Appendix 4 - view of Stormwater Loading Rates from the Available Literature

(1) Stormwater Loading Rates . December 1989. FDER Nonpoint Source Section staff.

Table A-8 was given to Northwest Florida Water Management District staff for comparison purposes. The numbers were not, however, documented. Values were included in the table matrix, but not considered for use in this study, since lack of documentation made it impossible to review data collection duration and watershed characteristics (homogeneity, etc.).

(2) Estimation of Loading Rate Parameters for the Tampa Bay Watershed 1989. Southwest Florida Water Management District [SWFWMD].

The literature search associated with this report was extensive, involving the review of approximately 100 reports and publications. Studies that were selected presented at least one year of data collection and measured only homogenous watershed areas. The report presents the results of a literature search and study of pollutant loading rates appropriate for selected parameters and land use types within the Tampa Bay watershed. The pollutant parameters selected for potential inclusion into a loading rate model were total nitrogen, total phosphorus, BOD, suspended solids, lead, zinc, total coliform, pesticides, and oil and grease. The selected land use types included low density residential (rural), single-family residential, multi-family residential, low density commercial, high density commercial, industrial, mining, agriculture, recreation/open space and water/wetlands.

Each selected study was evaluated for adequacy of the database with special attention given to factors such as length of study, number of runoff events monitored and monitoring methodology. The applicability of hydrologic and land use conditions were evaluated and data judged appropriate for the Tampa Bay area was used for the final estimation of pollutant loading rates.

The report included a summary table of mean annual loading rates for total nitrogen, total phosphorus, BOD, and suspended solids. Values were available for each of these parameters for each of the land use categories. Because of the vast amount of information researched for this report and the attention paid to data collection (sampling duration and watershed homogeneity), many of the final loading rate estimates for the St. Marks and Wakulla Rivers nonpoint assessment were obtained from this report.

(3) Nonpoint Source Waste Load Study for Indian River, Titusville, Florida. September 1985. Dyer, Riddle, Mills and Precourt, Inc.

The purpose of this study was to estimate the quantity of NPS pollutants in lbs/year entering the Indian River in Brevard County from 28 drainage basins. The following sources were used to determine the loading rates for total nitrogen, total phosphorus, BOD and suspended solids for each respective land use:

a) USGS studied a residential area in Broward County and multiplied the loads (lb/acre/in) by 56-inches of annual rainfall in Titusville;

b) East Central Florida Regional Planning Council (ECFRPC) provided loads (then multiplied loads using the rainfall ratio 56/50.5);

c) USGS studied commercial area in Broward County and multiplied loads (lb/acre/in) in by 56-inches of annual rainfall in Titusville;

d) ECFRPC "Middle St. Johns River Basin Study and Lake Tohopekaliga Agricultural Runoff Plan with correction to 56-inches of annual rainfall;

e) ECFRPC - Nonpoint Source Evaluation, Chapter 10, Table 10-2 with correction to 56-inches of annual rainfall.

The results of the calculations do not take into account the impact of the city's stormwater management and conservation ordinance. NPS data from developments constructed before and after the ordinance would be required to determine what impact the ordinance has on loading rates. Also included were the estimated loads to the Indian River for the proposed land use for the City of Titusville as a prediction of what can be expected if the future land use is carried to completion. This report was the source for the St. Marks and Wakulla Rivers watershed loading rate estimates for the land use categories cropland/pasture, forest natural upland, and transportation/utilities.

(4) Agriculture Nonpoint Source Element - State Water Quality Plan. June 1979. Florida Department of Environmental Regulation.

This report contained site-specific analyses of the five watersheds chosen from the top 20 ranked within the state based on their potential for water quality problems due to agricultural activities. In this publication total annual loadings for the five specific watershed areas were calculated from land use and areal loading rates. Also contained in the report were site-specific analyses for each watershed, methods used for determining the extent of water quality problems caused by agricultural activities and management practices for pollution control. The five watersheds included in the study were Escambia River Basin (Canoe Creek), North St. Lucie Basin (Ten/Eleven Mile Creek), Upper Oklawaha River Basin (Yale-Griffin Canal), Middle Suwannee River Basin (Little River), and the Chipola River Basin (Spring Branch). For the Canoe Creek study, a cautionary note that stated the pollutant yields per-acre may not be accurate enough as they are based on literature values. It was also stated that a wide variability exists between the estimates and the actual loadings observed in the field. It was recommended that a receiving water analysis be performed in order to demonstrate that the loads were, in fact, producing sizable water quality problems.

(5) Nonpoint Source Effects. January 1976. M.P. Wanielista.

The results of a literature review on loading rates for BOD, suspended solids, nitrogen, and phosphorus for several land uses are included in this document. The average value is used to assess the potential quality problems resulting from NPS. It should be emphasized that studies from Florida only were used in determining the loading rates. As much data as possible, relative to local or regional conditions, should be used. Also, the loading rates are average values and are adjusted for an average rainfall quantity of 52 inches, which is the approximate average for the State of Florida. The range of the loading rates for suspended solids was wide, but expected because of variable erosion conditions. Additional research should be done on quantification of loading rates per unit area and time.

This report states that land use and precipitation are the two most important variables affecting NPS pollution. Assuming precipitation is of sufficient intensity, duration, and quantity, runoff and infiltration quantity and quality are dependent on ground cover, frequently expressed in terms of land use such as, urban, agricultural, pasture, forested (woodland) wetlands, etc. Each land use has certain permeable and impermeable characteristics that determine, to a great extent, the quantity of runoff, evapotranspiration, and groundwater infiltration. The type of land use is frequently equated to a runoff coefficient and used in predicting runoff. Rainfall and the type of land use are relatively easy variables to quantify in Florida. If large areas, such as major basin segments, are to be used, then the use of average conditions for rainfall and land use would produce data that are reasonable. After justifying the need for more accurate nonpoint data, a program to collect additional information is necessary. This information would include data in the general areas of topography, land management, water quality impacts, and meteorology.

General goals for the management of nonpoint source effects due to the land and water use modifications are presented below. These goals are the basis for the work presented in the report. The goals providing this basis include the following:

a) to minimize the deleterious effects of runoff and infiltration from land and water alteration activities on receiving waters by:

-minimizing the alteration of natural drainage patterns and conditions;

-minimizing soil loss due to erosion;

-maximizing the use of the soil's infiltration and percolation capacity to reduce the unnatural loss of surface waters within the individual watersheds under development;

-maximizing the use of management practices to reduce the levels of contaminants in the waters to be discharged;

-encouraging the gradual release of runoff and infiltration into the receiving bodies of water.

b) to encourage the development and implementation of all plans to be consistent with comprehensive regional water quality/quantity management plans;

c) to encourage information flow and to foster an understanding of the impact of NPS on water quality and to develop implemental solutions to these problems;

d) to keep abreast of developments in the state-of-the-art techniques of managing pollution from NPS.

 

(6) Water Quality Characteristics of Urban Runoff and Estimates of Annual Loads in the Tampa Bay, FL. 1984. M.A. Lopez and R.F. Giovannelli.

The purpose of this report was to describe the water quality characteristics of urban runoff in the Tampa Bay area and to provide a method for estimating loads of substances contained in runoff from urbanized watersheds under existing and future conditions. From 1975 to 1980, an urban runoff data collection program, including streamflow, climatic, physiographic, and water quality data, was established at nine watersheds ranging from beginning to advanced stages of urban development. Gaging stations were installed to monitor rainfall and runoff for each watershed. Physiographic features that consist of size, shape, and slope of the watershed, type of land use, degree of land use, area of impervious surfaces, type of storm drainage, soil types, and surface area of lakes or detention ponds were compiled from aerial photographs, US. Geological Survey topographic maps, planning agency data, and field observations.

Regression equations were developed for estimating loads of substances contained in runoff from unaged urban watersheds in the Tampa Bay area. Equations were developed for BOD, COD, total nitrogen, total organic nitrogen, total phosphorus, and total lead. Use of the regression equations requires watershed-based rainfall. The stormwater loads were computed as the product of instantaneous values of discharge rates and concentrations of water quality constituents sampled. A time factor was included, converting the discharge to incremental volumes, and summed to get the total stormwater runoff volume. Some precautionary measures are required for valid use of the regression equations for computing runoff volumes and water quality constituent loads. They are recommended to be applied only to Tampa Bay urban watersheds. Since the equations are empirically derived, they are only applicable in instances where values of the independent variables fall within range of values used in their formulation.

The same daily rainfall data was used in computations of runoff and annual loads for all watersheds. The differences in the computed runoff from the various watersheds were due to variations in land use and watershed characteristics. The differences in runoff are reflected in the magnitude of the computed basin loads of the various parameters involved (runoff volume enters directly into load calculations).

Selected water quality constituents computed by the Tampa Bay area regression equations were compared with those computed by other methods for other parts of the country. The annual loads computed, using the Tampa Bay area regression equations and the Broward County land use load factors, are of the same order of magnitude. Another comparison was made in the application of the U.S. Environmental Protection Agency (USEPA) screening procedure (Heaney and others 1976) to land use data at the St. Louis Street drainage ditch site. As determined by use of the Tampa Bay regression equations, the annual load of nitrogen was about the same, BOD was about one-half the magnitude, and total phosphorus was about one order of magnitude greater than loads estimated by the screening process. The Tampa Bay regression equations, because of their empirical nature, presumably reflect the high natural phosphorus content of bay area streamflow and fallout from local phosphate processing plants.

(7) Boggy Creek Water Quality Management Study, Final Report. 1988.South Florida Water Management District and East Central Florida Regional Planning Council

This study used a land use/nonpoint source model to develop control strategies on a regional scale. Developed by Camp, Dresser and McKee (CDM) (1988), this model simulated total nitrogen and total phosphorus loadings from the existing relatively rural watershed and then predicted loadings in future scenarios where the watershed was presumed to be highly developed, where residential, commercial and industrial uses predominate. The model was also run for low density future conditions. A range of BMP nutrient removal efficiencies and calibration coefficients that represent the extremes of the probable ranges for these coefficients was applied in the model. As part of the study, the South Florida Water Management District (SFWMD) developed a maximum assimilative capacity (MAC) model to be used to compare the maximum assimilative nutrient capacity of the receiving waters with the nutrient loadings modeled by CDM.

Three land use/growth scenarios were developed by the East Central Florida Regional Planning Council utilizing current data and future projections as supplied by local governments within the basin. These scenarios include the following:

1) Existing land uses in the basin were modeled (1984) which included certain phases of Developments of Regional Impact (DRIs) which were present, and any other pre-existing development;

2) " Low Intensity" scenario - Assumes a slower growth rate and includes all future development proposed within a DRI by the year 2000. This scenario also includes all future development associated with the DRIs proposed, but not necessarily within the DRI;

3) "High Intensity" scenario - Assumes an intensified development rate over that of the low intensity scenario concentrating primarily in the rural southern and eastern fringes of the basin with the inclusion of the proposed southern connector beltway around the City of Orlando, and the ensuing growth associated with it.

In both future growth scenarios (high and low), wet detention was utilized in the model as the minimally-acceptable method of stormwater management, since it meets existing SFWMD and DEP permitting criteria and is currently the most commonly employed stormwater management practice.

Data collection consisted of gathering the information necessary to run both water quality impact models (East Lake Toho MAC model and Land Use/Nonpoint Source model). Nutrient loadings were determined using data from a USGS-maintained daily stage/flow recorder located in the southern portion of the basin below the Boggy Creek swamp combined with data received from the SFWMD monthly water quality station located downstream from the recorder. Nitrogen and phosphorus values were utilized in the model calibration of current land use runoff loadings for the entire Boggy Creek basin. Rainfall values (daily) were measured from the NOAA rain-gage at the Orlando International Airport (the monthly cumulative values for the period 1981-1985 were used in the study). For runoff calculations, the USGS stages were converted to flow rates using the rating curve.

Description of Boggy Creek Model

Maximum Assimilative Capacity (MAC) Model - SFWMD used this modified Vollenweider (1976) model (Federico et. al 1981) which was applied to East Lake Toho to derive a preliminary approximation of the maximum assimilative capacity for phosphorus (the limiting nutrient) within the lake. This was based on per-basin surface area loading appropriation. The basic concept of the MAC model was to determine whether East Lake Toho would maintain its current mesotrophic status or become degraded upon urbanization. The MAC model has been used with success by SFWMD for Lake Okeechobee. This model was calibrated utilizing data collected during a three-year period. For output, the model determines the current and critical loading rates from the East Lake Toho basins. The current loading rate refers to the existing phosphorus contributions and the critical loading rates which are derived from the model for each basin and compared with the actual measured loads generated to determine whether a particular basin is exceeding its critical phosphorus loading rate. If the critical phosphorus loading rate is being exceeded, it is then possible to calculate the load reduction which would be required to stay within the means of the maximum assimilative capacity of the system.

The Description of the Land Use/Nonpoint Source Model (CDM)

The intent of this study was to develop a model capable of determining NPS control strategies on a regional scale. A relatively simple spreadsheet model, previously developed by CDM, was employed using the LOTUS Symphony program (version 1.1) for an IBM compatible computer. Data used in the preparation of this model was provided by SFWMD and ECFRPC. This land use/nonpoint source model simulated total nitrogen (TN) and total phosphorus (TP) loadings resulting from the existing rural watershed and then predicted loadings in future scenarios where the watershed was presumed to be highly developed, where residential, commercial and industrial land uses predominated. TN and TP were the only parameters used in the model. The major inputs needed to calculate nutrient loads included nutrient load calibration coefficients (Fn and Fp), rainfall conditions, areal extent of land, and BMP nutrient removal efficiency factors (REFs). The REFs assume that properly managed stormwater systems using BMPs will reduce nitrogen loadings by 30 percent and phosphorus loadings by 50 percent (from a particular land use type). Event mean concentrations (EMCs), which are standardized concentrations of nitrogen and phosphorus expected from a particular land use type, are combined with total runoff and nutrient load calibration coefficients to generate nutrient loadings from the current and future scenarios. The EMCs used were previously employed in other central Florida area studies. The purpose of the calibration is to more accurately predict nutrient loadings for the three development scenarios by comparing measured loadings to modelled loadings. The ratio of these is the nutrient loading calibration coefficient. Prior to model calibration, nutrient loadings were determined using the existing land use, and then re-running the model to predict future low and high intensity development scenarios. Output consisted of annual loadings for each land use type (pounds/acre/year) which impacts Boggy Creek and East Lake Toho. The model was run for the current, low density future, and high density future conditions with a range of BMP nutrient removal efficiencies and calibration coefficients that represent the extremes of the probable ranges for these coefficients. The predicted current and future nutrient loading rates in the land use/nonpoint source model were analyzed for sensitivity to changes in the BMP nutrient removal efficiencies and the nutrient loading calibration coefficients.

Several management alternatives were suggested for local governments to consider in an effort to reduce nutrient loadings to Boggy Creek and East Lake Toho:

1) The development and implementation of master stormwater management plans;

2) The establishment of regional stormwater systems;

3) The amendment of local government stormwater regulations to include stricter controls in certain land use types, the promotion of local land use controls such as land acquisition and the protection of the extensive Boggy Creek swamp.

(8) An Assessment of Urban Land Use/Stormwater Runoff Quality Relationships and Treatment Efficiencies of Selected Stormwater Management Systems. 1988.- South Florida Water Management District

This assessment was initiated to address two objectives relative to stormwater runoff and its treatment. The first objective was to assess reported stormwater runoff quality for differing land uses throughout the U.S., with a focus on data collected within the State of Florida. The second objective of this publication was to evaluate the data reported in the literature concerning the treatment efficiencies associated with the various stormwater management systems.

One conclusion of this report was that for selected constituents, runoff water quality varies with land use. The land use types that were evaluated and compared in this assessment included residential, commercial, light industrial, roadway, and mixed urban. Statistical differences between runoff water quality parameters and land use classification were evaluated by using the Duncan's multiple-range test. This study concluded that higher nutrient loads were generated by residential land uses than commercial, mixed urban, light industrial, or roadways. Metal contamination was more widespread from commercial and roadway projects than from residential, light industrial, or mixed urban land uses. Residential and roadway areas demonstrated higher export potential for chemical oxygen demand. There were no discernible trends for suspended solids export as a function of land use. Urban highway projects generally had higher overall pollutant loadings than rural roadway projects. Limited data indicated that organic contamination in the form of polycyclic aromatic hydrocarbons were comparable for residential, commercial, and highway sites, although significantly higher levels were found at a heavy industrial site. For any monitoring program designed for a specific land use, this report recommended utilizing the above information when designing a sampling strategy.

(9) Generalized Watershed Loading Functions for Stream Flow Nutrients. June 1987. D. A. Haith and L. L. Shoemaker.

The Generalized Watershed Loading Functions Model (GWLF) is based on simple runoff, sediment, and ground water relationships combined with empirical chemical parameters. According to the authors, this model is unique in its ability to estimate monthly nutrient fluxes in streamflow without calibration. Validation studies in a large New York watershed indicated that the model possesses a high degree of predictive accuracy. Although better results could perhaps be obtained by utilizing more detailed chemical simulation models, such models have substantially greater data and computational requirements and must be calibrated from water quality sampling data.

The concluding remarks indicate that the GWLF model has several limitations. Peak monthly nutrient fluxes were potentially underestimated by as much as 22 percent. Since nutrient chemistry is not modelled explicitly, the model cannot be used to estimate the effectiveness of fertilizer management or urban stormwater storage and treatment. The model has only been validated for a largely rural watershed in which agricultural runoff and ground water discharge provided most of the nutrient load. Although the urban runoff component is based on a widely used model (STORM), model performance in more urban watersheds is uncertain.

10) Assessing Regional Nonpoint Source Problems with the Aid of a Watershed- Based Simulation Model. 1989. B.M. Evans, N.F. Parks, G.M. Baumer, G.W. Petersen.

In 1979, the Pennsylvania Bureau of Soil and Water Conservation (PABSWC) designated a number of large watersheds throughout the state as "high priority" watersheds for agricultural NPS pollution assessment. In recent years, many county conservation districts have conducted watershed assessments for the high priority areas. One of the primary goals of these assessments is to satisfy the requirements of the "first level screening" process outlined by PABSWC for securing additional funding through the Chesapeake Bay Program to implement Best Management Practices. To date, most of these assessments have been accomplished using an "indexing-type" methodology for identifying critical areas within a watershed. With this approach, variables such as percentage of agricultural land in row crops, percentage of agricultural land in cover crops, animal density, average slope, mean soil erodibility, and drainage density are weighted and evaluated on the basis of their relative impact on NPS contribution within each watershed. While this cost-effective approach has considerable merit, some criticism has been directed toward it because it does not directly produce estimates of nitrogen, phosphorus, and sediment loadings within a certain area.

The intent of this demonstration project was to illustrate how a computerized simulation model could be used to enhance the results of watershed assessments performed in Pennsylvania. Particular emphasis was placed on how a GIS could be used to derive parameters required by the simulation model and display the resulting output. A recently developed model called Agricultural Nonpoint Source (AGNPS) (Young et al. 1985) was used to simulate erosion, sedimentation, and surface runoff processes within a watershed in central Pennsylvania.

The original scope of work planned for a detailed analysis of the simulation modelling results be performed. Specifically, the original plans called for the following:

1) A comparison of model results with results obtained from a previous watershed assessment;

2) An evaluation of predicted nutrient and sediment loadings against water quality sampling data;

3) A sensitivity analysis to determine the relative change in model output with respect to changes in model input.

NPS models may be categorized in numerous ways, but of interest in this study were programmable, mathematical models commonly referred to as numerical models. Numerical models can be further classified as being either empirical or physical process models. Empirical models are generally cause-and-effect models in which a mathematical expression transforms a set of input variables into a description of the output without trying to describe the process taking place. Regression models (Universal Soil Loss Equation related), and statistical time-series models, and several "indexing-type" models are good examples of empirical models. Empirical models are generally simpler and require less data than physical process models, so they are cost-effective to use. Unfortunately, they are difficult to improve, cannot normally be extended beyond the range of data used in their development, are easily misapplied, and can be misleading about cause and effect. Physical process models are more sophisticated than empirical models in the sense that they attempt to simulate the physical, chemical and/or biological processes that take place in a given natural system. Such models, however, require volumes of data and extensive research to test. NPS models such as Chemicals, Runoff, and Erosion from Agricultural Management Systems Model (CREAMS) (Knisel 1980), Areal, Nonpoint Source Watershed Environmental Response Simulation Model (ANSWERS) (Beasley et al. 1980), and more recently, AGNPS (Young et al. 1985) are typical physical process models.

The AGNPS model was developed by the Minnesota Pollution Control Agency, the Minnesota Soil and Conservation Board, the U.S. Soil Conservation Service, and the U.S. Agricultural Research Service (Young et al. 1985). The AGNPS computer program actually contains two sub-models which are used to analyze both sediment and nutrient transport within a watershed. The two sub-models used are single-event based models intended to simulate sediment and nutrient transport from primarily agricultural watersheds. However, land use/cover conditions, in addition to cultivated land, are considered as well. Proceeding from the headwaters of the watershed to the outlet, the pollutants are routed in a step-wise fashion so the flow at any point may be examined. In using this program, a watershed is first sub-divided into square cells. The basic components of the model provide for the analysis of hydrology, erosion, and sediment transport, and nitrogen, phosphorus and chemical oxygen transport.

Data required for model execution is normally obtained through field data collection, maps (topographic and soils), and various technical publications, tables, and graphs. Input data can be classified into two categories: watershed data and cell data. Watershed data includes information pertaining to the entire watershed and to the storm event to be simulated. Cell data includes physical information describing each of the cells, as well as parameters based on the land practices within the cell.

Once the watershed segmentation step was completed, the data input file was established. Compilation of the various types of data and deriving the necessary input parameters to drive the model can be very time consuming, so instead of using this manual approach many of these computations were done with the aid of a GIS. This approach was believed to be less expensive and more accurate than manual data compilation methods.

As a result of the final AGNPS model runs, a detailed watershed summary was generated for the Bald Eagle watershed. The results include analyses for hydrology, with estimates for nitrogen, phosphorus, and chemical oxygen demand in concentration and mass units. Preliminary output includes watershed and cell areas, storm precipitation and erosivity, and estimated values at the watershed outlet for runoff volume, peak flow rate, and a detailed analysis of the sediment and nutrient yields. The detailed sediment analysis include area-weighted erosion rates for both uplands and channels, sediment delivery ratios, mean sediment concentrations, area-weighted yields, and net sediment yields. These values are given for each of the five particle classes, as well as the total. The detailed nutrient analyses include the nitrogen and phosphorus mass per unit area for sediment-absorbed nutrients, the soluble N, P, and COD mass per unit area in runoff, and the N, P, and COD concentration in the runoff. Additional runoff, sediment, and nutrient analyses are given for each cell. Runoff analyses for each cell provide estimates of drainage area, runoff volume, percent of runoff volume entering the cell from above and peak runoff rate. Sediment analyses for each cell provide estimates of upland erosion rate, amount of sediment generated within the cell, the amount entering the cell from above, the sediment yield leaving the cell, and the percent deposition in the cell. The detailed nutrient analysis for each cell provides estimates of absorbed and soluble nutrients in mass per unit area and the concentration of these nutrients in the runoff.

To aid in the evaluation of AGNPS model output, a program called AGNPSOUT was developed. Output maps which can be generated with AGNPSOUT now include the following:

1) Runoff, including drainage area, volume, generation both above and within the cell and peak rate;

2) Sediment delivery, including cell erosion, generated above, generated within, yield, and deposition;

3) Nitrogen production, including sediment delivered (both within cell and at cell outlet), water soluble (both within cell and at cell outlet), total production (sediment plus water), and concentration;

4) Phosphorus production, including sediment delivered (both within cell and at cell outlet), water soluble (both within cell and at cell outlet), total production (sediment plus water), and concentration;

5) Chemical oxygen demand within cell, cell outlet, and concentration.

In summary, it was the intent of this demonstration project to illustrate how AGNPS could be used for watershed assessments performed in Pennsylvania. In this study, the relatively new watershed model (AGNPS) was used to simulate erosion, sedimentation, surface runoff and nutrient transport processes within the Bald Creek watershed. Particular emphasis was placed on showing how a GIS could be used to derive model input parameters and display the resulting output. Additional work to complete the project is to be done at a later date as the model needs to be tested and calibrated for use in Pennsylvania.

(11) Silviculture, Hydrology and Water Quality in the Lower Coastal Plain of the U.S.A. 1988. H. Riekerk.

The loading rates for the land use categories Tree Plantations and Forest Regeneration areas for the St. Marks and Wakulla Rivers nonpoint assessment were derived from the results associated with this study. The study was conducted from 1976- 1985 for three artificial watersheds to study the effects of intensive pine flatbeds forest management practices. The study site was 40 km northeast of Gainesville, Florida, in the Bradford Forest which belonged to a timber company. Slash pine flatwood plots were harvested and regenerated with high and low levels of disturbance, and a 40-year-old forest was left undisturbed as a control. The low-disturbance watershed was cut manually, bucked, and removed leaving considerable slash. To minimize disturbance, the site was quickly regenerated by slash chopping, bedding, and machine planting. The high disturbance was machine harvested, followed by slash burning, windrowing, harrowing, bedding, and planting. Each site was monitored for hydrologic variability as well as water quality impacts due to the silvicultural practices for an eight-year time period. In order to complete the loading rate matrix, it was decided to use this data for the category Tree Plantations as well as Forest Regeneration Areas. The first three years of data from the high-disturbed experimental watershed was averaged for the Forested Regeneration Areas' loadings for nitrogen, phosphorus and suspended sediment. It was decided that during the first three years after a watershed had been disturbed to this degree, the area would closely approximate what has been defined by the "Florida Land Use, Cover and Forms Classification System" (FLUCCS) as a forest regeneration area (evidence of windrows and site preparation). The Tree Plantation loadings for nitrogen, phosphorus, and suspended sediment were taken from the entire eight-year period for the combination of high and low disturbed areas. This decision was based on the idea that tree plantation runoff loadings can correspond to the high/low disturbed areas when cut and prepped, but can also approximate Natural Upland Forests during the non-disturbed growth stages. BOD loadings were not found in any of the silviculture studies researched, so the loadings for Natural Upland Forests and Agriculture were utilized in order to complete the loading rate table.

The following is a summary of loading rates assembled. The loading estimates were obtained from the aforementioned studies.

TN..........TP.......... BOD.......... SS.......... REF #

(lb/ac/yr)

2.25...... 0.34 ..........6.86 .......27.50 ...........(2)

1.50...... 0.26 ..........4.70 .......28.00 ...........(3)

14.27 ....7.60 ........48.89 ......902.26 ..........(1)

4.43 ......0.47 ........10.80 ............................(6)

5.09 ......0.79...........(7)

7.54 ......1.93........ 42.10 ......206.00.......... (2)

6.60...... 2.00 ........32.10 ......344.10 ..........(3)

17.26.... 9.01........ 62.94 ....1081.40 ..........(1)

5.88...... 0.87 ..........(7)

12.03 ....2.10....... 69.65...... 474.50........... (2)

11.30 ....1.10 .......63.00 ......647.00 ...........(3)

10.95 ....5.13 .......47.89...... 725.18 ...........(1)

5.34...... 3.12 .......44.52 ......300.93 ...........(5)

5.45 ......0.51......... (7)

8.00...... 2.00 .......53.90...... 586.50........... (4)

13.60 ...3.00........ 87.40 ......980.90 ..........(3)

10.29 ...6.06 ........91.53 ....1508.95 ..........(1)

4.50..... 0.36 ...........(7)

8.00 .....2.00 .........53.9 .......586.50...........(4)

8.56 .....3.70....... 46.80 .......696.00 ..........(2)

11.30 ...1.10....... 63.00 .......647.00 ..........(3)

12.02 ...8.43....... 47.21..... 1385.25.......... (1)

4.86.... 0.46.......... (7)

1.31... 0.16......... 2.48........... 23.70........ (2)

2.32... 0.21......... 3.70........... 29.12........ (1)

2.55 ...0.32.......... (7)

1.50... 0.10 ........1.60............ 12.00........ (4)

2.60 ...0.11.........7.40.............15.10........ (4)

3.30... 0.15 .......6.70............ 16.30........ (4)

4.52 ...0.67....... 7.05 .............8.60 ........(2)

5.40...0.22.......15.40 ...........32.30........ (3)

4.98.. 0.88 ......16.13........... 33.77........ (1)

4.90...0.20 ......13.90 ...........24.30 ........(4)

7.20 ....1.30 ....16.20 .........434.40 .......(3)

4.70.... 0.27 ....10.00 .........750.00....... (4)

7.54.... 4.60 ....46.02....... 1223.28....... (1)

4.72 ....0.26 ......9.79 .........747.89....... (5)

2.26.... 0.26 .......(7)

11.00.. 0.50 ....14.60 ........393.00 .......(4)

5.60 ...1.00 .....15.10 ........391.70...... (4 )

4.49..... 0.88..... 4.10 .......18.20 .....(2)

4.40..... 0.22 ...16.20 .......27.80..... (3)

23.00... 0.94 ..16.00 ....3747.00..... (4)

23.15. ..0.93 ..16.03 ....1997.03 .....(5)

8.30..... 1.86 .....(6)

34.60... 2.11 ..16.06 ....2721.00 ....(4)

2.30 ...0.10..... 2.60....... 27.80....... (3)

2.76... 0.90 .....4.45....... 87.25 .......(5)

3.30... 0.15..... 6.70.......16.30....... (4)

2.80... 0.09..... 4.50...... 87.00 .......(4)