The city of Indian Harbour Beach is located on central Florida’s eastern barrier island at the convergence of the Banana River and the Indian River Lagoon (IRL), which are actually bays. The city abuts the Atlantic Ocean on the east and the Banana River on the west. Essentially, the whole city is urbanized and developed, with only a few vacant parcels of undeveloped lots. The population is estimated to be over 1,800 people.
The IRL was designated in 1991 as one of the first National Estuaries in the country. The IRL—and, in particular, the subsection named the Banana River—is an important economic resource for Indian Harbour Beach. Over the last 20 years, the health of the river has markedly declined, with significant losses of critical seagrasses and fisheries. The principal cause of degradation of the Banana River has been polluted stormwater runoff.
To address stormwater pollution problems in the IRL, the Florida Department of Environmental Protection (FDEP) is implementing the total maximum daily load (TMDL) program with the following goals:
- Evaluation of the health of the Indian River Lagoon
- Determination of the pollutant loadings entering the water body
- Estimation of the assimilative capacity of the water body to receive pollutants without degradation
- Calculation of the reductions in pollutant levels necessary to avoid exceeding the assimilative capacity of the water body
- Assigning pollutant load allocations/reductions to communities and stakeholders within the watershed
To comply with proposed load allocations, communities will be required to undertake several steps to reduce their stormwater pollutants. These steps include retrofitting existing stormwater infrastructure to add stormwater treatment facilities to the systems, launching public education programs, increasing development regulations, and complying with National Pollutant Discharge Elimination System (NPDES) permits.
FDEP’s Verified List of Impaired Waters for the Banana River indicates that nutrients (phosphorus and nitrogen) and mercury are the principal sources of impairment. In April 2007, the EPA established TMDLs for the IRL and the Banana River. In the Banana River adjacent to the city, the EPA’s TMDL is a 63% reduction of total nitrogen (TN) and 67% reduction of total phosphorus (TP) mass annual loadings from stormwater systems. There are no point-source discharges or septic tanks in Indian Harbour Beach, meaning that the city’s load allocation reductions must come entirely from its municipal separate storm sewer system (MS4).
FDEP is also in the process of establishing TMDLs for the IRL, which will probably mirror EPA’s TMDLs. At this point, community compliance with the TMDL program is voluntary, but FDEP plans enforcement of TMDLs by opening a community’s NPDES Phase II permit and inserting TMDL allocation reductions into that permit.
Recognizing that FDEP would be developing TMDLs in the near future, the city undertook a stormwater quality study with Allen Engineering and Stormwater Solutions Inc. (SSI) to model and quantify stormwater pollutant loadings, compare the pollutant loadings to TMDL allocations, and propose stormwater retrofit projects necessary to comply with TMDL goals. The city was unusual in that it had virtually no flooding problems due to its proximity to the beach and bay; this was strictly a water-quality study.
Pollutant Loading Assessment
Indian Harbour Beach encompasses approximately 1,323 acres. The city has 98 stormwater outfalls to the Banana River and is well blanketed with traditional stormdrain pipes, ditches, curbs, and gutters. There are currently 19 outfalls that have stormwater treatment systems, with the remaining 79 outfalls discharging untreated stormwater to the Banana River.
The city has predominantly residential properties, with a few commercial sites. There are 22 existing stormwater treatment systems in the city, serving 483 acres with either wet detention or dry retention ponds.
The EPA and FDEP used the Pollutant Load Assimilation (PLASM) spreadsheet model and the Hydrologic Simulation Program–Fortran (HSPF) model for calculating pollutant loadings for the TMDLs. These models were used on a region-wide basis, but were not discrete enough to determine individual basin or citywide loadings. While these are highly accurate models that are appropriate for establishing scientifically defendable pollutant load allocations, the expense, complexity, and calibration effort of using these models was beyond the budget of a small town that does not have a geographical information system (GIS) database.
The EPA’s TMDL allocation was for a percent removal of existing stormwater loadings, rather than an effluent limitation or specific mass load. As such, an accurate determination of existing loads and associated model calibration were not necessary. Therefore, SSI recommended the use of a simplified spreadsheet model similar to PLASM that could be used to calculate uncalibrated mass annual loadings of TN, TP, and total suspended solids (TSS) from the city’s stormwater system. A spreadsheet model also enabled easy calculation of load reductions from proposed retrofit projects. Use of a relative spreadsheet model provided the city with considerable savings in model development that will be applied toward implementing proposed projects.
FDEP is in the process of refining Florida’s stormwater design criteria based on the study “Evaluation of Current Stormwater Design Criteria within the State of Florida” (Harper and Baker 2007). The methods used by Harper and Baker to calculate pollutant loadings were used in this study to ensure that the modeling methods would be similar to FDEP’s TMDL study.
Allen Engineering delineated and mapped the city’s drainage basins. Within each drainage basin, the land uses, soil types, C factor, soil curve number (CN), and “Non Directly Connected Impervious Area” were determined. Potential pollutant loads for TP, TN, and TSS for each of the 93 basins was calculated, using the following formula:
PPL parameter = Area × Loading
where
PPL = potential pollutant load (kg/year) for each parameter
Area = drainage basin area (acres)
Loading = pollutant load (kg/acre/year) for each parameter
Localized mass annual pollutant loading rates are critical factors for use of this type of model. Harper’s report allows determination of pollutant loading rates based on local rainfall records, soil types, and land uses for all parts of Florida. Annual runoff volume was calculated by multiplying the mean annual rainfall depth by the area and runoff coefficient C.
Existing stormwater treatment systems and their associated sub-basins were then identified in each basin. FDEP has a BMP database showing event mean concentration (EMC) removal efficiencies based on years of monitoring data in Florida. BMPs in this database were constructed to current design standards, as were the existing BMPs in the city. Therefore, using FDEP’s BMP database provided reasonably accurate estimates for removal efficiencies of the existing BMPs.
Because of the obvious difficulties in quantifying BMP effectiveness when there is no effluent discharge, dry retention ponds are not listed in FDEP’s database. FDEP recommends using Harper’s approach of assuming that 100% of all pollutants from the design storm would be trapped and filtered into the ground. He performed a statistical analysis of rainfall events for various regions of Florida to determine the frequency of storms exceeding the design storm (generally 1 inch of runoff), enabling an annual weighted mass calculation based on how much water percolates into the ground with 100% treatment for small storms versus how much runoff bypasses the system with no treatment in larger storms. Harper developed a series of charts for a range of design storms from 0.25 inch to 4.0 inches of runoff storage. The full report can be viewed at www.florida-stormwater.org.
For analysis of existing dry retention ponds in this report, mean annual mass removal efficiencies for 1.00 inch of retention (the normal design standard) were used from Harper’s study. The removal efficiencies were a function of meteorological zone, CN, Non Directly Connected Impervious Area CN value, and Percent Directly Connected Impervious Area.
The resultant existing pollutant loads per contributing basin of each BMP were calculated with the following equation to account for existing BMP pollutant reductions:
EPL parameter = Σ [Loading parameter × (1 – Removal parameter)]
where
EPL = existing annual pollutant load for each parameter
Loading = potential pollutant load (kg/year) for each parameter
Removal = removal efficiency for each BMP for each parameter
Subtracting the removed mass pollutant load from the potential pollutant load gives the existing stormwater pollutant loads.
The result of the calculations shows that there are 3,209 kg/year of TN, 536 kg/year of TP, and 84,362 kg/year of TSS discharged into the Banana River from the city’s stormwater system.
TMDL Goals
Based on the EPA’s load allocations, Indian Harbour Beach’s targeted pollutant removals will be 2,022 kg/year for TN and 359 kg/year for TP. It is significant to note that the TMDL reduction goals are citywide, meaning that development with existing treatment facilities are not excluded from load allocations. The burden of pollutant reductions is equally spread across all land uses. If properties with existing stormwater systems are excluded from retrofitting, then the remainder of the properties will shoulder additional burdens to remove even more than 63% TN and 67% TP from their runoff.
Proposed Retrofit Projects
To meet the large TMDL load reductions, it was obvious that significant levels of stormwater retrofitting would be required. SSI evaluated existing groundwater elevations, soils types, and land uses to generate a list of appropriate structural BMPs for Indian Harbor Beach. Four types of BMPs were recommended:
- Dry retention ponds/swales in areas with Type A soils
- Wet detention ponds on deep major outfalls
- Exfiltration trenches on streets with Type A soils, low groundwater, and sidewalks
- Inlet traps on numerous small drainage basins where other BMPs could not be installed
Every outfall and drainage basin in the city was inspected to develop a list of retrofit projects to reduce nutrients. All vacant properties were evaluated for acquisition and potential pond construction. Projects were targeted on all city and county properties. Every road with medians and type A soils was a potential site for retention swales or exfiltration trenches. On streets with no sidewalks and type A soils, curb cuts were proposed with retention swales to be cut behind the curbs. A preliminary list of proposed retrofit projects was generated.
Initial pollutant load modeling showed that the proposed BMPs would not generate targeted pollutant removals. To provide even more pollutant reductions, treatment trains consisting of 15,806 linear feet of exfiltration trenches and six dry retention projects were proposed for the upstream basins of the Gleason Park wet detention pond. Those projects would reduce the volume of water entering the Gleason Park pond, making it more effective in treating the runoff that did make its way to the pond.
There were numerous outfall pipes that had no economically feasible retrofit opportunities. In those basins, inlet traps were specified to at least remove gross solids of sediment and organic debris. A project list of 19 structural BMPs plus 105 inlet traps was proposed.
Proposed Pollutant Loads
The determination of the pollutant loads, assuming implementation of the proposed alternatives, started with the existing pollutant loads. Because the city was fully built out, all land uses in the proposed conditions were considered the same as in the existing conditions. Proposed pollutant loadings after BMP implementation were calculated with the following methodology.
Sub-basin areas contributing to each of the proposed BMPs were determined. It was assumed that the pollutant loads over the entire basin were homogeneous. Therefore, a ratio of sub-basin size to basin size was used to calculate existing loadings for the sub-basins. The formula shown below was used for calculating the sub-basin loadings.
EPL parameter = PB Loading per parameter × SB Area / PB Area
where
EPL = existing sub-basin annual pollutant load for each parameter
PB loading = parent basin pollutant load (kg/acre/year) for each parameter
SB area = sub-basin area (acres)
PB area = parent basin area (acres)
Proposed BMP Effectiveness
Each individual proposed dry retention pond and swale was modeled by calculating the retention volume and by using the appropriate tables from Harper’s study as discussed previously.
The design principle for exfiltration trenches is the same as for dry retention ponds: a designed volume of water is percolated into the ground, treating 100% of all associated pollutants. Removal efficiencies for exfiltration trenches were calculated in the same manner as for dry retention ponds.
The proposed wet ponds would not meet standard design criteria for 1 inch of retention and therefore would not have the pollutant removal effectiveness shown in the FDEP database. Hence, pollutant removals for the proposed projects were calculated by using the methodology from Harper’s study, which indicates that a wet pond’s effectiveness is more accurately related to the permanent pool volume. FDEP recommends using the linear regression equations for calculating wet detention pond removal efficiencies. These methods are now promulgated by several of Florida’s water management districts for many permitting conditions.
Inlet traps remove organic debris and sediment from stormwater runoff, which are different components of pollutants not normally measured in the water column with autosamplers. Research for FDEP by Smith and England (2007) indicated that in the Rockledge, FL, area, gross pollutants accounted for 22.7% of the TN and 16.8% of the TP mass annual loadings (including EMC loads) from a residential drainage basin. In a study by England (2001), a grated inlet trap was shown to remove 79.3% of grass and sediment loads. Pollutant removals for inlet traps were calculated as:
RETN = 0.793 × 22.7% = 18%
RETP = 0.793 × 16.8% = 13.3%
RETSS = 79.3%
where
RE = removal efficiency for each parameter
Using the above methods showed each proposed BMP to have a unique, calculated removal efficiency for each pollutant of concern.
The basic methodology for treatment train calculations was as follows:
- Calculate the existing mass annual pollutant load within a sub-basin for each parameter based on land use, soils characteristics, and existing treatment systems.
- Use the calculated treatment efficiency for each parameter to determine the pollutant load moving to the next downstream sub-basin.
- Calculate the pollutant loading for each parameter for the next sub-basin downstream by adding the pollutant load leaving the upstream basin to the downstream sub-basin load to give the total pollutant load entering the next BMP.
- Using the BMP removal efficiency for the second BMP, multiply the removal by the load calculated in step 3 to give the mass load removed in the second BMP.
- Continue this stepwise process for each sequential BMP until the receiving water is reached.
Acknowledging that the removal efficiency numbers used in this method may be controversial when using vault boxes to treat TSS, it was felt that the method was sound when using the selected BMPs for nutrient and sediment (not TSS) removals. Remember that there are no TSS TMDLs for this region.
The pollutant load removals for each BMP were summed to give a total pollutant load removal for each basin. The total proposed pollutant removals for the entire city are 996 kilograms per year for TN, 248 kilograms per year for TP, and 51,797 kilograms per year for TSS. This equates to a 31% removal for TN, 54% removal for TP, and 59% removal for TSS, which is considerably less than the TMDL allocations.
A cost per kilogram of pollutant removed for each project is important information when prioritizing projects in a cost benefit analysis. A cost per kilogram of pollutant removed should be an important selection criterion used for all BMP selections.
Although implementation of the proposed structural BMPs would provide significant load reductions from the city’s stormwater system, these improvements would still fall short of probable TMDL reduction goals. Further pollutant load credits can be obtained from FDEP by implementing a combination of so-called soft, or nonstructural, programs of ordinance revisions requiring higher levels of BMP use for development activities, street sweeping, public education, and enacting an ordinance to reduce the use of nitrogen bearing fertilizers. The city is already pursuing some of these programs as part of its NPDES Phase II permit.
These types of soft BMPs are effective methods for source control that reduce the pollutant loadings entering the city’s MS4. It is generally more cost-effective to prevent pollutants from entering stormwater than to remove them from stormwater once they are dissolved.
Indian Harbour Beach has a stormwater utility with an equivalent residential unit rate of $4 per month. Annual revenues generated are approximately $200,000. At the current funding levels, it would take about 35 years to implement the proposed projects in this study. To effectively implement TMDL-mandated improvements in a reasonable time frame, the city should investigate additional sources of funding for its stormwater program.
Conclusions
Recognizing impending TMDL mandates, Indian Harbour Beach proactively engaged Allen Engineering and Stormwater Solutions Inc. to develop a water-quality master plan for the city’s stormwater system. Master planning for water quality, as opposed to creating a flood control master plan, allowed Indian Harbour Beach to specifically address TMDL concerns.
Because of the nature of the TMDL being expressed as a percent reduction of TN and TP loadings, SSI was able to develop a spreadsheet model of the city’s stormwater discharges much more economically than using a full-blown hydrologic and hydraulic (H&H) model. Even though this type of model is not calibrated, the selection of proposed projects with relative removals will suffice for TMDL-compliance purposes. H&H models are necessary for establishing TMDLs, but may not be necessary for a city to demonstrate TMDL compliance.
SSI generated an existing conditions model of the city’s stormwater system, selected 20 potential retrofit projects, and modeled the resultant pollutant reductions. A spreadsheet-based model showed that with the proposed BMPs, the city would reduce the TP loads by 31% and TN loads by 54%. Costs for these projects were estimated at $6,803,251. Through the use of spreadsheet data summaries, individual project costs, pollutant removals by parameter, and costs per kilogram of pollutant removed were developed to assist the city in prioritizing project construction sequences.
Because the use of a spreadsheet model is not calibrated to actual runoff concentrations, actual pollutant loadings may be different from model predictions. It is recommended that the city undertake a sampling program to obtain actual runoff concentration data that could be used to calibrate the model. Such a monitoring program for long-term pollutant sampling generally takes several years of storm event (not background) sampling. To meet FDEP’s request for a list of retrofit projects as part of the TMDL process, the spreadsheet model allowed quick development of pollutant loadings and retrofit projects that will take many years to implement. During the ensuing years, collected sampling data can be used to refine the model, but the basic project needs will not change much.
Even though these removals would fall far short of the expected TMDL allocations, further significant reductions through structural BMP implementation would not be fiscally feasible. To achieve additional load reductions, it was also recommended that the city undertake a series of nonstructural BMP initiatives that would result in further reductions of nutrient loads through source control programs.
