Internal loading in shallow lakes


Material Information

Internal loading in shallow lakes
Shallow lakes
Physical Description:
vii, 191 leaves : ill. ; 28 cm.
Pollman, Curtis Devin, 1951-
Publication Date:


Subjects / Keywords:
Lakes   ( lcsh )
Limnology   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1983.
Includes bibliographical references (leaves 178-190).
Statement of Responsibility:
by Curtis Devin Pollman.
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 - 000365770
notis - ACA4587
oclc - 09882701
System ID:

Full Text








Support from numerous sources assisted me during the course of my

studies. Charles Hendry, Tom Belanger, Carl Miles, and Charles Fellows

all sacrified time and effort in accommodating me during the field

portion of my studies. Further appreciation is extended to Charles

Hendry for introducing me to some of the subtle nuances of the

analytical equipment critical to the success of this research.

I am deeply grateful for the support of my chairman, Dr. P.L.

Brezonik whose continued faith, gentle proddings, and guidance helped

provide the inspiration to complete my research. I also wish to

acknowledge the other members of my graduate committee, Dr. Wayne

C. Huber, Dr. Donald Graetz, Dr. Edward S. Deevey, and Dr. Thomas

L. Crisman, whose critical review and comments helped produce the final

manuscript. Dr. A.J. Mehta and Dr. Gregory M. Powell both provided

invaluable help in approaching the development of a sediment

resuspension model.

Eileen Town, who typed the initial and final manuscripts, deserves

special recognition for her professional competence and perserverance.

Appreciation is also extended to Carla S. Jones and Marlene A. Hobel for

putting up with the neurotic antics of an expectant author and providing

the finishing touches to the manuscript.

I particularly wish to thank my parents, Mr. and Mrs. Ralph

C. Pollman, and my grandmother, Mrs. Harold G. Lengerich, for

spiritually and financially sustaining me during the ebb moments of my

graduate career. My wife Kathleen sacrificed much for the sake of my

work and is especially deserving of recognition. Without her support

and encouragement, this work doubtlessly would have remained

uncompleted. Finally, I wish to dedicate this work to the memory of my

twin brother, Chris, whose recent departure from life instilled in the

author a sense of necessity. His spiritual guidance was prodigious and

is profoundly missed.







Forms of Sedimentary Phosphorus. . 2
Mechanisms of Nutrient Release . 6

Bioturbation . .. 6
Gas Ebullition . . 8
Sediment Resuspension. . ... 10
Molecular Diffusion. . ... 13

Summary. . . .. 15


. 17

Physical Description and Limnology
Physical Description and Limnology
Materials and Methods. .

of Lake
of Lake
. .

Apopka. .

Physical and Chemical Characterization .
Sorption Experiments . .


Physical Characteristics . .
Chemical Characteristics . .


Turbulence Study . .

Fresh Sediment . .
Partially Dessicated Sediment. .

In Situ Studies. . .

. 57

. 71



Development of a Langmuir Sorption Model ...... .97
Application of Langmuir Model to Lake Apopka and
Lake Okeechobee Sediments. . ... 108

Equilibrium Phosphorus Concentration (EPC) ... .118
Desorption . . ... .. .123
Effect of pH on P Sorption by Lake Okeechobee
and Lake Apopka Sediments. . ... 127

Development of a Sediment Resuspension Phosphorus
Release Model for Lake Okeechobee and Lake Apopka. 131

Nutrient Release Submodel. . .
Sediment Resuspension Model. .
Calculation of Wind-Induced Stress at the
Sediment-Water Interface . .
Determination of Critical Shear Stress .
Calculation of Sediment Resuspension Rates .

Sediment Dispersion Model. .
Application of the Integrated Model to
and Lake Okeechobee. .

. 131
. 133

Lake Apopka

. 134
. 137
. 142
. 144

. 150


Discussion of Internal Loading .
Summary . .



. . 178

. 191

. 161
. 174

1 I I

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



Curtis Devin Pollman

April 1983

Chairman: Patrick L. Brezonik

Major Department: Environmental Engineering Sciences

Internal loading of phosphorus may be significant in shallow lakes

where physical and biological processes can accelerate rates of solute

exchange across the sediment-water interface. Sediments were collected

from Lake Apopka and Lake Okeechobee located in central and south

Florida, respectively, and characterized with respect to physical and

chemical attributes of the sediment that control the dynamics of

phosphorus exchange between sediments and overlying water. A series of

experiments were conducted on bulk sediment to determine rates of

phosphorus exchange across the sediment-water interface as a function

of turbulence and physical state of the sediment.

The shallow nature and broad fetch characteristic of Lake Apopka

(mean depth 1.7 m) and Lake Okeechobee (mean depth 2.8 m) suggest that

sediment resuspension and the concomitant release of sorbed phosphorus

may be important in the cycling of phosphorus. As a result, a model

was developed incorporating physical, chemical, and hydrodynamic

components to predict internal loading of phosphorus in shallow lakes.

Phosphorus releases are predicted on a single event basis in response

to wind-driven turbulent mixing in the water column. The model

synthesizes a wave-hindcasting submodel to determine sediment resus-

pension rates and a phosphorus release submodel derived from Langmuir

sorption theory. Langmuir constants used as model inputs were derived

from adsorption-desorption isotherms conducted at controlled pH.

Under ambient conditions in the water column, both Lake Apopka and

Lake Okeechobee sediments are predicted to release phosphorus upon

resuspension. Results indicate that a moderate wind event (approxi-

mately 8.9 m/s) can more than double mid-lake concentrations of ortho-

phosphorus in Lake Apopka from 20 to 61 ug P/L at pH 8.3. Phosphorus

release from Lake Okeechobee sediments is substantially lower

(approximately 3 to 8 ug P/L). Release is primarily from desorptive

processes; however, in Lake Apopka entrained pore fluid contributes

approximately 30 percent of the net release. The cumulative effect of

cyclonic and convective disturbances during the course of a year

results in internal loading in Lake Apopka easily exceeding external

loading rates. Sediment resuspension is less significant in Lake

Okeechobee, and the nutrient regime appears to be controlled primarily

by external loading.



Within the past two decades, the widespread recognition of lake

degradation as an issue of major environmental concern has spawned

accelerated efforts to better protect and manage lake resources (Born

1979; U.S. EPA 1979). Phosphorus has been implicated as the most

important nutrient limiting phytoplankton growth and production in many

freshwater lakes (Vollenweider 1968; Vallentyne 1970; Likens 1972;

Schindler 1977). Consequently, many lake restoration schemes have

focused on curbing external or cultural input of phosphorus to

lacustrine systems.

In the decade following the introduction of Vollenweider's (1968)

now classic paper, great emphasis was placed by limnologists on modeling

one factor that influences productivity--external nutrient loading. In

spite of the emphasis given this narrow focal point, interest also has

been sustained in observing the contribution of phosphorus from lake

sediments to the overlying water--"internal loading" (cf. Golterman

1977). Much of this work was inspired by early studies such as those

conducted by Mortimer (1941, 1942) on Esthwaite Water in the English

Lake District. More recent work by Fee (1979) has rekindled speculation

and debate on the importance of lake morphometry to lake trophic state

and, in particular, the area of sediment in direct contact with the

trophogenic zone.

Lacustrine sediments may consist of up to 0.75 percent (by weight)

phosphorus (Jones and Bowser 1978). This nutrient pool represents a

potential source of supply to the overlying water column approximating

8 g P/m2 for each centimeter of the sediment column actively engaged

in exchange across the mud-water interface. In view of the work of

Vollenweider (1968, 1975, 1976) and others in defining critical areal

nutrient loading rates, it becomes quite evident that the availability

of sedimentary phosphorus for exchange and the factors that mediate the

transfer of the available pool, both abiotic and biotic, may have a

profound influence on the nutrient regime of a lake. The following

section is thus devoted to a discussion of the chemical nature of

sedimentary phosphorus and its availability for exchange and biological

uptake; factors that control the rate of exchange across the sediment-

water interface are discussed in subsequent sections.

Forms of Sedimentary Phosphorus

A variety of analytical techniques has been used to fractionate

sedimentary P with the ultimate objective of correlating a defined

fraction with biological response (i.e., nutrient uptake). Early

fractionation schemes borrowed heavily from the pioneering work in soils

by Chang anc Jackson (1957). Perhaps the most definitive work on

lacustrine sediments to supersede the efforts of Chang and Jackson

(1957) in soils was that of Williams et al. (1971). Williams et al.

(1971) operationally defined four distinct species of sedimentary P:

(1) inorganic P present as orthophosphate ion sorbed on surfaces of P

retaining minerals (nonoccluded P); (2) inorganic P present as a


coprecipitate or minor component of an amorphous phase (occluded P);

(3) orthophosphate P present in discrete phosphatic minerals such as

apatite Cal0(PO4)6X2, where X = OH-, F-, or 1/2 C03-2; and (4) P present

as an organic ester or directly bonded to carbon atoms (organic P). The

chemical nature of sedimentary P has been the subject of rather

intensive reviews by both Jones and Bowser (1978) and Armstrong (1979).

In more recent work by Williams et al. (1976), the categorization

of sedimentary P has been reduced to three major types: (1) organic

phosphorus, which encompasses all P associated with organic material via

C-O-P or C-P bonds; (2) apatite P comprised of orthophosphorus present

within the crystalline lattice of apatite grains; and (3) nonapatite

inorganic phosphorus (NAI-P) which consists of the remaining ortho-

phosphorus fraction including that dissolved in the interstitial

solution. This latter category constitutes the most labile fraction and

is generally associated with exchangeability. In the development of

these categories, Williams et al. (1976) assumed that nonorthophosphatic

forms of inorganic phosphorus, e.g., polyphosphate ions (including

pyrophosphate) are present in sediments only in negligible quantities.

Isotopic exchange studies conducted by Li et al. (1972) using P-32

as a tracer provide direct evidence that a substantial fraction of the

total inorganic pool in sediments is exchangeable. Exchangeable P

comprised 19 to 43 percent of the total inorganic content of sediments

derived from four Wisconsin lakes. Total exchangeable P was found to be

relatively insensitive to changes in redox status; this was interpreted

by Li et al. (1972) as consistent with the retention of inorganic P in

sediments by an iron-rich gel complex (Williams et al. 1971;

Shukla et al. 1971). Li et al. (1972) hypothesize that although the

equilibrium position between sorbed and free inorganic P for the iron

gel complex is altered by a shift in redox potential, the total

exchangeable pool is not appreciably increased. Conversely, if

inorganic P existed primarily as a discrete crystalline iron phosphate

such as strengite, total exchangeable P should increase significantly

upon dissolution of the solid phase during anaerobiosis.

The high rate of exchangeability observed by Li et al. (1972)

provides further qualitative evidence of the retention of inorganic P as

a nonoccluded form through sorption to an amorphous iron hydrous oxide

gel. Incorporation of P into a crystalline lattice is anticipated to

yield a reduction of exchangeability primarily because of a reduction in

surface area. Additional experiments conducted by Li et al. (1972) with

sediments equilibrated with solutions spiked with orthophosphorus

demonstrated equivalent degrees of exchange for sorbed and native

inorganic P. The authors suggested that native inorganic P is retained

in sediments by the same mechanism as sorbed P.

Alternative fractionation schemes to define the labile fraction of

sedimentary P available for algal uptake have involved chelating agents,

anion and cation exchange resins (Wildung and Schmidt 1973; Huettl

et al. 1979), and desorption potential (Schaffner and Oglesby 1978)

Golterman (1973) used 0.01 N NTA (nitrilotriacetic acid) to extract P

associated with calcium (Ca) and iron (Fe) constituents in sediments.

Golterman (1977) found essentially no difference between NTA extracted

phosphorus and calculated P uptake by the alga Scenedesmus utilizing the


same sediment as its solitary source of P. Grobler and Davies (1979)

found NTA-extractable P to generally be a more applicable measure of

algal available P than total inorganic P. Although both fractions were

correlated with algal uptake, the most consistent relationship was

observed between P availability and NTA extractable P. A non-linear

relationship between algal available P and sediment inorganic P was

demonstrated by Grobler and Davies (1978); the fraction of inorganic P

comprised by algal available P varied with sediment type and was quite

variable, ranging from 4 to 98 percent.

As a means for differentiating between Fe-bound and Ca-bound P,

Golterman (1977) proposed an initial extraction with 0.01 M Ca-NTA.

According to Golterman, complexation of the added Ca by NTA prevents the

solubilization of sedimentary calcium and extraction of Ca-P; conse-

quently only iron-bound P is extracted. Subsequent work by Williams

et al. (1980), however, does not support Golterman's contention that

Ca-NTA selectively extracts Fe- and aluminum (Al)-associated P, leaving

Ca-P intact. Extraction with a neutral, noncomplexing agent (1 M NaCI)

yielded extraction efficiencies comparable with Ca-NTA. Extraction with

Ca-NTA rarely exceeded 10 percent of the NAI-P fraction. Uptake of

sedimentary P by S. quadricauda was closely correlated with the NAI-P

fraction by Williams et al. (1980) and constituted approximately

75 percent of NAI-P; apatite-P was virtually unavailable for algal

growth. These results qualitatively confirm the earlier observations of

Grobler and Davies (1978), who noted that algal available P generally

exceeded NTA-extractable P by a factor of approximately 5 to 6.

Mechanisms of Nutrient Release

As detritus accumulates in surficial deposits, metabolism of

accreting organic material results in the buildup of interstitial

phosphorus concentration to levels that exceed those in the overlying

water by as much as several orders of magnitude.. Release to the

overlying water is accomplished by burrowing and irrigation activities

of benthic organism (i.e., bioturbation), by gas ebullition from

anaerobic decomposition processes, by scouring of the interface by

wind-induced waves, and simply by molecular diffusion across the

concentration gradient that usually exists between sediment pore water

and the overlying water.


Based on analysis of Pb-210 and Cs-137 profiles in recent sediments

of the Great Lakes, Robbins and Edgington (1975) suggested that post-

depositional movement of material occurred within the surficial 10 cm of

sediment. Robbins and Edgington (1975) postulated that the observed

redistribution of sediment particles was the result of either physical

mixing or bioturbation. Activities of the benthic community, such as

burrowing, irrigation, and feeding, result in fluid and particle

transport near (and across) the sediment-water interface. Bioturbation

increases the influx of water into sediments; this is balanced by an

equal volume of interstitial solution flushed from sediments (Petr

1977). Hargrave and Connolly (1978) observed a great deal of spatial

variability in nutrient and gas fluxes across the sediment-water inter-

face of a subtidal flat and attributed the variability to heterogeneity

in the distribution of benthic macrofauna. Berger and Heath (1968)

suggested that homogenization of the upper sediments is complete when

sediment displacement by benthic infauna is comparable to or exceeds the

sediment accumulation rate. As an extension of this assumption,

Krezoski et al. (1978) used average defecation rates for Tubifex tubifex

to estimate sediment displacement, and they calculated that oligochaete

densities in profundal sediments of Lake Huron were sufficient to

displace the upper 3 to 6 cm. Neame (1977) observed a total phosphorus

flux of approximately 650 ug/m2-d across the surficial 3 cm of

sediment in Castle Lake, California, and attributed this flux to

chironomid activity in the oxidized zone of the sediments. Under anoxic

conditions, the phosphorus flux across the interface decreased by two

orders of magnitude.

Aller (1978) reported that benthic infauna activity affects the

dynamics of the sediment-water exchange process by altering the geometry

and kinetics of molecular diffusion in sediments. In the absence of

benthic activity, Aller (1978) successfully used Berner's (1976) one-

dimensional diffusion model to describe the vertical distribution of

interstitial phosphorus in marine sediments. In the presence of the

highly mobile protobranch, Yoldia limatula, pore-water transport was

characterized by a non-steady-state two-layer model in which an

effective (or biogenic) diffusion coefficient acts in the zone of

feeding and molecular diffusion controls transport in underlying

sediment. The effective biogenic diffusion coefficient for pore-water

transport by Yoldia is approximately 1 x 10-5 cm2/s; this compares

to a tortuosity-corrected molecular diffusion coefficient of

2.4 x 10-6 cm2/s for HP04-2 in undisturbed sediments. Lerman (1979)

indicated that effective diffusion coefficients in the mixed layer are

on the order of 5 to 100 times greater than the molecular diffusion


Several other types of models have been developed to describe the

effects of bioturbation on sediment-water solute exchange. Grundmanis

and Murray (1977) and McCaffrey et al. (1980) suggested that transport

can be simulated by assuming that interstitial water is biogenically

advected between discrete well-mixed reservoirs within the sediment and

overlying water. Aller (1978, 1980) proposed an idealized model which

assumes that specific changes in the average geometry of molecular

diffusion result from the presence of irrigated tube and burrow

structures. Average interstitial concentrations, solute flux, and

apparent one-dimensional diffusion coefficients are influenced by both

the size and spacing of burrows.

Gas Ebullition

The evolution of gas bubbles within anaerobic sediments also may

enhance the flux of dissolved constituents across the sediment-water

interface. Two mechanisms have been proposed by Klump and Martens

(1981) to explain the influence of gas ebullition on exchange. Bubble

tube structures alter the geometry of the sediment-water interface and

increase the surface area available for chemical exchange. If flow

patterns at the interface permit unrestricted import and export of

overlying water within tube structures, the total flux across the

interface is increased. Conversely, restriction of flow may allow

solute concentrations within the tube to reach high concentrations,

establishing a sharp vertical concentration gradient across the

interface. Mixing of the pore water in the tube by rising bubbles can

mix the pore water along the length of the tube, creating a sharp

gradient immediately at the interface. Enhancement of diffusive

transport consequently would result from the increased gradient.

The critical concentration at which bubbles form is dependent on

hydrostatic pressure; thus, resistance to bubble formation increases

with increasing depths (Hutchinson 1957). Gas ebullition generally

occurs in shallow aquatic systems such as marshes and littoral zones,

where surficial sediments are the site of intensive anaerobic metab-

olism. In hypereutrophic Wintergreen Lake, a shallow (T = 3.5 m),

hardwater lake in Michigan, Strayer and Tiedje (1978) measured an

average methane ebullition rate of 21 mmol/m2-d during the period

late May through August. Maximum rates of methane loss by bubble

evolution (35 mmol/m2-d) occurred in late summer. In comparison,

Martens and Klump (1980) measured an average methane ebullition of

16.8 mmol/m2-d from intertidal sediments on the North Carolina

coast between June and October. Ebullition was initiated by the release

of hydrostatic pressure that accompanied low tide. High fluxes of total

dissolved phosphorus were observed; rates up to 120 umol/m2-hr were

observed from these sediments during the summer months. The high rates

were attributed to increased mass transport associated with bubble tubes

maintained by methane gas ebullition.

Sediment Resuspension

A third mechanism that may contribute to internal loading is the

resuspension of sediments in shallow lakes by wind-induced currents

(e.g., Sheng and Lick 1979). Fee (1979) correlated rates of primary

production in unfertilized ELA lakes in western Ontario to the fraction

of epilimnetic surface area in direct contact with bottom sediments.

Fee concluded that epilimnetic nutrient recycling is dominated by

processes occurring at the sediment-water interface and not within the

sediment. Chapra (1982) recently showed that resuspension becomes an

important mechanism in lakes with mean depths less than 9.2 m. Lastein

(1976) showed that approximately 24 percent of the material collected in

sediment traps in Lake Esrom (Sweden) (z = 12.3 m) was from resuspended

sediment, and he hypothesized that wind-induced circulation alone was

sufficient to explain the phenomenon. In Lake Uttran (Sweden), a

shallow (T = 5.7 m), eutrophic lake, Ryding and Forsberg (1977)

correlated aqueous total phosphorus concentrations with the force and

duration of winds blowing lengthwise across the lake, and they suggested

that resuspension played a dominant role in regulating water quality.

In a study of six shallow Danish lakes, Andersen (1974) concluded that

wind-induced mixing of surficial sediments is the most probable

mechanism for the high phosphorus release rates he observed (up to

1.2 g/m2-month). Concomitant laboratory studies indicated that

maximum rates of release from quiescent, undisturbed sediments were of

the order of 0.25 g/m2-month.

Surface waves produce an oscillatory motion which translates in

shallow water to elliptical orbits extending to the sediment-water

interface (Bascom 1964; U.S. Coastal Engineering Research Center 1977).

Near the boundary layer, the elliptical motion is reduced to a simple,

reciprocating horizontal motion, and the maximum horizontal velocity

associated with the oscillating motion is given by

u = (I-I)
m Ts sinh (27 d Ld) (

where Hs = significant wave height;
Ts = significant wave period;
d = depth of the water column;
Ld = wavelength for the particular depth, d; and
um = maximum orbital velocity (Komar and Miller 1973).

When the shear force exceeds the bulk shear strength of the

surficial sediment deposits, sediment resuspension occurs (Alishahi and

Krone 1964; Komar and Miller 1973, 1975; Terwindt 1977; Sternberg 1972).

Scour may be accomplished either by detachment and entrainment of

individual flocs (surface erosion) or removal of aggregated material

(mass erosion) (Lonsdale and Southard 1974; Terwindt 1977). In deposits

where shear strength increases with depth, erosion continues to the

depth where the applied stress is equivalent to the bed strength (Krone

1976). Below the threshold or critical shear velocity, the bed remains

intact and no resuspension occurs. Migniot (1968) demonstrated that for

various types of cohesive sediments, the critical shear velocity may be

expressed as an inverse function of water content. These findings

subsequently were confirmed for marine sediments by Southard et al.

(1971) and Lonsdale and Southard (1974).

The critical shear velocity also is influenced by the mechanical

structure of the sediment, which in turn is related to the solute

characteristics and ionic strength of the interstitial solution and the

eroding solution (Arulananden et al. 1975). For example, critical shear

stress decreases with increasing sodium adsorption ratio (SAR)

(Arulananden et al. 1975):

SAR = [N (1-2)
([Ca2+ + [Mg2])1/2

This observation correlates with the Schulze-Hardy valence rule,

which indicates that the critical coagulation concentration of mono-,

di-, and trivalent ions are in the ratio of z-6, where z is the

charge of the ion (Stumm and Morgan 1981). Increasing the solute

concentration of the pore fluid destabilizes colloidal suspensions

(i.e., agglomerates them and tends to remove them from solution). The

net effect is an increase in bed strength (Arulanandan et al. 1975).

The effects of electrolyte concentration and SAR on critical shear

stress and erosion rates were reviewed in detail by Terwindt (1977).

Mixing or agitation of surficial sediments due to wave action

entrains interstitial (pore) water, resulting in release of solutes such

as nitrogen and phosphorus species to the overlying water (Gahler 19b9;

Lam and Jaquet 1976). For example, resuspension of the upper 1 cm of a

flocculent sediment with a porosity of 80 percent and a pore water

phosphorus content of 1 mg/L would result in a release of 8 mg P/m2.

This is sufficient phosphorus to supply a shallow lake (e.g., 7 = 3 m)

with an additional 2.7 ug P/L from this source alone. In addition, a

substantial fraction of the inorganic phosphorus sorbed to lake

sediments participates in rapid solid phase/aqueous phase exchange

reactions. This reservoir of phosphorus may be more important in

controlling the phosphorus flux resulting from sediment resuspension

than is the pore water phosphorus.

Lam and Jaquet (1976) developed an empirical model to predict the

upward flux of TP from resuspended sediment in Lake Erie. The flux is a

function of shear stress at the sediment-water interface (TB) from

wind-induced waves; TB was calculated from linear wave theory using

the deep-water Sverdrup-Munk-Bretschneider (SMB) wave-hindcasting method

to estimate wave height and period. Regeneration was assumed to arise

primarily from organic and nonapatite inorganic phosphate fractions, in

conjunction with a small quantity of soluble interstitial phosphate.

Apatite [Calo(P04)6(OH)2] was not considered in the regeneration

process because of its relatively high density and large grain size.

Sheng and Lick (1979) developed a similar sediment resuspension

model for Lake Erie; a principal point of departure from the Lam and

Jaquet model was the use of a shallow-water SMB wave-hindcasting model

to predict bottom orbital velocities.

Molecular Diffusion

The existence of concentration gradients across the sediment-water

interface implies that molecular diffusion may be an important mechanism

in describing solute transport. Theoretically, the flux across the

sediment-water interface constituent dissolved in the interstitial

solution arises from advection due to pore water buried by sediment

deposition, as well as from diffusion (Imboden 1975; Lerman 1979).

Neglecting the advective component, the flux equation reduces to Fick's

first law (Berner 1971):

J oDs () (1-3)
z = 0

where Js = flux from the sediment to the overlying water
(g/cm2-s with "s" denoting bulk sediment),

0o = porosity at the interface,
Ds = bulk sediment diffusion coefficient at the interface
(cm2/s), and
(--) = pore water (pw) concentration gradient at the
3z pw
sediment-water interface (i.e., z=0).

This equation is tractable if the assumption is made that Ds is

both spatially and temporally invariant. The flux of a dissolved

species across the sediment-water interface then can be calculated from

the concentration gradient at the interface and molecular diffusivity

(e.g., McCaffrey et al. 1980; Thibodeaux 1979; Vanderborght et al.


Fluxes as high as 7 mg P/m2-d attributed to diffusion have been

reported by Ulen (1978) for sediment cores from Lake Norvikken in

central Sweden. The average diffusive flux for Lake Norvikken was

1.0 g/m2-y; this contrasts with an external phosphorus loading rate

of 0.53 g/m2-y. Ulen (1978) obtained good agreement between

empirically measured fluxes and calculated rates of exchange using a

bulk sediment diffusion coefficient of 4.4 x 10-6 cm2/s. This value

was first derived by Tessenow (1972) and probably represents an upper

limit for bulk sediment diffusivities (Tessenow 1972, in Kamp-Nielsen

1974). Data of Li and Gregory (1974) indicate that the diffusivities of

H2PO at 25'C and infinite dilution are 7.34 x 10-6 cm2/s and

8.46 x 10-6 cm2/s, respectively; Manheim (1970) suggested that

these values are diminished by factors ranging from 0.05 to 0.5 in

unconsolidated sediments, as a result of porosity and path tortuosity


According to Kamp-Nielsen (1974), the aerobic release of phosphorus

from sediments in Lake Esrom (Sweden) is controlled by both adsorption

and diffusion. Berner (1976) modified the expression for Fickian

diffusion to quantify the retarding influence of adsorption:

D 0
s o ac
J () z = 0 (1-4)
s 1 +K 3z

where K = nondimensionalized linear adsorption coefficient.

Excellent agreement was obtained by Krom and Berner (1980b) between

empirically-derived effective diffusivities for phosphate, ammonium, and

sulfate ions and molecular diffusivities corrected for tortuosity and

adsorption. The influence of adsorption on the effective or apparent

diffusion coefficient can be quite significant; for example, Krom and

Berner (1980a) cite values for K ranging up to 5,000 for oxic oceanic



Sediments in lakes are generally characterized by high concentra-

tions of nutrient and other chemical constituents (such as heavy metals)

resulting from the deposition and accretion of detrital material. As a

result, sediments may actively participate in regulating the cycling of

a particular constituent in the water column. The extent that the

sedimentary reservoir can ultimately influence material cycling within

the water column depends on the availability of the substance for

exchange. Exchange of nutrients and other solutes across the sediment-

water interface into the overlying water is controlled by four principal

mechanisms: bioturbation, gas ebullition, resuspension of bulk sediment

by wind-induced turbulence, and molecular diffusion. In shallow lakes

characterized by extensive fetches, sediment resuspension is likely to

be a major mechanism for internal nutrient cycling. This research

focuses on the significance of sediment resuspension on rates of inter-

nal nutrient release in two lacustrine systems of similar morphometry,

but quite different trophic state--hypereutrophic Lake Apopka and

mesotrophic Lake Okeechobee. In conjunction with this objective, a

corollary objective of this study was to develop an event-dependent

model to predict the release or removal of phosphorus by wind-induced

sediment resuspension.


This research focused on two rather large and shallow lakes

important as aquatic resources in Florida--Lake Apopka in central

Florida and Lake Okeechobee in south Florida. By virtue of the broad

fetch and shallow nature of each basin, both Lake Apopka and Lake

Okeechobee are subjected to periodic sediment resuspension due to wind-

induced currents. The trophic state of these two lakes, however, is

quite different. Until the early to mid-1950's, Lake Apopka achieved

national reknown as a bass fishing lake. Currently and primarily

because of cultural inputs of nutrients, Lake Apopka is now widely

regarded as one of the most eutrophic lakes in Florida. Conversely,

Lake Okeechobee is intensively used as a recreational resource and is

critical to the hydrologic and ecological stability of south Florida.

The following section briefly summarizes the limnological characteris-

tics of these two systems and details the specific analytical procedures

used in this study.

Physical Description and Limnology of Lake Okeechobee

After the Laurentian Great Lakes, Lake Okeechobee is the largest

freshwater lake in the United States. Located in south central Florida

between latitudes 26'41' to 27013'N and longitudes 8036' to 81007'W

(Figure II-i), Lake Okeechobee has a surface area of 1,890 km2 at

normal stage (approximately 4.6 m above sea level). The broadest

expanse of open water lies along the north-south axis and approximates

56 km; the maximum east-west dimension is 48 km. The shoreline

development index, which is the ratio of the lake shoreline length to

the circumference of a circle with an equivalent surface area is 1.12

(Brezonik et al. 1979). Summarized in Table II-1 are some of the

important physical characteristics of the watershed.

Lake Okeechobee, which is considered moderately eutrophic (Brezonik

et al. 1979), is quite shallow throughout the entirety of its basin. At

normal stage, the depth of the lake averages only 2.8 m and has a

maximum depth of less than 5 m (Figure II-i). Because of its shallow

depth, Lake Okeechobee is characterized by an extensive littoral zone.

Macrophyte communities occupy nearly 500 km2 or 26 percent of the

total lake surface area along the southern and western shores of the

lake; however, no important macrophyte communities are established on

the eastern portion of the basin (Brezonik et al. 1979).

Lake Okeechobee in its present state originated approximately

6,000 years ago (Brooks 1974). It occupies a shallow depression which,

according to Hutchinson (1957), was formed by epeirogenetic uplift of an

irregular marine surface. Brooks (1974) has hypothesized that

differential subsidence of the thick underlying sequence of Miocene

clays is responsible for the deeper portions of the lake. Extensive

discussions of the origin and early history of Lake Okeechobee have been

presented by Brooks (1974) and Brezonik et al. (1979).

O II 0 A
S T'-""" -A--

I fi o *io -K-

Figure II-i. Location Map of Lake Okeechobee.

Table II-I. Physical Characteristics of Lake Apopka and Lake Okeechobee.

Lake Apopka*

Lake Okeechobeet

Lake Surface Area (km2)

Maximum Depth, zm (m)

Mean Depth, (m)

Volume, V (m3)

Annual Inflow (m3)




2.14 x 108

5.92 x 107

Theoretical Retention Time (years)

Maximum Length (km)

Maximum Width (km)

Shoreline Length (km)

Shoreline Development Index, DL

Watershed Area (including lake) (km2)

Watershed Land Area (km2)











5.24 x 109

1.98 x 109 to
4.68 x 109

1.12 to 2.65







* From Brezonik et al. (1978).
t From Brezonik et al. (1979) and Federico et al. (1981).

Excluding the lake surface, the watershed of Lake Okeechobee drains

approximately 11,116 km2. The watershed comprises a series of both

natural and man-made sub-drainage basins (Figure 11-2). Flow is

primarily from north to south; the Kissimmee River, which drains

6,048 km2, is the most important tributary (Davis and Marshall 1975).

Other natural basins of importance include Fisheating Creek (1,194 km2)

and Taylor Creek which drain 1,194 and 477 km2, respectively. Since

the early 1900s, a 785 km2 area of the Everglades immediately south

of the lake has been added to the watershed via an extensive system of

canal and pump stations. This region, otherwise known as the Everglades

Agricultural Area (EAA), is dominated by Everglades muck soils and is

intensively farmed. Flow in the EAA is conducted along the North New

River, Hillsboro, and Miami and, depending on flood control and water

conservation requirements, can be either imported from or exported to

the lake. Hydraulic export from Lake Okeechobee is also conducted by

the St. Lucie, West Palm Beach, and Caloosahatchee canals, whereas

outflow historically occurred as sheet flow over the southern rim into

the Everglades.

Land use within the Lake Okeechobee watershed has been compiled by

McCaffrey et al. (1976). Results from the compilation are presented in

Figure 11-3 and Table 11-2. In general, the watershed is dominated by

some type of agricultural activity. Agricultural lands constitute

48 percent of the watershed and comprise primarily improved and

unimproved pasture land in the northern and northwestern portions of the

basin and croplands in the south. Wetlands are the next most signifi-

cant land use type (18 percent), followed by surface water (18 percent

Figure 11-2. Surface Water Drainage Basins in the Lake Okeechobee Watershed.

Note: Areas designated by dashed lines contribute only small amounts of water
at irregular intervals. Drainage basins south of the lake have been
added by man. From Federico et al. (1981).

Figure 11-3. Land Use Within the Lake Okeechobee Watershed. From
McCaffrey et al. (1976).

0O *. I OI
Sj o i I I I
4.- a' i C14 CM a
u 0 l) Iu

o -
4O*- 1 -t-t Ic I I I 0
co c) D.0 0 I 00 I I 1
o Ccu C 10 C- I cJ
00 o sj ccO u' t
m 0 w m 00 C-T

0O M
"O j 4 CB
V) C- CI I a
a < o *1 I I i
uE-i L n c L 4
- UiL CN en
U C- 0) -
r- P. 0

00 00 I0 1 1 0
o r-. o0 a I -t I I I o
" Cc L N 0 IT


ca a.


(Ua) A CV) C 14 C14 M 00
0-C ca u -- 4c'J0'Co -^ cI C4 Co

Sw -It c c -
0) 4 -i -

ca 4a

m a '-0% (C) C O\ O Z u' m o c
CO V a) C.J a'C,-4' Lo r C '.0
cc o O E cO .j 0''.0irmcJ cri oO r -
a) m1 o I 00 -n 0 U
u .c 0 "a v o -1 0 C M N 0 P

) u u

0 0
Ud -a

) IV Jj Cu

= ( u o m -tC0 al 0o 0 o

cO' uE -3 r- -0n 00 o r-? '
- c. -1- CM -
o w (U 0

Cu a)
-a ,OC O '.0a
*4 U CCU (LWCI Ca ,'

*-4 -^

'0 I. -
CuQ u4 I-. C

41 u 5 m -
u 0. -4 V) c Cau 3r 3c
-a -4 l-4 '0 I" V ^j

M 3 0 Q m U3


including Lake Okeechobee) and forests (6 percent). Urban areas

(5 percent), which include the towns of Clewiston, Belle Glade, and

Okeechobee City, represent the smallest area of the total watershed.

Physical Description and Limnology of Lake Apopka

Lake Apopka is the fourth-largest lake in Florida, encompassing a

surface area of 124 km2. Located approximately 25 km northwest of

Orlando between latitudes 28*33' to 2841' and longitudes 8131' to

81'42', Lake Apopka forms the headwaters of the Oklawaha chain of lakes,

and ultimately the Oklawaha River (Figure 11-4). As indicated by its

shoreline development index (1.39), the lake is somewhat circular in

shape and is characterized by relatively broad fetches extending along

both the north-south and the east-west axes (14.4 and 13.8 km,

respectively). Important physical characteristics of the watershed are

summarized in Table II-1.

Lake Apopka resides at the southernmost end of the physiographic

region known as the Central Valley, which is bounded on the west by the

Lake Wales ridge and to the east by the Mount Dora Ridge. The lake

itself occupies an extremely shallow basin with an average depth of only

1.7 m at normal stage (20.3 m above MSL) (Figure 11-5); consequently,

Lake Apopka is continually well mixed by wind-driven currents. Despite

its shallow nature, the lake is characterized by the virtual non-

existence of rooted macrophytes. This apparently stems from compression

of the photic zone to much less than 1 m (approximately 40 to 60 cm)

because of high algal densities characteristic of the hyper-eutrophic

lake and the instability of the highly flocculent sediments.


Location Map of Lake Apopka.

Figure 11-4.


Figure 11-5. Bathymetric Features of Lake Apopka. Depth Contours Given
in Feet.

The watershed of Lake Apopka is rather small in relation to the

lake surface area, and excluding the lake surface, drains only 187 km2

No natural surface influents of any consequence convey water to the

lake. The major hydraulic input is precipitation falling directly on

the lake surface; the other important natural source is artesian flow

from the Floridan Aquifer at Gourd Neck Springs in the southwest corner

of the lake. During periods of intense rainfall, Johns Lake located

2.0 km to the south may overflow into Lake Apopka. Several other small

streams in the southeastern basin Lake Apopka may flow into the lake;

these streams are also ephemeral, and their existence is a function of

antecedent rainfall.

Similar to Lake Okeechobee, the watershed of Lake Apopka has been

increased by cultural activities. Around 1940, a vast (7,200 ha) marsh

was drained and converted into highly productive muck farmland. Flow,

which historically has been south to north, was diverted from the marsh

by an earthen dike and conducted northward through the Apopka-Beauclair

Canal. A series of canals and pump stations was constructed to regulate

water level within the muck soils. As a result, backpumping for the

for the muck farms has constituted a significant hydraulic (and

nutrient) source; however, since 1976 the magnitude of this source has

diminished considerably with the institution of discharge abatement


Land use estimates within the Lake Apopka drainage basin have been

developed by the East Central Florida Regional Planning Council (1973)

and are included along with projections for 1990 in Table 11-2. At the

time of the above study, 8 percent of the land area was urbanized, while

citrus growing and muck farming activities occupied 43 and 39 percent of

the land, respectively. As can be seen from Figure 11-6, muck farming

is limited to the northern perimeter of Lake Apopka, and most of the

remaining land in the lake basin is occupied by citrus groves. Urban

areas are primarily limited to sections on the south and west shores of

the lake and include the towns of Winter Garden and Montverde.

Materials and Methods

Physical and Chemical Characterization

Samples were collected at 14 stations within Lake Okeechobee

(Figure 11-7) to determine physical and chemical characteristics of

surficial sediments. Five stations (01, 03, 04, 08, 09) were located

within the littoral zone; five stations (06, 07, 010, 011, 012) were

open-water stations; and the remaining four stations (02, 05, 013, 014)

may be described as transitional. Samples were collected during three

sampling efforts in the summer and fall of 1980. Sediment samples from

Lake Apopka were collected quarterly at nine open lake stations (Al

through A9) and one station at Gourd Neck Springs (A10) (Figure 11-8)

between March 1977 and March 1978. All samples were collected with a

Ponar dredge and placed in wide-mouth one-liter polyethylene bottles,

stored on ice until return to the laboratory and then refrigerated until


All sediment samples were homogenized prior to analysis. Water and

volatile solids content were determined on a percent basis according to

standard methods (APHA 1976). Particle size distribution was evaluated

by wet sieve analysis for sand-sized fractions (>63 um) and by pipette

I /

*:- -

Citrus Gro e
Apopka -Beauclair Citru o .
i ; : : .: :-. .

U.S.G.S. Drainage Boundary .

Figure 11-6. Land Use Within the Lake Apopka Watershed.

Figure 11-7. Sediment Sampling Locations in Lake Okeechobee.


Sediment Sampling Locations in Lake Apopka.

Figure 11-8.

analysis for silt-clay fractions (Guy 1969). In order to characterize

the effective particle size of the whole sediment, sedimentary organic

matter was not removed by oxidation. Effective particle size is

necessary in determining the threshold shear required to induce sediment

resuspension for a particular sediment.

The phosphorus content of selected sediment samples was analyzed in

terms of three fractions: apatite inorganic phosphorus (AI-P),

nonapatite inorganic phosphorus (NAI-P), and organic phosphorus. The

latter fraction was determined as the difference between the total

phosphorus (TP) content and the sum of two inorganic fractions. The

inorganic fractions were estimated on fresh undriedd) sediments by the

analytical method described by Armstrong (1979). This method involves a

two-step sequential extraction scheme: the sample is initially

extracted with 0.1 N NaOH for 16 hours to obtain NAI-P followed by an

extraction with 1 N HCI for 16 hours to determine AI-P. TP was

determined on fresh sediments by the acid-persulfate digestion technique

(APHA 1976). Digested samples were filtered, neutralized, and analyzed

for soluble reactive phosphorus (SRP) using the single reagent

molybdenum blue technique (U.S. EPA 1976).

Interstitial water was extracted at 4C by centrifugation of fresh

sediment samples at 10,000 rpm for 30 minutes. The supernatant was then

filtered through GF/C glass micro-fibre filter paper, and the filtrate

was analyzed for SRP by the single reagent molybdenum blue technique

(U.S. EPA 1976).

Sorption Experiments

A series of adsorption-desorption studies were conducted to

quantify phosphorus exchange dynamics using sediment samples representa-

tive of littoral transitional and profunded zones. Sorption isotherms

were evaluated at 220C; samples were buffered at pH 8.3 with NaHCO3

at an ionic strength of 8.5 x 10-3 M and 5.0 x 10-3 M. These conditions

reflect the average pH and ionic strength observed in Lake Okeechobee

(Federico et al. 1981) and Lake Apopka (Pollman et al. 1980), respec-

tively. Aliquots of wet sediment equivalent to 0.5 g of oven-dried

material were allowed to equilibrate with 100 mL of buffered water for

24 h on a rotary shaker table. The buffered solution was adjusted with

KH2PO4 to yield initial orthophosphorus concentrations of 0, 20,

50, 100, 250, 500, 1,000, and 5,000 ug P/L. After 24 h of shaking, the

suspensions were filtered through GF/C glass micro-fibre filter paper

and subsequently analyzed colorimetrically for SRP using the method

described previously. The quantity of phosphorus adsorbed or desorbed

was determined from the difference between the final and initial

concentrations and was corrected for contributions of SRP from

interstitial water.


Physical Characteristics

Erosion, transportation, and deposition of sediment particles in a

lake are governed by many physical factors, including effective fetch,

water depth, bottom depth, and lake morphometry (Hakanson 1981). The

sediment parameter that probably best indicates the dynamic conditions

prevailing at the sediment-water interface is the water content of

surficial sediment. For example, Migniot (1968) demonstrated an inverse

relationship existed between critical shear stress and sediment water

content for a series of lacustrine muds. Hakanson (1977) derived an

empirical relationship that describes erosion and deposition of

sediments in Lake Vanern (Sweden) as a function of sediment water

content and maximum effective fetch (Peff) (U.S. Coastal Engineering

Research Center 1977). The water content of surficial sediments in

Lakes Okeechobee and Apopka thus was determined to evaluate the physical

nature of the sedimentary environment, i.e., whether a particular

location within a lake basin is a zone of relative quiescence or an area

subject to turbulence and resuspension.

Results for water content analyses for Lake Okeechobee and Lake

Apopka are summarized in Tables III-1 and 111-2, respectively.

Sediments from the central portion of Lake Okeechobee had water contents

ranging from 82.3 to 91.1 percent. With the exception of Station 09,

C I r- n 0
0 I C 4. -CP

00 r- qTC o
-I 0 -I 00 00
C,4 't M e

a%0 Ca -It ( It r-> -It 00 ON
C t ~< o 00 ir --4 in cn
M n r~.l-r


0 =



Cc to
Ir P










u M



OOOc c



- o-

-) 0 CM

s*o *

o 0 CM4 r-
S0O -- 0
- 000





<* -i -i CM n

00 0 IT T 0

0 -000


1 0 0 0

00 0 ci I
-- ON -'T I

- oo 0

0 o

cnI0 o0 -* 0
It C4C-


< L
CL l

.M 4-1

r- Cj %O CM
CM 3- r- 0




-i C

00 O% 'T en

0 C-M CM



r,0Nt0UN C

o C4o c 0
CM C- r- 00 N
C M 00cw



Cu U~

a) -4



a 0
ii e

).-4 -



2 2j-



ni '

*r- *

!< u
o d
l- 0 ^-
& (U


M 2

a7' CM 00

0 0
0 r- .0.
r0 ,0 rC

00 '. o


c O oo uL- t

-00---C --CM

- a

0 3

a D a'n ao 'o 'D a'a
.C .

the water content of littoral sediments within Lake Okeechobee was

significantly lower (p <0.05) than that of the central sediments and

averaged only 49.6 percent. Station 09 was in the vicinity of beds of

rooted macrophytes and emergent vegetation, which effectively dampen

wave energy, thus permitting the accumulation of finer particles and

detritus. Water content of sediments from transitional zones was quite

variable, ranging from 26.7 to 88.6 percent, and reflected the gradation

from a highly dynamic environment to more quiescent conditions. The

difference between central-lake and littoral-zone sediments in Lake

Okeechobee is further illustrated by the relationship between the

fractional content of silt- and clay-sized particles and water content

(Figure III-i). A close correlation (r = .94) was found between

sediment water content and particle size distribution:

W = 41.16 + 12.66 (In B) (III-1)

where W = percent water content, and
B = cumulative mass percent (sediment particles < 63 um in

Thus, the near-shore sediments within Lake Okeechobee tend to be large

grained with low water content, while the central-lake sediments are

fine grained with high water content. A similar relationship for Lakes

Vanern and Ekoln (Sweden) was derived by Hakanson (1977), who approxi-

mated water content as a function of mean particle size. In general,

water content (or more specifically porosity) increases with decreasing

particle size, as a result of increasingly effective electrostatic

(repulsive) forces between clay minerals (Berner 1971, 1980).





0 "m
IIn U,

\ fO

iLI u

: to

o wUI .-

\ -^

\L L i :
\, e ^ ,
^^^^ ^
^^^^^^ h-
---- ---- ---- ----*^* 1 ^
0 0 0 0 0 o vxm
<^ ~~~ ~ OO l< O l
(%)~~~ 1N1NOd3V

In contrast to Lake Okeechobee, the sediments of Lake Apopka are

relatively homogeneous with respect to water content; this trend implies

a somewhat uniform hydrodynamic environment. Lake Apopka is charac-

terized by extensive deposits of highly flocculent sediments with water

content ranging between 95.1 to 97.0 percent. According to Schneider

and Little (1969), these unconsolidated deposits cover 90 percent of the

lake bottom with an average thickness of 1.5 m. Excluding the immediate

shoreline, stable sediments comprising sand and shells are restricted to

four small areas. The largest of these areas is a 240-ha region in the

northern portion of the lake basin; the second largest region

encompasses 120 ha adjacent to Crown Point near the eastern shoreline.

Sly (1978) suggested a simple approach to evaluate the occurrence

of sediment suspension. According to Sly, the depth at which

wave-induced shear stress exceeds the critical shear stress (Fc) of

bed materials corresponds to 25 percent of the available wavelength

(based on significant wave period). For Lake Okeechobee, the deepest

portion of the basin is approximately 4.3 m deep and has a maximum fetch

of 20 km. Applying linear wave theory (U.S. Coastal Engineering

Research Center 1977) to Sly's (1978) model, one calculates that sedi-

ment transport can be initiated in the deepest part of Lake Okeechobee

at continuous wind velocities of 13.1 m/s (29 mph) or greater. It is

thus apparent that sediment erosion and transport can be induced

throughout Lake Okeechobee and that sediment accumulation tends to be

restricted to relatively protected areas along the southern edge of the

main basin (South and Pelican Bays) and the subbasin in the northern

portion of the lake. The relatively low water content (<90 percent)

generally observed throughout the lake and especially in the deeper

portion are indicative of substantial wave energy extending to the

sediment surface. For example, according to Hakanson's (1981) model for

sediment transport, sediments at a depth of 1 m in accumulating regions

of Lake Okeechobee should exhibit water contents exceeding 80 percent.

In Hakanson's model, the critical water content defining the limit

between sediment transport and accumulation was found empirically to


Wc = Wmax 10 (111-2)

where Wc = the critical water content, and

Wmax = the water content at the maximum depth.

Of the three samples collected at this approximate depth (Stations 03,

08, and 014) only Station 014 located within South Bay had a water

content suggesting active sediment deposition. Despite the expansive

fetch that extends to South Bay during wind events from the north

(nearly 58 km), South Bay apparently is sufficiently sheltered by Rocky

Reef, macrophyte vegetation, and by the shoreline configuration to serve

as a settling basin. Moreover, prevailing winds originate from the

east, resulting in more limited fetches for South and Pelican Bays.

According to Hakanson's (1981) model, the only other station within

Lake Okeechobee with sediment water content indicative of relatively

undisturbed accumulation is Station 09, which is located within a

protective stand of rooted macrophytes. The central (open-lake)

stations experience a mixture of deposition and erosion, resulting in

the accumulation of silt- to sand-sized particles (Table III-l). In

shallow lakes such as Lake Okeechobee, continued circulation and

upwelling prevent deposition of extremely fine particles, as Sly's

(1978) model predicts.

Excluding several relatively deep and narrow trenches along the

western and southern shores (4.9 and 5.5 m, respectively), the deepest

portion of the major basin of Lake Apopka is only 3.0 m in depth and has

an effective fetch of 7.4 km for winds originating from the east.

Under these conditions, Sly's (1978) model indicates that resuspension

can be initiated at wind velocities of 9.4 m/s (21 mph) or greater.

This result suggests the ease with which sediment resuspension occurs in

Lake Apopka. Because of its morphometry, quiescent regions of

undisturbed deposition are essentially nonexistent. Indeed the water

column of Lake Apopka often assumes a brownish cast because of

wind-induced scouring of the bottom sediments; for example, turbidities

in Lake Apopka often exceed 40 NTU (cf. Brezonik et al. 1978, 1981;

Tuschall et al. 1979; and Pollman et al. 1980) and have been observed as

high as 120 NTU (see Chapter IV).

Use of Sly's (1978) model as a guideline indicates that virtually

the entire basin of Lake Apopka is subjected to resuspension; the only

regions falling outside Sly's (1978) threshold criterion are the

previously mentioned trenches along the western and south shores and the

relatively protected subbasins comprising Gourd Neck Springs. Given the

extremely flocculent nature of the sediments, the existence of such

trenches is unusual. Bush (1974) explained their existence as a result

of scouring created by wave setup. Wind-induced surface currents are

maximized along the longest reaches of a lake, resulting in a piling-up

of surface water toward the leeward shore. The pressure thus exerted

induces a return flow along the bottom in the opposite direction,

scouring the bottom of muck.

These results conflict with a priori conceptions concerning

sediment water content and lake hydrodynamics. The consistent and

extremely high water content of Lake Apopka sediments suggests (in the

absence of other information) that quiescent conditions prevail

throughout the basin. A similar conclusion is reached from Hakanson's

(1981) model, which indicates that water contents in excess of

90 percent reflect an undisturbed depositional environment. It is

apparent from the previous discussion of Sly's (1978) model that the

morphometry of the Apopka basin precludes undisturbed deposition. The

question then arises as to why such high sediment water contents are

observed throughout Lake Apopka.

Lake Apopka, by virtue of its shallow nature and inherently high

rates of algal productivity, constitutes a rather unique depositional

environment that falls beyond the framework of Hakanson's (1981) model.

Average rates of net primary production as high as 465 mg C/m3-hr

have been observed for Lake Apopka, with an annual average rate of

140 mg C/m3-hr reported by Brezonik et al. (1978). Compared with

the threshold level of 95 mg C/m3-hr reported by Brezonik and

Shannon (1971) for eutrophic lakes, these results show that Lake Apopka

is highly eutrophic. As evidenced by sedimentary concentrations of

volatile solids, high rates of deposition of organic detritus are a

direct -consequence of high rates of algal production. Average sedimen-

tary organic content (as defined by volatile solids) ranged between 56.0

and 77.4 percent in Lake Apopka, and the lakeside average was

64.5 percent. Organic detritus, which has a bulk density approaching

water, is readily resuspended and transported by wind-induced currents,

and its accumulation in sedimentary deposits generally reflects not only

rates of biogenic material production but also the vertical distribution

of hydrodynamic forces. In most lakes, then, transitional hydrodynamic

environments occur that result in selective deposition of particles as a

function of size and density. Consequently, sediment-focusing or high

rates of deposition of organic-rich material are observed in the deeper

reaches of these systems (cf. Davis and Ford 1982). However, because of

the configuration of Lake Apopka's basin, relatively undisturbed

depositional zones are non-existent. Consequently, a virtually uniform

layer of highly flocculent, organic sediments prevails throughout the


Unlike Lake Apopka, the distribution of organic material in Lake

Okeechobee clearly demonstrates a transitional hydrodynamic environment.

Organic content varied with station location in the basin, ranging from

0.3 to 42.7 percent (dry weight) (Table III-i). Sediments in littoral

areas (excluding Station 09) exhibited very low concentrations of

organic matter, averaging only 2.2 percent. Organic content of open-

lake sediments was significantly higher and ranged between 12.7 and

35.1 percent.

Two classes of open-lake sediments in Lake Okeechobee can be

identified from the results of this study. The first class consists of

highly organic muck (Stations 07 and 012), with sediment organic content

approximately 34 percent. These sediments tend to predominate in the

mid to north central portion of the lake basin. The second sediment

association, which occurred at Stations 06, 010, and 011, is less

organic in nature, averaging 15.1 percent, and was found in closer

proximity to the shoals of the western basin. Sediment organic content

was highest at Station 014 in South Bay, with organic matter

constituting 43.2 percent of the total solid content. South Bay, which

previously was identified as a settling basin, receives substantial

portions of inorganic nutrients from the Everglades Agricultural Area

(Federico et al. 1981).

Further evidence supporting the idea that detrital accumulation in

Lake Okeechobee is influenced by hydrodynamic factors is supplied by

evaluating the relationship between organic matter and water content in

Lake Okeechobee sediments. A logarithmic correlation (r = .92) was

observed between water content and organic content (Figure 111-2):

W = 43.21 + 12.37 (In VS) (111-3)

where VS = percent volatile solids dry weight.

A similar logarithmic relationship was noted by Nisson (1975) for

sediments from Lakes Harney, Jessup, and Monroe, on the St. Johns River

in central Florida.

If detrital accretion in sediments primarily reflects hydrodynamic

factors, a correlation also should be observed between particle size and

organic content. This relationship stems from the fact that the

conditions that favor the deposition of fine-grained sediments are also

conducive for the accretion of detritus. For sediments in Lake






0 0 o
0 X) 0 -

0 *



O .6

0 0
C w

0- a


> O n w


a)) a)
(0 01

0 >

V 36-4
0 0 0u

\ O- 0
XO tn >e
^L V~o aa3N0 eg~'

Okeechobee containing 23 percent organic matter or less, this is indeed

the case (Figure 111-3). In this range, organic content and silt-clay

content are linearly related. With organic content exceeding about

23 percent, however, the fractional silt-clay content remains constant

(approximately 20 percent), suggesting that in quiescent waters local-

ized rates of primary production assume a greater role in controlling

the magnitude of detrital deposition. The fractional silt-clay and

organic content of Station 06 does not conform to this relationship;

unique morphometric and hydrodynamic factors may permit greater

deposition rates of colloidal silt-sized particles at this site.

Chemical Characteristics

The chemical nature of sedimentary phosphorus defines its

availability for exchange and biological uptake. Differential dissolu-

tion schemes have been developed that correlate algal availability with

a particular fraction. Such fractionation schemes were applied to Lake

Okeechobee and Lake Apopka sediments, and results are described in this

section. Total P analyses, on the other hand, provide an estimate of

the total pool of phosphorus that ultimately is available for exchange.

Total phosphorus (TP) concentrations in Lake Okeechobee sediments

were highly variable and ranged from only 142 ug P/g (dry weight) to

1,440 ug P/g. Conversely, the distribution of TP in Lake Apopka

sediments was fairly homogeneous and reflects the uniformity of the

depositional environment in the lake. Concentrations in sediments

within the main basin ranged from 749 ug P/g at Station Al to 1,170 ug

P/g at Station A8. TP concentrations in Station A10 sediments at Gourd




ac o

o o
I O-.



0 m
c *

ca t"

>o 3o


Q) x N

S0 0 0

1711S S AV70 1N3083d
& en

.71 0 ^t1 (B3t~

Neck Springs were somewhat higher (1,320 ug P/g) because of the

deposition of large quantities of decaying water hyacinths (Eichornia

crassipes). Phosphorus concentrations in lacustrine sediments may range

as high as 7,500 ug P/g, but typically are less than 2,500 ug P/g (Jones

and Bowser 1978). In a survey of mineralogical constituents comprising

profundal Minnesota lake sediments, Dean and Gorham (1976) observed an

average TP concentration of 1,400 ug P/g in calcareous lakes. Although

it is apparent that TP concentrations in Lake Okeechobee and Lake Apopka

sediments fall somewhat within the lower end of the spectrum, the

observed concentrations approach and in some cases exceed the

1,000 ug P/g threshold level cited by Frink (1967) as indicative of


Within Lake Okeechobee, the highest TP concentrations generally

were observed in the pelagic reaches of the lake basin. TP concentra-

tions in pelagic sediments ranged from 502 to 1,440 ug P/g with an

average value of 866 ug P/g; concentrations in littoral sediments were

significantly lower (p <0.05), averaging only 294 ug P/g. These results

imply that selective size sorting processes influence sedimentary P

distribution; similar results have been reported by Frink (1967 and

1969) and Hakanson (1981), as well as others. Based on the relatively

poor correlation between water content and TP in Lake Hjalmaren

sediments (r = .68), however, Hakanson (1981) concluded that accretion

of sedimentary P is mediated by anthropogenic and/or chemical factors in

addition to hydrodynamic processes. The weak correlation (r = .57,

Table III-3) between TP and water content in Lake Okeechobee sediments

implies a similar conclusion.

Table III-3. Summary of Sedimentary Component Interrelationships
with Total Phosphorus Concentrations in Lake
Okeechobee Sediments.


NAI-P vs. TP .46

AI-P vs. TP .81

Interstitial SRP vs. TP .19

Volatile Solids vs. TP .76

Water Content vs. TP .32

As indicated by linear regression analysis, the accumulation of

sedimentary P in Lake Okeechobee is dictated to a significant extent

(p <0.05) by the deposition of apatite-inorganic P (AI-P) (Table III-3).

With the exception of organic content, correlations between TP and other

sediment parameters were weak and insignificant at the 95 percent

confidence level.

Mechanistically, the functional dependency of sediment TP content

on AI-P and organic content may be explained by evaluating the relation-

ship between calcium-carbonate solubility, calcium-phosphate solubility,

and pH changes induced by primary production. The premise of this

hypothesis is that deposition of AI-P and detrital material are causally

linked to the same forcing function--primary production. For example,

calcium carbonate precipitation may be induced by increases in pH that

result from net primary production (Kelts and Hsu 1978); calcite

surfaces act as heterogeneous nuclei for calcium phosphate minerals

(Stumm and Leckie 1971; Griffin and Jurinak 1973), resulting in removal

of phosphorus from the water column. Similarly, calcium-phosphate

minerals tend to become less soluble with increasing pH (within the

range for most natural waters) owing to a shift in phosphorus speciation

and, in the case of hydroxyapatite CalO(P04)6(OH)2, an increase in

hydroxide concentration. To test the hypothesis that AI-P deposition

may be influenced by biogenically induced pH shifts, the stability of

various calcium phosphate mineral phases in Lake Okeechobee was

evaluated by comparing measured ion activity products (IAP) for the

species with thermodynamic solubility products (Ks0 values). For

purposes of this exercise, two pH values were used: 8.10, the average

pH and 9.10, the maximum pH observed by Federico et al. (1981) for a

series of stations near South Bay. Reported average Ca concentrations

approximated 46.5 mg/L. Average orthophosphorus concentrations at these

stations ranged between 0.020 mg P/L and 0.025 mg P/L; for the purposes

of this exercise a concentration of 0.0225 mg P/L was assumed. Of the

three mineral phases evaluated [CaHPO4, Ca4H(PO4)3, and hydroxy-

apatite, Cal0(PO4)6(OH)21, only hydroxyapatite is stable under ambient

conditions in Lake Okeechobee (see Table 111-4). During periods of high

photosynthetic activity (i.e., high pH), the water column is greatly

supersaturated with respect to hydroxyapatite (IAP/KsO >108), and this

may result in the direct precipitation and subsequent incorporation of

this species in the sediments.

Similar relationships between sedimentary TP and sediment inorganic

P were observed by Wildung et al. (1974) in Upper Klamath Lake, a non-

calcareous lake located in Oregon. Unfortunately, inorganic P as

reported by Wildung et al. (1974) includes both non-apatite inorganic P

(NAI-P) and as AI-P, thus precluding a direct comparison of results. In

direct contrast to the results of this study, however, Williams et al.

(1980) found a weak, negative correlation between TP and AI-P in

sedimentary material derived the drainage basins of Lakes Ontario and

Erie. Williams et al. (1980) demonstrated strong correlations between

variability in TP concentrations and changes in both NAI-P and organic P

fractions; in Lake Okeechobee sediments, however, neither parameter

yielded a significant (p <0.05) correlation with sediment TP.

NAI-P, which is indicative of the sedimentary P reservoir directly

available to algae upon resuspension (Williams et al. 1980; Allan et al.

Table III-4.

Solubility Product Constants (Ks0) and Calculated Ion
Activity Products (IAP) for Several Calcium Phosphate Mineral
Phases as a Function of Ambient Ca (1.16 x 10-3 M), SRP
(7.26 x 10-7 M), and pH Levels in Lake Okeechobee.

pH 8.10 pH 9.10
Solid Phase log Ks0* log IAP log (IAP/Ks0) log IAP log (IAP/Ks0)

CaHPO4 -6.6 -10.1 -3.5 -10.1 -3.5

Ca4H(P04)3 -46.9 -55.1 -8.2 -54.9 -8.0

Ca(P04)6(OH)2 -114 -111.6 2.4 -105.4 8.6

* From Stumm and Morgan (1981).

1980), generally is low in Lake Okeechobee sediments. Concentrations

ranged from 6 to 96 ug P/g and generally constituted less than

9.0 percent of the total P content. From vertical profiles of nutrients

in Lake Okeechobee sediment cores Brezonik et al. (1981) inferred a

sediment suspension or bioturbation depth of 10 cm; for Station 07

sediments this implies a labile P pool of approximately 1.04 g P/m2

which is sufficient to increase existing levels by 305 ug P/L (of labile

P) if the entire reservoir was released in a single pulse event.

Despite the relatively low concentrations of NAI-P, this calculation

illustrates the inherent potential for photosynthetic stimulation

induced by nutrient exchange across the sediment water interface.

Concentrations of NAI-P in Lake Apopka sediment were relatively

uniform and ranged from 83 to 111 ug P/g. Similar to Lake Okeechobee,

this fraction comprises only 10 percent of the total P reservoir.

However, the inherent potential for this reservoir to support primary

production in the water column is considerable. For example, at

Station A5 with average NAI-P and interstitial SRP concentrations of

108 ug P/g and 0.80 mg P/L, respectively, the total potential release of

inorganic phosphorus to the water column by resuspending only 1 cm of

sediment is 49 mg P/m2. This is equivalent to an incremental

addition of 29 ug P/L to the water column. Of this total, the NAI-P

fraction contributes about 84 percent. It is important to note that

resuspension of sediment to a depth of 4.7 cm has been observed in Lake

Apopka during convective disturbances (see Chapter IV) and most likely

extends more deeply into the sediment.

The distribution of SRP concentrations in sediment pore fluids for

both Lake Okeechobee and Lake Apopka was somewhat erratic, and no trends

were discernible between sediment type and interstitial SRP. Inter-

stitial SRP concentrations in Lake Apopka were quite high and ranged

from 0.40 to 2.80 mg P/L. The source of this variability is apparently

a combination of physical, biological, and chemical factors. For

example, the ability of Lake Apopka sediments to further adsorb

phosphorus (and hence buffer interstitial SRP concentrations within a

relatively narrow concentration range) is virtually exhausted and is

limited to extremely high concentrations. Under these conditions, small

changes in the concentration of phosphorus adsorbed to sediment particle

surfaces can result in very large changes in equilibrium SRP concentra-

tions in the pore fluid (see Chapter V). In addition, interstitial SRP

levels increase as organic phosphorus is mineralized by microbial

heterotrophs. Pore fluid concentrations apparently continue to increase

until the sediments are disturbed and this pool is exchanged with the

overlying water. The sediments then reequilibrate at a lower but still

highly elevated (relative to the water column) concentration.

Interstitial SRP concentrations in Lake Okeechobee sediments

ranged from 0.002 mg P/L to 0.942 mg P/L with the maximum concentration

found in the nutrient-enriched, highly organic sediments of Station 014.

On the other hand, the pore water content of SRP in Station 07

sediments, which also were quite organic, was only 0.012 mg P/L, and the

interstitial SRP content of the relatively inorganic, sandy sediments of

Station 04 was substantially higher (0.267 mg P/L). Despite the

variability in concentrations, it is clear that SRP concentrations in

the interstitial fluids were generally quite low, and this fraction

constitutes less than 0.1 percent of the total sedimentary P. The

ability of calcareous sediments (and specifically calcium carbonate) to

adsorb P is well documented (e.g., Griffin and Jurinak 1973 and 1974),

and Morse and Cook (1978) suggested that this mechanism accounts for the

low levels of dissolved interstitial SRP found in carbonate-rich marine

sediments. The small size of the interstitial reservoir implies that

this pool is not a significant source of phosphorus to the overlying

water via advective exchange across the interface. For example,

resuspension of 10 cm of sediments from Station 07 would result in

addition of only 0.3 ug P/L to the water from entrained pore water.


A series of passive exchange studies were conducted to evaluate P

release over a broad range of water content for Lake Apopka sediment.

Sediment was collected from Station A2 and placed in five small aquaria

with surface areas approximately 305 cm2. Aquaria were filled with

fresh sediment to a depth of approximately 13 cm and were exposed to

varying intensities of artificial light from two 150 watt spotlights to

generate a gradient of sediment dessication. Initial water content was

95.1 percent, final values ranged from 13.8 to 88.3 percent (Table IV-1).

As dessication proceeded, surface areas decreased correspondingly and

ranged from reductions of 18.5 to 78.9 percent. Following dessication,

the aquaria were refilled with phosphorus-free 6.0 x 10-4 M CaC03

to simulate ambient alkalinities observed in Lake Apopka and were

allowed to remain quiescent for a 9-day period. Samples of the

overlying water were taken on a daily basis to determine the quantity of

SRP released.

Phosphorus release by simple diffusion is illustrated in

Figure IV-1. The appropriate theoretical (Fickian) expression for

describing the flux of phosphorus across the sediment-water interface is

given by the solution to Fick's first law of diffusion:

J = 0 Ds -- (IV-l)

Table IV-1.

Final Volatile Solids and Water Content of Partially
Dessicated Lake Apopka Sediments Used in Passive
Exchange Experiments.

Water Content Volatile Solids
Aquarium (percent) (percent)

1 88.30 69.99

2 86.18 70.58

3 34.31 68.58

4 13.87 70.90

5 13.82 70.94


c( >

-4 W a)

0 *

S- -O o .

0.- 0 0B

0 *

\) ,Q) 4 0

> 0 0 ca
-4 4-

S'' C m

to CI N -W4

/d ~ ~ O ca d

o 0 0 0
10 \J d C- no



where J = flux of phosphorus across the interface (g P/cm2-s),
S= porosity of the sediment dimensionlesss),
Ds = the whole sediment diffusion coefficient (cm2-s),
dz= the concentration gradient at the interface (g/cm4).

This simple model assumes that the flux across the interface is strictly

due to molecular diffusion and that deposition or compaction of sediment

particles is negligible (cf. Lerman 1979; Berner 1981). The change in

concentration in the water column at any particular point in time is

simply the flux across the interface divided by the depth of the water

column; this of course assumes that the water column is well mixed and

is a reasonable assumption for the depth scale of these microcosm


Solution of Equation IV-1 requires that the concentration gradient

at the interface is known at a particular point in time; consequently,

application of Equation IV-1 to the experimental results is extremely

difficult. A more tractable approach is to fit an empirical model to

the data which may then be differentiated to describe the instantaneous

change in concentration as a function of initial concentration. In all

cases, release over the duration of the experiment (9 days) can be

described by the following logarithmic expression:

Ct = a + b In t (IV-2)

where Ct = the uniform concentration (mg/L) in the overlying
water at time t (days), and
a (mg/L) and b (mg/L/d) are both empirically derived constants.

The rate of change in concentration is simply

dC 1
b (IV-3)
dt t

Agreement of the data with the model, expressed in terms of the

correlation coefficient, r, is presented in Table IV-2. Results from

Aquaria 4 and 5 were combined because of the virtually identical water

contents. Equation (IV-3) can be applied to evaluate the instantaneous

change in concentration, and through mass balance considerations, the

instantaneous areal release rate for a specified time. Average and

instantaneous fluxes evaluated at the conclusion of the experiment are

presented in Table IV-3.

Release rates averaged over the 9-day exchange period ranged from

1.57 to 22.7 mg P/m2-d and were inversely related to water content.

These results were somewhat surprising and did not conform to a priori

expectations; that is, if the concentration gradient is independent of

the degree of dessication, sediments with greater water contents should

yield higher diffusive fluxes because of the decreased tortuosity

(Manheim 1970; Lerman 1979). Enhanced rates of detrital mineralization

with increasing dessication and a concomitant elevation of interstitial

SRP levels can account for the observed changes in flux; however,

analyses conducted by Fox et a!. (1977) of interstitial water SRP

content in fresh and dried Lake Apopka sediments demonstrated essen-

tially no difference between the two sediment types and thus do not

substantiate this mechanism. A more likely mechanism is given by

considering the change in sediment mass exposed on an areal basis as

dessication proceeds. As pore water evaporates and the sediments

Table IV-2.

Summary of Empirical
Parameters (Equation

Time-Dependent Concentration Model
IV-2) and Goodness of Fit.

Water Content
Aquarium (percent) a b r2

1 88.30 0.0268 0.0417 .995

2 86.18 0.0231 0.0415 .992

3 34.31 0.0416 0.1137 .986

4 and 5 13.85 0.0032 0.1951 .919

Table IV-3. Calculated Average and Instantaneous Phosphorus Flux at
Time = 9 Days During Passive Exchange Studies.

Average Flux Instantaneous Flux
Aquarium (mg P/m2-d) (mg P/m2-d)

1 1.57 0.55

2 2.18 0.77

3 13.9 5.12

4 and 5 22.7 9.71

consolidate, increasing quantities of underlying sediment particles

become directly exposed at the surface. A change in water content from

88 to 13 percent increases the exposed mass by approximately a factor of

10, which is close to the observed differences between the average flux

from sediments 4 and 5 and from 1 and 2.

Additional experiments to quantify the passive exchange of P

between Lake Apopka sediment and water were conducted using homogenized

sediment from Station A5. Sediment was transferred to a series of three

cylinders in sufficient quantity to yield a final depth of 14.5 cm. The

diameter of the cylinders was 10.2 cm, resulting in a surface area of

81.7 cm2 for exchange. The experiment was also designed to quanti-

tatively evaluate the effects of dessication induced by lake drawdown on

exchange; consequently two of the cylinders, B and C, were partially

dessicated by insolation to yield final water contents of 91.66 and

89.14 percent, respectively. The final cylinder, A, was allowed to

incubate at room temperature (22C) for the duration of the dessication

period (1 week). The water content of Cylinder A sediments was

94.89 percent. Following the dessication period, water was added to all

three cylinders simultaneously. One liter of filtered, aged lake water

from Lake Apopka, with an SRP concentration of 0.004 mg P/L, was added

to each cylinder to serve as the solution phase; this maximized the

concentration gradient across the sediment water interface and also

removed any algae present.

During the incubation period the sediments in Cylinder A began to

compact, lowering the sediment-water interface by approximately 0.7 cm

and forming a residual interfacial pool of SRP enriched water.

Subsequent extraction, filtration and analysis of 10 ml of the

interfacial water indicated an average concentration of 1.35 mg P/L.

The interfacial water comprised a labile pool of approximately 64 ug P

for immediate exchange upon addition of filtered lake water. Care was

taken not to disturb the interface when the filtered lake water was

added to each cylinder. Since the photic zone rarely extends below a

depth of 0.9 m in Lake Apopka, the diffusional phase of the experiment

was conducted in the dark. Free exchange with the atmosphere was

permitted, however, to prevent the system from going anaerobic.

Changes in SRP concentration with time are shown in Figure IV-2 for

all three cylinders. Cylinder A, representing fresh sediment, began the

initial phase of the experiment with the highest concentration of SRP

(0.192 mg P/L). Approximately 32 percent of this pulsed release was

engendered by mixing of the residual interfacial pool with added lake

water; the remaining release may have been induced by a slight

disturbance of mud-water interface and subsequent entrainment of

interstitial P and/or desorption from sediment particles. Because of

this pulsed release, the concentration gradient was reduced and

Cylinder A, consequently, showed the lowest rate of release. Calcula-

tions of release rates were based upon observed changes in SRP in the

overlying water during the initial phase of the experiment, when the

concentration gradient was most extreme and the resulting overlying

water concentrations approximated ambient SRP levels observed in situ.

Release rates were essentially linear during this phase, and as seen in

the previous exchange studies with A2 sediment, release increased as an

inverse function of sediment water content. Release from Cylinder A

approximated 4.3 mg P/m2-d while release from Cylinder C was

calculated to be 28.8 mg P/m2-d, (Table IV-4). Observed rates of


SI 4a)


I 0
I< I D "

f \ 'oo

I \ a

I \.-

"/ a6 )

I \

I I -
/ O

o 0 0 0 0)

I n eN


Q "u > m M 01 00

> W C.

3 3 .lOE




" >2

0 '
LU -1 0 ,

-4) X m 00 0

ou U o

m o
0 ca

o o

C-0 Q

m J 01 0 .r1
00 0
cu( 0. 0


w 00 1 0
4J 1-4 LM t -.T
aa u* 3-T C


>^ oe w
0 0

4 0 Q a-


0 0 X

all 0' LW <



E "V
2 0
~ (*0

release for Cylinders A and B are comparable to other values reported

in the laboratory for undisturbed cores (Table IV-5). In addition to

the previously described hypothesis regarding sediment exposure and

dessication, the observed flux in Cylinder A is also undoubtedly lower

because of the reduced concentration gradient across the sediment-water

interface. This effect can be derived directly from Fick's first law of

diffusion where flux is proportional to the concentration gradient. This

effect has also been empirically verified by Rippey (1977) who observed

that maintenance of low concentrations in the overlying water of

undisturbed cores increased areal release rates by 82 percent.

For Cylinders A and B, the linear phase of rapid release was

confined to an extremely short timeframe of 6 hours beyond which

decreases in concentration approaching initial levels occurred.

Concentrations in Cylinder A remained relatively stable for the next

42 hours, but by 72 hours they showed a further decline to a minimum

concentration of 0.172 mg P/L. After 72 hours, release was again

essentially linear in Cylinder A, and approximated 0.6 mg P/m2-d.

Similar erratic behavior was observed in Cylinder B; between 6 hours and

48 hours SRP levels fell to 0.007 mg P/L and then proceeded to increase

towards a secondary maximum of 0.023 mg P/L at 48 hours. Between 121

and 362 hours, SRP concentrations steadily increased. Release during

this period equalled 0.3 mg P/m2-d. Although the concentration

gradient was more pronounced in Cylinder B, the reduced flux after the

initial equilibration stage is more consistent with the decline in

overall porosity and the formation of a relatively impermeable crust

caused by partial dessication.


0. 0C

-4 -4 -








en iD

00 0






.-< 4.1
a ao




0 0 "a

wn c cuo

0- Cl 0

CM -. ( '



S.0 .0 .0

-u U (U O

(U Q) o *-< "-4 AJ
-4 -4 ~ -

-C a

CO 0

CU a c
U) CUa L L
ZC CU 4_ CO 7 CU C
O U C J U0) 0 4w ca
CU li (U .0 CU CU C .4.4 CU
O c I a a) *- *- C
- )3' U X Z 3 l: M

0 0


0 O

c a

0 C
o 0

o% r-,


a) Cis a)
0 0

a) oo Cc
j .,4 a)
c m z

i- I

'- I-.

o a

(U *
ca r
,4,1 U).
U) .-4
Co p




o m

0 0

= o
o N


- 00

cm C

N NnO m

I 'CMC-4

o^ '-'CM




W a
s ^


I Io
f1 0





o 04m




0' u'

- 0



C ..
6 3


U )

0 0

0 0 4-*

I I m l
0 0CU -4

Fox et al. (1977) conducted similar experiments with Lake Apopka

sediments to determine the effects of dessication on overlying water

quality. A general pattern of increasing P release with extent of

dessication was observed. Release as a function of dessication for both

this study and that of Fox et al. (1977) is presented in Figure IV-3 for

equilibration times of 144.5 hours and 1 week, respectively. Linear

regression analysis indicates a highly significant relationship between

the two variables (r2 = .74, p <0.025), yielding the following model:

SL = 1.44 0.01 W (IV-4)

where SL = concentration (mg P/L) observed after the stated
equilibration times, and
W = percent water content of the sediment.

The most consistent behavior with respect to release was observed

for the most dessicated sediment, Cylinder C. Release was relatively

constant over the initial 24 hours and began to decrease, apparently in

response to the diminished concentration gradient. Net release

continued until a state of dynamic equilibrium was achieved by Day 4,

when uptake essentially balanced release. Equilibrium was maintained

for approximately the next 93 hours following which time concentrations

dropped steadily for the duration of the experiment. The subsequent

decline in SRP is probably biological in nature; the longer period of

exposure to insolation and incident UV radiation undoubtedly had a

pronounced, deleterious effect on the benthic microflora in Cylinder C

relative to the other cylinders. If benthic microbial activity

0 0 00
CM 0 0
- .0



3Z w

I -* u>
-0 0 u
3 I I Io,

c c0
LJ -4 0)

a,. u Ia
o c

** *
U .-4 0

0 u 0)

\ ._ n > c
o LU

S 4. t

S -4 0

0 41 O OO
0 00 0

w r--

So o oo oo
\ 0*1

d oc 6o 0

(id oW 6d 0S 6
H/d 6^)ddS

increased as the experiment progressed a concomitant increase in SRP

uptake should have been observed; this mechanism is consistent with the

observed decline in SRP. It is interesting to note that concentrations

of SRP in both Cylinder A and B remained relatively constant during the

latter stages of the experiment; this may indicate that the benthic

microbial community was stable from the outset.

Calcium and magnesium concentrations were monitored during the

course of the study. Magnesium levels were essentially constant when

corrected for evaporative effects whereas corrected calcium levels show

a substantial increase of approximately 14 mg/L by the end of the

experiment. Sedimentary CaCO3 is apparently solubilized as a result

of heterotrophic activity producing CO2 (Lee et al. 1977). The

magnitude of Ca release for all three cylinders was approximately the

same, and thus solubilization of hydroxyapatite was not the primary

mechanism for release since the amount of P release differed markedly

among the cylinders.

Turbulence Study

As an extension of the quiescent nutrient release study conducted

with sediments from Station A5 in Lake Apopka, the effects of turbulence

on SRP exchange also were evaluated. After equilibration periods of 35

to 45 days to dampen phosphorus concentration changes associated with

diffusion, Cylinders A and C (representing fresh and partially

dessicated sediments, respectively) were subjected to varying degrees of

turbulence with a variable speed rotor with a paddle attachment. The

paddle encompassed a surface area of 21 cm2 and was situated 4.4 cm

above the sediment-water interface. The depth of the water column for

each system was maintained at approximately 11.2 cm.

Samples for SRP were withdrawn immediately preceding the start of

each experiment; subsequent samples were taken at measured time

intervals. All samples were filtered through GF/C glass microfibre

filter paper. Supplemental aliquots also were removed at periodic

intervals to determine suspended sediment concentrations. Equivalent

quantities of aged and filtered water from Lake Apopka were introduced

into each cylinder after each sample was withdrawn in order to maintain

a constant volume.

Fresh Sediment

Results of the turbulence experiment with fresh Lake Apopka

sediment are depicted in Figures IV-4 and IV-5. The initial phase was

conducted at a stirring velocity of 10.5 to 12.2 revolutions per minute

(rpm) sustained for 62 minutes. This velocity was sufficient to entrain

isolated, loose sediment particles at the interface, although the

interface remained essentially intact. Upon lowering the velocity to

9.0 rpm, the suspended particles began to settle. A net decline

(corrected for sample removal) of 0.002 mg P/L was observed during this

period, indicating virtually no effect of stirring at this rate on

equilibrated sediment.

The system was allowed to remain quiescent for another 67 minutes

before initiation of a second turbulence period. Rotor velocity was

increased to a maximum of 67.6 rpm, causing immediate scour of the

interface. Turbulence was maintained for exactly 1 hour. After a



a a.
0- -N I

Z cl

O o

2 bw) N/ *-
C. +--
JJ C CJa s

(0 B

I .







0 000

lw) al



4z I I

preliminary reduction in concentration of SRP in the overlying water at

the start of turbulence, SRP levels increased to 0.484 mg P/L and

suspended sediment concentrations reached 12.61 g/L by the conclusion of

the turbulent phase. Mass balance calculations indicate that between

25 to 68 percent of the net release in SRP was the result of entrained

interstitial water; the remaining fraction was apparently desorbed from

suspended particles and was equivalent to desorption of 12.3 to

28.4 ug P/g. More exhaustive studies concerning sorption phenomena are

presented elsewhere (see Section V, Adsorption-Desorption Studies). The

contribution of the pore water to release was calculated by assuming

that the concentration of SRP was between 0.50 to 1.35 mg P/L; the

latter value represents the interfacial water concentration at the

beginning of the diffusion study following exposure whereas the former

value constitutes the interstitial water content of the fresh sediment

prior to exposure. If release during the quiescent study is assumed to

be strictly diffusive, then 33 percent of the observed net release can

be attributed to entrainment of the pore fluid and the magnitude of

desorption approximates 25.6 ug P/g.

After stirring was concluded, resuspended sediment particles began

to settle rapidly, establishing a sharp interface between a clear

supernatant and the settling particles. Migration of the interface was

relatively slow (2.57 x 10-3 cm/s) during the first 7 minutes after

cessation of stirring because of turbulence within the water column

(Figure IV-6). After this equilibration period, water motion within the

cylinder became dampened and the settling velocity was much more rapid.

As the interface approached the point of origin, velocities decreased in




70. o

W 50-

? 40-

S 30- e

20- e


10 20 30 40 50 60 70 80

Figure IV-6. Migration of Interface Between Suspended Sediment and
Clear Supernatant with Time Following Cessation of

response to resistance from the compacted particles. The maximum

observed settling velocity was 1.95 x 10-2 cm/s. Some readsorption

of phosphorus occurred as the entrained sediment particles settled;

however, a significant increase in concentration relative to initial SRP

levels was observed after equilibrium had been established. The final

concentration was 0.355 mg P/L compared with an initial concentration of

0.159 mg P/L. These results conflict with Holdren and Armstrong (1980)

who observed that P release from Lake Mendota sediments generally

increased with increased stirring up to the point where sediment suspen-

sion occurred. Resuspension was found by Holdren and Armstrong (1980)

to decrease overlying water SRP concentrations to levels below those in

unstirred cores. Conversely, resuspension of Lake Apopka sediments

results in a substantial release of SRP, while stirring in the absence

of resuspension has virtually no effect on release. Indeed, in the

experimental system employed in this study, no change in SRP levels was

anticipated with stirring up to the point of bed failure because the

system was allowed to equilibrate for an extended timeframe. In other

words, the initial conditions of the experiment set ac/9t = 0, which

according to Fick's second law implies that any change in the diffu-

sivity constant caused by stirring will not alter the concentration of

SRP in the overlying water. In the cores used by Holdren and Armstrong

(1980), the derivative of the concentration gradient is non-zero; thus,

an increase in diffusivity induced by stirring will result in a

comcomitant increase in cc/lt.

Partially Dessicated Sediment

The effect of turbulence on SRP release from partially dessicated

sediments (Cylinder C) was minimal. As in the preceding experiment on

Cylinder A, the sediments of Cylinder C were subjected to two discrete

periods of turbulence separated by a quiescent phase to isolate any

observed effects. The initial turbulent phase was conducted at an

average rotor velocity of 13.9 rpm maintained for 1 hour. Concentration

changes were somewhat erratic during this phase (Figure IV-7) and when

corrected for sample removal and replacement with filtered Lake Apopka

water were equivalent to a net release of only 0.001 mg P/L. Rotor

velocity was insufficient to disturb the sediment-water interface,

although some unconsolidated particles were entrained.

After a stabilization period of nearly 2 hours during which the

overlying water SRP concentration remained essentially constant,

turbulence was again induced. Rotor velocity was increased to 115.5 rpm

and was allowed to continue for exactly 2 hours. Despite the extreme

amount of turbulence, the interface remained intact except for some

erosion from a small hole initially created by the removal of a sediment

subsample for analysis. A steady-state suspended sediment concentration

of only 0.0720 g/L was observed after 90 minutes. No phosphorus

desorption was detected in association with the minor amount of sediment

resuspended; however, a small but discernible decrease in concentration

of SRP in the overlying water did occur. The net decrease when

corrected for sample removal was 0.004 mg P/L and corresponds qualita-

tively to the initial decrease observed in Cylinder A when resuspension

was induced. The apparent mechanism for the loss from solution is



O c

S -.



z CL
z M
I | 8 -

0 1-
I- /I

00000 000 0
d roo r E

(-l/d bW) d6S
0 y

o / ? E^

1_ / -IJ
h- / 4-. >
QC / y-i ud

<~/ Yu ) dt

adsorption by resuspended particles derived from the surface; due to a

period of desorption and diffusive loss during the month long quiescent

equilibration phase, the surface particles apparently have less

surface-sorbed P than underlying sediments. Because of the reduction in

surface coverage of available binding sites (see Section V, Sorption

Characteristics), the relatively impoverished surface particles are free

to adsorb P upon resuspension. Further scour of underlying sediments

entrains surface-P rich particles which then undergo desorption (e.g.,

Cylinder A). The extent of scour in Cylinder C was apparently not

sufficient to resuspend these enriched particles; thus, a net decline in

SRP was observed.

The results of this experiment indicate that even partially

dessicated sediments are very stable and will not be resuspended even

under conditions of extreme turbulence. However, if the oxidized

surface layer is breached, exposing the less compacted sediments lying

underneath, localized scour is likely to occur.

In-Situ Studies

Water quality and limnological monitoring of Lake Apopka was

conducted between January 1977 and January 1981 as an integral part of

the Lake Apopka restoration program. The results of the study have been

presented elsewhere for individual years (Brezonik et al. 1978; Tuschall

et al. 1979; Pollman et al. 1980; and Brezonik et al. 1981) and will not

be discussed here except to relate observed chemical characteristics

with the phenomenon of internal loading. Monitoring within Lake Apopka

focused on three open-water stations (A2, A5, and A8) and Gourd Neck

Springs for the bulk of the study after a comprehensive survey of

11 stations during the first year showed that adequate spatial

characterization was provided by three sampling points. Sampling was

conducted monthly except for the first year when the frequency was


To evaluate the degree of interaction between SRP concentrations

and other chemical components, a correlation matrix was developed using

SAS (Statistical Analysis System) (Barr et al. 1976). Correlation

analysis compares parameters by way of a linear model to determine if a

significant relationship exists. The degree of linear correlation

between the two variables is given by the correlation coefficient, r,

which by definition must lie between 1 and -1. Selected results of the

analysis are presented in Table IV-6. Using the 5 percent confidence

level as the threshold of significance, significant relationships

between SRP and other abiotic variables were established for only two

parameters: turbidity and inorganic carbon (IC). Although a simple

correlation between two variables is not conclusive evidence of a cause

and effect relationship, this result does suggest that SRP levels within

Lake Apopka are influenced by the resuspension of sediments.

Indirect evidence to substantiate the role of sediment resuspension

of phosphorus dynamics is offered by evaluating the interaction between

turbidity and ammonia variability. Correlation analysis (Table IV-6)

indicates these two parameters are significantly related. Ammonia,

which is generated by heterotrophic bacteria as organic matter is

Table IV-6.

Intercomponent Relationships Between Selected Variables in
Lake Apopka, 1977 to 1980.

Turbidity IC K Ammonia Color

SRP 0.28915 0.38099 0.30150 0.12823 -0.14958
(0.0487) (0.0082) (0.0621) (0.3904) (0.3156)

Turbidity 0.65573 0.72893 0.32722 0.01153
(0.0001) (0.0001) (0.0248) (0.9387)

IC 0.90591 0.10961 -0.20465
(0.0001) (0.4633) (0.1676)

K -0.00954 -0.10756
(0.9541) (0.5145)

Ammonia -0.03957

Note: ( ) = significance of relationship.

decomposed, accretes in sedimentary pore fluids to quite significant

levels. Concentrations as high as 49.0 mg N/L have been observed in

Lake Apopka sediments (Brezonik et al. 1978). Average interstitial

concentrations exceed typical levels in the overlying water of Lake

Apopka by over two orders of magnitude; thus even minor disturbances of

the sediment-water interface can exert a profound influence on the

nutrient status of the water column. Seasonal dynamics of nitrogen in

Lake Apopka are stochastic in nature, apparently as a result of episodic

disturbances and scouring of the bottom sediments.

SRP was strongly correlated (p = 0.0082) with inorganic carbon and

to a lesser extent (p = 0.0621) with potassium. Similar to ammonia,

inorganic carbon accumulates in the interstitial fluid as detrital

material is metabolized. Likewise, the close relationship between

turbidity and inorganic carbon (r = .656, p = 0.0001) (Table IV-6)

supports a resuspension type of mechanism for controlling to a signifi-

cant extent the concentration of this solute in the overlying water.

Potassium, which in turn is very closely related to inorganic carbon

(r = .729, p = 0.0001), is essential for aerobic life forms (Lehninger

1970) and has been used as a tracer for the migration through soils of

leachate derived from decomposing vegetative material (Ellis 1980).

Wetzel (1975) has observed that potassium levels can be depleted in the

trophogenic zone of highly productive lakes presumably due to biotic

uptake and in turn accumulate in the tropholytic zone. It consequently

seems reasonable to label K as an indicator of internal nutrient cycling

and, in particular, nutrient release induced by sediment resuspension.

If the internal dynamics of phosphorus cycling in Lake Apopka are

dominated by external loading, a simple correlation between external

loading (or a corollary parameter) and SRP can be anticipated. A

nutrient budget developed for the 1977 water year has identified main-

tenance backpumping of the muck farms bordering the northern perimeter

of Lake Apopka as the principal external source of phosphorus. Because

of the high organic content of the muck soils, the nutrient-enriched

soil water is highly colored. This characteristic provides a convenient

marker with which to trace the discharge of agricultural wastes and

associated nutrients to the system. The following example illustrates

this point. During late summer 1979, the Apopka drainage basin

experienced heavy rainfall and high rates of backpumping to the lake and

the Apopka-Beauclair canal were observed. The effect of agricultural

discharge on Lake Apopka and the downstream Oklawaha lakes was striking,

as illustrated in Figure IV-8 which delineates the spatial distribution

of both color and SRP during September 1979. Concentrations in both SRP

and color show an increase with location moving towards the north end of

the lake. Once within the Apopka-Beauclair canal, SRP and color

concentrations increased dramatically. Color and SRP concentrations at

the outfall of a discharge pipe on the Apopka-Beauclair canal were

101 chloroplatinate units (CPU) and 0.523 mg P/L, respectively; this is

contrasted with substantially lower color (65 CPU) and SRP (0.026 mg P/L)

concentrations at the outlet of the lake. Figure IV-8 shows that SRP

concentrations and color appear to be intrinsically linked; both

parameters decrease monotonically with distance downstream from the

source until the Dora canal is reached. Similar observations were made

during August and October 1979.

- -

-- --- -

C '

o 0 %C

(nd3) H031 0

SRP (mg P/L)

O 0
O 0
o o
g d


C 3




zo 0
0 0
c -





f ) o'

a< e

If color is a reasonable label to trace long-term influences of

external loading on SRP dynamics, correlation analysis for the 4-year

data base implies that the relationship is insignificant. A weak,

negative correlation (r = -.14958, p = 0.3156) was derived between color

and SRP, suggesting that SRP variability within Lake Apopka is

independent of external inputs.

On July 11, 1979, a brief but intense convective disturbance was

encountered on Lake Apopka at the beginning of a routine sampling

effort. The storm, which originated from the southwest, began at

approximately 1540 hours. Wind velocities were estimated to range

between 9 and 11 m/s at the advent of the storm and had diminished to

approximately 6.7 m/s by the time the last sample was taken 1 hour

later. As the storm swept across the lake surface, bottom sediments

were immediately resuspended to the surface. The visual change was

rather remarkable; for example, sediment resuspension was observed as a

sharp front normal to and moving across the lake in the same direction

as the wind. Behind the front the water surface assumed a muddy brown

appearance from entrained sediments while on the leeward side, a

characteristic greenish cast typical of the highly productive water

column was observed.

Turbidity and suspended solids measurements from a protected point

along the western edge of the lake near Gourd Neck Springs showed

concentrations of 31 NTU and 0.0865 g/L of suspended material at the

surface. These values were considered to be representative of back-

ground or pre-storm conditions. Surface samples from the open-water

stations (A2, A5, and A8) readily showed the effects of resuspension;

turbidities ranged from 43 to 46 NTU and suspended sediment

concentrations ranged from 0.2053 to 0.2553 g/L.

Composite samples were taken from the surface and 1-m depth at the

open-water points and analyzed for turbidity, SRP, and TP. Results are

presented in Figure IV-9. Turbidity levels at all three stations were

substantially higher in the composite samples relative to surface

specific samples, thus indicating that resuspension was not uniform

throughout the water column but existed in the form of a concentration

gradient. Lakewide average turbidity and SRP concentrations both

exceeded levels observed the previous month during quiescent conditions

by a factor of three (Pollman et al. 1980). SRP levels were particu-

larly high at Station A5 where a concentration 0.277 mg P/L was

observed. Although the composite turbidity concentration was lower at

this station (90 NTU) relative to Stations A2 and A8 (100 and 120 NTU,

respectively), sampling was conducted at this station during the waning

moments of the disturbance. Wind velocity and wave activity were much

reduced at Station A5, allowing entrained particles to resettle.

Kinetic factors may also account for the lower SRP concentrations

observed at Stations A2 and A8.

Although suspended sediment concentrations are unavailable for the

composite samples, an empirical relationship derived in situ between

turbidity and suspended solids for Lake Apopka (Figure IV-10) enables an

estimate of the quantity of sediment resuspended to be made. The

following linear model was derived from least squares regression

analysis (r2 = .96, p <0.025):

Figure IV-9.

0.326 PA

0.277 T
0.206 0.7645




Suspended Sediment (g/L) and Total Phosphorus (TP)
(mg P/L) Concentrations in Lake Apopka During a Convective
Disturbance, July 11, 1979. TP and Suspended Sediment
Concentrations are Presented as the Upper and Lower
Figures, Respectively, at Each Sampling Site.




w a
o o

0 0


SZ o.

0^ 0

( N N ) A l(JI
^^^. 3

Turbidity = 12.04 + 141.2 (SS) (IV-5)

where SS is the concentration of suspended solids (g/L), and turbidity

is given in NTU. On the basis of this model, the sediment concentration

yielding the observed composite turbidity levels can be determined.

Using data from Station A5 as an example, a turbidity concentration of

90 NTU equates to a suspended solids concentration of 0.5521 g/L. The

suspended sediment concentration at 1-m depth must therefore be equal to

0.8973 g/L. Assuming the concentration of suspended matter increases

linearly with depth gives an average concentration for the water column

(z = 176 cm) of 0.8145 g/L.

Correcting for the background contribution from ambient phyto-

plankton and suspended sediment concentrations (0.0696 g/L) yields a

net resuspension of 0.7449 g/L. The areal loss across the sediment

water interface is therefore 1,311 g/m2. Since the average bulk

density and water content of Lake Apopka sediments is 1.024 g/cm3

and 96.04 percent, respectively, the depth of the sediment entrainment

is estimated to be 3.2 cm. Calculated depths of sediment resuspension

for Stations A2 and A8 are summarized in Table IV-7.

Table IV-7.

Estimated Depth of Sediment Resuspension of Lake Apopka During a
Convective Disturbance, July 11, 1979.

Calculated Composite Surface Suspended Calculated
Composite Suspended Sediment Sediment Depth of
Depth Turbidity Concentration Concentration Resuspension
Station (cm) (NTU) (g/L) (g/L) (cm)

A2 180 110 0.6937 0.1616 4.7

A5 176 90 0.5521 0.2069 3.2

A8 150 120 0.7645 0.2553 3.5


A series of adsorption-desorption isotherm studies were conducted

in the laboratory to determine net release or uptake of phosphorus as a

function of ambient levels of SRP in the water under well-mixed

(equilibrium) conditions. Experimental results indicate that Lake

Okeechobee sediments vary appreciably in their ability to adsorb (or

desorb) phosphorus. A characteristic curvilinear relationship between

initial concentration and the magnitude of sorption was observed for all

stations except Station 013, which showed little ability to desorb or

absorb phosphorus over the experimental range of concentrations

(Figures V-1 through V-4). Similar curvilinear results were reported by

Harter and Foster (1976) and Mayer and Gloss (1980). A station-by-

station comparison of adsorption at an initial concentration of

1,000 ug/L suggests that Station 07 sediments have greatest affinity for

phosphorus at high concentrations; these sediments also have the

greatest silt-clay content, and thus the greatest surface area available

for exchange. Because of their relative homogeneity with respect to

chemical and physical characteristics, sorption studies with Lake Apopka

sediments were limited to sediment from Station A5. A curvilinear

relationship between the extent of sorption and initial concentration

was also observed for Station A5 sediment; the results, however,

m 0


C) --
\ c0

\) a
E o
\ U



\ o


O 0 00 0

(6/d Br-r) NOIldIOS