management is fundamentally linked to

concepts of ecosystem dynamics. The second chapter in this work compares and

contrasts alternative concepts of ecosystem dynamics from which the hypotheses

and objectives of the dissertation are derived. The third chapter presents the

results of an attempt to invalidate the hypothesis that the entire Everglades

drainage basin contributed water to the park. The "upstream area" hypothesis

was tested using a state-of-the-art ecosystem model that couples vegetation and

hydrologic dynamics. The fourth chapter contains the results of analyses of a

series of data sets that are used to test the second hypothesis, based on an

alternative concept, and seeks to understand system dynamics based upon a

recognition of the role of discontinuities in both structure and processes The

final chapter contains a summary that compares and contrasts the understanding

of system dynamics and water deliveries that were developed in chapters three

and four and presents implications of these results on water management and

policy.

This introductory chapter is devoted to describing both the natural and

human components in the Everglades ecosystem. First, the components and

processes of the pristine or natural Everglades ecosystem are described. The next

three sections are historical accounts, documenting the increasing human

involvement with the system. These historical accounts include a brief history of

relevant human activities, a review of how water management developed in

southern Florida, and finally, how the understanding and policies of water

deliveries to Everglades Park have changed. This chapter concludes with a







description of the linkages between management, policy and understanding of

system dynamics, and how these relationships have evolved in this complex

wetland system.


The Everglades Ecosystem

The Everglades is a distinct physiographic region located in southern

Florida, and its natural features have been described for over 100 years (Heilpren

1887, Willoughby 1898, Harshberger 1914, Harper 1927, Davis 1943, Craighead

1971, Gunderson and Loftus In Press). Prior to intervention by man, the

Everglades encompassed approximately 10,500 km2 of freshwater marshes,

sloughs and hardwood tree islands (Davis 1943). The system was approximately

210 km along the north-south axis, bordered by Lake Okeechobee on the north

and Florida Bay on the south. The widest east-west dimension was 77 km, from

the higher Atlantic coastal ridge on the east to the Big Cypress Swamp on the

west (Figure 1).

The wetland complex is a result of a large arcuate trough in the
underlying limestone bedrock. Three surficial formations are recognized, and all

were formed by shallow marine accumulations during the Pleistocene Era,

primarily the Sangamon interglacial stage (Parker and Cooke 1944). The Fort

Thompson formation underlies the northern Everglades, and is comprised of

marine and freshwater marls beds interleaved with limestone and sandstone.

The Anastasia formation is found in the northeastern Everglades, and is

characterized by sandy limestone, calcareous sandstone. The surficial feature of

the southern Everglades is Miami limestone, comprised of oolitic and bryozoan

facies (Hoffmeister 1974).























ATLANTIC
OCEAN


BIG
CYPRESS
SWAMP


GULF OF
MEXICO


Florida Bay


Figure 1. Location of the Kissimmee River, Lake Okeechobee and Everglades
Drainage Basin in Southern Florida.


Historic Freshwater
Everglades
.-- Kissimee Lake
S Okeechobee-Everglades
Drainage Basin

I 0 16 32 km
S 0 10 20 mi


67 ev4





5


The wetland soils of the Everglades are Holocene sediments, categorized

as peats, mucks and marls, and are biogenic. The oldest soils in the Everglades

are approximately 5500 years old (Gleason et al. 1984), dating back to an

approximately 3 m transgression of sea level (Robbin 1984). Peats and mucks are

histosols, named by the dominant recognizable plant remains from which the

soils are derived, and accumulate under extended periods of inundation. The
marls are a calcitic mud, produced by reprecipitation of calcium carbonate from

saturated water during photosynthesis by blue-green algae (Gleason 1972).

The topography of both the bedrock and the soil surface is flat,

characterized by almost no relief with extremely low gradients. The maximum

elevations in the northern Everglades are approximately 5.3 m above the national

geodetic vertical datum (NGVD), and now occur in the Arthur R. Marshall

National Wildlife Refuge (Figure 2). The elevational gradient is mostly north to

south, with an average slope of 2.8 cm/km (Parker et al. 1955). The variation in

elevation is attributed to the underlying bedrock structure and accumulations of

organic sediments. The microtopographic variation is caused by and contributes

to differences in vegetation cover and type.

The vegetation of the Everglades region is a complex of gramineous and

woody wetland associations. The spatially dominant communities are sawgrass

marshes, wet prairies, and hardwood swamp forests (Davis 1943, Craighead

1971, Olmsted et al. 1980, Gunderson and Loftus In Press). Sawgrass marshes,

monotypic stands of sawgrass, Cladium jamaicense found over peat and marl, are

the ubiquitous, characteristic association of the Everglades. Wet prairies over

peat are sparsely vegetated, generally dominated by either spikerush Eleocharis

cellulosa, or maidencane, Panicum hemitomon. Wet prairies on marl are diverse

association, dominated by sawgrass and muhly grass, Muh-lenbergia filipes, and

contain over a hundred other species (Olmsted et al. 1980).












20 m
20 nru


FLORIDA
BAY


Figure 2. Current broad scale land-use designations in the historic freshwater
Everglades drainage basin.







The hardwood swamp forests of the Everglades are called tree islands,

descriptive of the isolated clumps of trees surrounded by the lower stature
wetland grass communities. Dominant species in the tree islands are mostly bay

trees; swamp bay, Magnolia virginiana, red bay, Persea palustris, dahoon holly, Ilex
cassine, wax myrtle, Myrica cerifera.
The Everglades is a unique wetland due in part to the spatial and
temporal patterns of the components of the hydrologic regime. Year to year

variation in the hydrologic cycle results in oscillating periods of flood and
drought. The intra-annual variation is also great, characterized by wet summers

and dry winters. Rainfall and overland flow are the principal inputs, yet the

relative contribution of each to the hydrologic budget is debated. The magnitude
of direct rainfall contribution to the hydrologic regime of the Everglades

distinguishes it from other large freshwater wetland systems such as the Llanos

in Venezuela, the Okavango in Botswana or the Pantenal in Brazil where most of

the marsh water originates from rivers.
The climate of southern Florida is subtropical, classified by Hela (1952) as

a tropical savanna, with insufficient rainfall during the summer months to

compensate for a winter dry season. The area has also been classified as

subtropical moist forest type (Dohrenwend 1977, Greller 1980) due to the high

annual rainfall and moderate annual biotemperatures.

Rainfall over the Everglades exhibits both spatial and temporal variability.

Annual rainfall over the system averages 130 cm (Thomas 1970, Bradley 1972,
MacVicar and Lin 1984). Annual rainfall extremes for the period of record 1940

through 1980 range from low of 95 cm in 1961 to a high of 270 cm in 1947

(MacVicar and Lin 1984). Thomas (1970), using spectral analysis, found about a
seven year pattern within the record, indicative of a cyclical pattern with this

return interval. Rainfall patterns exhibit distinct seasonality; approximately 85%







of the annual average rain falls between May and October. Thomas (1970) found

that rain totals during the wet season were bimodally distributed, with peaks in

June and September. Spatial analysis of rainfall records indicate that the coastal

region receives on the average 30 to 35 cm more than the interior marshes. The

northern Everglades and southern Everglades receive more rainfall than the

central regions (MacVicar and Lin 1984).

The rainfall patterns can be related to different processes which influence

the timing and amount of precipitation. The summer rainy season is attributed

to convective thunderstorms, which are linked to mesoscale land-sea breeze

patterns. During the summer, insolation results in differential heating of the air
over the land mass compared to air over the water. The heated air over the land

rises, creating low pressure and establishing a pressure gradient along which

maritime air flows toward the center of the Florida peninsula. The moisture-

laden air rises, cools adiabatically, condenses, and forms convective

thundershowers. This process has been described as the 'rain machine' (Pardue

1982, Yates 1982). Some authors suspect that rainfall totals have decreased

because drainage and development have altered the net radiation budget

(increased reflectance due to a higher albedo of developed areas) which in turn

decrease the rate of convection (Gannon 1978, Pardue 1982). Statistically higher

rain amounts measured during September have been explained by the greater

incidence of tropical cyclones during this month. Rain during the winter dry

season is associated with the passage of cold fronts that pass on the average

every seven days. Annual variations in frontal passage have been linked to jet

stream location.

Historically, surficial flow left the Everglades system through a number of

pathways. In the northeast, water moved through the cypress-dominated

Hungryland and Loxahatchee Sloughs. Major rivers that carried Everglades







waters through the coastal ridge include the New River, Little River and Miami

River. Surface water moved through the higher coastal ridge in a series of

transverse glades. Water flow in the southern Everglades occurred in the broad
shallow depressions of Taylor and Shark Sloughs, named for the rivers that
received the bulk of the flow.

Although the Everglades are recognized as a distinct physiographic

region, it is part of a much larger drainage system, containing a number of river
systems and Lake Okeechobee to the north (Figure 1). Prior to development,
hydrologic connections were traceable to central Florida where the Kissimmee

River originates. The Kissimmee is the largest of the rivers and other smaller

creeks and sloughs which empty water into Lake Okeechobee. Surface water
entered the Everglades from the southern boundary of the lake at two points
when stages exceeded 4.5 m NGVD, and a 52 km long spillway when stages

exceeded 5.6 m NGVD (Parker 1984). The entire system has been referred to as
the Kissimmee-Lake Okeechobee-Everglades (KLOE) system.

The hydrology of the historic Everglades ecosystem can be summarized as

follows. The system is a wide, shallow flat basin, with an overall small

topographic gradient. The primary hydrologic input is rainfall, and although

averages 1.3 m/yr, is characterized by wide spatial and temporal variability.

Other inputs occurred as surface and subsurface flow from Lake Okeechobee and

from wide sloughs in the Big Cypress Swamp. Evapotranspiration is the primary
avenue of water loss, estimated to be 80% of rainfall. Remaining water in the

system flows slowly to the south either to the east recharging the surficial aquifer

of the coastal ridge or the southwest entering the estuarine mangrove zone prior

to reaching Florida Bay. The seasonal patterns of rainfall and evapotranspiration
interact to yield distinct annual wet and dry periods as well as variations in

overland flow.







History of Human Use

Evidence of human inhabitance in southern Florida dates to well over

10,000 years. B.P. (Carr and Beriault 1984), prior to the existence of the vast

wetland ecosystem. Written accounts, which date back almost 500 years,

describe native humans using the resources of the wetland ecosystem. Nunez de

Cabeza (1514) relays descriptions of the fierce native Indian tribes that inhabited

the coastal portions of southern Florida and the peaceful Mayami tribes which

colonized the edge of Lake Okeechobee. These early Americans probably burned

the Everglades (Robertson 1954), and used the area for hunting and fishing

purposes.

The name Everglades first appeared on British maps in the early 1800s

(Vignoles 1823) probably a contraction of "Never a glade," descriptive of the large

treeless expanses. With the expansion of European derived settlers throughout

the southeastern coastal plain, native Americans translocated from the Carolinas

to southern Florida. The term "Seminole," which is the name for the major tribes

of current native populations that persist today, means "runaway." The Seminole

term for the area is "Pa-hay-okee", which loosely translates into "grassy lake",

again, descriptive of a non-forested wetland system. These native Americans

used the elevated tree islands for homesites and cultivation of crops, as well as

hunted and fished throughout the system. The United States fought a series of

wars with the Seminoles during the mid 1800s, restricting their territory to a few

reservations through south Florida. One remains within the Everglades proper,

where the Miccosukee Indians still retain land use rights.

The latter part of the nineteenth century marked the first influx of white

settlement and attempts at "reclamation" of the wasteland known as the

Everglades. Soon after Florida became a state in 1845, early settlers and their

governments embarked on programs to drain the Everglades for habitation and







agriculture. Buckingham Smith was commissioned by the U.S. Senate to

reconnoiter the Everglades for development potential (Smith 1848). In 1850,

under the Swamp and Overflowed Lands Act, the federal government deeded

7500 mi2 to the state, including the Everglades. The Florida legislature

established the Internal Improvement Fund, whose board was to sell and

improve these swamp lands through drainage. Attempts at manipulation of the
water were ineffective in the 1800s, as the magnitudes of the variations in
hydrology were far greater than the minor control structures could handle.
By 1900, initial colonization of the coastal regions east of the Everglades

was underway. The population of Palm Beach, Broward and Dade counties in

southern Florida in 1900 was 28,000 (U. S. Dept. of Commerce 1990). By 1920, the

major land uses now found in southern Florida had started. Urban development

was occurring along the railroad line down the east coast. Agriculture was
developing in the peat lands south of Lake Okeechobee. Conservation of the
natural resources had begun with the formation of Royal Palm State Park in 1917

in the southern Everglades.

During the period 1920 through 1990, the spatial extent of these land uses

grew, in large around these three general loci. In the 1940s, 283,000 ha of the

northern Everglades was designated as the Everglades Agricultural Area (EAA).

By the mid 1940s Water Conservation areas were designated in the central

regions of the glades to manage water resources for multiple purposes.

Conservationists work started during the 1920s came to fruition in 1935 with the

establishment of Everglades National Park in the southern Everglades, although

the park was not formally dedicated until 1947. The park area was increased in

1989 from 1.4 million to 1.6 million acres by the addition of Northeast Shark

Slough. Urban development along the east coast has followed the exponential

increase in population, and resulted in the drainage and colonization of former







wetland areas. As of 1990, 5.1 million people live within the confines of the

historic drainage basin (U. S. Dept of Commerce 1990). The current

configuration of the Everglades ecosystem depicting agricultural areas in the

north, water conservation areas in the central areas and Everglades National Park

in the south is shown in Figure 2.

Through the past century, the spatial extent of the historic Everglades

ecosystem has been slowly whittled away, to the degree that perhaps one-half of

the original system has been irrevocably converted to specific land uses. As of

1985 the historic Everglades ecosystem as defined by Davis (1943), was

partitioned into at least five major use types. Gunderson and Loftus (In Press),

estimate that 32% of the historic Everglades is in areas designated for water

management, 27% in agriculture, 17% for preservation of natural resources, 12%

has been developed for urban purposes, and 12% remains as drained,

undeveloped lands. Davis et al. (In Press), estimate that only half of the original

land area of the Everglades is still in native vegetation types, and that certain

landscape types, including a large pondapple forest in the north as well as

marshes and cypress forests in the east, are gone. The remaining natural areas

have probably been hydrologically altered, and their future viability is largely

dependent upon water management actions.


Water Management

Water management within the Everglades is accomplished by physical

structures and operational criteria. The physical structures consist of levees or

dikes, canals, water control gates (mainly weirs), and pumps. The operational

criteria are constructed around the multiple objectives of the system. The two

primary objectives are flood protection and water supply, having evolved with

the changes in land use within the system and the nature of the historical








ecosystem. The history of water management appears to be one where natural

events or crises precipitated plans and activities that resulted in more
infrastructure and attempts to control the variation in the natural system.
Reactions to natural crisis have resulted in changes and development of two
components of water management; the physical structures of water manipulation

(Table 1) and the policies and programs by which water is managed (Table 2).
Canal construction typified the earliest period of water management in the

Everglades. The first large canal in the system was completed in 1882, when

dredges excavated a channel between Lake Okeechobee and the Caloosahatchee
River. Water levels in the Lake were reported to have declined approximately 50

cm (Johnson 1958). During the next 45 years, canal construction proceeded

sporadically as a result of intermittent funding. By 1917, four major canals, the
Miami, North New River, Hillsboro and West Palm Beach had penetrated the

interior of the Everglades, probably resulting in some drainage of the wetland

system.
Hurricanes during the 1920s devastated human developments along the

east coast and south of Lake Okeechobee. Earthen dams which had been

constructed to exclude waters of the Lake were breached during the hurricane of

1928, resulting in extensive flooding and a loss of about 2400 lives (Blake 1980).

In response, the federal government funded the construction of the Hoover Dike

around the Lake, which was completed by 1938, in order to contain floodwaters.

During the 1940s, federal and state laws established the system of water
management as it now exists. Rainfall during this decade varied wildly, creating








History of major water management structures in the Everglades
ecosystem that influenced water deliveries to Everglades National
Park.


STRUCTURE


RESULT


Construction of Miami,
North New River, Hillsboro
and Palm Beach Canals

Construction of
Caloosahatchee Canal

Lake Okeechobee levees
Muck levee constructed
Hoover dike constructed

Construction of Tamiami Trail


Everglades Agricultural
Area levees completed

Water Conservation Areas
1, 2, and 3 enclosed by levees

S-12 structures complete


L-67 canal and levee


Drainage of coastal areas
and interior Everglades
wetlands

Lowering of water level in
Lake Okeechobee

Impound water in Lake
Okeechobee, control water
movement to south

Alteration of flow patterns
channel through culverts

Control of water movement
in northern Everglades

Control of water movement
in middle Everglades

Flow spatially constricted to
four flowways

Canal to deliver water into
center of Shark Slough


Table 1.


YEAR


1917


1924



1926
1938

1928


1959


1962


1962


1967








History of major water management policies that influenced water
deliveries to southern Everglades and ENP. (Blake 1980, Wagner
and Rosendahl 1985).


POLICY


PURPOSE


Everglades Drainage District


Flood Control Act PL 80-858



Deliver water from WCA 3A,
based upon stages



Deliver water based upon
stage in Lake Okeechobee


- Minimum delivery schedule
(PL 91-282 guarantee park a
amount of water)

- Flow through plan
(PL 98-181 allowed for
experimental deliveries)


1985 -
present


Rainfall plan


To drain Everglades for
agricultural and development

Ameliorate flood effects by
construction of conservation
areas, levees

Store water in WCA 3A,
park to receive after
storage requirements met


Increase flow to park during
hurricane season, restrict
flow during drought

To assure ENP 260,000 acre-
feet/year, and share in certain
drought adversity

Allow S-12 structures to
remain open, no regulation
schedule

Deliver water based upon
upstream climatic conditions


Table 2.


YEAR


1907


1948



1962
1966



1966
1970


1970
1982


1982
1985







conditions which prompted action. The early 1940s were extremely dry,resulting

in saline intrusion into the freshwater aquifers of the coast and subsequent salt

dam construction. Extensive flooding occurred during 1947, following an

extremely wet summer and the passage of two cyclonic storms. Over 105 inches

(270 cm) of rain was reported to have fallen (MacVicar and Lin 1984) during

1947. This flood resulted in the passage of the federal Flood Control Act in June

1948 (PL 80-853). The act authorized the U.S. Army Corps of Engineers to

develop a plan known as the Central and Southern Florida Project for Flood

Control and Other Purposes, which would address the water management needs

of the area. The plan contained three basic elements: 1) designation of the EAA,

2) construction of water conservation areas in the central Everglades and 3)

construction of an eastern levee. The purposes of the water conservation areas

were to protect the east coast and agricultural areas from flooding, recharge

regional aquifers and prevent salt water intrusion. In 1949, the state legislature

created the Central and Southern Florida Flood Control District (FCD) to act as

local sponsors for the federal project. The FCD was renamed in 1977 as the South

Florida Water Management District, at which time an additional objective,

enhancing environmental resources, was added to the above mentioned

purposes.

Construction of the physical structures of the project began in the early

1950s and continues to be modified to date. Three water conservation areas

(WCA) were surrounded by levees (Figure 2). Water conservation area 1, was

also given designation as the Loxahatchee National Wildlife Refuge in 1951. (In

1984, the area was renamed the Arthur R. Marshall National Wildlife Refuge in

honor of an eminent ecologist). Water Conservation Areas 2 and 3 were divided

into subunits A and B, primarily to decrease infiltration losses in the southeastern

portions of these areas. By 1962, the conservation areas were closed in and







functionally intact. Canal construction to date has resulted in approximately

1400 miles of canals.

Operational criteria for water management in the southern Everglades

revolves around the stated regulation schedules for the water conservation areas.

The schedules are target stages which vary over the year, which tend to revolve

around two objectives: 1) minimizing flood risk during the hurricane season

(June-October) and 2) maximizing storage during the dry season (November-

May). When levels are below regulation schedule, water outflow is minimized to

allow stages to increase to the regulation level. When the schedule is exceeded,

water is released to lower levels. Modifications to these schedules have been

made during recent years. The schedule for WCA 3A has been modified to allow

zones around a certain stage value, within these zones water input and outflow

are moderated so that rapid movement of water is negated. The regulation

schedule of WCA 2 has been modified to allow periodic drydown (Worth 1987).

Currently, the schedule for the Marshall NWR (WCA1), is being evaluated for

changes that would improve wildlife habitat.

The water conservation areas are not only spatially central, but

functionally central to water management in the Everglades. These areas are

designed to be used for many purposes, primarily flood control and water

supply. These areas act as surge tanks in receiving water during flood periods.
Runoff from agricultural areas to the north is placed in these areas. Water in the

WCA's is also kept from flowing into areas to the east, in order to lessen flood

impacts. During dry periods, water is also stored in order to meet demands

along the coast and to the south, especially Everglades Park.







Water Deliveries to Everglades National Park

Estimates of pre-drainage water flow into the area now in Everglades park

are tenuous due to at least two reasons. No measurements of flow were made
prior to 1940 and by 1940 many upstream canals were in place and may have

siphoned upstream waters to the coast. The Miami Canal was cut through the

ridge as of 1917 (Blake 1980) thereby removing water from the area immediately

north and east of the park. Historic (pre-drainage) average annual flows to the

area of the park were calculated to be 2 to 2.5 million acre feet. (Parker 1984). The

U.S. Army Corps of Engineers (1968) calculated a smaller mean value,

approximately 1.25 million acre feet. These flows were estimated to be the

amount of overland flow into the southern Everglades. Smith et al. (1989) using

a correlation between freshwater flow and the annual band width of a coral in

Florida Bay, estimated that during the period 1881-1939 annual flow averaged

1.15 million acre feet (1.4 billion cubic meters), whereas flow during 1940-1986

was estimated to be 0.47 million acre feet (0.5 billion cubic meters). Dynamic
flow models (Walters et al. 1992) driven by actual rainfall during the period 1960-

1987, predict flow to have varied between 0.5 and 2.5 million acre feet (0.62 and

3.1 billion cubic meters ), depending upon rainfall.

Overland flow has entered the area now in northern Everglades Park

(primarily Shark Slough) through man-made structures since about 1928 when a

series of round and square culverts were placed beneath the roadbed of Tamiami

Trail (US Highway 41). Most of these culverts are still in place and deliver water

to northeast Shark River slough. As part of the plan to enclose southern WCA

3A, a levee (named L-29) was constructed on the border between WCA 3A and

ENP. This effectively altered the distribution of flow through the western half of

the historic Shark Slough. Four sets of gates (designated S12A through S12D)

were placed in Levee-29 to allow water movement between the conservation area








and the park. Each of the four gates is comprised of six 25 foot wide vertical lift
gates. Each set of gates is designed for a maximum flow of 8000 cfs (226 m3/sec),

with a maximum headwater stage of 12.4 ft. and maximum tailwater stage of 11.9
ft. (U.S. Army COE 1968). The L-29 borrow canal provides the headwater to the
gates. Other structures that were constructed for various reasons to direct flow

in the Shark Slough, but no longer used, include the L-67 extension canal, S-12-E,
S-12-F and S-14 (Wagner and Rosendahl 1982). The alignment of park boundary

also bisects the other main drainage basin (Taylor Slough) from its headwater.

The physical structures that deliver water at the boundary into Taylor Slough
include a pump station (S-175) that delivers water out of canal L-31 W.
There have been at least eight different time periods each with varying

hydrologic regimes under which water has flowed into the Shark Slough area of

the park. Prior to initiation of construction of L-29 and the S-12 structures, water
flow into the park was unregulated in the sense that water across the boundary

was dependent upon hydraulic gradients within the upstream marshes and only

restricted by the capacity of the culverts. Starting in 1961, overland flow was

entirely cut off to Shark Slough while construction was underway, marking the

second flow regime. From December of 1963 through March of 1965, water was

moved from WCA 3A only after regulation schedule was met, that is, the park

only received excess water after upstream storage were met (Wagner and
Rosendahl 1985). During 1965 and 1966, three zones within WCA 3A were used

to deliver a monthly amount of water. From the period of March 1966 through

September 1970, the stage in Lake Okeechobee was used to determine water
deliveries to the park, with totals scaled from no delivery if the stage was below
12.5 ft, 150 cfs if the stage was above 12.5 and below 13.5, and 1000 cfs if the stage

was greater than 13.5 ft. (Wagner and Rosendahl 1985).







During the 1960s the park experienced low rainfall years and was

concerned about the quantity of water it received in context of increasing urban

demands. Two studies defined the water needs of the park using existing flow

and stage data. Dunn (1960) analyzed data for the period 1947-1952 and found
that the median annual flow into the Shark Slough area was 273,000 acre feet

(3.36 x 108 m3), a value that he recommended be adopted as the minimum flow

requirement. Hartwell et al. (1964) developed stage-duration curves for station

P-33 in the park and stage-discharge correlation between P-33 and flow into the

park, and used these relationships to determine an annual discharge requirement

of 243,000 ac. ft. (2.97 x 108 m3). A crude average of these two figures was

incorporated into a congressional act in 1970 (PL 91-282) which guaranteed the

park an annual minimum delivery of 315,000 ac. ft. (3.85 x 108 m3) or 16% of the

water in the system. These annual deliveries were to be partitioned into the three

flow sections into the park. Shark Slough was to receive a minimum of 260,000

ac. ft (3.18 x 108 m3) annually, 37,000 ac. ft.( 0.45 x 108 m3) were to be delivered

into Taylor Slough, and 18,000 ac. ft. (0.22 x 108 m3) into the eastern panhandle

area of the park (Wagner and Rosendahl 1985). This law established the legal

right of the park to a minimum amount of water and to share adversity

associated with periods of drought. During the 1970s the minimum delivery
concept was altered from a minimum threshold to one of a static portion of water

allocated to the park each year. The annual flows through the S-12s were

regulated tightly, and during the years 1970 through 1982, met minimum

delivery requirements, but tended to release water over the schedule during

summer months.

In 1983, following a wet year and changes in the operating criteria for
backpumping into Lake Okeechobee, the park service requested alterations to the
"minimum" delivery schedule. Fearing too much water would come into the







park, a number of alterations to the structures of the system were requested,
along with changes to methods of delivery. In order to remove water from

eastern and southern WCA 3A through pathways other than into the park,

culverts were placed in Levee 28 in order to allow water to flow into the Big

Cypress. Other outlets for WCA 3A were requested but not implemented. Water
was to be diverted into WCA 3B. In response to these requests, Congress passed
a law (PL 98-181) that allowed for experimental water deliveries to the park. For

the next two years (1983-1985), the S-12 structures were left entirely open, so that
water would enter the park as a function of hydraulic gradients between WCA

3A and the park. Although the flow-through plan may have achieved objectives

of restoring the natural timing of flow, the situation of leaving the gates open did

not bode well with water managers faced with the necessity of storing as much
water as possible in WCA 3A for meeting other needs on the coast.

The latest act in the unfolding play of water deliveries to the park was the

Rainfall plan developed by Tom MacVicar of the SFWMD and staff of the COE.

They analyzed rainfall-runoff data from the period 1940 through 1952, and

developed a statistical model which predicted weekly flow based upon net

rainfall (rainfall minus evapotranspiration) over WCA 3A from the previous ten

week period and the previous weeks' discharge. The model achieved two
objectives; the timing and quantities of deliveries were linked to upstream

weather conditions, and the flow would be re-distributed spatially as it was prior

to the construction of the water conservation area. The regulation schedule was
also modified to allow for variation in water level conditions within WCA 3A in

order to avoid the rapid releases of water into the park (MacVicar and VanLent

1984, MacVicar 1985, Neidrauer and Cooper 1988). In essence, the rainfall plan
limits the source basin of the park to WCA 3A by directly timing delivery to

rainfall over the area. Buried in this delivery plan is the unknown contribution







of other areas in the Everglades (and Lake Okeechobee) to water in WCA 3A and
eventually to the park. This is manifest in the supplemental deliveries, by which

more water than the rainfall formula predicts is delivered. The supplemental

deliveries are linked to a modified regulation schedule. The key elements of the

rainfall plan are 1) to link timing and quantity of baseline deliveries to upstream
rainfall, 2) to increase quantity of flow during periods of high water, 3) to

decrease quantity of insured deliveries during dry periods, and 4) to supplement
the baseline quantity of water depending upon a wider range of water depths

(regulation schedule) within WCA 3A.

Major determinants to the constantly changing methods and policies of

water delivery to the park have been the observed degradation of biological

resources in the southern Everglades and Everglades National Park. Dry years
and accompanying fires during 1962 and 1971 prompted the appeal for more
guaranteed water. Increased mortality of alligator young (Kushlan and Kushlan

1980) was attributed to rapid water level rises associated with regulatory releases
during the early dry season. Browder (1985) developed relationships between

flow into the estuary of Florida Bay and shrimp production. The most attention

has been drawn to a dramatic decline in the number of wading birds; nesting

success of wading birds has decreased by 95% of levels in 1930s (Robertson and

Kushlan 1984). Reasons for the declines have been intimately linked with
decreases in flow through the park (Ogden 1978, Ogden 1987, Powell et al. 1989,

Walters et al. 1992). Other authors believe that too much water in the Everglades
has contributed to the population decreases (Kushlan 1987) and that the park
should receive less water.

The preceding review of the Everglades ecosystem has followed two

separate paths: one recounts the natural history and the other the human






41








history. These histories intertwine, and are linked by the ways in which humans
perceive, understand and react to nature, the subject of the next section.


Managing Ecological Systems

One interpretation of the history of water management in the Everglades

is that it appears to follow a pattern of crisis and reconfiguration (Light et al, In
press). The crises arise from dramatically unexpected system behavior, such as
floods, droughts and fires. Crises in the past have appeared suddenly, as

surprises, and the subsequent responses have dramatically changed the way the
system has been managed (Table 1). The central reconfiguration occurred

following the flood of 1947, when the Central and Southern Florida Project was

spawned. Since then, other crises have occurred with subsequent changes in the

policy and practice of delivering water to the park. The reaction by humans to

these surprises takes the form of policies and management actions. The
responses are shaped by perceptions and interpretations of how nature operates.
At least two concepts are involved in the interpretation of nature that

create the basis of policy and management formulations. The first concept relates

to various views of system stability. The second concept deals with the assumed

or perceived uncertainties associated with either system understanding or

impacts of management actions. At least three views of system stability have

been abstracted: equilibrium, dynamic and evolutionary (Holling 1987). An
equilibrium view is defined as one dominated by the assumption that key

response variables always return to a point or set of points. The dynamic

perspective recognizes that system variation occurs within and between a range
of stability regions so that system behavior appears at times constant, other times

continuously changing, and at times jumping abruptly into another stability

regime. In the evolutionary view, the stability landscape can change, implying







fundamental structural and organizational changes in the system. Dealing with

the inherent and fundamental uncertainty associated with shifts within and

among these stability domains is at the heart of adaptive management (Holling

1978, Walters 1986). During the last decade both the policy and management
philosophy in the Everglades crossed thresholds involving both changes in views

of stability domains, and in strategies of management.

Policy in the Everglades is still largely rooted in the equilibrium-centered

perspective, although the dynamic view has been recognized and partially

incorporated into management schemes in the mid 1980s. Water movement is

largely determined by regulation schedules in the different components of the

system (Lake Okeechobee and the water conservation areas). These regulation

stages reflect an equilibrium view of water management, that is, it always returns

to an ideal stage within a retention pool. The modifications to WCA 3A schedule

associated with the rainfall delivery plan, however, indicate a shift to a dynamic

viewpoint, allowing variability in the managed system.

Water management in the southern Everglades during the last decade has

developed more attributes of adaptive environmental management (Holling

1978, Walters 1986). Within the last decade, programs such as the iterative
testing plan (Light et al. 1989) have been applied using the concept that water

management necessarily has some experimental attributes. This has even been

codified, by the adoption of PL 98-181 which allows for experimental deliveries

to the park. The rainfall plan can be classified as a passive adaptive technique

(Walters 1986), whereby historical data are used to construct a model that guides

management plans. Two problems with this technique are that 1) environmental

and management effects are confounded, and 2) little opportunity exists for

improving the model or testing new models (Walters 1986, Walters and Holling

1990).







Summary

In this chapter, the key pieces of the interplay between the natural and

human dimensions of the Everglades are described. The undisturbed Everglades

ecosystem can be characterized as an oligotrophic, sub-tropical wetland system
with high temporal variability in rainfall input. The landscape is flat, yet

supports a complex spatial mosaic of marsh and woody vegetation plant
associations. Humans have interacted with the system for as long as it has been a

wetland. Dramatic changes have occurred during this century, within which

time about half of the land area has been converted to agriculture and urban

development. Over the past 50 years, Everglades National Park has been
established, as has one of the largest water management infrastructures in the
world. Water management and deliveries to Everglades Park have undergone
dramatic, non-linear changes resulting from recurrent crises and surprises. The

foundations for policy and management development during these periods of

reconfiguration are intimately linked to and dependent upon our understanding

of ecosystem dynamics. Understanding ecosystem dynamics and the different

paradigms regarding system organization is the point at which the first chapter

ends and the second chapter begins.












CHAPTER 2.
POSING THE QUESTIONS


Using all the weapons of our logical, mathematical and technical armoury we try to prove
that our anticipations were false--in order to put forward, in their stead, new unjustified
and unjustifiable anticipations, new 'rash and premature prejudices' as Bacon derisively
called them

-K.R. Popper



As indicated in the preceding sections, the problem of water deliveries to

Everglades Park has many dimensions, including how ecosystems vary over

different time spans, and the subsequent reactions and adaptations of people to

these fluctuations in the system. Resource policy and management is

fundamentally related to how humans perceive and attempt to comprehend the

vagaries of nature. Even though institutional and human dynamics of the system

are important, they are fundamentally rooted in basic paradigms about how

ecosystems function. The study of ecosystems can be particularly difficult

because of the variety of components, processes and variables. Attempting to

incorporate all variables makes the problem overwhelmingly complex (Gallopin

1991). Simplifying assumptions allow for these studies to become tractable. The

remainder of this chapter will outline and contrast three existing approaches to

simplify understanding of ecosystem dynamics and a description of a new

emerging paradigm. Hopefully, this theoretical background will lay the

framework from which hypotheses and objectives of this work are derived in the

concluding sections of this chapter.







Views of Ecosystem Structure and Function

Understanding and interpreting ecosystem structure and function are

based upon underlying methodological assumptions and paradigms held by the

observer. At least three such concepts are recognized, while a new one is

emerging to account for the paradoxes that emerge from applying the first three.

These viewpoints can be characterized and contrasted by two components of the

paradigms: 1) the factors or variables that are important in the system and, more

fundamentally, 2) the manner in which these variables interact. The first

assumption is that variables interact in such a way that the strength or

significance of the interaction can be tested against a null model that is random.

The second paradigm is based upon a view that the world is structured in a

hierarchical manner, with distinct levels of causation defined by the observer.

The third view of ecosystem science is rooted in mathematical modeling, and

concentrates on system dynamics across a limited range of scales. The newest

belief, emerging because of limitations in the other views, is cross-scalar in scope

and implies a world of lumps and discontinuities in which a small number of key

processes determine function, each over its own range of scales. None of these

views are wrong; indeed, all represent partial truths and continue to thrive

because they are useful. Following a brief characterization of each existing

assumption and their limitations, the emergence of a new view will be presented.

Ecosystem science as practiced by the 'Stochastics' is characterized by

multi-variate statistical approaches. The implicit assumption is that the variables

are operating within similar domains in space and time and therefore have

correspondingly similar ranges of variation. Explicit assumptions include that

the variables are essentially derived from continuous distributions and therefore

have certain properties that can be estimated from sampling. In the extreme,

variation in response variables is partitioned to either the variation of other







variables or to a random error term. In all cases, the null hypothesis is a random

one; relationships can only be inferred by rejection of the null. Examples of work

in the Everglades ecosystem of this type include analyses by Smith et al. (1989)

and Browder (1985) in correlating freshwater flow in the system with biotic

responses in Florida Bay. Indeed, the current rainfall formula developed by

MacVicar (1985) is a statistical approach, whereby flow through the southern

Everglades is regressed against rainfall and stage. This approach is powerful,

because the tools are readily available, and only a statistically significant number

of samples are necessary for application.

The hierarchical view of ecosystems, on the other hand, while a powerful

concept, is still struggling for widespread application after being introduced at

least 50 years ago. The basic framework offered by hierarchists is one of

partitioning variables and interactions into distinct levels or "holons" (Allen and

Starr 1982). Variables that operate at similar scale ranges occupy the same level

within a hierarchy and interact more than with variables between levels. One

feature of hierarchies is dubbed as asymmetry, where the variables at slow levels

constrain the variables at fast levels. The underlying assumptions of variability

are similar to the random approach, in that the variables are assumed to have

continuous distributions and that these distributions are predictable. Another

common belief of hierarchists is that hierarchies are relatively stable, static and

operate near equilibrium. The most current view of hierarchists is that the world

should be partitioned into the appropriate hierarchical level such as landscape,

ecosystem, population, organism for study, analysis and understanding (Allen
and Hoekstra 1992). To my knowledge, no applications of this theory has been

applied to studies of the Everglades.

A tremendously rich understanding of ecosystem structure and function

has been achieved by modellers who apply the third approach to explanation.







Although inductive, the approach can improve understanding by testing

dynamic interactions among variables. Assumptions regarding variables and

interactions can be clearly stated by mathematical formulae translated into
computer code. The utility or power of modeling also carries related costs.

Empirical rules such as "parsimony in the selection of variables" (Clark et al.

1979), the "power of two" (Walters 1986) or the optimum trade-off between

articulation and predictability (Costanza and Sklar 1985) all attest to constraints

on modeling.
The limitations imposed on each of these views has to do with issues of

scaling. All of these approaches to explanation treat both variables and

interactions as scale invariant. Scale invariant means that the behavior of

variables and the rules or properties of interaction do not change within the scale

limits imposed by the observer. The power of these approaches comes from the

knowledge of a rule set, and how far (over what range of scales) the rules apply.

Limitations related to scale are a result of underlying assumptions (stochastics),

theoretical frameworks (hierarchists) and of practical experience modellerss). For

variables to be analyzed, compared or contrasted using the available statistical

approaches, they must change within a similar manner. If there are dramatic

differences in the space or time dynamics of variables, then statistical methods

either give erroneous conclusions or, flat out, don't work. An example of this is

the rainfall formulation (MacVicar 1985) mentioned above, where the regression

analysis indicated that no statistical relationship existed between rainfall and

flow!

Similar problems of the mismatch between variable spatial or temporal

domain arise in hierarchical theories and in the application of modeling

techniques. The hierarchists (Allen and Starr 1982, O'Neill et al. 1986) recognize

that "slower" variables constrain "intermediate" variables while "faster" variables







are essentially meaningless, or noise. Little progress has been made with linking
these variables together other than in a conceptual or qualitative sense.

Modellers have come to essentially similar conclusions, as expressed in the

practical "Rule of Two". The empirical rule states that the best models never

extend more than two orders of magnitude in either space or time. The basic

approach of scaling in modeling, is to treat "slower" variables as constants, and to

treat "faster" variables as random or stochastic events. Existing space-time

models of hydrodynamics in the Everglades fall within this guideline and will be

described later in the modeling section. The result of limitations imposed by

these various approaches is the breakdown of understanding, as evidenced by

inherently unpredictable system behavior (Holling 1986). These limitations and

inevitable surprises helped prompt the development of a theory that attempts to

embrace the cross-scale dynamics of ecosystems.

This emerging cross-scale theory has roots in both the hierarchical and

modellers perspective and can be traced to a review and synthesis of the

dynamics of a number of ecosystems. Holling (1986) compared the dynamics of

23 ecosystems, and concluded that the essential behavior of the system could be

traced to three or four sets of variables, each of which operated at distinctly

different rates. The sample ecosystems were categorized into one of four classes

of systems: forest insects, forest fires, grazing in savannas and aquatic harvesting.

Models of the reviewed ecosystems all generated complex behavior in space and

time that qualitatively correlated with observations of the actual systems. The

essential dynamics of the systems could be attributed to a small number of

keystone variables. The speeds of each keystone variable differed from each

other by as much as an order of magnitude, so that the time constants were

discontinuous in distribution (the hierarchists would consider each keystone

variable as a part of different levels or holons) or as a small number of nested







cycles. The results of the review led to the hypothesis that ecosystem dynamics

are organized around the operation of a few key variables, each operating at

distinct speeds.
The next critical step in development of theory was the proposed

hypothesis that the system should be structured in such a way that the keystone
variables entrain other variables. The entrainment should occur in both spatial
and temporal dimensions creating structural features that exhibit distinct gaps
and lumps and temporal processes that exhibit distinct periodicities. An overt

manifestation of a lumpy, discontinuous world should be expressed by attributes

of the animals that live in these systems. This hypothesis was challenged by a
series of tests using data from three biomes (Holling 1992). The tests using adult
body mass of birds and mammals from the boreal forests, prairies and pelagic

ecosystems, indicated the presence of discrete gaps that defined groups (Holling

op cit.). Alternative hypotheses using developmental, historic or trophic

explanations for the groupings were all invalidated, leaving only the strong

inference that ecosystems (abiotic and biotic components) were similarly

organized (Holling op cit.) into discreet lumps.


Hypotheses

Two hypotheses are posed in this work and arise from two of the

approaches mentioned above. Both are aimed at improving understanding of the

structure and function of the Everglades ecosystem that specifically relates to

system dynamics and flow to Everglades Park. The first hypothesis derives from

the approaches that understanding complex system behavior can be induced

from modeling non-linear interactions among continuous keystone variables
within a constrained range of scales. The second hypothesis is developed from







the cross-scale arguments, and attempts to invalidate the lumpy, discontinuous

view of the world.


Water Deliveries to Everglades National Park The First Hypothesis Set

A dynamic water budget approach is a powerful conceptual tool for

evaluation of the factors influencing deliveries to Everglades Park. The amount

of water in the southern Everglades at any time is a net result of changes between

inputs (rainfall and inflow), and outputs (outflow, evapotranspiration, and

groundwater infiltration). Theoretically, the rates of flow and evapotranspiration

are related to vegetation type and structure. Since Everglades National Park is

situated at the downstream end of the historic ecosystem, it is dependent upon

water from upstream sources. The water that enters the park comes from two

sources: rainfall over the park and overland flow from the north. Assuming that

local rainfall contributions to the park water budget are relatively unchanged, the

first hypothesis deals with the contribution of upstream sources to the water

requirements of the park.


Hypothesis: The effective drainage basin that supplied water to

Everglades Park was the entire Kissimmee, Lake Okeechobee and

historical Everglades ecosystem. Implicit in this hypothesis is that

overland flow is a dominant pathway of water movement, and that

hydrologic continuity throughout the historic system is crucial to

maintaining water supply to the park.


Corollary: In an area as flat as the Everglades, the vegetation and

hydrology are intimately coupled. The structure and type of vegetation

affect both evapotranspiration rates and resistance to overland flow.







Vegetation type and structure are in turn, affected by water depths and

hydroperiod. The coupling between hydrology and vegetation

determines the relationship between the amount of water that flows

through the system and the amount that evapotranspires to the

atmosphere.


Null: The effective drainage basin was a much smaller geographic area.

Overland flow into the park system originates from an effectively smaller

drainage basin. This is because evapotranspirative losses are high relative

to rainfall, therefore water would evaporate before moving very far

downstream. Other users in the system can remove water without major

disruption of the flow that entered the park historically.


Cross-Scale Patterns In The Everglades Ecosystem The Second Hypothesis Set

The processes that influence flow to the southern Everglades cover a wide

range of space and time scales. Rainfall results from atmospheric processes,

ranging from meso-scale (Florida peninsula) to global dynamics. Vegetation

structure can be identified at scale ranges from parts of individual plants (stems,

leaves) to the organization of plant associations in the landscape. The combined

processes of evaporation and transpiration occur from the level of leaf stomata to

entire ecosystems. Other processes have similarly wide ranges of variation in

space and time.
In order to attempt to invalidate the first hypothesis, the methodology

requires that the world of the Everglades be "squeezed" into a framework of fixed

spatial and temporal domains. The second hypothesis is based upon the

emerging theory which suggests that across scale ranges, ecosystems are







organized in such a way so that dumps and gaps appear in structural features

while a small number of cycles and harmonics occur in temporal features.


Hypothesis: The Everglades ecosystem is structured by a small number of

processes, of which hydrology and vegetation are one set of keystone

variables. Over a range of scales, patterns produced by vegetative and

hydrologic processes should have a few characteristic domains in space or

time that are separated by discontinuities. Within the time domain, a few

dominant frequencies or cycles should emerge in the hydrologic processes

of rainfall, water level, flow and evapotranspiration. Within the spatial

domain, a few groupings of object size (such as vegetation patches), or

texture will emerge that correspond to levels in a spatial hierarchical

structure. Other ecosystem level processes, such as topography and fire,

will exhibit similar patterns.


Null: Over a range of scales, the temporal patterns of hydrology and

spatial patterns of vegetation in the Everglades will exhibit structures that

correspond to underlying continuous distributions No dominant or

nested cycles will appear in the time patterns of hydrologic processes such

as rainfall, water level, flow or evapotranspiration. No breaks or

discontinuities in the spatial patterns will be found, and patterns will be

self-similar over a wide range of scales.


Objectives
The aim of this work is to develop new understanding of ecosystem

dynamics by testing hypotheses regarding water and vegetation dynamics that

are rooted in two different viewpoints. There are two main objectives: 1) use







"scale-bound" modeling techniques to help understand factors influencing water

deliveries to Everglades Park, and 2) apply cross-scale analyses to Everglades

data sets to test for breaks or discontinuities in patterns.
The first objective will involve the construction of a spatially and

temporally explicit model to capture the dynamics of the system in order to test

the first hypothesis. The model will couple dynamics of hydrology and
vegetation within a spatial domain of two orders of magnitude and a temporal

domain of almost three orders of magnitude. The model will be used to attempt

invalidation of the proposal that the entire basin contributed water to the

southern Everglades. This objective is the focus of Chapter 3.

The second objective will be to develop, test and apply a variety of cross-

scale methods to identify patterns in keystone variables in the Everglades
ecosystem. Since the theory of cross-scale interactions is just emerging, a great
deal of this work has been devoted to the development of new methods and

methodology. Fortunately, this work was able to reap the benefit of data sets of

many variables that have been collected over the years on different portions of

the Everglades system. The methodology, and results of the cross-scale analyses

are the subject of Chapter 4.













CHAPTER 3.
MODELING THE "RIVER OF GRASS"


When your only tool is a hammer, the answer to every problem is a nail.

-R. Yorque


Almost 50 years ago, Marjory Stoneman Douglas created a dramatic image

when she described the Everglades as a "River of Grass" (Douglas 1947).

Technically, neither of these terms are appropriate. The system is hardly a river

because there is no defined water course and water flows very slowly, only about

60 kilometers a year. The "grass" in the title refers to sawgrass, which is properly

classified as a sedge. However, the metaphor is still appropriate because it can

be interpreted to depict the coupling of the vegetation and hydrology in this

ecosystem that at one time, was a united ecosystem.

In this chapter, the test of the first hypothesis that the entire Everglades

system provided water to Everglades National Park and test of the corollary

hypothesis regarding the coupling of vegetation and hydrology, are presented.

These hypotheses were tested with a model that incorporates coupled vegetation

and hydrologic dynamics over time within an explicit spatial array. Hydrologic

models of the Everglades system have been used to evaluate management

options within the current system configuration (MacVicar 1985) or to create

views of the system prior to human intervention (Walters et al. 1992, Perkins and

MacVicar In prep.). Such models provide a robust methodology from which the

contributions of upstream areas to flow into the southern Everglades can be








evaluated. Key uncertainties in these models include information about

overland flow resistance and evapotranspiration and infiltration to groundwater.
In a system as flat as the Everglades, vegetation influences surficial

hydropatterns by mediating resistance to flow and controlling
evapotranspiration. Hydrologic regimes also influence the vegetation pattern

(Davis 1943, Craighead 1971, Gunderson 1989). The corollary hypothesis posits
that the interactions between hydropatterns and vegetation patterns are coupled

and create feedback loops. None of the previously developed hydrologic models
of the Everglades incorporate complete feedbacks between hydrology and

vegetation. The SFWMD models (MacVicar 1985, Perkins and MacVicar, In

prep.) vary flow and evapotranspiration by land cover types, but the land cover

types do not change as a function of hydrology. The model of Walters et al.

(1992) changes vegetation types as a function of hydrology and other factors, but
has spatially fixed flow and evapotranspiration rates. The approach in this work,
therefore, is to couple vegetation and hydrologic feedback dynamics in the

framework of existing models to test hypothesis about upstream area

contributions to the park.
This chapter is divided into four sections: background, model

development, results and summary. A fair portion of this chapter is devoted to
improving understanding of the interactions among evapotranspiration, flow

and vegetation since they are key uncertainties in the model. The background
section will develop a theoretical base for understanding these processes and the
results of studies that compares evapotranspiration rates among vegetation

communities of the Everglades in order provide a foundation for the linkages in

the model. Following the background section, the model is described including

components and their interactions. The section following the model description

presents the results of testing the upstream area hypothesis and the corollary







hypothesis regarding the coupling of vegetation and hydrologic processes. This

chapter is concluded with a summary of modeling the "River of Grass" and

implications of the results to policy and management.


Background

Evapotranspiration in Wetlands

Evapotranspiration is the combined processes of water flux into the

atmosphere by evaporation from water or soil surface and transpiration from

vegetation. In an area such as the Everglades, both processes are in effect, as

there are areas of relatively sparse vegetation (open water marshes), grassy

wetlands and forested wetland communities. In addition to the interaction

between evapotranspiration and vegetation, other physical variables influence

the rate of water flux. All of these variables, and measures of evapotranspiration,

appear to vary across scales of interest. Previous studies on evapotranspiration

in the Everglades have been made at different scales, and will be reviewed below

in relation to spatial and temporal groupings, but first a review of theoretical

background.

Evapotranspiration is a component of the energy budget. A general

formulation for steady state system at a specific location is shown in equation (1),

modified from Brutsaert (1984), and Viessman et al. (1989). The amount of net

solar radiation (ambient minus reflected) determines the amount of energy

available for other processes. The energy can be used to increase the

temperatures (sensible heat) of both the atmosphere and the soil substrate. Some

of the energy is used in photosynthesis, and some may be moved by advection

(wind) to other areas. The other energy is used for the phase transition of water

from liquid to vapor. Since the evaporative process requires energy for the phase

transition, the energy is not measurable or latent. The latent heat of evaporation







times the rate of evaporation describes the evaporation term in the energy budget

equation.


(1) Rn = LeE+H+G+P+A


Rn = specific flux of incoming net radiation

Le = latent heat of evaporation
E = rate of evaporation
H = specific flux of sensible heat into the atmosphere

G = specific flux of sensible heat into the ground

P = rate of energy used in photosynthesis
A = rate of lateral advection of energy


Due to difficulties in measurement of the latent heat of evaporation and

the confounding effects of vegetation influences on the processes, a number of

techniques have been developed to measure evapotranspiration. The techniques

fall into two categories, those that derive from energy budget, with certain

simplifying assumptions and those that are empirical. Penman (1948) derived a

formula that includes a wind advectionn) term with assumptions of constant

Bowen ratio (ratio between sensible and latent heat) that allows for measurement

at one level. Other techniques derive from the Dalton formulation that estimates

a vapor gradient between the surface and the atmosphere. Empirical methods

include formulations by Blaney-Criddle (1950) who related evapotranspiration

with average temperature; Holdridge (1967) who related vegetation form to

temperature and potential evaporation and simple techniques such as pan

evaporation or lysimetry.








Evapotranspiration is influenced by a mixture of processes that occur at

different scales in space and time. Solar insolation at a spot on the earth

fluctuates on daily, annual and multiple year cycles. The solar radiation is also

influenced by fast (on the scale of minutes) fluctuations in processes, such as

cloud cover. Vegetation both directly and indirectly influences the

evapotranspirative processes. The fastest controls vegetation occur at the level

of the stomata, where water flux is linked to photosynthesis (Jarvis and

McNaughton 1986). Individual plant species' genetic composition and

adaptations influence the size and density of stoma, leaf orientation, responses

to various changes in insolation, wind, humidity and temperature (Jarvis and

McNaughton 1986). The composite architecture of the canopy in wetlands can

influence the reflectance of both short and long wave radiation, with implications

to net energy and water use (Odum 1984, McClanahan and Odum 1991). At the

landscape or regional level, the vegetation cover type influences the reflectance

or albedo.

The preceding paragraphs gave a brief review of evapotranspiration

theory to provide a basis for understanding the various approaches and

techniques for measurement of this complex process. The following section will

present the results of previous investigations and present published

measurements of rates of evapotranspiration at different spatial and temporal

scales.


Measurements of Evapotranspiration in southern Florida

Measures of evapotranspiration in southern Florida have also been made

at different scales ranging from local up to the entire peninsula. For convenience,

these can be grouped for discussion into broad-scale measures that include the

region and basins, medium scale measures (less than 10 m on a side) typified by







lysimeters, and evaporation pans and small scale measures that examine losses
from leaf surfaces.
For the region, Dohrenwend (1977) used an empirical formula developed

by Holdridge (1967) that related evaporation and mean annual biotemperature to
calculate an annual evapotranspiration of about 1000 mm. Input-output analyses

of basins in and around the Everglades calculate a range of values of
evapotranspiration that are similar. Allen et al. (1982) estimated annual
evapotranspirative losses from 890 to 1040 mm from Taylor Creek basin north of

Lake Okeechobee. Leach et al. (1971) estimated evapotranspiration from the
Water Conservation Areas at 965 mm/yr. Shih et al. (1983) estimated water
losses from the Everglades Agricultural area at 1018 mm/yr using a water
budget approach and also compared a number of other techniques and found

annual means ranging from 1018 to 1035 mm.

Most of the smaller scale investigations involve the use of lysimeters

(tanks with planted vegetation), evaporation pans or shallow wells to derive

monthly and annual estimates of evapotranspiration. Clayton et al. (1949)

planted sawgrass in lysimeters and found monthly ranges of 78 to 208 mm and

mean annual losses of 1735 mm. Parker et al. (1955) estimated

evapotranspiration from pan evaporation and reported annual values from 1016
to 1143 mm. The Army Corps of Engineers (1968) reported monthly values from

63 to 135 mm. Shih (1981) compared average monthly and annual

evapotranspiration among sugarcane, alfalfa, and bahiagrass plants planted in
lysimeters with data from Clayton (1949). Shih (1981) found a range of monthly

values from 35 mm to 212 mm for sugarcane. Other crops were within these

monthly averages, and had lower annual totals. Carter et al. (1973) derived

annual estimates of 1100 mm for the Big Cypress Swamp area, immediately west

of the Everglades.








A number of studies attempted to develop relationships between
evapotranspiration (from lysimeters) and climatological data. Most assume

constant Bowen ratio; that is, a constant proportion between the sensible heat

flux and the evapotranspiration. Stephens and Stewart (1963) found that best

approximation to lysimeter values of potential evapotranspiration were based
upon a fraction of ambient radiation. Shih (1981) found good correlation

between lysimeter losses and monthly temperature corrected for cloudy days, a

modification of the Blaney-Criddle technique. Stewart and Mills (1967) and Shih

(1983) both measured decreasing evapotranspiration rates with declining,

subsurface water levels.

Perhaps the fewest studies have been done at the level of the individual or

on a daily basis. Brown (1981) studied pondcypress in southern Florida, and

measured average daily losses at 1.3 mm/day. Dolan et al. (1984) working to the

north measured daily marsh evapotranspiration from 0.5 to 10 mm/day. One
small scale study (Alexander et al. 1976) compared evapotranspiration rates of

potted seedlings of sawgrass and Melaleuca..

Different variables operating at different space and time scales have been

shown to influence evapotranspiration. Some variables, such as radiation are

spatially global; that is, they remain the important input to the process across all

scales. Other variables, such as leaf stomata, dictate fine scale (leaf level) control,

but cease to be important at the regional scale (Jarvis and McNaughton 1986).

More work has been done on scaling measures of evapotranspiration over the

time domain by increasing the extent or window (day to month to year).

Authors working in southern Florida (Stephens and Stewart 1963, Shih 1981, and

Carter et al. 1971) all recognized a decrease in variation as the time unit is
increased.








In reviewing the available literature, few measurements of evaporation or

transpiration from the native plant communities in the Everglades have been

published. In order to incorporate community level measures into the model, a

series of studies were done to develop measures of water loss by community

type. The studies were important for two reasons, 1) to establish that
evapotranspiration rates varied among the major vegetation communities, and 2)

to estimate the magnitude of any differences. If there was no difference among
the vegetation types, then evapotranspiration could be modeled by a variable

that only changed in time and not over space. If there was a difference in water

loss rates among the vegetation communities, then maps of vegetation

communities could be used to develop spatial patterns of evapotranspiration.
The next section of this chapter reports on the summary of two studies

using two different methods to calculate rates of evapotranspiration among the

dominant native plant communities. The first study uses transpiration rates
reported in Herndon and Gunderson (1989), and, with a few assumptions,

attempts to aggregate from the leaf level to the community. The second study

uses data reported by Gunderson and Stenberg (1990) on evapotranspiration

from two wet prairie sites. The measures reported in these two studies will be

summarized for use in the model.


Transpiration from Three Everglades Plant Communities
The rates of water flux from the leaves of dominant species in three

community types; sawgrass, tree island and marl prairie. (wet prairies had no

macrophytes) were measured. The measurements of leaf transpiration were

multiplied by the leaf area per vegetation type to yield community estimates.

Water flux rates (millimoles/cm2/sec) from the leaf surfaces were measured
using a steady state porometer (Li-Cor Model 1600). During the period from







December 1984 through February 1986, measures were made at 10 am, 12 pm

and 2 pm local time during one day every other month. At each sample period a

total of thirty measurements were made at random locations within the

community. The thirty measurements were combined to give an average flux for

each period. The rates of water loss were integrated over the day in order to

yield a water loss per day. To translate or scale these measures to a community

level, estimates of leaf surface area were made for each community type. All

leaves within a one meter square area were counted and leaf areas measured to

yield a leaf area/ m2.

Transpiration rates did not vary much among the vegetation types

sampled. Mean and range of water fluxes (transpiration) from the leaf surface

were determined from the field measures. Sawgrass transpiration ranged from a
low of 0.9 mmol-m2.sec-1 during December 1985 to a high of 3.2 mmol-m2.sec-1 in

July of 1985 (Table 3). Muhly grass, Muhlenbergia filipes, the co-dominant species

in the marl prairie, had similar rates, ranging from 1.6 to 2.4 mmolm2.sec-1 (Table

3). The swamp forest species showed similar transpiration, with values ranging

from 1.8 to 3.0 mmol.m2.sec-1 (Table 3). A one-way analysis of variance was

performed, and indicated no significant difference in daily transpiration among

vegetation type. A second analysis of covariance that removed the seasonal

trend in the data, also indicated no significant difference in transpiration rates

among the species monitored.

Daily water losses per community type were estimated by first calculating

daily transpiration and multiplying by leaf area per community type. Daily

transpiration was calculated by multiplying a six hour day length times the leaf

transpiration rates. A six hour day was thought to reflect an average period of

daily metabolic activity through out the year and probably underestimates daily

transpiration in the summer and overestimates wintertime values. Estimates of












oO O M 10 C) an in .o .- .N C' C M M C.o M
oocoo6ooo6ooooooocoooo



0 0 0 0 0 0 0 0 0 0 0 0 0
ddddddddddddddddddddddd









M O M 0 M ) 00 M CN 00 a0 M ma
cc CCcCcdddddddddddddddddddddddc


~< ~rNNN cNN oococo
c- . o 6 6 6 6 6 6 6 OO


(N- C\ N 00 00 (O ON ir) M 00 Nt N (O (O 4 4 N clO C) C4

--o






"t \D M M CD0 (O M m (N 0 Om

o 2



Z s













r-








z u
0) Up.. a:
0 41----------
" Z
)- L]m ^ b G ^ h ^' ^ ii ^ ' ^ Lr U i ^ tn Lr t ^
-" M o e ^ ^ o o o o cpg o o o o g oo o c oog d

>- S-( O i t C- ^i C - C 1 ; -


o E
u
0
0a


[-\







leaf area per m2 of ground area were multiplied times the daily transpiration to

yield a community water loss.

Seasonal trends in community transpiration among the three types were

most evident in the swamp forest (Figure 3), and less evident in the marl prairie.
Transpiration was measured from all three vegetation types throughout the year,

indicating year round metabolic activity and no dormant period.

Analyses indicated a significant difference in water loss among the

vegetation types, due to differences in leaf area. A one-way analysis of variance
and one-way analysis of covariance (removing seasonal trend) both indicated a
significant difference in daily water loss among the three vegetation types.

Water loss rates from the marl prairie vegetation was the lowest, with an annual

mean of 0.016 cm/day. Average loss from the sawgrass marsh was 0.16 cm/day

and from the swamp forest 0.45 cm/day. Posteriori contrasts indicated that these

three types were significantly different.

In summary, estimates of water loss using the technique of scaling from

small scale transpiration to community values are dependent upon the

vegetation structure more than the flux rates. The water flux from the leaf

surfaces tend to vary seasonally, and do not exhibit differences among vegetation

type. Significant differences in water loss do appear to exist among community

types and appear to be related to the amount of leaf area present.


Measurements of community evapotranspiration

Other estimates of community evapotranspiration were made using

recorded tracings of water levels in shallow wells at two sites in Everglades

National Park. One well is designated P33 and is surrounded by wet prairie on

peat vegetation type. The other well is designated P37 and is situated in a wet

prairie on marl.

























1

0.9

0.8

- 0.7 -

0.6 -
v^-
S0.5 -

S0.4

S0.3

0.2
0.1
0 I I I

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
MONTH













Figure 3. Mean daily transpiration rates from sawgrass, tree island and marl
prairie plant communities.


B Sawgrass

] Tree Island

SMarl Prairie







Evapotranspiration was estimated by comparing nighttime water losses
with daytime water losses, similar to the technique reported by Dolan et al,

(1984). The method is based upon the assumption that the only difference

between daytime and nighttime recession rates is due to evapotranspiration. The

technique is not useful on days with rain.

Rates of community evapotranspiration were different between the wet

prairie on peat and marl prairie sites. Mean daily water loss rates from the marl

prairie, calculated for each month of available data, ranged from a low of 0.10

cm/day during December to a high of 0.28 cm/day during June (Figure 4). This

translates to a mean annual total water loss of about 77 cm at the marl prairie

site. Daily rates were higher at the peat site. Anomalously high rates were

observed during June 1985, when the mean daily rate was 1.15 cm/day. This

occurred during a period of high temperatures, little rainfall and low ambient

water levels. Annual water loss at the peat wet prairie site was about 114 cm.

The mean difference between sites was 0.10 cm/day, (significant at P = 0.001),

indicating dramatically higher water loss rates at the peat site than at the marl

prairie site.

Mean daily evapotranspiration rates at the marl prairie (P37) followed a

smooth sinuous pattern over the time course of a year (Figure 4), whereas the

peat prairie site had dramatic anomalies during the early part of the summer.

The variation over time was summarized as percentages of annual water loss for

each month for use in the modeling section.


Summarization of Evapotranspiration Studies

The transpiration studies indicate that a difference in community
transpiration exists among the types studied, that comprise a hierarchy of water



























00


o,6


00




'-U)

m m 00



00
00,






C13 0"
V)V





CZ0
00.0
















00 00. \0
0\
ZO


o cz
00,


Pm CZl




















(~~p/tn3)~kr U~lJdUf~~~~ppUA
005



>E

00X
.6,
IT C4 00IT C4 C

-4 6 C5 C
(X~/Lo)uopiisuzoeAQXpp uo







use in the landscape. The variation in community rates is due to dramatic

differences in leaf area among the vegetation types, more than losses per leaf

area. The marl prairie has low rates of transpiration, has a small total leaf area

per unit ground area, and is the most oligotrophic of the sites. The sawgrass sites

on peat have higher leaf area and community transpiration rates. The highest

transpirative water loss is from the tree island/swamp forest community, due to

higher rates and highest leaf areas. The swamp forest appears to be the most

eutrophic of the sites.

The studies of community evapotranspiration, indicate that differences

among types exist, the components of evaporation and transpiration vary among

the types. The transpiration and community evapotranspiration data both

indicate the fluctuation in rates over an annual cycle. Even though the rates

fluctuate seasonally, a mean daily rate will be used as the basis of comparison

among types and pathways of water loss. The P37 site is the only one of the sites

to have both transpiration and community estimates. The transpiration estimates

(0.02 cm/da) were about an order of magnitude lower than the community

estimates (0.2 cm/day), indicating that transpiration is not a large pathway of

water loss. The wet prairie on peat site (P33) had a higher daily rate (0.3

cm/day) than the marl prairie. The wet prairie rates were higher than the daily

transpiration rates in the sawgrass and lower than the transpiration rates at the

tree island site.

Evapotranspiration is one major pathway of water flow out of wetland

ecosystems. The other major outflow is via overland flow. The theoretical and

practical work studying this process will be reviewed in the next section.







Flow in Wetlands
Surface water flow in wetland ecosystems has been studied using

principles of applied hydraulics. The theoretical foundations for flow arise from
hydraulic principles, although there appears to be disagreement in the literature

as to the fundamental nature of flow regimes. A unique functional feature of
wetland systems is the periodic flooding and flow followed by periods of no
flow. This periodicity involves transitions among flow regimes from no flow to

laminar flow to turbulent flow. Most studies of wetland flow (Ree 1949, Petryk

and Bosmajian 1975, Lin and Shih 1979, Rosendahl and Probst 1980, Rosendahl

1981, Shih and Rahi 1982) assume a turbulent flow regime and utilize Manning's
formula (Manning 1890). Kadlec (1990) argues that laminar flow is dominant in
wetlands due to the low energy gradients and suggests a formulation such as the

one given in Equation (2) be used. Wetland flow is of the magnitude that

momentum terms in flow equations are usually ignored (Kadlec 1990, Hammer

and Kadlec 1992). The disagreement involving flow regimes can be partially

resolved as one of parameter values, as shown in Equation (2). Equation (2) is a
generalization of the Manning equation relating flow velocity as a function of

hydraulic radius and hydraulic slope. The critical parameters are a and 0 and

are assigned values of a = 0.5 and 3 = 0.67 in the Manning formula. Kadlec

(1990) reports values of a range from 0.4 to 1.0 and 3 from 2.5 to 3.75 for laminar

flow in wetlands.


(2) V = K S Rhe

V = velocity
K = general resistance coefficient

S = Hydraulic slope (difference in elevation potential)







Rh = Hydraulic radius (Cross-sectional area/wetted perimeter: Note

for very wide channels, Rh is approximated by water depth)
aX = slope exponent

3 = hydraulic radius exponent


In Manning's formula, K is represented by 1/n in SI units (1.49/n in

English units), and n is referred to as a roughness coefficient or as Manning's n.

Assuming turbulent regimes, most prior work on wetland flow has

attempted to develop better estimates of Manning's n, and particularly how this

coefficient varies with factors of water depth and vegetation density. The earliest
work relates n with water depth (Ree 1949) and the product of velocity and water

depth (Palmer 1945). Ree worked with flow in short grasses and found an
increase in n with depths up to the height of the grass, then a decrease, similar to

Palmer. Petryk and Bosmajian (1975) laid much conceptual framework, and

related n as a function of vegetation characteristics (primarily vegetation

density), boundary roughness and hydraulic radius. Petryk and Bosmajian

(1975) thought that vegetation resistance was much greater than boundary

roughness, and hence related n to vegetation density and inversely to hydraulic
radius; for n to remain constant vegetation density had to decrease if depth (Rh)

increased. Shih and Rahi (1982) using principles of Petryk and Bosmajian (1975),

developed estimates of n from 0.16 to 0.55 for marshes in the Kissimmee basin,

where n varied with seasonal changes in vegetation density.

Studies of flow in the marshes of the Everglades date back to the 1940s,
with the earliest work (Parker et al. 1944) measuring decreased flow in canals

resulting from infestations of water hyacinths. The U.S. Army Corps of
Engineers developed estimates of Manning's n in design memoranda for the

C&SF project, that averaged 0.035 and ranged inversely with water depth (U.S.







Army 1954). Leach et al. (1971) investigated data from a series of years and

found maximum flow rates of 1600 ft/day (0.6 cm/sec), which translate to a
cumulative annual distance of 50 miles (81 km). Rosendahl and Probst (1980)

and Rosendahl and Rose (1981) measured flow rates and resistance coefficients in

sawgrass and open marshes in Everglades National Park and reported greater
flow rates; from 0 to 0.022 ft/sec ( 0.67 cm/s) in dense sawgrass strands and
from 0 to 0.034 ft/sec (1.0 cm/s) in open marshes. Rosendahl (1981) calculated a
range of values of Manning's n between 0.4 and 2.4, with a mean of 0.99 and

found little correlation with depth.

Most of the models of water flow in the Everglades marshes use
Manning's equation and with varying reports as to the sensitivity of model

output to variations in the roughness coefficient. Lin and Shih (1979) used values

of n between 0.4 and 1.2, with an inverse relationship between n and depth. Lin
and Shih (1979) found that seasonal variation in n was necessary to achieve

model calibration, with lower values in the dry season and higher values in the

wet season. MacVicar et al. (1983) relate flow coefficients as a function of land

cover type, with values ranging from 0.1 to 2. Perkins and MacVicar (In press)

did a sensitivity analysis using very low value of n (0.05) and very high (2.0) and
found more effect on flow volume than stage, recommending development of

better coefficients with vegetation type. Walters et al. (1992) developed a
coefficient of flow equivalent to K in equation 2, of 2, which translates to a

Manning's n of 0.75.

A review of previous studies involving flow in wetland systems can be

summarized as follows. Although there is still uncertainty regarding the nature

of the flow regime, a generalized form of Manning's equation can be used. The
equation equates flow velocity as a function of water depth, hydraulic slope, and

a resistance coefficient, such as Manning's n. The flow coefficient can be







estimated from vegetation density, defined as the total cross-sectional area of

vegetation per unit length of flow (Petryk and Bosmajian 1975)

In the preceding sections, the processes influencing the measures of

evapotranspiration in south Florida and flow relationships in wetlands were

discussed in context of scale. The way in which these processes are incorporated

into a landscape model is primarily a scaling issue. The details of how this

scaling process was done is described below in the section on modeling

methodology. A brief review of some general concepts and approaches to

scaling in ecological models is included as a final piece in this background

section.


Ecological Models and Scale

Most descriptive landscape or ecosystem models have explicit domains in

space and time. Temporal domains are defined at the small end by the time step

and at the large end by the time horizon. Similarly, spatial grain is defined by

the size of grid cell and extent by the number of grid cells. Bounding the model

along spatial and temporal dimensions, defines what is inside and outside the

model domain in terms of scale. The empirical rule of thumb is that models

cover no more than two to three orders of magnitude in either space or time. The

rule is probably not related to technical constraints such as computer processing

power (Costanza and Maxwell 1992). The scale restriction may be related to

practical factors, such as debugging problems, validation criteria (Clark et al.

1979 ), or understanding the model (Costanza and Sklar 1985).

By using an ecological model with a fixed domain, decisions must be

made about what to do about processes that occur at different scale ranges. The

common approach in model construction is to treat processes that occur at slower

speeds and over broader ranges as constants. For example, if a model is







constructed to examine seasonal dynamics in sea level, then the global processes

that created a dramatic sea level rise between 5 and 10 thousand years ago, are

assumed to not change much over the course of a few years and therefore are

treated as constant. Faster processes are generally treated as noise or random

fluctuations within the system and can be averaged. Using the same example,

tidal influences on sea level occur over a short term and therefore can be treated
as noise over time spans of a year. The short term (daily) influences are averaged

to study seasonal or annual dynamics.

Components both inside and outside model structures are dealt with by

processes of aggregation and disaggregation. The simplest form of aggregation

is linear scaling. Scaling is defined as the translation of units based upon a fixed

relationship or ratio among metrics used in measurement. For example,

temporal metrics of minutes, hours, and days have a fixed relationship, therefore

one can trivially determine that one day is equal to 1440 minutes. Broadly, the

issue of aggregation has been dealt with either in terms of applying standard

statistical methods to derive "best" estimates and minimize error (O'Neill et al.

1986, Gardener et al. 1982), or by assuming linear aggregations among complex

variable sets (Iwasa et al. 1986, 1989). Basically, aggregation works if the

assumptions and rules used remain valid over the scales of translation.

Incorporation of evapotranspiration, flow and vegetation dynamics into a

spatially and temporally explicit hydrologic model of the Everglades involves

"fitting" these into a model with explicit bounds in time and space. The next

section describes the structure and development of the model used to investigate

hydrodynamics in the system.







Model Description and Development

The framework for the model was developed during a series of

workshops held between 1989 and 1990. The initial objective of the model was to

improve communication among scientists, engineers and practitioners in order to

discuss issues related to Everglades restoration. The model resisted a series of

attempts at invalidation (Walters et al. 1992, Richardson et al. 1990) and hence

has become a credible tool for examination of movement of water across the

landscape. The model framework depicts the time dynamics of the hydrology

within approximately 800 4 x 4 km grid cells (Figure 5) that cover the historic

Everglades ecosystem and surrounding areas. The model is bounded by Lake

Okeechobee to the north, the Atlantic coastal ridge to the east, Florida Bay to the

south and the Big Cypress National Preserve to the West (Figure 5). The basic

framework reported by Walters et al. (1992) was modified to include coupling of

vegetation and hydrology. The hydrology and vegetation components of the

model are described in the next two sections, followed by the development of

subroutines of evapotranspiration and flow that link the hydrologic and

vegetation components.


Hydrologic Components

Water depths within a cell change over time due to inflow associated with

rainfall, losses via evapotranspiration, net flux associated with overland runoff

from adjacent cells and net flux of water into or out of canals. The model is

driven by historic rainfall data, covering the period from January 1960 through

December 1988. Input data are of total monthly rainfall, averaged over the entire

basin area. Even though spatial gradients exist in the system (MacVicar and Lin

1984) equal amounts are added to each cell at the beginning of each simulated

month. Annual rainfall during the model time period ranges from 100 to 150


















































Figure 5. Model grid used to depict hydrologic and vegetation dynamics in
Everglades ecosystem.







cm/yr. The input to a cell is actually as net rainfall; that is, the recorded rain

total minus evapotranspiration.

Water is moved to adjacent cells based upon the Manning formula, where

velocity is a function of hydraulic slope, water depth and a resistance coefficient

(Equation 2). The alpha and beta values from equation 2 are 0.5 and 0.67.

Hydraulic slope is determined by difference between adjacent cells in the sum of

ground elevation and water depth. Levees in the model stop flow movement.

Water management is incorporated in the model by a series of water

management schedules within each conservation area and rules of water

movement around the schedules. If water levels at index cells are above the

monthly schedule, then water is removed from certain output cells. If water

levels are below scheduled levels, then water is retained. Water is diverted to

coastal areas by removal from key index cells to simulate removal via canals.

Target diversion rates are a function of maximum diversion (flow allowed in a

canal) and water depth.

Only surface water movement is calculated; no losses to groundwater are

included. In the peat areas of the system, with high infiltration resistance, this is

not considered to be a major source of error. In the transitional, sandy and marl

areas, movement from surface to groundwater is considerable, resulting in

overestimates of water depths.

The modeled area exchanges water with the surrounding areas. Water is

input from Lake Okeechobee, based upon decision rules and schedules within

the Lake. The Big Cypress regions to the west receives the same rainfall inputs as

areas over the Everglades proper. Water exchange with this area is only

constrained by structures in the model. No exchange (other than management

diversion) is made with the east, even though under historic conditions water

moved through the coastal ridge through a number of rivers, sloughs and finger








glades. Boundary conditions at Florida Bay vary seasonally from a low value in

winter to a high value in late fall, to reflect the annual dynamics of sea level.

Information on water depth and flow over time can be output for each

grid cell. Target cells that correspond to the locations of sites P33, P35, P37 and
P38 (Figure 6) were used as key indicators of model results. Cumulative annual

flow amounts were determined for three flow sections. One is for the set of cells

that coincide with the Tamiami Trail Flow section (Figure 6). The other sections

are at the boundary cells at the mouth of Shark River Slough, and at the mouth of

Taylor Slough.
The preceding paragraphs summarized the hydrologic variables and

interactions within the model framework. The next section deals with the

structure of the vegetation component of the model.


Vegetation Components

A total of 26 cover types were created to capture the variety of vegetation

patterns in the landscape of the model area (Table 4). The types were based upon

combinations of plant associations, as a grid cell of four kilometers generally

covers a non-homogeneous combination of plant communities. Some of the

vegetation types, such as sawgrass, can be the only type within some grid cells.

Other types, such as sloughs and tree islands are always smaller than the grid

cells. To determine the percent cover of vegetation communities within the

dominant landscape types, and how robust the measures of percent cover were

with a change in scale, the following exercise was done.

Vegetation cover was measured in a series of subsamples from the two

classified sixteen kilometer SPOT satellite scenes used in the vegetation map

section above. A total of thirty two samples were made (sixteen from each

scene), for window sizes of 500, 1000, 2000 and 4000 m. Percent cover was




















































Figure 6. Location of sample rain and stage gauges, flow sections and pan
evaporation sites within the Everglades region.


* RAIN GAUGE
* WATER LEVEL GAUGE
< FLOW SECTION

O PAN EVAPORATION
TEMPERATURE





61




Table 4. Description of vegetation categories (landscape units) used in
Everglades model.






Map LANDSCAPE UNIT
No. Description of Components

1 sawgrass, slough, tree island
2 sawgrass/slough
3 slough, with periphyton
4 slough, no periphyton
5 sawgrass marsh
6 sawgrass with woody invasion
7 sawgrass,cattail
8 wet prairie, on peat, tree islands
9 wet prairie, on peat, native woody plants
10 wet prairie, on peat, exotics
11 wet prairie, on marl and tree islands
12 wet prairie, on marl, muhly grass
13 wet prairie, on marl,native woody plants
14 wet prairie, on marl, exotics
15 Pine, hammock
16 Pine, hardwood
17 Pine, prairie
18 Upland hardwood scrub
19 tall mangrove forest
20 short mangrove forest
21 mangrove prairie
22 dwarf cypress
23 cypress dome
24 cypress hardwood
25 agriculture
26 no vegetation







calculated from each sample for each of the three major vegetation types;

sawgrass, tree island and wet prairie. The percent cover of sawgrass, wet prairie

and tree island communities in the southern Everglades did not vary over the

sampled window sizes. For windows sizes from 500 through 4000 m, the cover

of each of the three types was relatively constant (Figure 7). The mean cover was

about 66% for sawgrass, 21% for wet prairie and 13% for tree island. A two-way

analysis of variance indicated no significant difference in cover among the

window sizes. Although variances tended to decrease with window size, the

differences were not enough to violate assumptions of homoscedasticity. These

results indicate that using the percent cover of plant communities to describe the

landscape units is very robust to changes in the cell size of the landscape units.

A map editor is established in the model to allow for creating the initial or

starting vegetation array within the active cells. The map editor creates an array

that consists of a vegetation type code for each cell that can be addressed and

updated during the simulation. The rules in the model for changing the

vegetation types are the subject of the next section.


Changes in Landscape Vegetation Types

The vegetation change module of the model is set up to switch among

vegetation types, based upon rules relating to hydroperiod, nutrient

concentration and fire. At the end of each year of simulation, the hydroperiod

(number of months per year that a site is wet) and soil nutrient concentration are

calculated for each grid cell. The annual hydroperiod value is used to update a

running average hydroperiod for each cell. Fire is a stochastic event, with

probability related inversely to the annual hydroperiod value. If a random

number is less than the assigned probability for the hydroperiod value, then a

fire event is said to have burned the cell. The average hydroperiod, soil nutrient




















500 1000 2000 4000
Window Size (m)


WET PRAIRIE


500


1000 2000
Window Size (m)


4000


500 1000 2000 4000
Window Size (m)






Percent cover of sawgrass, wet prairie, and tree island in various
window sizes.


80

60

40

20


Figure 7.







concentration and a fire event are all used as critical values to make changes

among vegetation types. If average hydroperiods are wetter or drier than a

certain threshold value, or if the soil nutrient concentration exceeds a critical

value, or if a fire occurs in the cell, then the vegetation cover type for that cell
changes. The threshold values of hydroperiods and target transition types are

user defined.

The transitions among vegetation types were determined using

information derived from a combination of sources, namely, literature and field

experience. The transitions involve the loss or gain of certain community types

within a cover type. The primary cover in the peat portions of the Everglades

system is a mosaic of sawgrass, slough and tree islands, designated as type one.
The hydroperiod ranges from 9 to 11 months in this type. If hydroperiods exceed

11 months, then the tree island and sawgrass types disappear, leaving only

slough (Craighead 1971, Worth 1987). If hydroperiods are less than 9 months,

then woody plants invade the sawgrass and slough (Craighead 1971, Gunderson

and Loftus In Press) resulting in a change to type 9. The other dominant peat

landscape type was a monospecific stand of sawgrass in the area now known as

the EAA (Davis 1943). If the sawgrass burns or dries out, woody plants invade,

(Wade 1980) resulting in type 6. The vegetation dynamics associated with a

change in nutrient status include a loss of periphyton in a slough system (Swift

1984), and a transition from sawgrass to cattail (Davis 1989).

The preceding paragraphs describe the "landscape" vegetation units used
in the model. These units are comprised of combinations of identifiable plant

communities. The transitions among landscape types involve the addition or

replacement of plant communities within each landscape unit. Since the plant

communities provide the "building blocks" for each landscape unit, they will be

used as the basis of relating flow and evapotranspiration.







Development of Flow Coefficients for Landscape Units
Two steps were involved in the determination of flow coefficients by
landscape unit. The first step related estimates of vegetation density for each of

the dominant plant communities to flow coefficients. The second step used the
percent cover relationships within a landscape type to develop a spatially
weighted flow coefficient.
To develop relationships between vegetation type and flow regimes,
estimates of vegetation density were derived. Vegetation densities were then
translated to Manning's flow coefficients using relationships developed by
Petryk and Bosmajian (1975) and Shih and Rahi (1982). Vegetation density was

determined for four vegetation types: sawgrass marsh, wet prairie over peat,

marl prairie and tree island. The literature was surveyed for measured values of

stem density for the gramineous vegetation types (marsh and prairie) and for

values of basal area for the forest type (tree island). Average stem densities in
the graminoid vegetation types were multiplied by the average stem size to yield

an average cross sectional area per length of flow. Total basal area was divided

by the stem density to yield an average tree size, then average cross sectional

area was determined per unit of flow length. The values of vegetation density

were then correlated to a Manning's n value based upon data compiled by Petryk
and Bosmajian (1975).

Stem density varied from a low value in the wet prairie (0.05 /m2)
to the highest (42/m2) in the marl prairie (Table 5). Even though the marl prairie
had the highest stem density, the sawgrass had the highest vegetation density.

The cross-sectional area of the plants comprising the marl prairie were much

smaller than sawgrass. The vegetation density reflects the total cross-sectional

area (m2) per ground area in the direction of flow (m3), and is in units of m-1.

The vegetation densities were highest in the sawgrass areas (0.15 m-1),










Table 5.


Vegetation density and related flow coefficients as a function of
depth for sawgrass, tree island, wet and marl prairie vegetation
types.


Vegetation Type Sawgrass Wet Prairie Tree Island Marl Prairie
Stem Density
(#/m2) 28 *(1) 3 *(2) 0.6 *(3) 42 *(4)
(#/ft2) 8.5 0.9 0.2 12.8
Stem Size (ft) 0.08 0.03 0.30 0.02
Stem Area
(ft2/ft2) 0.71 0.03 0.05 0.27
Depth
Vegetation (ft)
Density 0.5 1.42 0.06 0.11 0.53
(ft2/ft3) 1 0.71 0.03 0.05 0.27
1.5 0.47 0.02 0.04 0.18
2 0.36 0.01 0.03 0.13
2.5 0.28 0.01 0.02 0.11
3 0.24 0.01 0.02 0.09
3.5 0.20 0.01 0.02 0.08
Mannings *(5) 0.5 1.12 0.22 0.31 0.68
n* 1 1.26 0.25 0.35 0.77
1.5 1.35 0.27 0.37 0.82
2 1.41 0.28 0.39 0.87
2.5 1.47 0.29 0.41 0.90
3 1.51 0.30 0.42 0.93
3.5 1.55 0.31 0.43 0.95
Model Flow *(6) 0.5 1.3 6.7 4.8 2.2
Coefficient 1 1.2 5.9 4.3 1.9
K 1.5 1.1 5.5 4.0 1.8
2 1.1 5.3 3.8 1.7
2.5 1.0 5.1 3.7 1.7
3 1.0 4.9 3.5 1.6
___3.5 1.0 4.8 3.5 1.6


Herndon et al. 1991
Goodrick 1984
Gunderson 1982
Olmsted et al. 1980
Petryk and Bosmajian, 1975
Walters et al., 1992


n = (depth)^0.67*(veg density)^0.5
K=1.49/n


REFERENCES








intermediate in the marl prairie (0.06 m-1) and tree islands (0.05 m-1), and lowest
in the wet prairie (0.01 m-1). The estimated flow coefficients for sawgrass was

about twice that of tree island and wet prairie, and substantially greater than in

the wet prairie.
A spatially weighted average flow coefficient was determined for each

landscape unit. First, a mean flow coefficient was calculated over a range of

depths for each plant community type. The percent cover of each plant

community in a landscape unit was used as the weighting factor. For example,
in landscape unit 1, sawgrass covers on the average 66% of a cell, wet prairie
covers 22%, and tree islands cover 12%. The depth averaged flow coefficient for

these three types are 1.09, 5.44 and 3.93, respectively. The flow coefficient for this

landscape unit is (0.66*1.09+.22*5.44+.12*3.93) = 2.4. The spatially weighted

coefficients are given in Table 6. Even though the values vary among plant
community types, the spatially weighted averages are similar among the

landscape units


Development of Evapotranspiration Coefficients for Landscape Units

A spatially weighted average evapotranspiration rate was also calculated
for each landscape unit. The plant community rates were derived from average

daily rates calculated from the evapotranspiration data. The values ranged from

a low of 0.22 cm/day in the marl prairie to 0.4 cm/day in the swamp forest

(Table 7). The values for cattail and melaleuca were estimated from other

sources (Koch, unpub. data; Woodall 1980). The calculation of annual totals is
extremely sensitive to the daily rates, a difference of 0.1 cm in the daily rates
results in an annual difference of 36 cm. The spatially weighted mean annual
evapotranspiration values ranged from 95 cm (marl prairie, type 11) to 159 cm for

the unit of exotic trees on peat (Table 7). Annual totals for each landscape type








Table 6. Spatially weighted flow coefficients for each landscape unit used in
the Everglades Model.



Spatially
Map LANDSCAPE UNIT Spatial % Weighted
No. Description of Components Components Average K

1 sawgrass, slough, tree island 66-22-12 2.4
2 sawgrass/slough 78-22 2.0
3 slough, with periphyton 100 5.4
4 slough, no periphyton 100 5.4
5 sawgrass marsh 100 1.1
6 sawgrass with woody invasion 88-12 1.4
7 sawgrass,cattail 50-50 1.3
8 wet prairie, on peat, tree islands 80-20 5.1
9 wet prairie, on peat, native woody plants 60-40 4.8
10 wet prairie, on peat, exotics 20-80 4.2
11 wet prairie, on marl and tree islands 80-20 2.2
12 wet prairie, on marl, muhly grass 80-20 2.2
13 wet prairie, on marl,native woody plants 60-40 2.6
14 wet prairie, on marl, exotics 40-60 2.6
15 Pine, hammock 100 1.8
16 Pine, hardwood 100 1.8
17 Pine, prairie 100 1.8
18 Upland hardwood scrub 100 1.8
19 tall mangrove forest 100 10.0
20 short mangrove forest 100 10.0
21 mangrove prairie 100 10.0
22 dwarf cypress 100 2.0
23 cypress dome 100 2.0
24 cypress hardwood 100 2.0
25 agriculture 100 2.0
26 no vegetation 100 2.0








Table 7. Average daily and annual evapotranspiration for plant
communities used to develop evapotranspiration coefficients for
Everglades model.

Plant Community Daily ET Annual ET Annual ET
(cm) (cm) (in)
Sawgrass 0.33 120 47
Wet Prairie/Slough 0.26 95 37
Tree island 0.42 153 60
Marl Prairie 0.22 80 32
Exotic-Melaleuca 0.48 175 69

% Spatially Relative
Map LANDSCAPE UNIT Spatial Weighted Annual Rate
No. Description of Components Components Annual ET X/114 cm
(cm)
1 sawgrass, slough, tree island 66-22-12 119 1.04
2 sawgrass/slough 78-22 115 1.01
3 slough, with periphyton 100 95 0.83
4 slough, no periphyton 100 95 0.83
5 sawgrass marsh 100 120 1.06
6 sawgrass with woody invasion 88-12 123 1.08
7 sawgrass,cattail 50-50 145 1.27
8 wet prairie, on peat, tree islands 80-20 107 0.93
9 wet prairie, on peat, native woody plants 60-40 118 1.04
10 wet prairie, on peat, exotics 20-80 159 1.40
11 wet prairie, on marl and tree islands 80-20 95 0.83
12 wet prairie, on marl, muhly grass 80-20 95 0.83
13 wet prairie, on marl,native woody plants 60-40 110 0.96
14 wet prairie, on marl, exotics 40-60 118 1.04
15 Pine, hammock 100 114 1.00
16 Pine, hardwood 100 114 1.00
17 Pine, prairie 100 114 1.00
18 Upland hardwood scrub 100 114 1.00
19 tall mangrove forest 100 114 1.00
20 short mangrove forest 100 114 1.00
21 mangrove prairie 100 114 1.00
22 dwarf cypress 100 114 1.00
23 cypress dome 100 114 1.00
24 cypress hardwood 100 114 1.00
25 agriculture 100 114 1.00
26 no vegetation 100 114 1.00







were expressed as a ratio of 114 cm/yr, the fixed coefficient for the model. As

with the flow values, there appears to be some spatial convergence of averages.

That is, a relatively constant percentage of plant community types within a

landscape unit combined with a difference in rates among plant communities,

appears to result in a global average for a grid cell.

The preceding section of this chapter described the hydrologic

components, vegetation components and linkages of flow and

evapotranspiration in the model. Water depths within each grid cell change

monthly as a function of historic rainfall, net flow, and evapotranspiration. At

the scale of the model, 26 landscape types comprised of plant communities, are

used to describe vegetation patterns. Flow and evapotranspiration rates are

linked to the "landscape" type. This linkage was done through two steps: first to

determine coefficients for the dominant vegetation communities, then to develop

a spatially weighted average coefficient for each landscape unit based upon the

percent cover of vegetation types within the landscape unit. The transitions

among the landscape units are a function of cumulative water depths

(hydroperiod), fire and nutrient concentration. The next section of the chapter

presents the results of sensitivity analyses and testing the hypotheses.


Results

The results section has three parts. The first part assesses the sensitivity of

key parameters to flow calculations and compares model output with historic

data. The second part presents attempts to invalidate the linkages between

vegetation and hydrology. The third portion of this section reviews the tests of

the upstream area hypothesis.







Sensitivity Analysis- Flow and Evapotranspiration
Tests were done to explore the sensitivity of the model output (primarily
flow) to uncertainties in parameters associated with flow and evapotranspiration.

The sensitivity analyses of flow coefficients were done by doubling and halving

all coefficients calculated in the above paragraphs, running the model for a full

28 year scenario under natural conditions (no water control structures in the
system). The sensitivity to evapotranspiration was tested by running a full 28
year natural scenario with the annual evapotranspiration rate set at 89, 102, 107,
and 114 cm (35, 40, 42 and 45 inches).
Doubling and halving the flow coefficients did not appear to have an

appreciable affect on flow through the Tamiami flow section (Figure 8). The

largest deviations among the three sets of coefficients occurred during wet years

(model simulation years 1967, 1969, 1970; Figure 8). During these periods, the

flow differed by about 500 x 106 m3/yr between either of the runs with adjusted
coefficients and the unadjusted coefficients. This difference during wet years,

between the adjusted coefficient flows and the unadjusted flow, was about 12%

of the unadjusted flow. Differences during dry years were much less. There was

no evidence that changes in vegetation landscape types during any of these runs

altered or confounded the flow relationship among the adjusted coefficients. The

flow results from the unadjusted run were always intermediate between the

higher flows calculated by doubling the flow coefficients, and the lower flows

associated with halving the flow coefficients.

The results of varying base evapotranspiration indicate counter-intuitive

effects on rates of annual flow through the Tamiami flowsection. The results are

unexpected because there is not a constant relationship between the

evapotranspiration rate and amount of flow. The lower evapotranspiration rates

should generate higher stages and higher flow. The simulated flow data start out











4000-

3500

t3000

2500

,2000
0
S1500

:1000

500

1960
1960


1965 1970 1975 1980 1985
YEAR


Figure 8. Effects of doubling and halving flow coefficients on simulated flow
through Tamiami flow section.


--- FLOW COEFFICIENT* 1/2

-0- FLOW COEFFICIENT 2

-A- NATURAL







with this relationship, but the relationship changes through the time course of

the model run (Figure 9). These discrepancies are the result of the linkages

between hydrologic conditions and the vegetation landscape conditions. If the

system gets too dry, such as the scenario of higher base annual

evapotranspiration (114 cm/yr), then woody species invade the landscape,
increasing the evapotranspirative loss and lowering flow rates. The system

appears to entrain flows at lower levels of evapotranspiration. The lower

evapotranspiration maintains "wetter" landscape types that result in higher flow

rates.
Even with these interesting and counter-intuitive effects of vegetation-

hydrology-evapotranspiration linkages on flow, the model is apparently more

sensitive to changes in evapotranspiration than to variations in flow coefficients.

Changing the base transpiration rate only 25 cm/yr, results in a variation of

annual flow on the order of 500 x 106 m3/yr (Figure 9). A similar effect was

achieved by doubling and halving the flow coefficients. After establishing the

sensitivity of changing the base settings on flow regimes, model output using the

base settings will be compared to historic data.


Agreement with Historic Data
The model output, both stage and flow, indicates periods of agreement

and divergence with measured data. The actual stage at P33 and P35 and flow

through Tamiami tend to be lower than the model output during the period from

1961 through 1965 (Figures 10 and 11). This is probably due to the management

policy in effect during this period, when little or no flow was delivered to the

park (Wagner and Rosendahl 1985, Gunderson 1989). During other years, the

actual and modeled stage data tend to qualitatively agree. Since there was no

groundwater component to the model, agreement was only possible with surface





















































Figure 9. Effects of varying base evapotranspiration rates on simulated flow
through Tamiami flow section.


4000

3500-

S3000-

0 2500

I 2000

S1500

1000

500


1960 1965 1970 1975
YEAR


- 89 ---- 102 107 -- 114

Annual Base Evapotranspiration (cm)


1980


1985












MANAGED ------- ACTUAL


Time series of simulated (solid lines) and actual (dashed lines)
stages at gauge P33 from 1960 to 1988.


100

80

60

40

L 20

0

H-20

-40

-60


-80 E
1960 62 64 66 68 70 72 74 76 78 80 82 84 86
Year


Figure 10.












- MANAGED --..... ACTUAL


Time series of simulated (solid lines) and actual (dashed lines)
stages at gauge P35 from 1960 to 1988.


80

60

S40
u

S20

0
H
5-20

-40


-60 II
1960 62 64 66 68 70 72 74 76 78 80 82 84 86
Year


Figure 11.







water conditions. The modeled flow through Tamiami trail also agreed fairly
well with measured flow in periods other than the early 1960s and early 1970s.

The vegetation patterns at the end of the simulation period of the natural
scenario appear to agree with early descriptions of vegetation in the Everglades.
The initial array of landscape units in the model grid consisted of a sawgrass

plain south of Lake Okeechobee, the tree island/sawgrass mosaic in the central
core of the system, and marl prairies units in the south. (Figure 12). At the end of
the run, the sawgrass plain and marl prairie units had persisted (Figure 13). The

tree island/sawgrass mosaic had changed to a sawgrass/slough type in the area

of the persistent pool described in the paragraph above. Davis (1943) and Jones
(1948) observed and mapped similar patterns; the tree island/sawgrass mosaic
was only mapped in the northeast and southwest portions of the central
Everglades and a tree-less marsh was in the topographically lower southeast

region. The tree-less sawgrass/slough vegetation type is also captured in the

native Americans description of the system as "Pa-hay-okee", which loosely

translates to a grassy lake (Douglas 1947).
The output from the model agrees fairly well with historic hydrologic and

vegetation information. Even with uncertainties and sensitivities to

understanding flow and evapotranspiration processes, the model captures key

aspects of hydrologic and vegetation dynamics of the system. The next section

puts the model at risk, and attempts to determine the bounds of the relationships

between the hydrology and vegetation.


Linkages between hydrology and vegetation

Since the model output agreed fairly well with historic data sets, a test was
developed to determine the limits of the influence of vegetation dynamics on the

relationship between rainfall and runoff. The test consisted of a series of model












5 5 5 5 5 5 5 5 5 5 5 5 5 1


23 923
9 24 24
24 24 23
24 24 23


23 24 23
24 24 24
24 23 23
232322


11 17 17 23 23 23 23
11 11 22 17 11 17 22
11 11 11 17 17 17 17
23 11 11 17 17 17 17
11 11 9 22 17 17 17
11 11 24 24 23 11 17
11 11 24 24 23 17 11
. 21 11 23 23 11 23
. 20 11 11 11 23 23


20 21 20
S21 20
S. 21
S. 21

. ,


232323
232323
23 24 23
23 24 23


17 17 22 22 22
2311222222
2323222222


21 11 11 23 23 23 22 22 22
20 21 11 23 23 17 22 22 11
21 20 21 11 11 17 22 11 11
19 20 20 21 17 17 11 11 11
21 21 19 21 21 11 11 11 11
19 21 21 19 21 21 11 11 11
19 19 21 19 19 21 11 11 13
. 21 21 19 21 19 20 11 1
. 19 19 19 21 21 20 1
S. 19 19 19 21 19 20 1
.19 21 21 19 20 1


222222
222222
222222
222222
11 11 11
11 11 11
11 11 13
13 13 1
13 13 1
1 1 1
1 1 1
1 1 1
1 122
1 1 17


. . . 19 19 19 19 19 20 1 17 22
. . . .. 19 19 19 19 19 19 19 21 11 11
. . . 19 21 19 19 19 19 19 19 21 21


..... 555
. 555
. . 5 5 5
. . . 5 5 5
. . . 5 5
. . . 5 5
. . . 5 5
5
5
5
23 22 23 23 22 22 22 5
. . . 5

. . . 5

23 232322 11 22 5 5
22232322 11 222223
2323232222222222
2323232222222222
2222222222222222
17 17 22 22 22 22 22 22
17 17 17 22 22 22 22 22
17 17 22 22 22 22 22 22
17 17 17 22 22 22 22 22


. . . 19 21 19 19 19 19 19 19 19 21 21 11
. . . 19 21 21 19 19 19 21 19 19 19 19 20 2
. . . .. 21 21 21 21 19 19 19 19 19 19 19
. . . . 21 21 21 19 19 . .191
. . . . . 19 19 . . .


5 5
5 5
5 5
5 5
5 5
5 5
5 5
5 5
5 5
5 5
5 5
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
2 1
11 1
1 1
1 1
1 1
1 1
1 1
1 11 1
11 17 1
17 15 1
17 18 1
13 13 1
22222
1322


Figure 12. Map codes for initial landscape vegetation types in the Everglades
model. Dominant code in the Everglades proper is type 1. Cover
types developed from Davis (1943). Explanation of codes is found
in Table 4.


5 5 5 5 5 5 5
5555555
5555555
5555555

5 5 5 5 5 5 5
5 5 5 5 5 5 5
5 5 5 5 5 5 1
5 5 5 5 5 5 1
5 5 5 5 5 5 1
5555555
5555555
5555555
5555551




5 5 5 5 5 5 1
5555551
5555551
5555551

1 1 15555551 1
1 1 1 1 1 1 1
1 1 1 1 1 1 1

1 1 1 1 1 1 1
1 1 1 1 1 1 1
1 1 1 1 11 1
1 1 1 11 1 1
1 1 1 11 11
1 1 1 1 1 11
1 1 1 1 1 1 1




1 1 1 1 1 1
1 1 1 1 1 1 1

1 1 1 1 1 11
1 1 1 1 1 1 1
1 1 1 1 1 1 1
1 13 11 13 8 11 1
1 11 11 13 13 11 11
1 11 13 13 13 11 11
1 12 1313 1311 11





8 11 13 11 8 11 11

9 20 20 20 20 21 21
20 19 19 19 20 20 20
.9 19 19 19 .
9 . . .


11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
111
111
11
11
























23 10
23 24
23 24
24 24


24 24 9
24 9 9
24 24 24
24 24 24
13 15 17
11 1313
111313
231313
11 13 9
11 13 9
11 13 24
.2113
.2013
202120
.2120
.21
..21


24 24 24 24 24
24 24 9 24 24
9 24 24 24 22
24 24 24 22 24
24 24 23 24 24
1513151313
15 15 15 15 15
17 15 15 15 15
13 15 15 15 15
24 24 14 1517
24 24 15 13 15
24 24 13 24 24
14 13 24 24 24
21 13 13 24 24
20 21 13 24 24
2120211313
19 20 20 21 15
21 21 19 21 21
19 21 21 19 21
19 19 21 1919
21 21 19 20
.191919
.19 19 19
S. 1920
S. 1919
.191919
S.19 21 19
S. 19 21 19
S. 19 21 21
S. 21 21
.. 21


Map codes of landscape vegetation types in grid cells of Everglades
model at end of 28 year simulation run. Note presence of type 3
codes in right -central portion of array. See Table 4 for explanation
of codes.


13 24 24
24 24 13
24 24 13
24 24 13
9 24 13
1313 13
15 14 13
15 15 13
15 13 13
15 15 13
151313
131313
24 13 13
24 13 13
151313
15 13 13
17 13 11
1313 8
21 8 8
21 8 8
19 20 3
202020
20 19 20
20 19 20
19 19 19
19 19 19
19 19 19
19 19 19
19 19 19
21 21 19
21 21 19


. 5 6 6 6
. 5 6 6 6
. 5 6 9 6
5 5 5 6
5 6 6
5 6 6
S. 5 55
. . 6 6
. 6 6
. 5 5
. . 5 5
13 13 13 6 5
1.5666







13 13 13 24 1
13 13 13 22 1
13 13 13 22 1
22 22 22 22 1
13 13 13 22 1
13 13 22 22 1
13 13 22 22 1
13 13 22 22 1
14 13 22 22 1
14 13 22 23 1
13 13 3 23 1
13 13 23 23 1
13 13 9 1 1
13 14 13 1 3
13 13 13 1 4
13 13 8 1 3
13 9 3 3 3
14 1 3 3 3
3 1 3 3 1
3 3 3 1 8
3 1 10 22 17
1 9 9 17 17
20 9 17 13 13
19 21 13 13 13
19 19 21 21 13
19 19 19 21 21
21 19 19 19 19
19 19 19 19 19
1.5666
.5696
.5556
..566
..566
..555
...66
...66
...55
...55
13 13 13 6 5
1313 6 6 1
13 13 13 24 1
13131322 1
131313 22 1
222222221
13131322 1
1313 22 221
13132222 1
13132222 1
14 13 22 22 1
14 13 22 23 1
1313 323 1
13132323 1
13 13 9 1 1
13 14 13 1 3
131313 1 4
1313 8 1 3

14 1 3 3 3


3 1 10 22 17
1 9 9 17 17
20 9 17 13 13
19 21 13 13 13
19 19 21 21 13
19 19 19 21 21
21 19 19 19 19
19 19 19 19 19
19 .


6 6 9
6 6 6
669
666
666
666
665
665
655
665
555
555
655
565






















111313
1 1 1
1 1 19
1 1 17



1 1 1
1 1 1























13 913
13 920
120 20 1
1 1 19
1 1 1
3 3 3
3 3 3
3 3 3

3 3 3
3 3 3
3 3 3



1 1 1
11 13 13
17 13 13
17 17 9
18 18 17
13 14 13
13 22 9
13 9 13
13 9 20
20 20 19
19 19 19
19 19 .


66
66
96
96
66
65
69
55
56
52
52
5 5
1 1
1 1
1 1
1 1
1 3
3 3
3 3
3 3
3 3
3 3
3 3
33
3 3
3 3
3 3
3 3
113
113
1313
1313
8 13
1113
1313
13 14
1313
1313
2020
19 19
19 19


Figure 13.


696
666
966
666
666
6 6 6
666
565
551
6 6 1
661
691
561
5 6 1
561
1 1 9
1 3 1
3 3 3
3 3 3
3 3 4
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3

1 1 3
13 13 1
3 11
13 8 3







13 14 3
13 13 13
13 9 14
10 9 13
9 13 13
8 13 13
9 11 11
20 21 21
20 20 20
333
333

333
333
333
333
333







13131
3811
1383
13 14 3
131313
13 914
10 913
91313
81313
9 11 11
20 21 21
202020


69
69
10 9
10 10
99
99
99
19
11
11
10 1
1 1
11
11
11
3 3
3 3
3 3
3 3
3 3
3 3
3 3
3 3
33
3 3
3 3
3 3
3 3
2 2
39
312
1311
1313
13.
8 .







runs. For each run, a set amount of rain was delivered each year of a 20 year

simulation. The rain varied seasonally, as modeled by composite sine waves to

emulate the natural annual pattern, but remove any interannual variation. Four

runs were made, ranging from a very dry year to a very wet year. The model

inputs were equivalent to annual totals of 105, 118, 142 and 176 cm (36, 46, 56 and

68 inches).

The relationships between rain and runoff appears constant over a wide

range of rainfall inputs, then dramatically shifts if the system becomes very wet.

The flow and stages at key stations all reached a seasonally oscillating

equilibrium with rainfall inputs less than 142 cm. At steady rainfall up to 142

cm, the landscape vegetation types remained constant, and hence, the

relationship between rainfall and runoff was linear. At an annual input of 176

cm, a dramatic shift occurred around year 6, when the landscape units shifted to

a treeless wet prairie. Without tree islands or sawgrass, the overland flow rates

were greatly increased, resulting in a new equilibrium of flow coming through

the Tamiami flow section (Figure 14). No such vegetation shifts occurred in the

mangrove areas and hence, no dramatic change in the flow relationships in either

the Shark Slough or Taylor Slough flow sections.

These results indicate that the vegetation-hydrology linkages can be

invalidated at an extreme. The key point is that the vegetation array is fairly

stable with constant levels of average rainfall input and the relationship between

rainfall-runoff is constant. However, if the system has a prolonged (at least five

year period) of surplus rainfall, then the vegetation structure is destroyed and a

different equilibrium in the rainfall-runoff relationship occurs. Since the

vegetation units are fairly stable with constant levels of average rainfall, the

relationship between rainfall and runoff may not be dramatically influenced by

the coupling of landscape vegetation dynamics with the hydrology.







5000
4500
4000
3500
3000
2500
2000
1500
1000


500
0


1 6 11
Simulated Year


Results of changing vegetation patterns on simulated flow through
three flowsections in the southern Everglades.


--- TAMIAMI

-0- SHARK SLOUGH

TAYLOR
SLOUGH


Figure 14.







The output of two models, one with and one without coupled vegetation-
hydrologic dynamics, was compared to see if addition of the linkages changed

the predicted hydropatterns. The output of the model with linkages qualitatively

agrees with the models without linkages (Walters et al. 1992, Perkins and

MacVicar In press) All models predict a persistent pool of water north and east

of Tamiami Trail in the area now known as the Pennsuco wetlands. Flow and

stage results are similar between the model with and without linkages (Figure
15). Key uncertainties to these results are in assumptions about the amount of

water that moved surficially and as groundwater into the coastal ridge, and

contributions from Lake Okeechobee.

The results in the preceding paragraphs lead to the conclusion that the

addition of the vegetation and hydrologic dynamics does not improve the

models' ability to predict stage and flow. Dramatic changes in rain and runoff
relationship occurs only after persistent wet conditions. Model output of flow

and stage is not different between models with and without vegetation linkages.

One of the reasons may be that the dominant influence on surficial hydropatterns

is the rainfall. Certainly this is partially true. Other reasons are related to

problems of scaling. The landscape units are composites of vegetation

communities. There is good evidence for dramatic differences in evapo-

transpiration rates and flow resistance coefficients among the vegetation

communities. However, when composite values are calculated at the landscape

levels, the differences are spatially homogenized and converge towards singular

coefficients. The net result of modeling at the landscape scale is that the addition
of more complexity at "smaller" scales does not dramatically improve accuracy of

the model to predict surficial hydropatterns. The failure to improve model

accuracy by the addition of complexity agrees with previous workers (Walters

1986, Clark et al. 1979, Costanza and Sklar 1985). In spite of these limitations, the




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs