Quantitative impacts of lake-level stabilization on sediment and nutrient dynamics


Material Information

Quantitative impacts of lake-level stabilization on sediment and nutrient dynamics coupling limnology with modeling
Physical Description:
viii, 197 leaves : ill. ; 29 cm.
Gottgens, Johan F
Publication Date:


bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1992.
Includes bibliographical references (leaves 160-173).
Statement of Responsibility:
by Johan F. Gottgens.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001796790
notis - AJM0503
oclc - 27483439
System ID:

Full Text








I wish to thank Drs. Clay L. Montague and Thomas L.

Crisman, chair and cochair of my committee. Both have been

mentors to me and our numerous, stimulating discussions

provided the inspiration to complete this research. I also

wish to acknowledge the other members of my committee, Dr.

Joseph J. Delfino, Dr. Frank G. Nordlie, and Dr. H.

Franklin Percival, whose critical reviews and comments

helped produce the final manuscript. Drs. Delfino and

Nordlie provided me with employment in teaching and

research which was financially essential and professionally

rewarding. Dr. Percival was helpful in securing the use of

an airboat, motorboat, and LORAN.

Drs. Peter Feinsinger and Carmine A. Lanciani helped

me develop a genuine appreciation for ecology. Dr. W.

Emmett Bolch and Ryan Richards introduced me to the

principles of y-ray spectroscopy. Dr. Danuta Leszczynska

and Ulrike Crisman assisted with the nutrient analyses, and

Dr. Mark Brenner allowed me the use of his coulometer for

the analysis of total carbon. Dan Peterson provided field

assistance in the, at times, forbidding Florida aquatic

environment. I also thank Eleanor Merritt and Sandra James


for their dependable help in maneuvering through an even

more treacherous, bureaucratic environment.

Throughout this work I have benefitted from a

productive working relationship with staff members from the

Florida Department of Natural Resources, the Florida Game

and Fresh Water Fish Commission, and the St. Johns River

Water Management District. Financial support from the

District for some of my unorthodox research ideas was


I particularly wish to thank my parents, Mr. F.L.J.

G8ttgens and Ms. M. G8ttgens-Veninga for teaching me an

appreciation for natural sciences and for their continued

encouragement. Finally, my wife Brigitte and my children

Ida and Leo deserve special recognition for tolerating the

demanding schedule of my graduate career. Most of all,

their caring support gave me the determination needed to

complete this work.




ACKNOWLEDGEMENTS..................................... ii

ABSTRACT.............................................. vi


1 INTRODUCTION................................ 1

Effects of Spillways....................... 1
Related Studies............................. 4
Research Methods: Coupling Limnology with
Modeling...... .... ..... .............. 6
Study Area................................. 9


Materials and Methods...................... 18
Collection of Sediment Cores........... 18
Radio-isotope Analyses................. 20
Bulk Density, Nutrient, and Carbon
Analyses ....................... 24
Results and Discussion..................... 25
Summary and Conclusions.................... 36

A SHORT-TERM DRAWDOWN................. 37

Materials and Methods....................... 38
Discharge, Rainfall, and Lake Stage.... 38
Water Quality Analyses ................ 40
Sediment Oxidation..................... 41
Sediment Compaction.................. 43
Results and Discussion .................... 44
Discharge, Rainfall, and Lake Stage.... 44
Removal of Particulate Matter and
Nutrients ......................... 48

Oxidation of Organic Matter in Exposed
Littoral Sediments................ 53
Consolidation of Littoral Sediments.... 56
Summary and Conclusions..................... 56


Materials and Methods ...................... 61
Sediment Cores ....................... 61
Bulk Density and Nutrient Analyses..... 64
Radio-isotope Analyses................ 64
Results and Discussion..................... 65
Littoral Cores......................... 65
Profundal Cores........................ 69
All Stations........................... 78
Uncertainty Analysis................... 79
Summary and Conclusions...................... 84


Methods.............................. ... .. 93
Feedback Dynamics........................ 93
Dynamic Hypothesis..................... 95
Results and Discussion...................... 124
Standard Run Output.................... 124
Sensitivity Analysis................... 130
Analysis of Alternative Management
Strategies........................ 143
Summary and Conclusions.................... 149

6 SUMMARY AND DISCUSSION...................... 153

LITERATURE CITED...................................... 160


A LEAD-210 (210PB) AND CESIUM-137 (137CS)
SPECTROSCOPY ......................... 174

B MODEL PROGRAM LISTING...................... 182

BIOGRAPHICAL SKETCH.................................. 197

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



Johan F. Gottgens

August 1992

Chair: Dr. Clay L. Montague
Cochair: Dr. Thomas L. Crisman
Major Department: Environmental Engineering Sciences

Spillways at lake outlets are used to reduce water-

level fluctuations and promote year-round lake access.

Such hydraulic alteration may cause accelerated

sedimentation in the lake. The impact of a 1967 spillway

was quantified in eutrophic Newnan's Lake (Florida) using a

model and experimental work.

A transect of sedimentary profiles, dated with 210Pb

and 137Cs by y-ray spectroscopy, showed threefold increases

in accumulation rates of organic matter, total Kjeldahl

nitrogen (TKN), and total phosphorus (TP) 1200 m lakeward

of the spillway since its construction. Concentrations of

TKN and TP increased 3.5 and 2.4 times, respectively, in

sediments deposited since 1967. These increases were


progressively less at stations farther from the spillway.

Postspillway accumulation of TP was focused toward the dam,

while recent TKN deposition was similar lakeside.

A 90-day spillway removal flushed 60 tons (dry weight)

of sediment from the lake. Discharge concentrations of

particulate organic matter, TKN and TP were highest during

the first month of the drawdown under adequate hydraulic

head. Storms stirred the water column and promoted

flushing of resuspended matter. Field and laboratory tests

did not show net oxidative removal of organic matter from

exposed lake bottom. Consolidated sediments remained

moderately firm after reflooding. Lowering the water depth

from 62 to 30 cm in the littoral zone removed 6.08

g m-2day-1 of organic littoral sediment, likely due to

increased wind-wave action on this substrate. Littoral

sediments with low bulk density eroded fastest and bulk

density of remaining substrate increased by an average of

250%. Redistribution of this material to deeper portions

of the lake was not demonstrated.

Results from experiments, literature and theory were

synthesized using feedback dynamics. The resulting

simulation model was most sensitive to changes in sediment

resuspension and actions of bacteria and benthic

heterotrophs on phosphorus dynamics in surface sediments.

Research aimed at quantifying these processes will enhance

confidence in the model. Separate impacts of the spillway


(decreased flushing, littoral zone vegetation, and sediment

resuspension) had minor implications for model behavior.

Combined, however, they increased seston, water-column

phosphorus, and sedimentation by 19, 39, and 27%,


In conclusion, after 25 years, intended benefits of

the spillway (lake access, navigation) have been replaced

with lake management "costs" such as increased turbidity,

primary production, and sedimentation.



Effects of spillways

Spillways have been built at the outlets of many large

shallow lakes in Florida during the past 50 years (Bishop

1967; Davis 1973). The majority of these water control

structures were designed to prevent low lake-stages during

the dry winter season characteristic of Florida's climate.

Year-round access to these lakes was promoted, and their

use for irrigation, recreation, and sportfishing maximized.

The structures varied from simple earthen dams to stoplog

weirs and concrete spillways with or without removable


The hydraulic modification produced by these dams may

produce short-term benefits for lake access and navigation,

but long-term consequences for lake ecosystems are poorly

understood. Stabilized lake levels may accelerate the

accumulation of sediment and nutrient-rich detritus in the

lake. This can occur in several ways.

First, the design of most spillways eliminates

discharge of bottom-water. This layer of water is rich in

solids due to sediment resuspended by wind-wave induced

currents (Sheng and Lick 1979; Paul et al. 1982). Outflow

is restricted to less turbid surface water flowing over the

crest of the spillway. During extended periods of low

rainfall when outflow of the lake is confined to evapo-

transpiration and losses to groundwater, suspended solids

are not exported.

Second, decreased water-level fluctuations reduce

exposure of littoral sediments to air, thus preventing

seasonal consolidation of this flocculent substrate and

reducing oxidative removal of organic matter. These

effects may be particularly pronounced in shallow lakes

with vast littoral areas. Such consolidation has been

documented when the water level in a lake was artificially

lowered for several months. Organic and nutrient-rich

sediments from Lake Apopka (Florida), with an initial water

content in excess of 80%, consolidated 40-50% after

exposure to rain and sun for 5 months (Fox et al. 1977).

Significant consolidation of the top 5-10 cm of exposed

littoral substrate was noted during the drawdown of Fox

Lake (McKinney and Coleman 1980). Oxidation of exposed

lake sediments has been suggested (Harris and Marshall

1963; Wegener and Williams 1974), but experimental evidence

is controversial. No significant oxidation was noted

during drying of Lake Apopka sediments (Fox et al. 1977),

while Holcomb et al. (1975) reported a considerable


reduction in organic matter content of exposed sediments in

Lake Tohopekaliga (Florida).

Third, stabilized lake levels may reduce wetland

habitat around the lake (Keddy and Reznicek 1985). Many

wetland plants require seasonal fluctuations in water level

for growth and reproduction (Gosselink and Turner 1978;

Bedinger 1979). Constant water levels result in a gradual

decline of such wetland taxa and may alter the wetland's

ability to filter nutrients and sediments carried into the

lake in runoff from the watershed.

Increased accumulation of flocculent sediment and

nutrient-rich detritus may significantly affect lake

ecosystems. Unconsolidated sediments are easily

resuspended (Lam and Jacquet 1976; Sheng and Lick 1979)

resulting in increased turbidity in the water column and

reduced light penetration. This may change plant

communities and reduce submerged vegetation. Periodic

resuspension of sediment and associated pore-fluid also may

lead to increased availability of nutrients for algal

utilization in the trophogenic layer (Holdren and Armstrong

1980; Pollman 1983). Organic and inorganic suspended

matter exert considerable biological and chemical oxygen

demand (Hargrave 1969; James 1974) which stresses

heterotrophic communities in the lake and promotes nutrient

release from the sediments (Mortimer 1971; Theis and McCabe

1978; Stauffer 1981). Deposits of flocculent sediment in

the littoral zone eliminate firm substrate for rooted

macrophytes, thereby reducing refuge and feeding habitat

for fish. Such accumulations may also reduce the

availability of preferred nesting habitat for centrarchids,

such as the largemouth bass (Micropterus salmoides) and

black crappie (Pomoxis niqromaculatus) (Bruno 1984), which

rely on firm substrate for deposition of eggs (Eddy and

Underhill 1978).

Related Studies

In spite of these plausible effects, little is known

about the long-term response of a lake to damming of its

outlet. Much work has focused on short-term responses

following drastic water-level manipulations such as the

filling of a reservoir by damming of a stream or the

drawdown of an existing impoundment. Many of these

investigations are reviewed in Baxter (1977) and Cooke et

al. (1986). Those applicable to this work are referred to

in Chapters 3 and 4.

Long-term investigations have concentrated on changes

in littoral flora following water-level stabilization and

the possible effects of such a reduction in high water-low

water perturbation on the ontogeny of aquatic ecosystems.

Symoens et al. (1988) reported a reduction in species

diversity in the littoral zone of eutrophic Lake Virelles

(Belgium) due to water-level stabilization by an outflow

control structure. Plants characteristic of the

periodically exposed shores disappeared. These included

sedges, rushes, and members of the buckwheat family.

Fluctuating water-levels increased the diversity of

vegetation types and plant species along the shoreline of

the Great Lakes (Keddy and Reznicek 1985). Low water

periods allowed many mud-flat annuals and emergent marsh

species to regenerate from buried seeds. Water-level

stabilization in Lake Tohopekaliga (Florida) eliminated a

considerable portion of the vegetated flood plain (Holcomb

and Wegener 1971).

Effects of stabilizing water-levels on the ontogeny of

aquatic ecosystems have been described for wetlands

(Gottgens and Montague 1988; Gunderson 1989; Walters et al.

1992), aquaculture systems (Odum 1971), and floodplain

vegetation (Goodrick and Milleson 1974). Freshwater

marshes, such as the Florida Everglades, appear to be

maintained at an early successional stage by seasonal

fluctuations in water levels and other pulsing factors

(e.g. fire, grazing). Aerobic decomposition of accumulated

organic matter may be promoted when sediments are exposed

during the dry season. Nutrients, released upon

reflooding, support a wet-season bloom in productivity. A

fluctuating water regime may be a dominant force in

preventing organic matter build-up and maintaining an

aquatic ecosystem in a state of pulsed stability (Odum


Research Methods: Coupling Limnology with Modeling

Two research approaches are convenient in an

investigation of long-term changes in a lake following

spillway installation. Both paleolimnology and ecosystem

modeling work with flexible time-frames and allow

investigation of long-term and slow-acting processes.

Paleolimnology is particularly useful when a detailed

historical data base is absent.

Paleolimnologists assume that lake sediments

accumulate in an orderly fashion through time. These

sediments contain a relatively stable, historical record of

past conditions of the lake-watershed ecosystem. This

record may be in the form of physical, chemical, and/or

biological signals. The timing of these signals requires

dating of levels in sediment cores which can be

accomplished, for instance, by measuring the content of

radioactive carbon (14C), lead-210 (210Pb), or cesium-137

(137Cs). The age of recent sediments (up to 120 years) can

be estimated by the content of short-lived isotopes (e.g.

210Pb) (Eakins and Morrison 1978; Krishnaswami and Lal

1978). In this work, paleolimnological techniques are

applied to quantify the impact of a 1967 spillway on

sediment composition and net rates of sediment and nutrient

accumulation in Newnan's Lake (Chapter 2).

Knowledge of past lake conditions based on

interpretations of the sedimentary record is not only

useful to evaluate rates of change in the lake (such as

loss of volume), but it is also helpful in the development

of lake management models, because it can assist in

reconstructing the reasons for past changes in the lake.

Models of different hypotheses of the causes of past lake

responses (such as an increase in the rate of

sedimentation) to perturbation (such as the construction of

a dam) may be used to predict the effect of future

management actions. One of the objectives of the

simulation model presented here (Chapter 5) is to evaluate

the long-term effect of spillway construction in the outlet

of Newnan's Lake on variables such as sedimentation rate,

lake depth, and concentrations of nutrients and suspended

matter in the water column. In addition, the model can

provide options for lake management by testing the effects

of alternative management strategies (such as a modified

dam or a drawdown) on intended uses of the lake.

The scheduling of a short-term, experimental drawdown

in Newnan's Lake in the spring of 1989 provided an

opportunity to investigate the impact of this management

strategy. This investigation focused on the effect of the

drawdown on littoral and profundal lake sediments for two

reasons. First, the accumulation of flocculent and

nutrient-rich sediments in the lake may have significant

implications for the lake ecosystem (as outlined earlier).

This material may be removed during a drawdown by flushing

of suspended matter and/or oxidation of exposed littoral

substrate. Both these processes were quantified in this

work (Chapter 3). Second, by focusing on the sediments,

the effectiveness of the drawdown to accomplish its main

objective may be determined. The main objective was to

improve littoral habitat for fish growth and recruitment

(Krummrich, Florida Game and Fresh Water Fish Commission,

pers. comm.). This improvement may occur when

consolidation, and/or removal of flocculent littoral

substrate promotes the establishment of littoral plant

communities which may, in turn, provide refuge and feeding

habitat for fish. Firm littoral substrate also provides

preferred nesting habitat for species of the sunfish family

(Centrarchidae) (Bruno 1984), economically important

sportfish in Newnan's Lake. This study measured both

consolidation of exposed littoral substrate, and erosion

and subsequent "sloughing" of this material to deeper parts

of the lake (Chapter 4).

Finally, studying processes in the lake with and

without the spillway complements the modeling exercise. It

assists in determining coefficient values for modeling the

effect of the spillway on the lake, and it allows field

testing of the model by comparing the results of the

drawdown study with the results of the same change made in

the model. As such, the limnological work assists in model

development and analysis. By forcing the quantification of

processes in the lake, the modeling exercise may reveal

gaps that exist in our understanding of this ecosystem and

identify necessary field or laboratory investigations.

Study Area

Newnan's Lake is located 8 km east of Gainesville,

Florida, within the Orange Creek Basin (Figure 1-1). The

geomorphology of this basin was first described by Pirkle

and Brooks (1959). It is dominated by the Hawthorne

formation, a marine deposit of Miocene age consisting of

phosphatic sands, clays, and limestone. It is relatively

impermeable compared with the underlying limestone and acts

as a confining layer. Surface flow, subsurface flow

through sands and clayey-sands, and direct rainfall are the

major sources of water for the lake (Canfield 1981). The

drainage area north of the lake supplies surface water

inflow via Hatchet Creek, Little Hatchet Creek, and several

smaller streams. The low mineral content of the lake water

suggests a small input from groundwater. Loss of water

from the lake is dominated by evapo-transpiration in

Florida's warm climate. Of the 130 cm of rainfall received

by the Orange Creek watershed annually (Adkins 1991), only

10% leaves as surface drainage (Clark et al. 1964).

Land-use in the basin is dominated by forest (Table 1-

1), including large tracts of commercial pine (Pinus

elliotti) plantations. Extensive areas of undeveloped pine

flatwoods also occur, with riverine hardwood-swamp and



0 2 4 1 1 10


Figure 1-1. Drainage map of the Orange Creek Basin.

cypress (Taxodium distichum) along the water courses. A

prominent fringe of cypress trees borders the lake.

Table 1-1. Land-use data for
the Newnan's Lake watershed.
Values in percent of total
watershed area.

Urban 8.6
Forest 1) 75.6
Agriculture 7.7
Open water 8.0
Wetland 0.1

Source: Huber et al. 1982.
1) Includes areas dominated by
cypress trees.

Newnan's Lake has a surface area of approximately 3000

ha and a maximum and mean depth of 3.6 m and 1.5 m,

respectively (Nordlie 1976) (Table 1-2). It is a flat

bottom basin with a few depressions where the water depth

exceeds 2.5 m at average lake stage (Figure 1-2). The lake

has a single major surface water outlet, Prairie Creek,

which drains to the south (Figure 1-1) through an area with

extensive marsh and bottomland hardwood communities. Prior

to the construction of Camp's Canal (in the early 1930s)

Prairie Creek discharged into Payne's Prairie (Figure 1-1).

This bypass canal was constructed around the eastern margin

of the prairie to control flooding and manage the land for

cattle grazing (Gottgens and Montague 1988). The canal












0 HP


directs the water around the prairie on a sluggish course

to the River Styx and subsequently to Orange Lake (Figure

1-1). The prairie has been managed as a state park and

marsh preserve since 1970 and a limited amount of water

from Prairie Creek has been diverted back into the


Table 1-2. Morphometry of Newnan's Lake.

Surface Area 3042 ha
Maximum Depth 3.6 m
Mean Depth 1.5 m
Development of Shoreline 1.09
Drainage Basin Area 308 km2
Lake Volume 58 x 106 m3
Lake Detention Time 0.6 yrs

The outflow from Newnan's Lake through Prairie Creek

has been regulated with a spillway since 1967. This 50 m

long structure consists of removable flashboards down to

the channel bed. The elevation of the top of the spillway

boards prevents surface discharge when the water level in

the lake drops below 20.1 m (MSL). During this study, one

of the top boards was missing, lowering the elevation of a

1.2 m section of the top of the spillway boards to 19.9 m


A comparison of pre- and postspillway stage-duration

curves indicates a reduction in the range of water-level

fluctuations of approximately 32 cm or 30% (Skoglund 1990;


Adkins 1991). This reduction is more pronounced if severe

prespillway droughts are included, during which 'o water

level records were kept (1954-56 and 1962-64). In this

shallow lake, with a very small elevation gradient in the

littoral zone, such reduced water-level fluctuations may

significantly affect seasonal exposure of littoral

sediments and the area of vegetated littoral habitat. The

spillway has raised the average stage of the lake by 13 cm,

i.e. it has increased the average water depth in the lake

by 9%.

Newnan's Lake may be classified as an eutrophic, soft

water lake (Canfield 1981). Water quality data show high

nutrient concentrations, high and variable chlorophyll a

values, and low Secchi disk transparency (Table 1-3). High

color (often in excess of 150 Pt units) occurs as a result

of input of water rich in dissolved humic materials from

the surrounding flatwood and cypress communities.

Temperature and dissolved oxygen data for Newnan's

Lake were also typical of shallow, productive systems.

Vertical changes in water-column temperature rarely

exceeded 3C, regardless of season (Crisman 1986a).

Consequently, thermal stratification rarely, if ever,

occurred. Daytime surface water during summer and early

fall was often supersaturated, while the water column in

the deeper portions of the lake below 3 m approached anoxia

-, 7


Sa o I.
Am 01t

00 ^ 0

66s .
00 d C

.o Do o
..- .- -

aoo U)

0R co0 0

-i -o

-o ad CT
0 O o co
CO) C6 0) 0

(0 CV 0 (00)
0 Cm

1 1 1CM
0) C00
I c0 00 CM

- 0
tc d d




ca- .53E



Ss i OI--o'
a. CL -01) ZZI- L

9. I

So O














as a result of intense decomposition in surficial sediments

(Crisman 1986a).

Using the data sources from Table 1-3, the ratio of

water-column N:P (in grams) varied from 12.4 to 27.6. This

ratio is approximately 7:1 in aquatic plant material

(Vallentyne 1970). Therefore, these water quality data

indicate that phosphorus is generally less available than

nitrogen relative to their content in aquatic plant

material, and that primary production in Newnan's Lake is

likely limited by phosphorus.

The lake had dense growths of filamentous algae

dominated by blue-greens (Cyanophyta) such as Spirulina

sp., Anabaena sp., and Aphanizomenon sp.(Nordlie 1976;

Crisman 1986a). At times, macrophytes dominated by exotics

such as water hyacinth (Eichhornia crassipes) and hydrilla

(Hydrilla verticillata) were abundant. Since the early

1970s, periodic applications of herbicides have been made

to restrict the spread of these exotics. The lake bottom

was covered with a homogeneous layer of highly flocculent,

organic sediment (Holly 1976). A review of existing

literature and an inventory of the existing data base for

Newnan's Lake is given in Gottgens and Montague (1987a),

and in its accompanying categorized bibliography (Gottgens

and Montague 1987b).


In Chapter 1, it was hypothesized that spillways

accelerate the accumulation of sediment and nutrient-rich

detritus in a lake and that such increased trapping of

material may have considerable long-term impact on the lake

ecosystem. Little is known, however, about increases in

the rate of sedimentation caused by damming of a lake

outlet. Measurement of this rate is essential for

determining loss in lake volume and quantifying mineral

cycling in lakes, including the modeling of nutrient

concentrations in water (Dillon and Rigler 1974:

Vollenweider 1976). It is also needed for the development

of a sediment or pollutant budget.

Material accumulation rates can be calculated by

determining geochronology in the sediment profile using

radioactive decay of fallout ("unsupported") lead-210

(210Pb) following burial in sediments (Appleby and Oldfield

1978). Matching the occurrence of cesium-137 (137Cs) in

the profile with the onset of widespread atmospheric

nuclear testing in the early 1950s (Ritchie et al. 1973)

provides an additional age-marker. These paleolimnological


techniques have been used widely to detect changes in

sediment accumulation rates caused by urban development in

a watershed (Smeltzer and Swain 1985), major storm events

(Robbins et al. 1978), and deforestation (Oldfield et al.

1980). They are particularly helpful in gaining insight

into past conditions in lakes when a long-term data base is

absent. Here, they were applied to quantify the impact of

a 1967 spillway on sediment composition and net

accumulation rates of organic matter, nitrogen, and

phosphorus in Newnan's Lake.

Materials and Methods

Collection of Sediment Cores

Core locations were arranged in a transect of

increasing distance lakeward from the spillway (Figure 2-

1). Cores 8, 6, and 1 were taken, respectively, 1200,

3000, and 4500 m lakeward of the structure at a water

depths of 115, 150, and 150 cm. The cores were collected

from the lake in the spring of 1989 using a Livingstone

piston corer (Livingstone 1955) equipped with 4.1 cm

diameter cellulose butyrate tubes. The piston was

positioned at the lower end of the coring tube which was

then lowered to the level at which sampling was to

commence. During this operation the piston cable was payed

out freely, but once the sampling level was reached the

cable was secured to the boat to prevent the piston from

moving any farther down. The tube was driven into the

Figure 2-1. Insert map of Florida showing location of
study area and map of Newnan's Lake with core sites
indicated (filled circles).


sediments by pushing on the extension rod while the piston

prevented the sample from being compressed or lost upon

retrieval. After the apparatus was hoisted to the surface,

appropriate caution was exercised to preserve the sediment-

water interface. Cores were kept refrigerated until they

were sectioned. The cores were sectioned in 1 cm

intervals, which were stored in plastic bags at 40C.

Radio-isotope Analyses

Concentrations of 210Pb and 137Cs were measured by

direct y-assay using a coaxial N-type, intrinsic-germanium

detector (Princeton Gamma Tech). The counting system used

for spectral analysis (Figure 2-2) is located in the

University of Florida's Department of Environmental

Engineering Sciences' Low Background Counting Room. This

room was designed to reduce the level of background

radiation interference during sample counting. An outer

shield (0.95 cm steel), main shield (10.1 cm lead), and an

inner lining (0.05 cm cadmium + 0.15 cm copper) were used

to reduce background radiation at the germanium detector.

The detector has a 38% efficiency (at 1332.5 keV) and

operates at a temperature near that of liquid nitrogen

(-196C), which is provided by a high-vacuum cryostat dewar

system (capacity 30 1). This thin-window (beryllium),

standard detector has a relatively low efficiency resulting

in the need for long count times. This type of detector,

however, counts over a large range of y-energies, is


41 .

ES4 0
Q.cB P






(1) 0.



a ...

.4 >


4 0



. -.4

*pr C
0 (

suitable for many different sample configurations, and

therefore more generally useful. It also detects

efficiently other gammas from the 238U decay series and can

thus be used to determine supported and unsupported levels

of 210Pb simultaneously (Appleby et al. 1986; Nagy 1988).

The electronics include a preamplifier (RG11B/C, Princeton

Gamma Tech, Inc.), amplifier (TC 242, Tennelec), bias

supply (5 kV, TC 950, Tennelec), power supply (TC 909,

Tennelec), and transformer (Sola). A Zenith (Z159)

computer with multi-channel analyzer ("Maestro" ADCAM 100,

EG & G Ortec) and mathematical spreadsheet (Quattro,

Borland Inc.) was used for data conversion and analysis. A

new procedure developed for y-ray spectroscopy using the

Department's recently acquired P-type, intrinsic germanium

well detector is given in Appendix A.

Samples for isotope analysis were dried at 950C for 24

hours, pulverized by mortar and pestle, weighed, and placed

in small plastic petri-dishes (#1006 Falcon). Core

sections were combined (up to 4 cm) to obtain an adequate

sample weight (generally more than 1 gram). Petri dishes

were sealed with plastic cement and left for 14 days to

equilibrate radon (222Rn) with radium (226Ra). Counting

times varied from 14 to 45 hours depending on sample

weight; small samples needed longer counting times to

minimize uncertainty. Blanks were counted for every two

samples to determine background radiation. Standards were


run with the same frequency to track efficiency (counts/y)

and to calculate a 226Ra conversion factor (pCi/cps).

Sample spectra were analyzed for activity in the 46.5 kev

(210Pb) and 662 kev (137Cs) peaks. Activities at 295 kev

(214pb), 352 kev (214Pb), and 609 kev (214Bi) representing

uranium series peaks were used to compute supported levels

of 210Pb.

Calculation of 210Pb dates followed the constant rate

of supply model (Goldberg 1963; Appleby and Oldfield 1978)

which is able to quantify changing sediment accumulation

rates. In this model, the cumulative residual unsupported
210Pb, At, beneath sediments of age t varies according to

At = A0 e-kt (2-1)

where A0 is the total residual unsupported 210Pb (pCi/cm2)

and k is the 210Pb radioactive decay constant. At and Ao

are calculated by numerical integration of the 210Pb

profile. The age of sediments of depth x is then given by

t= in (2-2)
k At

The sedimentation rate (r) can be calculated directly

(Appleby & Oldfield 1978):

r= A (2-3)

where C is the concentration of unsupported 210Pb (pCi/g)

in the sediment layer of interest. This model appears

applicable to Florida lakes, particularly because 210Pb

residuals match both the known atmospheric flux of this

isotope as well as the residuals of nearby cores (Binford

and Brenner 1986; Gottgens and Crisman 1992).

Bulk Density. Nutrient, and Carbon Analyses

Bulk density (mg/cm3) and organic matter content

(mg/mg) were determined on 1 cm3 subsamples for as many as

45 intervals per core. Bulk density was measured by drying

of subsamples at 95C for 24 hours, cooling under

dessication and weighing. Organic matter content was

estimated by weight loss on ignition at 5500C for 1 hour

followed by rehydration with distilled water and re-drying

at 950C for 24 hours (method 209D, A.P.H.A. 1985).

Nitrogen was measured as total Kjeldahl nitrogen (TKN)

using a Technicon-II semi-automated manifold after

digestion following Bremner and Mulvaney (1982), but

modified to exclude selenium as a catalyst. The digestate

was also used for total phosphorus (TP) determinations.

Liberated orthophosphate was determined with the ascorbic

acid method (method 424F, A.P.H.A. 1985) using a Milton Roy

spectrophotometer (Model 20) with a 1 cm light path.

Between 15 and 20 intervals per core were analyzed for

these nutrients. Total carbon in the sediments was

measured by combustion of the sample at 900C followed by

titration of the evolved CO2 (Lee and Macalady 1989) using

a model 5011 coulometer (Coulometrics, Inc.). Total

organic carbon was measured using a combustion temperature

of 5000C.

Results and Discussion

Results of paleolimnological analyses may be presented

in units of concentration or as rates of accumulation.

Concentration, expressed as a relative measure of sediment

composition (e.g. mg/g dry weight), is the conventional way

of expressing sediment stratigraphy (Shapiro et al. 1971;

Pennington 1973; Griffiths and Edmondson 1975; Kramer et

al. 1991). Such data, however, are vulnerable to

variations in sedimentation of other components in the

profile. For example, increased deposition of allogenic

inorganic material lowers the concentration of organic

matter in the sediment. Effects of such dilution can be

eliminated from the calculations using ratios of components

in the sedimentary matrix. Accumulation rates, on the

other hand, are normalized to time thus avoiding the

problem of co-variance among different sedimentary

components. Detailed sediment dating with acceptable

analytical precision, necessary for this procedure, can now

be done with direct 210Pb low-background y-assay (Appleby

et al. 1986; Gottgens and Crisman 1992). This technique

eliminates uncertainties associated with 226Ra supported

activity and provides additional independent data markers

in the profile via a simultaneous assay for other y-

emitting isotopes, including 137Cs. Nonetheless,

accumulation rate calculations are numerically sensitive to

small errors in 210Pb measurements. The uncertainty

associated with this may be reduced by averaging

accumulation rates over longer periods of time. In this

study, both concentration and accumulation rate data are

given to provide complementary information about the recent

history of the lake.

Cores consisted mostly of homogeneous brown material

with organic matter content generally greater than 50% and

bulk density less than 70 mg/cm3 throughout the top 30-40

cm (Gottgens and Crisman 1992). Abrupt increases in the

organic matter, TKN, and TP profiles of core 8 at 33 cm

depth coincided with a 210Pb determined sediment age of

approximately 1967, i.e. the time that the lake outflow was

dammed (Figure 2-3a). Concentrations of TKN and TP

increased 3.5 and 2.4 times, respectively, in sediments

deposited since 1967. This signal was progressively less

at stations farther from the spillway (Figures 2-3b and c).

Increased nutrient concentrations at a 21oPb date of

approximately 1952-56 (Figure 2-3a, b, and c) matched the

driest year (1954) in north Florida's recent history. This

Organic Matter
(mg/9 d'y w*ght)
0 200 400 600


65 -
... /



Total PhosDhorus
(mg/g dry eiqht)
1 0 20 3.0

Total K. Nitrogen
(mQ/q dry weight)
0 5 10 1S 20 25




200 '00 600 20
0 1r

25 ------- ---- --



30 0 5.0 20 30 40 50

200 400 600 2.0 30 0 5.010 20 30 40 50

--------------------------------- *--------------- -----~-----
,.. 1954 / /

60 "
8 / /'

/5 i "_ U _






0 E




Figure 2-3. Core sites 8 (a), 6 (b), and 1 (c), Newnan's
Lake: Profiles for organic matter, total phosphorus, and
total Kjeldahl nitrogen. Spillway installation (---) and
first occurrence of 37Cs marker (***) indicated.




may have resulted from reflooding of exposed littoral

sediments and subsequent release of nutrients to the water


The first occurrence of detectable 137Cs, marking the

onset of widespread atmospheric nuclear testing in the

early 1950s, matched well with the determined 210Pb

chronology (Figures 2-3a, b, and c). This agreement is

remarkable considering the shallow character of the lake

and the flocculence of the bottom substrate which makes

these upper sediments vulnerable to physical disturbance.

Eutrophic systems such as Newnan's Lake, however,

accumulate sediment rapidly, so that these periodic

physical disturbances likely affect short time-intervals.

The bulk density profile for core 8 (Figure 2-4)

demonstrated both the fluid nature of the deposits and the

sudden, fivefold decrease in bulk density after spillway

installation. Bulk density in the top 33 cm of sediment

(e.g. 24 years) was less than 60 mg/cm3, and average net

accumulation is 1.3 cm per year. In most sediment

profiles, bulk density increases with depth due to

compaction. Yet, the profile illustrates a significant

transition at the time the lake outlet was dammed.

Unconsolidated sediments may be resuspended easily by

wind-induced currents in shallow systems such as Newnan's

Lake. The resulting organic turbidity increases sediment

oxygen demand (Bowman and Delfino 1980), particularly in

Bulk Density (mg/cm3)

D 500



S ..- Spillway installation
r ~ --"-----------------
-35+ 5

-75- "
0! =~


0R ;






- 60


- 70

Figure 2-4.

Bulk density profile for core site 8, Newnan's




subtropical lakes where high temperatures stimulate

bacterial respiration in organic substrate (McDonnell and

Hall 1969). In addition, resuspended sediments may

contribute significant quantities of nutrients to the

overlying water column through desorptive processes or

entrained pore fluid (Bengtsson 1975; Cooke et al. 1977).

Pollman (1983) reported a more than twofold increase in

mid-lake orthophosphorus concentrations under a moderate

wind event (approximately 8.9 m/s) in a central Florida

lake with a mean depth and sediment composition similar to

Newnan's Lake. Pronounced increases in the nitrogen and

phosphorus content of surficial sediments deposited since

spillway installation may enhance internal nutrient cycling

and, therefore, change plant communities and lake

productivity. For example, using data from Pollman's

(1983) sediment-dispersion nutrient-release model,

resuspension of the upper 1 cm of flocculent sediment at

station 8 would cause a 70% increase in the phosphorus

concentration of the overlying water through entrainment of

interstitial water and desorption from suspended sediments.

Carbon:nitrogen:phosphorus ratios (C:N:P) demonstrated

both temporal and spatial changes in the composition of the

deposits (Table 2-1). Measurement of total carbon (TC) and

total organic carbon (TOC) demonstrated that almost all

sedimentary C is organic (TOC = 0.996 x TC; standard error

= 0.003). Postspillway deposits were low in P compared


with material deposited prior to 1967. This reduction in P

content was most pronounced relative to N content at

stations 6 and 1 (Table 2-1). While many factors could be

considered in interpreting this difference, including

changes in fractionation of sedimentary P and variations in

Fe content of the sediments (Engstrom and Wright 1984),

such information was not collected for this work. The

reduced P content of recent sediments may be a reflection

of a change toward P-limitation of primary production in

the lake.

Table 2-1. Total carbon (C) to total
Kjeldahl nitrogen (N) to total phosphorus
(P) ratios (weight/weight) for cores 8,
6, and 1 pre- and postspillway
installation, Newnan's Lake (FL).


8 6 1

Post C:P 82 91 91
Pre C:P 40 44 43

Post C:N 13 10 8
Pre C:N 7 6 6

Post N:P 6 10 11
Pre N:P 6 7 7

Spatially within the lake, the amount of P in the

sediments relative to C and N was higher in "downstream"

areas (e.g. site 8). This may result from rapid

entrainment of P and reduced N:P ratios during windy

conditions in shallow systems (Hamilton and Mitchell 1988)

followed by downstream transport and density-dependent

settling of the fine particulates with which P is

associated (Engstrom and Wright 1984). High C:N ratios in

core 8 deposits may reflect a somewhat higher degree of

decomposition of bottom material in the lower portion of

the lake. The low C:N ratios lakeside (between 6 and 13)

imply an autochthonous origin of the organic matter in the

sediment matrix (cf. Wetzel 1983). The contribution of

non-Kjeldahl nitrogen to this ratio is small in anaerobic

sediments (Reddy and Patrick 1984).

Net accumulation rates of sediment, organic matter,

and nutrients have increased greatly close to the spillway

since its construction (Figure 2-5). This increase was

progressively less at sites farther from the spillway,

where accumulation rates corresponded well with mid-lake

values reported for lakes of similar trophic state (Deevey

et al. 1986: Binford and Brenner 1986). The highest rates

were found consistently at the station most distant from

significant material inflow in the northern portion of the

lake. A 45% higher sedimentation rate at station 8 amounts

to an additional 11 cm of material accumulation and a 9%

reduction in water depth in this shallow system during the

24 years since dam construction.

r-I %

41 0




0 r

04) 4

-U4 9
4) 0)-
m .5r-I
#-4 0 t

0 4 U' 4



U'- Q)H

4) U) .v-I

$41: w '

O H 0
-q0 r.
4-) 4) .--

0 134)
0 k'

to $4

4) m'--
z 0 4)
N f-I


f.1 to
rT4 w

CM 0
cl o

rl- c Cn
r,~ ul NM x

I --- a>





-- c 1
a o oI



0 o b
( _-U uu o



-a------------- ajD.c -uoil-ln wn--




. M





o 0

- 0




A pronounced increase in the accumulation rate of P at

station 8 was observed (3-3.5 times the measured rate at

stations 6 and 1), while recent N deposition is similar

among the three stations (Figure 2-5). This "focusing" of

P in sediments in the lower portion of the lake parallels

the decreased C:P and N:P ratios at this downstream

station. The higher rate of entrainment of P (compared to

N) by wind-induced wave action (Maceina and Soballe 1990)

likely contributes to P focusing in this shallow, exposed

lake. Its short water residence time and obstructed

outflow promote horizontal transport and subsequent

accumulation close to the dam.

Other factors may have contributed to increased

sedimentation, but their effect would have resulted in

changed sedimentation rates lake-wide, rather than

concentrated toward the spillway. These factors include

herbicide treatment of aquatic macrophytes and added

nutrient runoff from the watershed. Between 1972 and 1977

a total of 130 ha (i.e. 4%) of lake surface area was

treated chemically to reduce the abundance of Eichhornia

crassipes (water hyacinth) and Hydrilla verticillata

(hydrilla). In November of 1988 an additional 110 ha

received chemical treatment (Hinkle, Florida Department of

Natural Resources, pers. comm.). Analysis of aerial

photographs (U.S. Geological Survey 1966, 1988) did not

show significant land-use changes in the basin during the

last 22 years. The number of residences less than 1 km

from the Newnan's Lake shoreline increased from 16.0 to

19.2 units/km2 between 1966 and 1988.

Summary and Conclusions

The Newnan's Lake spillway, intended to stabilize lake

stage, changed material transfer between water and

sediment. This change, operating over a period of decades,

can be quantified using paleolimnological techniques. A

transect of 210pb dated profiles perpendicular to the dam

at the lake outlet demonstrated a distinct impact from its

installation in 1967 on net rate of material accumulation

and sediment composition. While net accumulation rate of

bulk sediment was only 45% higher, resulting in a 9%

reduction in water depth, rates for organic matter and

nutrients tripled compared to prespillway conditions. P

accumulation was particularly enhanced close to the

spillway compared with other stations. If allowed to

persist, increased material transfer between water and

sediments may produce loss of lake volume and changes in

benthic habitat. The threefold increase in the trapping of

nutrients may change plant communities and lake



A three-month gravity drawdown of the lake was

initiated on April 24, 1989 by removing all spillway

flashboards. The objectives of this drawdown were to

improve fish growth and recruitment and to flush nutrient-

rich detritus from the lake. Accumulation of flocculent

and nutrient-rich sediments in the lake (Chapter 2) and the

presence of this flocculent substrate in the littoral zone

may reduce the availability of preferred nesting habitat

for centrarchids, such as the largemouth bass (Micropterus

salmoides) and black crappie (Pomoxis nigromaculatus) (Eddy

and Underhill 1978; Bruno 1984). Exposing the littoral

zone lake-bottom during drawdown may allow oxidation and

consolidation of this substrate, which may improve this

habitat for future fish spawning. Seed germination and

firm substrate promote the establishment of littoral plant

communities, which may provide refuge and feeding habitat

and promote fish recruitment.

Research on other Florida lakes has quantified effects

of drawdowns on fish populations (Wegener and Williams

1974), aquatic invertebrates (Wegener et al. 1974), and

littoral plant communities (Holcomb and Wegener 1971;

Goodrick and Milleson 1974; Hestand and Carter 1975).

However, no information is available on the impact of

drawdown practices on the flushing of organic matter and

nutrients from Florida lakes. The objectives of this part

of the study were: (1) To measure the flushing of organic

matter, nitrogen and phosphorus in surface discharge from

Newnan's Lake prior to and during the gravity drawdown and

(2) to determine the effect of exposing littoral zone

sediments to air on oxidative removal of organic matter and

long-term consolidation of the substrate.

Materials and Methods

Discharge. Rainfall, and Lake Stage

Discharge through Prairie Creek, rainfall, and lake

water-level were measured starting February 20, 1989, 9

weeks prior to drawdown, and continued throughout the

drawdown period until July 31, 1989, 2 weeks after the

spillway boards were re-installed.

Surface flow from the lake was measured weekly using

an Ott flow-meter (Type C2-10150) approximately 30 m

downstream from the dam in Prairie Creek (Figure 3-1). The

frequency of sampling was increased during storm events to

include a minimum of 1 prestorm flow, 1 peak flow, and 1

poststorm flow measurement. The creek width was divided in

1 m subsections, and flow was measured in each section at

Figure 3-1. Map of Florida showing study area location and
map of Newnan's Lake. Open circles indicate the locations
of field enclosures, and closed circles indicate the
sediment collection sites for the oxidation/consolidation
laboratory experiments.

0.6 depth (when channel depth < 30 cm) or at 0.2 and 0.8

depth and averaged (when channel depth > 30 cm). Discharge

was computed by multiplying flow in each subsection by its

cross-sectional area. Total discharge equals the sum of

all subsection discharges. If the measured discharge

differed from the discharge computed using USGS

stage-discharge ratings by more than 10%, then discharge

was remeasured and averaged with the first measurement. A

calibration flume (Department of Civil Engineering,

University of Florida) was used to determine the accuracy

of the Ott flow-meter.

The Gainesville Flight Service Station (National

Oceanic and Atmospheric Administration) kept daily rainfall

records during the study period using a standard (20.32 cm

diameter) U.S. Weather Service raingauge, located near the

center of the watershed. Lake levels were recorded daily

by the U.S. Geological Survey (station # 02240900) located

on the west side of the lake.

Water Quality Analysis

Replicate water samples were collected concurrently

with discharge measurements 20 m upstream from the spillway

in the center of the creek. Water samples were collected

in acid washed Nalgene containers, immediately stored on

ice, and frozen within 1 hour of collection. Samples were

analyzed for total suspended solids (TSS), particulate

organic matter (POM), total Kjeldahl nitrogen (TKN), and


total phosphorus (TP). The analyses were completed within

90 days of collection.

Analysis for TSS was according to Standard Methods

(method 209C, A.P.H.A. 1985) using pre-weighed and

pre-muffled glass fiber filters (Whatman, 934-AH, 4.25 cm

diameter) with an effective particle retention of 1.5 gm.

Replicate analyses were done on 0.5 liter samples. POM was

measured as weight loss on ignition (at 5500C for 1 hour,

followed by rehydration with distilled water and drying at

950C for 24 hours) (method 209D, A.P.H.A. 1985) of oven-dry

samples (95C for 24 hours). TKN analyses were according

to E.P.A. method 351.2 (E.P.A. 1979) using a 40-cell block

digestor and a Technicon II semi-automated manifold. The

digested sample was also used for TP determination. The

liberated orthophosphate in the digested samples was

determined using the ascorbic acid method (method 424F,

A.P.H.A. 1985). Absorbance of the samples was read at 880

nm on a Bausch and Lomb spectrophotometer (Model 21) with a

light path of 2.5 cm. Both nitrogen and phosphorus

analyses were done on replicate samples and averaged.

Sediment Oxidation

Oxidation rates of organic matter in exposed littoral

zone sediment were measured in the laboratory and the

field. Littoral surface sediment was collected from 3

stations in the lake (Figure 3-1) and homogenized in a

blender in the laboratory. Subsamples (n=25) of 100 ml


each were incubated in crucibles at room temperature for 78

days under 3 treatments. One set was allowed to dry

completely and remained dry during the test period; one set

was kept wet without standing water by adding up to 8 ml of

distilled water per week, and a third set was kept

inundated (with 2-3 cm of distilled water). The duration

of this experiment (78 days) was chosen to coincide with

the expected length of time that littoral substrate would

be exposed in the field. Crucibles in the laboratory were

rotated every 48 hours to promote equal exposure to light

conditions (generally between 20-30 Lux, measured with a

Li-Cor photometer, Model LI 188). Organic matter content

before and after incubation was determined using weight

loss on ignition (at 5500C for 1 hour followed by

rehydration with distilled water and drying at 950C for 24

hours) of oven-dry samples (950C for 24 hours).

PVC enclosures (surface area = 180 cm2) were installed

in pairs at 3 locations in exposed littoral zone sediments

(Figure 3-1). The enclosures were pushed through the top

layer of wet organic substrate to a point 5-10 cm into the

underlying sand. Enclosed sediment was shielded from

precipitation by an elevated plexiglass roof. This design

prevented removal of enclosed substrate by water and/or

wind erosion. Hardware cloth (1.25 cm mesh size) was put

on top of the enclosure to deter animal disturbance.

Organic matter content at the beginning of exposure

was quantified by average weight loss on ignition of 3

"core" samples adjacent to the enclosure. Each sample

covered an area equivalent to the enclosure and was

analyzed in its entirety down to sand. After 28 days of

incubation, the organic matter content of the enclosed

substrate down to sand was determined. Net oxidative

removal (g m2day-1) was computed as the difference between

pre- and postincubation measurements.

Sediment Compaction

A small-scale sediment compaction study was also

performed to evaluate changes in substrate bulk density,

water content and percent organic matter at various times

after exposure and after reflooding. Four vessels (glass,

1 liter) containing 500 ml of homogenized littoral

substrate were allowed to dry in the laboratory for 51

days, then inundated (to a depth of 10 cm) for 149 days,

simulating the water regime in the field. Incubation of

the substrate occurred at room temperature, indoor light

conditions, and approximately 40% relative humidity. The

locations from which sediment samples were collected are

shown in Figure 3-1. Only the top 5 cm of substrate was

used. Ten water content (percent) and organic matter

content (percent) determinations were made at 5 times

during exposure, at the time of reflooding, and 4 times

during reflooding.

Results and Discussion

Discharge. Rainfall, and Lake Stage

During the 3.5 months prior to drawdown lake stage

dropped from 20.33 m (MSL) to 20.07 m (MSL) due to lack of

rain. The boards in the spillway were removed on April 24

(1989) after which the water level in the lake dropped from

20.07 m (MSL) to 19.80 m (MSL) in 1 month (Figure 3-2).

Lake stage remained between 19.70 and 19.80 m MSL for 8

weeks. This stage was the lowest recorded during the last

25 years, with the exception of a brief period during the

drought of 1981 (when the lake reached 19.65 m MSL). The

maximum drop in water level during drawdown was 36 cm.

Severe lack of rain during the months prior to drawdown

(Table 3-1) resulted in lake water-levels approximately 41

cm below average for early spring. This reduced the

amplitude of the drop in lake water-level. Small elevation

gradients in the basin and the gradual build-up of

obstructions in the lake outflow upstream and downstream

from the spillway since its construction (Crider 1972;

personal observation) likely prevented a more dramatic

drawdown of water level. Those obstructions to the flow in

Prairie Creek may have resulted from the impact of the

spillway on stage level and discharge volumes in the creek.

Low-flow and low-stage conditions downstream from the

spillway are much reduced compared to the natural

hydroperiod of the creek prior to dam construction in 1967

(Gottgens 1987).

S-Weir Opn --Weir Closed


S 20.1-

I 20.0-



(Source USGS]
19.7 I I I I I i
Jan Feb Mar Apr May June July Aug Sep

Figure 3-2. Water level record for Newnan's Lake for the
period of January-September 1989 [Source: U.S. Geological
Survey station # 02240900].

Table 3-1. Average rainfall (1897-1987) and rainfall
during first 7 months of 1989. Records given in cm.

Jan Feb Mar Apr May Jun Jul Total

Average 7.7 9.0 9.1 7.6 8.9 16.7 18.2 77.2
1989 2.6 3.0 5.5 7.4 5.2 24.4 10.0 58.1
Deficit 5.1 6.0 3.6 0.2 3.7 -7.7 8.2 19.1

[Source: National Oceanic and Atmospheric Administration, Gainesville
Flight Service Station, Florida)

Assuming a reduction of the mean water depth from 1.5

m to 1.2 m during the drawdown, it is estimated that 20% of

water volume of the lake was removed during the drawdown

period. Assuming unaltered base-discharge rates from the

lake through Prairie Creek during the drawdown period,

integration of the discharge curve (Figure 3-3) and then

subtracting base-discharge yields an estimate of 1.75 x 106

m3 of water flushed from the lake by removal of the

spillway. Base-discharge is defined as the discharge

flowing through the channel with the spillway in place, and

is estimated by measuring average discharge prior to dam

removal and after re-installation of the flashboards. In

both cases an averaging period of 2 weeks was used. Below

average rainfall during the drawdown period (Table 3-1) and

high evapo-transpiration contributed considerably to the

decrease in lake stage. Discharge through Prairie Creek

dropped to approximately twice the assumed base-discharge

within 6 weeks after spillway removal and remained at that


1.0 .

E .0.8- -


0.2 -

o.o4 -----I I----j ----- -
Feb March April May June July


Figure 3-3. Discharge (m3/sec) through Prairie Creek @
Florida S.R. 20 for the period of February-July 1989.

rate until closing of the weir (Figure 3-3). In spite of

low total rainfall, 5 storm events of low to moderate

intensity were included in the sampling period(Table 3-2).

Table 3-2. Precipitation and wind conditions during
sampled storm events, Newnan's Lake.

Dates (1989) Amount* Wind conditions*
(cm) (km/hr)

Pre- March 22-23 1.09 0-18 from N
drawdown April 14-15 2.03 0-12 from N or W

During May 29 2.26 0-40 misc. dir.
drawdown June 18-19 4.52 0-16 misc. dir.
July 16-17 1.42 0-16 misc. dir.

Measured at spillway with standard 2.54 cm glass raingage and Dwyer
handheld windmeter

Removal of Particulate Matter and Nutrients

Increased flow through Prairie Creek after spillway

removal increased discharges (kg/day) of particulate

organic matter (POM), total Kjeldahl nitrogen (TKN), and

total phosphorus (TP) (Figure 3-4a,b, and c). POM was a

consistent fraction of total suspended solids during the

sampling period (POM = 0.76 x TSS; N=64; R2=0.96).

Assuming base-discharge rates (kg/day) through Prairie

Creek during the drawdown period, integration of total

discharge rates and subtraction of base-discharge rates

yields an estimate of the amount (kg) of POM, TKN, and TP

flushed from the lake owing to drawdown (Table 3-3).

Particulate Organic Matter

Total Kjeldohl Nitrogen

Total Phosphorus

Total Kje!cchl Nitrogen

-ewr ooen weir clOfOs

Total Phosphorus

2 J

-** E

.o C



Feb Mar Apr May Jun Jul Feb Mar Apr 'lay Jun Jul

Figure 3-4. Characteristics of the discharge from Newnan's
Lake through Prairie Creek. Opening and closing of
spillway indicated. Discharge rate (kg/day) of POM, TKN,
and TP (Figure 3-4a, b, and c). Concentration (mg/1) of
POM, TKN, and TP in discharge (Figure 3-4d, e, and f).
Arrows indicate sampled storm events.


4 00





a 6

Particulate Organic Matter


& 4

Table 3-3. Amounts of total suspended solids, particulate
organic matter, total Kjeldahl nitrogen, and total
phosphorus flushed in excess of base-discharge from
Newnan's Lake during partial drawdown.

kg dry weight mg/m2 lake area
removed by removed by
drawdown drawdown

Total suspended solids 59,247 2,043
Particulate organic matter 46,537 1,605
Total Kjeldahl nitrogen 8,840 305
Total phosphorus 290 10

Computed on a m2-basis, removal during drawdown is

small compared to the likely stores of flocculent sediment

in the lake, with an average estimated thickness of 70 cm

(Skoglund 1990). Two factors may account for the

relatively low removal rates of particulate material from

the basin. First, the lower incidence of storm events

during the period of drawdown may have precluded

considerable resuspension of flocculent bottom material

(Pollman 1983). Wind-induced wave action will stir the

water column and may promote flushing of resuspended matter

(Maceina and Soballe 1990). Second, the low water-level at

the onset of drawdown results in small hydraulic head and

discharge rates. Low discharge rates depress removal of

resuspended material.

Concentrations of POM, TKN, and TP in the lake

discharge are shown in Figure 3-4d, e, and f. The data

points represent the average value of two replicate

samples. "Error bars" averaged 0.40 mg/l for POM (range

0-1.63), 0.15 mg/l for TKN (range 0-0.42), and 0.01 mg/l

for TP (range 0-0.03). The concentrations in the discharge

were higher during the first month of drawdown compared

with the pre-drawdown period (Figure 3-4d, e, and f).

These differences are statistically significant (P<0.05,

P<0.05, P<0.01, respectively) using a pooled analysis of

variance (Byrkit 1975) and assuming no autocorrelation

among data points (Table 3-4). After this period, when

discharge decreases to about twice the assumed base-

discharge (see Figure 3-3) concentrations of POM, TKN, and

TP drop to near predrawdown levels. At these low lake

stages, the outflow from Newnan's Lake may be restricted to

the less-turbid surface layer of water released through the

shallow creek.

Table 3-4. Concentrations of particulate organic matter
(POM in mg/1), total Kjeldahl nitrogen (TKN in mg/1), and
total phosphorus (TP in gg/l) in surface discharge from
Newnan's Lake prior to and during the first month of
drawdown. Number given are the mean, number of
measurements (N), and standard error of the mean (S.E.)

Predrawdown During 1st month
of drawdown

Mean N S.E. Mean N S.E.

POM 9.48 22 0.25 13.61 8 0.47

TKN 1.72 22 0.05 3.02 8 0.03

TP 55.81 22 2.88 83.13 8 11.42

Considerable changes in water quality of the outflow

occurred during the sampled storm events. Three sampled

storm events during the drawdown period produced increases

in the concentration of POM, TKN, and TP (Figure 3-4d, e,

and f). Other peaks in these time patterns may have been

associated with wind events which were not sampled.

Although a more detailed analysis (incorporating daily wind

data and in-lake water quality) would provide better

evidence, it appears that sampled storms were effective in

resuspending surficial flocculent sediment in Newnan's

Lake. This corresponds with findings in other shallow lake

systems (Sheng and Lick 1979; Somly6dy 1982). Mixing of

these deposits in the water column enhances their removal

through flushing. The lack of high-intensity wind events

(e.g. in excess of 30 km/hr) during the study period

contributed to the low particle flushing rates encountered.

Two storms of low intensity, sampled prior to removal of

the spillway, produced small or no increases in particulate

matter and nutrient concentrations of the flow through the

section of the dam with the missing top-flashboard.

Evidently, storms are effective in flushing particulates

and nutrients from the system. The flushing, however, is

typically significant when obstructions such as the

spillway are removed from the lake outlet.

Oxidation of Organic Matter in Exposed Littoral Sediments

A second objective of the short-term drawdown of

Newnan's Lake was to improve littoral zone habitat.

Exposure of littoral zone lake bottom to air may allow

oxidative removal of organic material (Wegener et al. 1974)

and should consolidate flocculent sediments (Fox et al.

1977). This improves the area for future sportfish

spawning and provides firm rooting for aquatic macrophytes

(Holcomb and Wegener 1971), which may contribute to a

higher standing crop of aquatic macro-invertebrates and,

eventually, fish (Wegener and Williams 1974). A vegetated

littoral zone may function as an effective nutrient trap

and reduce nutrient input from runoff into the lake (Mickle

and Wetzel 1978). Numerous authors have suggested that

aquatic macrophytes can inhibit the development of algae

(Canfield et al. 1984; Crisman 1986b). Accumulation of

algae are perceived as a persistent problem in Newnan's


No significant net oxidation of littoral substrate was

noted after 28 days of in situ exposure (Figure 3-5). The

difference in organic matter content between pre- and

postexposure enclosures was statistically not significant

(paired t-test, P=0.05). Field observations showed

production of organic matter in the form of germinating

seeds, roots, and above-ground plant biomass inside the

enclosures. If oxidation of the sediments occurred, it may


I Pre-Exposure 11.05
c^ 3 Post-Exposure
IN Nwunb=Ae of inc rr Organe MY~r (g/m2 .d)
6 5000 0.79

c 2500-

..34 0.95
0.08 0.61

o nPI MM M no I __
West Shore East Shore North Shore
Enclosures Enclosures Enclosures

Figure 3-5. Organic matter content (g/m2) at the onset
("pre") and after 28 days of exposure ("post"), and rate of
increase in organic matter (g m'2d"1) in Newnan's Lake
littoral sediments (field data).

have been masked by this organic matter production. It is

possible that oxidation rates may have been quite low,

particularly if the organic sediments were largely humic

compounds, which are relatively resistant to microbial

degradation (Sederholm et al. 1973).

The laboratory experiments did not provide evidence of

oxidation of these sediments either. Oxidation rates were

extremely low and not significantly different between dry,

moist, and permanently inundated substrate (Table 3-5).

Table 3-5. Mean oxidation rates of incubated littoral
sediments in the laboratory under 3 different treatments.

N Mean oxidation rate S.D. Range
(mg g-ld1)
Inundated 9 0.23 0.05 0.17-0.30
Moist 9 0.24 0.05 0.16-0.35
Dry 7 0.19 0.05 0.13-0.26

Production of organic matter in these laboratory chambers

under the low-light regime was less likely than in the

field enclosures. Hence, these experiments suggest that

oxidation rates were indeed low. These results support

findings by Fox et al. (1977), who noted no significant

decomposition of organic matter in sediment from Lake

Apopka, Florida, using a series of laboratory experiments.

Consolidation of Littoral Sediments

In the laboratory consolidation-experiment, water

content of exposed sediments decreased from 90 to 50

percent (Figure 3-6) and remained moderately compact 149

days after reflooding. Organic matter content remained

constant throughout the duration of the experiment (Figure

3-6). Firmer substrate will reduce rates of resuspension

during periods of high winds.

Summary and Conclusions

The drought of 1989 produced a low lake-stage (41 cm

below average) at the start of the drawdown period and a

low incidence of storm events during the drawdown. This

reduced the discharge rates of particulate organic matter

and nutrients from the lake.

Concentrations of POM, TKN, and TP in the lake

discharge were significantly higher during the first month

of drawdown than during the pre-drawdown period. These

concentrations dropped to near predrawdown levels at lower

lake stages, when the sill depth at the mouth of Prairie

Creek may have restricted the outflow from the lake to the

less-turbid surface layer of water.

Storm events produced increases in concentration of

particulate organic matter and nutrients in the discharge

from the lake during drawdown. Storms sampled prior to

opening of the spillway did not cause such increases.

Storms resuspended fine particulate deposits in this



50 100



Duration of Exposure/Reflooding (Days)

Figure 3-6. Changes in water and organic matter content
(weight/weight ratio) in Newnan's Lake littoral sediments
upon exposure and after reflooding in a laboratory setting.








O-oI-----Organic Mttro--o-----

53SW if a Our a



shallow, exposed lake and promoted flushing of this


The sill at the mouth of Prairie Creek reduces the

likelihood of a gravity drawdown to a stage much lower than

that accomplished (19.70 m MSL). At this stage, the water

level is barely lakeward of the cypress tree fringe and

drying/consolidation of lake bottom is limited to a narrow

littoral zone fringe.

Field and laboratory tests did not show oxidative

removal of organic matter from exposed areas of the lake

bottom. Consolidated sediments remained moderately firm

after reflooding in a laboratory experiment and may,

therefore, provide improved substrate for rooted aquatic



Water-level drawdown is a well established lake

management technique. It has been used to influence the

abundance and composition of aquatic plant communities

(Holcomb and Wegener 1971; Hestand and Carter 1975; Cooke

1980; Tarver 1980), increase fish standing crop (Lantz et

al. 1964; Wegener and Williams 1974), and consolidate

littoral sediment (Holcomb et al. 1975; McKinney and

Coleman 1980). Drawdowns have also produced higher

densities of littoral macroinvertebrates (Wegener et al.

1974), increased nutrient concentrations in the water

column (Serruya and Pollingher 1977), and reduced dissolved

oxygen levels by disrupting the thermal stratification of

the water column (Richardson 1975). The intent of the 1989

short-term drawdown in Newnan's Lake was to improve

littoral habitat for fish growth and recruitment by

allowing oxidation and consolidation of exposed littoral

zone lake-bottom (Chapter 3).

In spite of the frequent use of water-level drawdown,

little is known about its impact on erosion of littoral

sediments. Low water levels during drawdown increase wind-

wave action on littoral substrate. The resulting physical

resuspension of bottom material may then be followed by

transport and gravitational settling of the entrained

particles in deeper areas (Bengtsson et al. 1990). As

such, material can be focused from the littoral zone to

profundal substrate (Davis 1968). This process may be

particularly pronounced in lakes with fine, organic bottom

substrate, i.e. those where water-level drawdown is most

commonly applied.

The effects of enhanced resuspension of bottom

material may include increased turbidity and reduced light

penetration in the water column. It may also exert

considerable oxygen demand (James 1974) and lead to

increased availability of nutrients in the overlying water

for algal utilization (Holdren and Armstrong 1980; Pollman

1983). "Sloughing" of eroded material to deeper parts of

the lake reduces maximum water depth and may lead to the

development of substantial shallow areas. This increases

the potential of the lake ecosystem to support extensive

growth of rooted macrophytes.

The objective of this part of the study was to

quantify the removal of flocculent sediments from the

littoral zone and determine their redistribution to deeper

areas of the lake during a short-term drawdown of Newnan's


Materials and Methods

Sediment Cores

Removal and redistribution of lake bottom material due

to drawdown was investigated using two series of nine

sediment cores. Marker horizons were identified in the

cores either from profiles of bulk density, unsupported

210pb, 137Cs, organic matter and nutrient content or from

direct field evidence of distinct stratigraphy. Gain or

loss of sediment at each site was then quantified by

matching marker horizons between pre- and postdrawdown


Figure 4-1 shows the location of the cores. An

electronic long-range navigation system (LORAN, Si-Tex 797)

was used to ensure agreement in location of pre- and

postdrawdown sampling sites. Additionally, triangular

compass measurements with permanent landmarks were used for

the littoral cores. Specifics of the core locations are

given in Table 4-1. Latitude and longitude records were

relative to the calibration site located near the

southwestern boat ramp (Figure 4-1).

The accuracy of this LORAN to return to a sampling

site is limited to a range of approximately 20 meters.

This inherently results in error when comparing pre- and

postdrawdown sediment stratigraphy. This error is reduced

when in-lake variability between nearby sites, in terms of

water depth and bottom stratigraphy, is low as in Newnan's

Figure 4-1. Map of Newnan's Lake with core locations
indicated (filled circles).

Lake with its rather flat bottom topography (Holly 1976)

and homogeneous deposits of soft sediment (Skoglund 1990).

By matching several marker horizons between cores, rather

than one single horizon, error may be reduced further.

Table 4-1. Newnan's Lake cores: Water depth
and location.

Core Depth (cm) Latitude Longitude

1-11 150 29 39 03 82 13 82
2-12 200 29 39 12 82 12 87
5-15 170 29 38 02 82 12 71
6-16 150 29 38 04 82 13 47
7-17 118 29 37 69 82 14 74
8-18 115 29 36 80 82 14 50
NE 65 29 39 27 82 12 05
SE 62 29 36 68 82 14 43
W 58 29 38 16 82 14 22

Note: Measured with Si-Tex 797 LORAN in degrees, minutes,
and (minutes/100). Calibration marker at southwest
boat ramp: 29 37 07/82 15 23.

Water depth at all stations was estimated by carefully

lowering a Secchi disk until contact between the disk and

soft bottom substrate was noticed.

Cores were collected and preserved as described in

Chapter 2. Visual observations of the cores were made in

the field and again during sectioning in the laboratory to

detect changes in sediment color or texture which could

serve as marker horizons.

Bulk Density and Nutrient Analyses

The cores were sectioned in 1 cm intervals to a depth

of 30 cm, into 2 cm intervals from 30 to 80 cm depth, and

into 4 cm intervals below that. The sections were sealed

in plastic bags and stored in a refrigerator. Bulk

density, organic matter content, total Kjeldahl nitrogen

(TKN), and total phosphorus (TP) were determined as

described in Chapter 2. Measurements were made at

intervals selected to give a representation of the entire

core profile.

Radio-isotope Analyses

Marker horizons from measurement of 210Pb and 137Cs

levels throughout the core profiles may aid in a

determination of sediment redistribution during drawdown.

210Pb profiles have been reliable in documenting sediment

removal due to major storm surges (Robbins et al. 1978).

Pennington (1981) recorded variations in 137Cs profiles in

a shallow lake resulting from episodic sediment

redistribution and deposition. Bengtsson et al. (1990)

successfully used settling sediment traps to investigate

redistribution of fine sediments in Swedish lakes.

Maintaining suspended sediment traps in the water column,

however, was not an option in a public use lake, such as

Newnan's. Use of 210Pb and 137Cs in the determination of

depositional markers has additional significant benefits,

because such profiles permit the calculation of the age of

deposited material in the cores (Eakins and Morrison 1978;

Appleby and Oldfield 1983).

Between 8 and 16 210Pb and 137Cs measurements were made

depending on the length of the core. 210Pb and 137Cs

concentrations were measured by direct y-assay as described

in Chapter 2.

Results and Discussion

Littoral Cores

Two series of three littoral cores were taken;

northeast, southeast and west cores (Figure 4-1).

Predrawdown cores were taken when water depth averaged 62

cm. Following the drawdown and natural rise of the water

level to predrawdown stage the second series of three cores

were taken at the same locations. Water level at these

sites was lowered from 62 to 30 cm for 8 weeks during the

drawdown. Sand underlying flocculent brown substrate

served as a suitable marker horizon.

Sediment depth to sand decreased by an average of 42%

after drawdown (Figure 4-2a), and bulk density of the

remaining substrate increased by an average of 250% at all

three sites (Figure 4-2b). Organic matter, TKN, and TP

(expressed as weight/weight ratios) decreased at two of the

three littoral zone sample sites following drawdown (Figure

4-2c, d, and e). Analogously, the amount of organic matter

overlying the sandy littoral bottom was reduced (Figure 4-

2f). Calculated removal rates are given in Table 4-2.

2 Pre W Post
'a a
S 20



e o
In c

E 20
0 002 n I

S 30

S 10
"or I
E 0
re d

a 1.5
2 1.0
f II

o 0.4 5
So. I
o 0.2
E 1 7 I = E

U 0.



Figure 4-2. Northeast, southeast and west littoral cores,
Newnan's Lake. Predrawdown (open bars) and postdrawdown
(filled bars); a) sediment depth to sand (cm), b) bulk
density (g/cm3), c) organic matter content (g/g), d) total
Kjeldahl nitrogen (mgN/gdr wt), e) total phosphorus
(mgp/gdrywt) and f) organic matter to sand (g/cm2).

Table 4-2. Material removal rates (g m2day-1)
from littoral cores during drawdown, Newnan's

Removal rates NE SE W

Organic matter 3.52 14.71 0.0
TKN 0.14 1.58 0.01
TP 0.01 0.11 -0.02

Removal rates from the west core deviated from the

observed pattern. This may have been due to the location

of this station in a rather quiet littoral cove, less

subject to wind-wave action. Sediment depth to sand (in

cm) decreased at this site following drawdown, but no

removal of organic matter (in mg/cm2) was noted (Figure 4-2f).

These observations may be interpreted in two ways.

First, sediment thickness may have decreased without actual

removal of material due to consolidation of the substrate

at this site. This appears unlikely, since this location

remained inundated by at least 30 cm of water at all times.

Second, erosion of material did in fact occur, resulting in

the observed increased bulk density of the remaining

substrate at this site. However, eroded organic material

may have been replaced by net primary production during the

period of drawdown. Increased organic matter and total

phosphorus content at this site (Figure 4-2c, and e)

support the latter interpretation. Enhanced primary

productivity of benthic algae is likely at low water depth

during the drawdown when more light reaches the littoral

bottom. This is particularly plausible at the protected

west core site, with lower light inhibition by suspended


Removal of littoral sediment likely resulted from

increased shear stress caused by wind-wave action. Such

surface waves generate periodic oscillations in the water

column, which attenuate with water depth (cf. Wetzel 1983)

but may reach the sediment-water interface in shallow

water. When this stress exceeds the bulk shear strength of

the surficial deposits, sediment resuspension occurs (Lick

1982). Waves are largely wind-induced, although wakes from

boat traffic may also be considerable in this public-use

lake. Resuspended material may then be moved from shallow

to deeper parts of the basin by water currents (Davis and

Ford 1982; HAkanson 1982; Bengtsson et al. 1990) and

eventually settle where water depth is sufficient to

eliminate stress from wind-wave action.

The amount of organic matter was not significantly

different between the pre- and postdrawdown cores although

the northeast and southeast cores demonstrated a

considerable reduction (Figure 4-2f). The reduction in

thickness of the sediment layer and the increase in bulk

density following drawdown, observed at all three sites,

were statistically significant (P=0.05, paired-t). The

assumption in the study was that these three sites were

representative of the littoral zone. A different approach

may be to make many, simple to perform measurements of

sediment-thickness-to-sand throughout the littoral zone

pre- and postdrawdown. Such sampling will represent better

the entire littoral zone, but it will not provide data on

removal rates of bulk sediment, organic matter and


Profundal Cores

The locations of the profundal core sites are shown in

Figure 4-1. Material in the cores consisted generally of

homogeneous, black-brown sediment of low bulk density (<100

mg/cm3 up to a depth of 30 cm).

Cores 1 and 11. The matching profiles for cores 1

(predrawdown) and 11 (postdrawdown) for bulk density,

organic matter, unsupported 210Pb, 137Cs, TKN, and TP are

shown in Figure 4-3 (left panel). During sectioning in the

laboratory, the first occurrence of clay was noted at 120

cm (core 1) and 132 cm (core 11) (Table 4-3). This

suggested an addition of 12 cm of material to the profile

at this station during drawdown. Profiles for bulk density

and organic matter content revealed a marker at a depth of

82 cm (core 1) and 94 cm (core 11). This also indicated a

gain of 12 cm during drawdown.

Unsupported 210Pb and 137Cs profiles, however, implied

removal of material postdrawdown. Profiles for TKN and TP

were inconclusive with widely fluctuating nutrient levels

Cores 1(o .........o) and 1(*--*) Cores 2(o -.o o) and 12(*--*)
0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140
- 180 180
I '50 150
120 1 120

2 90 V .... 90

: 40 .% ..- --
3 030
a '----- -- ',----'r --'- '

S 0 1. L 60
- ..*, ,: ....-*. **** -* ^ ./.- .._............

. 1 .

I ... ... .--.'. Y. ..

S3.0 3 ..""-""". 0
20 0

2.0 20

S40 40

" 30 0./ ./
S20 20

10 .. ..... 10
F" F
d ao d own ,s-o-i-d n ) -R i .'.h--t---l--o ...... a d 4,o
4-a0 .e a. 1 ( t ... .

2. 2.0
E. I
"1 0 ....- 1.0
00 0.0
O0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140
Depth in Core (cm) Depth in Core (cm)

Figure 4-3. Depth profiles for Newnan's Lake cores. Left
panel; cores 1 (predrawdown, dashed line) and 11 (post-
drawdown, solid line). Right panel; cores 2 (predrawdown,
dashed line) and 12 (postdrawdown, solid line). Arrows
indicate horizons discussed in the text. A, bulk density in
mg/cm3; B, organic matter content in mg/mg; C, unsupported
Pb-210 in pCi/g; D, Cs-137 in pCi/g; E, total Kjeldahl
nitrogen in mg/gdry wt; and F, total phosphorus in mg/gr 4.

in both cores. Average TKN concentrations in the top 120

cm of cores 1 and 11 were 33 and 27 mggTK/gdry weight

respectively. TP levels averaged 3.4 and 3.0 gTp/gdry weight'

Examination of the six profiles did not demonstrate a

distinct removal or gain of bottom material due to drawdown

for this station.

Table 4-3. Length of retrieved cores and
depth of sand and clay horizons in
Newnan's Lake cores.

Core Core length Depth (cm) to
site (cm) sand(s)/clay(c)

1 129 120c
11 148 132c
2 155
12 140
5 113 85s,113s,95c
15 92 76s,86c
6 135 135c
16 115
7 135 100s
17 148 100s
8 72 38s,64s
18 61 36s

Cores 2 and 12. At station 2, bulk density and

organic matter profiles (Figure 4-3, right panel) revealed

clear markers at 68 cm predrawdown (core 2) and 78 cm

postdrawdown (core 12). Peaks and dips in the unsupported

210Pb and 137Cs profiles displayed a similar gain of

approximately 10 cm of material in the postdrawdown core.

Predrawdown peaks in the 210Pb profile at depths of 10, 18,

and 44 cm occurred in the postdrawdown core at 16, 24, and

54 cm. A similar gain was evident from the 137Cs profiles.

This indicated an addition of approximately 9 cm during

drawdown at this station.

Based on the bulk density of the top 10 cm layer of

the postdrawdown core, this gain was equivalent to 0.18

g/cm2. This translated into an additional 1.1 years of

sediment deposition during the 245-day drawdown period

compared to normal, non-drawdown sedimentation rates

(discussed later). TKN and TP profiles were inconclusive.

TKN levels fluctuated between 25 and 35 mgTN/gdry weight

with a few measurements below 20 mgTN//gdry weight- Averages

for both cores approximated 32 mgTN/gdry weight and

corresponded to levels in cores 1 and 11. Fluctuations in

TP were even more pronounced with an average concentration

between 3.0 and 3.5 mgTp/gdry weight.

Cores 5 and 15. Visual analysis of cores 5

(predrawdown) and 15 (postdrawdown) during sectioning

revealed distinct sand layers at depths of 85 and 113 cm

for core 5 and at 76 cm for core 15. First occurrence of

clay was at 95 cm (core 5) and 86 cm (core 15) (Table 4-3).

Based on these marker horizons, 9 cm of sediments were

removed during drawdown. Bulk density and organic matter

profiles appeared to display a phase difference of a

minimum of 10 cm (Figure 4-4, left panel) also showing

Cores 5(o 0.. o ) ond 15(*--*)
0 20 40 60 80 100

Cores 6(0 o) and 16(*-*) 73
20 40 60 80 100 120

0 20 40 60 80 100 0 20 40 60 80
Depth in Core (cm) Depth in Core (cm)

100 120

Figure 4-4. Depth profiles for Newnan's Lake cores. Left
panel; cores 5 (pre-drawdown, dashed line) and 15 (post-
drawdown, solid line). Right panel; cores 6 (pre-drawdown,
dashed line) and 16 (post-drawdown, solid line). Arrows
indicate horizons discussed in the text. A, bulk density in
mg/cm3; B, organic matter content in mg/mg; C, unsupported
Pb-210 in pCi/g; D, Cs-137 in pCi/g; E, total Kjeldahl
nitrogen in mg/gdry wt; and F, total phosphorus in mg/gdry wt.

material removal during drawdown. Unsupported 210Pb and
137Cs profiles indicated a distinct removal of

approximately 8 cm during drawdown. In addition, TKN and

TP profiles suggested material removal postdrawdown. This

implied that the drawdown removed a minimum of 8 cm, or

0.26 g/m2, from the sediment profile. This equated to

approximately 4.7 years of sediment removal at this site

during drawdown (discussed later).

The profiles also demonstrated the different character

of the substrate at station 5-15 compared with other

sampling sites. Bulk densities were significantly higher

throughout the profiles, while organic matter and nutrient

concentrations were depressed compared with deposits at

similar depths in other locations. Field observations

concurred with these data in that substrate appeared

firmer, with a relatively high sand content.

Cores 6 and 16. Sectioning of the cores in the

laboratory revealed a clay horizon in core 6 (predrawdown)

at 135 cm (Table 4-3), well below the bottom section of

core 16 (postdrawdown). Bulk density and organic matter

profiles (Figure 4-4, right panel) showed a rather

homogeneous top 100 cm for both cores containing flocculent

material (bulk density<100 mg/cm3) and an average organic

matter content of 60% (weight/weight ratio). Firmer

substrate started at 105 cm depth in core 6 and at

approximately 110 cm in core 16. A horizon of inorganic

material was encountered first at 105 cm depth in core 6.

This layer may have occurred immediately below the last

sampled segment in core 16.

Isotope profiles for both cores also displayed a shift

in phase of several centimeters, while nutrient profiles

appeared inconclusive. Average TKN and TP levels coincided

with cores 1-11 and 2-12 (30 mgTK/gdry weight and 3.0-3.5

mgTp/gdry weight). In summary, these profiles provided weak
support for the conclusion that 5 cm, or 0.05 g/cm2, of

material was deposited at this site during drawdown. This

equated to approximately 0.04 years of sedimentation in

excess of "normal" sedimentation occurring during the

period of drawdown (discussed later).

Cores 7 and 17. Core sectioning in the laboratory

demonstrated 1 m of black, soft, organic muck overlying a

distinct firm, sandy layer in both cores 7 (predrawdown)

and 17 (postdrawdown) (Table 4-3). Water and organic

matter content increased again below this layer. This was

reflected in the bulk density and organic matter profiles

at this core site (Figure 4-5, left panel).

The additional similarity of unsupported 210Pb and

nutrient profiles for pre- and postdrawdown cores suggested

no significant redistribution of material at this site.

The absence of a 137Cs peak below 40 cm in the postdrawdown

core is unexplained, but has been verified by repeated

analysis of the sample material. Average TKN and TP

Cores 7(o .......o) and 17(*--*)
0 20 40 60 80 100 120

I. 900

o so-.
S-i-)-- -^----

6. 4,0


6.0 D

.U ~

9 3.0



Depth In Core (cm)

Cores 8(o ... o) and 18(0-*)
0 10 20 30 40 50 60

10 20 30 40 50 60
Depth in Core (cm)

Figure 4-5. Depth profiles for Newnan's Lake cores. Left
panel; cores 7 (pre-drawdown, dashed line) and 17 (post-
drawdown, solid line). Right panel; cores 8 (pre-drawdown,
dashed line) and 18 (post-drawdown, solid line). Arrows
indicate horizons discussed in the text. A, bulk density in
mg/cm3; B, organic matter content in mg/mg; C, unsupported
Pb-210 in pCi/g; D, Cs-137 in pCi/g; E, total Kjeldahl
nitrogen in mg/gdry wt; and F, total phosphorus in mg/gdry wt.

concentrations in the top 100 cm corresponded to levels

found in other cores (except site 5-15).

Cores 8 and 18. Minor sand was encountered at 38 cm

in core 8 (predrawdown), with a clear, pronounced sand

layer at 64 cm. Core 18 (postdrawdown) showed clear sand

at 36 cm (Table 4-3). Both the bulk density and organic

matter profiles displayed these markers (Figure 4-5, right

panel) with bulk density values in excess of 1200 mg/cm3

and organic matter content less than 5%. Likewise, TKN and

TP concentrations approached zero at 65 (core 8) and 36 cm

(core 18). Evidence from these four profiles suggested

that approximately 28 cm of material eroded from the

profile at this station during drawdown.

The absence of the 210Pb peak in the top 10 cm of the

postdrawdown core also suggested such removal, but to a

much lesser extent. Analogously, elimination of the 137Cs

peak in the top 20 cm of deposits in core 18 and the onset

of detectable levels of this radionuclide about 10 cm

deeper in the predrawdown profile implied removal of

material during drawdown. Such massive removal rates,

however, seemed implausible in light of the small flushing

rates of particulate matter through the nearby outflow

recorded during drawdown (Gottgens and Crisman 1991). It

is more likely that either the exact location of the pre-

and postdrawdown cores did not match or that a significant

disturbance of the sediment profile occurred at this

station during drawdown. Hence, no inference on gain or

removal of bottom material at this site was made.

All stations

Rates of removal or gain of material for the profundal

sampling stations are summarized in Table 4-4. The 245-day

time period covered started immediately

Table 4-4. Profundal cores: Gain/removal (-) rates of bulk
sediment, organic matter, TKN, and TP (g m-2day-1).

Gain/removal(-) Core station
rate 1-11 2-12 5-15 6-16 7-17 8-18

Bulk sediment 7.2 -10.6 2.2 0.0 *
Organic matter 4.3 -5.9 1.5 0.0 *
TKN 0.2 -0.3 0.1 0.0 *
TP 0.02 -0.03 0.01 0.0 *

*) No clear record or conflicting evidence.

prior to removal of the spillway and lasted until the

natural return of the water level to predrawdown levels

following spillway re-installation.

Consequently, out of six profundal cores analyzed for

this study, two showed evidence of added sediment

deposition during drawdown, one showed no gain or loss, one

showed removal, and two were inconclusive. No conclusion,

therefore, can be reached by this study's data regarding

flocculent sediment accumulation in the profundal zone of

Newnan's Lake during the drawdown.

While sediment removal from the littoral zone was

indicated by the results of this study, a quantitative

record of this transfer to profundal sites was difficult to

obtain. This is particularly true when profundal substrate

consisted of a thick pack of near homogeneous material

without clear marker horizons. Furthermore, the small drop

in lake level during this drawdown did not create large

areas of erosion in the littoral zone. This reduced the

magnitude of potential redistribution of sediments in the

lake which reduced the signal in profundal core profiles.

Using the 210Pb profiles, sedimentation rates were

computed for each profundal station. Calculations followed

the constant rate of supply model (Goldberg 1963; Appleby

and Oldfield 1983). These accumulation rates were then

compared to the gain or loss of material at those sites due

to drawdown. As such, this gain or loss was equated to a

time period of "normal" sedimentation. For instance, the

gain of material due to drawdown at station 2-12 was

equivalent to 1.1 years of sedimentation (Table 4-5).

Uncertainty analysis

Because inferences were made from 21oPb profiles in

different aspects of this work, an assessment of the level

of confidence in these profiles is appropriate. This was

particularly pertinent since profiles were established in

soft lake sediment with the potential of disturbance in the

chronology of deposits. Three independent observations

aided in such an assessment.

Table 4-5. Newnan's Lake profundal cores: Comparisons of
recent (5 yrs. B.P.) dry-sediment accumulation rates with
gain/loss (-) of material during drawdown, and with
calculated dry-sediment accumulation rates.

Core Recent Gain/loss Gain/loss Calc.recent Cumulative
site sed.rt. during during sd.rt. residual
drawdown drawdown -2ys. 0Pb
(g cm yr"1) (g/co) (yrs) )g CM T ) (pCi/l )

1-11 0.06 0.05 30.2
2-12 0.10 0.18 1.10 0.10 26.4
5-15 0.07 -0.26 -4.70 0.07 17.4
6-16 0.08 0.05 0.04 0.07 28.0
7-17 0.06 0.00 -0.70 0.05 27.0
8-18 0.07 0.07 18.9

*) no clear record or conflicting evidence.
1) sedimentation normally occurring during the 245 days between pre-
and post-measurements is subtracted.
2) Using a model developed by Binford a Brenner (1986)A Measured
average cumulative residual unsupported 'Pb-24.65 pCi/cm'; Flux for
10Pb-fallout-0.77 pCi cmyr'.

First, recent 210Pb based deposition rates were not

statistically different (P=0.05; two-tailed correlated

test) from calculated values using an earlier,

independently developed model (Binford and Brenner 1986)

(Table 4-5). In this model fallout 210Pb is used as a

dilution tracer to compute net accumulation rates of any

material in surface mud according to:

r= F210Pb x A-1


where F210Pb equals the flux of fallout 210pb (pCi cm"2y-I)

and A is the activity of 210pb in the sediment sample

(pCi/gdry weight)* Furthermore, the different cores from
Newnan's Lake had comparable 210Pb residuals (i.e. total

residual unsupported 210Pb contents) despite differences in

accumulation rates (Table 4-5). The 210Pb residuals of the

cores reflected the 210Pb fallout from the atmosphere.

Since this fallout lies in the range 0.5-0.9 pCi cm-2yr-1

(Nozaki et al. 1978), depending on locality, the 210Pb

residuals should lie in the range 16-30 pCi/cm2 (the 210Pb

radioactive decay constant=0.03114 yr-1). This

corresponded to the measurements in Newnan's Lake cores

(Table 4-5).

Second, recent 210Pb based dry-sedimentation rates

correlated well with water depth (R2=0.81; N=5), when

station 5-15 is excluded. The proximity of this site to a

fish attractor (e.g. submersed brush attached to an

anchored buoy) and, hence, higher boat traffic and boat

wake may have disturbed the sediments and depressed

sedimentation rates. Consequently, direct comparisons

between this sampling site and others in the lake may be

misleading. The close relationship between water depth and

material accumulation rates is well-established (Evans and

Rigler 1980). Including the 5-15 site, 57% of the

variability in dry-sedimentation rates was explained by

depth of the water column. While this relationship does

not necessarily underwrite the accuracy of recent 210Pb

levels, it does demonstrate that these levels correlate

well with each other (i.e. their precision is supported).


Finally, counting statistics and error prediction were

applied to the recorded unsupported 210Pb data to compute

statistical precision. This error analysis only addressed

internal uncertainty associated with the accuracy of y-ray

detection. It did not consider uncertainty controlled by

external factors such as error associated with the (sub)

sampling design, smearing of the core (Chant and Cornett

1991), bulk density determinations, and post depositional

mobility of constituents in the core (Anderson et al.

1987). "Error bars" associated with the experimental data

are illustrated for cores 1 and 2 in Figure 4-6. Other

cores showed a similar magnitude of error. The length of

the error bar equals one standard deviation (a) on either

side of the point (i.e. 68.3% confidence limits), which is

standard practice in expressing uncertainty in nuclear

measurements (Wang et al. 1975). Because the recorded

counts in nuclear counting experiments follow a Poisson

distribution (Knoll 1979), the predicted standard deviation

is the square root of the mean number of counts. To arrive

at this mean (counts/hr), samples were counted from 14 to

45 hours depending on sample weight. Since

Net counts = Total counts Background counts (4-2)

uncertainty in the net counts is propagated according to


0 20 40 60 80 100 120

0 0

(1I 1
5____ ?- -^
0 I\ T N

0 0
-10 -1 A

I/ .

5 r

0T O 0s--.sT T
0 00 --------

M i I/ 0

0 20 40 60 80 100 120
Depth in Core (cm)

Figure 4-6. Unsupported Pb-210 profiles for core 1 (top) and
core 2 (bottom). The length of the error bar equals la on-
either side of the data point.

an = (at2 + o 2) (4-3)

where an, at, and ab are, respectively, the standard

deviations of the net count, total count, and background


The effect of the computed internal uncertainties on

the interpretation of the profile in terms of marker

horizons or sediment accumulation rates was small. For

instance, recent sediment accumulation rates for core 1

varied between 0.049 and 0.055 g cm-2yr-1 when mean

unsupported 210Pb concentrations (5 years B.P.) plus and

minus 1 a, respectively, were substituted into equation (4-

1). Analogously, these rates for core 2 ranged from 0.085-

0.105 g cm-2yr-1. Error analysis, such as Monte Carlo

simulation, may be used to estimate the effect of internal

uncertainty on the assigned dates for an entire core

profile. Binford (1990) applied this technique to 12 cores

from north Florida lakes and found 95% confidence intervals

ranging from about 1-2 years at sediments 10 years of age,

10-20 at 100 years, and 8-90 at 150 year old deposits. No

estimates of the effect of external uncertainty on 210Pb

derived sediment chronology have been documented.

Summary and Conclusions

An average of 6.08 g m-2day-1 of organic matter (0.58

gTKN m 2day-1 and 0.03 gTp m-2day-1) eroded from the littoral

zone by lowering the water depth from 62 to 30 cm for 8

weeks. Littoral sediments with low bulk density eroded

fastest, and bulk density of remaining substrate increased

by an average of 250%. In the absence of direct field

evidence of distinct stratigraphy, transfer of this

material to deeper areas of the lake was measured using

sedimentary cores with marker horizons. The latter were

derived from measurements of bulk density, organic matter,

radionuclides, and nutrients. The quantitative record of

this transfer, however, was unclear. The small drop in

lake level during drawdown may have contributed to this.

Since the drawdown did not create large areas of erosion in

the littoral zone, the magnitude of potential

redistribution of sediments in the lake and, thereby, the

signal in profundal cores, is reduced.

Two options may now be identified for further work.

First, measurements limited to the littoral zone can be

followed by calculations of deposition rates in the

profundal if the extent of the zones of erosion, transport,

and accumulation of sediment in the lake are known.

Second, use of settling sediment traps to collect

resuspended and transported material may give additional

insight. In Newnan's Lake, however, installation of these

traps on the profundal bottom is difficult because of the

soft nature of the substrate. Suspension of the traps in

the water column may jeopardize public use of the lake.


Instead, this technique may be used in other systems.

Since resuspension and erosion of fine sediments may

influence the ecology of a lake through habitat alteration,

release of nutrients, high turbidity and enhanced sediment

oxygen demand, erosion processes must be considered when

drawdowns are attempted.


Analysis of field data demonstrated that the spillway

caused accelerated accumulation of flocculent and nutrient-

rich sediment in Newnan's Lake (Chapter 2). Plausible

consequences of such increased material transfer between

water and sediment for shallow, productive lake ecosystems

were discussed in the introduction (Chapter 1). They

included: (1) increased resuspension of flocculent

sediments by wind-wave action leading to high turbidity and

diminished light penetration in the water column, (2)

enhanced biological and chemical oxygen demand from organic

and inorganic suspended matter (Hargrave 1969; James 1974),

and (3) expansion of flocculent substrate into the littoral


Each of these consequences, in turn, has implications

for lake management criteria such as water clarity, level

of primary production, extent of the littoral zone, lake

access, and standing stock of sportfish. When acting

simultaneously, however, their net effect on these lake

management criteria is difficult to determine. A few

examples may illustrate this. Increased suspended matter

in the water column, for instance, decreases light

penetration and habitat for submerged vegetation, but

accelerated deposition of sediments and the resulting

increased rate of filling of the lake may actually lead to

a larger area that can be invaded by bottom vegetation.

Increased suspended matter may also reduce water

clarity and decrease the aesthetic value of the lake. This

opposes management objectives designed to secure the lake

for water recreation. Enhanced primary production may

follow when higher rates of wind-wave induced resuspension

of bottom material increase internal loading of nutrients

in the lake through entrainment of nutrient-rich

interstitial (pore) water from the surface sediments (Lam

and Jaquet 1976) or desorption from suspended particles

(Pollman 1983). Oxygen demand from increased suspended

solids lowers dissolved oxygen level in the water column,

which may promote nutrient release from the sediments

(Mortimer 1971; Theis and McCabe 1978).

Low concentrations of dissolved oxygen may stress fish

and other heterotrophs and selectively favor species with

physiological and behavioral mechanisms that aid them in

surviving under such conditions. The Florida gar

(McCormack 1967), bowfin (Johansen 1970), and bullheads

(Loftus and Kushlan 1987) are able to utilize atmospheric

oxygen, while small fishes with upturned mouths can extract

oxygen from the well-oxygenated surface film (Lewis 1970).

Centrarchids, a significant component of the standing stock

of sportfish in Florida lakes, are usually the first fishes

to die under low-oxygen stress (Kushlan 1974). This would

conflict with lake management strategies aimed at promoting

sportfish. Expansion of flocculent sediments into the

littoral zone may also reduce nesting habitat for

centrarchids (Bruno 1984), which rely on firm substrate for

deposition of eggs (Eddy and Underhill 1978).

Changes in components of the Newnan's Lake ecosystem

have been documented, such as the consistent decline in the

number of harvestable sportfish per hectare from 1850 to

227 between 1982 and 1988 (Florida Game and Fresh Water

Fish Commission 1982-1989). However, no work has been done

to analyze the response of the lake to simultaneously

acting factors and assumptions such as those described

above. In other systems, dynamic simulation models have

been used successfully for such analysis by synthesizing

the best current understanding of succession in a prairie

ecosystem (Gutierrez and Fey 1975), eutrophication in a

lake (Anderson 1973), and predator-prey dynamics in a

natural two-species system (Montague et al. 1982). The

computer model keeps track of all interrelationships in the

system once they have been identified. It then builds on

the reliable part of our understanding of the system, while

compensating for the "unreliable" part by requiring an

explicit statement of assumptions and showing the

consequences of these in a model analysis. Hence, the

first objective of this model is to integrate existing

information concerning the Newnan's Lake ecosystem and to

program this information on the computer in the form of a

testable hypothesis.

A second purpose of the model is to identify knowledge

gaps that are most critical in improving the understanding

of the behavior of this ecosystem. This is accomplished by

systematically changing the value of the parameters used to

develop the model and determining the effect of this

systematic change on model output. Those parameters

generating considerable changes in one or more model output

characteristics require further field and/or laboratory

study to increase confidence in the model. As such, the

model helps to guide limnological research, while, in turn,

research data can be used to further develop the model.

Finally, the model aims to evaluate the consequences

of alternative management actions in Newnan's Lake. Models

of different hypotheses of the causes of past lake

responses (such as an increase in the rate of sediment

deposition) to perturbation (such as water-level

stabilization) may be used to make inferences about changes

to be expected following future lake management actions.

Different hypotheses by which the dam may influence

sediment and detritus accumulation are illustrated in

Figure 5-1. They include spillway-induced changes in lake

tor V M
I3 4C) r
4 N > mC
*4 -H 0 *-4 (0
S(a 4 u a)
41 4Q) k 4V
A to ) 0 -*
V 4- > 4 W
0 4) k 1 0
r V to a04

4-) kaV(
4- a ) r .P 0)

*aC ) 4 C-

M >i >0 *4
S- *4 4 0
a)H 4 4-i
S'to a) a)4
a) 1 yo) 0 0>i
r-4 (C a)
0 0 tc 4)
44U ) 0 0) >
:e *4 m #
C C .3 *

1-r4 4 ( ) a)
*0 0 m IMP
(0 r- C 0) 4 '4 9

0O) 0- 0
43 UN 0(D 4 a)
4 .J 0 1go

0 4- ) k 0
43 4 W 4P
(0 >i C
C 4x 4 *r
a ) 0 ) 4-4 o
a) *Z ) CO 0 )
a) -4 k to< 0
MQ 0 U r-

0 a) a) i-4
04iU a) )

S4 4) '0 4- V

to 9 0 4) 0
0 ) r4 4 O--I
S3 -H 4 4 4) 4-)
k4 0 V4 0 0 0
o O 4 ) $4 54
-* 4 U ) C 43 -
1r. (0 4 04- a

i+ i ai

*i '