Measurement and modeling of the hydrology of cypress wetlands-pine uplands ecosystems in Florida flatwoods

MISSING IMAGE

Material Information

Title:
Measurement and modeling of the hydrology of cypress wetlands-pine uplands ecosystems in Florida flatwoods
Physical Description:
xviii, 340 leaves : ill. ; 29 cm.
Language:
English
Creator:
Sun, Ge, 1965-
Publication Date:

Subjects

Subjects / Keywords:
Cypress swamp ecology -- Florida   ( lcsh )
Wetland ecology -- Florida   ( lcsh )
Wetland hydrology -- Florida   ( lcsh )
Forest Resources and Conservation thesis, Ph. D
Dissertations, Academic -- Forest Resources and Conservation -- UF
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1995.
Bibliography:
Includes bibliographical references (leaves 326-339).
Statement of Responsibility:
by Ge Sun.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 002066468
oclc - 34351137
notis - AKQ4667
System ID:
AA00002049:00001

Full Text









MEASUREMENT AND MODELING OF THE HYDROLOGY
OF CYPRESS WETLANDS PINE UPLANDS ECOSYSTEMS IN FLORIDA
FLATWOODS













By

GE SUN


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






























This dissertation is dedicated to my parents, brothers and sisters at home, who encouraged
me to reach high ...














ACKNOWLEDGMENTS


wish


express


deep


appreciation


to Dr.


Riekerk,


chairman


of my


supervisory committee for all the guidance and help he has given me.


His personality has


made my studies in the USA much easier, and his academic attitude and working habits have


set an invaluable example for me both in school and in my future career.


His many extra


hours spent on this work are especially acknowledged.


Dr. K.


Campbell served as advisor for my minor and provided significant


assistance in my formal training in hydrological sciences.


Dr. G. R. Best introduced me to


the complexity of wetland ecosystems and inspired my interest in wetland hydrology.


H. L.


Gholz encouraged me to think more eco-logically and systematically.


Dr. R.


Mansell led me to the world of soil physics.


Mr. L.


V. Korhnak maintained field equipment and supervised the data collection.


I appreciate his help on the occasions when I got stuck either mentally or physically along


the way in this project.


Mr S.


Liu and Mr.


W.F


Wood helped with the vegetation survey


and the canopy interception studies.


Mrs.


G. L.


Geninger assisted the editing of this


dissertation.

I also wish to express my thanks to the Intensive Management Practices Assessment








funded this project and provided me with a graduate assistantship.

Last, but not least, the on-site help and love from my wife, Haiyan Tong, has made


this work more enjoyable and kept it on schedule.


Her sacrifices and patience were


indispensable to the success of this program and will continue to be in those that come.


















TABLE OF CONTENTS


ACKNOW LEDGMENTS .. . . . . . . . . . . iii


LIST OF TABLES . . . . . . ... .... .... . . . viii


LIST OF FIGURES .. .. .. ... ..... ...... ..... ........ ... x


ABSTRACT . . . . . . . . . . . . . .


CHAPTERS


INTRODUCTION ..


Research Background


Objectives


Structure of the Dissertation


LITERATURE REVIEW .


. . . . . .... . . . .... 1
. . . . . . . . .... . 3


. . . . . . . . . 5


Experimental Field Studies ...........
Wetland Hydrology Studies ......
Upland Hydrology Studies .......
Integrated Wetland-Upland Studies
Simulation Models ..................


RESEARCH APPROACHES .......


* a a a a a S S S S S C S C S S S a a a
. . . . . . . 1
. . . . . . . 1
. . . . . . . 1


. . . . . ... .. 1


Physical Setting, General Experimental Design and Instrumentation


Gator Nationals Forest (GNF) Site
Bradford Forest (BF) Site ....


* C
* a a a a


. . . . . . 19
. . . . . . 24


Methods for Field Data Acquisition and Analyses


EXPERIMENTAL RESULTS AND ANALYSES .... ...... ........ 5


. . 1










Water Table Dynamics . . .. .. .. ..
Temporal Water Level Dynamics ..............
Spatial Distribution of the Water Level ..... ......
Water Table Responses to Storm Events and Daily ET


Groundwater Flow


Groundwater flow direction
Groundwater flow velocity


Deep Seepage .....
Surface Flow ....


* S 9


* a S C S C
* S S 9 5 9


Runoff Characteristics


Variable Source Areas in Flatwoods


. . . . . . . . . 120
. . . . . . . . . 129
. . . . . . . . . 134
. . . . . . S C S S S C 137


. . . . . . . 14 0


Water Balances


Cypress Wetlands Water Balances
Pine Uplands Water Balances ..............
Pan ET Models for Cypress Wetlands and Pine Uplands


Summary and Discussion ...

THE FLATWOODS MODEL

Model Development. . .
Model Objectives ..
Model Structure ....
Governing Equations
Model Inputs and Outputs .
Model Inputs .....


Model Outputs


S S .


.


. .*. . 158


. . . . . .. . . . 172
. . . . . . . . . . 172
. . . . . . . . . . 176
. . .. .. .. .. .. .. .. . .. .. .. .. . 181
. . . . . . . . . a 195
. . . . . . . . . . 195
. . . . . . . . . . 195


Model Calibration and Verification


. . . . . . . 197


Model Calibration and Verification Schemes ... ........... 197
Model Calibration and Verification Results ........ .. .... 199
Model Sensitivity Analysis and Application .......... .... 219
Model Sensitivity Analysis and Application Schemes .......... 219


Simulation Results


. . . 223


SUMMARY


, CONCLUSIONS, AND RECOMMENDATIONS


Summary


a. . . . . . 239


Conclusions


Recommendations for Additional Research


Field Research


Mnln olinno R esrrh


* .U 72


. . . .. 88
. . . . 91


946


165









Groundwater Maps for K, C, and N Wetlands of 1993-


Listing of the FLATWOODS Model Source Code


1994


. . 248


. . . . . 258


REFERENCE LIST


BIOGRAPHICAL SKETCH ................... ................... ......
















LIST OF TABLES


pa~ge


Selected forest hydrologic models.


Site characteristics and instrumentation at the Gator Nationals Forest Site.


Site characteristics and instrumentation at the Control watershed (WS3),


Bradford Forest


Average soil moisture content (SMC %) and water table depth during June


1993


- January 1994 in two transects of the C system.


Model parameters describing the relations between soil moisture content and
water table depth. . . . .. ...... . ... . . .


Hydroperiod of cypress wetlands in 1992.

Hydroperiod of cypress wetlands in 1993.


* S S U S S S S S S S S S S S S S S S S

* S S S S S S S S S S S S S S 5 6 5 5 5 5 S S S


Annual duration of water table above critical depths in uplands (months).


Factors affecting water table response to storm events.


A listing of 10 wetlands and the interpreted conditions of water flow between


the wetland and the surrounding pine plantations.


Saturated hydraulic conductivity of flatwoods soils.

Deep seepage from cypress wetlands in 1993-1994.


* S S S S S S S S S a * S S S

* S S S S S S S S S S S S S S S S S S S


4.10.


Maximum source area of wetlands during January 1993


- April 1994.


S S


. .


Table









4.12.


4.13.


Wetland boundary parameters used in water balance calculation.


Calculation of monthly wetlands water balances with uncorrected field data.


4.14.


Annual wetlands water balances during May 1993


- April 1994 (mm).


4.15.


Monthly water balances for pine uplands during May 1993 to April 1994.


A list of model inputs required to run the FLATWOODS model..


A list of model outputs from the FLATWOODS model.

Data sources for the model calibration and verification.


. S S S S S S 4 S S S 4

. S S S S S S S S S S S S S S S


Simulation results for the calibration and verification period before treatments


(1992-1994), Gator Nationals Forest.


Simulation results for the calibration and verification period after treatments


(1994-1995), Gator Nationals Forest.


Summary of parameter values for the Gator Nationals Forest site calibrated by
the FLATWOODS model from pre-treatment and post-treatment data. ..


The FLATWOODS model calibration and verification results for the Bradford


Forest Watershed.


Summary of parameter values for the Bradford Forest watershed calibrated by
the FLATWOODS model from pre-treatment and post-treatment data. ..


Simulation schemes to study the potential harvesting effects in the first year
and long-term on pine flatwoods hydrology at the Gator Nationals Forest and


Bradford Forest sites.


5.10.


Simulated hydrologic effects of forest harvesting under three climatic


conditions in pine flatwoods.















LIST OF FIGURES


Location of the research sites in Florida

Topographic map of the Gator Nationa


. a s Forest site i. .n Alachua County, .
ls Forest site in Alachua County,


Florida.


Monthly precipitation in 1992, 1993,


(193


1994 at GNF and the long term average


-1992) of Gainesville, Florida.


. . . . 2 1


Geology of the Austin Cary Control Dome (Gillespie, 1976).


. . . . 23


Topographic map and experimental design at the Bradford Forest site in


Bradford County, Florida.


Hydrologic components in a wetland-upland system.


. . . . . .... 26


. . . . . . 2 8


A vertical view of instrumentation at the C wetland/upland showing transects
for soil moisture measurement, water table wells and a deep well. .....


The calibration curve for soil moisture measurements with TDR method.


Instrumentation in the C wetland in transects of shallow wells radiating from


the wetland, wetland margin, to the surrounding upland.


3.10.


Instrumentation in the K wetland in transects of shallow wells radiating from


the wetland, wetland margin, to the surrounding upland.


. 35


Instrumentation in the N wetland in transects of shallow wells radiating from


the wetland, wetland margin, to the surrounding upland.


An PvrmnlP nftkhi rlrl-nlk tinn nf thp ctnlrntPd rAnnhllr.tivitv (K determined


.. 20


Figure


page









the C wetland.


3.14.


The palmetto line is indicated by the arrow


Relations between water level elevation, and pond surface area and water


volume for the K wetland


The palmetto line is indicated by the arrow.


S SS '6


Relations between water level elevation, and pond surface area and water


volume for the N wetland.


The palmetto line is indicated by the arrow.


. Weekly average air temperature and cumulative rainfall for 1992,
1994 at the study site . . . . . . . .


1993,
S S S S 9


. and*

and
0 9 0 0 0 5


Relations between precipitation and throughfall for cypress wetlands and pine


uplands


Relations between canopy interception rates (Interception/Precipitation) and
precipitation for cypress wetlands and pine uplands ......... ......


0S5


Seasonal variation of interception rates for wetlands and uplands during 1993-


1994


A hypothetical soil profile in pine flatwoods showing the unsaturated water


zone, capillary fringe, water table, saturated zone, and soi


distribution


Soil moisture content (%) dynamics for the ES transect (#1).

Soil moisture content (%) dynamics for the NW transect (#2).


moisture content
. 0 4 0 0 S 5


. . . . 61

. . . . 63


Relationships between soil moisture content and water table depth in the ES


transect (#1).


Relations between soil moisture content and water table depth in the NW


transect (#2).


4.10.


Generalized relations between soil moisture content and water table depth in


a wetland-upland system.


. Water level dynamics in the C wetland and the surrounding upland from 01-


-1992 to 05-31-1994.


.. 73


.. 68









4.13.


Water table dynamics in the N wetland and the surrounding upland from 01-


01-1992 to 05-31


-1994.


4.14.


Frequency distributions of water level in wetlands.


. . 8


Frequency distributions of water table in pine uplands.


4.16.


. 4 . .... 83


A comparison of the water table fluctuation in the wetland, the wetland
margin and upland areas to show: a) the different responses to rainfall and
evapotranspiration; b) reversed water table gradients in the spring season.


Seasonal distributions of groundwater table in uplands.


4.18.


The overall spatial distribution of the groundwater table relative to the ground
surface; each data point represents the average water table elevation of three


systems during 1992-1994..


4.19.


4.20.


The spatial distribution of the groundwater table elevations from the wetland
margin to the pine upland in single systems. ...... .... ... .... .....


The spatial distribution of the groundwater table depth from the wetland
margin areas to the pine upland areas to show the drier conditions in pine


uplands.


Water table profiles in the cross section AA' to show both the spatial
distribution and seasonal dynamics of the groundwater table in a wetland-


upland continuum.


.. 94


4.22.


Water table profiles in the cross section, BB'


to show both the spatial


distribution and seasonal dynamics of the groundwater table in a wetland-


upland continuum.


4.23.


Water table hydrographs plotted from recording charts with a 2-hour time
interval to show the spatial and temporal responses of the groundwater tables


.. 95


to storm events and evapotranspiration. .....


.. 97


4.24.


Seasonal variation of the water table response (RWR) at two pine upland sites


and one wetland margin site during 1993-1994.


Analvqsi of the ornundwater table fluctuation with a soil water balance


84


92








4.26.


Two water table profiles in the K wetland to show the groundwater recharge
induced by reversed hydraulic gradients during the wet-dry transition season
(refer to Figures 3.10, 4.23 and 4.25). . . . . . . .


Topographic maps of the C, K, and N wetlands with the same contour interval
of 10-cm as water table maps in Figures 4.28, 4.29 and 4.30. .........


4.28.


Water table maps for the C wetland/upland system constructed from weekly
measurements to contrast water flow directions in a wet month and a dry


month.


4.29.


Water table maps for the K wetland/upland system constructed from weekly
measurements to contrast water flow directions in a wet month and a dry


month.


4.30.


Water table maps for the N wetland/upland system constructed from weekly
measurements to contrast water flow directions in a wet month and a dry


month.


4.31.


4.32.


Spatial distribution of the saturated hydraulic conductivity (K) determined by
the Bouwer and Rice method at the study sites. .............


Well coordinates for finite element analysis of groundwater flow direction and


velocity.


4.33.


Estimated groundwater flow direction and velocity of the solute movement in


two finite elements of the K wetland.


4.34.


Estimated groundwater flow direction and velocity of solute movement in


three finite elements of the C wetland.


4.35.


Estimated groundwater flow direction and velocity of solute movement in a


finite element of the N wetland.


4.36.


Nonlinear relationships between the hydraulic gradients and water table


elevations in finite element #2, K wetland .....


. Nonlinear relationships between the hydraulic gradients and water table
elevations in finite elements #1, #2, and #3, C wetland ..........









4.39.


The water table variation in the deep and shallow wells in 1993.


. 132


4.40.


Schematic of various runoff processes and their controls (adapted from


Dunne, 1978).


Hydrographs of surface outflow from wetlands from 01-01


1994.


4.42.


Surface runoff from wetlands increases with the expansion of the variable


-1993 to 05-30-


139


source areas. .


4.43.


Comparison of the available water storage in a upland and three wetland
systems under different water table conditions....... .... ......


Monthly water balances of the C wetland from May 1993 to April 1994.


4.45.


Monthly water balances of the K wetland from May 1993 to April 1994.
.. . . . . .. 150


Monthly water balances of the N wetland from May 1993 to April 1994.


S C 5 1


Annual hydrologic inputs and outputs of the three wetlands during May 1993


-April 1994.


4.48.


Annual hydrological inputs and outputs of the C wetland during May 1993


April 1994.


4.49.


Annual hydrological inputs and outputs of the K wetland during May 1993


April 1994....


4.50. Annual hydrological inputs and outputs of the N wetland during May 1993


April 1994.


. Estimated evapotranspiration of the C wetland and upland estimated by the


water


balance


method is


fitted


evaporation model


with


coefficient of 0.70.


4.52.


Double mass analysis to compare evapotranspiration of the C wetland and C
upland estimated by the water balance method and Pan evaporation model


with n eneffiRient nf A 70


. 136


149


. 153


. 155


161


162


------~~-~-~-~-~









0.70.


Structure of the FLATWOODS model showing the grid system and modeling


units of pine uplands and cypress wetlands.


Hydrologic components of the three submodels in the FLATWOODS model.
...-...-.-..... . ........ .. ... 178


Physical information of the modeling units of the Gator Nationals Forest


site. .


The FLATWOODS model calibration (April 6


- December 31,


1993) and


verification (January


1994


- May


1994) with measured average


groundwater table data during the pre-treatment period. .............


Simulated daily runoff and evapotranspiration from the Gator Nationals Forest
site during model calibration and verification periods under pre-treatment


conditions.


Model verification results (June


1994


- May


1994) with the post-


treatment data (after harvesting) using the calibrated parameters of the pre-
treatment period, showing an over-prediction of the water table level....


. The FLATWOODS model calibration with measured groundwater table data
from three separate treatment blocks during June 1 December 31, 1994.
The NW block (B 1), SW block 2 (B2) and SE block (B3) represent wetlands


clear-cut,


control


wetlands


uplands


clear-cut


treatment


. 207


respectively.


Simulated and measured average groundwater tables during the calibration


(June 1


- December 31, 1994) and verification (January 1


-May 31, 1995)


periods under post-treatment conditions.


Simulated daily runoff and evapotranspiration from the Gator Nationals Forest
site during model calibration and verification periods under post treatment


conditions.


5.10.


C11


Physical information of the modeling units of the Bradford Forest site.


Thl 1T ATWlfl'K mnelpl rnlihrntinn with me2Qnred rinnffla n arnindwnter








5.12.


FLATWOODS


model


calibration with measured runoff from


Bradford Forest Watershed during 1980-1982.


5.13.


The FLATWOODS model verification with measured runoff data (1983


1992) from the Bradford Forest Watershed.


5.14.


Simulated dynamics of the groundwater tables (a) and runoff (b) in a normal
year showing different treatment effects of three harvesting methods....


5.15. Simulated dynamics of the groundwater tables (a) and runoff (b) in a dry year
showing different treatment effects of three harvesting methods. .......


5.16. Simulated dynamics of the groundwater tables (a) and runoff (b) in a wet year
showing different treatment effects of three harvesting methods. ......


Comparison of treatment effects on annual average groundwater table
elevations; (a) annual total runoff, (b) annual total evapotranspiration, and (c)


under three climatic conditions.


TM1


= treatment 1


, wetlands harvesting


only;


TM2


= treatment


uplands harvesting only,


TM3


= treatment 3,


wetlands and uplands harvesting.


5.18.


Simulated long-term effects of forest clear-cutting on groundwater table


dynamics during a 15-year stand rotation.


The treatment is assumed to be


completed by January 1, 1978.


Simulated long-term effects of clear-cutting on annual average groundwater


table elevations during a 15-year stand rotation.


The treatment is assumed to


be completed by January 1, 1978.


5.20.


Simulated long-term effects of clear-cutting on daily runoff during a 15-year


stand rotation.


5.21.


The treatment is assumed to be completed by January 1, 1978.


Simulated long-term effects of forests clear-cutting on annual runoff and


evapotranspiration during a 15-year stand rotation.


The treatment is assumed


to be completed by January 1, 1978.


5.19.


229














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

MEASUREMENT AND MODELING OF THE HYDROLOGY OF CYPRESS


WETLANDS


- PINE UPLANDS ECOSYSTEMS IN FLORIDA FLATWOODS


By

Ge Sun


December 1995

Chairperson: Hans Riekerk
Major Department: Forest Resources and Conservation

The goals for this dissertation were to document the hydrology of cypress wetlands-


uplands


ecosystems


in north


central


Florida and


develop


a simulation


model


(FLATWOODS).


Three cypress wetland/pine upland systems embedded in a


flatwoods


landscape


were


extensively


instrumented


monitored.


Annual


average


canopy


interception rates on a monthly basis for cypress wetlands and pine uplands were 12% (8-


40%) and 14% (6-43%) of rainfall, respectively.

describe the high dynamics of wetland hydroperio


classified into four categories.


An quantitative index was presented to

d. Groundwater table fluctuations were


Shallow groundwater flow velocity was low (<10 cm/day)


due to low hydraulic gradients (<1.0 %), although the soil hydraulic conductivity was high


(i 5_lAn m/dAI


lrrnnmnwanter flnwc fnllnwpd the nverall tnnnolrnhic aradlients during the








water tended to recharge the surrounding uplands for some distance.


Surface runoff was


generated only from saturated variable source areas (VSA) associated with cypress wetlands


in winter.


The VSA increased 8-32% in a single storm. Monthly water balances showed that


rainfall and/or surface inflow were the major inputs to the wetland/upland ecosystems, while

ET and surface outflow were the main outputs.


The existing COASTAL model with the following modifications


was used to


construct the FLATWOODS model: 1) algorithms for the calculation of initial soil moisture

content in unsaturated zones were added; 2) algorithms for unsaturated water drainage from

unsaturated zone to the surficial aquifer were modified; 3) evapotranspiration procedures


were


modified


adding plant leaf area index


(LAI) and root components;


4) soil


evaporation was set as a function of groundwater table depth and LAI; 5) smaller grid


systems, dynamic aquifer parameters of


hydraulic conductivity and specific yield were


employed to account for the surface-groundwater interactions; and 6) surface flow routing


procedures for each modeling unit were added.


In addition, FLATWOODS was calibrated


and verified with field data of ground water table and daily runoff from two different

research sites.

Simulation by the FLATWOODS model showed that clear-cutting both pine uplands

and wetlands had the greatest effect on groundwater table and runoff, especially during the


first six years after treatment.


Following the regrowth of young plantations, the hydrologic


regime may recover to the pre-treatment conditions within about ten years. Partial harvesting













CHAPTER 1
INTRODUCTION


Research Background


In modem times, environmental quality has become an important component of

national concern. Protection of the environment and wise use of natural resources have never

been so frequently addressed by the governments as well as citizens, in developed countries


as well as in developing countries on this planet.


In the USA, the environmental influences


of forest changes were hypothesized by Noah Webster as far back as 1799 (Kittredge, 1948).


The book "Man and Nature"


, and its later version "The Earth as Modified by Human


Action" (Marsh, 1965; Marsh; 1884) has been regarded as the most comprehensive summary

of the viewpoints on human activities, including forest water relations, during that time.

With the advancement of modem technology, mankind has increasing powers to alter abiotic


as well as natural biotic systems.


For example, more mechanical and automated equipment


is available for silvicultural practices such as harvesting, reading and ditching.


Fertilizer,


pesticide, and herbicide use for vegetation management have become nearly routine not only


in agricultural fields but also on forest lands.


In Florida, tourism has been one of the major


sources of income for the State, and this industry has imposed a high priority and more








2

management practices with near-agricultural intensity (short rotation, clear-cutting, burning,


fertilization) have been adopted to gain more yield per unit land area.


Forest lands are no


longer seen as 'pure green' lands without any environmental problems, but are listed as one

of the sources of non-point pollution along with agricultural and urban lands (Binkley and

Brown, 1993). Since the 1970s, values of forested wetlands have been recognized in terms


of timber production, water quality protection and flood control, and wildlife habitat.


On a


global scale, wetlands may be significant in balancing nitrogen, sulfur, carbon and other


important


elements


biosphere


(Mitsch


Gosselink,


1986).


Protection


biodiversity as well as forest environments characteristic of unique ecosystems have been


major


additional


goals


recent


forest


ecosystem


management


perspective


(Greenburg,


1993).


Conflicting strategies over the same


lands have generated many


concerns among environmental protection agencies as well as forest industry companies

about the degradation of environmental quality, and consequently forest productivity in the

long run.


In 1975,


with the goal to evaluate the ecological effects of silvicultural practices


commonly employed in the southeastern United States, the Intensive Management Practices

Assessment Center (IMPAC) in Florida was formed in a cooperative effort by the University

of Florida's School of Forest Resources and Conservation (SFRC), the U.S. Forest Service's


Southeastern Experiment Station, and various members of the regional forest industry.








3

ecosystems, a research project with emphasis on the water flow relations between cypress

wetlands and slash pine uplands was initiated by the Soil and Water Sciences Department

and the SFRC with funding from the National Council of the Paper Industry for Air and


Stream Improvement (NCASI).


The author has assisted in both the IMPAC project and


NCASI wetland studies in field measurement and hydrologic model development since 1991.

Due to the hydrologic characteristic of high variability in space and time, the delayed

forest ecosystem response to disturbance, and the long life cycle of trees, it usually takes a


long observation time to make definitive conclusions.


Computer simulation techniques


provide alternative ways to help solving environmental problems (Freeze, 1971; Swift et al.,


1975;


Mercer and Faust, 1980; Decoursey, 1985).


Our approach to hydrologic studies of


flatwoods has been comprehensive, including both extensive field measurements in the

NCASI and the IMPAC sites for detailed information of each hydrologic component and its


processes, and computer modeling based on field parameters.


The ultimate goals for


developing a simulation model were to achieve improved understanding of forest ecosystem

and provide a scientific basis for forest water management.


Objectives


This dissertation study focused on hydrologic dynamics of the shallow groundwater


table and evapotranspiration


in pine flatwoods landscape emphasizing surface-subsurface








4

forest management practices on wetland/upland systems, and therefore to provide basic

information for forest ecosystem management in pine flatwoods.

Hypotheses were as follows:


Pine


flatwoods


are heterogeneous


hydrologic


systems;


cypress


wetlands


have


significantly different water regime from pine uplands.


Pine flatwoods are storage-based hydrologic systems;


groundwater tables are controlling


factors for surface flow in pine flatwoods and surface runoff is generated by the

Variable Source Areas (VSA).

3. Rainfall and evapotranspiration are the major water input and output respectively, of pine

flatwoods ecosystems.

4. Water movement in pine flatwoods ecosystems can be modeled adequately by a coupled

surface-subsurface forest hydrological model.

Specific tasks were as follows:

1. Determine the surface-subsurface water flow pathways and variable source areas.

2. Quantify the water balances (precipitation, canopy interception, percolation,

evapotranspiration, surface runoff and subsurface flows) of wetland-upland

ecosystems;


Develop/adapt


a distributed,


multi-dimensional,


process-based


hydrologic


simulation model for cypress wetland/pine upland systems in the flatwoods











Structure of the Dissertation


As described in the objectives of this dissertation, it essentially includes two integral


parts; experimental data analysis and model development and testing.


Chapter


Literature


Review,


identifies research needs


in flatwoods


hydrology,


both


with respect to


experimentation and computer simulation.


Chapter 3,


Experimental Methods, introduces


the research sites where experimental data were collected for model development, and


instrumentation and data analysis methods. Cha

for each process (precipitation, soil moisture,


.pter 4 presents the field experimental results

water table dynamics, groundwater flow,


surface water flow and deep seepage) of the complete hydrologic cycles of three replicate


wetland-upland systems.


The hydrologic processes were first analyzed separately and then


combined by the water balance equation, and as a result, monthly water budgets were


established for wetlands and uplands respectively.

the new computer simulation model FLATWI


Chapter


O0DS,


introduces the theory behind


which is a modification of the


published COASTAL model.


Then the model calibration and verification results based on


field data from two experimental sites, Gator Nationals Forest and Bradford Forest are


presented.


The model sensitivity analyses combined with model application to several


scenarios of common forest management practices are presented in the last section of


Chapter


In the Appendices, detailed water table maps and the FLATWOODS model














CHAPTER


LITERATURE REVIEW


Past studies relevant to flatwoods hydrology by experimental methods and selected

modeling efforts for this landscape have been discussed in this review.


Experimental Field Studies


Wetland Hydrology Studies


Hydrogeology of cypress wetlands has been extensively studied by the Center for

Wetlands and the Geology Department at the University of Florida over the past two

decades, to investigate the feasibility of waste water treatment by wetlands (Cutright, 1974;


Smith, 1975; Gillespie, 1976; Ewel and Odum, 1984).


Those workers concluded that the


water table aquifer and surface water of study sites in Alachua County were hydrologically


isolated from the Floridan Aquifer.


Furthermore, it was suggested that the major water


discharge


from


wetlands was through evapotranspiration, and that wetlands could be


recharge areas during wet periods but discharge areas during dry periods.


of hydrology


Also, the integrity


geology and biology of cypress domes and surrounding upland forests was


recognized.









7

although most of the domes investigated were not under typical conditions due to the


addition of artificial water sources.


Surface water in cypress domes was concluded to be


closely coupled with groundwater in surrounding pine uplands, and cypress swamps could

be recharged or discharge water depending on the position of the surrounding groundwater


table relative to the surface water level.


Three groundwater flow patterns between ponds and


their surroundings were postulated, namely flow-in, flow-out and flow-through, but only the


last two types were documented.


Riekerk (1992) employed a heat-pulse groundwater flow


meter to directly measure groundwater flow velocity at the margins of cypress ponds and

found tremendous variation both in space and time possibly due to fractal effects of forest


soils.


Although this method has an advantage over the Darcian gradient method in that it


does not need the parameter of hydraulic conductivity, significant measurement errors may


mask the relative low flow rates of groundwater in these systems.


Phillips et al. (1993) and


Phillips and Shedlock (1993) investigated the hydrologic relations between surface water and

groundwater of small seasonal ponds in a forested Coastal Plain drainage basin in Delaware.

They concluded that the groundwater table was not a subdued expression of land surface


topography,


wetlands


could


behave


as seasonal


recharge


or discharge


areas.


Hydrochemical sampling showed that the water quality of shallow seasonal ponds was

strongly influenced by the adjacent groundwater systems with the upland-wetland margins


as the most dynamic part.


Winter and Carr (1980) documented flow-through type wetlands










which average annual rainfall was 1510 mm.


was lower than that from uplands and therefore


groundwater recharge (Ewel, 1990).


It was implied that ET from cypress wetlands


cypress ponds were beneficial for

using daily hydrographs of water


Estimation of ET


levels showed that the total water loss from cypress ponds was slightly lower than that from


pine uplands (Ewel, 1993).


Brown (1981) surveyed several cypress ecosystems in Florida


to investigate how inputs of water and nutrients influenced the structure and functions of

cypress wetlands. It was found that water loss from cypress forests increased with increasing


phosphorus inputs.


It was also found that transpiration rates from scrub cypress forests (1.0


mm/day) and cypress domes (3.1-3.8 mm/day) were lower than evaporation rates from open


water bodies while those from the floodplain forests (


.6 mm/day) were similar.


This


suggested that cypress domes conserved water and used water more efficiently.


Rushton


(1994) reported that annual ET from a marsh with vegetation consisting of mainly pickerel


weed and arrowhead in South Florida approximately equalled annual rainfall.


Monthly ET


was close to 80% of pan evaporation in the Spring and Fall but higher in the Summer,


suggesting more water loss from wetlands than from open lakes without vegetation.


Lake


evaporation is often taken as 0.7*Class A pan evaporation measurements (Veihmeyer, 1964).

By this definition, evapotranspiration from five experimental tanks covered with different


aquatic


plants,


Panicum


regidulum,


Juncus


effusus,


Carex


lurida,


Alternanthera


philoxeroides, and Justicia americana,


were found


0.94 times








9

intermediate areas would fall into a regulatory wetland definition due to a shortage of


quantitative hydrologic information (Comerford et al., 1995; Segal et al.,


1987).


In general,


wetland hydrologic information is relatively scarce compared to the other two criteria of


vegetation and soils for wetlands delineation (Best et al.,


1990).


Wetland hydrology, the


primary driving force in the formation and maintenance of wetlands, has been stressed as


basic


to understanding,


quantifying and


evaluating wetland


functions and ecological


processes (Ivanov, 1981


Gopal et al.,


1982; Rykiel, 1984; LaBaugh,


986; Gilvear et al.,


1993).


Carter


etal.


(1979) and Carter (1986) identified five areas of research for wetland


hydrology studies:

1. The need for improving, refining, and perhaps simplifying existing

techniques for hydrologic measurements;

2. The need for making accurate measurements of all the hydrologic inputs

and outputs of representative wetland types and estimating the errors inherent

in various measurement techniques;

3. The need to improve basic understanding and quantify the soil-water-

vegetation relationships of wetlands;

4. The need to make detailed, long-term studies of different wetland types

under different environmental conditions;

5. The need to continue developing models based on hydrologic data for a










Upland Hydrology Studies


There is more literature on upland hydrology than wetland hydrology.


The study by


Asmussen and Thomas (1974) in a lower coastal site in Georgia documented overland flow


as only 2% of the annual rainfall.


The shallow phreatic aquifer was therefore a controlling


factor affecting the behavior of the Coastal Plain streams.


Sainju (1982) concluded that the


surface


groundwater


was


relatively


uniform


varied


associated


topography, and that pine trees could benefit from groundwater for most of the studied area


of pine flatwoods in north central Florida.


In a similar site, Phillips (1987) studied soil water


movement in a young flatwoods forest stand on a similar site and recognized an spatial


heterogeneity of water tables due to even small topographic variations.


This study suggested


that two-dimensional models instead of one-dimensional models were needed to better

represent water pathways in flatwoods. In a study on the hydrologic effects of adding treated

sewage effluent in cypress ponds on surrounding soil moisture and nutrients, Okorie (1976)


found


to 20 m from the pond edge of the treatment site had significantly higher soil


moisture content at 30.5 cm depth, than the control (no sewage water). He suggested that

the pond water treatment had significant influence on upland soil moisture. Most forest


ecological studies have included hydrological components, because transpiration by plants

and availability of soil water and nutrients are closely related to tree growth and forest

productivity (Wickramasinghe, 1988; Gholz et al., 1990; Sands and Mulligan et al., 1990;










was recorded for a 25-year old slash pine stand in Florida (Golkin, 1981).


Butler et al.


(unpublished data, personal communication) compared estimates of ET from a mature slash

pine forest by different methods including the Penman-Monthieth equation, Priestley-Taylor


equation and a water table drop (WTD) method.


ET estimated by WTD method generally


was within the range of other methods, but the uncertainty of deep drainage and lack of soil

moisture data in the unsaturated zone caused significant errors, especially when the water


table was deep.


Tremwel and Campbell (1992) investigated the hydrology and surface water


quality of four pasture sites characteristic of Spodosols in the Okeechobee Basin during


1989-1991.


Tremwel (1992) reported that the average surface runoff varied from zero (the


plot with deep sandy soil) to 22% with an average of 11% of total rainfall.


Annual ET


calculated by the Penman-Monteith equation varied from 74% to 90% of rainfall, which


amounted to 110 cm/year on the average. Capece

by standard, widely used mathematical models.


(1984) studied the storm flow mechanisms

He concluded that flatwoods hydrologic


systems were storage-based and that the groundwater table was crucial to predict runoff


volume.


Results of tracer experiments in pasture flatwoods showed that groundwater flow


velocity was in the range of 1.9-12.0 cm/day and surface runoff dominated the transport

mechanism for the relatively flat sites (< 0.3% land slope), while the groundwater flow was


primary mechanism at the sites with higher slopes (Capece, 1994).


Integrated Wetland-Upland Studies








12

hydrologic effects (water quantity and quality) of intensive forest management have been


carried out for slash pine flatwoods systems since


1978 by the Intensive Management


Practice Assessment Center (IMPAC) of the University of Florida (Riekerk et al., 1979;


Rodriguez,


1981; Pratt,


1978).


Three small artificially diked watersheds (Max., Min.


disturbance and Control) were monitored for 17 years collecting weather, runoff,

quality and water table data at the Bradford Forest, 50 km north of Gainesville, FL.


water

At the


Gator Nationals Forest, 15 km north of Gainesville, FL, six small watersheds (plots) have

been monitored to study the hydrologic and water quality responses to annual fertilization


(Riekerk and Korhnak, 1984).


At the University of Florida Austin Cary Forest,


13 km


northeast of Gainesville, FL, two small watersheds were used as demonstration of forest


management effects.


Overall, it was documented that clear-cutting of 40-year-old flatwoods


forests significantly increased runoffto 150 mm or 150% of predicted runoffin the first year.

Water table levels were higher after harvesting due to reduced ET, and more pronounced


effects were found when the water table was deeper in dry years (Riekerk, 1989).


increase dropped to 65% in the 6t year.


The runoff


Annual average ET over a 5-year period from a


young slash pine flatwoods watershed was estimated by an actual/potential ET ratio method


as about 1050 mm, which was 82% of the total rainfall (Riekerk, 1989).


However, these


studies designed to evaluate the cumulative environmental effects of management practices

and have given little attention to the heterogeneity of the system and hydrologic processes










Forest may


provide in-depth information on hydrologic dynamics


in pine


flatwoods,


especially


interaction


between


wetland


surface


water


upland


groundwater.


Preliminary results showed that shallow groundwater flow generally followed the landscape

gradients and most of the cypress wetlands were flow-through types, though local ground

water flows may change direction during the annual dry-wet cycles (Crownover et al, 1995;


Comerford et al., 1994; Sun et al., 1995).


The implication of these experimental results was


that management practices around wetlands may not be as problematic as surmised in terms

of chemical loading to wetlands due to low groundwater flow rates and small source areas.

A two-year study was focused on evapotranspiration of the slash pine plantations and

associated cypress wetlands, and suggested there were no significant differences between

these two ecosystems, but the composition of evapotranspiration (evaporation, transpiration

and canopy interception) varied significantly between cypress wetlands and pine uplands


(Riekerk et al., 1995).


This study questioned the proposition that cypress wetlands conserved


more water than uplands, and suggested that


high variability of ET


existed in cypress


wetlands ecosystems. Different stand structures of pine plantations could also contribute to

this variability.


Simulation Models


Although numerous forest hydrologic models currently exist in the USA and in the








14

are relevant to the present study were compared and have been summarized in Table 2.1.

A simulation effort was made by Guo (1989) to test the applicability of the simulator

VSAS2 to the pine flatwoods watershed at Bradford Forest for the prediction of stormflow.


simulation of five storm events showed that the original simulator significantly


overestimated the peaks of the stormflow hydrographs.


It was suggested that pond storage


in this flat and complex landscape played a significant role in the runoff generation.


Some


models specifically developed for simulating high water table conditions in agricultural


landscapes, such as CREAMS-WT (Heatwole et al. 1987), DRAINMOD (Skaggs, 1980,

1984), and FHANTM (Tremwel and Campbell, 1992), were applicable to horizontally


relatively homogenous landscapes.

forested flatwoods is not known.


How those models perform in spatially heterogeneous

Models developed to study wetland hydrology often


included only the wetland itself without surrounding upland components (Hammer and


Kadlec, 1986; Scarlatos and Tisdale, 1989; Kadlec, 1993b).


Current ecological models to


study water and nutrient or carbon fluxes in pine flatwoods employed a lumped approach,

and placed significant emphasis on the vegetation components with various simplification

of the soil components and water pathways in the ecosystems (Golkin and Ewel, 1985; Ewel


and Gholz, 1991).


A series of models have emerged to investigate the interactions between


surface water and groundwater with various simplification in the evapotranspiration and


unsaturated


water movement in


hydrological


systems (Winter,


1976;


Cheng and


- --- --















Table 2.1.


Selected forest hydrologic models.


NO Title Author Distinct features


SHE (Systeme Hydrologique
Europeen)

WSHS (a mathematical model for
the watershed hydrologic system)


VSAS2


Peatland Hydrologic Impact Model-
new version


BROOK2 (BROOK90)


PROSPER


CREAMS-WT


VS2DT (Variably Saturated, two
Dimensional Transport)

DRAINMOD-forest hydrology


Abbott, 1986


Al-Soufi, 1990


Bemier, 1982


Barten,


Federer, 1983, 1

Goldstein, 1972

Heatwole, 1987


Lappala, 1987


McCarthy, 1990


Distributed physically based to simulate effects of human
interference

3-D, physically based for soil-water-plant continuum


Distributed, process based, stormflow. To test Variable
Source Area Concept.
Lumped, physically-based, peatland, upland hydrology
simulated

Lumped forest hydrology model for the eastern USA

Lumped, process mixed soil-water-vegetation model


Lumped, modified from CREAMS to apply to high
table condition in Florida


water


Physically based. Un- and saturated water and solute
transport. It is being modified for wetland-upland system.

Lumped, subsurface drainage and runoffwith high water


version


MODFLOW


UNSAT2 -modified version


A Distributed Numerical Model for


Watershed Hydrology

COASTAL-a distributed


hydrological simulation model for
lower coastal watershed in Georgia
Hydrological Modeling Using
Variable Sources Areas

Hydrology-Vegetation Model


ANSWERS-forest hydrology version


SWIF (Soil Water in Forested
Ecosystems)


Integrated Surface Water and
Groundwater Model


McDonald and
Harbaugh, 1988

Neuman, 1974
Ahmad, 1991


Sabur,


Sun, 1985


Whitelaw


Wigmosta,


Thomas,


Tiktak,


Yan and Smith,


3-D distributed, physically based groundwater flow model


2-D, un- and saturated finite element model; 2-D root
distribution function included

Distributed, physically based for low relief, infiltration
dominated areas


& ET


2-D, most physically based, distributed model for flat terrain
with high groundwater table; Coupling of MODFLOW and
BROOK.


3-D, physically based, upgraded from


VSAS2


and SHE


Distributed, ET emphasized, surface-subsurface flow, large
complex watershed to evaluate effects of the climate change


Distributed, nonpoint pollution


Lumped, soil moisture simulation for forest ecosystems
lowland areas


Distributed, modification of MODFLOW and South Florida
Water Management Model










for comprehensive large-scale (watershed to


global) ecosystem studies,


which the


hydrology is one of the most important components (Pierce et al., 1987), and accelerated by


the increase of computation power.


Another factor may be that more spatial field data are


available with the aid of Geographic Information Systems (GIS).


The Florida flatwoods


landscape includes a mosaic of cypress wetlands and forest uplands, so the hydrology of


flatwoods is inherently complex.


Examples are the slight spatial changes in topographic


elevation causing significant changes in the water regime, and obstructive soil layering


because of the spodic and argillic horizons in the soil profile.


The heterogeneous vegetation


cover of wetlands and uplands and associated phenology may further complicate the


interactions between surface water and groundwater.


Due to the complex geologic


formation of flatwoods, the preferential water pathways in this system have not been either


well documented or understood.


A new distributed flatwoods forest hydrologic model is


needed to study the hydrologic processes of wetland/upland systems and provide a tool for

water management, specifically for this landscape. It is worthwhile to mention that the South


Florida


Water Management District (SFWMD)


has proposed a conceptual integrated


hydrologic model for Dade County in south Florida characterized by flat topography, sandy

soils, a shallow groundwater table, and well developed canal systems (Yan and Smith, 1994).

The model was intended for use in regional planning for water supply, but no further action

has been taken (Yan, personal communication, 1994).















CHAPTER 3
RESEARCH APPROACH


Basically


research


approaches,


experimental


field


studies


computer


simulation methods,


were used in this study.


As in most cases, on the one hand, field


experiments generate more realistic data but are limited by manpower and financial resources

for studies of large-scale systems, while on the other hand, computer modeling methods are


less costly but often restricted by available real data for calibration and verification.


design of this study is a compromise between these two approaches.


methods are introduced in this chapter,


Chapter


Experimental field


while simulation methods will be discussed in


along with model development.


Physical Setting. General Experimental Design and Instrumentation


Two research sites were selected as the basis for data collection for this study of

typical flatwoods landscapes in north central Florida. One was the Gator Nationals Forest,


located 1


km north of Gainesville in Alachua County (Figure 3.1).


This 42-ha flatwoods


site has been studied by several projects by different disciplines (Comerford, et al. 1994;


Riekerk et al.


,1994).


Another site is the Bradford Forest on industrial lands in Bradford


Ca. *n~q r l'n. .4 CAirn .rr t i. '


r'L .L





















BRADFORD COUNTY


-4.














I
p
I
I

I
/
I *

I,
I
I


a-e










evaluate environmental effects of different forest harvesting methods (Riekerk, 1989).


this dissertation, wetlands refer to cypress domes or cypress ponds delineated by the saw


palmetto line, while uplands refer to the surrounding thinned pine plantations.


The saw


palmetto line was defined as the distinct vegetation break from wetland vegetation to saw


palmetto and slash pine plantations.


line out to about


The wetland margin area extends from the saw palmetto


m into the lower pine upland areas.


Gator Nationals Forest (GNF) Site


This site was located in the plateau physiographic region in Alachua County where


Plio-Pleistocene


terrace


deposits


Hawthorne


Formation


dominate


geology.


Topographical slopes ranged from 0 to


1.6% (Figure 3.2).


Surface flow from this site


generally was facilitated by ditches leading into interconnected wetlands which drained into


the Santa Fe River to the north of Alachua County.


The average annual temperature was


21C, with a mean monthly low of 14 C in January and high of 27 C in July.


Average


annual rainfall was about 1330 mm with dry periods during the spring and the fall.


average monthly rainfall patterns during the study period from 1992 to 1994 and the 61-year


average have been presented in Figure


A 50*50 meter grid system was overlaid over


the entire 42 ha experimental area and each grid point was marked and labeled with a steel

post and related to a reference coordinate (0, 0) set by an arbitrary elevation of 30.48 m (100

























"""shff""""'


Flume


-1 0 1 2 3 4 5 6 7 9 10 11 1213 14
West-East











350

300

250
E
E 200

.S150

100

50

0


Annual Rainfall:


Year 1992
Year 1993
Year 1994
Long term


= 1552 mm
= 1103mm
= 1307 mm
= 1330mm


1 2 3 4 5 6 7 8 9


Month


Figure 3.3.


Monthly precipitation in 1992,


1993


,1994 at the GNF site and the long term average (1932-19


of Gainesville, Florida.


S1992 1993 E 1994
-a- Long term Three-year avg.


Cr









22

detailed description of the site and instrumentation of the grid system can be found in the

annual research reports submitted to NCASI (Comerford, 1993).

Approximately 35% of this 42-ha site was in cypress swamps with sizes ranging from


a few square meters to more than


plantations.


ha, while the remaining upland areas were in slash pine


The 27-year-old upland plantations were 5S-row thinned in 1986 to a stem


density


of 500 stems/ha (Riekerk et al,


1992).


Pond cypress (Taxodium ascendens)


dominated the wetland species along with slash pine (Pinus elliottii) and black gum (Nyssa


sylvatica var.


biflora).


The dominant tree canopy in uplands was slash pine with an


understory of saw palmetto (Serenoa repens) and gallberry (Ilex glabra) shrubs.


geology of this site has

drilled in the ponds. Ho


not been investigated in detail although several bore holes were


wever, the geologic characteristics and formation of cypress domes


at similar sites in the region were extensively studied to determine the applicability of


wetlands for waste water treatments (Cutright, 1974; Smith, 1975; Gillespie, 1976).


Austin


Cary


Control


Dome


located


southeast


GNF


was


cited


representative of the geologic characteristics of a cypress wetland and its surrounding (Figure


3.4).


It was suggested that the surficial aquifer consisted mainly of sands, and that the


underlying 60-m thick Hawthorn Formation of white dolomite with phosphate pebbles and


gray-green clays isolated it from the deep karstic Floridan Aquifer.


At the GNF site,


impermeable blue-green clays approximately


m below the bottom of


the wetlands






















CROSS


SECTION


THE


AUSTIN


CARY


CONTROL


DOME


SII


a p-- -** - a- a2..
ar a **


-a-


1L


-A-.


sat urn m~lnms


p ucta


I oTLL UMWI SS


-

-a-


a*


d2r


6.6 g~ I.


_. r, -_*,.- a.^ -.?i,
S a a a n.--..a
*-T -" -*" a* a --t
* .- a - f': a *.**'E


* 3


* a -~ a a -t
a a -
--- -------
- a----- pS87,jJ~
- - -


-3 -U


-a-


manY


I .am UUAgt


RAO C t sU.C.UY
*m Ms -Y mMO


K*d AM9 vttag
MANGa CLAW


-a-


* T1


-U


-a-


Wan
use's


I f nflnt MoISt



n~vest,, Ms
a1 ITn LoIf


WSCCWFOAmflY


Figure 3.4


Geology of the Austin Cary Control Dome (Gillespie, 1976).


'4

*.


_ r_ ~~I ~ I


- -- --











Table 3.1.


Site characteristics and instrumentation at the Gator Nationals Forest site.


System K C N

Surface areas of wetlands ha 0.55 0.64 0.46

Maximum difference in elevation 42 39 33
between the wetland bottom and
upland (palmetto line), cm

Number of shallow wells used for 27 31 17
weekly data collection

Number of water table recorders 3 2 1
used for continuous data collection

Elevation at the pond recorder (m) 29.468 29.61 28.124

Elevation of the palmetto line (m) 29.89 30 28.449

Elevation at the upland recorder (m) 30.2 29.718 28.932


' Measured along the palmetto line; : the elevation at the margin recorder


30.108 m.


poorly drained. The soil type was primarily a typical Spodosol (Pomona fine sand, Utic

Haplaquods; sandy, siliceous, thermic).


Within this site, three representative cypress wetland/pine systems, namely C,


were extensively instrumented (Table 3.1).


K, and


This work reports the detailed studies of the


three systems since most of the process-related data were collected and analyzed from these


by the author.


Data from Bradford Forest site were collected by Korhnak and Riekerk


(Personal communication), and


used mainly for model calibration and verification.


Bradford Forest (BF) Site









25

that of the GNF site, but the 0-4 m thick clay aquitard below the surficial water table aquifer

was more leaky as suggested by Pratt (1978).


In 1978, three watersheds (WS 1


WS2, and WS3) designed to evaluate the effects of


different treatments of low disturbance, high disturbance and control were created artificially

by enhancing the pre-existing drainage divides with perimeter roads and ditches (Figure.


Treatment experiments were started in April 1979.


water level recorder in the main cypress pond and one in the


Each watershed contained one

upland. All shallow wells and


deep wells were monitored only during the first year of the project, but selected wells were


monitored continuously.

recording flume. Air tei


Runoff from each watershed was measured by a long-throated


mperature and rainfall data were collected from a central weather


station and network of seven rain gauges (Riekerk, 1989).


Brief information about the


control watershed (WS3) has been described in Table


Table 3
Forest.


Site characteristics


and instrumentation at the Control watershed (WS


), Bradford


Total surface area, ha


Surface area of cypress wetlands:


Maximum difference in elevation between
the wetland bottom and the upland, cm

Number of shallow wells (1 m deep)

Number of deep wells (7 m deep)












Bradford Forest


Watershed
Boundary

Ditch

Weather
Station


F'-#


Ground Surface Contours,
in Feet Above Mean Sea
Level

500 Feet


S R. 100


Figure 3


Topographic map and experimental design at the Bradford Forest site in Bradford County, F


-145-


t











Methods for Field Data Acauisition and Analyses


As stated earlier, this dissertation only includes field data from the GNF site, and the


following discussion is specifically for this site.


The guide for investigations of different


hydrologic components was the water balance, but focused on wetland-upland interactions

for the present study (Figure 3.6).

The water balance of a watershed can be expressed as Equation 3.1 for any time scale

from an hour to a year:


xA + Si + G,


= A(V+V


+VJ


-ET


= A( f,(wl) + f2 (wl) + f3(wl, 0))


= change in water storage (L3);


= net precipitation


= precipitation (P)


- interception by vegetation (P,) (L);


= surface area of the target system ( L


i = surface inflow across the system boundaries (L3);


= ground water inflow or subsurface inflow across watershed boundaries (L3)


' = evapotranspiration including soil and pond surface evaporation and transpiration from
vegetation (L);

= surface outflow across watershed boundaries (L3);

= lateral ground water outflow or subsurface outflow across watershed boundaries (L3);


o-G,
































Transpiration


Rtaif~l


Groundwater
Inflow


Deege
Seepage


\nd~rr'










= volume of water in unsaturated soil (L3);

= water table level in the system (L);


= soil moisture content in the unsaturated zone;


fl, f2 and f3 are functions of water level or soil moisture content (6) in the unsaturated zone.

This equation may be used in any system as long as the boundary conditions are


determined.


For a flatwoods system, surface flow and groundwater flow between wetland


and surrounding upland represent one way of the energy and matter interactions of these two


ecosystems.


The change of surface water area in wetlands during a storm event or from


season to season reflects the variable source of storm water runoff.


Data Acquisition


Net precipitation (P.)

A standard weather station was established near the K pond, and hourly rainfall,

average air temperature, humidity, and solar radiation data were recorded by a data logger.

One standard rain gauge was placed in an open area adjacent to and one gauge inside of each


pond to manually collect weekly rainfall and throughfall.


Ten 10 cm wide and


cm long


throughfall troughs made from polyvinyl chloride half-pipes were installed in each wetland


and upland plot to determine rainfall interception by forest canopies.


Six additional troughs


were placed under the palmetto layer to account for understory interception.


Throughfall










distribution as well as by groundwater table depth.


Soil moisture content was measured


using Time Domain Reflectometry (TDR) method (Topp, 1980; Ledieu et al., 1986) during


June 1993


- December 1993 on a weekly basis.


The objective of these measurements was


to establish a relationship between water table depth and volumetric water content of the


unsaturated zone for both uplands and wetlands.


This relationship will be used to determine


initial soil moisture conditions for the simulation model, and for ET estimation by the water

balance method.

Two transects were established from the wetland into the upland with measurement

spots at 25-m intervals. At each spot, paired steel rods were pushed laterally into a vertical

soil pit wall as sensors for direct measurement of volumetric soil water content by the TDR


method (Figure 3.7).


The number and layering of rods depended on the clay depth and water


table conditions of the particular sites.


During the excavation of the pits that were 0.5 m


wide and 1 m deep, intact soil samples for each layer were collected with copper rings

in diameter and 4 cm in height for analysis of soil physical properties.


The equation for the calculation of volumetric soil moisture content from


TDR


readings was adopted from Topp (1980):


8= -0.053 + 0.0292


- 0.00055


x K2 + 0.0000043xK3


where, K is the reading from the TDR equipment.

Calibration of the above equation was made by a gravimetric soil moisture sampling





























Groundwater table


Wetland shallow
well


Figure 3.7.


Wetland
Margin


Cypress Pond


Pine Upland


Soil moisture content


sensors


Deep well


Upland shallow
well


.. i-&


A vertical view of instrumentation at the C wetland/upland showing transects for soil mois


measurement, shallow water table wells and a deep well.




































10 20


SMC by


TDR (X)








33

A water table well was installed down to the clay horizon nearby each spot for soil

moisture content measurement to measure the water table level during each visit.

Groundwater flow (GjGo)

More than 20 shallow groundwater wells were installed for each of the three selected


wetland-upland systems to determine groundwater flow gradients.


each measurement, was constructed with a 10-cm contour interval.


The water table map for


cm diameter and


m long polyvinyl chloride (PVC) wells were installed in holes made with a 5-cm


diameter hand auger.


The lower 1 m of the pipe had well-screening and was closed at the


bottom.


In most cases, a clay layer was encountered about 1


m below the ground surface.


In each wetland, a 8-cm diameter hole was dug with a power auger and a 5-cm well with a

1-m long screen installed at 7 m depth to determine deep seepage from the perched water


table to the secondary aquifer as sketched in Figure 3.7


. Soil samples obtained during the


drilling process were brought to the laboratory for examination and re-construction of the


geologic profile. Four to six 50-m transects w

wetland edge to the upland were established


wells at 25-m intervals radiating from the


for each system (Figures 3.9, 3.10, and 3.11).


The water level near the center of each wetland was recorded continuously with a 15-minute


punch tape recorder.


The groundwater table changes in an adjacent upland well about 50 m


from the wetland margin was recorded on weekly charts with a Stevens Type F recorder.

Water table data were collected since January 1992 for the K and N wetlands, and March










50*50 m grid point
Contour line
Weather station
Flume
Water level recorder
Shallow well
Soil moisture measurement point
Transect for water table profile


25 50 75 100 125 150 175 200 225


-East


Figure 3.9.


(m)


Instrumentation in the C wetland in transects of shallow wells radiating from the wetland, wel


margin, to the surrounding upland.


,_ B'


.25

V




250













200


175


150


125


100


75


50


25


0


50*50 m grid point
Contour line
Weather station
Flume
Water level recorder
Shallow well
Soil moisture measureme
Transect for water table j


25 50 75 100 125 150


- East


(m)


gure 3.10. Instrumentation in the K wetland in transects of shallow wells radiating fromthe wetland, wetland ma


to the surrounding upland.






















1ooI~


50*50 m grid point
Contour line
Weather station
Flume
Water level recorder
Shallow well


Soil moisture measurement
Transect for water table pro


50 75


-East


Figure


(m)


Instrumentation in the N wetland in transects of shallow wells radiating fromthe wetland, wetland ma
I. 1 l


to me surrounding upland.


z-











system.


Since the water table in wetlands varied dramatically during the year, the boundary


for groundwater flow calculations as well as water balance was delineated by the minimum

polygon which can cover the wetland surface area during the wettest season.

The method to calculate groundwater flow quantity was based on Darcy's equation

(Fetter, 1988):


= KxA,xI,


(3.4)


N


where,


= groundwater flow recharge/discharge through the vertical


wetland


boundary segment i (m3/day);

= hydraulic conductivity of the water table aquifer (m/day);

= the ith cross-sectional area of groundwater flow (m2), product of the


length


a segment of the


boundary


line (m)


the aquifer


thickness (m);

Ii = hydraulic gradient (m/m) at a right angle across the boundary segment,
positive from the upland to the wetland;

T = time interval for the calculation (day);

G = total groundwater flow amount into area of a wetland during the time
period of T days (mi3).


(3.5)











the segments.


Normally, four measurement periods were used for calculation of monthly


averages.

A finite element analysis method was employed to further explore the dynamics of


water flow direction and velocity at certain particular points of interest (Keen, 1992).


In the


finite element theory (Segerlind, 1976, 1984), the hydraulic gradient of a two dimensional

domain can be expressed as:


bb
cc


(3.6)


where, the determinant 2A is defined as:


and X and Y are the coordinates of any three points i, j, and k of a triangular element.


the study of groundwater flow, these three points may represent well locations.


coefficients a, b c are calculated from the X and Y values.


b,=Y


=Xk


bj=


1










Writing the differential


gradient matrix


with two discrete equations,


flow


direction and hydraulic gradient (If) (m/m) can be evaluated by:


Ah2
(-Ax
Ax


Alh2
S(-A)
Ay


(3.7)


The average linear velocity with which water actually moves through pore spaces was

defined as (Fetter, 1988):


K cdh
ndl

where,


(3.8)


= groundwater velocity (cm/day);


= effective porosity;

= saturated hydraulic conductivity (cm/day);


If = dh/dl


= hydraulic gradient (water table gradient) (m/m).


The flow direction a (0-360 degrees counter clockwise) was defined by the three well


locations, i, j, and k.


It was expressed as:


a= arctan [ (Ah/Ay)/(Ah/Ax) ]


(3.9)


This finite element method provided a convenient way to study the variability of

water flow properties over a continuous time period if three water table recorders were used,

as was the case for the K wetland, or for different seasons for those locations where water

table data were collected on a weekly basis.









40
water table across the wetland-upland systems.

The saturated hydraulic conductivity (Ks) was determined in situ by the Bouwer and

Rice method (Bouwer and Rice, 1976; Rice, 1989). Although this method may have a 10% -

25% measurement error, its accuracy has been justified as being acceptable and has been


widely used in estimating K, for unconfined aquifers (Brown and Narasimhan, 1995).

expression of K1 was derived as:


r, Ln(R/r)
r w


- In(-)


(3.10)


where,

Ks= saturated hydraulic conductivity (LT');

re = inside radius of the well (L);


= the length of well perforation (L);


Re = the effective radius over which y is dissipated (L);

r, = well radius (L);

Yt = the vertical distance between the water level in the well and the equilibrium water table

in aquifer (L);


= y(to)= initial maximum head change (L);


= y(t)


= the head change at time t after the weight was removed from the well (L).












Ln (R/r)


1.1


C


(3.11)


where, H


= L, and the constant C can be found in Bouwer and Rice (1976).


Field measurements of K3 for each existing shallow well were conducted during the


winter of 1994 when the groundwater table elevations were relatively high.


A pressure


transducer and displacement weight were placed in the well under the water table (Figure


3.12).


After re-equilibrium of the water table, the weight was pulled out to create an


instantaneous change of gradient that was recorded at a 5-second interval. This data set was

used to calculate saturated conductivity of the soil in the water table aquifer. An example


of the calculation of Ks is presented in Figure 3.12.

Surface flow (SL/So)

Topographical gradients are slight in natural flatwoods and soils have high infiltration


capacities, both of which promote infiltration.


However, shallow sheet flows occur after


storms during the wet season when the water table reaches the ground surface. The sheet

flow meanders through cypress ponds in a diffused fashion toward the streams. This means


that the surface flow farther away from the streams in flatwoods could be ephemeral.


result, field measurements of surface flow in flatwoods is extremely difficult.


As a


Ditching


between wetlands and reading around wetlands may alter the topography to facilitate surface











Wel


KSE1


n K Wetland


.54 cm


L=64 cm


t=25


Yt=5.67 cm; Yo


=10.77 cm


C=2.0


.57 m/day


100


t (sec.)


Displacement Weight


Recording Pressure Transducer


Groundwater


Table


Impermeable Layer


i-i











wetlands was useless because of the very low gradient.


Instead, a smooth PVC culvert with


a 28-cm diameter was used in combination with two water table level recorders for each


inlet.


Flow velocity through these pipes was measured with a low-velocity flow meter to


calibrate the stage-discharge relationship.


Both the N and K


wetlands had inflow and


outflow during the wet season while the control C wetland only had outflow during extreme


wet periods and no significant surface inflow has been observed.


To concentrate surface


flow pathways and avoid bypassing the flumes, 30-cm high sheet metal boundary wings were

buried 10 cm into the ground surface adjoining the inlet and outlet recorders.

Eight-hour intervals were used to calculate daily total flow volume from the stage


hydrograhs.


The equation for calculating surface outflow rate in the flumes was


= 0.001428 H 2.7087


(3.12)


where,


= flow rate (m


); and,


H = stage height of surface water level in the flume (m).
Deep seepage (Di)

Deep seepage represents vertical water loss/gain through the bottom boundary from

the water table aquifer to deeper aquifers. A deep well about 7 m long as sketched in Figure

3.4 was installed in each wetland to estimate the magnitude of possible deep seepage by

comparing the pond water level to that in the deep well.










where,


= vertical seepage rate (mm/day);


= restrictive layer's vertical hydraulic conductivity (m/day);


s = hydraulic head at the top of the restrictive layer (m);

S= hydraulic head at the bottom of the restrictive layer (m);

= the thickness of the restrictive layer (m).


Change in storage (A V)

As shown in the water balance equation, the water storage component includes

surface water in the pond (Vp), and soil water in both the saturated (Vs) and the unsaturated


zones (V).


The change in water storage was evaluated on a monthly basis by calculating the


three components separately. A 10 m by 10 m grid system was established in each pond and

water table depths were measured at each intersection at a time when the water table was


high.


The bathygraphy of each wetland-upland system was determined by the computer


software SURFER (1989)


elevations.


from this water table depth data and associated grid point


The relationships between pond water volume (Vp) or pond water surface area


(Ap), and water level (wl) were estimated with this information (Figures 3.13, 3.14, 3.15).

The water storage in the unsaturated zone (V.) was calculated assuming that the soil water

content was a function of water table level, although the water content in wetland soils did


not change significantly.


The water storage in the saturated zone (Vj was the product of the







2500


2000

1500


1000

500

0



2500


2000

1500


1000

500


0


C Wetland


Volume


29.6 29.7 29.8 29.9 30.0


30.1


Pond Area


29.6


29.7 29.8 29.9


30.0


30.1


Water level elevation (m)









12000



8000



4000



0


K Wetland


Volume


29.2


29.4


29.6


29.8


30.0


30.2


30.4


45000



30000


15000



0


Pond Area


29.2


29.4


29.6


29.8


30.0


30.4


Water level elevation (m)









4500




3000


1500




0


N Wetland


Volume


28.0


28.2 28.4 28.6


28.8


16000



12000



8000



4000



0


Pond Area


28.0


28.2 28.4


98 R


9Mg


-V -. .









48
Evapotranspiration (ET)

There are four categories for ET measurement, namely, the water balance category,

the meteorological category, the physiological category, and the empirical equation category.


water balance category


be represented by the soil moisture method in


unsaturated root zone (Kirsch, 1993) or saturated zone (White, 1932; Freeze and Cherry,

1979; Gerla, 1992), the lysimetry method (Patric, 1961; Fritschen et al., 1977; Dunin et al.,


1985; Riekerk,


1985), and the watershed method, depending on the scale of interest.


Meteorological methods are designed to quantify the change of water from liquid to vapor

and vapor movement in the lower atmosphere, and to combine these processes with the


energy budget of the system.


These methods are represented by aerodynamic method, eddy


correlation and


Bowen ratio energy


balance method


(Fritschen and Simpson,


1985).


Physiological methods have been used in the study of water vapor loss (transpiration) along


with


photosynthesis


other


plant


physiological


processes


are represented


porometer, chamber and sap-flow methods. Empirical equation methods include the Penman

and Thomrnthwaite equations with a crop coefficient, Class A pan evaporation corrected by

a coefficient, and correlations among air temperature, solar radiation and ET (Jones et al.,


1984).


Although no one approach is appropriate to all situations, under certain conditions


one may be more effective than others.


For example, the watershed water balance method


is fairly accurate if the watershed boundaries are well defined and impermeable, and other








49
flatwoods uplands is limited by the underlying clay layers and the resulting shallow ground


water table in the rooting zone is very sensitive to ET.


These conditions are favorable for


estimating ET by examining the water level or water table hydrographs,


which are the


dynamic signatures of change in storage, resulting from change in the water balance.


However,


following


considerations may make this method less accurate or even


unjustifiable for estimation of ET at high resolution (e. g. daily): (1) Lateral drainage or


degassing processes may


immediately


be significant but are difficult to quantify


after storm events;


during storms or


Water content in the unsaturated zone decreases


significantly when the water table is low; (3) Surface flow or near-surface quick flow occurs

during wet seasons but is difficult to measure due to poorly defined boundaries; and (4)


Uncertainty of groundwater loss by deep seepage (leakage).


Nevertheless, the water balance


method remains a useful and relatively low-cost method to study the importance of each

hydrologic component in a complete water cycle in pine flatwoods.

In this study, monthly ET was estimated for both wetlands and uplands based on the

water balance equation for the entire soil profile including the unsaturated soil zone,


saturated soil zone and/or wetland surface water.


Longer time intervals reduced errors


caused by above considerations #1 and #2, and careful selection of the location of water table

recorders reduced the problems in considerations #3 and #4, especially for the calculations

for pine uplands.














CHAPTER 4
DATA RESULTS AND ANALYSES


The following data analyses were based on data collected since 1991


but complete


data sets for the comparison of wetlands and uplands water balances were from April 1993

to May 1994.


Precipitation and Vegetation Interception


Weather Data


Alachua


convective


County,


systems due


rainfall


to land surface


primarily


heating


generated


combined


mechanisms:


with a double sea breeze


convergence in summer months, and (2) frontal weather systems in the winter and spring


months (Dohrenwend, 1978).


Hurricane and tropical storms modified these annual rainfall


patterns. An annual rainfall of 1330 mm with air temperature of 21C was normal.


During


the three years of study, we had a wet year of 1992 and two somewhat dry years of 1993 and


1994 (Figure 4.1).


The year of


992 had an annual rainfall amount of 1


mm and a lower


average air temperature of 17 C.


The years of 1993 and 1994 had annual rainfall amounts


of 11 04 and 1 240 mm rtneeptivalv and a nArmal avernae annual air temnratnire nfe2Ao













Gator Nationals Forest


Jan Feb Mar Apr May


Jun Jul Aug Sep
Time (day)


Oct Nov Dec


Gator Nationals Forest


Jan Feb Mar


Apr May Jun Jul Aug Sep Oct
Time (day)


Nov Dec








52

Vegetation Interception


Vegetation interception (Pi) represents that part of rainfall intercepted by vegetation

surfaces that does not reach the soil surface during storm events, and eventually returns to


the atmosphere by direct evaporation.


Canopy interception has been documented as a


significant component in the forest canopy energy balance, influencing plant transpiration,

and water budgets of forest stands (McNaughton and Jarvis, 1983; Lankreijer et al., 1993).

In this study, interception was calculated as the difference between rainfall and throughfall.

Previous studies suggested that the stemflow component represented less than 2.5% and 1%

of rainfall for cypress wetlands and pine uplands respectively (Heimburg, 1976; Gholz et al.,


unpublished data).


Stemflow was not evaluated due to its insignificance, however, for the


upland forests on our site with low tree density (500 trees/ha), stemflow may have been

higher.

The canopy interception rate (IR) was defined as:


P-P P
IR' = --'
P P (4.1)


where, P is rainfall above forest canopies and P, is throughfall (mm).


Weekly


throughfall


closely


followed


precipitation as shown in


regression


equations for both uplands and wetlands of the three sites combined (Figure 4.2).


However,


the correlation between the interception rate and precipitation was poor for both uplands and













Cypress Wetland (1993-1994)


120

-100
E

--80
ILj


120

^ 100
E
E
v 80
u.
bc


20 40 60 80 10
Precipiation (P) (mm)






Pine Upland (1993-1994)


20 40 60 80
Precipitation (P) (mm)












Cypress Wetlands (1993


-1994)


20 40 60 80
Weekly precipitation (mm)


Pine Uplands


(1993


-1994)


20 40 60 80 100


Weekly precipitation (mm)








55

The high variability of IR for a certain forest stand may be due to many factors, such

as storm size, rainfall intensity, wind speed, air temperature and leaf area index (LAI) during


each measurement.


For deciduous tree species, such as pond cypress, the leaf area is


especially significant for canopy interception.


The regression equations suggested that the


maximum canopy storage was 0.56 mm and 0.02 mm for cypress wetlands and pine uplands


(excluding shrubs), respectively.


Gholz et al. (unpublished data) reported 0.77-mm of


canopy water storage for mature slash pine stands with a tree density of 190 +


118 trees


hat', twice as high as in the present research study. Apparently, the low storage value of pine


plantations on the project site reflected the low tree density and low LAI.


The deciduous


nature of cypress trees caused low canopy storage capacity in winter months from mid-

November to end of March.


Figure 4.4


shows the seasonal


distribution of the


interception rate of weekly


measurements during 1993-1994.


Deviation, STD


The average interception rates were 16.9% (Standard


= 11.1%) and 9.3% (STD=10.0%) for wetlands and uplands, respectively.


Those figures intended to demonstrate the high variability of IR during different seasons and


storm events.


The monthly summary of interception loss from wetlands and uplands will be


discussed in the 'Water Balance' section along with other hydrologic components.


Soil Moisture Content
















Cypress Wetlands (1993-1994)


,50
ta
'U
~4O


0
a~ou
.00


Xx


x x

-


Julian Day


Pine Uplands (1993-1994)


o


^ 0


o o
x,,o


-o0


I


Julian Day


*ml


I.









































I

Pressure head

h< 0








58

molecules and soil particles, and thus water pressure is less than the atmospheric pressure.

The capillary fringe is the transition zone from the unsaturated zone to the saturated zone or


the water table aquifer.


In a micro-scale view of a soil column, this zone is not of uniform


thickness because the capillary rise of water depends on the radius of soil pores as modeled


by Romanov (1968).


In the literature, the capillary fringe is also termed as the tension-


saturated zone since it is quasi-saturated but the pore water is held under tension by


molecular forces (Freeze and Cherry,


1979; O'Brien, 1982).


A rapid water table rise


following a minor application of moisture was observed when this tension-saturated zone


extends to the soil surface under high water table conditions (Gillham, 1984).


The water


table is defined as the surface in the soil/groundwater system where pore water is at


atmospheric pressure.


The water in the saturated zone is under hydrostatic pressure greater


than atmospheric pressure and the volumetric water content is approximately equivalent to


soil porosity.


However, under most field conditions entrapped air prevents water content


actually reaching the soil porosity.


Water table levels represent the lower boundaries of the


capillary zone when measured in wells that function as very large soil pores.

The importance of soil moisture and the capillary fringe on runoff generation and


solute transport has been recognized and emphasized (O'Brien,


Gillham, 1988).


1982; Novakowski and


Tracer tests showed that 50-90% of storm water was from pre-event 'old'


water rather than the new rain water, and the miscible displacement process of translator










unsaturated water flow and


'interflow' through connected macropores is not expected to be


significant in flatwoods due to the low topographic gradients, except in wetland margin


areas.


However, the response of the shallow groundwater table to storm events or the


vertical movement of water may be greatly affected by initial soil moisture content.


Spatial Distribution


The spatial distribution of soil moisture content was investigated in lateral direction

at 25-m intervals from the cypress wetland to the pine upland and vertical direction from 20


cm down to 100 cm in two transects established in the C wetland/upland system (Figures


and 3.9).


Only data collected under unsaturated conditions were used for comparisons


between different layers and sites.

Duncan's test by ANOVA for the two transects showed that the soil moisture content

at the 25-cm depth in the wetland site was significantly higher at a 95% confidence level than


that of the wetland margin and upland sites.


Overall the average soil moisture contents of


the wetland margins and upland sites were not statistically different from each other although


the values at the wetland margin spot was much higher than in the upland in transect #1.


wetland margin sites showed significantly higher values in the transect #2 while transect #1


did not.


moisture content at the


75-cm


depth


wetland margin site


significantly higher than that in the upland.


As a result, the average annual soil moisture


was











Table 4.1.


Average soil moisture content (SMC %) and water table depth during June 1993 -


January 1994 in two transects of the C system.


ES Transect


Layer


SMC


Water table
depth (cm)


STDt


NW Transect


SMC


Water table
depth (cm)


sm>


Wetland 25 cm 23.9 4.1 44 --- --- ---

Margin 25 cm 21.5 4.8 12.5 9.2

45 cm 12 3.9 91 22.2 5.6 85.5

75 cm 29.2 2.7 --- ---

Upland 25 cm 7.0 2.4 9.4 4.3

45 cm 7.9 2.3 11.6 7.3 88.2

75 cm 19.7 4.8 >100 19.2 4.2

100 cm 24.4 6.4 --- --


silyU


= standard deviation.


the wetland margin site in transect #1, was much higher than in the 25-45 cm layer (Table


4.1 and Figure 4.6).


This might be explained by: (1) the surface layer had higher organic


matter and lower bulk density than the second layer thus containing more soil moisture;

and/or (2) plant roots (mainly saw palmetto) in the 25-45 cm layer extracted more water from


the lower layer than the upper layer.


The margin site of transect #2 did not show this


difference, possibly due to a higher water table most of the time and low vegetation density

at this particular point.


Wetland soils,


which had higher water holding capacity due to higher organic












C Wetland, ESO, 1993


m3m al e (Se3Z~


| saturated


Jun. 15


Julian Day


Dec. 21


C Wetland Margin, ES1


290

E
240 2
19o>


0

1o 8
1 0
40

-10 -


Jun. 15


Julian day


Dot, 15


C Upland, ES2, 1993


160 187 179 RA 917 fll 'Ao 8i *XI 2ia 2c ,ic ,4


290

E
240 o

190



-10
140O^

90 8
.0
40

-10


~~Id


~e~


(


thy ,,








62

evaporation rates of cypress wetlands were lower than that of the uplands if the transpiration


rates of the two systems were similar (Riekerk et al.,


1995).


Two mechanisms might


. contribute to this observation: (1) less air movement and lower temperatures in dense

wetlands than in open uplands; and/or (2) inhibition by undecomposed litter mulch layer on

the pond soil surface.


Seasonal Patterns


Seasonal patterns of soil moisture content closely traced the water table depth

(Figures 4.6 and 4.7) with the low in the summer and the high in winter months during 1993.

The soil moisture content of the 0-25 cm layer of the upland had lower average values and

lower variability, suggesting that the saturated water table zone had little effect on this layer.


Water loss due to either ET


or internal drainage from the first layer was rapid and soil


moisture content was not favorable for plant use during most of the year.


In contrast,


because of the high water table, soil moisture content at the wetland margin sites was mainly


controlled by the water table.


Later in this chapter, it has been demonstrated that during dry


seasons wetland margin areas could be recharged by wetland surface water, thus more water


is available for water loss.


In upland soils, the second layer at the 45-cm depth had patterns


similar to the first layer when the water table was below 100 cm in the summer, but the

moisture content increased dramatically when the water table was above the 100-cm depth


















C Wetland Margin,


NW1


1993


217 223 231 240 249 255 261 268 276 283 291 296 304 310 316 332 355
ug. 4 Jday


-290
=-- E
240 &

190 C
00
140

-90 _C
C)
,~o




D5 -10
365
Dec. 15


C Upland, NW2, 1993


217 223 231 240 249 255 261 268 276 283 291 296 304 310 316 332 355 365


Aug. 4


290

240 o
a)

1 -O
0


-10
140 **
(0
90 .
.0
^.
40

-10


Dec. 15


/


[WT -- 20 cm 45cm -+-75cm








64

groundwater table and 10-cm depths were sensitive to rainfall events, and soil moisture

decreased slowly presumably due to less water loss by root extraction and close to the water

table.


Relationships between Soil Moisture Content and Water Table Depth


In a hypothetical soil moisture profile, such as under a free drainage condition, the

soil moisture content increases from the soil surface to the water table level (Figure 4.5).


However,


the shape of this curve may vary significantly in field conditions depending on


the water recharge and discharge history of the soil profile.


According to Phillips (1987),


when the water table dropped below 50 cm, the upward water flux from the water table to

the unsaturated zone was less than ET demand, and soil moisture content decreased more

rapidly in the surface layer, eventually altering the soil moisture distribution. For sandy soils

with high hydraulic conductivity, the water redistribution of infiltrated rainfall water is rapid.

If the water table is close to the soil surface, infiltrated water first reduces water tension in

the unsaturated zone until the soil moisture content reaches the soil field capacity, then extra


rain water will add to the saturated zone and raise the water table level.


In an opposite


scenario, when ET exists in the upper layers, upward water flux may satisfy ET demand by


reducing water table level but not soil moisture content (steady state).


When the water table


is deeper, ET may greatly reduce water content in the upper layers (0-45 cm) but less in the









65

specified in simulation models. In this study, the Van Genuchten equation (1980) of the soil

moisture characteristic curve was modified to model the soil moisture content distribution


in a soil profile.


Van Genuchten's analytical equation was developed to describe the


relationship between soil moisture content and soil water suction.


The advantage of the Van


Genuchten equation instead of other methods (Campbell, 1974; Brooks and Corey, 1966)

was that the closed form of the formula (Equation 4.2) allowed the matric potential to

approach zero or the gradient dh/d(0) to become 0:


- Fl =[*'~l


(4.2)


where,


= effective degree of water saturation;


0 = volumetric soil water content;


= saturated soil water content;


0, = residual soil water content;

h = matric potential;


= parameters.


Although previous studies suggested that soil moisture content in different soils had

linear or nonlinear relationships with water table depth (Guo, 1989; Young, 1969), the Van

Genuchten type nonlinear equation was considered to perform better for sandy soils,


cr, m, n









66

generalize the different responses of soil moisture in different soil layers to water level

fluctuations resulting from rainfall events and ET.

The matric potential, h, in Van Genuchten's formula was replaced by d-L to satisfy

the basic requirements of this equation, i.e. when h approaches 0, 0 equals saturated water


content ,.


The variable d is the water table depth and L is a constant specified for different


layers (i.e., 20 cm,


45 cm and 75 cm).


The relationship between water content and water


table level was developed as:

a) for d>L,


for dcL,


O =0

where,


(4.4)


1 I/a,

= minimum soil moisture content for the average climatic conditions during the study


period.


The lower layers had higher values of 0mi, than the upper layers.


The three


parameters, a, b, and p were fitted from observed data.


Because the water table was above


cm in the upland site in the NW transect (#2) most of the time, no modeling was done for


this layer.


The parameters fitted by the least square method have been summarized in Table


(e,-emb)[l+(p (d-L))"J48,









67
The model fitted well to the observed data as indicated by the high correlation


coefficients (R2), but a few exceptions were noticed.


The values of soil moisture content of


the wetland margin site in transect #1 showed more scatter (Figure 4.8).


The upper portion


of the data points with higher values represented a time series of rainfall recharge, and the


Table 4


Mode


parameters describing the relations between soil moisture content and


water table depth.


Location Layer 0Ojin 03 p a R2

Wetland 20 cm 0.16 0.35 0.0561 1.952 0.59
Margin Transact 1 20 cm 0.12 0.33 0.0386 2.198 0.95

45 cm 0.05 0.20 0.0887 1.955 0.84
Transact 2 20 cm 0.05 0.035 0.0711 2.268 0.79

45 cm 0.16 0.33 0.0786 1.733 0.82
Upland Transact 1 20 cm --- --- -- --- --

45 cm 0.05 0.2 0.042 2.538 0.89

75 cm 0.16 0.32 0.0566 1.59 0.78
Transact 2 20 cm 0.05 0.3 0.0593 2.474 0.91

45 cm 0.05 0.26 0.1207 1.715 0.88











Transect #1 68
Wetland ESO, 20 cm depth


40 60 80
Water table depth (cm)


Transect #1


Wetland Margin


,20 cm depth


0.15 -
20


Water table depth (cm



Transect #1


Wetland Margin


,45 cm depth


- w


* Data Model










Transect #1


Upland ES2,


0.25


0.2 -


0.15 -


0.1 -


0.05
40


45 cm depth


Water table depth (cm)





Transect #1


Upland ES2,


75 cm depth


0.15 V
70


90 100 110
Water table depth (cm)










Transect #2


Wetland Margin NW1


20 cm depth


0.4

0.35 --

0.3 -

0.25 -

0.2 -

0.15 -
*
0.1 -

0.05
20


60 80


Water table depth (cm)






Transect #2


Wetland Margin NW1


45cm


,depth


0.4 -


0.35 -


0.3 -


0.25 -


0.2 -


0.15
50


70 80 90 100 110


Water table depth (cm)


* Data Model











Transect #2


Upland NW2,


20 cm depth


40 60 80 100 120


Water table depth (cm)





Transect #2


Upland NW2,


45 cm depth


a 0.1
(0,


80 100


Water table depth (cm)










lower portion of the data represented a time series of water table decline due to ET.


This


'hysteresis' phenomena at the pond margin sites reflected the higher retention capacity of

organic soils in the surface layers as compared to the sandy upland soils.

The fitted parameters were useful not only for modeling purposes, but also to

demonstrate that the characteristics of soil moisture variation in the vadose zone followed


water table changes in the saturated zone.


The SMC-WT relationships have been plotted in


Figure 4.10 using the average values of the parameters in Table 4.2.


Those curves were used


in the hydrologic model as the input parameters to initialize soil moisture content of each soil

layer.


Water Table Dynamics


The temporal and spatial dynamics of water tables in the three wetland-upland

systems and interactions between the surface water and the groundwater are presented in this


section.


Also


, controlling factors for the water table rise were analyzed and determined.


Because the water table level has a close relationship with'the soil moisture content in the

unsaturated zone, it may have profound effects on the biological activities (e.g. redox


potential, root respiration, decomposition, seed germination), physical processes (e.g.


solute movements, mineralization) and chemical transformation (e.g. denitrification) in


flatwoods systems.


The lateral groundwater movement in this flat terrain essentially is
















Wetland 20cn

Margin 20cm
-a-
Margin 45cm

Upland 20cm
-U-pad
Upland 45cm

Upland 75cm


100


120


140


Water table depth (cm)


Figure 4.10.


Generalized relationships between soil moisture content and water table depth in a wetland-upland system








74
respectively, are equally important for amphibians to breed in flatwoods ecosystems (Tanner,


1995).


Similarly, seed germination in cypress wetlands needs unsaturated but moist soil


conditions, and long term inundation may adversely affect cypress wetlands' regeneration

(Brandt and Ewel, 1989).


Temporal Water Level Dynamics


Wetlands


To compare water levels in wetlands, uplands and other significant elevations more

effectively, water level elevation (reference to an arbitrary benchmark level for this research

site) rather than water level depth was used.

Average daily water level elevations of the three wetlands in 1992, 1993 and 1994


(before May 1) have been presented in Figures 4.11,


4.12 and 4.13.


Water levels in the wet


year of 1992 had less variation compared with those in the year of 1993, which showed a


significant dry-wet pattern.


The water level elevation in K wetland showed higher values


than in the upland during the spring from March to May and during the late fall in November


but lower values during other months.


In north central Florida, the spring was normally the


driest season due to less rainfall and high evapotranspiration demand.


cypress trees started to leaf out and transpire.


Around late April


The C upland water table levels were always


lower than those in wetlands because the upland well was on a topographically lower site.









C Wetland


1992-


1994


30.5


29.5


29


28.5


28


27.5


F M


AM


JL A


SON


FM


A M


JLA S ON D


FM


A M


1992


1993


1994


Wetland


Upland


Figure 4.11


Water level dynamics in the C wetland and the surrounding upland from 01-01


-1992 to 05-31


-199.









K Wetland


1992-


1994


30.5


30



29.5



29


ci)

S28.5



28


FM


JL A


FM


JL A


FM


1992


1993


1994


Wetland


Upland


Margin


Figure 4.12.


Water level dynamics


in the K wetland


, K margin,


and the surrounding upland from 01-01


-1992 to 05-31


m


-








N Wetland


1992-1994


28.5


28


27.5


26.5


FM


AM


JJLA


FM


J JLA


FM


AM


1992


1993


1994


Wetland


Upland


Figure 4.13.


Water level dynamics in the N wetland and the surrounding upland from 01-01


-1992 to 05-31


a!








78
elevations between wetlands and uplands implied that the groundwater flow direction


reversed during different seasons.


More detailed analyses will be presented later in this


Chapter.

Wetland hydroperiod was defined as the duration and frequency of water inundation


or flooding during a year.


It defines the rise and fall of a wetland's surface and subsurface


water (Mitsch and


Gosselink,


1986).


Although


it has


been


widely used


wetland


delineation and identification as the most important hydrologic index of wetland hydrology

(US Federal Interagency Committee for Wetland Delineation, 1989), it is still an ambiguous


and misleading concept (Best, 1993, personal communication).


While the duration (in days)


of inundation affects wetland functions, many other properties of the hydrology,


continuity


such as the


of each inundation period, and the residence time and water depth during


inundation are also very important for sustaining and developing wetlands ecosystems.


describe the hydrologic characteristics of a dynamic wetland system, such as the cypress

domes of the present study conditions, several hydrologic variables require consideration.

A comprehensive index of a combination of three factors for wetland hydroperiod is


presented below.


These factors are:


1) Duration of wetlands inundation (DI


= days of flooding annually);


2) Number of times that continuous inundation (CI) is present annually; and

3) Average depth of inundation of surface water annually (WT in cm).










magnitude of inundation.


If daily water table data are available, these three parameters can


be calculated by a graphical method.


The impacts of forest management practices on the


wetland hydroperiod may be quantitatively evaluated with these three variables.

The palmetto line is the distinct transition line from the wetland vegetation to the saw


palmetto, an understory of upland pine plantations.


The abrupt change of vegetation from


wetland to upland species (shortened ecotone) could have been caused by bedding and other

surface alterations of upland forest management practices (Clavid, personal communication,


1994).


However, Riekerk et al.


(1995) suggested that palmetto boundary lines reflected a


long term intolerance of the saw palmetto to periodic inundation and/or represented fire

boundaries. Field observations reveal that when the water level is higher than the palmetto


line, surface flow occurs.

of a cypress pond. Hyc


The palmetto line might represent the high water table boundary


Iroperiod parameters associated with the palmetto line were also


included to study maximum inundation characteristics of cypress wetlands.

The three factors for the three wetlands in 1992 and 1993 are summarized in Table

4.3 and Table 4.4, and show that the wetlands had longer duration of inundation (DI), fewer

chances of soil surface exposure (CI) and deeper water levels (WT) in a wet year than in a


somewhat dry year.


Those three indices quantitatively and efficiently demonstrate the


variation of wetland hydroperiods from year to year.


The hydroperiod analysis with respect


to the palmetto line characterized the extreme water table conditions and suggest that wetland












Table 4.3.


Hydroperiod of cypress wetlands in 1992.


Reference Indicator C K N Average
Level

Pond Bottom Duration of inundation -- 340 340 340
(DI), day
Number of continuous 1 1 1
inundation (CI)

Average water depth -- 45 32 39
(WT), cm
Palmetto Line Duration of inundation -- 119 101 110
(DI), day

Period of continuous --- 4 7 6
inundation (CI)

Average water depth --- 7 2 5
(WT), cm




Table 4.4. Hydroperiod of cypress wetlands in 1993.


Reference Indicator C K N Average
Level

Pond Bottom Duration of inundation 237 212 201 217
(DI), day
Number of continuous 6 4 4 5
inundation (CI)

Average water depth 16 21 16 18
(WT), cm
Palmetto Duration of inundation 55 80 62 66
Line (DI), day

Period of continuous 1 1 1 1
inundation (CI)
Average water depth 4 13 2 6
(WT), cm











K Wetland
1992 and 1993


20 40 60 80


C Wetland.


20 40 60 80


N Wetland
1993


loam










Uplands


Daily water table dynamics at the three upland sites during May 1993


- April 1994


are also shown in Figures 4.11


4.12 and 4.13.


Associated frequency distributions are


presented in Figure 4.15.

Water tables in the upland and margin areas indicate much higher fluctuations than

in wetlands when the water levels in wetlands were above the pond bottom. An amplified


example


given


in Figure 4.16


to compare


the different responses


to rainfall


evapotranspiration in the K wetland-upland system.


Three hypotheses have been postulated


to explain this observation.

1) Lisse effect and Wieringermeer effect;

2) Upland soils have lower storage capacity; and

3) More ET on uplands.


Lisse


effect


was


named


1947


Hooghoudt,


observed


disproportionate water table rise after rain storms in the village of Lisse in Holland (Heliotis


and DeWitt, 1987).


Rapid water table rise was attributed to air entrapment.


Rain water,


acting as a tightly


closing lid during the infiltration process caused air to become trapped,


resulting in an increase of air pressure above the capillary fringe.


Therefore, the water level


in recording wells or places open to the atmosphere was pushed up to compensate for the air


pressure increase in the unsaturated soil zone.


Lisse effects are expected to be transient and