tmdl and plrg modeling of the lower st. johns river...

$: TMDL and PLRG Modeling of the Lower St. Johns River ...waterinstitute.ufl.edu/WorkingGroups/TCAA/downloads/7-ICM-ExtLd... · Distinction of Labile and Refractory Organic Nutrients$
TMDL and PLRG Modeling of the

Lower St. Johns River

Technical Report Series Volume 1:

Calculation of the External Load

By:

John Hendrickson, Environmental Scientist

St. Johns River Water Management District

Nadine Trahan, GIS Analyst

Jones, Edmunds and Assoc.

Emily Stecker, Environmental Scientist

BCI Engineers and Scientists

Ying Ouyang, Ph.D., Environmental Scientist

St. Johns River Water Management District

May 2002

2

Table of Contents

INTRODUCTION AND APPROACH ........................................................................................................ 7

Project Area Description ........................................................................................................................... 8

Justification of the Conceptual Approach to the LSJR External Load .................................................... 10

Distinction of Labile and Refractory Organic Nutrients and Carbon ...................................................... 12

Organic Matter Composition Effects on Biodegradability And Nutrient Bioavailability ....................... 13

Objectives ................................................................................................................................................ 16

METHODS ................................................................................................................................................. 17

Separation of Labile and Refractory Organic Carbon, Nitrogen and Phosphorus .................................. 17

Overview ............................................................................................................................................. 17

Calculation of the External Load For the Lower St. Johns River ............................................................ 27

Point Source Load Estimation ............................................................................................................. 28

Non-Point Source Load Estimation ..................................................................................................... 29

Determination of the Upstream Load to the LSJR .............................................................................. 46

Determination of the Atmospheric Deposition Load ........................................................................... 53

RESULTS ................................................................................................................................................... 56

Tributary Organic Carbon and Nutrients ................................................................................................. 56

Calibration Data Set Summary ............................................................................................................ 56

Watershed Model Calibration .............................................................................................................. 58

Point Source Organic Carbon and Nutrients ........................................................................................... 70

Upstream Concentrations of Organic Carbon and Nutrients ................................................................... 72

Reconstruction Of The Upstream Natural Background Load ................................................................. 81

Natural Background Concentrations of Small Order, Undeveloped Streams ...................................... 81

Role of Spring Inputs ........................................................................................................................... 85

Historic Data for the St. Johns River ................................................................................................... 89

Total LSJR Load Estimates ..................................................................................................................... 97

Atmospheric Deposition ........................................................................................................................ 105

DISCUSSION ........................................................................................................................................... 106

Literature Cited ......................................................................................................................................... 113

3

FIGURES

1. The Lower St. Johns River Basin ....................................................................................9

2. Comparison of Inorganic and Non-Inorganic Nutrient Fractions for Black Creek and

the Lower St. Johns River at Racy Point .......................................................................13

3. Rates of Exertion of BOD for Organic Substrates Typical of Northeast Florida Surface

Waters ............................................................................................................................19

4. Tributary Water Quality Sampling Stations for Watershed Modeling Set-up and Skill

Assessment ....................................................................................................................20

5. Organic Carbon:Nitrogen Ratio as a Function of the Percent Labile Organic Carbon

.......................................................................................................................................25

6. Organic Carbon:Phosphorus Ratio as a Function of the Percent Labile Organic

Carbon ...........................................................................................................................26

7. Relationship Between PLSM Predicted Runoff:Observed Runoff Ratio and Measured

Seasonal Whole Watershed Runoff Coefficient ............................................................34

8. Relative Position of the 1995-1999 Time Interval in the Historic Long Term Flow

Record ...........................................................................................................................35

9. Development of Hydrologic Correction Factors for the PLSM Runoff Coefficient .....38

10. Comparison of Original, Seasonal-fixed and Long-Term Rain Ratio Adjusted Runoff

Coefficients ...................................................................................................................39

11. Comparison of Original PLSM-Predicted, Long-Term Rain Ratio Adjusted and

Observed Cumulative Discharge Curves for LSJRB Calibration Watersheds, 1995-99

.......................................................................................................................................40

12. Monthly Mean Concentrations and 95% Confidence Intervals for Color and Total

Organic Carbon for 24 Unimpacted Blackwater Streams in Northeast Florida ............43

13. Watershed Model Input Areas for Nutrient Load Compilation ....................................46

14. Tributaries Forming the Lower St. Johns River ............................................................48

15. Flow chart for differentiation of laboratory analytical fractions into CE-QUAL-ICM

state variables for the lower St. Johns River upstream boundary at Dunns Creek and

Buffalo Bluff .................................................................................................................50

16. Comparison of Corrected Chlorophyll a and Algal Biovolume for Combined LSJR

Freshwater Water Quality and Plankton Analysis, 1995 – 2001 ..................................52

4

17. Comparison of POC:Algal Biovolume (a) and POC:Total Chlorophyll a Ratios to

Total Biovolume and Total Chlorophyll a Concentration for LSJR Freshwater Samples

.......................................................................................................................................54

18. Relationship Between Refractory Dissolved Organic Carbon and Color for Blackwater

Streams of the LSJR Basin ............................................................................................55

19. Comparison of Observed to Simulated Flow-Weighted Concentrations of Carbon,

Nitrogen and Phosphorus Forms for the December through March Season .................61


Nitrogen and Phosphorus Forms for the April through July Season .............................63


Nitrogen and Phosphorus Forms for the August through November Season ...............65

22. Partitioned Nitrogen Concentrations at (a) Buffalo Bluff and (b) Dunns Creek, Dec.

1994 - Nov. 1999 ...........................................................................................................74

23. Partitioned Phosphorus Concentrations at (a) Buffalo Bluff and (b) Dunns Creek, 1995

– 1999 ............................................................................................................................75

24. Partitioned Organic Carbon Concentrations at Buffalo Bluff and Dunns Creek, 1995 –

1999 ...............................................................................................................................76

25. Loads of Nitrogen Forms Entering the Lower St. Johns River at Buffalo Bluff and

Dunns Creek, 1995-99 ...................................................................................................79

26. Loads of Phosphorus Forms Entering the Lower St. Johns River at Buffalo Bluff and

Dunns Creek, 1995-99 ...................................................................................................80

27. Continuous Probability Density Functions for Total and Inorganic Nutrient Mean

Concentrations for Streams in Northeast Florida ..........................................................84

28. Time-Series Concentrations of Nitrate+Nitrite-N and Orthophosphate-P in Major

Springs Discharging to the St. Johns River That Exhibit Nitrate+Nitrite Trends .........86

29. Comparison of Present Day and Predicted Natural Background Concentrations of

Total Nitrogen and Total Phosphorus in the Lower St. Johns at Buffalo Bluff, 1995-99

.......................................................................................................................................88

30. Population growth within the 14 Counties of the St. Johns River Basin, 1890 – 2000

.......................................................................................................................................90

31. Comparison of Monthly Mean Water Quality Parameters for 1995-99 (solid boxes) to

the Data Collected by Pierce (1947) in 1939-40 (open diamonds) for the St. Johns

River near Buffalo Bluff................................................................................................95

5

32. Comparison of Total and Bioavailable Nitrogen Forms in Runoff from Natural

Forested and Mixed Urban/Commercial/Residential Watersheds ..............................108

6

TABLES

1. Tributary Water Quality Station Locations Employed in Determination of Labile and

Refractory Organic Nutrients ........................................................................................22

2. Point Source Facilities Included in the Calculation of the Lower St. Johns River

External Load ................................................................................................................23

3. Seasonal Runoff Coefficients for Application of the Pollution Load Screening Model

to the LSJR Basin ..........................................................................................................33

4. Seasonal Water Quality Coefficients Used in the PLSM to Predict Non-Point Source

Loads to the LSJR .........................................................................................................41

5. Total Organic, Labile Total Organic and Refractory Total Organic Carbon Land Use

Category Concentration Coefficients ............................................................................45

6. Mean Total, Inorganic, and Calculated Labile and Refractory Organic Nutrient and

Carbon Mean Annual Flow-Weighted Concentrations for Tributaries sampled within

the lower St. Johns River Basin .....................................................................................57

7. Pearson Correlations, Slopes, and Confidence Intervals of the Slopes for Intercept-Fit

Regressions Between Calibration Station Measured Flow-Weighted Concentrations

and Contributing Area Modeled Runoff-Weighted Concentrations .............................68

8. Summary of Point Source Mean Effluent Water Quality Concentrations ....................71

9. Total Phosphorus Concentrations Determined for Selected Locations in St. Johns

River Basin in 1952 .......................................................................................................92

10. Summary of Mean Annual Loads to the Lower St. Johns River, 1995 .........................98

11. Summary of Mean Annual Loads to the Lower St. Johns River, 1996 .........................99

12. Summary of Mean Annual Loads to the Lower St. Johns River, 1997 .......................100



15. Summary of Overall Mean Annual Loads to the Lower St. Johns River, 1995-99 ....103

7

INTRODUCTION AND APPROACH

Accelerated eutrophication arising from nutrient enrichment of estuaries represents one of the

most significant water quality problems faced by near coastal waters worldwide (National

Research Council, 2000). Within the United States, part of this problem rests in the standards-

based approach to water quality control, in which the potential harm incurred by sources is

evaluated based upon effluent and near-field concentrations of pollutants. In this approach,

cumulative loads of substances, in particular nutrients, have been overlooked, with the result that

receiving water assimilative capacities have been overwhelmed. This situation has increasingly

lead water managers to resort to the TMDL process (CWA Section 303(d)) as a means of

eutrophication control. To address problems associated with accelerated eutrophication in the

lower St. Johns River estuary (LSJR), both the Florida Department of Environmental Protection

(FDEP) and the St. Johns River Water Management District (SJRWMD) are jointly executing a

strategy for nutrient pollution control that fulfills their respective responsibilities for the

establishment of TMDLs for impaired water bodies and stormwater PLRGs (F.A.C. Chapter 62-

40).

A generally accepted approach has evolved for addressing estuarine eutrophication in which the

sources, magnitude and timing of the external nutrient load are linked to the effects of the

receiving water body. Because of the temporal and spatial disconnect between the entry of

nutrient loads and the manifestation of eutrophication effects, and the need to predict the levels

of expected improvement with various nutrient reduction strategies, dynamic water quality

process models have become invaluable tools in estuarine nutrient management efforts. Such

models “process” the external load in a time-sequential fashion in a 2 or 3-dimensional

discretized grid that approximates the morphology of the water body. In the context of

eutrophication, relevant “processes” are the biological processes photosynthesis and algal carbon

fixation, community respiration (as both a loss of organic carbon and exertion of oxygen

demand), and organic nutrient re-mineralization, as well as physical processes such as oxygen

reaeration, substance advection, molecular dispersion, solar light absorption, and sedimentation.

8

In this modeling approach to the establishment of nutrient pollution reductions, two large

investigative efforts must be undertaken: one to quantify the timing, magnitude, and spatial

nature of the incoming nutrient load, referred to as the “external load”, and another to determine

the effect of this load on the receiving water body. This report describes the first element of this

intricate undertaking for the Lower St. Johns River, that of the derivation of the external load.

Project Area Description

The St. Johns River is one of the largest blackwater rivers of the southeast U.S. The river is

located in northeast Florida and drains about 1/5th

of the state, encompassing a 9,562 square mile

drainage area. The river is slow moving, with a slope of only 1.4 in/mi (Toth, 1993), and is

essentially at sea level for its final 125 mi. The lower St. Johns is the estuarine portion of the

river, formed at the confluence of the middle St. Johns and the Ocklawaha River, and

encompassing a 2,750 square mile area (Figure 1). Within this reach, the St. Johns River is

slightly more that 100 miles long and has a water surface area, including tributary mouths below

head of tide, of 85,000 acres. The lower St. Johns can be differentiated into three riverine

salinity and limnologic zones: a fresh tidal lacustrine zone which extends from the city of Palatka

north to approximately the mouth of Black Creek; a predominantly oligohaline, lacustrine zone

extending from the mouth of Black Creek northward to the Fuller Warren Bridge (I-95) in

Jacksonville; and a mesohaline/polyhaline, riverine zone downstream to the mouth. The slow

moving, lacustrine nature of the river facilitates phytoplankton primary production, and spring

and summer algal blooms in this nutrient-rich river often exhibit chlorophyll a concentrations

exceeding 100 g/L.

The southern portion of the lower basin is largely rural, with predominant land uses in forestry

and row crop agriculture. The northern portion of the basin is distinguished by the heavily

urbanized cities of Jacksonville, Orange Park and Middleburg. Roughly three quarters (64 to 82

percent) of the basin’s highly developed land uses (medium and high residential, high intensity

commercial and industrial) drain to the oligohaline and mesohaline lower St. Johns. In contrast,

62 to 98 percent of the basin’s agricultural land uses drain to the fresh tidal reach.

9

Figure 1. The Lower St. Johns River Basin.

10

The existence of poor water quality in the LSJR has been identified in a number of reports dating

back to at least 1947 (Florida State Board of Health 1947). Because of these problems, the

establishment of TMDLs and PLRGs for the lower St. Johns River are a high priority, and an

aggressive schedule has been established that seeks the identification of river assimilative

capacity and general allocation to major sources by the end of 2002.

Comprehensive external nutrient load assessments have been performed twice previously for the

LSJR. In 1976, the firm of Atlantis Scientific (Atlantis Scientific, 1976), under authorization of

the 1972 Clean Water Act Section 315, undertook a computation of the external load and

concluded that point source comprised the majority of this load. Hendrickson and Konwinski

(1998) also computed the external load to the river for 1993-94, and concluded that nitrogen and

phosphorus were 2.5 and 6 times greater than natural background, with augmented nutrient loads

(that load above natural background) approximately evenly split between point and nonpoint

sources.

Justification of the Conceptual Approach to the LSJR External Load

By virtue of its long term presence in the St. Johns River and its frequent project partnerships

with the SJRWMD, the U.S. Army Corps Jacksonville District has brought valuable assistance to

the river TMDL and PLRG development. This partnership has provided the assistance of the

U.S. Army Engineer Research and Development Center (ERDC) at Vicksburg, MS, to assist in

the examination of the nature of the interaction between river processes and the external load.

To quantify this interaction, the ERDC has adapted its water quality model, CE-QUAL-ICM

(Corps of Engineers Water Quality Integrated Compartment Model), to the LSJR. CE-QUAL-

ICM (hereafter referred to as just ICM) was developed to study eutrophication processes in

Chesapeake Bay (Cerco and Cole, 1994), however, as ICM simulates the fundamental processes

related to algal (and plant) growth, death and decomposition, its robust model formulation is

applicable to a wide range of water bodies and even wetlands. ICM differs from another widely

used water quality model, WASP, in that it predicts eutrophication effects – transparency loss,

dissolved oxygen sags, and sedimentation - through the use of a carbon budget, rather that

11

relying on the input and internal formation of biochemical oxygen demand and chlorophyll a.

Because ICM allows for the distinction of carbon and nutrients compartmentalized within labile

and refractory forms, it is in theory particularly useful for applications in blackwater river

estuaries.

Along with the mixture of inorganic and nutrient-bearing organic substrates (such as animal and

human waste, industrial process effluents, and algae or algal detritus) that are the focus of

anthropogenic nutrient enrichment, blackwater rivers and streams also exhibit significant nutrient

content of natural origin. Large portions of this natural nutrient load, as much as 40 percent of

the phosphorus, and over 90 percent of nitrogen, are contained within the organic fraction.

Strong relationships between total organic carbon and color suggest that the majority of this

natural, organic nutrient load is contained within colored, dissolved organic matter (CDOM) of

terrestrial and riparian vascular plant origin. Although natural CDOM is generally believed to be

resistant to microbial decomposition and largely unavailable for utilization by phytoplankton in

typical estuarine residence times, these heterogeneous, humic substances contain a substantial

amount of nitrogen (N) and phosphorus (P) in their structures (DeBusk et al, 2001), and hence

the sheer volume of the material with respect to other OM pools dictates its relevance be

considered.

With the capabilities of ICM come fairly rigorous requirements on the detail of the external

nutrient and carbon load that must be input for model simulations. The most difficult of these

determinations is the separation of the external organic nutrient and carbon load into labile

(easily decomposed and utilized) and refractory (slowly decomposed) components based upon

readily available water quality monitoring data. This technique for separation needs to extend

also to the river water quality monitoring calibration data set for ICM. In order to predict the

changes in the external load with various nutrient reduction strategies, it is not sufficient to only

characterize the incoming labile and refractory carbon and nutrient load; the relationship between

land development and organic carbon and nutrient bioavailability must also be described. An

additional relationship must be addressed between the concentration of colored dissolved organic

matter (CDOM) and water column transparency in order for the appropriate functioning of the

ICM light attenuation algorithm in the algal photosynthesis calculation.

12

Distinction of Labile and Refractory Organic Nutrients and Carbon

It is generally understood that dissolved, inorganic forms of nutrients (NO2+3, NH4, and PO4), as

well as some low molecular weight organic compounds such as urea, are immediately available

for algal growth, while organic nutrient forms, which must first undergo desorption (if

particulate bound), hydrolysis, bacterial decomposition or photo-decomposition (Bushaw et al.

1996) for inorganic nutrient regeneration and subsequent utilization by phytoplankton, are less

readily available. Organic nutrient bio-availability for aquatic primary production is dependent

upon the utilization preference of the parent organic substrate by general microbial heterotrophs

(DeBusk et al., 2001), which must first decompose this substrate in order to liberate mineral

nutrient forms. With regard to organic carbon and nutrient bioavailability, a general working

hypothesis has evolved that partitions organic carbon and nutrients into two pools: a labile pool,

that can be utilized in time frames relevant to water quality processes of interest in the receiving

water, and a refractory pool, that is decomposed very slowly and essentially inert for relevant

time frames (Wetzel, 1983). The bioavailablility of the organic nutrient pool represents an

important issue in the assessment of nutrient enrichment in blackwater rivers of the southeast

U.S. coastal plain, as much of the total phosphorus (TP) and most of the total nitrogen (TN)

enters the river as an organic or non-inorganic form (Figure 2).

Relatively little attention has been paid to differences in organic nutrient bioavailability in

assessments of external loads to eutrophic water bodies (Stepanauskas et al., 1998). This may be

due to the predominance of inorganic nutrients in river flow to intensely studied temperate

estuaries, leading most authors to not further differentiate the organic nutrient pool (Magnien et

al. 1992; Goolsby et al. 2001) or even to distinguish it from the inorganic nutrient-dominated

total nutrient pool (Boynton et al. 1995; Jaworski et al. 1992; Valiela et al. 1992). This lack of

differentiation extends also to land use-loading rates applied in watershed load indexing models

(Hartigan et al., 1982; Adamus and Bergman, 1995; Harper, 1994; EPA, 1984), to the commonly

used, process-based watershed models such as HSPF, to agronomic field scale models such as

GLEAMS, and to water quality process models such as WASP.

13

Figure 2. Comparison of Inorganic and Non-Inorganic Nutrient Fractions for Black Creek and

the Lower St. Johns River at Racy Point.

Organic Matter Composition Effects on Biodegradability And Nutrient Bioavailability

Research in aquatic microbiology has elucidated several patterns pertaining to organic matter

utilization by general aerobic microbial heterotrophs, and concomitantly organic nutrient

bioavailability. Two generally accepted precepts form the basis of our understanding of the

biodegradability of biogenic organic compounds, 1) recently produced, undecomposed OM is

more biodegradable than material that has undergone diagenetic alteration through repeated

decomposition cycles, and 2) OM produced by non-structural aquatic plants and algae is more

biodegradable than that produced by terrestrial, lingo-cellulosic vascular plants. Carbohydrates,

(a) South Fork Black Creek

0

0.5

1

1.5

2

1 2 3 4 5 6 7 8 9 10 11 12

Month

Co

nc

en

tra

tio

n, m

g/L

Total Inorganic N

Total Organic N

(b) St. Johns River - Racy Point

0

0.5

1

1.5

2

1 2 3 4 5 6 7 8 9 10 11 12

Month

Co

nc

en

tra

tio

n, m

g/L

Total Inorganic N

Total Organic N

( c ) South Fork Black Creek

0

0.05

0.1

0.15

0.2

1 2 3 4 5 6 7 8 9 10 11 12

Month

Co

nc

en

tra

tio

n, m

g/L

Orthophosphate

Non-PO4 P

(d) St. Johns River - Racy Point

0

0.05

0.1

0.15

0.2

1 2 3 4 5 6 7 8 9 10 11 12

Month

Co

ncen

trati

on

, m

g/L Orthophosphate

Non-PO4 P

14

proteins, lipids, nucleic acids and pigments are decomposed in relatively short time frames, while

humic substances are less readily decomposed and in some cases essentially inert (Wetzel, 1983;

Moran and Hodson, 1990; although this assertion is contradicted in the work of Volk et al., 1997,

who find similar utilization of humic substances). The biodegradability of natural OM that

occurs in aquatic systems and its bioavailability of incorporated C, N and P can be viewed as

dependent largely upon two factors: 1) whether the material is allochthonous or autochthonous in

origin, and 2) whether or not the OM has undergone some degree of decomposition and

diagenetic alteration prior to its entry to surface waters. Thus it generally holds that

autochthonous OM is more labile than allochthonous OM, and that the humic fraction of DOM is

less bioavailable on a mole carbon than non-humic DOM (Kaplan and Newbold 1995; Moran

and Hodson, 1990; Moran et al., 1999). In their work on piedmont and blackwater river OM in

the southeast U.S., Sun et al. (1997) demonstrate that the compositional changes that accompany

diagenetic condensation relate directly to bioavailability, with blackwater stream OM appearing

the most refractory per mole carbon. This assertion is in congruity with work that has shown

some forms of soil humus in the allochthonous organic carbon pool to be decades to hundreds of

years old (Raymond and Bauer, 2001).

Surprisingly, this difference in OM bioavailability runs contrary to the “smaller is better”

nutrient utilization paradigm, in that particulate organic nutrients in surface waters, in the form of

algal cells or relatively undecomposed plant detritus, are generally more readily available than

dissolved forms. Within the dissolved organic matter pool (<0.45 m diameter), high molecular

weight organic compounds (> 10,000 nMW) also have been to found to be more bioavailable

(Tranvik, 1990; Amon and Benner 1996; Gardner et al., 1996; Mannino and Harvey 2000) than

low molecular weight DOM (< 1000 nMW). In the Amazon River, Hedges et al. (1994)

considered DOM to be the most profoundly degraded material, hence the most refractory,

mobilized through a process of “selective solubilization”.

Also fundamental to the bioavailability of organic nutrients for primary production is whether or

not OM decomposition will result in nutrient regeneration (e.g., an increase in water column

inorganic nutrients) or nutrient immobilization to meet bacterial growth needs. Goldman et al.

(1987) postulated that if the substrate C:N and C:P ratios are sufficiently low such that, when

15

corrected for carbon gross growth efficiency (the fraction of the total carbon decomposed that is

retained as bacterial biomass), N and P remain in excess of bacterial growth needs, then these

nutrients will be regenerated in the inorganic form and be potentially available for incorporation

by phytoplankton. Because labile OM is high in proteins, amino acids and cellular metabolic

organic compounds that exhibit relatively low C:N and C:P ratios, decomposition of labile

substrates in the aquatic environment tends to lead to the regeneration of N and P. Conversely,

substrates with a high C:N, such as aquatic humic OM (averaging 50:1 molar; Thurman, 1985),

will tend to immobilize inorganic N and P (Mann, 1988; Strauss and Lamberti, 2000). Not

surprisingly, organic substrates with high C:N ratios that typically exist in the aquatic

environment exhibit low biological availability, and concomitantly a low likelihood that bacterial

decomposition of this substrate can regenerate mineral nutrients for autotrophs (Bushaw et al.,

1996). Because of these general differences, C:N and C:P ratios have been a commonly used

proxy for bioavailability of OM.

No clear definition exists on what constitutes labile verses refractory, and whether or not the

range between the two extremes exists as a continuum or as discrete states. Labile substrates

have been described as those utilized within timeframes of one to two weeks (Sondergaard and

Middelboe, 1995); as utilization through the exponential growth phase to the stationary phase

(approximately 2 days; Stepanauskas et al., 1999; approximately 4 days for DON of the

Delaware River (Seitzinger and Sanders 1997)); or in-situ bioreactor residence time (4 to 18

hours; Volk et al., 1997). The first order decay coefficient of 0.075 day-1

used by ICM (Cerco

and Cole, 1995) yields a duration of 9.2 days for 50% utilization of the original labile substrate,

and 30 days for 90% utilization. Moran and Hodson (1989), in their investigation of fresh and

salt marsh plant ligno-cellulose, observed what appeared to be distinct rates of utilization,

suggesting distinct, uniform chemical classes driving separate utilization rates. Similarly, Ogura

(1975) determined that two distinct pools of dissolved organic compounds existed in most

aquatic systems.

In various examinations of surface waters exhibiting a range of human impact, the biodegradable

percent of the total OC pool has been found to vary between 1 and 86 percent (Sun et al., 1997).

For most rivers dominated by allochthonous OC, the range is closer to between 7% to 25%

16

(Sondergaard and Middelboe, 1995; Volk et al., 1997), with blackwater rivers exhibiting the

lowest relative amounts of labile OC (Moran et al., 1999). Stepanauskas et al. (2000), in their

study of Scandinavian rivers, estimated the percent of labile dissolved organic nitrogen (DON) as

between 19 and 55%.

Objectives

The objectives of this report are

1) describe the approach to partitioning organic carbon and nutrients in the external load to

the LSJR, for the purpose of distinguishing the relative bioavailability of these forms and

hence the impact on eutrophication;

2) determine the relationship between land development factors and the relative

bioavailability of carbon and nutrient forms in runoff at a watershed scale, for the

purpose of modeling the external carbon and nutrient load, and predicting the changes in

this load with changes in land development patterns;

3) determine the load to the LSJR from all sources – upstream, within basin point and non-

point sources, and atmospheric sources – in order to assess the relative effects of these

sources on eutrophication; and

4) reconstruct the natural background load to the LSJR, for the purpose of putting present-

day loading rates in perspective, and for gauging the baseline level of productivity of the

LSJR in its pre-development state.

17

METHODS

Separation of Labile and Refractory Organic Carbon, Nitrogen and Phosphorus

Overview

To partition labile and refractory organic carbon and nutrients in the external load calculation, a

two-step empirical approach was developed. First, a conceptual model was developed relating

the rate of oxygen consumption during decomposition to total organic carbon to determine

overall decomposition rate, based upon partial decomposition rates for labile and refractory

organic material already established within ICM. This model was then applied to a water quality

data base of tributary sampling stations and point source effluents to partition labile and

refractory organic carbon. A multiple regression approach was employed to establish specific

land use, labile and refractory organic carbon runoff concentrations, and these specific land use

organic carbon concentrations are then applied to a watershed model to develop whole-basin

labile and refractory organic carbon loads. The second step partitioned organic N and P based

upon the relative amounts of labile and refractory organic carbon. Again employing the tributary

water quality monitoring and point source effluent data, C:N and C:P ratios were related to the

percent of labile organic carbon, and this relationship used to predict refractory and labile C:N

and C:P ratios. These ratios were then used to sub-divide the previously modeled organic

nitrogen and non-orthophosphate phosphorus (TP-PO4) loads into labile and refractory fractions.

Model for Organic Carbon Partitioning

Organic carbon found in surface waters is borne in a mix of organic matter from multiple

sources. This complex mix of organic substrates is assumed to be composed of fractions that are

readily available for microbial decomposition (labile) and relatively unavailable (refractory).

The degree of organic carbon lability in oxygenated surface waters should, in theory, be reflected

in carbonaceous biochemical oxygen demand (CBOD), with labile substrates consuming more

oxygen per mole of carbon in the test period (typically 5 days) than refractory substrates.

18

Consumption of organic carbon by bacterial heterotrophs has generally been found to adhere to

first-order exponential decay. Chapra (1977) provides this relationship in the following form, in

which the maximum amount of CBOD that can be exerted on a substrate, CBODultimate, is related

to that amount consumed in time t, by the relationship:

Ct = Co(1-e-kt

)

where Ct is the oxygen (or carbon) consumed at time t, Co is the BODultimate, and k is the

substrate-specific decomposition coefficient.

In practice, measurements of ultimate BOD are rarely performed, although total organic carbon,

a frequently measured constituent, should in theory be related to ultimate BOD. The molar rate

of O2 consumption per CO2 production has typically been set at 1:1 in computations of

community respiration (Wetzel and Likens, 1990; as per the respiratory quotient (RQ) of

Strickland, 1960). In the computations here, a RQ of 1 mole O2 consumed per 1 mole of OC

consumed, or mass ratio of 2.67:1, was used.

Figure 3 demonstrates the theoretical rate of change in BOD exerted over time for various

homogeneous categories of substrates common to surface waters of northeast Florida, and their

associated decay coefficients. Algal biomass and domestic waste organic matter appear to be

highly labile substrates, while pulp mill effluent and colored dissolved organic matter in runoff

of native, undeveloped blackwater streams appear to be relatively refractory. Decay coefficients

established in the CE-QUAL-ICM model of 0.075 day-1

for labile substrates and 0.001 day-1

for

refractory substrates appear representative of the range in decay coefficients for the aquatic

organic substrates in Figure 3. In comparison, decay rates determined by Moran et al. (1999) for

5 rivers of the southeast U.S., expressed as first order decay coefficients, ranged from 0.003 day-1

to 0.001 day-1

.

19

Figure 3. Rates of Exertion of BOD for Organic Substrates Typical of Northeast Florida

Surface Waters. Model of the form Ct = Cu(1-e-Kt), adapted from Chapra (1997). Cu

= ultimate BOD exertion, estimated from TOC from 2.67:1 mass ratio of O2

consumption to OC consumption and respiratory quotient =1. Algae: Determined

from mean BOD data from lake Dora, FL; phytoplankton organic carbon determined

form 50:1 carbon:chlorophyll a. Value should be considered the sum of algal

respiration and bacterial decomposition. Secondary WWTP effluent from a sampling

of 23 point sources of the lower St. Johns River basin. Pulp and paper determined

form a large mill in the lower St. Johns River basin. Native DOM developed from the

mean of undeveloped blackwater streams in northeast Florida.

Determination of Labile and Refractory Organic Carbon

To partition labile and refractory organic carbon, tributary runoff and point source effluent water

quality monitoring data collected between 1993 to 1999 within the lower St. Johns River basin

were compiled to create a data base of BOD, nutrients and organic carbon. Tributary station

descriptions and number of events sampled are included in Table 1, and the locations of these

tributaries and their contributing areas are shown in Figure 4. Point sources are listed in Table 2.

Stations were included in the analysis if the sample constituent suite included CBOD, total

0

10

20

30

40

50

60

70

80

90

100

0 10 20 30 40 50 60

Time, days

Pe

rce

nt

of

Ult

ima

te B

OD

Exe

rted

, o

r P

erc

en

t o

f

To

tal

Org

an

ic C

arb

on

Co

ns

um

ed Algae

K = 0.094 day-1

2ndary STP

K = 0.0386 day-1

Pulp & Paper Eff.

K = 0.0096 day-1

Native DOM

K = 0.0022 day-1

20

Figure 4. Tributary Water Quality Sampling Stations for Watershed Modeling Set-up and Skill

Assessment.

21

organic carbon, total phosphorus, orthophosphate, total ammonia and total nitrate+nitrite

nitrogen. In all, 789 samples were available for 28 surface water stations and 22 point sources.

Sample total organic carbon was considered to be the sum of carbon within labile substrates

(labile total organic carbon, or LTOC) and refractory substrates (refractory total organic carbon,

or RTOC), the proportions of which can be determined through the simultaneous expression of

their rates of decomposition, as indicated by oxygen consumption in the 5-day biochemical

oxygen demand (BOD5) test. Using the rates of decomposition of the first-order decay model of

0.075 day-1

for labile substrates, and 0.001 day-1

for refractory, a pair of equations for the

simultaneous solution of labile and refractory portions can be set up in the form:

(1) TOCt=5 = RTOC(1-e-(0.001)*5

) + LTOC(1-e-(0.075)*5

)

(2) TOCt= = RTOC(1-e-(0.001)* ) + LTOC(1-e

-(0.075)* )

where RTOC = refractory total organic carbon, and LTOC = labile TOC. In equation (1), the

moles of TOC decomposed at t=5 was assumed to be in unity (RQ = 1) with the moles of oxygen

consumed (CBOD5) and was converted to TOC consumed by dividing by 2.67. When all TOC is

consumed, at t = , (analogoug to ultimate BOD) the exponent term in parenthesis goes to zero,

and TOC = RTOC + LTOC. The above paired equations were simplified for computation

through the following steps:

(1) (CBOD5/2.67) = RTOC*(0.005) + LTOC*(0.3127)

(2) TOC = RTOC*(1) + LTOC*(1)

(1) 200*[(CBOD5/2.67) = RTOC*(0.005) + LTOC*(0.3127)]

(2) TOC = RTOC*(1) + LTOC*(1)

(1) CBOD5*74.906 - LTOC*(62.54) = RTOC

(2) TOC - LTOC = RTOC

22

Table 1. Tributary Water Quality Station Locations Employed in Determination of Labile and Refractory Organic Nutrients

Station

Abbreviation Station Description Latitude Longitude

River

Mile

Entry

Point Samp. No.

Urban,

Commercial,

Residential

Fraction

High

Intensity,

Livestock

Fraction

Row Crop,

Citrus, Low

Intensity

Fraction

Forested

Fraction

16MCRK 16 Mile Creek at Deep Crk Rd. W 293932.27 812741.76 11 0.0 0.0 84.9 15.1

ARLRM Arlington River Near Mouth Below Pottsburg Ck 301917.00 813558.00 20 6 57.8 0.1 10.9 27.7

BC218 BRADLEY CREEK @ 218 300035.28 814824.12 7 0.0 0.0 0.9 99.1

BC739 BRADLEY CREEK @ 739 300246.86 814705.10 20 5.6 0.0 25.2 69.2

BLC Black Creek at Hwy 209 300455.00 814835.00 44 47

BRDRM Broward River Near Mouth at Hecksher Drive 302500.00 813608.00 3 5

BSF South Fork of Black Creek at Hwy 218 300337.00 815218.00 44 44 3.5 1.1 9.8 84.8

CCR Clarkes Creek at US 17 295242.00 813950.00 56 5

CEDSJ Cedar River Above San Juan Blvd 301654.00 814426.00 26 19 67.5 0.2 12.8 17.7

DBR Dog Branch 50 meters downstr. County Rd. 207 A 294143.00 813450.00 67 29 4.4 0.0 81.5 13.8

DCH Deep Creek headwaters 294034.00 812800.00 19 0.0 0.0 73.4 26.0

DPB Deep Creek at Railroad Bridge 294345.00 812914.00 67 47 0.7 0.0 74.1 24.6

DUNCM Dunn Creek Near Mouth at Hecksher Drive 302516.00 813509.00 3 4

GC16 Governors Creek at Hwy 16 Near Green Cove 295902.00 814211.00 22 8.8 0.4 24.0 66.8

GC315 Green's Creck above County Rd. 315 295438.00 814740.00 3 0.0 0.9 3.6 95.5

GOV Governors Ck Near Mouth @ Seaboard Coast Rr Bridge 300014.00 814133.00 6

ML209 MILL LOG @ 209 300344.72 814525.66 15 0.6 3.6 28.1 67.2

MLRMC Mill Log Creek @ Russell Missionary 300305.00 814546.00 19 0.4 7.5 50.9 40.0

MOB Moccasin Branch On SR 13 294617.00 812850.00 30 1.3 0.3 39.5 58.7

NBC North Fork of Black Creek at SR 21 300432.00 815150.00 44 48 4.6 1.1 13.5 79.2

OHD Outlet of Hastings Drainage District 294249.00 813243.00 31 1.3 0.0 38.8 59.7

ORTCR Ortega River @ Collins Road 301203.00 814351.00 1

ORTTM Ortega River Above Timaquana Road 301451.00 814236.00 26 13 35.0 0.2 14.8 48.5

PCRHR Peters Creek at Rosemary Hill Rd 300025.00 814454.00 24 1.1 0.8 5.5 92.6

PTC Peters Creek at Hwy 209 300200.00 814329.00 44 82 1.5 4.7 13.3 79.9

SMC Sixmile Creek at SR 13 295732.00 813237.00 51 56 2.0 2.1 20.4 75.1

TRC Trout Creek at SR 13 295905.00 813358.00 51 12 1.6 0.1 9.1 88.5

TRTRM Trout River Near Mouth Below Main St Bridge 302337.00 813856.00 15 4

Total 629

23

Table 2. Point Source Facilities Included in the Calculation of the Lower St. Johns River External Load

Facility ID Facility Name Data

Freq.

Service Area

(Ac.)

Design

Capacity

(MGD)

1997-98

Mean Flow

(MGD)

Connect % Facility

Latitude

Facility

Longitude

Model Grid

#

IC JC

R-Seg.

FL0023493 MANDARIN WWTF Daily 30690 7.50 4.81 57.0 30.17903 -81.62241 1504 100 28 Oligohal

FL0026000 BUCKMAN STREET WWTF Daily 73143 52.50 33.06 79.0 30.35232 -81.62898 1121 55 24 Mesohal

FL0026441 ARLINGTON EAST WWTF Daily 63576 11.00 10.85 49.0 30.34665 -81.54316 1061 39 48 Mesohal

FL0026450 JAX DISTRICT II WWTF Daily 67105 10.00 4.32 89.0 30.42293 -81.61842 338 36 24 Mesohal

FL0026468 SOUTHWEST DISTRICT WWTF Daily 48194 10.00 5.86 43.0 30.23276 -81.72250 1422 90 20 Oligohal

FL0000400 STONE CONTAINER CORPORATION Monthly 20.00 8.85 N/A 30.41900 -81.60420 183 28 21 Mesohal

FL0000892 JEFFERSON SMURFIT CORPORATION Monthly 7.00 5.79 N/A 30.36670 -81.62500 1035 51 24 Mesohal

FL0002763 GEORGIA PACIFIC, PALATKA Monthly N/A 50.00 34.24 N/A 29.68247 -81.68278 2027 171 20 Fresh

FL0020231 JACKSONVILLE BEACH Monthly 3.07 874 31 81 Mesohal

FL0020427 NEPTUNE BEACH WWTF Monthly 1321 1.50 0.94 97.0 30.31558 -81.42007 874 31 81 Mesohal

FL0020915 GREEN COVE SPRINGS, CITY OF Monthly 4083 0.75 0.46 85.0 30.00724 -81.69646 1679 121 20 Fresh

FL0022489 WESLEY MANOR RETIRMNT VILL-JAX Monthly 0.1 0.05 30.11390 -81.60610 1573 110 31 Oligohal

FL0023248 BUCCANEER WWTF Monthly 1785 1.30 1.00 95.0 30.36976 -81.41157 874 31 81 Mesohal

FL0023604 MONTEREY WWTF Monthly 3684 3.60 3.02 55.0 30.33060 -81.60116 1158 59 27 Mesohal

FL0023621 HOLLY OAKS SUBDIVISION Monthly 3803 1.00 0.00 72.0 30.35752 -81.52208 1105 43 54 Mesohal

FL0023663 SAN JOSE SUBDIVISION Monthly 2225 2.25 2.09 88.0 30.24698 -81.62258 1430 94 28 Oligohal

FL0023671 JACKSONVILLE HEIGHTS Monthly 2.50 1.19 30.24100 -81.75670 1384 86 12 Oligohal

FL0023922 ORANGE PARK, TOWN OF Monthly 2694 2.50 1.34 99.5 30.18241 -81.70981 1511 103 21 Oligohal

FL0024767 SAN PABLO WWTF Monthly 1260 0.50 0.46 84.0 30.27763 -81.43065 1343 53 78 Mesohal

FL0025151 MILLER STREET WWTP Monthly 8471 5.00 3.41 65.0 30.17820 -81.71228 1511 103 21 Oligohal

FL0025828 ORTEGA HILLS SUBDIVISION Monthly 191 0.22 0.14 89.0 30.21869 -81.70962 1452 92 12 Oligohal

FL0026751 ROYAL LAKES Monthly 2.40 2.33 30.21389 -81.54440 1458 96 29 Oligohal

FL0026778 BEACON HILLS WWTF Monthly 2266 1.30 0.75 98.0 30.38379 -81.52166 750 31 57 Mesohal

FL0026786 WOODMERE SUBDIVISION Monthly 1106 0.50 0.35 97.0 30.37987 -81.60245 712 44 27 Mesohal

FL0030210 SOUTH GREEN COVE SPRINGS WWTF Monthly 3526 0.50 0.27 85.0 29.98259 -81.66759 1723 125 21 Fresh

FL0032875 FLEMING OAKS WWTP Monthly 5159 0.49 0.30 65.0 30.07463 -81.70457 1629 115 22 Oligohal

FL0038776 ATLANTIC BEACH WWTF Monthly 2218 3.00 1.70 92.0 30.33551 -81.40882 874 31 81 Mesohal

FL0040061 PALATKA, CITY OF Monthly 4724 3.00 2.76 95.0 29.61582 -81.65123 2175 182 42 Fresh

FL0041530 ANHEUSER BUSCH MAIN ST. LAND APP. Monthly 1.46 30.45278 -81.65000 89 29 11 Mesohal

FL0042315 CITY OF HASTINGS Monthly 0.06 29.72500 -81.50000 1927 154 29 Fresh

FL0043591 JULINGTON CREEK WWTP Monthly 6141 1.00 0.21 56.0 30.10634 -81.62597 1613 113 30 Oligohal

FL0043834 FLEMING ISLAND SYSTEM WWTP Monthly 8878 1.50 0.69 65.0 30.09279 -81.71982 1616 113 22 Oligohal

FL0117668 UNITED WATER FL - ST. JOHNS NORTH Monthly 0.00 30.09556 -81.61089 1613 113 30 Oligohal

FLA011427 USN NS MAYPORT Monthly 0.98 30.39690 -81.39750 558 31 94 Mesohal

FLA011429 USN NAS JACKSONVILLE Monthly 1.09 30.24138 -81.67580 1432 91 20 Oligohal

Brierwood S/D - Beauclerc STP Monthly 0.78 0.00 1445 95 29 Oligohal

24

Solving these 2 equations for LTOC produces:

LTOC = (CBOD5*74.906-TOC)/61.54

And;

RTOC = TOC – LTOC

In calculations, 2 of the 88 point source samples and 6 of the 702 tributary samples had CBOD5

values that indicated decay rates less than 0.001 day-1

; conversely, 3 point source samples in the

data set exhibited CBOD5 values that when converted to TOC exceeded the TOC at the

maximum decomposition rate of 0.075 day-1

. These values were omitted from subsequent

calculations.

Determination of Labile and Refractory Organic Nutrients

To determine labile and refractory organic nitrogen and phosphorus in tributary runoff and point

source effluents, the relationships between labile organic C content and organic C:N and C:P

ratios were examined to partition organic nitrogen (TON = TKN – NH4) and non-orthophosphate

P (TNOP = TP – PO4) into these respective pools. In this partitioning scheme, it is assumed that

the majority of nitrogen not accounted for in the separate analysis of inorganic nitrogen (NH4

and NOX) is in either dissolved or particulate organic matter. The same cannot be said for non-

orthophosphate phosphorus forms, as a significant proportion of this analytical fraction may in

the form of calcium or magnesium phosphates. For this reason, in fraction of total P not in

orthophosphate is referred to as “total non-PO4-phosphorus”, and abbreviated as TNOP.

Organic C:N and C:TNOP ratios for the tributary and point source data set were plotted against

percent labile organic carbon [(LTOC/TOC)*100] to determine the relationship between

proportional nutrient content and lability. One data point from stream runoff draining a large

dairy and intensive pasture lands in which the TOC:TNOP was 4225:1 was omitted from this

analysis. These log – log plots (Figure 5 and 6) demonstrate significant partitioning of carbon to

nutrient ratios based upon their content of labile organic carbon, with samples high in labile

organic carbon exhibiting low organic C:N and C:TNOP ratios.

25

To determine the TOC:TON and TOC:TNOP for hypothetical, purely labile or refractory

substrates, polynomial regressions of these log – log relationships were solved for the TOC:TON

and TOC:TNOP values corresponding to the %LTOC = 0% and when %LTOC = 100% (Figures

5 and 6). This yielded an TOC:TON mass ratio of 37 for a completely refractory substrate, and a

ratio of 4.5 for a completely labile substrate. In the case of non-orthophosphate phosphorous, the

TOC:TNOP mass ratios obtained were 617 for refractory OM and 27 for labile.

Figure 5. Organic Carbon:Nitrogen Ratio as a Function of the Percent Labile Organic Carbon.

for LSJR tributary and point source effluent samples; n = 763. Ratios in boxes

identify the organic C:N for the hypothetical conditions of 0 and 100 percent labile

organic carbon composition.

26

Figure 6. Organic Carbon:Phosphorus Ratio as a Function of the Percent Labile Organic

Carbon. for LSJR tributary and point source effluent samples; n = 727. Ratios in

boxes identify the organic C:P for the hypothetical conditions of 0 and 100 percent

labile organic carbon composition.

Determining refractory and labile nutrients directly as the product of the respective TOC:nutrient

ratio x TON or TNOP would likely result in departures from the laboratory analytical

determination of TON (TKN-NH4) and TNOP (TP-PO4). Because it is more plausible to

constrain the sum of labile and refractory nutrient within the laboratory analytical determination,

a proportional compartmentalization of nutrients was achieved by balancing labile and refractory

forms within the already separated organic carbon fractions. The form of this calculation was:

27

LTON =

{[LTOC/(TOC*4.5)]/[RTOC/(TOC*37) + LTOC/(TOC*4.5)]}*TON.

Following this calculation, RTON could be calculated by difference with the relationship

RTON = TON – LTON,

or with the complimentary partitioning equation of the form

RTON =

{[RTOC/(TOC*37)]/[RTOC/(TOC*37) + LTOC/(TOC*4.5)]}*TON.

Similarly, TNOP was partitioned with the relationship

RTNOP =

{[LTNOP/TOC*27)]/[RTNOP/(TOC*617)+LTNOP/(TOC*27)]}*TNOP.

This approach also had the advantage in that it could be applied to the existing watershed

modeling constituent breakdown, which provides TON and TNOP (Hendrickson and Konwinski,

1998). Thus, instead of deriving specific land use loading rates for LTON, RTON, LTNOP and

RTNOP, it was only necessary to establish new model specific land use water quality

coefficients for LTOC and RTOC. From these, labile and refractory ON and NOP could be

determined outside the model framework.

Calculation of the External Load For the Lower St. Johns River

Several different statistical estimation approaches have been relied upon to calculate the external

load to the LSJR. In general, the approach used is adapted to suit the inherent variability of the

load source, and the monitoring data available for estimation, calibration and verification.

28

Point Source Load Estimation

To perform the point source load estimation, six separate data sets were utilized to gain available

information on concentration, flow, point of discharge and service area. These data sets included

1) hard copy monthly operating report files maintained at the FDEP Northeast District Office; 2)

NPDES electronic files obtained from FDEP Tallahassee; 3) Discharge quality data maintained

by the Jacksonville Electric Authority; 4) Fifth-year synoptic surveys performed by FDEP

Tallahassee or by contractor as part of permit renewal process or WQBEL studies; 5) a special 2

year sampling program conducted jointly by FDEP-NED, SJRWMD and Duval County RESD;

and 6) a GIS data base of locational information compiled by contractor. Table 2 lists the point

source facilities included in the data base, their permitted volume, and location of entry into the

WQ model grid.

Point source data were compiled into two files based on sampling frequency. The JEA data base

in most cases contained daily data on flow and water quality concentration for the 5 largest

facilities in Jacksonville, and these data were the core of one data set. Remaining facilities with

monthly or quarterly reporting data were compiled into a second data base.

Data coverage for the JEA facilities was excellent, with almost complete daily coverage for the

entire 1995 through 1999 time interval. Data coverage for the remaining facilities was fair to

good, with data coverage increasing through time as permit monitoring requirements increased to

cover nutrients in effluent. The most serious data deficiency occurred for total organic carbon,

and data from a short term, joint sampling program from 1995 through 1996 were heavily relied

upon to supply typical values for this constituent.

To calculate daily loads for facilities with monthly reporting data, mean monthly flow was

multiplied by monthly grab sampling or flow composite water quality data when available.

Generally, monthly nutrient concentrations were not available, as quarterly nutrient sampling

was typically the case for these facilities, so the mean of the sampling record was used. Some

questions arose regarding data representation, as many nutrient values were recorded in the

FDEP WAFER data base as “monthly maximum value”. However, after conferring with FDEP-

29

NED staff, it was concluded that these data were invariably fixed-interval grab sampling data,

and could be used to generate representative mean values.

Non-Point Source Load Estimation

Unlike point source effluent loads, nonpoint source loads enter at so many locations and exhibit

such large temporal variation that a direct monitoring approach is infeasible except for the

largest, most significant inputs. At all other nonpoint entry points, statistical watershed modeling

is relied upon to complete the external load budget.

Land development influences the delivery of water quality constituents to surface waters in two

fundamental ways. Through fertilization, lawn maintenance, manure spreading, septic tank

operation, vehicular use, etc., nutrients and other pollutants are added to the land surface or to

shallow groundwater in excess of natural land cover conditions (i.e., native forest, wetland).

Unlike the situation that tends to predominate on developed lands, natural land covers are highly

conservative of essential growth nutrients, and thus labile nutrient forms tend to be retained

within these terrestrial ecosystems. In addition, the creation of impervious surfaces, drainage

development, and the destruction of near stream wetlands increases the amount of rainfall that

ultimately ends up as runoff, thus increasing the pollutant exporting capability in developed

landscapes. Thus, the process of nonpoint source pollution has both chemical and hydrologic

components.

The watershed modeling approach used for the LSJR TMDL and PLRG development utilizes the

relationship between land use development and alteration in water quality and quantity to

perform a spatial extrapolation of whole basin nonpoint source load. The formulation of this

statistical model has its roots in the spreadsheet watershed load screening model, referred to as

the Pollution Load Screening Model (acronym PLSM; Adamus and Bergman 1995), which

utilizes a computer-driven geographic information system framework to calculate constituent

loads as the product of water quality concentration associated with certain land use practices, and

runoff water volume associated with those same practices. The model’s nonpoint source

30

pollutant export concentrations are specific to one of 15 different land use classes. Water

quantity is determined through a hybrid of the SCS curve number method, and is the product of

rain volumes and a coefficient (referred to as the runoff coefficient, or RC, with values ranging

from 0 to 0.9) relating the propensity of various land use and soil hydrologic group combinations

to generate runoff. The computational approach of the PLSM is similar to that of the Surface

Water Management Model (SWMM) screening level tool.

In the initial application of the PLSM to the LSJRB, Hendrickson and Konwinski (1999) made 4

major modifications to the model’s original framework: 1) the model time step was shortened to

seasonal, rather than annual average loading rates, to account for seasonal differences in specific

land use export concentrations and runoff quantity; 2) total nutrient forms were subdivided to

provide orthophosphate and total inorganic nitrogen, and by difference, TON and TNOP; 3)

land-use loading rates were adjusted to monitoring data collected within the LSJR basin using a

linear multiple regression best-fit approach based on contributing land-use fractions in

calibration watersheds; and 4) runoff coefficients were varied by season to account for intra-

annual variation in rainfall and evapotranspiration patterns. In the original application,

nonpoint source loads were predicted for the time period from 1993 to 1995, relying on land use

information compiled from 1989-90.

In this application of the PLSM, the modifications described above were maintained, and in

addition 1) LTOC and RTOC, were added to the specific land-use loading rate water quality

coefficients; and 2) runoff water quantity was varied based upon deviations in the long term

rainfall patterns. From the model output of labile and refractory organic carbon, LTON, RTON,

LTNOP and RTNOP were differentiated based on the proportional nutrient ratio weighting

described previously. A parameter referred to as the long-term rain ratio (LTRR) was developed

as a weighting factor to adjust the PLSM runoff coefficient based on antecedent watershed soil

water conditions, and is described below. Two versions of the PLSM were utilized. One version

ran within ARC Info, and calculated loads directly as the sum within contributing areas of

overlying grid coverage-products of runoff and water quality concentration. A second version

was run within Microsoft Excel, and calculates area-weighted runoff coefficient and runoff-

weighted concentration based upon the area within contributing watersheds of unique land us

31

and soil hydrologic group combinations. Though different in computational approach, both

models are theoretically identical and provided only slightly different results owing to the areal

weighting applied to rainfall input data in the Excel model application.

Watershed NPS Hydrologic Set-Up and Calibration

One of the principal deficiencies of the original PLSM hydrologic algorithm for predicting time-

varying load is its inability to account for short term changes in antecedent soil moisture

conditions that lead to changes in the propensity for rainfall to generate runoff. To account for

patterns in antecedent soil moisture associated with normal intra-annual patterns in rainfall and

evapotranspiration, the previous PLSM application to the LSJR utilized a set of runoff

coefficients for each land use-soil hydrologic group combination that varied by season (Table 3).

While this adjustment helped to simulate seasonal runoff patterns, model responsiveness to long-

term deviations from the normal seasonal rainfall pattern, for which model runoff coefficients

were originally established, was still poor. The underlying nature of this insufficiency is readily

shown in Figure 7, which plots the ratio of the PLSM-predicted runoff to measured runoff to the

seasonal whole-watershed water yield (the ratio of the measured runoff volume to the watershed

seasonal incident rainfall volume). When watershed water yield is low, resulting from prevailing

drier than normal conditions, the fixed runoff coefficient of the PLSM formulation over-predicts

measured volume (PLSM/observed > 1). When watershed yield is high due to high rainfall

seasons, the opposite is observed (PLSM/observed < 1). Because of the large range in flow

conditions occurring in the 1995-99 TMDL/PLRG modeling time window (the relative ranking

of mean annual flow rates for the 1995 – 1999 is shown in Figure 8), it was critical to obtain a

better time-varying estimate of runoff, and hence, load.

To adjust for this model insensitivity, a correction factor was developed based upon two

concepts. First, as rainfall and runoff patterns deviate form the long term norm, the ratio of the

fixed, PLSM runoff coefficient to the measured, basin yield (rain volume/runoff) varies in a

predictable manner, as was shown in Figure 7. Because extended drought-induced changes in

the rainfall-runoff relationship may occur over several seasons, the calculation to devise what is

32

referred to as the long-term rain ratio (LTRR) incorporated rain excess or deficit for one year (3

seasons) with the following equation:

LTRR = [RAINCS/LT RAINCS + (RAINCS-1/LT RAINCS-1)/2 +

(RAINCS-2/LT RAINCS-2)/3]/LT RAINCS

where:

RAINCS, RAINCS-1, and RAINCS-2 = current season rain, the previous season’s rain, and rain of

season prior to previous season; and LT RAINCS, LT RAINCS-1, and LT RAINCS-2 = 30-year,

long term mean rain for the corresponding seasons above (Rao et al., 1997).

Second, the rate of deviation of the ratio of PLSM runoff : Measured runoff with drought or

wetness is a function of the degree of impervious surface area of the basin. Under extended

durations of low rain, basins with a large degree of impervious surface area are less affected by

dry soil moisture conditions, and as a result still return a relatively large amount of their rainfall

as runoff.

33

Table 3. Seasonal Runoff Coefficients for Application of the Pollution Load Screening Model

to the LSJR Basin. Values represent the fraction of that produces runoff.

Land Use Soil Hydrologic Group

A B C D

Well Poorly

Drained Drained

Season 1: December through March

Low Density Residential 0.05 0.12 0.18 0.25

Medium Density Residential 0.5 0.6 0.7 0.8

High Density Residential 0.6 0.7 0.8 0.9

Low Density Commercial 0.5 0.6 0.7 0.8

High Density Commercial 0.7 0.8 0.9 1

Industrial 0.5 0.6 0.7 0.8

Mining 0.05 0.12 0.18 0.25

Miscellaneous Agriculture 0.05 0.12 0.18 0.25

Pasture 0.05 0.12 0.18 0.25

Row Crop 0.401 0.401 0.401 0.401

Citrus 0.05 0.12 0.18 0.25

Livestock Feedlots 0.05 0.12 0.18 0.25

Forestry, Silviculture, Range, Barren 0.05 0.12 0.18 0.25

Water Surfaces 1 1 1 1

Wetlands 0.95 0.95 0.95 0.95

Season 2: April through July

Low Density Residential 0 0 0.05 0.1




High Density Commercial 0.4 0.5 0.6 0.7

Industrial 0.2 0.3 0.4 0.5

Mining 0 0 0.05 0.1

Miscellaneous Agriculture 0 0 0.05 0.1

Pasture 0 0 0.05 0.1

Row Crop 0.392 0.392 0.392 0.392

Citrus 0 0 0.05 0.1

Livestock Feedlots 0 0 0.05 0.1

Forestry, Silviculture, Range, Barren 0 0 0.05 0.1


Wetlands 0.75 0.75 0.75 0.75

Season 3: August through November

Low Density Residential 0.05 0.15 0.25 0.35




High Density Commercial 0.7 0.8 0.9 1

Industrial 0.55 0.65 0.75 0.85

Mining 0.05 0.15 0.25 0.35

Miscellaneous Agriculture 0.05 0.15 0.25 0.35

Pasture 0.05 0.15 0.25 0.35

Row Crop 0.512 0.512 0.512 0.512

Citrus 0.05 0.15 0.25 0.35

Livestock Feedlots 0.05 0.15 0.25 0.35

Forestry, Silviculture, Range, Barren 0.05 0.15 0.25 0.35


Wetlands 1 1 1 1

34

Figure 7. Relationship Between PLSM Predicted Runoff:Observed Runoff Ratio and Measured

Seasonal Whole Watershed Runoff Coefficient. Seasonal Whole Watershed Runoff

Coefficient = Incident Rainfall Volume/Runoff Volume

Figure 7. Relationship Between PLSM Predicted Runoff:Observed Runoff Ratio and

Measured Seasonal Whole Watershed Runoff Coefficient. Seasonal Whole

Watershed Runoff Coefficient = Incident Rainfall Volume/Runoff Volume

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00

PLSM Q/Observed Seasonal Q

Ob

serv

ed

Seaso

nal W

ho

le W

ate

rsh

ed

Rain

Vo

lum

e/R

un

off

PLSM =

Measured Q

35

Figure 8. Relative Position of the 1995-1999 Time Interval in the Historic Long Term Flow

Record. (From Hendrickson and Magley, 2002.)

St. Johns R. at Deland Nth.Fork Black Creek1960 7463.8 1964 447.1

1953 5402.8 1959 411.5

1959 4896.4 1947 363.6

1947 4835.8 1948 356.1

1941 4800.1 1979 338.5

19 9 5 4 70 4 .1 1966 316.5

1966 4628.4 1953 299

1964 4458.3 1960 283.8

1948 4376.9 1970 283.4

19 9 8 4 2 2 8 .4 1965 273.3

1979 4149.8 1991 272.8

1945 4095.5 1983 268.8

1994 4054.1 1992 267

1982 3970.3 1969 266.5

1968 3926 1968 261.1

1969 3918.6 1994 260.7

1934 3910 1984 242.6

1983 3851.6 1973 242.3

1978 3662.8 1944 238.9

1949 3649.1 19 9 5 237.4

1991 3481 19 9 8 236

19 9 6 3 3 4 3 .6 1946 232

1974 3259.5 1963 228

1954 3212.3 1974 224

1992 3202.6 1950 222.3

1957 3195.9 19 9 6 220.6

1984 3152.6 1933 220.5

1944 3148.8 1949 218

1936 3111.6 1987 208.8

1985 3026.9 1972 206

1942 2984.2 19 9 7 203.9

1963 2980.1 1982 199.8

1973 2968 1961 198

1946 2895 1988 197.7

1970 2842.7 1980 194.9

1952 2795.5 1941 191

1943 2704 1942 189

1988 2688.5 1945 187.3

1950 2654.3 1978 186.5

1976 2610.9 1967 175.2

1951 2597.7 1958 170.2

1987 2573.6 1937 156.4

1958 2573.1 1985 151.1

1956 2562.4 1986 150.7

1965 2383.3 1957 143.1

1935 2362.5 1971 142.8

1937 2360 1934 140.8

1967 2326.4 1956 137.8

19 9 7 2271 1993 134.3

1939 2259.6 1989 130.2

19 9 9 2226 1938 120.8

1972 2186.5 1975 120.7

1993 2122.3 1977 117

1955 2107.2 1940 115

1975 2103.6 1962 107.4

1986 2094.4 1939 107.3

1940 1958.1 1976 99.6

1938 1920.7 1935 92.9

1977 1920 1952 89

1989 1804 1932 86.4

1962 1715.4 1981 85.6

1961 1714.2 1943 84.2

1990 1505.4 1955 60

1971 1307.7 1951 58.1

1980 1174.2 19 9 9 48.9

1981 859.3 1954 48

1990 41.9

1931 12.5

1995

1998

1996

1997

1999

36

Figure 9 demonstrates this 2-step analysis process. Analysis was confined to the 4 most reliable

flow gauging stations; the Deep Creek gauge, due to poor performance of the acoustic flow

meter at this site, was excluded. The original model-predicted, area-weighted PLSM runoff

coefficient for each season and year for these four calibration watersheds (60 seasonal values

total) were placed into 4 runoff coefficient classes. The rate of change (slope) in the measured

yield : PLSM predicted runoff coefficient ratio as a function of the LTRR2 was determined for

each of these classes through zero-intercept simple linear regression (Figure 9(a)). As PLSM-

predicted runoff coefficient class increases, the rate of change of the observed:PLSM runoff

coefficient ratio, as LTRR increases, declines. Stated more simply, as watershed impervious

surface area increases, the degree to which variation in the long term rainfall pattern leads to

deviations in the fixed, PLSM runoff coefficients, decreases. To account for lower response in

watersheds with higher amounts of impervious surface, regression was again used to quantify the

rate of PLSM runoff coefficient deviation with changes in watershed impervious area,

approximated by the original PLSM runoff coefficient (Figure 9(b)). This relationship was then

integrated into an adjustment for the PLSM fixed, seasonal runoff coefficients based on the

LTRR of the form:

Observed RC/PLSM RC = LTRR2*(0.3228*PLSM-RC

0.6206)

This relationship was multiplied through by the PLSM-predicted runoff coefficient to derive a

long term, rain-adjusted runoff coefficient:

RUNOFF COEFFICIENTadj = PLSM-RC*[LTRR2*(0.3228*PLSM-RC

0.6206)]

Figure 10 compares the original and adjusted PLSM runoff coefficient and seasonal flow volume

to the measured watershed yields and seasonal flows. The adjustment removes bias and

improves precision in both the PLSM runoff coefficients and flows, moving slopes of the

regressions between observed and simulated to near one and zero, respectively. The correlation

coefficient improves from 0.12 to 0.43 in the case of runoff coefficient and from 0.59 to 0.80 in

the case of seasonal flow volume. Cumulative discharge curves for the seasons from December

1994 through November 1999 for the measured, original PLSM and PLSM-adjusted discharge

37

volumes (Figure 11) show that, with the exception of the South Fork of Black Creek, the long

term rain-ratio adjustment of PLSM seasonal simulations greatly reduces the cumulative over-

prediction of the measured flow.

Water Quality Set-Up and Calibration

In the original application of the PLSM to the LSJR basin, specific land use water quality

coefficients were developed for total nitrogen (TN), total inorganic nitrogen (TIN, or NOX +

NH4), total phosphorus (TP), orthophosphate (PO4), biochemical oxygen demand (BOD), and

total suspended solids (TSS). These coefficients, shown in Table 4, were left unchanged from

the earlier application, so the comparison of measured to simulated values here represents a skill

assessment for these constituents. (The term skill assessment is used here, rather than

verification, as the calculated flow-weighted concentrations that are used to compare to

watershed model simulations encompass the data used in the original calibration, collected from

1990-95, and newer data from 1996-2000.) New to this iteration of model development are total

organic carbon (TOC), labile total organic carbon (LTOC), and refractory total organic carbon

(RTOC). To calculate the labile and refractory portions of organic nitrogen and non-PO4-

phosphorus, the proportioning equations described earlier were applied utilizing the simulated

LTOC or RTOC operating on the difference of simulated TN-TIN and TP-PO4. The use of the

simulated TON and TOP values assures that TIN+RTON+LTON and PO4+LTOP+RTOP will be

equal to TN and TP.

38

Figure 9. Development of Hydrologic Correction Factors for the PLSM Runoff Coefficient. (a)

Linear regressions relating changes in seasonal rainfall pattern, expressed as

(LTRR)2, the ratio of observed watershed yield to PLSM predicted. (b) Relationship

between the slope of the (LTRR)2 and the watershed area-weighted runoff coefficient.

(b) Slope LTRR2 x OBS/PLS vs. Mean WS Runoff Coefficient

y = 0.3228x-0.6206

R2 = 0.9065

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700

Watershed Area Weighted RC

Slo

pe L

TR

R x

O

bs/P

LS

M

(a) Long Term Rain Ratio x PLSM RC/Observed Yield

y = 0.8024x

R2 = 0.3786

y = 0.6152x

R2 = 0.486

y = 0.4596x

R2 = 0.2111

y = 0.4534x

R2 = 0.8376

0

0.5

1

1.5

2

2.5

3

0 1 2 3 4 5 6

LTRR2

OB

S/P

LS

RC=.199-.288

RC=.344-.402

RC=.441-.510

RC=.607-.674

39

Figure 10. Comparison of Original, Seasonal-fixed and Long-Term Rain Ratio Adjusted

Runoff Coefficients (a) and Total Seasonal Discharge (b) for Flow Calibration

Watersheds Within the LSJRB.

(a) Observed vs. Predicted Runoff Coefficient

y = 0.2588x + 0.3423

R2 = 0.1236

y = 0.9155x + 0.0423

R2 = 0.434

0.000

0.200

0.400

0.600

0.800

1.000

1.200

1.400

0.000 0.200 0.400 0.600 0.800 1.000

Observed Watershed RC

Pre

dic

ted

RC

PLSM Original

Adjusted RC

(b) Measured x Simulated Seasonal Flow - All Stations

y = 0.9654x + 1E+06

R2 = 0.8037

y = 0.6602x + 2E+07

R2 = 0.5881

0.E+00

5.E+07

1.E+08

2.E+08

2.E+08

3.E+08

0.E+00 5.E+07 1.E+08 2.E+08 2.E+08

Observed

Sim

ula

ted

Adjusted Runoff

Original PLSM

40

Figure 11. Comparison of Original PLSM-Predicted, Long-Term Rain Ratio Adjusted and

Observed Cumulative Discharge Curves for LSJRB Calibration Watersheds, 1995-

99.

Big Davis Creek

0

2

4

6

8

10

12

14

95s1

95s3

96s2

97s1

97s3

98s2

99s1

99s3

Cu

mu

lati

ve M

ean

Dis

ch

ag

e, m

3/s Observed

PLSM

PLSM-Adjusted

Deep Creek

0

5

10

15

20

25

30

35

40

95s1

95s3

96s2

97s1

97s3

98s2

99s1

99s3

Cu

mu

lati

ve M

ean

Dis

ch

arg

e, m

3/s

Ortega River

0

5

10

15

20

25

30

35

95s1

95s3

96s2

97s1

97s3

98s2

99s1

99s3

Cu

mu

lati

ve M

ean

Dis

ch

arg

e, m

3/s

Rice Creek

0

5

10

15

20

25

30

35

95s1

95s3

96s2

97s1

97s3

98s2

99s1

99s3

Cu

mu

lati

ve M

ean

Dis

ch

arg

e, m

3/s

South Fork Black Creek

0

10

20

30

40

50

60

70

80

95s1

95s3

96s2

97s1

97s3

98s2

99s1

99s3

Cu

mu

lati

ve M

ean

Dis

ch

arg

e, m

3/s

North Fork Black Creek

0

20

40

60

80

100

120

95s1

95s3

96s2

97s1

97s3

98s2

99s1

99s3

Cu

mu

lati

ve M

ean

Dis

ch

arg

e, m

3/s

41

Table 4. Seasonal Water Quality Coefficients Used In the PLSM to Predict Non-Point Source

Loads to the LSJR. All Values represent flow-weighted concentrations in mg/L.

Season 1: Dec. 1 through Mar. 31

Land Use TN TP BOD SS TIN TPO4

Low Density Res. 0.8 0.08 1 6 0.04 0.06

Medium Density Res. 1.4 0.25 2 15 0.35 0.1

High Density Res. 1.8 0.3 4 20 0.4 0.13

Low Dens. Commercial 1.1 0.2 2 15 0.35 0.1

High Dens. Commercial 1.2 0.3 4 25 0.4 0.13

Industrial 1.2 0.25 2 25 0.4 0.1

Forest, Range/Open, Barren 0.7 0.06 1 25 0.02 0.04

Pasture 3.9 0.75 4 15 1 0.6

Row Crop, Misc. Ag 2 0.38 1 15 0.7 0.2

Livestock 4.5 1.3 6 15 1.5 1

Water (Atmos. Wetfall) 0.28 0.017 0 0 0.28 0.015

Wetlands 0.7 0.06 1 3 0.02 0.04

Season 2: Apr. 1 through Jul. 31


Low Density Res. 0.8 0.07 1 6 0.04 0.05


High Density Res. 2 0.5 4 40 0.3 0.12


High Dens. Commercial 2 0.5 4 50 0.4 0.12

Industrial 1.2 0.3 2 40 0.4 0.1


Pasture 3 1.1 4 10 1 0.85

Row Crop, Misc. Ag 10.7 1.8 1 54 4.7 0.6

Livestock 6 1.3 6 30 1.2 1


Wetlands 0.7 0.05 1 3 0.04 0.03

Season 3: Aug. 1 through Nov. 30


Low Density Res. 0.8 0.09 1 5 0.06 0.07


High Density Res. 1.7 0.53 4 35 0.55 0.22


High Dens. Commercial 1.7 0.53 4 35 0.55 0.22

Industrial 1.3 0.22 2 25 0.3 0.13


Pasture 3 2 4 15 1.2 2.1

Row Crop, Misc. Ag 4.4 2.2 1 26 0.55 1.6

Livestock 5 2.6 6 30 1.7 2.4


Wetlands 0.7 0.07 1 3 0.04 0.05

42

Following the calculation of LTOC and RTOC for the water quality data set, specific land use

flow-weighted mean concentrations were developed following the procedure outlined in

Hendrickson and Konwinski (1998). Mean flow-weighted concentrations for LTOC and RTOC

were computed by sub-dividing water quality samples for each season and station into low flow

(samples collected within the 1 – 50th

percentile flow days) and high flow (samples collected

within the 51 – 100th

percentile flow days) and multiplying mean concentrations within flow

class by the proportion of total seasonal flow volume occurring within the class, and summing

these two products. Seasonal, flow-weighted concentrations were used to set specific loading

rates, as observations from blackwater stream monitoring stations suggest that TOC

concentrations in runoff change significantly due to season (Figure 12). Seasons used were not

the Julian seasons but modified seasons corresponding to hydrologic and meteorological patterns

of northeast Florida, and were comprised of a cool, moderately wet winter season from

December through March characterized by regular frontal storm events; a hot, dry

spring/summer from April through July; and a hot, wet summer/fall from August through

November characterized by afternoon convective thunderstorms and tropical systems.

A multiple regression approach was then used to objectively assign seasonal, specific land-use

concentrations. Land uses were aggregated into four broad classes: 1) row crop agriculture, of

which the majority was cabbage and potatoes on seepage-irrigated flatwoods soils; 2) dairy,

including associated feedlots, loafing areas, improved pastures and manure spray-fields; 3)

medium to high intensity residential, commercial and industrial; and 4) silviculture, native forest

and wetlands. Multiple regression models for each season were set up in the form:

1*LU-FR + 2*LU-FD + 3*LU-FU + 4*LU-FF = [TOC, LTOC, RTOC],

43

Figure 12. Monthly Mean Concentrations and 95% Confidence Intervals for Color and Total

Organic Carbon for 24 Unimpacted Blackwater Streams in Northeast Florida.

(a) Color

0

50

100

150

200

250

300

350

400

450

500

1 2 3 4 5 6 7 8 9 10 11 12

MONTH

ME

AN

CO

LO

R,

Pt-

Co

Un

its

(b) Total Organic Carbon

0

5

10

15

20

25

30

35

40

1 2 3 4 5 6 7 8 9 10 11 12

MONTH

ME

AN

TO

C,

mg

/L

44

where LU-FR, LU-FD, LU-FU, and LU-FF were the fractions of watershed area upstream of

sampling stations in the aggregated row crop, dairy, urban and forested categories. Regression

statistics were developed using Minitab Statistical Software. Determination of water quality

coefficients in this manner generally produces concentrations lower than those determined from

specific land use monitoring on small catchments of homogeneous land use, such as the type

developed for the National Urban Runoff Program monitoring (U.S. EPA, 1984), presumably

due to instream assimilation and sedimentation that leads to an attenuation of concentration from

source to mouth. However, such “implicitly attenuated” loading rates are viewed as superior for

whole watershed determinations of constituent loads, as would be required for input to a

receiving water body eutrophication model.

TOC, LTOC and RTOC concentrations within these aggregated land use categories were then

disaggregated into a broader set of land uses corresponding to those employed in the PLSM

framework of Adamus and Bergman (1995). The set of final water quality coefficients for all

land uses is shown in Table 5.

In their development of the EFDC hydrodynamic model for the lower St. Johns, Sucsy and

Morris (2002) identified 63 discrete entry points and their contributing areas for water volumes

entering the model grid (Figure 13), and these have been retained for the nonpoint source load

entry points to ICM.

45

Table 5. Total organic, labile total organic, and refractory total organic carbon land use category concentration coefficients.

Aggregated class model coefficients in bold, with individual class concentrations estimated by the trend with land use

intensity. RTOC concentrations determined from TOC - LTOC. Values in parenthesis are regression model coefficients.

Constit-

uent

Urban

Great

Class

Med.

Dens.

Resi-

dential

High

Dens.

Resi-

dential

Low

Intens.

Comm.

High

Intens.

Comm.

Indus-

trial

Dairy

Great

Class

Feed-

lot,

Loafing

Pas-

ture

Row

Crop

Great

Class

Row

Crop

Citrus,

Tree

Crops

Miscel.

Ag.

Forest

Great

Class

Low

Dens.

Resi-

dential Mining

Range/

Open

Space

Upland

Forest

Silvi-

culture

Wet-

lands Barren

F

Statistic

Season 1

TOC 11.7 14 11 12 9 12 34.3 45 34 17.3 17.3 17.3 17.3 18.0 16 16 18 20 22 30 14 43.8

LTOC 4.6 3 4 5 6 5 4.6 15 4 1.5 1.5 1.5 1.5 0.5 1 0.5 1 0.5 0.5 1 0.5 43

RTOC 7.1

(11.6)

11 7 7 3 7 29.7

(27.5)

30 30 15.8

(13.4)

15.8 15.8 15.8 17.5

(15.9)

15 15.5 17 19.5 21.5 29 13.5 36.5

Season 2

TOC 8.7 12 8 10 7 10 18.4 30 18 11.4 11.4 11.4 11.4 18.1 16 16 18 20 22 30 14 47.5

LTOC 5.1 4 5 5 6 5 2.6 15 3 2.7 2.7 2.7 2.7 1.4 2 1 2 1 1 2 1 15.2

RTOC 3.6

(4.1)

8 3 5 1 5 15.8

(17.6)

15 15 8.7

(6.5)

8.7 8.7 8.7 16.7

(13.9)

14 15 16 19 21 28 13 20.8

Season 3

TOC 7.8 10 7 8 7 10 47.2 55 45 27.5 27.5 27.5 27.5 22.1 20 20 22 20 24 30 18 58.9

LTOC 3.4 3 3 3 4 4 4.4 14 4 2.4 2.4 2.4 2.4 0.9 1.5 1 1.5 0.5 1 2 0.5 15.9

RTOC 4.4

(6.5)

7 4 5 3 6 42.8

(45.1)

41 41 25.1

(25.2)

25.1 25.1 25.1 21.2

(19.8)

18.5 19 20.5 19.5 23 28 17.5 32.3

Notes:2) Loading rates for forested land covers increased slightly due to: a) dry conditions of the calibration data set, and b) the preponderance of sandy ridge sampling sites.3) Specific land use breakouts based on 2 premises: a) Forested covers export more TOC than grassed or annual plant covers; and b) Suwannee River WAM BOD loading rates were used to proportion LTOC on the development intensity continuum.

46

Figure 13. Watershed Model Input Areas for Nutrient Load Compilation.

47

Determination of the Upstream Load to the LSJR

Background

The upstream load is composed of the three large tributaries that make up the lower St. Johns:

the middle St. Johns River, the Ocklawaha River, and Dunns Creek (Figure 14). These three

tributaries make up approximately 61, 21, and 18 percent of the long-term annual mean river

discharge at Palatka. Because of autochthonous production in upstream lakes, the upstream load

differs greatly from watershed loads that enter within the LSJR basin.

Because of the large amount of the entire LSJR flow that enters upstream (roughly 60%), a direct

monitoring approach, featuring continuous measurement of discharge, and bi-weekly collection

of water quality samples, has been used to determine its constituent load. This is in contrast to

the watershed modeling approach that has been used to develop the downstream tributary load.

Along with water quality monitoring, phytoplankton (algae) monitoring has also been performed

at these inputs to determine the amounts and types of algae entering the LSJR, and how they

change throughout the year and from year to year.

The presence of phytoplankton, and the preponderance of nutrients and carbon that is contained

in phytoplankton, is the single most important feature that separates the upstream inputs in the

whole LSJR external load. This poses an additional level of difficulty in the

compartmentalization of nutrients and carbon. Other tributaries to the LSJR (with the exception

of Doctors Lake, which in the LSJR TMDL modeling is included within the model grid) contain

insignificant amounts of algae, and distinguishing algal-borne nutrients and carbon in the

particulate labile organic pool has not been done in the calculation of the watershed nonpoint

source load.

48

Figure 14. Tributaries Forming the Lower St. Johns River.

Crescent

Lake

Lake

George

Lake

Ocklawaha

Palatka

Dunns Creek

Station

Buffalo Bluff

Station

Jacksonville

Middle St.

Johns R.

Ocklawaha R.

Dunns Cr.

Lower St.

Johns R.

Approach to Upstream Boundary Nutrient and Carbon Partitioning

The combined incoming load from the middle and upper St. Johns and Ocklawaha River was

determined from the LSJR sampling station at Buffalo Bluff, a location 11 miles downstream of

the Ocklawaha River mouth. Direct monitoring was also used to characterize the load entering

the LSJR from the Crescent Lake Basin, at the sampling location in Dunns Creek, just

downstream of the Buffalo Bluff site. Both locations are instrumented with acoustic Doppler

current profilers, capable of measuring velocity and discharge in conditions of reversing flows.

Biweekly water quality monitoring has been performed at these locations under two programs,

with only slight differences in constituents. An ambient monitoring program, collecting samples

during the second week of each month, includes BOD in its suite, necessary for the empirical

separation of labile and refractory organic carbon. A second monitoring network samples these

locations on the fourth week of the month, and includes the analysis of particulate organic

49

carbon, and a complete plankton analysis featuring taxonomic classification, cell densities and

cell biovolume for major algal taxonomic groups. Both programs analyze complete suites of

chlorophyll and dissolved and total organic and inorganic nutrients and carbon.

The flow chart of Figure 15 illustrates the mapping of water quality chemical analysis results to

respective biological compartments of the LSJR WQ model. In this calculation cascade, one of

the most important determinations that must be made (and that distinguishes this boundary

determination from downstream tributary inputs) is the content of the total nutrient and carbon

load in phytoplankton. In this approach, the inorganic forms (for nitrogen or phosphorus) are

distinguished from the organic forms directly through laboratory analysis as shown. Next,

organic carbon is subdivided into labile or refractory forms based upon the procedure described

previously for tributary nonpoint sources, utilizing BOD converted to carbon units and the

established first-order decomposition rates. Labile organic carbon and nutrients are then further

subdivided into algal and non-algal, under the assumption that all algal organic matter is labile.

The determination of algal carbon content is described below, and algal nutrient composition was

determined from this algal carbon estimate, assuming adherence to the Redfield stoichiometric

proportions of 106:16:1 C:N:P molar.

50

Figure 15. Flow chart for differentiation of laboratory analytical fractions into CE-QUAL-ICM state variables for the lower St.

Johns River upstream boundary at Dunns Creek and Buffalo Bluff.

Total Form

TN, TP, TOC

Inorganic

(NOX, NH3, PO4)

Organic

(TKN-NH3, TP-PO4, POC+DOC)

Labile

(TOC consumption rate = 0.075/day)

Refractory

(TOC consumption rate = 0.001/day)

Particulate

(RTOC - RDOC; nutrients by

stoichiometric partitioning)

Dissolved

(RDOC determined by Color; nutrients by stoichiometric

partitioning)

Particulate

(LTOC - LDOC; nutrients by


Dissolved

(DOC - RDOC; nutrients by


Non-algal

(LPOC - Algal POC)

Algal

(By biovolume or chlorophyll a;nutrients by Redfield ratio)

Cyanobacteria Diatoms Other algal sp.

51

A data set of coincidently collected particulate organic carbon, nitrogen and phosphorus,

chlorophyll a, and algal biovolume (Phlips and Cichra, 2001) were relied upon to determine algal

carbon content and confirm nutrient ratios. Direct measurements of particulate organic carbon

using the high temperature combustion coulombmetric method were used to establish algal

C:chlorophyll a and C:algal biovolume ratios. Estimation of POC by difference of TOC – DOC

with measurements performed by the carbon gas analyzer were not used due to their much lower

precision and poor fit with algal biomass. Coulombmetric POC data are limited in coverage,

having been performed only once monthly from August 1998 to the present. Thus, to calculate

TOC for subsequent compartmentalization to labile and refractory forms for dates on which POC

was not determined, TOC was considered to be at least equal to DOC + algal POC, with algal

POC determined either through algal biovolume or chlorophyll a, preferentially selecting

biovolume data in the calculation when available. At most, TOC was considered to be equal to

DOC + (total suspended solids 0.4), representing a situation in which all TSS were organic

particles with a 40% composition by weight of carbon.

Comparison of algal biovolume to corrected chlorophyll a for LSJR freshwater sampling stations

(Figure 16) shows a strong correlation between these two variables, as would be expected.

Regression R2 values for the relationship range from 0.75 to 0.83 for the 6 stations examined,

and intercepts are near zero. A similar correlation is obtained in regressions between

uncorrected chlorophyll a and biovolume, with the exception that the intercept values are higher,

suggesting the presence of chlorophyll-containing algal detritus that is not accounted for in the

biovolume determination.

52

Figure 16. Comparison of Corrected Chlorophyll a and Algal Biovolume for Combined LSJR

Freshwater Water Quality and Plankton Analysis, 1995 – 2001. Plankton taxonomy

and biovolume determinations performed by Phlips and Cichra, 2001.

Both chlorophyll a and algal biovolume were strongly correlated to measurements of particulate

organic carbon, though linear regression R2 values were typically driven by a small number of

large values in the data set associated with algal bloom events, when a large amount of the POC

is expected to be in algal biomass. To help distinguish carbon to biomass relationships, sample

C:chlorophyll and C:biovolume ratios were viewed as boundary relationships, in which the

approach of POC:biomass values the true C:biomass measure occurs as non-algal carbon

declines (Figure 17). In these boundary relationships, it can be seen that minimum C:chlorophyll

ratios approach the commonly cited carbon:chlorophyll a ratio of 50:1, shown in red in Figure

17(a). In the case of the biovolume comparison, a similar boundary analysis yields an estimate

of 3.3 m3/ml per mg/L POC.

Buffalo Bluff

Chla = 0.0034*(Biov) + 9.5556

R2 = 0.8188

Federal Pt.

Chla = 0.005*(Biov) + 3.3355

R2 = 0.8258

Crescent L.

Chla = 0.0051*(Biov) + 5.3773

R2 = 0.8212

Deep Cr. Cove

Chla = 0.0045*(Biov) + 5.706

R2 = 0.7664

L. George

Chla = 0.0036*(Biov) + 8.739

R2 = 0.8104

Palatka

Chla = 0.0056*(Biov) + 5.2055

R2 = 0.7508

0

20

40

60

80

100

120

140

160

0.0 5000.0 10000.0 15000.0 20000.0 25000.0 30000.0

Total Biovolume, (micro-m3/ml)*1000

Co

rrecte

d C

hlo

rop

hyll a

, m

g/m

3

Buffalo Bluff

Deep Cr. Cove

Federal Point

L. George

Palatka

Crescent L.

53

To perform the partitioning of organic carbon when reliable TOC measurements were not

available, RDOC was first derived through a relationship with color (Figure 18), and LDOC

determined as the DOCmeasured – RDOCcolor-derived. LTOC and RTOC were then calculated by

way of the carbon equivalency BOD decomposition-derivation described in the watershed model

coefficient development, using a TOC estimated by DOC + algal OC. For Dunns Creek, this

approach suggested that a large fraction of the organic carbon pool is contained in RDOC. In the

case of the Buffalo Bluff station, however, RDOC concentrations were lower, hence relatively

large amounts of DOC were attributed to the LDOC partition. These high LDOC concentrations

(mean = 3.7 mg/L; range = 0 to 6.1 mg/L) in most cases exceeded the calculated non-algal

LTOC concentration when non-algal LTOC was calculated as TOCmin – RTOC – algal LTOC.

To adjust for this inconsistency, LTOC concentration was increased such that non-algal LTOC =

LDOC, resulting in non-algal LPOC = 0. In cases where the sum of LTOC + RTOC exceeded

DOC + TSS*0.4 (TOCmax; occurring in 12 of the 51 sampling events in between Nov. 1996

through Dec. 1998), RTOC was adjusted downward.

Determination of the Atmospheric Deposition Load

The calculation of the atmospheric deposition load to the lower St. Johns River was performed

by the private consulting firm Tetra Tech, Inc. (Pollman, 2003). Estimates were derived for total

wet and dry deposition of nitrogen and phosphorus species. Wet nitrogen deposition

concentration was determined by the monthly mean of the 3 nearest National Atmospheric

Deposition Program (NADP) sites to the LSJR, FL99 (Kennedy Space Center), FL03 (Bradford

Forest), and GA09 (Okeefenokee Swamp). Volumes were computed as the product of the

monthly Thiessen rainfall polygon rainfall volumes (the same spatial rainfall data set as that used

for watershed modeling) directly falling on the ICM grid surface area. Monthly dry deposition

nitrogen flux estimates were developed using the method of Poor et al. (2001). Phosphorus wet

deposition was determined by the volume weighted mean deposition estimates collected at the

Lake Barco monitoring station, approximately 20 miles west of Palatka.

54

Figure 17. Comparison of POC:Algal Biovolume (a) and POC:Total Chlorophyll a Ratios to

Total Biovolume and Total Chlorophyll a Concentration for LSJR Freshwater

Samples. Red Lines in graph (a) correspond to a C:chlorophyll a of 50:1, and in

graph (b) corresponds to C:biovolume of 0.0003 (for biovolume expressed as

µm3/ml x 1000).

(a ) POC:Total Chlorophyll a

0

20

40

60

80

100

120

140

160

0 100 200 300 400 500

POC:Chlorophyll a ratio

To

tal

Ch

loro

ph

yll

a,

mg

/m3

(b) POC:Algal Biovolume

0

5000

10000

15000

20000

25000

30000

0 0.002 0.004 0.006 0.008 0.01 0.012

C:Biovolume Ratio

Bio

vo

lum

e,

mic

ro-m

3/m

l x

10

00

55

Figure 18. Relationship Between Refractory Dissolved Organic Carbon and Color for

Blackwater Streams of the LSJR Basin. From Hendrickson et al., 2001.

Color = 0.4125(RDOC)2 + 4.633(RDOC)

R2 = 0.9939

0

200

400

600

800

1000

1200

1400

1600

0 10 20 30 40 50 60 70

Mean RDOC by Color Class, mg/L

Co

lor,

Pt-

Co

Un

its

Cuthbert and del

Georgio, 1992

Pulp Mill

Eff luent

Rasmussen et al 1989

56

RESULTS

Tributary Organic Carbon and Nutrients

Calibration Data Set Summary

Calculated seasonal flow-weighted concentrations of organic carbon, nitrogen and phosphorus

forms in field sampling data of tributary runoff are listed in Table 6. Organic carbon in samples

was considered to be primarily allochthonous in origin. Of the 789 total samples available, 570

were also analyzed for chlorophyll a. Of these, 82% had chlorophyll a concentrations less than 5

g/L, suggesting low contributions to the total organic carbon pool of internally-derived,

autochthonous organic carbon. Of the 103 samples with chlorophyll a concentrations above 5

g/L, 71 were obtained at on two tributaries, Sixmile and Peters Creeks, with sampling sites in

quiescent, open embayments below the head of tide, where autochthonous production potential

was high.

Using the BOD partitioning approach, labile total organic carbon (LTOC) was found to comprise

a relatively small fraction of the TOC pool in stream runoff. This is not surprising, as southeast

U.S. coastal plain blackwater streams are by nature high in refractor, colored dissolved organic

matter. The highest absolute concentrations of LTOC were found to occur in tributaries draining

urbanized Jacksonville. These tributaries also exhibited lower TOC concentrations, presumably

do to the increase in impervious surface area and the concomitant reduction in runoff leached

through soils on the flow-path to the stream, as well as a reduction in the amount of natural forest

cover in the basin to supply organic detritus. Catchments with high contributions of dairy land

use exhibited relatively high LDOC concentrations, but had calculated RDOC concentrations

similar to undeveloped watersheds.

57

Table 6. Mean Total, Inorganic, and Calculated Labile and Refractory Organic Nutrient and Carbon Mean Annual Flow-Weighted Concentrations for Tributaries sampled within

the lower St. Johns River Basin. Sorted in order of increasing land use intensity. Values in boxes lacked BOD sampling data and were calculated using simulated

values.

Land Use Percent in Class

Forest Dominated n Urban Dairy

Row

Crop Forest TN TIN

Labile

TON

Refrac.

TON TP TPO4

Labile

TNOP

Refrac.

TONP TOC

Labile

TOC

Refrac.

TOC

Bradley Creek Upstr. 7 | 1 6 0 93 | 0.490393 0.013565 0.085395 0.391433 0.017946 0.01022 0.002321 0.003883 15.42371 0.401154 15.02256

Peters CreekUpstr. 24 | 3 4 1 92 | 0.640559 0.027803 0.155494 0.456867 0.029919 0.013087 0.008189 0.008642 21.23662 0.833601 20.06556

Black Creek No. Fork. 48 | 4 1 1 92 | 0.802675 0.119433 0.195487 0.487386 0.059442 0.026052 0.018489 0.014363 24.08545 1.030466 20.64633

Black Creek So. Fork 44 | 13 5 1 81 | 0.7094 0.054821 0.169452 0.484908 0.115486 0.070977 0.0225 0.021265 23.20996 0.892373 20.70435

Trout Creek 12 | 4 2 8 86 | 1.162648 0.081714 0.342241 0.738693 0.073858 0.027308 0.030129 0.024262 24.89197 1.253613 22.77622

0.76 0.059 0.190 0.512 0.06 0.030 0.016 0.014 21.77 0.88 19.84

Dairy Dominated

Peters Creek Dwnstr. 82 | 4 8 7 80 | 1.073664 0.282677 0.323955 0.467032 0.210378 0.143282 0.04446 0.022637 17.00254 1.325701 15.53925

Bradley Cr. Dwnstr. 20 | 5 24 0 70 | 4.113989 3.439598 0.35236 0.386799 0.093123 0.049962 0.03238 0.010782 11.23281 1.171885 10.06092

Trout River 82 | 22 20 2 56 | 1.77 0.391 0.766 0.618 0.43 0.398 0.023 0.001 27.03 2.29 17.25

Mill Log Br. Dwnstr. 15 | 2 25 0 73 | 1.988097 0.547011 0.651904 0.789182 0.96523 0.76857 0.136757 0.059903 28.76962 2.595772 26.17384

Mill Log Br. Upstr. 19 | 1 54 0 45 | 2.872906 0.801499 0.995905 1.075502 1.715412 1.534645 0.129663 0.051104 32.63283 3.372127 29.26071

Governors Creek 22 | 10 24 0 66 | 3.319618 2.144658 0.685129 0.489831 0.218639 0.157146 0.049441 0.012694 18.46127 2.684303 15.70752

2.52 1.268 0.629 0.638 0.60 0.509 0.069 0.026 22.52 2.24 19.00

Row Crop Dominated

Hastings Drainage Dist. 31 | 2 13 26 59 | 1.808419 0.473851 0.697115 0.63843 0.472736 0.281731 0.141746 0.049259 23.18233 2.693788 21.05455

Moccasin Branch 30 | 7 6 28 59 | 1.86147 0.553435 0.758721 0.547453 0.468344 0.304752 0.128775 0.034816 31.36513 3.420644 22.59039

Deep Creek Dwnstr. 47 | 2 33 30 35 | 1.942408 0.688538 0.480799 0.777022 0.557594 0.420494 0.081452 0.047354 24.37006 1.732058 23.70844

Dog Branch 29 | 12 4 72 13 | 2.500268 1.123014 0.958052 0.570389 0.525746 0.276305 0.202008 0.047432 17.08496 2.863868 13.35062

2.03 0.710 0.724 0.633 0.51 0.321 0.138 0.045 24.00 2.68 20.18

Urban Dominated

Cedar Creek 78 | 25 9 4 62 | 1.021877 0.136099 0.398539 0.487238 0.09942 0.06413 0.024671 0.010619 13.93 1.277845 12.65216

Ortega River 13 | 38 8 5 50 | 0.965428 0.154752 0.279993 0.530465 0.125436 0.064639 0.041893 0.028275 18.45483 1.190488 18.6259

Arlington River 6 | 68 3 2 27 | 1.008321 0.183998 0.549518 0.266673 0.153877 0.077913 0.062494 0.013487 13.58393 2.755236 11.10736

Cedar River Dwnstr. 19 | 67 9 1 24 | 1.166174 0.328866 0.638752 0.195467 0.190901 0.097539 0.086623 0.009593 12.38027 3.446573 8.667477

McCoy's Creek 69 | 89 7 0 4 | 1.659184 0.43134 1.112342 0.115503 0.430823 0.169996 0.251564 0.009263 7.863333 4.075591 3.787743

1.16 0.247 0.596 0.319 0.20 0.095 0.093 0.014 13.24 2.55 10.97

OVERALL AVERAGE 17 6 11 66 1.64 0 .60 0.53 0.53 0.348 0.248 0.076 0.024 20.31 2.07 17.44

58

Based on the empirical partitioning scheme derived here, organic N was on average evenly split

between labile and refractory forms, while non-orthophosphate P tended to predominate in labile

forms. These proportions differed widely based upon the level and type of development within a

sampling basin. For undeveloped basins, with over 90% of their area in natural land covers or

silviculture, LTON was on average 27% of TON. For these same basins, LTNOP was, at 51%,

on average a higher proportion of TNOP. It should though be noted that these streams exhibited

such low concentrations of TP that absolute LTNOP values are very low, and the precision of the

partitioning approach should also be considered low. Conversely, for tributaries draining basins

with over 30% of urban land area, LTON averaged 70% of TON, and LTNOP was on average

86% of TNOP. Despite relatively low TOC concentrations in comparison to streams draining

undeveloped watersheds, urban development-dominated watersheds exhibit some of the highest

concentrations of LTOC, LTON, and LTNOP.

Streams draining agriculture-dominated watersheds also exhibited proportionally high levels of

LTOC, LTON and LTNOP. Within this category, several streams draining watersheds with

large areas of dairy land use exhibited some extremely high concentrations of TP and TN. The

high levels of these nutrients were not consistent within any given stream, suggesting various

unidentified dairy practices influence water quality in different ways. Bradley Creek and

Governor’s Creek exhibited high TN levels, at 6.39 and 5.13 mg/L, respectively, with a large

portion of this as nitrate+nitrite-N, but had relatively low (for dairy runoff) TP concentrations.

Another dairy-influenced stream, Mill Log Branch, had elevated but much lower levels of TN, at

2 mg/L, while mean TP concentration was 1.36 mg/L! This lack of consistent pattern with

respect to elevated nutrient concentrations led to difficulties in establishing model concentration

parameters.

Watershed Model Calibration

Hydrologic and water quality coefficients were calibrated separately and together (as load) to

provide a relative sense of their individual and combined precision. Hydrologic calibration was

59

discussed in the previous methods section, therefore only the quality and load calibration are

covered here. As TN, TIN, TP, and PO4 were calibrated in the first LSJR development of the

PLSM (Hendrickson and Konwinski, 1998) and remain the same in this application, the results

can be viewed as a verification of the established model coefficients. For the newly created

constituents TOC, RTOC and LTOC, this comparison represents an assessment of the

performance of the coefficient assignment process through the multiple regression approach.

Figures 19, 20 and 21 compare the seasonal measured and simulated flow-weighted

concentrations of relevant water quality variables. Large differences are evident for monitoring

stations downstream of dairy land uses (shown as open squares in Figures 19 - 21. Part of this

appeared to be due to the inconsistencies identified in the previous section with regard to the

appearance of N or P in runoff. However, the consistent under-prediction suggests that either 1)

water quality concentrations assigned to improved pasture and confined animal feeding zones are

too low; 2) runoff coefficients assigned to these areas are too low; or 3) the area dedicated to

dairy is under-reported in the GIS land use layer. As the present PLSM nutrient concentrations

assigned to improved pasture and concentrated feeding areas range from 3.0 - 3.9 mg/L and 4.5 –

6.0 mg/L for TN and from 0.75 – 2.0 mg/L and 1.3 – 2.6 mg/L for TP, considerably greater that

the 2.48 TN and 0.349 TP mg/L mean concentrations reported by Harper (1994) for central and

south Florida, it does not appear at first that model concentrations are too low. After some

inspection, it was discovered that the GIS land use data in many cases either incorrectly coded

land cover associated with dairy, identifying intensive pasture or manure sprayfields as row crop

or miscellaneous agriculture, or failed to account for it at all. Also, as the PLSM in its present

form assigns the same runoff coefficients to pasture and dairy as undeveloped land use, this too

may be a source of under-representation. Darkened squares in Figures 19 - 21 identify these

simulated values for these same watersheds derived by assigning all land are under the

“miscellaneous agriculture” category to dairy, and increasing runoff coefficients to the same

values observed for the tri-county agricultural area. These modifications appear to bring organic

nutrient concentrations more in line with observed data, but still leave simulated inorganic

nutrient concentrations (and hence, total nutrient) low. Because these manipulations are

hypothetical, the sampling stations within the three watersheds with substantial dairy

60

development (Governors Creek, Mill Log Branch and Bradley Creek) are excluded from the

quantitative calibration described below.

61

Figure 19. Comparison of Observed to Simulated Flow-Weighted Concentrations of Carbon,

Nitrogen and Phosphorus Forms for the December through March Season. All units

in mg/L. Open squares represent existing dairy watershed predictions, and solid

squares represent miscellaneous agriculture added to dairy land and runoff

coefficients increased to match tri-county ag area row crop values.

Total N

0

1

2

3

4

5

0 1 2 3 4 5Observed

Sim

ula

ted

Total Inorganic N

0

0.5

1

1.5

2

2.5

3

3.5

0 0.7 1.4 2.1 2.8 3.5

ObservedS

imu

late

d

Total P

0

0.4

0.8

1.2

1.6

2

0 0.5 1 1.5 2

Observed

Sim

ula

ted

Total PO4

0

0.3

0.6

0.9

1.2

1.5

1.8

0 0.3 0.6 0.9 1.2 1.5 1.8

Observed

Sim

ula

ted

Total Organic C

0

5

10

15

20

25

30

35

40

45

0 5 10 15 20 25 30 35 40 45

Observed

Sim

ula

ted

Total Suspended Solids

0

10

20

30

40

0 10 20 30 40

Observed

Sim

ula

ted

62

Figure 19 (cont.)

Labile Total Organic C

0

1

2

3

4

5

6

0 2 4 6Observed

Sim

ula

ted

Refractory Total Organic C

0

5

10

15

20

25

30

35

40

0 10 20 30 40

Observed

Sim

ula

ted

Labile Total Organic N

0

0.25

0.5

0.75

1

1.25

1.5

0 0.5 1 1.5

Observed

Sim

ula

ted

Refractory Total Organic N

0

0.25

0.5

0.75

1

1.25

1.5

0 0.5 1 1.5Observed

Sim

ula

ted

Labile Total Non-PO4P

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0 0.1 0.2 0.3 0.4

Observed

Sim

ula

ted

Refractory Total Non-PO4P

0

0.02

0.04

0.06

0.08

0.1

0 0.02 0.04 0.06 0.08 0.1

Observed

Sim

ula

ted

63


Nitrogen and Phosphorus Forms for the April through July Season. All units in

mg/L. Open squares represent existing dairy watershed predictions, and solid

squares represent miscellaneous agriculture added to dairy land and runoff


Total N

0

1

2

3

4

5

0 1 2 3 4 5

Observed

Sim

ula

ted

Outlier:Dog BranchSim. = 2.8Obs. = 7.2

Total Inorganic N

0

0.5

1

1.5

2

2.5

3

3.5

0 0.7 1.4 2.1 2.8 3.5

ObservedS

imu

late

d

Total P

0

0.4

0.8

1.2

1.6

2

0 0.5 1 1.5 2

Observed

Sim

ula

ted

Total PO

0

0.3

0.6

0.9

1.2

1.5

1.8

0 0.3 0.6 0.9 1.2 1.5 1.8

Observed

Sim

ula

ted

Total Organic C

0

5

10

15

20

25

30

35

40

45

0 5 10 15 20 25 30 35 40 45

Observed

Sim

ula

ted


0

10

20

30

40

0 10 20 30 40

Observed

Sim

ula

ted

64

Figure 20 (cont.)


0

1

2

3

4

5

6

0 1 2 3 4 5 6Observed

Sim

ula

ted


0

5

10

15

20

25

30

35

40

0 10 20 30 40

Observed

Sim

ula

ted


0

0.25

0.5

0.75

1

1.25

1.5

0 0.5 1 1.5Observed

Sim

ula

ted

Dog Br.Calc = 1.0Sim = 2.9


0

0.25

0.5

0.75

1

1.25

1.5

0 0.5 1 1.5Observed

Sim

ula

ted


0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0 0.1 0.2 0.3 0.4Observed

Sim

ula

td

Dog Br.

Calc = 0.2

Sim = 0.7


0

0.02

0.04

0.06

0.08

0.1

0 0.02 0.04 0.06 0.08 0.1

Observed

Sim

ula

ted

65


Nitrogen and Phosphorus Forms for the August through November Season. All

units in mg/L. Open squares represent existing dairy watershed predictions, and

solid squares represent miscellaneous agriculture added to dairy land and runoff


Total N

0

1

2

3

4

5

0 1 2 3 4 5

Observed

Sim

ula

ted

Total Inorganic N

0

0.5

1

1.5

2

2.5

3

3.5

0 1 2 3Observed

Sim

ula

ted

Total P

0

0.4

0.8

1.2

1.6

2

0 0.5 1 1.5 2

Observed

Sim

ula

ted

Total PO4

0

0.3

0.6

0.9

1.2

1.5

1.8

0 0.3 0.6 0.9 1.2 1.5 1.8

Observed

Sim

ula

ted

Total Organic C

0

5

10

15

20

25

30

35

40

45

0 5 10 15 20 25 30 35 40 45

Observed

Sim

ula

ted


0

9

18

27

36

45

0 10 20 30 40

Observed

Sim

ula

ted

66

Figure 21 (cont.)


0

1

2

3

4

5

6

0 1 2 3 4 5 6

Observed

Sim

ula

ted


0

5

10

15

20

25

30

35

40

45

0 10 20 30 40 50

Observed

Sim

ula

ted


0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0 0.5 1 1.5 2

Observed

Sim

ula

ted


0

0.2

0.4

0.6

0.8

1

1.2

1.4

0 0.25 0.5 0.75 1 1.25 1.5

Observed

Sim

ula

ted


0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0 0.1 0.2 0.3 0.4

Observed

Sim

ula

ted


0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1

0 0.02 0.04 0.06 0.08 0.1

Observed

Sim

ula

ted

67

Two statistical tests were performed to characterize the level of agreement between the

calculated flow-weighted concentrations and the PLS model simulated, runoff-weighted

concentrations, and the results of these are shown in Table 7. Pearson product-moment

correlations were calculated to estimate general agreement between observed and simulated

values, and zero-intercept regression slopes and confidence intervals, in which the observed

flow-weighted concentration represented the independent variable, were calculated to assess

potential bias in model water quality coefficients. Variables exhibiting the poorest correlations

between the calculated flow weighted and simulated values include TPO4 for the December –

March season, RTON for the April – July.

68

Table 7. Pearson Correlations, Slopes, and Confidence Intervals of the Slopes for Intercept-Fit Regressions Between Calibration

Station Measured Flow-Weighted Concentrations and Contributing Area Modeled Runoff-Weighted Concentrations. Boxed

confidence bounds indicate seasons and constituents for which agreement between measured (independent) and modeled

(dependent) concentrations are significantly different than 1:1. Parameter TN TIN LTON RTON TP TPO4 LTNOP RTNOP TOC LTOC RTOC TSS

December - March

Pearson Correlation 0.656 0.796 0.73 0.754 0.638 0.399 0.7 0.907 0.62 0.846 0.6 0.657

p-value 0.004 0.000 0.001 0.000 0.006 0.113 0.002 0.000 0.008 0.000 0.011 0.004

Slope 0.742 0.400 0.982 0.870 0.625 0.545 0.611 0.521 1.095 0.986 0.992 0.870

Slope Standard Error 0.066 0.051 0.104 0.056 0.080 0.102 0.088 0.050 0.057 0.076 0.074 0.100

Slope Upper 95% C.I. 0.601 0.292 0.762 0.751 0.454 0.328 0.424 0.416 0.973 0.824 0.836 0.657

Slope Lower 95% C.I. 0.882 0.508 1.201 0.989 0.795 0.761 0.798 0.627 1.216 1.148 1.149 1.082

April - July


p-value 0.000 0.000 0.007 0.164 0.000 0.000 0.003 0.000 0.008 0.003 0.006 0.001

Slope 1.625 1.800 1.315 1.390 1.517 0.922 1.837 2.307 1.150 1.106 1.410 1.421


Slope Upper 95% C.I. 1.284 1.513 0.928 0.934 1.171 0.702 1.219 1.396 0.976 0.896 1.146 1.058

Slope Lower 95% C.I. 1.965 2.088 1.702 1.846 1.862 1.142 2.455 3.217 1.324 1.316 1.674 1.784

August - November


p-value 0.002 0.025 0.002 0.097 0.000 0.000 0.003 0.650 0.001 0.168 0.001 0.218

Slope 0.764 0.651 0.630 0.783 0.722 0.701 0.896 0.861 0.853 0.570 0.894 0.661


Slope Upper 95% C.I. 0.633 0.428 0.478 0.564 0.523 0.497 0.612 0.347 0.726 0.375 0.750 0.361

Slope Lower 95% C.I. 0.895 0.873 0.782 1.001 0.921 0.905 1.179 1.375 0.980 0.764 1.039 0.962

Overall


p-value 0.000 0.000 0.000 0.014 0.000 0.000 0.000 0.237 0.000 0.000 0.000 0.000

Slope 1.020 1.019 0.924 0.920 0.844 0.715 1.162 0.871 0.987 0.808 1.006 0.929


Slope Upper 95% C.I. 0.857 0.799 0.757 0.772 0.695 0.601 0.894 0.580 0.905 0.684 0.902 0.742

Slope Lower 95% C.I. 1.182 1.239 1.091 1.068 0.993 0.829 1.430 1.162 1.070 0.933 1.110 1.117

69

season, RTON, RTNOP, LTOC and TSS for the August – November season, and RTNOP for the

overall, all season comparison. In all other comparisons, Pearson correlation p-values are less

than 0.05. In 22 of the possible 36 (3 seasons x 12 constituents) seasonal comparisons, the slope

of the regression line relating the observed, flow-weighted concentration to the simulated value

was found to be significantly greater than or less than 1, suggesting some bias in the model

seasonal water quality coefficients. Many of the seasonal biases are compensating, and when all

seasons are combined, 9 of the 12 constituent slope confidence intervals contain 1, and the

remaining 3, TP, PO4 and LTOC, upper confidence bounds are close to 1. Seasonal oscillation

in observed verses simulated slopes in several cases results from a few data points belonging to

streams (principally Dog Branch and Deep Creek) draining the tri-county agricultural area that

exhibit much higher concentrations and thus heavily influence the regression line. When high

flow samples from these streams, influential in the flow-weighting calculation, were recorded

near the end of the December – March WQ coefficient seasonal break, observed concentrations

are high relative to model predictions, producing a observed x simulated slope of less that 1.

In most cases, the newly-added model constituents for TOC, LTOC and RTOC perform well,

however, it should be kept in mind that in the case of these constituents the same data set was

used for assigning land use water quality coefficients and calibration, so statistics only reflect the

skill in assigning coefficients.

To assess the accuracy of load prediction (the product of concentration and discharge volume),

daily loads calculated at water quality sampling stations that also have established stream-flow

gauging installations were compared to watershed model simulated daily loads. Actual daily

load was assumed to be equivalent to the product of the instantaneous water quality sampling

concentration and the daily discharge. Because water quality concentrations can vary within a

day, this should be considered to be an estimate of daily load. Simulated daily load was

calculated as the seasonal, watershed area-weighted constituent concentration x seasonal water

quantity x the daily flow fraction. Daily flow fraction is the fraction of the given daily flow of

the total seasonal flow, and was the statistic used to disaggregate seasonal watershed model loads

to daily loads for input to the LSJR water quality processes model.

70

Three tributary stations in the LSJR meet the criteria of having both water quality and quantity

monitoring, and comparison of their simulated and observed daily loads are shown in Figure 21.

The plots demonstrate the tendency for baseflow conditions to be over-sampled relative to

stormflow or high-flow conditions in fixed-interval ambient monitoring data. This phenomena is

manifested in the plots by many sampling points clustered near the origin, with few high-flow

(and thus high load) data points to discern the relationship between observed and simulated.

Considering the uncertainty associated with estimating the true value of the daily load, and with

fractionating a daily load from the watershed model seasonal load, simulated and observed daily

loads are in good agreement. Of particular interest are the dual data points of Figure 21. Open

diamonds represent the comparison using the North Fork’s daily flow fractions to disaggregate

the seasonal load, while the solid triangles represent the simulated daily load disaggregated with

the daily flow fractions from the South Fork. This latter condition is a closer representation to

the approach used for most sub-basins, which did not have flow gauging stations, and were

disaggregated with nearest-neighbor daily flow fractions. In general, inorganic nutrient fractions

appear to be more variable than total, and larger, homogeneous watersheds (exemplified by

South Fork Black Creek) have better agreement than heterogeneous watersheds with varying

land uses and management that leads to variations in pollution runoff (exemplified by Deep

Creek).

Point Source Organic Carbon and Nutrients

Mean point source effluent discharge volumes and water quality concentrations for 1997-98 are

summarized in Table 8. On average, 135.4 million gallons per day of treated domestic waste,

and 59.9 million gallons per day of treated industrial waste were discharged during that time

period, with the majority of this entering the oligohaline and mesohaline portions of the river

north of Julington Creek.

71

Table 8. Summary of Point Source Mean Effluent Water Quality Concentrations

Facility Name

Mean

Monthly

Discharge

(MGD)

CBOD

(mg/l)

TIN

(mg/L)

TN

(mg/L)

LTON

(mg/L)

RTON

(mg/L)

Ortho-P

(mg/l)

Total P

(mg/l)

Labile,

Non-

ortho P

(mg/L)

Refract.

Non-

ortho P

(mg/L)

TOC

(mg/l)

LTOC

(mg/L)

RTOC

(mg/L)

TSS

(mg/l)

Inorg. SS

(mg/L)

Buckman Street 32.5 11.7 8.869 11.617 2.477 0.271 3.652 4.621 0.938 0.031 17.77 14.01 3.76 20.08 2.46

Arlington East 10.8 8.6 12.157 14.492 2.229 0.106 2.072 2.593 0.515 0.006 12.40 10.26 2.14 14.11 1.18

Southwest 5.8 5.1 7.214 10.469 2.748 0.507 1.226 1.436 0.200 0.009 12.71 5.99 6.72 18.36 5.23

Mandarin 4.8 4.7 8.962 10.835 1.554 0.319 2.022 2.458 0.410 0.026 10.99 5.54 5.45 4.78 0.10

District II 4.3 2.6 21.867 23.601 1.325 0.409 4.537 5.788 1.132 0.119 10.00 3.05 6.95 6.76 0.08

Miller St. 3.4 3.3 3.627 4.343 0.622 0.094 2.022 2.226 0.193 0.010 8.56 3.82 4.74 6.57 0.00

Jacksonville Beach Average 3.1 2.5 7.368 9.664 1.700 0.596 1.779 2.136 0.320 0.036 11.27 2.89 8.38 2.73 0.00

Monterey 3.0 2.4 9.866 11.021 0.890 0.266 1.488 1.908 0.383 0.037 9.55 2.72 6.84 9.38 0.00

City of Palatka 2.8 6.1 15.193 16.538 1.237 0.073 2.251 2.409 0.155 0.003 9.86 7.23 2.64 7.27 0.00

Royal Lakes 2.3 4.4 6.600 7.800 1.112 0.088 2.833 3.564 0.713 0.018 8.65 5.22 3.43 6.55 0.00

San Jose 2.1 3.9 10.782 12.285 1.324 0.179 1.700 2.000 0.287 0.013 9.48 4.61 4.87 7.30 0.00

Atlantic Beach 1.7 2.1 8.859 10.136 0.990 0.288 1.518 1.845 0.300 0.028 8.87 2.44 6.43 2.64 0.00

Orange Park 1.4 2.6 9.936 10.391 0.407 0.048 1.859 2.688 0.798 0.031 5.82 3.05 2.77 3.21 0.00

Julington Creek 1.2 2.8 7.400 7.800 0.365 0.035 2.299 2.839 0.525 0.016 5.66 3.29 2.38 2.19 0.00

Jacksonville Heights 1.2 1.1 9.060 10.000 0.589 0.351 2.285 2.596 0.262 0.049 7.83 1.21 6.61 0.45 0.00

USN Naval Air Station 1.1 2.4 10.887 11.883 0.766 0.230 1.192 1.350 0.143 0.015 8.07 2.78 5.29 1.61 0.00

Atlantic Beach - Buccanneer 1.0 5.1 8.814 10.817 1.845 0.159 1.839 2.259 0.409 0.011 10.66 6.00 4.66 3.11 0.00

Mayport Naval Air Station Average 1.0 5.6 6.727 7.937 1.078 0.131 1.620 1.903 0.272 0.011 10.61 6.67 3.94 28.52 13.48

Neptune Beach 0.9 4.9 7.254 8.285 0.988 0.043 1.618 1.934 0.311 0.004 8.13 5.87 2.26 3.95 0.00

Beacon Hills 0.8 2.8 9.933 11.335 1.160 0.242 1.821 2.215 0.369 0.025 9.21 3.24 5.98 3.71 0.00

Anheuser Busch Land App. 0.7 8.2 2.889 5.297 2.353 0.054 3.132 3.976 0.839 0.005 11.58 9.79 1.79 21.93 0.00

Fleming Island 0.6 3.9 2.530 6.100 2.319 1.251 2.185 2.660 0.391 0.083 14.24 4.56 9.68 2.52 0.00

Holly Oaks 0.5 3.7 7.505 8.774 1.105 0.164 2.433 2.686 0.241 0.011 9.66 4.29 5.37 11.43 0.00

San Pablo 0.5 1.1 5.200 6.350 0.701 0.449 2.268 2.810 0.450 0.092 8.50 1.25 7.25 1.72 0.00

Green Cove Springs 0.5 3.0 9.351 11.690 1.850 0.489 1.915 2.371 0.420 0.036 11.38 3.52 7.87 7.16 0.00

Woodmere 0.3 3.0 10.496 12.396 1.499 0.401 1.203 1.402 0.181 0.018 10.43 3.46 6.97 6.06 0.00

Brierwood S/D - Beauclerc 0.3 2.9 6.697 7.778 0.967 0.115 1.447 1.730 0.273 0.010 7.16 3.40 3.76 4.70 0.00

Fleming Oaks 0.3 4.0 2.586 3.201 0.534 0.080 1.890 2.283 0.373 0.020 7.62 4.69 2.93 1.90 0.00

South Green Cove Springs 0.3 2.5 11.745 12.434 0.599 0.090 1.852 2.229 0.359 0.017 6.90 2.99 3.91 2.95 0.00

St. Johns North - United Water 0.2 5.5 11.043 11.240 0.196 0.000 2.390 2.977 0.587 0.000 6.63 6.59 0.04 11.10 0.00

Orteg Hills 0.1 2.2 16.600 17.601 0.803 0.197 1.795 2.153 0.332 0.026 8.03 2.55 5.47 2.50 0.00

City of Hastings 0.1 5.5 6.764 9.330 2.111 0.455 0.513 0.560 0.043 0.004 12.21 6.53 5.68 11.80 0.00

Wesley Manor Retirement Village 0.1 5.1 3.776 6.094 2.108 0.209 1.276 1.505 0.222 0.007 11.32 6.08 5.24 5.02 0.00

Domestic Waste Average 135.4 4.1 8.744 10.289 1.289 0.254 1.998 2.427 0.404 0.025 9.75 4.84 4.91 7.40 0.68

%Bioav. TN %Bioav. TP

Georgia Pacific Corp. 34.4 10.8 1.419 4.427 1.654 1.354 0.546 1.136 0.464 0.126 93.31 11.68 81.63 15.62 0.00 69.4142877 88.9007792

Stone Container Corp. 8.8 18.5 1.112 6.038 4.163 0.762 0.493 0.911 0.394 0.024 51.45 21.73 29.73 23.91 0.00 87.3726384 97.4117881

Jefferson Smurfit Corp. 5.4 30.5 8.567 14.135 4.802 0.767 0.959 1.497 0.508 0.030 65.34 36.06 29.28 125.22 3.86 94.5744953 97.9804699

Industrial Waste Average 59.9 20.0 3.699 8.200 3.540 0.961 0.666 1.181 0.455 0.060 70.03 23.16 46.88 54.92 1.29

72

In contrast to tributary surface waters, inorganic nutrients are the predominant form of domestic

waste effluent TN and TP concentrations. Based upon the BOD-partitioning of the organic

nutrient fraction, domestic waste exhibits a high proportion in the labile form, on average 84%

for LTON and 87% for LTNOP.

Pulp and paper processing waste is the only industrial effluent (excluding dairy runoff)

discharged directly to the lower St. Johns River. During the TMDL and PLRG water quality

model calibration time period, three pulp and paper mills were operating within the basin,

however, since this time one has ceased operation (Jefferson Smurfit), and another, Georgia

Pacific, has considerably reduced its effluent volume. Thus it should be understood that these

statistics are in transition, and relevant only for calibration of the water quality model and for the

starting point in subsequent load allocation calculations.

The most noteworthy characteristic of pulp and paper process waste is the very high levels of

TOC. While largely refractory, LTOC concentrations are quite high, suggesting high potential

oxygen demand. These effluents also appear to carry a large proportion of their organic nutrient

forms in the labile fraction, with 78% of TON as LTON, and 88% of TNOP as LTNOP.

Upstream Concentrations of Organic Carbon and Nutrients

Nutrients and organic carbon entering the lower St. Johns (Figure 14) exhibit an annual pattern in

compartmentalization linked to algal production and upstream runoff. Lake George is the

predominant hyro-morphological feature that shapes the upstream load entering at Buffalo Bluff,

and the co-occurrence of increased residence time and increased day length and temperature in

the spring and early summer transforms the river from an allochthonous (external carbon-

supplied) system to one which is principally autochthonous (internally-generated carbon supply).

This pattern results in a sequence of nutrient sequestration within the lake that starts with

inorganic nitrogen disappearance in the spring, leading quickly to nitrogen-fixing blue green

algal dominance.

73

Figures 22, 23, and 24 show the patterns in N, P and organic C form concentrations at Buffalo

Bluff and Dunns Creek from 1995 through 1999. Algal organic carbon concentrations increase

each year in May, and remain elevated through August. Peak algal organic carbon

concentrations coincide with refractory organic carbon (largely composed of allochthonous

colored dissolved organic matter) annual minimums. Annual peak algal organic carbon

concentrations range between 2.4 mg/L (June 1995) to 5.9 mg/L (July 1999) at Buffalo Bluff,

and between 2.0 mg/L (July 1995) and 4.0 mg/L (May 1999) at Dunns Creek. Low inorganic N

concentrations exiting Lake George and at Buffalo Bluff suggest a condition of low N stress on

the algal community is a normal occurrence by late spring. If this is the case, then the drawdown

of non-algal labile N that occurs in May and June may be an important regulator of algal growth

at this time.

By June of most years, evidence for blue-green algal nitrogen fixation can be seen at Buffalo

Bluff. These can be sees as peaks in non-algal LTON, absent of flow events that would be

expected to import allochthonous organic N. In 1996, the summer total Kjeldahl nitrogen peak is

2.05 mg/L, occurring on June 12; in 1997, it is 2.32 mg/L on July 23; and in 1999, it is 2.8 mg/L

on May 25. Peaks in LTON in Figure 22 near these events suggest either low C:N ratios due to

luxury acquisition of nitrogen (in which case this should more correctly be attributed to the algal

N compartment), exudation of labile N by the actively-growing N-fixing community, or the

presence of a large amount of algal detritus.

74

Figure 22. Partitioned Nitrogen Concentrations at (a) Buffalo Bluff and (b) Dunns Creek, Dec.

1994 - Nov. 1999.

(a) Buffalo Bluff

0.000

0.200

0.400

0.600

0.800

1.000

1.200

1.400

1.600O

ct-

94

Jan

-95

Ap

r-95

Jul-

95

Oct-

95

Jan

-96

Ap

r-96

Jul-

96

Oct-

96

Jan

-97

Ap

r-97

Jul-

97

Oct-

97

Jan

-98

Ap

r-98

Jul-

98

Oct-

98

Jan

-99

Ap

r-99

Jul-

99

Oct-

99

Jan

-00

Co

nc

en

tra

tio

n, m

g/L

as

N

Non-Algal LTON

Algal ON

TIN

RTON

(b) Dunns Creek

0.000

0.200

0.400

0.600

0.800

1.000

1.200

1.400

1.600

Oct-

94

Jan

-95

Ap

r-95

Jul-

95

Oct-

95

Jan

-96

Ap

r-96

Jul-

96

Oct-

96

Jan

-97

Ap

r-97

Jul-

97

Oct-

97

Jan

-98

Ap

r-98

Jul-

98

Oct-

98

Jan

-99

Ap

r-99

Jul-

99

Oct-

99

Jan

-00

Co

ncen

trati

on

, m

g/L

as N

Non-Algal LTON

Algal ON

TIN

RTON

75

Figure 23. Partitioned Phosphorus Concentrations at (a) Buffalo Bluff and (b) Dunns Creek,

1995 - 1999.

(a) Buffalo Bluff

0.000

0.010

0.020

0.030

0.040

0.050

0.060

0.070

0.080

Oct-

94

Jan

-95

Ap

r-95

Jul-

95

Oct-

95

Jan

-96

Ap

r-96

Jul-

96

Oct-

96

Jan

-97

Ap

r-97

Jul-

97

Oct-

97

Jan

-98

Ap

r-98

Jul-

98

Oct-

98

Jan

-99

Ap

r-99

Jul-

99

Oct-

99

Jan

-00

Co

nc

en

tra

tio

n, m

g/L

as

P

Non-Algal LTNOP

Algal OP

PO4

RTNOP4/11/95, 0.102

5/2/95, 0.107

(b) Dunns Creek

0.000

0.010

0.020

0.030

0.040

0.050

0.060

0.070

0.080

Oct-

94

Jan

-95

Ap

r-95

Jul-

95

Oct-

95

Jan

-96

Ap

r-96

Jul-

96

Oct-

96

Jan

-97

Ap

r-97

Jul-

97

Oct-

97

Jan

-98

Ap

r-98

Jul-

98

Oct-

98

Jan

-99

Ap

r-99

Jul-

99

Oct-

99

Jan

-00

Co

nc

en

tra

tio

n, m

g/L

as

P

Non-Algal LTNOP

Algal OP

PO4

RTNOP10/18/95,

0.088

76

Figure 24. Partitioned Organic Carbon Concentrations at Buffalo Bluff and Dunns Creek, 1995

- 1999.

(a) Buffalo Bluff

0

3

6

9

12

15

18

21

24

27

30

33

Oct-

94

Jan

-95

Ap

r-95

Ju

l-95

Oct-

95

Jan

-96

Ap

r-96

Ju

l-96

Oct-

96

Jan

-97

Ap

r-97

Ju

l-97

Oct-

97

Jan

-98

Ap

r-98

Ju

l-98

Oct-

98

Jan

-99

Ap

r-99

Ju

l-99

Oct-

99

Jan

-00

Co

nce

ntr

ati

on

, m

g/L

as

C

Non-Algal LTOC

Algal OC

RTOC

(b) Dunns Creek

0

3

6

9

12

15

18

21

24

27

30

33

Oct-

94

Jan

-95

Ap

r-95

Ju

l-95

Oct-

95

Jan

-96

Ap

r-96

Ju

l-96

Oct-

96

Jan

-97

Ap

r-97

Ju

l-97

Oct-

97

Jan

-98

Ap

r-98

Ju

l-98

Oct-

98

Jan

-99

Ap

r-99

Ju

l-99

Oct-

99

Jan

-00

Co

nce

ntr

ati

on

, m

g/L

as

C

Non-Algal LTOC

Algal OC

RTOC

77

Algal organic carbon at Buffalo Bluff as a fraction of TOC peaks during the months of July and

August, when algal biomass accounts for roughly 20 to 25% of water column TOC, and RTOC

reaches its annual minimum (Figure 24). A decrease in allochthonous OC import, dilution with

artesian spring flow, and perhaps photodecomposition, are suspected to be the principle factors

leading to the annual summer decline in RTOC. On average, total labile OC (algal plus non-

algal labile OC) represents 23 percent of the total organic carbon concentration entering the

lower St. Johns at Buffalo Bluff. In contrast, LTOC concentration was on average 10 percent of

TOC exiting the Crescent Lake Basin (Dunns Creek).

While not shown in Figure 24, the further partitioning of Buffalo Bluff LTOC into dissolved

(LDOC) and particulate (LPOC) forms resulted in assigning nearly all of the non-algal labile

carbon in the dissolved form, as a consequence of maintaining LDOC = DOC – RDOC. It

cannot be determined if this exists as an artifact of the carbon partitioning process, or represents

a real phenomena. The presence of high concentrations of LDOC might be possible if the source

of the LDOC was algal exudates associated with high levels of algal production. Between July

and September, peak Buffalo Bluff LTOC (algal + non-algal) concentration ranges between 5

and 9 mg/L, exceeding the maximum concentrations calculated for any of the tributary sampling

stations used in the development of watershed model water quality coefficients by 2-fold. In

comparison to the regression-determined specific land-use loading rate LTOC concentrations,

the Buffalo Bluff LTOC is comparable to runoff from the highly urbanized tributaries of

Jacksonville and northern Clay County. These high LTOC concentrations should be viewed in

context with the low dissolved oxygen concentrations in this reach of the lower St. Johns, which

can exhibit frequent and prolonged excursions below 5 mg/L between July and October. From

this analysis, it appears that the capability for generation of LTOC from river autochthonous

production far exceeds that of terrestrial export in watershed runoff.

Incoming nutrients and organic carbon to the lower St. Johns River from Dunns Creek appears to

resemble that of a large blackwater stream, with algal productivity in the upstream Crescent Lake

exerting less influence over carbon and nutrient partitioning that is observed for Lake George.

RTOC concentrations of Dunns Creek are much higher than that of Buffalo Bluff, exceeding 30

78

mg/L in the aftermath of the 1997-98 El Nino winter. In March of 1998, a color measurement of

800 Pt Co units was recorded, however, this value stands alone as an outlier, and most color

measurements during this period were between 400 – 500 Pt-Co units – still very highly colored.

In the dry-down following the El Nino 1997-98 winter, RTOC levels fall to a low of around 10

mg/L in September. Mean bio-available TP (PO4 + LTNOP) concentration of Dunns Creek is

similar to that of Buffalo Bluff, 0.057 mg/L, compared to0.056 mg/L. Mean total bio-available

nitrogen (TIN + LTON), at 0.732 mg/L, is somewhat less than that for Buffalo Bluff, which was

1.072 mg/L.

Figures 24 – 26 compare 1995-99 time series loads entering the LSJR from Buffalo Bluff and

Dunns Creek. Because of the much greater flow from the mid and upper St. Johns and

Ocklawaha Rivers, the Buffalo Bluff load dominates the total upstream load to the LSJR. From

April 1998 through November 1999, the net discharge from Dunns Creek was zero, resulting in

no net load to the LSJR.

79

Figure 25. Loads of Nitrogen Forms Entering the Lower St. Johns River at Buffalo Bluff and

Dunns Creek, 1995-99.

(a) Buffalo Bluff

-10

-5

0

5

10

15

20

25

30

35

40

Oct-

94

Jan

-95

Ap

r-9

5

Ju

l-95

Oct-

95

Jan

-96

Ap

r-9

6

Ju

l-96

Oct-

96

Jan

-97

Ap

r-9

7

Ju

l-97

Oct-

97

Jan

-98

Ap

r-9

8

Ju

l-98

Oct-

98

Jan

-99

Ap

r-9

9

Ju

l-99

Oct-

99

Jan

-00

Me

an

Da

ily L

oa

d,

MT

Non-Algal LTON

Algal ON

Total Inorganic N

Refractory TON

(b) Dunns Creek

-10.0

-5.0

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

40.0

Oct-

94

Jan

-95

Ap

r-95

Ju

l-95

Oct-

95

Jan

-96

Ap

r-96

Ju

l-96

Oct-

96

Jan

-97

Ap

r-97

Ju

l-97

Oct-

97

Jan

-98

Ap

r-98

Ju

l-98

Oct-

98

Jan

-99

Ap

r-99

Ju

l-99

Oct-

99

Jan

-00

Me

an

Da

ily L

oa

d,

MT

Non-Algal LTON

Algal ON

Total Inorganic N

Refractory TON

80

Figure 26. Loads of Phosphorus Forms Entering the Lower St. Johns River at Buffalo Bluff and

Dunns Creek, 1995-99.

(a) Buffalo Bluff

-0.25

0.00

0.25

0.50

0.75

1.00

1.25

1.50

Oct-

94

Jan

-95

Ap

r-95

Ju

l-95

Oc

t-95

Jan

-96

Ap

r-96

Ju

l-96

Oc

t-96

Jan

-97

Ap

r-97

Ju

l-97

Oct-

97

Jan

-98

Ap

r-98

Ju

l-98

Oct-

98

Jan

-99

Ap

r-99

Ju

l-99

Oct-

99

Jan

-00

Me

an

Da

ily L

oa

d,

MT

Non-Algal LTNOP

Algal OP

Diss. PO4

Refractory TNOP

12/6/94, 1.5

12/21/94, 1.9 11/5/95, 1.6

(b) Dunns Creek

-0.25

0.00

0.25

0.50

0.75

1.00

1.25

1.50

Oct-

94

Jan

-95

Ap

r-95

Ju

l-95

Oct-

95

Jan

-96

Ap

r-96

Ju

l-96

Oct-

96

Jan

-97

Ap

r-97

Ju

l-97

Oct-

97

Jan

-98

Ap

r-98

Ju

l-98

Oct-

98

Jan

-99

Ap

r-99

Ju

l-99

Oct-

99

Jan

-00

Me

an

Daily L

oad

, MT

Non-Algal LTNOP

Algal OP

Total PO4

Refractory TNOP

81

Reconstruction Of The Upstream Natural Background Load

To provide the endpoint for the continuum in water quality responses to nutrient enrichment, it is

necessary to estimate or reconstruct the characteristics of the natural background load. It should

be emphasized that achieving the natural background condition is not the objective of the TMDL

process per se, though it is assumed to be a desirable condition. Rather, the quantification of this

condition anchors one end of the water quality impairment – development continuum and allows

for the interpolation to the point of water quality criteria adherence and hence tolerable

anthropogenic load. Estimation of the natural background condition is also desirable as it helps

in envisioning the estuarine ecosystem prior to human impacts, which can help define restoration

goals and set realistic water quality targets.

To estimate the natural background load of nitrogen, phosphorus and organic carbon that enters

the lower St. Johns from within the immediate (i.e., lower St. Johns) basin, the watershed

modeling approach use was followed, with all present-day developed land used reverted to

natural forest cover. Because similar modeling has not been performed for the Ocklawaha and

St. Johns River basins upstream of the lower St. Johns, and perhaps more importantly, water

quality modeling has not been performed to estimate the transformation of this load that would

occur in the numerous, large upstream lakes, estimation of the upstream boundary natural

background load is more difficult. In the following analysis, the load delivered to the inlet of the

lower St. Johns Basin is estimated through a simple source-proportionality approach and by

applying a Vollenweider model phosphorus settling term for Lake George. This estimate is then

corroborated by comparison to available historic water quality data.

Natural Background Concentrations of Small Order, Undeveloped Streams

Nitrogen, phosphorus and carbon forms engage in almost continuous flux in surface waters from

the moment of dissolution or detachment from the terrestrial environment. In streams

contributing to the St. Johns River, the time between export from the terrestrial environment to

82

the river’s predominantly lacustrine environment is relatively short (when compared to the river),

on the order of hours to days. The lower light availability typical of stream environments limits

photosynthesis, nutrient uptake, and autochthonous carbon accrual. The result is that the

transformation of exported N, P and C is relatively low in flowing water stream settings and is

generally limited to sedimentation of larger, denser particles, bacterial decomposition of labile

organic matter, dilution with lower concentration phreatic ground water within influent reaches,

and abiotic gas exchange. Rates of transformation associated with autochthonous production

increase substantially when river systems become more lacustrine, and in such settings biological

processes of assimilation and mineralization often dominate the speciation of N, P and C.

Aside from development impact, three environmental gradients dominate the composition of

tributary water quality in northeast Florida (Hendrickson 1993, unpublished data): stream order,

watershed physiography, and annual cycles of temperature, rainfall and plant litter fall. The

PLSM construct attempts to incorporate annual temporal patterns and watershed development.

Stream order is implicitly handled by calibrating streams at their 2nd

to 4th

order reaches, the size

typical of most contributing stream’s terminus in the river. On a broader scale, stream water

quality can vary based on physiographic features such as the topography of the landscape and the

underlying parent material.

If one assumes that the continuum in concentration increase for these streams results largely

from development intensification, it follows that some low percentile rank reflects the

concentration resulting solely from natural watershed factors, such as parent material, native

vegetation cover and endemic faunal communities. In recent work on the establishment of

nutrient criteria (U.S. EPA, 2002) and identification of reference sites for developing

bioassessment metrics, the 25th

percentile has become an accepted benchmark in this continuum.

In both of these endeavors, the assumption is that this point in the continuum represents a

tolerable level of anthropogenic impact. Thus, implicit in the 25th

percentile concentration is

some small impact, and true natural background concentrations of N and P should be considered

to be lower. The 25th

percentile concentration can therefore be considered to represent an upper

range favoring low nonpoint source effect.

83

The 25th

percentile rank (Figure 27(a)) for streams of the type within the LSJR basin corresponds

to 0.77 mg/L of total nitrogen (TN) for Pleistocene Ridge streams (streams that drain the higher,

sandy ridges that predominate on the western banks of the St. Johns valley), and 0.88 mg/L TN

for Atlantic Coast Flatwoods streams (streams draining the low-lying region on the eastern banks

of the St. Johns). For total inorganic nitrogen (TIN = NH4 + NO2+3), the 25th

percentile value for

Pleistocene Ridge streams is 0.041 mg/L, and for Atlantic Coast Flatwoods streams is 0.058

mg/L. Total phosphorus (TP) and orthophosphate (PO4), concentrations corresponding to the

25th

percentile for Pleistocene Ridge and Atlantic Coast Flatwoods streams are, respectively:

0.038 mg/L TP, 0.023 mg/L PO4, and 0.066 mg/L TP and 0.039 mg/L PO4 (Figure 27(b)). In

comparison, concentrations assigned to undeveloped land uses (timberlands included) in

watershed model development for the lower St. Johns River Basin (Hendrickson and Konwinski,

1998; Hendrickson et al. 2002), shown in Table 4, are similar, producing a somewhat lower

concentrations of 0.7 mg/L for TN and values of between 0.02 and 0.04 mg/L for TIN, but

comparable and values between 0.05 to 0.07 mg/L for TP and between 0.03 and 0.05 mg/L for

PO4. Recent draft nutrient criteria developed by the EPA’s convened National Expert’s

Workshop places criteria thresholds (considered upper acceptable levels and thus incorporating

some anthropogenic contribution) for streams and rivers of this ecoregion (Region 12, northern

peninsular Florida) at 0.9 mg/L for TN and 0.04 mg/L TP.

84

Figure 27. Continuous Probability Density Functions for Total and Inorganic Nutrient Mean

Concentrations for Streams in Northeast Florida. ACF = Atlantic Coast Flatwoods

Streams; PPR = Plio-Pleistocene Ridge Streams; TN = Total Nitrogen; TIN = Total

Inorganic N (NOX+NH4); TP = Total Phosphorus; PO4 = Orthophosphate.

(a) Nitrogen

0

10

20

30

40

50

60

70

80

90

100

2 3 4 5 6 7 8 9 10

Ln(Total N, mg/m3as N)

Pe

rce

nti

le R

an

k

ACF - TN

PPR - TN

ACF - TIN

PPR - TIN

TN 25th %ile =

0.77 - 0.88 mg/L

TIN 25th %ile =

0.041 - 0.058 mg/L

(b) Phosphorus

0

10

20

30

40

50

60

70

80

90

100

0 1 2 3 4 5 6 7 8

Ln(TP, mg/m 3as P)

Pe

rce

nti

le R

an

k ACF - TP

PPR - TP

ACF - PO4

PPR - PO4

TP 25th %ile =

0.038 - 0.066 mg/L

PO4 25th %ile =

0.023 - 0.039 mg/L

85

Role of Spring Inputs

Artesian spring flow represents a significant proportion of the baseflow of the St. Johns,

comprising approximately 38% of the mean annual discharge at the basin’s inlet at Buffalo

Bluff. Thus, the influence of artesian spring N and P concentrations cannot be ignored in the

present-day and natural background levels of river and estuary productivity.

Today, many springs exhibit elevated concentrations of nutrients as a result of contamination of

deep Floridian ground water over the past half century. This contamination has occurred by

several means, including percolation of surficial aquifer water that has been enriched through

fertilization, land application of waste and nitrogen-enriched atmospheric deposition; stormwater

drainage wells; and sinkhole inputs of eutrophic surface waters. In most cases this enrichment

has resulted in increases in spring concentrations of the highly soluble and mobile nitrate-

nitrogen form (Figure 28), though in some cases direct entry through drainage wells and

sinkholes may have also resulted in elevated nitrogen and phosphorus concentrations (Figure

28). Trends in nutrient concentrations suggest that spring contamination via direct entry through

storm drain wells may be on the decline (as in the case of Rock and Wekiva Springs, Figure 28),

while nitrate contamination through surficial aquifer downward percolation may still be

increasing (as in the case of DeLeon, Blue, Silver and Gemini Springs). Presumably, because of

the tendency for phosphorus to be depleted in deep percolating groundwater from adsorbtion and

assimilation in surface soils, downward percolation does not represent a significant pathway for

phosphorus contamination.

86

Figure 28. Time-Series Concentrations of Nitrate+Nitrite-N and Orthophosphate-P in Major

Springs Discharging to the St. Johns River That Exhibit Nitrate+Nitrite Trends.

Data From USGS, Odum (1953) and SJRWMD.

(a) Nitrate + Nitrite

0.000

0.500

1.000

1.500

2.000

2.500

Fe

b-6

3

Fe

b-6

6

Fe

b-6

9

Fe

b-7

2

Fe

b-7

5

Fe

b-7

8

Fe

b-8

1

Fe

b-8

4

Fe

b-8

7

Fe

b-9

0

Jan

-93

Jan

-96

Jan

-99

Jan

-02

NO

X, m

g/L

Blue

DeLeon

Gemini

Rock

Silver

Wekiva

(b) Orthophosphate

0.000

0.020

0.040

0.060

0.080

0.100

0.120

0.140

Feb-6

3

Feb-6

6

Feb-6

9

Feb-7

2

Feb-7

5

Feb-7

8

Feb-8

1

Feb-8

4

Feb-8

7

Feb-9

0

Jan-9

3

Jan-9

6

Jan-9

9

Jan-0

2

To

tal P

O4, m

g/L

Blue

DeLeon

Gemini

Ro ck

Silver

Wekiva

87

Because of the absence of trends in orthophosphate concentrations for major springs contributing

to the St. Johns River (Wekiva and Rock Springs excluded), the present day, flow-weighted

orthophosphate concentration of 0.041 mg/L has been used to represent the natural background

concentration. For nitrate+nitrite-N, the 25th

percentile occurrence of all springs has been used,

which corresponds to a concentration of 0.01 mg/L. Spring TP and TKN data are also available,

though in lower numbers and temporal coverage. Spring-flow mean TP concentration was found

to be 0.061 mg/L, while the mean concentration of dissolved TP was 0.04 mg/L, suggesting

particulate P as a large component of non-PO4-phosphorus. The mean concentration of total

Kjeldahl nitrogen was 0.081 mg/L. Fourteen of the 35 available measurements for TKN were

remarked as below the method detection limit of approximately 0.08 mg/L. Due to the

uncertainty regarding the concentrations of TP and TKN, the relatively low concentrations, and

the possibility for organic N and non-PO4-phosphorus to arise from spring source other that the

artesian boil, the nitrate+nitrite and orthophosphate concentration have been used to represent

total N and total P from artesian springs.

Using the 25th

percentile values of TN and TP observed for ridge and flatwoods tributary runoff

to represent background terrestrial flow concentrations, the 25th

and percentile concentration of

spring NO2+3, and the mean flow-weighted spring PO4 concentration, natural background

concentrations of TN and TP were projected for the St. Johns River at Buffalo Bluff (Figure 29).

The proportions of surface and groundwater were determined through a relationship between

present-day refractory total organic carbon and the calculated ratio of terrestrial:artesian flow

occurring during the 2 week period prior to the time of sampling. This approach was used to

dampen the oscillations in discharge that occur at the Buffalo Bluff due to intermittent and

extended reverse flows (Sucsy and Morris, 2001). TP concentration was reduced by the

sedimentation rate for Lake George by applying the Vollenweider calculation using the

formulation of Chapra (1997). Calculated Lake George TP sedimentation rate was found to be

1.0 m/yr, and although unusually low, is in agreement with the sedimentation rate determined for

Lake Apopka, another large, shallow sub-tropical lake (Coveney, 1997). The final natural

background annual mean TP concentration delivered to Buffalo Bluff was determined to be

0.037 mg/L, for the hydrologic conditions existing from 1995-99. The actual 1995-99 mean TP

concentration at Buffalo Bluff was 0.063 mg/L.

88

Figure 29. Comparison of Present Day and Predicted Natural Background Concentrations of

Total Nitrogen and Total Phosphorus in the Lower St. Johns at Buffalo Bluff, 1995-

99.

(a) Total Nitrogen

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Se

p-9

4

Ma

r-9

5

Se

p-9

5

Ma

r-9

6

Se

p-9

6

Ma

r-9

7

Se

p-9

7

Ma

r-9

8

Se

p-9

8

Ma

r-9

9

Se

p-9

9

TN

mg

/L

Natural Background

Current Condition

(b) Total Phosphorus

0.00

0.02

0.04

0.06

0.08

0.10

0.12

Se

p-9

4

Ma

r-9

5

Se

p-9

5

Ma

r-9

6

Se

p-9

6

Ma

r-9

7

Se

p-9

7

Ma

r-9

8

Se

p-9

8

Ma

r-9

9

Se

p-9

9

TP

, m

g/L

Natural Background

Current Condition

89

To develop an estimate for natural background TN concentration, the terrestrial to spring flow

ratio approach was again used, for both bioavailable (considered to be the sum of inorganic,

labile organic and algal fractions) and refractory nitrogen partitions. Labile and refractory

nitrogen concentrations were separately estimated to account for phytoplankton bio-available

nitrogen deficit that would presumably be made up through algal nitrogen fixation.

Concentrations were calculated corresponding to each sampling day of the bi-weekly sampling

record between 1995-99. RTON was assumed to be the same as the present day concentration,

while time-varying LTON was calculated using the present-day RTON-LTON relationship,

reduced by the mean ratio of historic to present day LTON. Natural background estimated algal

P was used to determine an algal ON based on Redfield stoichiometric equivalency (7.2:1 N:P

mass ratio), and for days on which the labile N was less than algal ON, the difference was added

to the labile N concentration. Labile and refractory N concentrations for each day were then

summed, to produce a mean 1995-99 natural background mean TN concentration of 0.687 mg/L.

By comparison, the actual 1995-99 mean TN concentration at Buffalo Bluff was 1.54 mg/L.

Mean natural background chlorophyll a concentration was estimated to be 14.0 mg/m3, while the

actual 1995-99 mean corrected chlorophyll a was 26.6 mg/m3.

Historic Data for the St. Johns River

Few water quality data exist for the St. Johns River prior to substantial levels of development.

While the point at which the St. Johns River Basin became developed to a degree that it exerted a

significant effect on water quality is difficult to ascertain, the sharp increase in the State’s

population that began subsequent to World War II appears to correspond the earliest reports of

water quality degradation (Figure 30). Population density within the basin upstream of

Jacksonville remained relatively low until 1940. Between 1940 and 1950, population within the

basin increased by 39% (Fl. State Board of Health, 1951).

90

Figure 30. Population growth within the 14 Counties of the St. Johns River Basin, 1890 –

2000. Data from Dietrich (1978) and University of Florida (2000). Significant

events with likely impact on nutrient status identified.

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

1880 1900 1920 1940 1960 1980 2000

Po

pu

lati

on

, mill

ion

s o

f p

ers

on

s

Water hyacinth

introduced

First of 5 major

Jacksonville Harbor

dredging projects

Upper St. Johns marsh

encroachment begun w /

Fellsmere grade

Earliest

WQ data

2,4-D First used to

control w ater hyacinth

Upper St. Johns Marsh

1/3rd of its original area

Rodman Dam

completed

Water hyacinth eliminated

follow ing operation “Clean Sw eep”

– Era of algal blooms begins

Blount Island

Cut Completed

Upstream point

source direct

discharge

eliminated

By the time of the earliest, regular surface water quality monitoring efforts, beginning in the late

1960’s, water quality in the middle St. Johns River appears to have already declined. Moody

(1970) attributed this decline in Lake George to the occurrence of regular, severe algal blooms.

Principle factors leading to these algal blooms were believed to be: 1) upstream development and

its concomitant nutrient enrichment through point and non-point source pollution, and 2) aquatic

weed spraying to eliminate floating water hyacinth. Moody notes that concentrations of blue-

green algae were in 1967, “present in much greater numbers . . . “ than reported in an earlier

study in 1939-40 (by E. Lowe Pierce (1947)). Two water quality sample events for Lake

George, which were collected in July of 1967 and December of 1969, are contained in his report.

91

The July sample contained 2.28 mg/L TN and 0.13 mg/L TP, while the December sample

contained 1.3 mg/L TN and again, 0.13 mg/L TP.

Three data sets have been identified that can help in understanding the nutrient status and hence

open water ecology of the St. Johns prior to substantial development. The first of these is a study

conducted by E. Lowe Pierce (1947) reporting on several aspects of the water quality and

plankton in the St. Johns River at locations in the Ocklawaha River mouth, in the river upstream

and downstream of the Ocklawaha from September 1939 to November 1940. The second is the

work of Odum (1953) characterizing the phosphorus concentrations of waters of the State. The

third study was published by the Florida State Board of Health in1948 to determine the effect of

untreated waste discharge in and around the City of Jacksonville.

The objective of the Florida State Board of Health study was to determine the effect of the

discharge of untreated sewage in the vicinity of Jacksonville, and sampling stations were

established in the St. Johns River near the city of Green Cove Springs to characterize upstream,

background conditions. Two surveys were performed, the first conducted in May and June of

1945, and a second conducted from September 1945 to May of 1946. Though nutrient analysis

was not performed, BOD and dissolved oxygen were examined and may be used to infer trophic

status. Many of the reports’ sampling stations were located in Jacksonville, which at the time

was already significantly impacted, principally from the discharge of raw sewage. However,

comparing the report’s BOD data from the upstream, “un-impacted” site near the Shands Bridge

to present day concentrations at the same location suggests that water column biodegradable

organic matter has increased over time. The May to June survey produced what was referred to a

“highest average” (statistical methods are not explained in the report) of 0.83 mg/L; in

comparison, the May - June mean BOD concentration at Shands Bridge from 1996 to 2000 was

2.14 mg/L. The Sept. 1945 to May 1946 survey produced a highest average concentration of 0.82

mg/L, while the 1996-2000 average is 1.43 mg/L.

Under present-day conditions in the LSJR, algal biomass is accounts for the majority of labile

organic carbon in the river, and the relationship between BOD and chlorophyll a is highly

significant, with chlorophyll a concentrations explaining over 70% of the variation in BOD (Chl-

92

a = 14.39*(BOD) – 4.11; R2 = 0.71). Based on this relationship, the 0.83 mg/L BOD measured

in 1945 corresponds to a chlorophyll a of about 8 mg/m3. Converting the present day mean river

color for this location to refractory organic nitrogen (assumes color has not changed; in reality,

river color probably has declined somewhat due to basin development), and adding in the

nitrogen content of algal biomass at 8 mg/m3 chlorophyll a, a mean total organic nitrogen content

of 0.48 mg/L can be calculated. With the inclusion of inorganic nitrogen, it would be expected

from these BOD data that total nitrogen was in the neighborhood of 0.6 mg/L, comparable to the

reconstructed historic Buffalo Bluff mean TN concentration of 0.687 mg/L. It should be noted

that Green Cove Springs is located downstream of where, even in 1946, potentially significant

inputs of nutrients from the city of Palatka (20 miles upstream, at that time with a population of

approximately 8,000) and the tri-county agricultural area, may have occurred.

The Odum (1953) report to the Florida Geological Survey extensively surveyed orthophosphate

and total phosphorus in surface waters around the State. Samples were collected at one time

from many different locations, so annual trends cannot be inferred. These data for locations in

the St. Johns River and its contributing streams are listed in Table 9. Because a significant

amount of development had begun to occur at the time of this study, these data must be viewed

selectively for the potential of anthropogenic nutrient contamination. For locations that likely

still represented unimpacted reaches of the lower St. Johns River in 1952 (assumed to be Lake

George and Crescent Lake), these data suggest a concentration of total phosphorus of around

0.04 mg/L. Upstream of Lake Monroe, the Odum data suggests a St. Johns River that was

remarkably low in phosphorus.

Table 9. Total Phosphorus Concentrations Determined for Selected Locations in St. Johns

River Basin in 1952. Data from Odum (1953).

Location Date Total P, mg/L

Black Creek, Route 17 Aug. 9, 1952 0.04

Deep Creek, Hastings, Route 207 Jul. 14, 1952 0.54

Crescent Lake, Andalusia Jul. 19, 1952 0.033

93

Doctor’s Lake, Route 17 Aug. 9, 1952 0.065

Lake George at Silver Glen Springs Aug. 14, 1952 0.044

Lake Monroe, Sanford Jun. 23, 1952 0.18

Ortega River, Route 21 Aug. 9, 1952 0.044

St. Johns R., Crows Bluff, Volusia Co. Sep. 3, 1952 0.117

St. Johns R., Palatka Jul 19, 1952 0.061

St. Johns R., Route 192 (Brevard Co.) Jun. 23, 1952 0.007

St. Johns R., Route 50 (Orange Co.) Jun. 23, 1952 0.015

St. Johns R., Green Cove Springs Jul. 16, 1952 0.119

The data of Pierce (1947), due to the length of his study and the comparatively large suite of

measurements, provide compelling evidence of a river that was dramatically lower in nutrients

and algal biomass. The graphs of Figure 31 compare the results of this study with the present

day mean concentrations observed from 1995-99. In 1939-40, Pierce reported blue green algae

(of the genera Anabaena, Raphidiopsis and Microcystis) ranging from too few to count for most

months, to 36,000 cells/ml in August of 1940. In comparison, the annual mean peak blue-green

cell count exiting Lake George for 1997-2000 (Phlips and Cichra, 2001) was 518,893 cells/ml.

The Pierce study also suggests a shift in the dominance of phytoplankton groups, with diatoms

(primarily the genera Coscinodiscus and Melosira) formerly making up a much greater relative

portion of the plankton.

The Pierce study also provides data on nitrogen forms throughout the year. (Unfortunately,

analysis for phosphorus forms was not performed.) Due to some differences in methodology and

uncertainties regarding sample handling and preservation techniques, only total nitrogen

concentrations are compared. Pierce reported mean annual TN as 0.41 mg/L in Little Lake

George (upstream of the Ocklawaha mouth), and 0.37 mg/L at Welaka (downstream of the

Ocklawaha), values that are roughly 1/4th

of present day concentrations. The Pierce TN numbers

are similar to the present day estimated mean concentration refractory total organic N at Buffalo

Bluff, of 0.46 mg/L. Because RTON theoretically reflects a relatively constant natural supply of

94

organic nitrogen, its concentration would be expected to remain constant, or perhaps even

decrease due to development, over time.

95

Figure 31. Comparison of Monthly Mean Water Quality Parameters for 1995-99 (solid boxes)

to the Data Collected by Pierce (1947) in 1939-40 (open diamonds) for the St. Johns

River near Buffalo Bluff. Error bars on 1995-99 data represent 95% C.I. Pierce

chlorophyll estimated from cell counts.

a) Total Chloride

0

90

180

270

360

450

Jul-39 Sep-39 Nov-39 Dec-39 Feb-40 Mar-40 May-40 Jul-40 Aug-40 Oct-40

mg

/L

b) Dissolved Oxygen

0

2

4

6

8

10

12


mg

/L

c) Total Nitrogen

0.0

0.5

1.0

1.5

2.0

2.5


mg

/l

d) Secchi Depth

0

0.5

1

1.5

2


me

ters

e) Chlorophyll-a (corrected)

0

20

40

60

80

100


mg

/m3

96

The Pierce (1947) total N concentrations are low when compared to those reconstructed here,

which suggest an annual mean concentration of 0.65 – 0.70 mg/L. Similarly, mean annual

chlorophyll a estimated from cell densities in the Pierce work, at 2 mg/m3, are considerably

lower than the reconstructed annual mean of 14 mg/m3. However, unaccounted for in the water

column measurements of these studies, but theoretically included in the runoff delivery

approach, is the sequestration of nutrient in water hyacinth (Eichornia crassipes). Water

hyacinth, introduced to the St. Johns shortly before 1900, quickly spread through the river, and

anecdotal accounts prior to 1940 (Rawlings, 193_) indicate widespread coverage. Annual

reports to Congress on the progress of hyacinth control in the St. Johns (USACE Annual Reports

to Congress, 1930 - 1967) indicate that between 3000 to 13,000 acres of hyacinth were removed

annually, suggesting that at least 5 to 10 % (based on the sum of lake surface areas of the St.

Johns from Lake Winder through Little Lake George) of the rivers water surface area typically

may have been covered.

To estimate the potential sequestration of nitrogen in water hyacinth, literature values for organic

matter dry weight and percent N and P content were combined with areal coverage estimates

determined by the U.S. Army Corps of Engineers. Using an 8% areal hyacinth coverage

estimate (U.S. Army Corps estimate from the only available comprehensive survey of hyacinth,

conducted in November of 1948, of 9,500 acres (USACE, 1948)) and assuming 1000 g/m2

organic matter, with 1% as N, 0.05% as P (Reddy and DeBusk, 1987; values represent minimum

in the areal organic matter and nutrient content ranges, as would be expected for uncultivated

hyacinth growing under natural and assumed nutrient-limited conditions), an additional 0.27

mg/L N can be calculated as sequestered in hyacinth tissue. Adding this concentration to the

Pierce mean TN concentration for the St. Johns below the Ocklawaha mouth produces a mean

total N concentration of 0.64 mg/L.

Differences in analysis methodology, the relatively short study durations, possibly different

hydrologic conditions, and assumptions in the conversion procedures warrant that these estimates

of background concentrations be used with caution. However, these studies suggest substantially

lower water column nitrogen and algal biomass, and marginally lower phosphorus concentrations

prevailed in the pre-development St. Johns River. If the concentrations of TN of 0.6 mg/L and

97

TP of 0.04 mg/L are accepted as representative of water column (including floating

macrophytes) natural background levels, then nutrients flowing into the lower St. Johns River

from the upper and middle St. Johns and Ocklawaha Rivers today appear to be elevated between

1.5 to 4 times above pre-development conditions.

Total LSJR Load Estimates

With the estimated of upstream, natural background loads, it is possible to estimate the total

incoming anthropogenic nutrient and organic carbon load to the Lower St. Johns River. Tables

10 - 15 summarize the mean annual loads from source categories to the LSJR by river ecozone

from 1995 through 1999, and the over all average. When comparing this compilation of loads to

that performed earlier by Hendrickson and Konwinski (1998) it should be noted that in this

assessment, the portion of river identified as predominantly oligohaline has been moved

upstream, to encompass all loads entering from Black Creek northward to the Ortega River. In

the previous assessment, loads entering from Orange Park northward were considered to

contribute directly to the oligohaline reach, and Black Creek loads were added to the freshwater

river. The reason for the move was the prevailing drier conditions encountered during the 1995-

99 assessment period, particularly for 1998 and 1999.

The separation of labile and refractory organic carbon and nutrients performed in this version of

the external load provides a dramatically different conclusion on the relative sources of problem

nutrients leading to algal blooms in the lower St. Johns. In the previous compilation of the LSJR

external load, Hendrickson and Konwinski (1998), in lieu of information on the utilization of

organic nitrogen and non-orthophosphate phosphorus forms by river algal communities,

estimated the bioavailable load as that composed of only inorganic nutrient (NH4, NOX, and PO4)

forms. The addition of labile organic forms to the “total bioavailable pool” of nutrients greatly

increases the apparent influence of upstream loads. Accounting for labile organic forms also

increases the relative importance of urban nonpoint source runoff in potential contribution of

problem nutrient loads.

98

Table 10. Summary of Mean Annual Loads to the Lower St. Johns River, 1995. All values in metric tons per year.

Total N Labile TON

Refractory

TON

Total

Inorganic N Total P Labile TNOP

Refractory

TNOP Total PO4

Total

Organic C Labile TOC Refractory TOC

Buffalo Bluff Total 10765.1 4336.2 5373.0 1056.0 511.2 207.3 96.9 207.0 138347.0 12976.2 125370.9

Natural Background 6659.7 1006.7 5432.5 220.5 290.8 97.9 97.3 95.6 131539.8 5727.6 125812.3

Dunns Creek Total 1372.5 290.7 919.0 162.8 108.2 33.6 35.4 39.1 23100.5 718.7 22381.9


Upstream Total 12137.7 4627.0 6291.9 1218.8 619.3 240.9 132.3 246.1 161447.5 13694.8 147752.7

Fresh Tidal NP Total 1068.4 371.7 505.6 191.1 211.2 54.9 22.7 133.6 23875.4 1709.8 22165.6

Natural Nonpoint 626.3 203.8 387.9 34.6 56.8 11.1 6.7 38.9 23213.7 1255.7 21957.9

Agriculture Contribution 384.1 126.1 124.7 133.4 136.9 35.4 13.4 88.1 217.9 173.7 44.2

Urban Contribution 53.2 39.3 -3.5 17.4 15.4 8.5 1.7 5.2 -507.8 171.4 -679.2

Other Nonpoint 4.7 2.5 -3.5 5.7 2.1 -0.1 0.9 1.3 951.7 109.0 842.7

Point Source 306.7 151.5 12.0 143.2 70.2 32.0 0.8 37.4 1417.6 814.6 603.0

Oligohaline NP Total 1141.3 517.8 447.3 176.3 186.8 79.2 16.7 90.9 25692.0 2584.6 23107.4


Agriculture Contribution 26.2 10.7 2.0 13.5 10.9 2.1 0.5 8.3 1.5 46.5 -45.0


Other Nonpoint -1.8 1.0 -13.5 10.7 0.5 0.0 0.2 0.3 901.5 96.0 805.5

Point Source 333.5 49.6 3.9 279.9 72.1 11.2 0.3 60.6 287.1 165.0 122.1

Meso-Polyhaline NP Total 440.2 223.2 120.6 96.4 92.0 44.0 6.6 41.4 6524.5 1002.8 5521.6




Other Nonpoint -1.5 -2.1 -8.2 8.8 0.2 -0.2 0.2 0.2 236.7 30.1 206.6

Point Source 1147.4 238.9 18.9 889.6 294.4 46.9 1.2 246.2 1920.7 1103.7 817.0

Total Atmospheric Dep. 219.5 2.8

LSJRB Summary

Total Natural Nonpoint 1591.0 508.7 995.3 86.9 142.2 28.0 17.5 96.7 57574.7 3054.3 54520.4

Total Augmented Nonpoint 1059.0 603.9 78.2 376.8 347.9 150.2 28.6 169.1 -1482.8 2243.0 -3725.8

Total Point Source 1787.6 440.0 34.9 1312.7 436.6 90.2 2.3 344.2 3625.4 2083.3 1542.1

Grand Total 16794.8 6179.6 7400.4 2995.2 1548.8 509.2 180.6 856.1 221164.9 21075.5 200089.4

Notes: N= Nitrogen; P=Phosphorus; C=Carbon. NP=Nonpoint Sources. LSJRB Summary sums loads for only the lower St. Johns Basin downstream of Dunns Creek.

99


Total N Labile TON

Refractory

TON

Total


Refractory

TNOP Total PO4

Total






Upstream Total 9507.9 5000.6 3848.1 659.2 427.5 253.2 61.7 112.5 120237.1 17550.3 102686.8



Agriculture Contribution 177.0 56.0 49.7 71.3 51.8 15.0 5.6 31.2 -24.3 88.7 -113.0


Other Nonpoint 5.2 0.3 1.0 3.9 1.3 0.0 0.6 0.7 479.7 39.2 440.4

Point Source 285.6 144.6 11.5 129.6 66.0 30.5 0.8 34.7 1340.4 770.2 570.1





Other Nonpoint 1.0 117.0 -6.2 5.8 0.8 0.1 0.2 0.4 592.0 57.8 534.2

Point Source 322.5 42.3 3.4 276.8 65.6 10.5 0.3 54.9 351.8 202.2 149.7





Other Nonpoint 0.4 -0.6 -6.2 7.1 0.8 -0.1 0.3 0.6 275.6 33.3 242.3

Point Source 1144.4 251.6 20.0 872.9 328.9 50.8 1.3 276.9 2199.7 1264.0 935.7


LSJRB Summary




Grand Total 13142.2 6136.9 4555.3 2246.3 1185.2 457.1 91.9 633.2 158640.3 23045.9 135594.5


100


Total N Labile TON

Refractory

TON

Total


Refractory

TNOP Total PO4

Total






Upstream Total 5782.7 3924.6 1625.5 232.6 233.1 175.7 28.6 28.8 72744.4 18232.8 54511.5





Other Nonpoint 5.7 -1.2 0.0 7.0 1.7 0.1 1.0 0.5 582.3 58.7 523.6

Point Source 299.6 86.6 73.1 139.7 69.1 24.0 7.0 38.1 4789.3 585.6 4203.6





Other Nonpoint -1.6 -0.3 -8.3 7.0 0.1 0.0 0.1 0.0 550.9 54.9 496.0

Point Source 341.3 45.9 9.8 285.6 73.6 11.5 0.7 61.5 321.6 143.8 177.8





Other Nonpoint -0.8 -1.9 -5.0 6.1 0.2 0.1 0.2 -0.1 144.1 17.0 127.1

Point Source 1187.7 251.1 33.7 902.9 334.6 71.4 3.1 260.1 2233.5 1354.5 879.0


LSJRB Summary




Grand Total 9910.4 5157.8 2563.7 1953.6 1051.8 418.7 75.4 554.7 122927.4 24346.0 98581.5


101


Total N

Labile TON

Refractory TON

Total Inorganic

N

Total P Labile TNOP

Refractory TNOP Total PO4

Total Organic C

Labile TOC

Refractory TOC

Buffalo Bluff Total 8561.5 4942.4 3175.9 443.1

341.8 201.8 42.5 97.4

127323.1 21218.1 106105.0

Natural Background 4428.1 1189.7 3175.9 62.5

246.4 140.0 42.5 63.9

112873.9 6768.9 106105.0

Dunns Creek Total 971.2 217.6 681.9 71.7

51.3 15.8 15.9 19.7

21379.6 778.7 20600.9

Natural Background 813.6 108.2 681.9 23.5

39.4 11.1 15.9 12.4

21216.6 615.7 20600.9

Upstream Total 9532.7 5160.0 3857.8 514.9

393.1 217.6 58.4 117.1

148702.7 21996.9 126705.9

Fresh Tidal NP Total 1652.2 480.2 935.0 237.0

222.9 53.8 31.1 138.0

44053.4 2272.1 41781.4

Natural Nonpoint 1188.3 284.7 864.2 39.4

103.4 17.3 16.7 69.4

43976.7 1525.1 42451.6

Agriculture Contribution 350.3 111.7 93.9 144.7

92.2 27.3 11.9 53.0

-257.6 256.3 -513.9

Urban Contribution 110.4 92.5 -21.4 39.4

25.8 13.4 2.8 9.6

-817.6 443.8 -1261.4

Other Nonpoint 3.1 -8.7 -1.7 13.6

1.5 -4.2 -0.2 5.9

1151.9 47.0 1105.0

Point Source 274.2 82.4 57.1 134.5

62.1 21.9 5.1 35.0

4154.4 582.3 3572.2

Oligohaline NP Total 1236.9 492.7 565.8 178.4

171.8 63.8 18.4 89.6

28792.1 2331.6 26460.5


72.4 12.1 11.6 48.7

29623.8 1041.0 28582.8

Agriculture Contribution 35.9 17.9 9.5 8.5

8.7 5.9 2.4 0.5

-51.2 53.9 -105.1


90.6 50.4 6.3 33.9

-1540.4 1200.2 -2740.6

Other Nonpoint -3.5 -6.9 -12.1 15.4

0.1 -4.5 -1.9 6.5

759.8 36.5 723.3

Point Source 301.3 53.7 9.6 238.0

81.4 13.2 0.7 67.5

363.5 184.3 179.2

Meso-Polyhaline NP Total 867.0 436.2 254.1 176.7

152.0 68.5 11.4 72.1

13343.1 1966.9 11376.3


37.7 6.5 5.7 25.5

14672.1 570.9 14101.2

Agriculture Contribution 38.3 7.3 -1.5 32.5

11.8 -3.1 -1.1 16.1

29.8 49.8 -20.0


101.5 59.8 4.8 36.9

-1741.7 1310.4 -3052.1

Other Nonpoint -1.8 3.6 -4.1 -1.3

0.9 5.3 2.0 -6.3

382.9 35.8 347.1

Point Source 1267.0 279.4 38.3 949.3

341.5 70.7 3.3 267.6

2468.4 1500.7 967.7

Total Atmospheric Dep. 278.2

3.8 LSJRB Summary

Total Natural Nonpoint 2444.9 594.0 1765.7 85.2

213.5 35.8 34.0 143.6

88272.7 3137.0 85135.7

Total Augmented Nonpoint 1311.2 815.2 -10.8 506.8

333.2 150.3 26.8 156.1

-2084.0 3433.6 -5517.6

Total Point Source 1842.4 415.5 105.1 1321.8

485.0 105.8 9.1 370.1

6986.3 2267.2 4719.1

Grand Total 15409.4 6984.7 5717.8 2428.7

1428.5 509.5 128.4 786.9

241877.7 30834.6 211043.1


102


Total N Labile TON

Refractory

TON

Total


Refractory

TNOP Total PO4

Total




Dunns Creek Total -166.6 -120.9 -45.0 -0.8 -8.9 -6.5 -1.9 -0.6 -1443.5 -401.8 -1041.7

Natural Background -80.4 -35.3 -45.2 0.2 -3.9 -2.0 -1.9 0.0 -1263.6 -201.0 -1062.6

Upstream Total 5113.6 3755.4 1223.0 181.2 174.5 143.7 15.3 17.4 61183.8 16762.3 44421.6





Other Nonpoint -3.3 3.0 -8.1 1.8 -0.2 0.0 0.0 -0.3 105.7 51.9 53.8

Point Source 275.3 144.0 11.4 119.8 64.5 30.3 0.8 33.4 1232.2 708.0 524.1





Other Nonpoint -6.3 3.8 -12.7 2.6 -0.8 0.1 -0.2 -0.7 71.4 57.3 14.0

Point Source 305.2 46.3 3.7 255.2 81.9 13.1 0.3 68.5 249.4 143.3 106.1





Other Nonpoint -1.8 0.8 -5.3 2.7 0.0 0.0 0.0 0.0 47.9 21.8 26.1

Point Source 1121.5 206.0 16.3 899.2 330.3 50.0 1.2 279.1 1401.0 805.1 595.9


LSJRB Summary


Total Augmented Nonpoint 260.2 158.5 -0.3 102.0 92.0 38.1 6.8 47.0 -721.5 622.8 -1344.2


Grand Total 7632.4 4416.8 1511.2 1575.7 782.1 281.4 29.2 470.9 76828.0 19626.4 57201.6


103

Table 15. Summary of Overall Mean Annual Loads to the Lower St. Johns River, 1995 - 1999. All values in metric tons per year.

Total N Labile TON

Refractory

TON

Total


Refractory

TNOP Total PO4

Total






Upstream Total 8414.9 4493.5 3369.3 561.3 369.5 206.2 59.3 104.4 112863.1 17647.4 95215.7





Other Nonpoint 3.1 -0.8 -2.5 6.4 1.3 -0.8 0.4 1.6 654.3 61.2 593.1

Point Source 288.3 121.8 33.0 133.4 66.4 27.7 2.9 35.7 2586.8 692.2 1894.6





Other Nonpoint -2.4 -0.2 -10.6 8.3 0.1 -0.9 -0.3 1.3 575.1 60.5 514.6

Point Source 320.8 47.6 6.1 267.1 74.9 11.9 0.4 62.6 314.7 167.7 147.0



Agriculture Contribution 17.3 5.2 0.5 11.6 6.2 0.2 -0.1 6.1 19.7 23.9 -4.1


Other Nonpoint -1.1 0.0 -5.7 4.7 0.4 1.0 0.5 -1.1 217.4 27.6 189.8

Point Source 1173.6 245.4 25.5 902.8 325.9 58.0 2.0 266.0 2044.6 1205.6 839.1

Total Atmospheric Dep. 222.2 222.2 3.0 3.0

LSJRB Summary




Grand Total 12577.8 5775.1 4349.7 2462.1 1199.3 435.2 101.1 663.3 164287.7 23785.7 140502.0


104

In the upstream nutrient load, total N and P in the 1995-99 load summary are similar to that of

1993-94, at 8,415 MT/yr TN and 370 MT/yr TP, compared to 7209 MT/yr TN and 317 MT/yr

TP. Upstream TIN load from 1995-99 was considerably less than that in 1993-94, at 414 MT/yr

compared to 639 MT/yr. Total PO4 load was marginally less in 1997-98, at 73 MT/yr as

compared to 99 MT/yr in 1993-94. The generally lower upstream inorganic nutrient loads may

be attributed to lower flow conditions that prevailed during the 1997-98 period, resulting in

longer residence time in upstream lakes and greater incorporation of inorganic forms in algal

biomass. Although the el Nino winter of 1997-98 resulted in one of the highest mean annual

flow rates for the St. Johns River at the upstream Deland gauging station, this extremely wet

interval was short, and immediately followed by one of the longest droughts in Florida

meteorological history. Hence the tendency for low algal standing stock associated with a high

flow/short residence time hydrologic pattern was not seen in 1998.

Within-LSJR basin anthropogenic loads are still predominantly attributable to point source

discharges. From 1995-99, point sources accounted for 1,783 MT/yr of TN and 467 MT/yr of

TP, as compared to the 1993-94 estimate of 1,886 MT/yr of TN and 530 MT/yr of TP. Applying

the labile/refractory partitioning algorithm to point source effluents suggests that most of the

organic nutrient portion of these effluents is in the bioavailable form.

Within-LSJR basin nonpoint source loads are lower that that assessed for the 1993-94 time

period, owing largely to the prevailing dry condition that decreased nonpoint source runoff in

general. An additional factor leading to this decrease is a substantial decline in tri-county row

crop area. In the previous assessment, the within-basin, nonpoint TN mean annual load was

determined to be 3,528 MT/yr, with 1,243 MT/yr of this representing the above background load.

In this assessment, the mean annual 1995-99 nonpoint TN load was determined to be 2,158

MT/yr, with 834 MT/yr of that the above-background load. For TP, the 1993-94 nonpoint load

was determined to be 665 MT/yr, with 468 MT/yr of this the above-background load, while for

1995-99, the mean load was determined to be 360 MT/yr, with 243 MT/yr of this the augmented,

above-background load. However, the relevance of the nonpoint source load as it relates to the

contribution of bioavailable nutrients, is considerably increased in this assessment, owing to the

distinction of labile organic pool. In the 1993-94 assessment, in which only the TIN load was

105

considered to represent the “bioavailable” load, total anthropogenic bioavailable nonpoint load

was assessed as 488 MT/yr of TIN and 248 MT/yr of PO4. With the inclusion of labile organic

nutrient forms, the 1995-99 augmented nonpoint source bioavailable N load increases to 802

MT/yr. The 1995-99 augmented nonpoint bioavailable P load was estimated as 223 MT/yr, less

than the estimated PO4 load of 1993-94; however, the inclusion of the labile organic fraction

represents a 97 percent increase over an estimate based upon PO4 alone.

Atmospheric Deposition

The estimate of directly-intercepted atmospheric deposition load has been considerably reduced

in the 1995-99 total load assessment. Using three NADP gauges, rather than the one used in the

1993-94 assessment, including dry deposition, and relying on tower measurements to determine

the P load, Pollman and Roy (2003) provided data that demonstrates the directly-intercepted

rainfall load to the LSJR to be roughly 222 MT/yr of TN, of which most is inorganic N, and 3

MT/yr of P. Thus, direct atmospheric deposition represents only 2.8 percent of the total

bioavailable N load and 0.5 percent of the bioavailable P load to the LSJR. The discrepancy

between this estimate and the earlier estimate, which found atmospheric deposition to comprise

14.7% of the bioavailable N load and 0.8% of the bioavailable P load can be attributed to two

factors. The first is the inclusion of labile organic N and P in the assessment of bioavailable

point and nonpoint source loads. The second can be traced to a units error in the Hendrickson

and Konwinski (1998) assessment, in which atmospheric deposition NO3 concentrations that

were reported as mg/L NO3, were assumed to be as mg/L N. The result was a 443 percent

overestimate in atmospheric N load.

Accounting for indirect deposition, that portion of wet deposition falling on terrestrial areas and

entering surface waters un-attenuated in runoff, can in some cases substantially increase the

contribution of atmospheric nitrogen. Timpe (1999) concluded that 28% of the non-point source

nitrogen loading from mixed urban and residential watersheds in the Tampa Bay Watershed

could be attributed to wet atmospheric deposition. Rushton (1997) found a significant

correlation between precipitation nitrogen concentration and the concentration of parking lot

106

runoff for individual storm events in a study of wet detention pond effectiveness. Employing

the wet deposition retention factors (the proportion of wet deposition that arrives in surface

waters for various land covers) of Paerl et al. (2001), the 1995-99 mean annual indirect

atmospheric deposition added by land areas in the lower St. Johns River basin is estimated at and

additional 222 MT/year. This doubles the overall contribution of atmospheric nitrogen to the

total N load of the LSJR to 5.5 percent. However, an arguably more relevant way to view the

contribution is to view it in the context of the urban nonpoint nitrogen pollution. If atmospheric

nitrogen deposition has roughly doubled over the past 150 years due to anthropogenic inputs

(Pollman, pers. comm.), then 84 MT per year of the indirect atmospheric deposition that is

exported from urban areas can well be viewed as anthropogenic atmospheric contribution. This

represents 15 percent of the anthropogenic labile nitrogen that enters the lower St. Johns River in

an average year.

DISCUSSION

In order to develop ecologically relevant nutrient budgets for estuaries, that portion of the

organic nutrient pool that is biologically available must be quantified (Seitzinger and Sanders,

1997). The results of this analysis suggest that a potentially wide range in bioavailability may

exist in the organic carbon and nutrient pools of tributary runoff from different land uses. For

the watersheds employed in the calibration of PLSM coefficients, LTOC was on average 10% of

TOC, ranging from 4% in waters draining undeveloped, forested streams to 54% for a highly

urbanized streams. In a study of mixed urban/ag/forested watersheds in southeast Pennsylvania,

Volk et al. (1997) found bioavailable DOC to constitute on average 25 % of DOC. As southeast

US blackwater streams would be expected to exhibit higher levels of RDOC, the values obtained

here appear consistent with that study. Due to the tendency for labile organic matter to possess

higher relative amounts of N and P, the fractions of labile nutrient within total organic nutrient

forms are higher than labile carbon fractions, on average 50% for LTON and 75% for LTNOP.

LTON ranged from 28% of TON for largely undeveloped watersheds to 78% for urbanized

streams. Due to the very low levels of TNOP observed in undeveloped watershed runoff, it

cannot be said with certainty what percent is composed of LTNOP, although it appears that

concentrations are probably less than 10 g/L. For urbanized streams, TNOP was found to be

107

92% composed of LTNOP. Thus it can be seen that anthropogenic nutrient loading, due to its

relatively greater amount of labile nutrient content, has a proportionally greater effect on

eutrophication than would be surmised by the absolute increase in total nutrient (Figure 32).

108

Figure 32. Comparison of Total and Bioavailable Nitrogen Forms in Runoff from Natural

Forested and Mixed Urban/Commercial/Residential Watersheds. Due to the higher

relative amounts of labile nutrients in developed landscapes, deleterious nutrient

load often exceeds that which would be inferred by absolute increases in nutrients

alone. Reductions in refractory organic nutrient-bearing colored dissolved organic

matter may also increase algal productivity by increasing transparency.

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

Natural Urban

Co

ncen

trati

on

. m

g/L

Total Inorg. N

Labile Org. N

Refractory Org. N

214%

Increase

in TN

663% Increase in

Bio-available N

109

The addition of organic carbon and nutrient forms provides the necessary enhancement to our

understanding of the external load to permit the development of hypothesis related to the natural

and altered productivity states of the LSJR estuary, in particular as it relates to the TMDL

response variables: dissolved oxygen, chlorophyll a and turbidity. The analysis here suggests

that in its natural state, the LSJR saw considerably lower levels of labile organic carbon and

nutrients. The un-altered condition of forested landscapes and contiguous riparian areas also

probably resulted in a river that was more darkly colored. Based on 1995-99 annual averages,

the present anthropogenic load of total labile nitrogen of 6,407 MT/yr represents a 374 percent

increase over the natural background labile nitrogen load of 1,720 MT/yr. For labile

phosphorus, the 1995-99 mean annual anthropogenic load of 820 MT/yr represents a 297 percent

increase over the natural background load of 276 MT/yr. One would assume that this degree of

loading over background would result in much higher levels of internal algal production,

although the amount of this increased production cannot be determined from the annual average

external load alone. With regard to the pre-development oxygen regime that the estuary

experienced, not only would lower levels of algal production have a positive effect, but the lower

incoming load of LTOC would also be expected to result in lower imported BOD.

A significant, seasonal component exists in the concentrations of organic carbon forms that occur

in runoff, and such seasonal effects have been observed elsewhere. In Scandinavian boreal

forests, DON concentrations were observed to increase as spring snow melt-induced water table

rise brings water table to surface (Stepanauskas et al., 2000), and Leff and Meyer (1991), in an

examination of DOC patterns in the Ogeechee River in Georgia, found concentrations to

increases with increasing flow. With regard to bioavailability, Moran et al. (1999) found the

highest per mole C rates of utilization in August for the Satilla River in Georgia. McLatchey and

Reddy (1998) attribute such seasonal changes as largely due to the annual pattern in soil

temperature and soil redox changes owing to saturation and decomposition. With declining free

energy (redox potential) there is decreased OM decomposition, hence greater OM export; greater

P mobilization; higher substrate quality of the exported organic matter; and enhanced NH4

mobilization. Other important factors may include the timing of forest litterfall, the timing of

110

temperatures leading to litter decomposition, or the occurrence of leaching rains following

litterfall.

In the examination of multiple regression-developed land use loading rates, it is clear (and

perhaps intuitive) that intensification in land use, in particular associated with urbanization, leads

to an increase in the concentration and export of labile organic matter and a decrease in the

refractory form. While the increase in labile organic matter is clearly detrimental, the

implications of a decrease in RTOC are less clear. RTOC is assumed to be synonymous with

colored dissolved organic matter, and the effects of CDOM on aquatic systems are to increase

color, decrease pH, as well as to possibly exert some antibacterial properties. Higher prevailing

color levels in the lower St. Johns River has been suggested to favor the dominance of the low-

light adapted native submersed grass Vallisneria americana (Dean Dobberfuhl, pers.

communication) allowing it to fix carbon at lower light levels and thus out compete

phytoplankton for nutrients.

While the high concentrations calculated for LTOC in dairy runoff are probably not surprising,

the relatively higher concentrations calculated for urban runoff are. Possible sources in the urban

and residential environment could include malfunctioning septic tanks, sanitary sewer line leaks,

pump station overflows, animal feces, grass clippings, industrial waste or perhaps even

hydrocarbons. The implications of this are that while urban runoff tends to exhibit moderate

levels of organic nutrient concentrations, it is likely that much of this is in the labile fraction and

hence potentially more detrimental to receiving waters. The differentiation and inclusion of

labile organic carbon and nutrients into the total bioavailable pool increases the relative harm

that this form of land use development perpetrates over what was previously believed.

The greatest change in our understanding of the relative sources of bioavailable nutrients to the

lower St. Johns that comes with the inclusion of labile carbon and nutrients is the apparent huge

contribution of the upper and middle St. Johns. Based on only inorganic nutrient concentrations,

Hendrickson and Konwinski (1998) estimated its contribution to be much lower, providing only

20% of TIN and 11% of PO4. With the inclusion of labile organic N and P, the upper and middle

St. Johns can be seen to contribute 61% of the whole LSJR bioavailable N and 28% of

111

bioavailable P. Much of this bioavailable N is produced internally through atmospheric nitrogen

fixation by blue-green algae in Lake George (Paerl, 2002; Phlips and Cichra, 2001) during long

residence times that favor the depletion of bioavailable N stores. Thus, while nitrogen control

has been one of the most widely touted management strategies in the oligohaline and mesohaline

portions of the LSJR to reduce the potential for nuisance algal blooms in this nitrogen-limited

estuary, the most successful approach to reducing nitrogen in the upstream load delivered to the

estuary may well be reduction in phosphorus load to the upper and middle St. Johns River.

The discovery of the potential sequestration of the greater part of the historic bioavailable

nitrogen (and presumably also phosphorus) supply to the St. Johns by water hyacinth

(Eichhornia crassipes) necessitates a fundamental re-definition of the “natural background” load

to the LSJR. Even prior to the introduction of water hyacinth, Bartram (1792) noted in his

travels of Florida in 1774, in the reach of the LSJR between Green Cove Springs and Palatka,

vast quantities of the floating aquatic plant water lettuce (Pistia stratiotes). He wrote:

“I set sail on early, and saw, this day, vast quantities of Pistia stratiotes . . .

forming most delightful green plains, several miles in length, and in some

places a quarter of a mile in breadth”

Accounts such as this indicate that vast areas of floating aquatic plants existed in the pre-colonial

LSJR, and that river ecology was based upon the substantial utilization of nutrients by

macrophytes, rather than phytoplankton.

Beginning in 1949, chemical aquatic weed spraying with 2,4-D began in an effort to eradicate

water hyacinth. It was not until the late 1960’s that the exotic weed, by this time wildly out of

control due no doubt by the increased nutrient load to which the river was now subjected, was

reduced to manageable levels. Today, hyacinth is kept in check through continuous aquatic

weed spraying. This undertaking, along with continued urban, industrial and agricultural

development within the basin, has undoubtedly transformed the river into the phytoplankton-

dominated system that it is today. The characterization of the incoming, natural background load

to the LSJR as phytoplankton-dominated should be viewed as a device to establish a common

112

currency for evaluating nutrient enrichment and control efforts against established,

phytoplankton chlorophyll-a based standards. The apparent importance of water lettuce in the

pre-colonial LSJR suggests that a balanced ecosystem restoration approach should not only

address nutrient reduction to reduce the severity of algal blooms, but should also assess to the

role of floating and submersed, rooted aquatic macrophytes in maintaining water quality and

organism health and diversity.

113

Literature Cited

Adamus, C.L. and M.J. Bergman. 1995. Estimating nonpoint source pollution loads with a GIS

screening model. Wat. Res. Bull. 31(4): 647-655.

Amon, R.M.W. and R. Benner. 1996. Bacterial utilization of different size classes of dissolved

organic matter. Limnol. Oceanogr. 41(1) 41-51.

Boynton, W.R., J.H. Garber, R. Summers and W.M. Kemp. 1995. Inputs, transformations and

transport of nitrogen and phosphorus in Chesapeake Bay and selected tributaries. Estuaries

18(1B): 285-314.

Bushaw, K.L et al. 1996. Photochemical release of biologically available nitrogen from dissolved

organic matter. Nature 381:404-407.

Cerco, C. F. and T. Cole. 1995. Users Guide to the CE-QUAL-ICM Eutrophication Model.

U.S. Army Corps of Engineers Waterways Experiment Station, Vicksburg, MS.

Chapra, S.C. 1997. Surface Water Quality Modeling. McGraw Hill, New York. 844 pp.

Coveney, M. 1997. P sedimentation rate for L. Apopka.

DeBusk, W.F., J.R. White and K.R. Reddy. 2001. Carbon and nitrogen dynamics in wetland

soils, In, Modeling Carbon and Nitrogen Dynamics for Soil Management , M.J. Shaffer, L.

Ma and S. Hansen, eds. CRC Press, Boca Raton, FL.

Dietrich, T.S. 1978. The urbanization of Florida’s population: A historical perspective on county

growth 1830-1970. Bureau of Economic and Business Research, University of Florida,

Gainesville, FL.

Florida State Board of Health. 1948. Report on the St. Johns River pollution survey. Bureau of

Sanitary Engineering, Jacksonville, Fla. 59 pp.

Florida State Board of Health. 1951. One River, One Plan – St. Johns River Basin: Report on

Water Pollution Control. Water Pollution Series #27. Tallahassee, FL.

Gardner, W.S., R. Benner, R.M.W. Amon, J.B. Cotner Jr., J.F. Cavaleto and F.R. Johnson. 1996.

Effects of high molecular weight dissolved organic matter on nitrogen dynamics in the

Mississippi River plume. Mar. Ecol Prog Ser. 133: 287-297.

Goldman, J.C., D.A. Caron and M.R. Dennett. 1987. Regulation of gross growth efficiency and

ammonium regeneration in bacteria by substrate C:N ratio. Limnol. Oceanogr. 32(6):

1239-1252.

Goolsby, D.A., W.A. Battaglin, B.T. Aulenbach and R. P. Hooper. 2001. Nitrogen input to the

Gulf of Mexico. J. Environ. Qual. 30: 329-336.

114

Harper, H. 1994. Estimation of Stormwater Loading Rate Parameters for Central and South

Florida. Environmental Research and Design, Inc. Orlando, FL.

Hartigan, J.P., T.M. Quasebarth and E. Southerland. 1982. Calibration of NPS loading factors.

J. Env. Eng. 109(6): 1259-1272.

Hedges, J.I., G.L. Cowie, J.E. Richey, P.D. Quay, R. Benner, M. Strom, B.R. Forsberg. 1994.

Origins and processing of organic matter in the Amazon River as indicated by

carbohydrates and amino acids. Limnol. Oceanogr. 39(4): 743-761.

Hendrickson, J.C. and J. Konwinski. 1998. Seasonal Nutrient Import-Export Budgets for the

Lower St. Johns River, Florida. Final report for Contract WM598, Florida Department of

Environmental Protection. 109 pp.

Hendrickson, J.C., C. Cerco and J. L. Martin. 2001. Photo-decomposition model for colored

dissolved organic matter in a blackwater estuary. In, Warrick, J.J., (Editor), American

Water Resources Association Annual Spring Specialty Conference Proceedings. “Water

Quality Monitoring and Modeling.” American Water Resources Association Conference,

Middleburg, Virginia, TPS-01-1, 284 pp.

Hendrickson, J., E. Lowe, D. Dobberfuhl, P. Sucsy and D. Campbell. 2003. Characteristics of

Accelerated Eutrophication in the Lower St. Johns River Estuary and Recommendation of

Targets for the Achievement of Water Quality Goals to Fulfill TMDL and PLRG

Objectives. Water Resources Department Technical Memorandum No. __, St. Johns River

Water Management District, Palatka, FL.

Jaworski, N.A., P.M. Groffman, A.A. Keller, J.C. Prager. 1992. A watershed nitrogen and

phosphorus balance: The upper Potomac river basin. Estuaries 15(1): 83-95.

Kaplan, L.A. and J.D. Newbold. 1995. Measurement of streamwater biodegradable dissolved

organic carbon with a plug-flow bioreactor. Water Res. 29:2696-2706.

Magnien, R.E., R.M. Summers and K.G. Sellner. 1992. External nutrient sources, internal

nutrient pools, and phytoplankton production in Chesapeake Bay. Estuaries 15(4): 497-

516.

Mann, K.H. 1988. Production and use of detritus in various freshwater, estuarine, and coastal

marine ecosystems. Limnol. Oceanogr. 33(4) 910-930.

Mannino, A. and H.R. Harvey. 2000. Biochemical composition of particles and dissolved

organic matter along an estuarine gradient: Sources and implications for DOM reactivity.

Limnol. Oceanogr. 45(4): 775-788.

McLatchey, G. P. and K. R. Reddy. 1998. Regulation of Organic Matter Decomposition and

Nutrient Release in a Wetland Soil. J. Env. Qual. 27: 1268-1274.

115

Meyer, J.L., R.T. Edwards, R. Risley. 1987. Bacterial Growth on Dissolved Organic Carbon

from a Blackwater River. Microb. Ecol. (1987) 13: 13-29.

Moody, 1970.

Moran, M.A., R.E. Hodson. 1989. Formation and bacterial utilization of dissolved organic

carbon derived from detrital lignocellulose. Limnol. Oceanogr. 34 (6):1034-1047.

Moran, M.A., R.E. Hodson. 1990. Bacterial production on humic and nonhumic components of

dissolved organic carbon. Limnol. Oceanogr. 35(8): 1744-1756.

Moran, M.A., W.M. Sheldon Jr., J.E. Sheldon. 1999. Biodegradation of Riverine Dissolved

Organic Carbon in Five Estuaries of the Southeastern United States. Estuaries 22(1): 55-

64.

National Research Council. 2000. Clean Coastal Waters: Understanding and Reducing the

Effects of Nutrient Pollution. Ocean Studies Board and Water Science and Technology

Board, Commission on Geosciences, Environment, and Resources. National Academy

Press, Washington D.C. 405 pp.

Odum, H.T. 1953. Dissolved Phosphorus in Florida Waters. Report to the Florida Geological

Survey, Report of Investigations, No. 9. Tallahassee, FL.

Ogura, N. 1975. Further studies on decomposition of dissolved organic matter in coastal

seawater. Mar. Biol. 31: 60-62.

Paerl, H.W. 2001. Production and Nutrient Limitation of Harmful Algal Blooms in the St. Johns

River. Progress report to the St. Johns River Water Management District under Contract

No. SD154RA.

Paerl, H.W., R.L. Dennis, D.R. Whitall. 2002. Atmospheric deposition of nitrogen: Implications

for nutrient over-enrichment of coastal waters. Estuaries 25(4b): 677-693.

Phlips, E.J. and M.F. Cichra. 2001. Plankton Communities of the Lower St. Johns River.

Annual Report to the St. Johns River Water Management District, Contract #97W165.

University of Florida Department of Fisheries and Aquatic Sciences, Gainesville, FL.

Pierce, E.L. 1947.

Pollman, C.D. and S. Roy. 2003. Examination Of Atmospheric Deposition Chemistry and Its

Potential Effects On The Lower St. Johns Estuary: Final Report Submitted To St. Johns

River Water Management District, Contract No. SE706AA. Tetra Tech, Inc. Gainesville,

FL.

Poor, N, R. Pribble, and H. Greening. 2001. Direct wet and dry deposition of ammonia, nitric

acid, ammonium, and nitrate to the Tampa Bay Estuary, FL, USA. Atmos. Environ. 35:

3947 – 3955.

116

Rao, D.V., S.A. Jenab and D. Clapp. 1997. Rainfall Analysis for Northeast Florida: Summary of

Monthly and Annual Rainfall Data Through 1995. St. John River Special Publication SJ97-

SP22.

Rawlings, M.K. 193_ . Hyacinth Drift.

Raymond, P.A. and J.E. Bauer. 2001. DOC cycling in a temperate estuary: A mass balance

approach using natural 14C and 13C isotopes. Limnol. Oceanogr. 46(3): 655-667.

Reddy, R. and _ Debusk. 1987. Hyacinth N content.

Rushton, B. 1997. Processes that effect stormwater pollution. Proceedings, 5th

Biennial

Stormwater Research Conference. Southwest Florida Water Management District,

Brooksville, FL.

Seitzinger, S.P. and R.W. Sanders. 1997. Contribution of dissolved organic nitrogen from rivers

to estuarine eutrophication. Mar. Ecol. Prog. Ser. 159: 1-12.

Sondergaard, M. and M. Middelboe. 1995. A cross-system analysis of labile dissolved organic

carbon. Mar. Ecol. Prog. Ser. 118: 283-294.

Stepanauskas, R., L. Leonardson and L.J. Tranvik. 1998. Bioavailability of wetland derived

DON to freshwater and marine bacterioplankton. Limnol. Oceanogr. 44(6): 1477-1485.

Stepanauskas, R., H. Laudon and N. O. G. Jorgensen. 2000. High DON bioavailability in boreal

streams during a spring flood. Limnol. Oceanogr., 45(6): 1298-1307.

Strauss, E. A. and G. A. Lamberti. 2000. Regulation of Nutrification in Aquatic Sediments by

Organic Carbon. Limnol. Oceanogr. 45(8):1854-1859.

Strickland, J.D.H. 1960. Measuring the production of marine phytoplankton. Bull. Fish. Res.

Bd. Canada 122. 172 pp.

Sucsy, P and F. Morris. 2001. Hydrodynamics report.

Sun, L., E.M. Perdue, J.L. Meyer and J. Weis. 1997. Use of elemental composition to predict

bioavailability of dissolved organic matter in a Georgia river. Limnol. Oceanogr. 42(4):

714-721.

Timpe, M.P. 1999. Impacts of atmospheric deposition on stormwater quality. Proceedings, 6th

Biennial Stormwater Research and Watershed Management Conference. Southwest

Florida Water Management District, Brooksville, FL.

Toth, D.J. 1993. Volume 1 of the Lower St. Johns River Basin Reconnaissance: Hydrogeology.

Technical Report SJ93-7, St. Johns River Water Management District, Palatka, Florida.

117

Tranvik, L.J. 1990. Bacterioplankton Growth on Fractions of Dissolved Organic Carbon of

Different Molecular Weights from Humic and Clear Waters. Applied and Environmental

Microbiology, 56(6):1672-1677.

Thurman, E.M. 1985. Organic Geochemistry of Natural Waters. Junk, Boston.

U.S.A.C.E. Annual Reports to Congress, 1930 – 1967.

U.S.A.C.E. 1948, November 1. Comprehensive Survey for Removal of Water Hyacinths and

Other Marine Vegetable Growths Interim Report, Serial No. 32. United States Army Corps

of Engineers South Atlantic Division Field Committee, Office, Division Engineer, South

Atlantic Division, Atlanta, Georgia.

U.S. EPA. 1984. Result of the Nationwide Urban Runoff Program: Volume 1 Final Report.

Water Planning Division Report No. PB84-185552.

U.S. EPA, 2002. Nutrient Criteria for SE Ecoregion Streams.

University of Florida. 2000. Florida Estimates of Population. Bureau of Economic and Business

Research. Gainesville, FL.

Valiela, I., K. Foreman, M. LaMontagne, D. Hersh, J. Costa, P. Peckol, B. DeMeo-Andreson, C.

D’Avanzo, M. Babione, C. Sham, J. Brawley and K. Lajtha. 1992. Couplings of watersheds

and coastal waters: Sources and consequences of nutrient enrichment in Waquiot Bay,

Massachusetts. Estuaries 15(4): 443-457.

Volk, C.J., C.B. Volk and L.A. Kaplan. 1997. Chemical composition of biodegradable dissolved

organic matter in streamwater. Limnol. Oceanogr. 42(1): 39-44.

Wetzel, R.G. 1983. Limnology. Second Ed., CBS College Publishing, New York, NY. 767 pp.

Wetzel, R.A. and G.E. Likens. 1990. Limnological Analyses. 2nd

Ed. Springer-Verlag, New

York. 391 pp.

tmdl and plrg modeling of the lower st. johns river...

Documents