LAKE CONWAY, FLORIDA: NUTRIENT DYNAMICS, TROPHIC
STATE, ZOOPLANKTON RELATIONSHIPS
ELDON CARL BLANCHER, 11
A DISSERTATION PRESENTED TC
IN PARTIAL FULFILLMENT OF
DEGREE OF DOCTOR
THE GRADUATE COUNCIL OF
THE IEOUIREMENTS FOR THE
UNIVERSITY OF FLORIDA
I want to extend my sincerest gratitude to my major professor,
Dr. Patrick L. Brezonik, and my co-chairman, Dr. Jackson L. Fox, for
their inspiration, guidance, support, and friendship during the past
three years. I also wish to thank the other members of my graduate
committee, Dr. Gabriel Bitton, Dr. Thomas L. Crisman, and Dr. Frank G.
Nordlie, whose advice and thoughtful suggestions helped produce the
Special acknowledgement is given to the U.S. Amy Corps of Engineers
Waterways Experiment Station and, in particular, L. Decelle and R.
Theriot whose support made this study possible.
Thanks are extended to Charles R. Fellows whose efforts were in-
valuable in producing this work and to Thomas D. Fontaine who provided
hours of thoughtful discussion. I also wish to acknowledge R. Conley,
F. Kooijman, D. Sompongse, R. Mestan, G. Comp, and T. Hall for their
valuable assistance with the field work.
My gratitude to Adele Koehler whose patience, speed, and competence
in typing made the final production of this dissertation proceed so
Finally, I wish to thank my wife, Madeleine, my son, Trey, and my
parents, Mr. and Mrs. Eldon C. Blancher for their combined moral support
and inspiration throughout my education.
TABLE OF CONTENTS
ACKNOWLEDGEMENTS . . . . . ..
LIST OF TABLES . . . . . ..
LIST OF FIGURES . . . . . .
ABSTRACT . . . . . . . .
1 INTRODUCTION . . . . .
The Nutrient Loading Concept and Nutrient Models. .
Trophic State Classification . . . . . ..
Zooplankton Communities and Eutrophication . ....
2 DESCRIPTION OF STUDY AREA . . . . . . . ..
Lake Conway System . . . . . . . ..
Regional Zooplankton Survey . . . . . .
3 METHODS AND MATERIALS . . . . . . . ..
General Limnological Characteristics . . . ..
Hydrology . . . . . . . . . . .
Nutrients . .... ....................
Trophic State Analyses . . . . . . ..
Zooplankton . . . . . . . . . .
4 HYDROLOGY OF THE LAKE CONWAY SYSTEM . . . . ..
Development of the Hydrologic Budget . . . ..
Development of the Dynamic Hydrologic Model . .
5 NUTRIENT LOADINGS AND PHOSPHORUS DYNAMICS OF THE
LAKE CONWAY SYSTEM . . . . . . . . . .
Development of a Nutrient Budget for Lake Conway. .
Aolean Sources ... . . . . . ...
Nutrients from Stormwater Runoff and Seepage .
Additional Sources of Nutrients to Lake Conway
Outputs . . . . . . . . . .
Phosphorus Loading Model .... ..............
A Dynamic Model of Phosphorus in Lake Conway ..
Nutrient Limitation in Lake Conway . . . ..
. . . . . .
6 TROPHIC STATE ANALYSIS OF THE LAKE CONWAY SYSTEM ..... 70
Seasonal Variations in Trophic Indicators ..... . 72
Determination of Trophic State for the Lake
Conway Ecosystem . .. ................ .. 74
Comparison of Trophic Indicators and Changes in
Phosphorus Loadings . . ... ............... 84
Conceptual Model of Phosphorus in Lake Conway. . 89
7 ZOOPLANKTON-TROPHIC STATE RELATIONSHIPS . . . .. 92
Zooplankton of the Lake Conway System . . .... 92
A Synoptic Zooplankton Survey of Several Florida
Lakes . . . . . . . . . . .. 106
Comparison of Florida Zooplankton Communities with
Those in Some North Temperate Systems ....... .114
Factors Affecting Zooplankton Communities ..... 116
8 SUMMIARY AND CONCLUSIONS . . . . . . . .. .122
LITERATURE CITED . . . . . . . . . . . .. 125
APPENDIX I . . . . . . . . . . . . .. 134
APPENDIX II . . . . . . . . . . . . .. 138
BIOGRAPHICAL SKETCH . . . . . . . . ...... 146
LIST OF TABLES
1-1. POSSIBLE SOURCES AND SINKS FOR N AND P FOR LAKES . . 4
2-1. MORPHOMETRIC FEATURES OF THE CONWAY SYSTEM . . . .. 14
2-2. FEATURES OF THE CONWAY DRAINAGE BASIN . . . . ... .16
2-3. SOME MORPHOLOGICAL FEATURES OF THE LAKES IN THIS STUDY 19
2-4. SELECTED PHYSICAL AND CHEMICAL CHARACTERISTICS OF THE
LAKES IN THIS STUDY . . . . . . . . . .. . 20
4-1. PRECIPITATION, PAN EVAPORATION, AND PAN COEFFICIENTS FOR
THE WATER YEAR 1976 (OCT. 1975 THROUGH SEPTEMBER 1976) . 31
4-2. SUMMARY OF STORMWATER RUNOFF HYDROGRAPHS COLLECTED AT
GATLIN AVENUE AND BUMBY DRIVE, ORLANDO, IN 1976 AND 1978 34
4-3. ESTIMATION OF TOTAL RUNOFF IN THE LAKE CONWAY SYSTEM . 39
4-4. ANNUAL HYDROLOGIC BUDGET FOR THE LAKE CONWAY SYSTEM FOR
THE WATER YEAR 1976 IN MILLIONS OF m3 . . . . ... .40
4-5. HYDROLOGIC BUDGET FOR EACH POOL IN THE CONWAY SYSTEM . 41
5-1. LOADINGS OF NITROGEN AND PHOSPHORUS TO LAKE CONWAY FROM
AOLEAN SOURCES . . . . . . . . . .. 48
5-2. NITROGEN AND PHOSPHORUS CONCENTRATIONS IN STORMWATER
RUNOFF ENTERING LAKE CONWAY . . . . . . ... .50
5-3. ANNUAL NITROGEN AND PHOSPHORUS BUDGETS FOR THE LAKE
CONWAY SYSTEM FOR THE WATER YEAR 1976 . . . . ... .53
5-4. MONTHLY LOADINGS OF NITROGEN ANDPHOSPHORUS TO THE LAKE
CONWAY SYSTEM FOR THE WATER YEAR 1976 . . . . ... .54
5-5. MAJOR SOURCES OF PHOSPHORUS LOADINGS TO LAKE CONWAY
DETERMINED BY BALANCING THE DYNAMIC MODEL . . . ... .65
5-6. RESULTS OF ALGAL ASSAY PROCEDURE PERFORMED ON LIMNETIC
WATERS OF LAKE CONWAY . . . . . . . . .. . 67
5-7. IN SITU ALKALINE PHOSPHATASE ACTIVITY OF VARIOUS PLANT
MATERIALS FROM THE LAKE CONWAY SYSTEM, 1977 . . .. 69
6-1. NATIONAL EUTROPHICAL SURVEY GUIDELINES FOR DETERMINING
TROPHIC STATE AND MEANS OF SEVERAL TROPHIC STATE IN-
DICATORS FROM THE VARIOUS POOLS OF THE CONWAY SYSTEM. . 78
6-2. SIGNIFICANT CORRELATIONS BETWEEN THE SIX TROPHIC IN-
DICATORS, PHOSPHORUS LOADINGS, AND TROPHIC STATE INDEX
OF CARLSON (1977) BASED ON SECCHI DISK FOR THE CONWAY
SYSTEM . . . . . . . . . . . . .. 79
6-3. SOME RESULTS OF THE PRINCIPAL COMPONENTS ANALYSIS ON
SIX TROPHIC INDICATORS FROM LAKE CONWAY . . . .. 81
6-4. RESULTS OF STEPWISE DISCRIMINANT ANALYSIS ON THE LAKE
CONWAY TROPHIC INDICATORS . . . . . . . 82
6-5. DISCRIMINANT FUNCTIONS DEVELOPED FROM THE LAKE CONWAY
TROPHIC INDICATOR DATA . . . . . . . ... 83
7-1. LIST OF ZOOPLANKTON COLLECTED FROM THE LAKE CONWAY
SYSTEM, APRIL 1976 THROUGH MARCH 1978 . . . . .. 93
7-2. AVERAGE AREAL ABUNDANCE OF MAJOR SPECIES AND TAXA IN
THE LAKE CONWAY SYSTEM . . . . . . . ... 98
7-3. AVERAGE ANNUAL ABUNDANCE OF ZOOPLANKTON COLLECTED FROM
SEVERAL NORTH-CENTRAL FLORIDA LAKES OF VARYING TROPHIC
STATE . . . . . . . . . . . .. !07
7-4. PERCENTAGES OF AVERAGE ANNUAL ABUNDANCE AND BIOMASS OF
THE MAJOR TAXA IN SOME FLORIDA LAKES . . . . . i
7-5. OCCURRENCE OF VARIOUS SPECIES OF ZOOPLANKTON COLLECTED
DURING THIS STUDY COMPARED TO SPECIES PRESENT IN
SEVERAL FLORIDA LAKES SAMPLED BY OTHER INVESTIGATORS. .112
7-6. INTRINSIC RATE OF INCREASE (r) AND BIOMASS TURNOVER
RATES OF VARIOUS TAXA . . . . . . . . .. 119
LIST OF FIGURES
1-1. A simplified model of nutrient pathways in a lake
ecosystem . . . . . . . . . . . . 5
2-1. A map of the Lake Conway system indicating the location
of sampling stations . . . . . . . . .. 12
2-2. Bathymetric map of the Lake Conway system . . .... .13
2-3. Map of Florida showing approximate locations of the
various lakes sampled during this study . . . ... 18
4-1. Hydrograph of stormwater runoff on 21 August 1976,
Orlando, Fl . . . . . . . . . . .. 35
4-2. Precipitation-total runoff relationship for Boggy
Creek . . . . . . . . . . . . .. 37
4-3. An information flow diagram for the dynamic hydrologic-
materials systems model . . . . . . . .... 42
4-4. Comparison of the variations in lake height of Lake
Conway to the simulated lake height from the hydrologic
model . . . . . . . . . . . . .. 45
5-1. Comparison of some initial simulations of the materials
model to total phosphorus concentrations in the lake . 60
5-2. Comparison of the final simulation of the materials
model to total phosphorus concentrations in the lake . 62
5-3. Phosphorus loadings to the Lake Conway system from
sediments, external sources and macrophytes ....... 64
6-1. Monthly variations in Secchi disk transparency and
specific conductivity in the Lake Conway System, April
1976 through March 1978 . . . . . . . .. . 73
6-2. Monthly variations in total nitrogen and total phos-
phorus in the Lake Conway system, April 1976 through
March 1978 . . . . . . . . . . .. 75
6-3. Monthly variations in gross primary productivity and
functional chlorophyll a in the Lake Conway system,
April 1976 through March 1978 . . . . . .... 76
6-4. Canonical variables resulting from the discriminant
analysis on the six trophic state indicators . . .. 85
6-5. Seasonal variations of Carlson's Trophic State Index
and phosphorus loadings to Lake Conway from external
sources . . . . . . . . . . . . .. 87
6-6. A model of major phosphorus flows in the Lake Conway
system . . . . . . . . . . . .. 90
7-1. Total zooplankton abundance for the Lake Conway system,
April 1976 through March 1978 . . . . . . .... 94
7-2. Areal abundance of the major zooplankton taxa collected
from the middle and west pools of the Lake Conway
system from April 1976 through March 1978 . .. .... . 96
7-3. Total biomass of zooplankton in the various pools of
the Lake Conway system . . . . . . . .. 1-I
7-4. Significant correlations between the major species of
zooplankton collected from the Lake Conway system ..... .102
7-5. Dendogram of the cluster analysis performed on the Lake
Conway zooplankton data . . . . . . . ... 104
7-6. Significant relationships between Secchi disk trans-
parency and the reciprocal of transparency to total
zooplankton for the various lakes sampled during this
study . . . . . . . . . . . . .. 109
7-7. Total crustacean abundance in some of the Great Lakes
and some Florida lakes . . . . . . . . .. 115
7-8. Hypothetical response surfaces of the major zooplankton
taxa to factor combinations of trophic state, temperature,
and predation pressures . . . . . . . ... 121
A. Stormwater hydrograph, 23 March 1978 . . . . .. .135
B. Stormwater hydrograph, 4 May 1978 . . . . . ... 136
C. Stormwater hydrograph, 16 June 1978 . . . . . .. 137
D. Stormwater hydrograph, 21 June 1978 . . . . .... 137
Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the Requirements
for the Degree of Doctor of Philosophy
LAKE CONWAY, FLORIDA: NUTRIENT DYNAMICS, TROPHIC
STATE, ZOOPLANKTON RELATIONSHIPS
Eldon Carl Blancher, II
Chairman: Patrick L. Brezonik
Co-Chairman: Jackson L. Fox
Major Department: Environmental Engineering Sciences
A study of external nutrient loadings to the Lake Conway ecosystem,
an interconnected series of three lakes located in Orange County, Florida,
showed that both nitrogen (2.6 g-N/m -yr) and phosphorus (0.22 g-P/m2 -yr)
inputs were within the range of loadings that leads to mesotrophic con-
ditions. The major external sources of both elements were atmospheric
inputs, urban runoff, and subsurface seepage. Experimental evidence
indicated that phosphorus became a limiting factor in the lakes briefly
during the spring and summer of 1977. A dynamic hydrologic-phosphorus
model demonstrated the relative magnitudes of nutrient loadings from
external sources and those from internal nutrient cycling by the macro-
phyte communities and sediments in this subtropical lake ecosystem.
Multivariate analysis of trophic state data by discriminant analysis
showed differences among the three lakes of the Conway system. Seasonal
trends varied concomitantly with changes in external nutrient loadings.
In both the Lake Conway ecosystem and five additional lakes located
in north-central Florida, total zooplankton abundance showed a strong
positive correlation (r = 0.87. a = 0.01) with trophic state indicators.
Zooplankton abundance averaged 1.0 x 105 organisms/m2 in oligotrophic
systems and up to 8.2 x 10 organisms/m2 in the eutrophic systems.
Seasonal variations in total abundance were greatest in the eutrophic
systems, where rotifers dominated and periodically produced sharp
population peaks (approaching 2.0 x 106/m2). In contrast, the more
oligotrophic systems had relatively stable levels of total abundance and
were dominated by copepods.
Diversities of the major taxa in the lakes were variable, with one
to four species of copepods, zero to four species of cladocera, and two
to seven species of rotifers occurring at any one time. Planktonic
cladoceran populations were often composed of only one or two species
indicating intense predation pressure on these animals.
A comparison of these subtropical zooplankton communities with
those from some temperate lakes showed higher abundances of crustaceans
in the colder climates. The lower crustacean abundance in the sub-
tropical systems may reflect increases in both temperature and predation
The process of eutrophication is caused by the increased availability
of plant nutrients. In most lake systems, the elements usually respon-
sible for the acceleration of this process are nitrogen and phosphorus.
The high concentrations of these nutrients in eutrophic lakes may reflect
either high loadings from external sources or alteration of nutrient
cycles within a lake. Such enrichment modifies the character of a lake
by increasing primary productivity and thus leads to problems associated
with increased growth of vascular plants or algae. Concomitant with the
changes in the plant communities are shifts in populations of zooplankton.
Since zooplankton are a major component in the energy and material flow
through aquatic systems, the relationship between these communities needs
careful consideration if nutrient loading-trophic responses are to be
The data presented in this work was collected during the first two
years of a comprehensive study by the U.S. Army Corps of Engineers to
test the efficacy of the grass carp, Ctenopharyngodon ideila, as a weed
control agent in Lake Conway, Florida. The analyses in this dissertation
will serve as a base from which the impact of the fish on non-target
biota can be ascertained.
There were three objectives of this investigation. The first of
these was to assess the nutrient loadings of the Lake Conway ecosystem
and develop a dynamic nutrient model which could be used to observe
seasonal changes in nutrient loadings and determine the relative con-
tributions of internal and external nutrient sources. The second
objective was to assess the seasonal variations in the trophic state of
Lake Conway and in turn relate these variations to changes in nutrient
loadings. The final objective was to observe the variability of zooplank-
ton communities in Lake Conway and some other north and central Florida
lakes, and to relate this variability to changes in trophic state. This
dissertation, then, summarizes and integrates the nutrient loading-trophic
state analyses and relates the results of those analyses to spatio-
temporal variations in zooplankton communities. Insights from this work
will help to significantly advance our knowledge of the nutrient loading,
trophic dynamics, and zooplankton ecology of these subtropical lake
The Nutrient Loading Concept and Nutrient Models
Of prime importance in our understanding of the eutrophication
process is the nutrient budget, a fundamental concept of theoretical and
applied limnology (Vollenweider 1969). General theories of nutrient
supply and losses, as well as the qualitative relationships between
nutrients and biological productivity, arewell known. Quantitative
relationships among these parameters are not yet fully understood.
The need for a quantitative method of predicting the response of
a lake to increased nutrient supply was expressed by Edmondson (1961)
in his study of Lake Washington. Since then, several workers proposed
models to predict rates of eutrophication. The earliest of these were
by Biffi (1963), Piontelli and Tonolli (1964), and Vollenweider (1965).
Improvement of these simple models led to the development of Vollenweider's
(1968) critical loading concept. Over the past few years, this concept
has been used, critically reviewed (Dillon 1974), and refined (Vollen-
weider 1975, 1976).
The construction of a nutrient loading model requires an analysis
of the sources and sinks of nutrients for a particular lake (Table 1-1).
When constructing a nutrient budget, sources and sinks generally refer
to the flow of nutrients to and from the water column. Figure 1-1 shows
a simple model of nutrient pathways to and within a lake. Nutrient
inputs may be either natural or cultural in origin and may flow into the
lake directly or indirectly as either surface or subsurface runoff from
the watershed. These materials are then recycled between the water,
biota, and sediments, and are ultimately lost in the outflowing waters,
buried in sediments, or lost to the atmosphere. Although there is a
constant exchange between the sediments and the other components, in
most lake systems the net flow is to the sediments. Thus lakes act as
traps by storing more nutrients in the sediments than are released to
the water column.
The basic equation of mass flows through a lake system is
Change in Storage = Inputs Outputs Loss to Sediments (1-1)
Expressed in differential formn the equation becomes
dt Qi[mli Qo[mlo km (1-2)
Table 1-1. POSSIBLE SOURCES AND SINKS FOR N AND P FOR LAKESa
Precipitation on lake
Deposition of plant litter
Wastes, domestic and industrial
Landfill leachates and runoff
Aquatic plant removal
aAfter Brezonik et al. (1969).
bApplies to nitrogen alone.
/\ BIOTA -- SEDIMENTS
^^^ ^ \^^ LAKE /
Figure 1-1. A simplified model of nutrient pathways in a lake ecosystem
(symbols after Odum 1971).
t= change of mass (m) in lake over time (t),
m = mass of substance m in lake,
Q0. = hydrologic inflows,
[m], = concentration of m in inflow,
Q = hydrologic outflows,
[mN]o = concentration of m in outflows,
k = first order loss term (sedimentation).
Vollenweider's model (1968, 1976) for a non-conservative substance was
based on a steady state solution of the input-output equation and re-
sulted in the following relationship:
Lm = (1-3)
qs + Z am
[m ] = mean outflow (or average lake cove) at steady state,g/m,
Lm = loading per unit surface area, g/m -yr,
z = average depth, m,
q = discharge height, m/yr,
am = sedimentation coefficient, yr.
Models developed by other limnologists (e.g., Dillon 1975; Imboden
1974; Snodgrass and O'Melia 1975) are basically adaptations and improve-
ments of this basic scheme. Dillon's (1975) model is essentially the
same as Vollenweider's although it uses the concept of a retention
coefficient (R), which is equal to the fraction of the loading not lost
via the outflow, a parameter more easily measured. It has been shown
that there is a functional relationship between Dillon's R and
Vollenweider's om (Dillon and Rigler 1975). Imboden (1974) and Snodgrass
and O'Melia (1975) developed two layer models from a similar input-
output scheme. These models were compared by Vollenweider (1976) and
Yeasted and Morel (1978), each of whom discusses various limitations
of the models.
Applied to the problem of phosphorus loading and trophic state
response, these models have advanced our understanding of the eutrophica-
tion process and our ability to predict the response of a lake to changes
in phosphorus loadings. They have resulted in a number of empirical
relationships between a lake's phosphorus loadings and the response of
its phytoplankton biomass (as chlorophyll a) (Vollenweider 1976; Larsen
and Mercier 1975; Schaffner and Oglesby 1978; Oglesby and Schaffner
1978; Rast and Lee 1978). However, as has been stated by Vollenweider
Also the trophic-dynamic interrelationships in the
sense of Lindeman (1942) requires more sophisticated
analysis . careful mass balance studies broken
down into monthly or even timely closer episodes
are scant. (p. 63)
It seems clear then, that research efforts should be directed to
the development of more detailed chemical and hydrodynamic models. While
more complex models will not supplant models of the Vollenweider type,
they will be of considerable heuristic value for our understanding of
Trophic State Classification
Important to all aspects of science is tne recognition of similari-
ties between objects, a process known as classification. Classification
involves description of the structure of constituent objects to each
other and to similar objects and simplification of these relationships
in such a way that general statements can be made about classes of
objects (Sokal 1974). Other purposes and benefits of classification
include (Brezonik and Shannon 1971; Sokal 1974; Brezonik 1976):
(1) identification--to achieve economy of memory, (2) ease of data
manipulation for better information retrieval, and (3) development of
hypotheses or theories on phenomena associated with a given class.
Limnologists have used nearly all lacustrine characteristics in
classifying lakes according to trophic state, but because of the great
diversity of lakes, no universal classification scheme based on a single
parameter has been found acceptable. The literature on this subject is
immense and several reviews on trophic indicators have been published
(Fruh et al. 1966; Vollenweider 1968; Hooper 1969; Rast and Lee 1978)
as well as on classification schemes (Brezonik 1976; Rast and Lee 1978).
Of the schemes available, the multivariate index of Brezonik and Shannon
(1971) and the single parameter indices of Carlson (1977) seem most
suitable for classifying Florida lakes.
While most studies determine the trophic nature of a large popula-
tion of lakes and assume steady state for individual lakes, the concept
of trophic state can also be viewed as a dynamic one. Individual lakes
respond to temporal variations in nutrient loadings, temperature, and
solar radiation. In order to follow trophic changes associated with
these seasonal fluctuations, it is necessary to visualize these varia-
tions on a more timely basis. For purposes of this dissertation, a
classification scheme suitable for the lakes in the sample group will
be developed to observe seasonal changes and in turn relate those changes
in nutrient loadings.
Zooplankton Communities and Eutrophication
A zooplankton community is the product of growth, reproduction,
competition between individual species for available resources, and
predation pressures. Opportunistic species with high rates of increase
are favored in rapidly changing environments while other species use
strategies that are adaptive in more constant conditions (Allan 1976).
Rotifers and cladocerans have short life cycles, are rather unspecialized
feeders, and develop large transitory populations. Copepods, on the
other hand, exhibit longer life cycles. Highly eutrophic lakes are
usually dominated by small herbivorous zooplankton rotiferss and ciado-
cerans) (Hrabacek et al. 1961; Brooks 1969; Coweil et al. 1975; Gannon
and Stemberger i978), whereas oligotrophic systems are typified by
populations of copepods (Allan 1976).
Since zooplankton community composition and structure is affected
by eutrophication, these communities have potential value as indicators
of changing trophic conditions. The effect of eutrophication on indi-
vidual species occurrence is well documented (Deevey 1942; Hasler 1947;
Patalas 1972), and a number of species have been proposed as trophic
indicators. However, use of species indicators can be severely limited
by regional specificity (Anderson 1974; Patalas 1972) and taxonomic un-
certainty (Gannon and Stemberger 1978). More recently, changes in the
major groups of zooplankton have been proposed as a more meaningful
indicator of trophic conditions (Gannon and Stemberger 1978). For
example, cladocerans and cyclopoid copepods are relatively more abundant
in eutrophic lakes than are calanoid copepods (also see Patalas 1972;
While studies of freshwater zooplankton in Florida are sparse, a
few thorough papers do exist with accompanying physico-chemical data
(Maslin 1969; Nordlie 1976; Cowell et al. 1975; Mallin 1978). With the
possible exception of Nordlie's (1976) work, however, there are no com-
parative studies over a large spectrum of lakes of varying trophic state
in this subtropical region. Thus, a further objective of this work is
to elucidate zooplankton community changes over a large range of
Florida lake types.
DESCRIPTION OF STUDY AREA
Lake Conway System
Three lakes consisting of five pools comprise the Lake Conway
system (Figure 2-1). From north to south the pools are Lake Gatlin,
the west and east pools of Little Lake Conway, and the middle and south
pools of Lake Conway proper. Located directly south of the city of
Orlando in the central portion of Orange county, these lakes form the
upper portion of the Boggy Creek drainage basin. Average elevation of
the lakes is 26.2 meters above mean sea level (MSL). Lake level is
controlled by a concrete and wooden dam at the southeast corner of the
south pool, and outflow from the system is through LaKe Warren and into
Boggy Creek via a canal. A bathymetric map of the system is given in
Figure 2-2 and morphometric features of the individual pools are de-
scribed in Table 2-I.
Soils of the surrounding watershed are of the Blanton, Charlotte,
and Orlando units and are composed of moderately to excessively drained
fine sands (Soil Conservation Service 1960, 1975). Infiltration of
rainfall into these soils is immediate and complete, resulting in
little or no surface runoff. Undifferentiated surficial sediments of
Pleistocene age are underlain by the discontinuous Hawthorn Formation
of Eocene age. Porous limestones in turn underlay the Hawthorn
- --- '7-. --. 5 <--- ,
" :,- :'./. : ., :
, <_-> :.J /^ ":; *- + ^ ^ ^- ....;==_-2.
5,,1v'. \ ,ii e :,* i itn ," "a "y", !!
"-' ..." \ ---", / .-
l ...;.- ..... /" "
.7 7 7 ~ -~
-Nsa. *. ... f l*,
7 ,-.'--. -^ :
-' __ A'
', '" ~-,, 7 dm,,,,ooa .. i
"- _._"t 1+ c.^ q
. ; .. .
2 ^ o ^"'"X
<-\ ^ ^ ^ ^L- s-
--, '*'> ,',*" '.^ \* O .i', /-"" 7__ -- *' ...
,-.. -- = , .^. .. 7 .
.. ... C ., ..... =. 0 ., _-
", ,"' ."
-. ,/; ... a "e I.
i,, .i=.._,ne Ca. a'tie*' I ,
S:: ">." '-^ "'A ~MIDDLE c .-
=' ":. . -~ .** ....,\., -.. ~_ : -
.,..J* A 1
!e e- -"- ,
_- ..: -. ,,- --. .N_ i . .
7-.' .- .3^3""' 0;\ *'.
- -" -," .. ', -" -s '? "
Figure 2-1. A map of the Lake Conway system indicating the loca Ton of
sampling stations. (0) indicates stations sampled from April
1976 through March 1978, 1* those sampled from AprIl 1976
through March 1977, and (0) those sampled from! April 1977
through March 1978. General location of the study area is
illustrated in Figure 2-3. (After USGS)
.- \,. ., ... " "- - r -.'
Figure 2-I. A map of the Lake Conway system indicating the location of
sampling stations. (.) indicates stations sampled from April
1976 through March 1978, (1) those sampled from April i976
through March 1977, and (o) those sampled from April 1977
through March 1978. General location of the study area is
illustrated in Figure 2-3. (After USGS)
A bathymetric map of the Lake Conway system.
are at 2 meter intervals.
Table 2-1. MORPHOMETRIC FEATURES OF THE
(km2) (m3 x 106)
Lake Conway, South pool 1.37 8.24
Middle pool 2.99 17.90
Little Lake Conway, East pool 1.27 5.76
West pool 1.48 8.46
Lake Gatlin .28 1.18
Total 7.39 41.54
Formation and constitute the upper portion of the Floridan Aquifer
(Lichtler et al. 1968).
The piezometric surface of the Floridan Aquifer around Lake Conway
averages 16.8 m above mean sea level (Lichtler et al. 1968). Since
lake level never falls below the piezometric surface of the aquifer,
flow from the aquifer to the lake is unlikely, but seepage of ground-
water into the aquifer is possible. The clay-rich sands of the
Hawthorn Formation usually retard vertical movement of water between
the water table and the limestones of the Floridan aquifer below. Since
this layer is discontinuous in Orange county, and the area is con-
sidered to be an area of good recharge for the aquifer (Lichtler et al.
1968), groundwater recharge through the lake bottom is probable.
Land use patterns for the Lake Conway watershed are presented in
Table 2-2. Residential development dominates the surrounding drainage
basin. A number of studies recently have been published on the Lake
Conway system in conjunction with the U.S. Army's Corps of Engineers
Project. These studies include a general description of the Lake
Conway area (Theriot 1977), and studies of its fishes (Guillory et al.
1977), aquatic macrophytes (Nall et al. 1977, 1978), plankton and
benthos (Fox et al. 1977; Conley et a!. 1978), water quality (Sawicki
1977), as well as system modeling (Ewel and Fontaine 1977; Fontaine
and Ewell 1978), nutrient budgets (Blancher et al. 1978; Blancher and
Fellows 1979), and a study of the nitrogen cycle (Sompongse 197S).
The reader is referred to these articles for an additional discussion
of these topics.
Table 2-2. FEATURES OF THE CONWAY DRAINAGE BASIH
Total Watershed Watershed I
Area (less lake
South pool 3.12 1.75
Middle pool 6.14 3.15
East pool 15.8 14.49
West pool 16.0 14.56
Gatlin 4.86 4.57
Areas reported in km2.
% Citrus or
% Citrus or
Regional Zooplankton Survey
In addition to the intensive survey of the Lake Conway system, 5
additional lakes were sampled for comparison of the zooplankton com-
munities. Ideally, to test the hypothesis of trophic state-zooplankton
relationships one would select lakes that are similar in physical and
chemical characteristics to reduce variations in zooplankton species
composition from these factors and still cover the trophic spectrum
from ultraoiigotrophic to hypereutrophic. Thus, a major basis for
selection of lakes-in this part of the study was similarity in physical
and chemical features. All study lakes were low in organic color
(< 75 mg/l as Pt), medium sized (100-1000 HA), and relatively shallow
(mean depths ranged from 2.3-7.3 m). The lakes were located in one
geographic region, north-central Florida (Figure 2-3), and can be con-
sidered as fairly representative of lakes in the area. Some morpho-
logical and physico-chemical features of these lakes are summarized
and compared to data from the Conway system in Tables 2-3 and 2-4.
The northernmost lakes (Kingsley and Sandhill) are located in Clay
County in the trail Ridge region of Florida. These lakes form, respec-
tively, parts of the Black Creek and Estonia Creek drainage basins.
On the basis of a trophic state index, both lakes were ranked in an
ultraoligotrophic group of lakes by Brezonik and Shannon (1971).
However, in their cluster analysis of uncolored lakes only, Kingsley
clusters out with some lakes that are decidedly mesotrophic. Lake
Wauberg is situated in Alachua county, 10 miles south of Gainesville.
This lake was classified by Brezonik and Shannon as a low color, soft
water, eutrophic lake. The lake is surrounded predominantly by
"-'" ...... ....... . + N -
,,- + ,: + "" .- ,,,,------ L...
KINGSLEY .. ;
.....^ r s I ^
WAUBERG mZ + .......... "-:c
MINNEOLA ^ ^ ^/ L- -*~
CONWAY --...\ \
= .It .... I + --^.S;- -
.+ j 4 -
Figure 2-3. Map of Florida showing approximate locations of the
various lakes sampled during this study.
Table 2-3. SOME MORPHOLOGICAL FEATURES OF THE LAKES IN THIS STUDY
Depth Depth Area Volume
(m) (m) (HA) (m3 x 106)
Table 2-4. SELECTED PHYSICAL AND CHEMICAL CHARACTERISTICS OF THE LAKES IN THIS STUDY
Secchi disc Total Functional
Transparency Conductivity pH Color Turbidity Hardness Chlorophyll a
(m) (pmhos/cmX) (PCU) (NTU) (mg/1) (pg/1) -
Kingsley 5.1 70 6.7 10.6* 1.0* 1.8*
Minneola 4.2 130 6.5 -- -- --
Sandhill 3.3 50 4.5 7.7* 1.4* 1.3*
Conway System 2.8 235 6.8 11.2 1.8 57.4 4.9
Wauberg 0.6 120 6.5 74.8* 4.8* -- 37.3*
Denhani 0.5 322 6.7 -- -- --
*After Brezonik and Shannon (1971).
forest, and the source of nutrients responsible for its eutrophication
is unknown. In all probability it is "naturally" eutrophic. Lake
Minneola is located in Lake county, Florida, and was classified as
oligotrophic by the National Eutrophication Survey (1973). Lake Denham
lies off the Oklawaha chain directly tributary to Lake Harris, and it
is hypereutrophic as a result of intensive agricultural activity in
the drainage basin (C. Biederman, Lake County Poll. Contr. Dept.,
pers. comm. 1976). Further details of all study lakes are provided
by Clark et al. (1964), Brezonik and Shannon (1971) (Kingsley, Sandhill,
Wauberg), and Brezonik et al. (1969) and the National Eutrophication
Survey (NES 1973) (Minneola).
METHODS AND MATERIALS
General Limnological Characteristics
Secchi disk transparency, dissolved oxygen, temperature, pH,
chlorophyll a, conductivity, turbidity, total hardness, and total al-
kalinity were routinely measured at all stations in the Lake Conway
system from April 1976 through March 1977. Temperature and dissolved
oxygen were measured in situ at 1 meter intervals using an oxygen-
thermistor thermometer probe (Yellow Springs Instruments), and samples
were obtained for additional oxygen analysis by the Winkler method.
Field measurements of pH were performed with a portable pH meter (Orion
Inc.). Samples for conductivity, turbidity, and total hardness and
alkalinity were returned to the lab, where analysis of these parameters
was carried out using standard methods (APHA 1975). After March 1977,
temperature, dissolved oxygen, pH, and conductivity measurements were
all performed in situ using a "Hydrolab Surveyor" (Hydrolab, Inc.).
All samples for chlorophyll analysis were stored in the dark at 4C
and returned to the lab within 48 hours where extraction and phaeophytin
corrected chlorophyll a determinations were performed (U.S. EPA 1973).
Monthly primary productivity measurements by the light-dark bottle
method (APHA 1975) were made at 1 meter depth in the center of each
pool of the Lake Conway system from June 1976 through March 1978.
Bottles were placed out during early morning and allowed to incubate
for 4 to 8 hours. Incoming solar radiation (langleys/min) was
simultaneously measured with a pyrheliograph. Because the productivity
measurements were made at only one depth, they are useful only as a
relative measure of trophic conditions among the various pools and
not as a measure of the overall plankton productivity of the system.
Lake and watershed areas were calculated from United States
Geological Survey topographical maps with an electronic planimeter
(Hewlett Packard 9810A-9864A). Extent of catchment area was determined
from topographic features and by consulting area tax assessment maps
at the Orange County Public Works Office for street drainage patterns.
Land use and development patterns for the watershed were estimated
from a recent (1975) aerial photograph supplied by county engineers
(Orange County Public Works, Orlando, FL).
A bathymetric map was produced using both depth profiles obtained
with a recording depth meter and an existing bathymetric map (Nall
et al. 1978). The area of each meter-depth interval was determined
with an electronic planimeter, and lake volumes were calculated by
summing the volumes of each one-meter interval. Mean depths of the
pools were computed by dividing their volumes by their respective
Monthly rainfall and evaporation data were obtained from U.S.
Weather Bureau climatological data reports (NOAA 1972, 1973, 1974,
1975, 1976, 1977, 1978). Rainfall data reported by the Orlando weather
station at McCoy Air Base were used because of the close proximity of
the airport to the lake. Averages of evaporation pan data from the
Lisbon and Lake, Alfred stations were assumed to be representative of
the Lake Conway area. Coefficients used for estimating evaporation
from the lake surface were those determined for Lake Okeechobee by
The volume of direct runoff was calculated by the Department of
Agriculture Soil Conservation Service (SCS) method (1975) and the
rational method (Chow 1964). The SCS method is based on an empirical
model which relates the amount of direct runoff to land use and
hydrologic soil characteristics. A weighted curve number (CN) is
developed from tabularized values for the different soil groupings and
the expected direct runoff is obtained from the model. The rational
method is based on the assumption that a percentage of the rainfall for
a given area will contribute to the direct runoff. By measuring the
amount of direct runoff during a storm event, a runoff coefficient can
be calculated by dividing the runoff by the total volume of rainfall
in the basin. To calibrate these methods a 128 ha area of the drainage
basin was identified from topographic maps and confirmed with Orange
County engineers for street drainage patterns. During an intense
thunderstorm in August i976, flow through an open channel draining into
Lake Conway from this area was measured. Flow rate was measured with a
Gurley flow meter at 15 minute intervals, and water samples were col-
lected for turbidity, conductivity, and nutrient determinations. Rain-
fall was simultaneously collected and measured. Additional hydrographs
were obtained during March, May, and June, 1978. Continuous measurements
of flow rates during these periods were performed using an ISCO bubble
flow meter and automatic recorder in conjunction with a 45 V notch
weir used as the primary device. During each storm event water samples
for nutrient analysis were obtained with a flow-actuated ISCO automatic
Subsurface seepage into Lake Conway was measured by Fellows (1978)
using the methods of Lee (1972, 1977). Drums enclosing 0.255 m2 of lake
bottom were pushed (open end down) 10 to 15 cm into the sediments. The
closed ends were vented by 0.9 cm diameter plastic tubing into either a
0.180 or 3.5 L bag. Flow was calculated by dividing the volume of water
collected during measured time. Water samples were collected from the
individual drum ends for subsequent nutrient analysis. Drum collectors
were placed on transects perpendicular to the shoreline since seepage
flow decreases with increasing distance from shore (McBride and
Pfannkuch 1975). Flow as a function of distance from shore was then
calculated and expressed as flow per meter of shoreline per unit time
(mn 3/m shoreline-day) when integrated.
Dry fallout and rainfall for nutrient analyses were collected
initially (August 1976-September 1976) using large (4 L) nalgene beakers.
After October 1977, collection of these samples was accomplished using
an automatic wet-dry precipitation collector (Aerochem Metrics Inc., Miami).
This collector consisted of two large buckets with a servo-operated lid
and a precipitation sensor.
Beginning in June 1976, monthly water samples were collected for
nutrient analysis at all stations in the Conway system at a depth of one
meter. Additional samples were obtained at 4 and 7 m at the deeper
stations in the center of each pool. After March 1977, sampling was
continued at the deep stations only with samples collected at each meter
interval. Supplemental nutrient data covering the period January 1976-
March 1978 were obtained from Orange County Pollution Control Department
(Ray Kaleel, pers. comm. 1978). All water samples were analyzed for
nitrogen [NH+, NO2, NO-, and Total Kjeldahl Nitrogen (TKN)] and
phosphorus (ortho and total phosphate) using methods described in
Standard Methods (APHA 1975) and the U.S. EPA Methods Handbook (1974).
Bioassay methods including Algal Assay Procedure (U.S. EPA 1971),
alkaline phosphatase assay (Fitzgerald 1968), and nitrogen fixation
(Stewart et al. 1968) were used to determine nutrient limitation in
the Conway system.
A hydrologic and phosphorus model was developed using a linear
systems model similar to one described by Rich (1973). The model was
simulated using the Continuous Systems Modeling Program (CSMP) (Speckhart
and Green 1976) on the Northeast Regional Data Center's Amdahl 460
computer. Functions for rainfall and evaporation were derived using
data from the National Weather Service (NOAA 1972, 1973, 1974, 1975,
1976, 1977, 1978), and lake height function was developed from United
States Geological Survey (USGS 1972, 1973, 1974, 1975, 1976, 1977,
1978). Data used for developing functions of phosphorus release from
both vascular aquatic plants and sediments were obtained from Fontaine
Trophic State Analysis
Trophic state of the Conway system and of the other lakes sampled
was analyzed in several ways. First, the trophic indices of Carlson
(1977) were calculated for each lake. These indices are transformations
of various trophic indicators (Secchi disk transparency, total P, and
chlorophyll a), and result in numerical scores ranging from 0 to 100
upon which an assessment of trophic state can be made.
Additional analyses of the trophic indicators of the Conway system
were accomplished using multivariate techniques. The indicators used
in this study were conductivity, total phosphorus, total nitrogen,
Secchi disk transparency, functional chlorophyll a, and primary pro-
ductivity. Principal components were computed using Biomedical Computer
Programs procedure P4M, and ordination of the various pools according
to trophic indicators was accomplished using stepwise discriminant
Monthly zooplankton samples were collected at littoral (< 3 m) and
limnetic (> 3 m) stations in each of the five pools of the Lake Conway
system following two sampling regimes. From April 1976 to March 1977,
16 littoral and 5 limnetic stations were sampled, and from April 1977
to March 1978, 13 littoral and 13 limnetic stations were utilized
(Figure 2-1). Quarterly zooplankton samples at two limnetic stations
from Lakes Denham., Wauberg, Minneola, Sandhill (Sumter-Lowry), and
Kingsley were collected from April 1977 to February 1978.
Zooplankton samples were collected at all stations by a vertical
haul with a U.S. Standard #10 (153 4 mesh) Wisconsin plankton net. Data
obtained by this method were expressed on both an area and a volumetric
basis. Beginning in October 1976, additional samples for determination
of nauplii and rotifers were collected with a U.S. Standard #20 net
(64 u mesh size) at one limnetic station in each pool of the Conway
system and at all stations in the other lakes. Zooplankton were rinsed
from the plankton nets into two-ounce sample bottles and preserved with
either 70 percent alcohol (April 1976 through September 1976) or 7 per-
cent formalin (October 1976 through March 1978). In the laboratory,
aliquots were placed in shallow dishes and identified using a setero-
scopic dissecting scope at a magnification of 20X to 50X with the aid
of taxonomic keys by Edmondson (1959), Pennak (1953), and Brooks (1957).
Significant differences exist in the volume of water that passes
through plankton nets of different mesh sizes. Generally, the smaller
the mesh size the lower the volume filtered. For this reason samples
obtained with different mesh sizes (#10 and #20) had to be weighted so
that comparisons between samples collected with different nets would
reflect real differences in the populations they represented. Vari-
ability in the counting technique was tested by taking successive
aliquots of a sample and counting each aliquot. Aliquots were chosen
so that each contained approximately 200 organisms, the minimum number
of organisms that was considered acceptable for an individual count.
To test for differences between replicate net hauls, two or three
successive net hauls of the same depth were taken on several occasions
during the study. Differences between counts were tested using the
"t" test and ANOVA statistical techniques.
Monthly average concentrations of zooplankton in the Conway system
were computed by averaging separately the littoral and limnetic samples
from each pool. A weighted mean was then computed by multiplying these
averages by the proportion of lake surface area that was represented by
the respective averages. A grand mean for the system was computed as
the simple mean of the five pool averages. Zooplankton averages of the
other lakes sampled were calculated as simple means of sample numbers,
weighted for net size.
Biomass was calculated by multiplying the weight of an average
individual of each species by the average number of individuals of that
species. The weight of individual plankters in ug/individual were
obtained from previously published values (Maslin 1969; Edmondson 1975).
If the weight of a particular species was not available from the litera-
ture, a weight was assigned using generic and size criterion. Shannon-
Weaver diversity index values were calculated using the machine method
of Lloyd and Gherhardt as per U.S. EPA (1973).
Multivariate classification and ordination of the zooplankton data
were performed using the stepwise discriminant analysis procedure of
the Biomedical Computer Program P series (BMDP77) (Dixon 1977). Trophic
state-zooplankton relationships were tested using simple linear and
curvilinear regression performed on the NERDC computer, utilizing
Statistical Analysis Systems (Barr et al. 1976). A cluster
analysis of the zooplankton data was also performed using CLUSTAN 1 C
HYDROLOGY OF THE LAKE CONWAY SYSTEM
Development of the Hydrologic Budget
Nutrient inputs to a lacustrine system are closely related to the
hydrologic budget of the system. Thus, critical to the development of
any realistic materials budget is the construction of an accurate water
The hydrologic budget of a lake may be expressed by the following
AV = Q. + P E Q + Q (4-1)
AV = change in water storage,
Q. = surface inflows,
Q = surface outflows,
Q = groundwater input or output,
p = precipitation, and
E = evaporation.
Monthly rainfall, pan evaporation, and pan coefficients for the
1976 water year are presented in Table 4-1. The volume of direct pre-
cipitation onto the Lake Conway surface was 8.87 x 10 m3. An annual
evaporation volume of 9.75 x 106 m3 was determined by multiplying the
Table 4-1. PRECIPITATION, PAN EVAPORATION, AND PAN COEFFICIENTS FOR
WATER YEAR 1976 (OCT. 1975 THROUGH SEPTEMBER 1976)
Mntih Precipitation* Evaporation Pan
(cm) (cm) Coefficient
October 1975 12.04 12.70 0.76
November 1.68 8.95 0.71
December 1.29 7.07 0.83
January 1976 0.94 7.54 0.77
February 2.11 10.23 0.69
March 4.37 14.93 0.73
April 5.49 17.09 0.84
May 26.31 17.01 0.82
June 25.22 16.34 0.85
July 17.90 17.61 0.91
August 8.26 17.25 0.91
September 14.91 14.70 0.85
Total 120.52 163.27
*For Orlando WSO, McCoy
Average for Lisbon and
Lake Alfred stations.
monthly pan coefficients (Kohler 1954) by the lake surface area (7.39
km) and the monthly pan evaporation and then summing the evaporation
Surface and subsurface inflows to the lakes were divided into two
major components each: stream inflow and surface of stormwater runoff;
subsurface seepage and groundwater inflows. Stream inflow to the Conway
system is primarily by small ditches that drain the surrounding water-
shed. Consequently, inflow from these ditches was considered as a
part of stormwater runoff. Total runoff was calculated from stormwater
runoff and seepage. Stormwater runoff is that water which enters stream
channels associated with storms and contributes directly to the
runoff hydrograph. Seepage is that portion of total runoff which
enters the lake following infiltration of soils in the watershed.
Groundwater inflows included those flows into the lake from the con-
In order to adequately assess these inflows, it was necessary to
calculate them using several methods, combine them to determine total
inflows, and then evaluate how they compared to expected total inflows
calculated from changes in lake stage. Thus a check of the accuracy
of the measured inflows could be made. The volume of stormwater runoff
can vary considerably within a drainage basin. Soil type, vegetative
cover, and extent of impervious surfaces as well as frequency, in-
tensity, and duration of rainfall play major roles in determining the
amount of runoff. It is therefore necessary to obtain hydrographs of
stormwater flows for the basin in addition to using standard empirical
methods for determining direct runoff.
Stormwater hydrographs for the Conway basin were calculated from
data collected on August 1976 and March, May, and June 1978, and are
summarized in Table 4-2. An example is given in Figure 4-1 and the
remaining hydrographs of these storms are presented in Appendix I.
Measurements of stormwater flows during 1977 were not possible due to
the paucity of rainfall. Of the precipitation that fell during those
storms, 3.5 percent on an average entered the lake as stormwater run-
off. Initiation of runoff and time interval to peak flow were generally
rapid as was return to base flow.
Estimates of stormwater runoff by the Soil Conservation Service
method resulted in a curve number (CN) of 56 for the Conway basin.
Using the SCS empirical model, this means that approximately 4.6 cm
of rainfall are necessary before any surface runoff can be detected.
For a 6.35 cm rainfall (2.5 in) approximately 0.15 cm of runoff or
2.4 percent of rainfall would result.
Since both the rational and SCS methods gave comparable runoff
coefficients, the annual volume of stormwater runoff was calculated by
multiplying the annual amount of precipitation in the basin by 0.035,
the runoff coefficient determined from the hydrographs.
In situ seepage measurements provide direct measurement of seepage
flow. Prior to the development of this technique, seepage flow was
estimated by subtracting the measurable outputs from the inputs. The
seepage meter method used by Fellows (1978) has provided integrated
values for seepage of 2.24 m3 x 106 with an estimated error margin of
+ 20 percent.
Estimation of the accuracy of the values for the total runoff com-
ponents were based on several methods of determining total runoff. For
Table 4-2. SUMMARY OF STORMWATER RUNOFF HYDROGRAPHS COLLECTED AT GATLIN AVENUE
IN 1976 AND 1978
AND BUMBY DRIVE, ORLANDO,
Volume of Antecedent
Total Basin Rainfall Rain Conditions Base Runoff
Date FlQw Rainfall Area in Basin Duration (Time Since Flow Coefficient
(m ) (cm) (ha) (m3) (hrs) Last Rain) (m3/s)
21 August 1976 986.30 1.98 128 25344 0.42 24 hrs 0.0 0.039
23 March 1978 45.43 0.18 43 774 1.25 24 hrs 0.0 0.059
4 May 1978 665.88 3.68 43 15824 2.00 24 hrs 0.0 0.042
16 June 1978 45.75 0.33 43 1419 0.90 48 hrs 0.0 0.032
21 June 1978 4.58 0.58 43 2494 0.33 5 days 0.0 0.002
Average Coefficient 0.035
S / \
1345 1800 1815 1830 1845 1900 1915 1930
TIME OF DAY
Hydrograph of stormwater runoff on 21 August 1976 at the
corner of Gatlin Avenue and Bumby Drive, Orlando, Fl.
the first of these it was assumed that correlation of total rainfall
in the basin with subsequent rise in lake level would provide an
additional estimate of total runoff. Well defined storm events follow-
ing a dry period were determined from the rainfall record, and subsequent
changes in lake storage and estimates of total runoff were calculated
(Table 4-3). From these calculations it was determined that an average
of 12.7 percent of the precipitation that fell in the watershed
(excluding the lake) entered the lake. Secondly, information from the
USGS (1974) indicates that the average annual outflow via Boggy Creek
is 17.8 cm. Considering that average total rainfall for the area is
132.1 cm, runoff is approximately 13.5 percent of total rainfall.
Variations in rainfall-runoff for the Boggy Creek, presented in Fig-
ure 4-2, indicates that when total rainfall was 120 cm approximately
12 to 13 percent of the rainfall that falls in the basin contributed
to total runoff.
By combining the estimated surface runoff and seepage input
estimated by Fellows (1978) and dividing by the precipitation in the
watershed (2.24 + 1.62/46.28 m3), a total runoff of 8.3 percent is
obtained. Compared to the calculated total runoff above of 12.13 per-
cent, this value is considerably lower than what would be expected for
An estimate of groundwater flow (from the confined aquifer) was
determined by considering the fall in lake stage during an extended dry
period and correlating this value with losses estimated from evaporation
alone. During three such periods the drop in lake levels could be
predicted by evaporation alone, showing no net inflow or outflow from
100 110 120
Precipitation--Total runoff relationship for Boggy Creek.
Closed circle (o) is the datum point for 1974, a year
when the rainfall record was incomplete, possibly ex-
plaining the apparent discrepency.
groundwater. These observations support the hypothesis that groundwater
flows (from the confined aquifer) to the Conway system are insignificant.
Surface outflow through the outlet during the study occurred only
during August, September, and October of 1976. From three measurements
obtained during those months, it was determined that annual surface
outflow was approximately 1.87 x 10 m. Additional outflows from the
lakes occur via seepage, evapotranspiration losses, and domestic pump-
age. No attempt to measure these flows was made, and it was assumed
that these flows constituted the difference between calculated inputs
From lake height information for the years 1974-1975 and 1975-1976
it was determined that the net change in storage in the system for the
water year 1976 was -0.53 x 10 m3. This amount was added to the
outputs as additional seepage out of the lake.
Utilizing all this information, a hydrologic budget for the Conway
system was constructed (Table 4-4). It is evident that the hydrologic
cycle of this system is dominated by precipitation-evaporation events.
Other significant inflows are stormwater runoff and seepage, with
latter the most important. Hydrologic budgets calculated for the
individual pools in the Conway system are presented in Table 4-5.
Development of the Dynamic Hydrologic Model
In order to verify the accuracy of the hydrologic budget, the
calculated flows were used in a dynamic model to predict change in
storage within the lake system. An information flow diagram for the
model is presented in Figure 4-3. The model was simulated for the
Table 4-3. ESTIMATION OF TOTAL RUNOFF IN THE CONWAY SYSTEM*
A Volume Precipitation Evaporation
in Lake on Lake from Lake
Net Total Percent
Precipitation Runoff Pct
*Reported in m3 x 106.
Table 4-4. ANNUAL HYDROLOGIC BUDGET FOR THE LAKE CONWAY SYSTEM FOR
THE WATER YEAR 1976 IN MILLIONS OF m3
Volume (m3 x 106)
(from confined aquifer)
TOTAL IN 12.73
Seepage Out and Pumpage
Change in Storage
*Obtained by difference.
Table 4-5. HYDROLOGICAL BUDGET FOR EACH POOL IN THE CONWAY SYSTEM*
Pool Precipitation Stormwater
Seepage Total In Evaporation Outflowst Seepage Total Out
South 1.64 .076 .395 2.11 1.72 0.18 0.21 2.11
Middle 3.59 .136 .669 4.39 3.75 0.20 0.44 4.39
East 1.52 .626 .488 2.63 1.59 0.78 0.26 2.63
West 1.78 .629 .433 2.84 1.86 0.70 0.28 2.84
Gatlin .34 .197 .164 0.70 0.35 0.28 0.07 0.70
*In millions of m
tObtained by difference.
I (t; --,,- dt -----" '-,,
qi (t) --- ^(-
Q- QoCo QIC k:1V
Figure 4-3. An information flow diagram for the dynamic hydrologic-
materials systems model.
period from December 1972 through March 1978 on a monthly basis.
Simulation of the model from the measured flows resulted in a large,
unrealistic increase in lake storage (i.e., the model was unstable).
Two possibilities could explain this discrepancy: Either the outflows
were too small or the inflows were too large. Since no reasonable
arguments could be presented to justify additional outflow from the
lake, it was assumed that the inflows to the lake were too large. The
controlling factors in the inflow equation left only two possibilities
where errors could exist; either the seepage and runoff coefficients
were too large or the basin area was too large. Since earlier arguments
indicated that the total runoff measured was already too small, the
most likely place of error was with the basin area. When the basin area
was reduced from 38.5 to 25.2 km the model behaved in a realistic
manner. The reduction in basin area can be justified. Even though the
area is well defined by drainage patterns and sewers, the effective
drainage area is smaller due to the high infiltration and depression
storage in the area. Additional evidence supporting a reduction in the
drainage area appears when we recalculate the percent of total runoff
for the area using the new basin area. Keeping total runoff the same
(1.62 + 2.24 mn ) and reducing the volume available (30.26 m3) we obtain
a new total runoff coefficient of 12.76 percent.
Additional refinement of the model was accomplished by incorporating
a factor that changed the runoff and seepage coefficients in order to
reflect (simulate) moisture conditions in the water table aquifer.
During dry periods (low rainfall), this factor decreased these coeffi-
cients and conversely increased them during wet seasons (high rainfall).
Also, minor changes in the coefficients were necessary to obtain the
best simulation. These adjustments resulted in a model that simulated
the observed lake height accurately for the period December 1972 to
June 1977. After this period, the model showed an increase in lake
height that was not observed. Repeated adjustments of the important
coefficients did not improve the response of the mode. Since a realistic
simulation was desired, an additional drain (approximately 3 percent
of inflows) was added to reduce the levels to those observed in the
lake. An explanation that this discrepancy was due to unpredictable
changes in outflow characteristics or recharge characteristics of the
basin during an exceedingly dry year is plausible.
The results of the model are compared with the actual changes in
lake height in Figure 4-4. A listing of the computer program in included
in Appendix II. Differences between the simulated and observed changes
in lake height are within 5 percent on a volume basis.
L- A -
E .~\'\ JI V
,, \ '1 ^ \
v 84 M ^
< 854 'IV
83UJ i f \ I 1 I
83 1 I I
0 10 20 30 40 50 60
Figure 4-4. Comparison of the actual variations in lake height of the
Lake Conway system (A) and the simulated lake height from
the hydrologic model (0) for the period December 1972
(month = 0) through March 1978 (month = 63).
NUTRIENT LOADING AND PHOSPHORUS DYNAMICS OF THE
LAKE CONWAY SYSTEM
The trophic dynamics of lake systems are the integration of in-
coming energies and allochthonous materials with the lake's biota. Thus,
the first step in elucidating factors affecting trophic structure is to
quantify some of these inputs. This is most effectively accomplished by
the development of a nutrient budget.
Depending on the type of information desired, two approaches for
developing budgets are available; the net budget method (Johnson and
Owen 1971) and the mass balance approach (Vollenweider 1975; Imboden
1974). Net budgets are based on the concentration of the input sources
entering a system and can have an input with a negative effect (i.e.,
if it has a dilution effect on the system). Net budgets are preferable
when the nature of the input sources is being investigated. In the
case of the mass budget, the quantity of materials is considered in-
dependent of the concentration and is always positive. Mass budgets
are more appropriate when the behavior of the individual nutrients
within the lake is the prime consideration (Burns 1976). For purposes
of this investigation the mass budget is the appropriate choice of
Development of a Nutrient Budget for Lake Conway
Atmospheric inputs of phosphorus and nitrogen originate from a
variety of sources, including fertilizer mining and manufacture, soils,
and combustion of various types (fuels, agricultural burning, and forest
fires). Eventually these materials are removed from the atmosphere as
dry fallout or scavenged by precipitation. Much of the material that
falls on land surfaces potentially can be resuspended into the atmosphere,
whereas fallout onto water surfaces is essentially irreversible. If
this does occur, then open bodies of water would receive a dispropor-
tionate share of atmospheric particulate matter (Murphy 1974).
Based on the rainfall and dry fallout samples collected (Table
501), annual loadings of 0.048 gP/m 2-yr and 0.36 gN/m 2-yr for rainfall
and 0.080 gP/m2-yr and 0.57 gN/m2-yr for dry fallout have been calculated.
The resulting bulk loadings (0.125 gP/m 2-yr and 0.93 gN/m -yr) compare
well with similar loadings of 0.105 gP/m2-yr and 1.0 gN/m--yr for
Gainesville, Florida (Hendry and Edgerton 1978, unpublished data,
University of Florida, Dept. of Environmental Engineering).
Nutrients from Stormwater Runoff and Seepage
The accumulation of particulate materials in urban areas and its
subsequent runoff into water bodies have been investigated recently
(Cowen and Lee 1976; Fieldet al. 1976; Barkdoll etal. 1977). An early
study by Weibel (1969) showed that runoff may contain suspended solids at
concentrations that exceed those in raw domestic wastewaters (Weibel
Table 5-1. LOADINGS OF NITROGEN AND PHOSPHORUS TO LAKE CONWAY FROM AOLEAN SOURCES. DRY LOADINGS ARE
REPORTED IN mg/mr-day AND WET PERCIPITATION AS mg/mm OF RAINFALL.
Total Nitrogen Total Phosphorus
Dp Wet Dry Wet
mg/m -day mg/mm rainfall mg/m2-day mg/mm rainfall
18 August 1976* 0- .30 0.080
19 November-i December 1977 -- 0.25 0.030
1 December-15 December 1977 -- 0.37 -- 0.004
15 January-11 February 1978 1.11 0.17 0.10 0.007 ,
19 March-3 May 1978 1.91 0.24 0.40 0.023
3 May-23 May 1978 1.58 0.35 0.18 0.093
23 May-12 June 1978 2.19 0.34 0.22 0.030
12 June-12 July 1978 1.32 0.32 0.28 0.018
12 July-30 July 1978 1.29 0.30 0.15 0.041
Mean Value 1.56 0.30 0.22 0.036
*One rain event only.
1969). Seven probable sources of contaminants in urban runoff were
listed by Sartor and Boyd (1972): 1) pavement, 2) vehicles, 3) atmo-
sphere, 4) vegetation, 5) litter, 6) domestic and industrial spills,
7) anti-skid compounds. Transport of these contaminants into nearby
waters often results in significant nutrient loadings and subsequent
water quality degradation.
Analysis of nitrogen and phosphorus concentrations in stormwater
runoff from the small catchment in the Lake Conway watershed described
in Chapter 4 yielded average concentrations of 3.25 mg/L and 0.39 mg/L
for nitrogen and phosphorus, respectively (Table 5-2). These compare
well to published values which range from 1.93 to 4.45 mg/L for N and
from 0.19 to 0.98 mg/L for P (Weibel 1969; Kluesener and Lee 1974;
Mattraw and Sherwood 1977). Multiplying the hydrologic loadings from
stormnwater runoff in the watershed (Table 4-4) by the measured concen-
trations and dividing by the surface area of the lake resulted in area
loadings of 0.069 gP/m -yr and 0.84 gN/m2-yr.
Since undeveloped land in the watershed surrounding Lake Conway has
a high infiltration rate, virtually no surface runoff should originate
from these areas. Since these types of areas are interspersed throughout
the "urban" area in the drainage basin, they are included in the pre-
vious calculations for residential sources.
Surface runoff from agricultural areas also is considered neglibible
for several reasons. Urban encroachment into these areas is very evi-
dent, and their classification as agricultural is questionable. Few
active citrus farms remain. Infiltration in these areas is also great
because of the paucity of impermeable surfaces.
NITROGEN AND PHOSPHORUS CONCENTRATIONS IN STORMWATER RUNOFF
ENTERING LAKE CONWAY. CONCENTRATIONS ARE FLOW WEIGHTED
AVERAGES EXPRESSED AS g/m3.
Number of Total Total Total Flow
Samples Nitrogen Phosphorus (m3)
21 August 1976 9 2.3 0.33 986.3
23 March 1978 27 4.7 0.45 45.5
4 May 1978 28 4.6 0.47 665.8
16 June 1978 12 4.0 0.42 45.8
21 June 1978 8 5.1 0.63 4.6
(flow weighted) 3.25 0.39
Nutrient inputs from subsurface seepage to Lake Conway were deter-
mined by Fellows (1978). Annual loadings from this source, calculated
as an average of three sites, were found to be 0.664 gN/m-yr and 0.024
Additional Sources of Nutrients to Lake Conway
Other "inputs" of nutrients to lake systems occur in the form of
internal loadings: nutrient regeneration from the sediments by chemical
and biological processes (burrowing insects and mixing by fishes);
nutrient "pumping" by vascular aquatic plants; and nitrogen fixation by
blue-green algae (and possibly by sediment bacteria). Inputs from the
sediments and aquatic plants are not considered for purposes of the annual
budget but they will be discussed in the phosphorus modeling section.
Detectable rates (up to 30 ngN/g dry weight-hr) of nitrogen fixation
occurred during August, 1977, in Lake Gatlin and in the east pool of
Little Lake Conway associated with benthic algal mats. Since N fixation
only occurred at that particular time and since heterocystic blue-green
algae were not common in the plankton, it was concluded that fixation of
nitrogen represented an insignificant portion of the total nitrogen
Discharge of nutrients through the outlet of Lake Conway accounted
for only a small portion of total nutrient loss. It was assumed that
outflow nutrient concentrations were equal to average surface
concentrations (see Figure 6-2), which were 0.87 mg/L for nitrogen and
0.024 mg/L for phosphorus. Multiplying by the total outflow (Table 4- )
and converting to annual areal rates of loss yields values of 0.22 g/m -yr
for nitrogen and 0.006 g/m2-yr for phosphorus.
Losses by fish and weed removal, volatilization, and evaporation
were not assessed. However, losses via these routes would probably
represent only a minor portion of the total. Groundwater recharge as a
nutrient sink for the Lake Conway system is considered to be insignifi-
cant, which follows from the conclusion of insignificant flow reached in
Chapter 4. Even if this conclusion were in error, nutrients being
carried out via this pathway would end up in the sediments of the lake
as either particulates or sorbed to clay and organic matter.
Subtraction of nutrient outputs from total inputs yields a value
which is considered to be representative nutrient loss by sedimentation.
Although this loss does not represent complete removal from the system,
sedimentation does eliminate a significant portion of nutrients from
biological or chemical cycling within the lakes.
Annual nitrogen and phosphorus budgets for the Lake Conway system
are presented in Table 5-3. The atmosphere was the major external source
of both nitrogen and phosphorus to Lake Conway, accounting for 37 percent
and 56 percent of the total, respectively. Urban runoff was the second
most important source of both nitrogen and phosphorus. Monthly nitrogen
and phosphorus loadings (Table 5-4) reflected differences in the
hydrologic loadings to the system.
ANNUAL NITROGEN AND PHOSPHORUS BUDGETS FOR THE LAKE CONWAY
SYSTEM FOR THE WATER YEAR 1976*
Combined Wet & Dry Precipitation
Undeveloped land runoff
Fish and weed removal
Volatilization and evaporation
are reported in grams per square meter
of lake surface area
Obtained by difference.
Table 5-4. MONTHLY LOADINGS OF NITROGEN AND PHOSPHORUS TO THE LAKE
CONWAY SYSTEM FOR THE WATER YEAR 1976*
Stormwater Seepage Aolean Outflow Storage Sedimentation'
October 1975 0.643 0.487 0.612 0.00 37.07 -8.44
November 0.075 0.067 0.382 0.00 28.00 +9.66
December 0.059 0.052 0.374 0.00 35.29 -6.74
January 1976 0.042 0.037 0.366 0.00 11.86 +23.94
February 0.094 0.086 0.392 0.00 18.65 -6.156
March 0.296 0.176 0.442 0.00 23.01 -3.48
April 0.347 0.221 0.467 0.00 22.14 +1.87
May 1.386 1.063 0.928 0.00 29.49 -4.11
June 1.138 1.029 0.904 0.00 3Q.77 +1.85
July 0.906 0.722 0.742 0.00 26.38 +6.72
August 0.480 0.333 0.538 0.38 42.68 -14.61
September 0.773 0.603 0.676 0.22 26.05 +18.80
October 1975 0.049 0.018 0.080 0.00 0.716 -0.133
November 0.014 0.002 0.053 0.00 1.08 -0.251
December 0.012 0.002 0.051 0.00 0.87 +0.275
January 1976 0.009 0.001 0.051 0.00 1.09 -0.159
February 0.013 0.003 0.054 0.00 1.13 +0.030
March 0.021 0.006 0.060 0.00 1.04 +0.177
April 0.026 0.008 0.063 0.00 1.06 +0.077
May 0.101 0.039 0.118 0.00 0.60 +0.718
June 0.092 0.037 0.115 0.00 0.75 +0.094
July 0.071 0.026 0.096 0.00 0.38 +0.287
August 0.030 0.012 0.070 0.003 0.37 +0.122
September 0.060 0.022 0.088 0.004 0.47 +0.070
*A11 loadings are reported in g x 106.
Septic tank inputs were treated as a constant inflow for these calcula-
tions. Positive sedimentation represents a loss from storage to the
Phosphorus Loading Model
If volumetric flow rates into and out of the system are unequal and
time varying, and a completely mixed situation exists, the budget for a
conservative substance (not transformed chemically or biologically or
removed by sedimentation) entering a lake would follow a fundamental mass
balance equation as follows:
dt- ii o (5-0)
V = volume of lake at time t,
Q. = inflows to lake,
Q0 = outflows to lake,
c.i = average concentration in inflow, and
co = average concentration in lake.
For a non-conservative substance such as phosphorus the basic equation
could be rewritten as:
dt -Qici Qoc0 kcV (5-2)
where K is a first order rate constant or a term for the loss of sub-
stances other than through the outlet (i.e., reaction and/or sedimenta-
Equation (5-2) above can be solved with the restrictive assumptions
that outflow concentration is equal to the mean lake concentration and
that sedimentation is a function of the mass of material under considera-
tion in the lake. Vollenweider (1968) obtained the solution:
m Lm 1 (5-3)
w qs 1+ am(qs)
[mw] = mean lake concentration at steady state,
Lm = yearly area loading of substance m,
m = sedimentation coefficient,
qs = discharge height in meters, and
z = mean depth in meters.
Generally, mean lake concentration is a function of residence time, areal
loading, and sedimentation characteristics. Using this relationship for
phosphorus in a number of Swiss and North American lakes, Vollenweider
(1975) developed empirical limits for critical loadings for lakes of
various flushing characteristics. He found that these limits were
0.1-0.3 gP/m -yr and 1.0-2.0 gN/m -yr in temperate lakes. Brezonik and
Shannon (1971) found higher limits of 0.28-0.49 gP/m -yr and 2.0-3.4
gN/m -yr for Florida lakes using a different approach, and suggested
that subtropical lakes may be able to assimilate somewhat larger nutrient
loadings. Phosphorus loadings to Lake Conway fall within the range of
critical loadings, whereas loadings for nitrogen exceed the acceptable
loadings of Vollenweider and are in the critical range according to
Brezonik and Shannon.
Two terms in the Vollenweider input-output model must be carefully
considered before utilizing Vollenweider's model: the water residence
time (rw) and the phosphorus sedimentation coefficient (ap ). Water
residence time has been recently shown to be one of the most important
factors in empirical loading models (Yeasted and Morel 1978). Hence,
critical to any application of such models is the definition of resi-
dence time. Although T appears to be a straightforward concept, defined
as the lake volume divided by the outflow, care must be taken on what
"outflow" means in terms of the models. Since this term represents the
flushing characteristics of a lake, outflow, especially in cases of
lakes whose hydrology is dominated by precipitation-evaporation events,
should be defined as total outflow less the losses due to evaporation.
Thus we obtain an "effective" residence time that represents the true
flushing characteristics of the system. All subsequent calculations
were performed using this "effective" residence time.
The second factor that must be critically examined is the phosphorus
sedimentation coefficient a (equivalent to k in Eq. (5-2)). Realizing
its importance in his relationship and the reality that it must be cal-
culated from the model itself, Vollenweider (1975, 1976) provided several
ways to calculate this coefficient. The first of these derived from
observations on lakes in his earlier studies:
In a = In 5.5 0.85 In Z (5-4)
Equation (5-4) results in a = 1.34 for Lake Conway. Use of this value
to predict the mean lake concentration [P-i from Eq. (5-3) results in a
value of 0.034 mg/L which is considerably higher than the observed value
of 0.024 mg/L. Working backwards from the observed concentration, I
obtained a value of 1.8 for up, indicating Vollenweider's equation may
underestimate a for Florida lakes.
Another relationship which can be used to calculate a p is the
P = (5-5)
where [P] represents average lake concentration over depth Z in mgP/m3
and F (Z) is the flux of P through horizon Z in mgP/m -yr (Vollenweider
1976). The major drawback to this equation is that F (Z) is unavailable
for Lake Conway and most lakes which Vollenweider studied.
Vollenweider (1976) then defined sedimentation velocity as
F (Z/[P]sP) and by working backwards from Eq. (5-5) using a sedimentation
coefficient from his model, he obtained an apparent settling velocity of
approximately 10 m/yr. As Vollenweider points out, the value obtained
from the model is considerably below experimentally obtained values be-
cause it is a net sedimentation velocity. Burns and Pashley (1974)
measured actual settling velocities in Lake Ontario and found that they
ranged from -0.4 to 2.0 m/day. Taking an average of their results,
calculated for a year, a value of 240 m/yr is obtained.
A Dynamic Model of Phosphorus in Lake Conway
Equation (5-2) in the preceding section was used in the systems
model to simulate the change in total phosphorus concentration over time
(see Figure 4-3). Although the model was simulated for a period of over
five years, total phosphorus concentrations were available only for the
period January 1976 through March 1978, and for comparative purposes only
those results are presented (Figures 5-1 and 5-2). However, the five
year simulation was justifiable for the purpose of testing the model's
Using only external inputs and a sedimentation coefficient derived
from Vollenweider's (1975) model, the model yielded for phosphorus
levels that are representative of conditions in the lake on an annual
basis (Figure 5-2A) but the simulations did not fit the observed seasonal
trends of total phosphorus concentrations. This was expected because of
the simplifying assumptions used in Vollenweider's derivation, i.e.,
that a constant, a p, can be used to represent the net sedimentation
coefficient for the system. While the assumption of a constant loss
coefficient may be adequate for an annual cycle it is an over-simplifi-
cation for a dynamic model. Hence other factors must be considered to
explain the seasonal events observed.
Functions were obtained of phosphorus loadings from vascular plants
and the sediments from Fontaine (1978). He estimated these functions
from productivity measurements and sediment leaching experiments per-
formed during 1976. Since these functions represent maximal expected
inputs from these sources, their actual contribution to the lake is con-
sidered as a fraction of the total. As an initial estimate, these
functions were run at half of their potential loadings (Figure 5-2B).
Reduction of the loadings was continued until the simulated phosphorus
concentrations fit the observed seasonal trends. This occurred when
the plants were reduced to 20 percent and sediments were reduced
to i percent of their respective maximal loadings. In order to obtain an
optimal fit, the sedimentation coefficient from Vollenweider's model had
to be increased from 0.15 to 0.70. While some justification of using a
higher sedimentation coefficient was presented previously in the dis-
cussion of loading models, it is necessary to examine the relationship
more closely. Using Vollenweider's relationship described earlier to
compute a settling velocity for Lake Conway using the sedimentation
coefficient from the dynamic model, a value of 44.1 m/yr is obtained.
-~ YLd ~ UN ~
-F-I'-r-- v I r r -r----i I- 1 1 I-r- -i- r r T rT 1i I
JF M A M J J AS 0 N D J F M A M J J A S 0 N D
1976 MONTH 1977
A comparison of some initial simulations of the materials model to total phosphorus
concentrations in the lake. Line A is the simulation using external sources only and a
sedimentation coefficient from Vollenweider's model. Line B is the simulation with
external loadings and 50 percent of plant and sediment loadings estimated by Fontaine
1978). Bars represent the standard error on either side of the mean of observed values
Although this is still well below the values reported by Burns and Pashley
(1974), it is probably realistic for a lake with an average mean depth
of only five meters.
As an additional reference parameter to check the calculations of
the loading model, Vollenweider (1976) also introduced the concept of
relative phosphorus residence time, Tr defined as
S/T w 1 (5-6)
717r Tp / W l/ w+ a +/T
The third term of Eq. (5-6) is developed from the steady state model and
the fourth term is a statistical approximation. For Lake Conway, ir
calculated from the third term is 0.053, but if calculated from the
fourth term it is 0.231. According to Vollenweider, this may indicate
that accumulation of phosphorus in the sediments of Lake Conway must
occur at a greater rate than in the lakes he used in his study. This
corroborates both the results of the dynamic model and the previous cal-
culations, as well as the findings of Brezonik and Shannon (1971).
Shallow subtropical lakes, such as those found in Florida, may oe able
to assimilate more nutrients than would be expected for temperate lakes,
justifying the increase in the sedimentation coefficient.
The final simulation is presented in Figure 5-2 along with observed
trends in total phosphorus concentrations. A correlation coefficient
(r = .25) was calculated for the two curves indicating a weak fit of the
model to the observed data. When the first five months in 1976 are not
included in the comparison, a more respectable fit is observed (r = .63).
The observed high phosphorus concentrations in early 1976 may have been
J F MA MJJ A S 0 N D J F M A M
- --- -- r t- -- -F1
J J A S O N D J F
A comparison of the final simulation of the materials model to total phosphorus
concentrations in the lake. Bars represent the standard error on either side of
caused by a large amount of phosphorus being recycled in the lake system
due to extensive herbicide treatment of the aquatic plants the pre-
Contributions of total phosphorus to the lake from the three major
input sources are presented in Figure 5-4 and Table 5-5. The relatively
small contribution of phosphorus from the sediments and the dominance
of macrophyte pumping in the model gives some insight to the role of
these components at least for the Conway system. The simulation suggests
that phosphorus released from the sediments during anoxic conditions in
the summer does not become entrained during overturn in the fall. This
hypothesis has been suggested previously by Fitzgerald (1970) who
observed that sorption of phosphorus by lake muds is rapid enough to
remove most of the phosphorus before it can be used to support algal
growth. An alternate hypothesis that the amount of P from the plants
is too high and the sediments are actually supporting the observed con-
centrations of phosphorus, cannot be ruled out entirely. However, the
mass balance of phosphorus determined by use of the dynamic model does
not support the sediment release hypothesis for Lake Conway.
Nutrient Limitation in Lake Conway
External nutrient loadings are often considered to be the major
factors controlling lake eutrophication. However, mechanisms of nutrient
cycling such as nutrient recycling from sediments and biological fixation
of N2 play important roles in lacustrine nutrient cycling. Recycling of
nutrients between the biota of a lake is also important and becomes
especially intense during periods when nutrient availability is low.
-Sediment + External
+ Sediment Loading
- -- i --' i ., 4.-.- z -'-
Oct Mar Aug Jan
Phosphorus loadings to the Lake Conway system from sediments, external sources, and
macrophytes. The area between the top and middle curves represents loadings from
macrophytes and the area between the middle and bottom curves is loading from external
MAJOR SOURCES OF PHOSPHORUS LOADINGS TO LAKE CONWAY DETERMINED
BY BALANCING THE DYNAMIC MODEL*
Year 1976 1977
all sources 0.24 0.25
macrophytes 0.66 0.68
sediments 0.02 0.02
TOTAL 0.92 0.95
*Reported in g/m -yr.
Thus, knowledge of which nutrients) is (are) limiting is essential to
understanding how a particular lake system behaves.
Results of algal bioassay (AAP) experiments performed quarterly
during 1977 (Table 5-6) indicate that phosphorus became limiting during
spring in Lake Gatlin. During August 1977, spikes of nitrogen and phos-
phorus combined and phosphorus alone caused significant increases in
algal growth in all pools except the south indicating phosphorus limita-
tion during that month.
Alkaline phosphatase activity measures the ability of plants to
utilize organically bound phosphorus. Enzyme activity increases when
inorganic orthophosphate becomes scarce (Fitzgerald and Nelson 1966).
Results from May and June, 1977, show high levels of activity for the
phytoplankton throughout the system (Table 5-7). Fitzgerald and Nelson
(1966) demonstrated that phosphorus-limited algae exhibit activities of
2,400 to 28,000 enzyme units or 5 to 25 times those of unlimited algae.
Activities in June suggest phosphorus limitation in all pools. By July,
however, alkaline phosphatase activities of all species tested were below
the levels indicative of limitation. The activity of the macrophytes
tested probably reflects the activity of the periphyton community and not
that of the macrophytes themselves.
Limitation of algal growth by nitrogen could not be demonstrated in
Lake Conway. Neither the results of algal bioassay (AAP) nor that of
the nitrogen fixation experiments described in the budget section indi-
cated limitation by nitrogen for the period tested.
Nitrification is a process whereby ammonium is oxidized to nitrite
and nitrate by a select group of aerobic autotrophic bacteria. Although
nitrification rates of up to 5.4 mg/L-day were obtained in experiments
Table 5-6. RESULTS OF ALGAL ASSAY PROCEDURE, 1977, PERFORMED ON LIMNETIC WATERS OF LAKE CONWAY
Month January April August November
STreatment NP N P N+P N P N+P N P N+P N P
Gatlin NS ++ NS + + NS + .
West Pool NS NS -- -- + NS + + --
East Pool NS NS -- ++ NS + NS --
Middle Pool NS -- NS + NS NS NS
South Pool NS NS -- NS -- + --
NS = Not Significant
+ = Significant (p = .05)
++ = Very Significant (p = .01)
-- = Test Not Performed
performed by Sompongse (1978) on Lake Conway, she concluded that Lake
Conway sediments do not add appreciable amounts of nitrate to water
except in well oxidized, stirred situations such as might be occurring
in shallow areas or during lake turnover.
These results show that phosphorus was limiting during spring and
summer of 1977. However, because of the conflicting results obtained by
AAP and alkaline phosphatase for the late summer (July-August), limita-
tion seems to be a transitory occurrence in this particular system.
Aerial loadings (56 percent) and urban runoff (31 percent) are major
sources of allochthonous phosphorus to the system. The early part of
1977 was extremely dry as evidenced by a delay in the peak of phosphorus
loadings (Figure 5-4). Thus the fact that phosphorus became limiting in
the early summer of 1977 is not particularly surprising. What may be of
more significance is the rapid response of the system to a change in
loadings, indicating again rapid assimilation and elimination of phos-
phorus entering the system.
IN SITU ALKALINE PHOSPHATASE ACTIVITY* OF VARIOUS PLANT
MATERIALS FROM THE LAKE CONWAY SYSTEM, 1977
Location Test Materials -
May June July
Lake Gatlin Phytoplankton (Net) + 7,000
Lyngbya + 10,222 --
East Pool Phytoplankton (Net) + 2,500 1,200
Oedogonium + 10,285 --
Hydrilla + -- 428
Vallisneria 0 239 --
Potamogeton 0 -- 200
Nitella 3,530 810
Middle Pool Phytoplankton (Net) + 23,636 591
Nitella 1,356 1,466
Potamogeton 0 -- 345
*Activity reported in alkaline phosphatase
units/mg dry weight-hr of plant
tDue to improper sample fixation all of the substrate was hydroylzed.
Thus rates could not be quantified for this month. (+) indicates a
positive response, i.e., complete hydrolysis; (0) indicates a negative
response; (-) indicates no sample was taken.
TROPHIC STATE ANALYSIS OF THE LAKE CONWAY SYSTEM
When Naumann (1919) used the term "eutrophic formation" in de-
scribing an algal assemblage in nutrient rich waters, he initiated a
debate that still goes on today. Although most limnologists agree that
the process of eutrophication is the enrichment of a water body with
concommitant changes in water quality and productivity, no universally
acceptable means exists to classify lakes according to trophic state.
Difficulties arise because the concept is so loosely defined, and to a
considerable extent classification requires some degree of subjectivity.
While numerous attempts at quantifying trophic state exist in the
literature, few approach the problem in a way that would prove suitable
for a large array of lakes. Some notable exceptions are the works of
Vollenweider (1968), Brezonik and Shannon (1971), and Carlson (1977),
each of whom used methods that are basically different in approach but
are rational alternatives for the classification of lake systems.
Vollenweider (1968, 1975) extensively reviewed the subject of eu-
trophication, viewed the problem from a causal approach, and developed a
mass balance model, as discussed in a previous section, that could be
used to classify lakes according to their nutrient loadings and hydraulic
characteristics. The mass balance model has provided much insight into
how lake ecosystems operate and it offers much in terms of overall
predictive ability. However, it gives little indication of actual water
quality of the water body. Additionally, relatively large amounts of
data are required to acquire a reasonable fit of the model.
Brezonik and Shannon (1971) saw the problem of trophic classifica-
tion as multivariate in nature, and they classified 55 Florida lakes
using the statistical techniques of cluster analysis and principal com-
ponents analysis. The latter was used to develop an index using seven
trophic indicators. An advantage of this approach is its inclusion of
many parameters representing the physical, chemical, and biological
nature of the lake in a single numeric index on which the lakes were
scaled. Disadvantages are its requirement of a large data base and the
difficulty in physical interpretation of the statistically-formed lake
groups and trophic ranking equations. Also, lakes outside the original
set used to formulate the index cannot be ranked easily since a new
model must be created to include all the lakes in a particular popula-
Carlson's (1977) trophic state indices (see Chapter 1) have some
of the advantages that the foregoing approaches lack, e.g., relatively
small data requirements and ease of computation. Thus, they can be
applied to a relatively large set of lakes. However, they suffer in that
each of the indices may not be good for use in some lakes due to problems
from turbidity, color, and toxins. Additionally, while the three indices
showed excellent agreement for the group of lakes Carlson analyzed, the
possibility of ambiguous and conflicting results from the indices is
great. Thus, the indiscriminate use of one or all of the indices is not
The objectives of this phase of the study were to observe seasonal
changes of the trophic state of Lake Conway and to find a way of quanti-
fying those changes. In order to accomplish that task, it was necessary
to utilize as much of the methodology that was available for determining
the trophic state of the system and to choose those parameters and
methods that best reflect the temporal variations in trophic state in
Seasonal Variations in Trophic Indicators
Six parameters were chosen as trophic indicators of the Conway
system representing the physical (Secchi disk transparency and conduc-
tivity), chemical (total nitrogen and phosphorus), and biological
(functional chlorophyll a and gross primary productivity) attributes of
the lakes. Monthly variations of these parameters for the period studied
are presented in Figures 6-1 through 6-3 for each of the pools.
Conductivity increased throughout the study in all pools, indicative
of the dry conditions observed in 1977 (Figure 6-1). A general trend of
increasing conductivity was evident moving north through the lakes to
Lake Gatlin which could be attributed in part to increased loadings from
Secchi disk transparency showed distinct seasonal variations, with
higher values observed during winter and lowest values occurring in fall
during periods of high algal biomass (Figure 6-1). As noted earlier,
Secchi disk transparency was observed to decrease northward through the
pools, corresponding to variations in other parameters.
AJA 0 FA J A O D F
Figure 6-1. Monthly variations in Secchi disk transparency and
specific conductivity in the Lake Conway system, April
1976 through March 1978.
* *, -
Total nitrogen (Figure 6-2) varied from 0.30 mg/L to 1.6 mg/L in
the various pools of the Conway system with the lowest values occurring
during the months of January and February. Higher values often were
noted in spring and fall. Sompongse (1978), whose study of the nitrogen
cycle in Lake Conway coincided with this study, presented a detailed
summary of the nitrogen cycle in these pools.
Consistent seasonal trends of total phosphorus (Figure 6-2) were not
evident in the individual pools, but lower concentrations occurred
primarily in the months of July and August, whereas higher values were
noted in spring and fall. Concentrations fluctuated around 0.02 mg/L
for all pools, with slightly higher values occurring in Lake Gatlin.
Primary productivity and functional chlorophyll a showed inconsis-
tent variations with respect to pool and month (Figure 6-3). The highest
values for both of these parameters were usually noted in late summer and
fall. Again, Lake Gatlin consistently showed the highest values of all
the pools. Productivity values were lowest during the winter months
(January and February), corresponding with lower temperatures and solar
radiation during those periods. Chlorophyll a also showed low values
during the winter, but even lower concentrations often were seen in mid-
summer in most pools.
Determination of Trophic State for the Lake Conway Ecosystem
Limnologists have determined general guidelines for determining
trophic state based on total phosphorus, chlorophyll a, and Secchi disk
transparency. Gakstatter et al. (1975) reviewed the literature and
established criteria used for classifying the lakes sampled for the
AJ A D F A JA0N J
AJ A 0 D F A J A 0 D F
Figure 6-2. Monthly variations in total nitrogen and total phosphorus
in the Lake Conway system, April 1976 through March 1978.
A JA O D F A J A 0 D F
0 CHLOROPHYLL a
Figure 6-3. Monthly variations in gross primary productivity and
functional chlorophyll a in the Lake Conway system,
April 1976 through March 1978.
National Eutrophical Survey (NES). Comparing their values with annual
average values of the five pools of the Conway system (Table 6-1, we see
that the south and middle pools border on oligotrophy, whereas the east
and west pools are decidedly mesotrophic. Using the NES criteria, Lake
Gatlin would be considered eutrophic. While this comparison illustrates
the overall (average) trophic states of these lakes, better resolution
over a shorter time interval is more useful in examining the mechanisms
and causative factors involved in the eutrophication process. This
resolution is necessary for any comparison of the importance of various
loading sources to overall lake response. As a first indication of a
relationship among the trophic indicators in the lake, correlation
coefficients for the parameters averaged over all the pools were cal-
culated (see Table 6-2). Conductivity was not significantly correlated
to any of the other indicators, and surprisingly, neither was total
phosphorus. Secchi disk transparency showed a highly significant nega-
tive relationship with total nitrogen, productivity, and functional
chlorophyll a; and the correlation between nitrogen and chlorophyll a
was also highly significant.
In order to elucidate temporal variations in trophic state and the
underlying structure and dependency of the trophic indicators, the
multivariate statistical techniques of principal components and stepwise
discriminant analysis were used to examine the individual lake data.
The method of principal component analysis can be used to determine
the dependence structure of multivariate data. This technique reduces
the dimensionality of the data set by forming linear functions (i.e.,
principal components) of the observed variables (Morrison 1967). The
NATIONAL EUTROPHICATION SURVEY GUIDELINES FOR DETERMINING TROPHIC STATE (AFTER GAKSTATTER
ET AL. 1975) AND MEANS OF SEVERAL TROPHIC STATE INDICATORS FROM THE VARIOUS POOLS OF THE
South Middle East West Lake
Pool Pool Pool Pool Gatlin
Total Phosphorus (rag/L) < 0.01 0.01-0.02 > 0.02 0.016 0.018 0.021 0.020 0.021
Chlorophyll a (pg/L) < 4.0 4-10 >10.0 2.8 3.7 4.14 4.47 9.49
Secchi depth (rn) > 3.7 2.0-3.7 < 2.0 3.46 3.30 2.70 2.66 1.86
SIGNIFICANT CORRELATIONS BETWEEN THE SIX TROPHIC INDICATORS, PHOSPHORUS LOADINGS, AND TROPHIC
STATE INDEX OF CARLSON (1977) BASED ON SECCHI DISK (TSISD) FOR THE CONWAY SYSTEM. NON-
SIGNIFICANT RELATIONSHIPS,INDICATED BY A DASH (--),HAD r VALUES OF LESS THAN 0.40.
Total Total Productivity Chlorophyll Secchi TSIS L L Ls
Phosphorus Nitrogen a Disk (SD) p e
Total Phosphorus -- -- -- -- 0.44* 0.43* .
Total Nitrogen -- 0.55** -0.64** 0.64* ..
Productivity -- -0.55** 0.54** ..
Chlorophyll a -0.58** 0.57*
Secchi Disk -0.99** ..
Lp = Phosphorus loadings from plants.
L = Phosphorus loadings from external sources
L = Phosphorus loadings from sediments.
*Significant at 0.05 level.
**Significant at 0.01 level.
first principal component is that linear combination of variables which
explains the maximum amount of the variance in a n-dimensional cloud of
data. The second principal component is the linear combination of the
variables that explains the maximum amount of the remaining variance,
and so on for as many variables there are in the data set. Ideally, one
would like to obtain a model where most of the variance is explained by
the first one or two principal components so that the dimensionality of
the data is reduced to managable levels.
Analysis of the six trophic indicators using principal components
analysis resulted in a model whose first three components explained 75
percent of the total variance (Table 6-3). Since the first principal
component explained only slightly over 40 percent of the total variance,
use of this model as a trophic state indicator would be tenuous at best.
The multivariate method of discriminant analysis provides linear
functions of the variables from a multidimensional data set. These
functions then allow the classification of the multidimensional data vec-
tor into one of several multivariate normal populations. A modified
procedure known as stepwise discriminant analysis (BMDP7M, using default
options) was used to identify the variables that would produce a "best
fit" model to discriminance between the populations. Each of the pools
in the Conway system was treated as a "population," and the observed
variables included the monthly mean values of the six trophic state
indicators. The procedure was performed using all six indicators, as
well as several subsets of the variables.
The results of the analyses (Table 6-4) show that conductivity and
Secchi disk transparency were the only variables necessary to discriminate
between the pools, and the resulting canonical variables explained most
SOME RESULTS OF THE PRINCIPAL COMPONENTS ANALYSIS ON SIX
TROPHIC INDICATORS FROM LAKE CONWAY
First and Second Principal Components (PC)
PC1 = 0.499
PC2 = 0.666
Conductivity + 0.636
0.729 Productivity +
0.789 Secchi Disk
0.265 Productivity -
0.281 Secchi Disk
Total N + 0.170 Total
0.764 Chlorophyll a
Total N + 0.654 Total
0.160 Chlorophyll a
Variance Explained by First Three Principal Components
Factor Variance Explained of Total Variance
Table 6-4. RESULTS OF STEPWISE DISCRIMINANT ANALYSIS ON THE LAKE CONWAY TROPHIC INDICATORS
Test Factors Considered Number Variable F' F .95(v *
of Groups Entered ".5vv
1 Conductivity, Secchi disk 5 Conductivity 15.88 4.00
transparency, chlorophyll a, Secchi disk 10.60 4.00
total nitrogen, total phosphorus,
product i vity
2 Conductivity, Secchi disk 5 Conductivity 15.86 4.00
transparency, chlorophyll a, Secchi disk 10.60 4.00
3 Secchi disk transparency, 5 Secchi disk 10.26 4.00
chlorophyll a, total nitrogen,
4 Conductivity, Secchi disk 3 Conductivity 32.11 4.00
transparency, chlorophyll a, Secchi disk 21.34 4.00
total nitrogen, total phosphorus,
*vI = G 1; v2 = N-G V (G = No. of groups, N = Total no. of samples, V = No. of variables).
Table 6-5. DISCRIMINANT FUNCTIONS DEVELOPED FROM THE LAKE CONWAY TROPHIC INDICATOR DATA. THE CLASSIFICA-
TION MATRIX SHOWS THE NUMBER OF SAMPLES PLACED IN EACH GROUP AND THE PERCENT OF CORRECT
Test 1 Test 4
Group South Middle East West Gatlin Soutdl East- Gatlin
Group Gatlin Middle West
Conductivity 0.311 0.313 0.339 0.346 0.409 0.317 0.349 0.416
Secchi disk 2.237 2.057 1.198 1.112 -0.177 2.194 1.189 -0.162
Constant -39.457 -39.273 -42.686 -43.929 -56.946 -39.506 -43.506 -57.359
Group Percent Number of Cases per Group Group Percent Number of Cases per Group
South 'Middle East West Gatlin South-Middle East-West Gatlin
South 50.0 11 1 5 5 0 South-Middle 59.1 26 18 0
Middle 18.2 10 4 2 6 0 East-West 56.8 11 25 8
East 18.2 4 3 4 8 3 Gatlin 86.4 0 3 19
West 40.9 1 3 6 9 3 Total 63.6 37 46 27
Gatlin 86.4 0 0 1 2 19
Total 42.7 26 11 18 30 25
of the variation in the data. The first analysis (Test 1, Table 6-4)
proved good for separating out the seasonal data for Lake Gatlin, but
showed poorer discriminating ability when trying to discriminate samples
from the other pools, especially between the west and east pools and the
south and middle pools (Table 6-5). This problem was due to the high
degree of similarity of the data of the contiguous pools. A considerable
improvement of the discriminant functions was obtained by combining
observations from the pools of the same lakes (i.e., the east and west
pools and the middle and south pools). The first two canonical variables
from this final analysis (Test 4) can be used to cluster the data into
three groups with some reliability as illustrated in Figure 6-4.
While conductivity was the best variable in terms of discriminating
the lake pools, this variable reflected changes in the hydrologic regimes
more than actual changes in lake trophic state. The discriminant analysis
showed that Secchi disk transparency was almost as good as a discrimina-
tor by itself and also reflected changes in the trophic condition in the
system with some regularity. For this reason it was felt that the
transparency alone could be used to indicate changes in the trophic
state of the Lake Conway system (see Figure 6-1). This is convenient
since an existing index, Carlson's index, can be used and the results
can readily be compared with other lake systems.
Comparison of Trophic Indicators and Changes
in Phosphorus Loadings
Since the primary purpose of developing a trophic index was to ex-
amine temporal changes in trophic changes in the system and relate them
to phosphorus loadings, an area weighted average of Secchi disk
,1,-/ 1 5 0 ,,// ,
.o,...I .. .. I .. . .o ). ...I..le t . 1 ,. .. .,. .... I .. .... e ... .o Qo B. . '. ..> .. 1 Q. . ... . . g .e ...* .. .. <.. .. .. I ..
-4.0 -2.0 0.0 2.0 4.0
CANONICAL VARIABLE 1
Figure 6-4. Canonical variables resulting from the discriminant analysis on the six trophic state
transparency for the entire system was calculated, and the corresponding
trophic state index was computed using Carlson's index. This index was
then compared to the phosphorus loading of the system as illustrated in
Figure 6-5. This comparison indicates that of the three sources of
loading derived from the model, the trophic response of the lake seems
to follow loading from external sources more closely. However, there
appears to be a time lag of approximately one month from the time of
maximal external loading to the peak in the trophic state index. What
was surprising was it appeared that the loading, due presumably from the
plants, does not seem to have an obvious influence on trophic state.
These observations are further supported by a correlation between
the trophic index to loadings from plants and external sources (Table
6-2). Total phosphorus and loading from plants show a significant
positive relationship, a finding which supports the dynamic model.
Significant negative correlations were found between external loadings
and loadings from plants and between total phosphorus and Carison's
TSI, the latter of which is indeed a surprising observation. These
findings suggest that while phosphorus loading from macrophytes does
play a major role in the total phosphorus cycle in the Lake Conway
ecosystem, the phosphorus released by the plants does not seem to be
readily available to phytoplankton communities. At first this seems to
be in conflict with the observations of authors (Nicholls and Keeney
1973) who showed that the compounds released by macrophytes are rapidly
mineralized to orthophosphate, the form readily utilized by phytoplank-
ton. A possible explanation is that some of the phosphorus released by
the plants is a refractory form. Several studies have indicated that
oExternal P loading
*Carlson's TSI (SD)
J J A SO N D J F MAM J J A S 0 ND J FM
1976 1977 1978
Figure 6-5. Seasonal variations of Carlson's Trophic State Index (TSISD) and phosphorus loadings
to Lake Conway from external sources.
much of the dissolved organic phosphorus fraction in natural waters is
a form that is not readily hydrolyzed. Herbes et al. (1975) tested
soluble unreactive phosphorus (SUP) fractions in lake water with several
phosphatases and found that only wheat bran phytase (a hexaphospho-
esterase) hydrolyzed the natural organic phosphorus in the water. This
confirmed the earlier observations (Rigler1961; Strickland and Solorzano
1966) that few monophosphatase hydrolyzable compounds are present in
natural waters for any length of time and that degradation of these com-
pounds is very rapid. Presumably, the group of compounds that are
responsible for this refractory nature are the hexaphosphate and penta-
phosphate esters of inositol (Eisenreich and Armstrong 1977). McKercher
and Anderson (1968) found that organic phosphorus in soils is made up
of 1 percent phospholipids, 5-10 percent nucleic acids, and their degra-
dation products and up to 60 percent as inositol polyphosphate compounds
conmmnonly known as phytin or soil phytate. This group of compounds is
readily synthesized by both plant and bacteria (Cosgrove 1975). Roberts
and Loewus (1968) found that most of C-14 labeled inositol ended up as
the hexaphosphate esters in an axenic culture of Wolfiella floridana,
demonstrating that this is an important metabolic pathway in plants.
Thus it is reasonable to expect that a portion of the organic released
by macrophytes are of this group of poorly hydrolyzed compounds. This
could explain why only 20 percent of the potentially available phos-
phorus from plants was necessary to balance the dynamic model. A large
portion of this phosphorus probably remains in a tight cycle in the
littoral zone, being rapidly hydrolyzed and absorbed by the epiphytic
community while the more refractory inositol polyphosphates are advected
into the limnetic portion of the lake. Since several groups of
microorganisms have been shown to produce the necessary phosphatases
(Aspergillus and Pseudomonas) (Kieslich 1976), the compounds will eventu-
ally be broken down and removed from the system but at a rate which
probably would not be fast enough to support large populations of
Conceptual Model of Phosphorus in Lake Conway
To integrate the information provided by the loading analyses,
phosphorus model, and trophic state analysis, it is best to utilize a
conceptual model of phosphorus dynamics for Lake Conway (Figure 6-6).
Only the major flows between the total phosphorus in the water column,
macrophytes, epiphytes, and sediments are shown. Minor flows and flows
to other ecosystem components are not shown for sake of simplicity. It
must be realized that while this model is of considerable heuristic
value, caution should be used when applying the results presented here
to other systems.
The major fluxes of phosphorus to the epilimnion are from external
sources and macrophytes, the latter being more important. While a large
flux of phosphorus is possible from the sediments, the flows from the
dynamic mass balance model would indicate that this is not the case. If
the model is correct, most of the phosphorus released from the sediments
in Lake Conway is returned, possibly via reabsorption and flocculation,
and it contributes only slightly to the epilimnetic phosphorus pool.
The model also suggests that most of the phosphorus released from
macrophytes is rapidly removed, most probably in the littoral areas.
Whether this phosphorus is taken up by the epiphyte community, trans-
located back to the roots or sedimented directly cannot be confirmed,
2.36* EPI- SEDIMENTS
Figure 6-6. A model of major phosphorus flows in the Lake Conway system. All flows are in g/m -yr.
*Combined flows. flncludes flows through other trophic levels.
xml version 1.0 standalone yes
ErrorText 90 CW: 0149-0154