Optimization Modeling of Phosphorus Removal in Reservoir and Stormwater Treatment Areas

Daniel Smith and Lewis Hornung

he Lake Okeechobee Watershed A combination of above-ground reser- Project (LOWP) is a major compo- voirs and Stormwater Treatment Areas Daniel P. Smith, Ph.D., P.E., is senior engi- nent of the Comprehensive (STAs) will be used to capture and treat T neer at the engineering and consulting Everglades Restoration Program (CERP). A runoff (Table 2). Configuration options are: firm Berryman & Henigar Inc. in Tampa. primary goal of the LOWP is to reduce phos- S Stand-alone (off-line) reservoirs that with- He is a member of the Water Environment phorus loadings to Lake Okeechobee from draw from a stream or canal and discharge Federation and the American Water Works contributing watershed planning areas north to the same or a different stream or canal. Association. Lewis Hornung is a senior of the lake (Figure 1). S Stand-alone (off-line) STAs that are simi- water resources engineering project man- Phosphorus must be reduced to restore larly configured, and reservoirs that with- ager with HDR Engineering, Inc. in West lake water quality and to meet the Total draw from a stream or canal and discharge Palm Beach. Maximum Daily Load (TMDL) of 105 metric to an STA, which then discharges to a tons per year (MTY). Best Management stream or canal. to simulate system performance: a reservoir Practices that are already planned will reduce A directly coupled reservoir/STA system model (RESOPT 3.0) and an STA simulation phosphorus loading from 433 to 235 MTY. is called a Reservoir Assisted Stormwater model. A stream matrix model was also The LOWP goal is to provide the additional Treatment Area (RASTA). For each of four employed to connect reservoirs and STAs 130 MTY needed to meet the TMDL (Table 1). basins (or Planning Areas), five Planning Area within the stream and canal network. The second LOWP goal is to provide sur- Alternatives (PAAs) were developed, consist- Simulation modeling was applied over a face water storage to help manage ecologically ing of combinations of off-line reservoirs, off- 36-year period of record to assess the effect of desirable lake levels and reduce the need for line STAs, and RASTAs (HDR Engineering, superimposing reservoirs and STAs (i.e, the damaging flood discharges from Lake Inc., 2005). Within each Planning Area, exist- LOWP) onto project Planning Areas. The Okeechobee to estuarine areas on the east and ing land uses and ecological values were used without-project conditions were the project- west coasts. The reservoirs will also provide a to pre-select land areas that could be used for ed future flows and loadings after implemen- continuous supply of water to the constructed construction of reservoirs and STAs. tation of planned BMPs, which were derived wetlands (Stormwater Treatment Areas). The number or reservoir and STA combi- from previous extensive modeling of the nations in a Planning Area was watersheds (Soil and Water Engineering potentially quite large. The Lake Technology, Inc., 2004). Okeechobee Combinatorial RESOPT 3.0 is a water budget and phos- Total P Loading Watershed Analysis Program phorus mass balance model that was devel- (Annual Average Mton/year) Planning Area (LOWCAP) was developed and uti- oped specifically to simulate LOWP reser- Future Target Future lized to evaluate the large number of (with LOWP with voirs. Phosphorus flux components in BMPs) Reduction LOWP possible combinations. LOWCAP RESOPT 3.0 are illustrated in Figure 2, and Lake Istokpoga/ computed storage volumes and water budget components are similar. Salient Indian Prairie 70.1 60.0 10.1 phosphorus load reductions, esti- features of RESOPT 3.0 are: Fisheating Creek 60.5 50.0 10.5 mated costs, and sorted alternatives S Reservoir water column control volume Kissimmee River 55.4 0 55.4 based on average annual values. S Completely mixed water column (0-D) Results were used to iden- Taylor Creek / S One-day timestep 49.0 20.0 29.0 tify a set of 20 cost-effective PAAs Nubbin Slough S Flow balance for reservoir that met the project goals. The next TOTAL 235 130 105 S Total phosphorus mass balance step was to perform more detailed S Total phosphorus (no speciation) evaluations of the PAAs to verify Table 1: Average Annual Phosphorus Loading from S Stream diversion rules storage capacities of reservoirs and Planning Areas to Lake Okeechobee. S Reservoir release rules phosphorus load reduc- For a given reservoir location, the point tion capacities and to of withdrawal from stream to reservoir and Management develop appropriate capac- Goal Approach the point of reservoir discharge were speci- Measure ities for pump stations, fied, as were capacity of withdrawal pumps. structures, and canals. Stream flow and phosphorus concentration Reduce Phosphorus Capture runoff and treat Stormwater Treatment Loading to Lake prior to discharge to the time series at the point of withdrawal were Okeechobee lake Areas Modeling Approach used to calculate the daily volume and phos- Provide improved Modeling tools were phorus mass routed into the reservoir for the management of lake Capture and detain peak Above ground reservoirs levels flows to lake needed to design reservoir specified withdrawal pump capacity. and STA systems that Precipitation time series were assembled Reduce freshwater Capture and detain peak could achieve project release to estuaries flows to lake; improved Above ground reservoirs from monitoring stations within each basin, lake management objectives within available and potential evaporation calculations pro- Table 2: LOWP Project Components. land area constraints. Two duced sinusoidally varying evaporation rates. coupled models were used Reservoir modeling included an over- 68 • JUNE 2005 • FLORIDA WATER RESOURCES JOURNAL Figure 2: Mass balance flux components for Total Phosphorus in RESOPT 3.0.

Figure 1: Lake Okeechobee Watershed Planning Areas. flow feature for high-precipitation events imposed on full or near-full reservoirs; excess water and phosphorus greater than maximum working depth (8.0 ft. for all LOWP reservoirs) were routed to a receiving stream or canal via a crest overflow. Phosphorus mass balance modeling (Figure 2) included sedi- mentation and resuspension. Phosphorus sedimentation was modeled using a constant settling velocity, with the parameterization resulting in a first- order, depth-dependent removal rate constant (Smith, et al, 2004). Phosphorus resuspension was estimated using the Shallow Water Wave Model (U.S. Army Coastal Engineering Laboratory, 2002), a quadratic relationship for bottom shear stress (Lijklema et al., 1994), and resuspension rate as a linear function of bottom shear stress (Sheng and Lick, 1979). Average daily wind velocity was the driving force in the resus- pension calculation. Generally, depths of less than 1.5 feet and wind velocities of greater than 15 miles per hour were needed to discern a noticeable resuspension effect. Daily direct phosphorus deposi- tion onto the reservoir surface was predicted using a rainwater phosphorus concentration that was estimated from the average areal total deposition rates over the Lake Okeechobee watershed. DMSTA, the Dynamic Model for Stormwater Treatment Areas, is a non-steady-state model of Stormwater Treatment Figure 3: Fisheating Creek RASTA: 3240 acre reservoir and 6-cell STA Areas that simulates the hydrologic water balance and phospho- (17,000 acre). rus removal processes in treatment wetlands (Walker and Kadlec, 2004). The basic function of DMSTA is to predict phosphorus removal efficiency of an STA. For a given STA area, STA cell con- Withdrawal Point figuration (number of parallel trains, number of cells in series), Fish Eating Creek LOWP-07 To Lake cell aspect ratios, and given influent flowrate and total phospho- Calcreach 4 Okeechobee l a rus concentration time series, DMSTA predicts the flowrate and w o w l f a r r e d v h t i total phosphorus concentration in STA effluent. O The DMSTA-predicted treatment efficiency of the STA is cal- W culated by the differences between influent and effluent phospho- rus mass on a cumulative basis over a selected averaging period. Reservoir DMSTA was applied to LOWP STAs using the biokinetic param- eters for emergent macrophyte wetlands, which was the STA veg- STA etation assemblage specified by the Project Delivery Team. Stream Calcreach

Node

Operating Rules Controlled Flow To Lake Operating rules were needed to simulate the performance of Okeechobee Overflow reservoirs and STAs. These included specifying rules for the rates of withdrawal from canals to reservoirs, release rates from reser- Figure 4: Fisheating Creek matrix. Continued on page 70

FLORIDA WATER RESOURCES JOURNAL • JUNE 2005 • 69

to maintain a minimum in- specified; STA influent flowrates were main- Annual Average Total P Withdrawal stream flow. tained at the maximum HLR whenever possi- vs. Maximum Withdrawal Flowrate An analysis of streamflow ble. 70 rates was performed to iden- Since the maximum HLR could be met

l Stream Total P Transport a 60 w w tify an efficient inflow pump only a small fraction of the time, the resulting a a r d

h capacity. Typically, this average HLR was much lower, and always less t 50 i W W ) ) capacity was approximately than 5 cm/day. Furthermore, water levels in r P FEC PAA Removal Goal a l 40 e e a t y /

o the 95th percentile stream STAs were never at excessively high levels that n T o t e e 30

g flow. For a stand-alone reser- could damage vegetation for continuous M ( ( a a r e

v voir, release rate was based on periods. 20 A l a a analysis of mass phosphorus Based on this analysis, reservoir release u

n 10 n n

A removal within the reservoir, rates to directly coupled STAs, and maximum

0 and also mass phosphorus pump sizes STAs directly receiving stream or 0 1000 2000 3000 4000 5000 6000 7000 8000 removal in a downstream canal withdrawals, were based on the STA Maximum Withdrawal Flowrate (cfs) STA, if it existed. area and 20 cm/day HLR. For a RASTA, the In order to prevent reser- STA would receive limited inflow if the reser- Figure 5: Phosphorus withdrawal from Fisheating Creek voir dryout, reservoir voir depth was less than or equal to 1.0 foot. versus pump capacity. depth/release rules were for- mulated. Initially, rules were Design Examples developed that decreased Continued from 69 Five alternative configurations of reser- allowable release rates in a stepwise pattern as voirs, STAs, and/or RASTAs were evaluated voirs, minimum reservoir depths, hydraulic reservoir depth decreased. for each of four Planning Areas. A summary loading rates to STAs, and minimum flows to The project delivery team decided to use of one alternative from two Planning Areas is STAs to prevent dryout. a minimum depth of one foot with a simple presented in this article. Central to the complexity of the system switching function. At depths greater than The initial example, the first alternative for were high daily variations in flowrates, along one foot, reservoir release was based on opti- Fisheating Creek (FEC-01), is a relatively with the accompanying variation in phos- mizing performance of the downstream STA. simple configuration. The second design, phorus loading. Also, hydrology in the At depths of one foot or less, reservoir release alternative 2 for the Lake Istokpoga/Indian Okeechobee basins is highly seasonal, and rate was limited to offset ET in the down- Prairie basin (ISTOK-02), is much more dry-season dynamics can be quite different stream STA. complicated. These two PAAs will reduce than wet-season dynamics. The project deliv- For STAs, the areal hydraulic loading phosphorus loading to Lake Okeechobee by ery team ultimately adopted a simplified rate (HLR) is an important design and oper- 110 MTY, which is 85 percent of LOWP tar- series of operating rules commensurate with ating parameter. HLR is the applied flowrate get reduction. the current level of planning effort. per surface area (L3/L2T). Simulation studies Operating rules were applied at the using steady-state flowrates and loadings Fisheating Creek Planning Area beginning of each one-day timestep. The indicated an optimum STA HLR of approxi- Fisheating Creek enters Lake maximum withdrawal rate from stream or mately 6 cm/day (Keller and Knight, (2004) Okeechobee from the northwest and trans- canal to reservoir was established based on to maximize phosphorus mass reduction. ports 60.5 MTY into the lake for the future the cumulative distribution of streamflow at Initially, flowrates to LOWP STAs were without project condition (Table 1). The the point of withdrawal. For any given day, calculated using a 6 cm/day HLR. It soon LOWP target phosphorus reduction for the the maximum withdrawal rate was equal to became apparent that that the steady-state Fisheating Creek Planning Area was 50 MTY. the lesser of two quantities: approximately analysis results had a limited utility for the Prior planning studies had identified a the 95th percentile stream flow (95 percent of LOWP STAs that operated under highly vari- number of land areas in the watershed as flows less than or equal to), or the actual daily able flowrates. potential sites for reservoirs and STAs. Due to streamflow, with the 15th percentile stream- An alternative modeling approach was concerns for the ecological and recreational flow (low streamflow) subtracted from each taken. A maximum HLR of 20 cm/day was

FEC PAA: 6 STA cells Average Total P Mass Transport Total P Removal vs. Withdrawal Pump Capacity Reservoir = 3600 acres STA = 17900 acres 160

60 t 140 r o p

s 120 50 n a l r a T

v 100 ) s ) r s o

40 y a a Reservoir a m d e M 80 / e y g / STA P k R

30 l ( n a

t 60 o P Total t o l T M a 20 ( t e 40 g o a T T r

10 e 20 v A 0 0 Stream Reservoir STA Cell 1 STA Cell 2 STA Cell 3 1500 2000 2500 3000 3500 4000 4500 Withdrawal Discharge Discharge Discharge Discharge

Withdrawal Pump Capacity (cfs) Location

Figure 6: Phosphorus fluxes through RASTA treatment system at Figure 7: Phosphorus fluxes through RASTA treatment system FEC. at FEC.

70 • JUNE 2005 • FLORIDA WATER RESOURCES JOURNAL Annual Average Phosphorus Fluxes (Mton/year)

Future Condition Without Project

Fisheating Creek 57.8

Precipitation

Lake PAA Land Area Okeechobee 20,240 acres 3.7

Total Phosphorus Loading = 57.8 + 3.7 = 61.5 Mton/year

Future Condition With FEC-01 PAA

Fisheating Creek 57.8 3.5

Withdrawal to Reservoir 54.3 Precipitation 0.6 Sedimentation 1.2 Lake Reservoir 53.7 Discharge Okeechobee STA Removal 48.7 Total Phosphorus Loading = 3.5 + 8.1 = 11.6 Mton/year Precipitation STA 3.1 Discharge 8.1

Figure 8: Effect of FEC-01 on phosphorus fluxes. values of Fisheating Creek, it was decided to use sites near the downstream end of the creek. The final available land area was located less than a mile from Lake Okeechobee, near the downstream end of the creek. The FEC- 01 alternative design consisted of a 3,240 acre reservoir directly coupled to a 17,000 acre 6- cell STA (Figure 3). The RASTA served as an end of pipe system for the entire Fisheating Creek watershed. A stream matrix (Figure 4) was devel- oped to represent the Fisheating Creek Planning Area. Water was withdrawn from the creek to the reservoir and STA treatment system. The STA discharges to Lake Okeechobee (or alternatively back to the last section of Fisheating Creek), and reservoir overflow is routed to the creek. The Fisheating Creek flowrate and total phospho- rus concentration time series at the point of diversion was used to generate a plot of phos- phorus withdrawal from the stream based on withdrawal pump capacity (Figure 5). In order to capture the total phosphorus mass required for the Planning Area, it was necessary to adjust the inflow pump capacity to capture more of the peak flood flows. Total phosphorus mass withdrawn from the creek increases rapidly up to 2,000 cfs pump capac- ity and becomes increasingly limited for greater pump capacities. Removal of 50 met- ric tons per year of phosphorus (the FEC removal target) would require a pump size of 2,400 cfs or greater. The phosphorus loading analysis shown in Figure 5 highlighted an important mini- Continued on page 72

FLORIDA WATER RESOURCES JOURNAL • JUNE 2005 • 71 Plan Area (acres) dicted by RESOPT 3.0 was due to a For the future without project condition, Reservoir 3,240 short average water residence time of average total phosphorus loading to Lake 2.4 days. Okeechobee from Fisheating Creek and the STA 17,000 Average daily total phosphorus land area of the FEC-01 PAA is 61.5 metric Total 20,240 mass transport through the FEC-01 tons per year. Withdrawing water from Withdrawal Pump Capacity (cfs) 3,500 treatment system are shown in Figure Fisheating Creek greatly reduces loading Reservoir 7. The average stream withdrawal of from the creek, and the majority of the with- Average Storage (acre-ft.) 1550 149 kg/day (54.4 Mton/year) compris- drawn phosphorus mass is removed in the Average Water Retention Time (days) 2.4 es 94 percent of Fisheating Creek treatment system. The FEC-01 alternative Flow Weighted Phosphorus Concentrations (ug/L) transport. reduces net phosphorus loading to Lake Influent 173.4 Most phosphorus entering the Okeechobee by an average of 49.9 metric tons Discharge 172.6 reservoir is discharged to the STA. per year. Stormwater Treatment Area Phosphorus mass reduction is greatest Hydraulic Loading Rate (cm/day) in the STA Cell 1, followed by smaller Lake Istokpoga/Indian Prairie Planning Area Maximum 20 reduction in STA Cell 2 and still small- The Lake Istokpoga/ Indian Prairie Average 1.24 er reduction in STA Cell 3. The decline Planning Area contributes 70.1 MTY of Total P Loading Rate including atmospheric deposition (kg/ha-yr) 8.26 in removal rates as the concentration phosphorus to the northwest area of Lake Total P Removal Rate (kg/ha-yr) 7.08 successively decreases through treat- Okeechobee (Figure 1) for the future without Average Annual Total Phosphorus Removal ment units in series is a common char- project condition (Table 1). Phosphorus (Mton/year) Reservoir 1.2 acteristic of biochemical treatment enters the lake directly through the C-41 and STA 48.7 processes. The total phosphorus trans- C-40 Canals and through the Kissimmee Total 49.9 port profile through cells in the paral- River through C-41A and other smaller lel treatment trains (Cells 4, 5, and 6) canals (Table 4). The LOWP target phospho- Table 3: FEC-01 Summary. is similar to that in the Cells 1, 2, and rus reduction for Lake Istokpoga/Indian 3. Prairie Planning Area was 60 MTY, or 86 per- A summary of FEC-01 design and cent of the future without project loading. operating characteristics is shown in The configuration of ISTOK-02, shown Continued from 71 Table 3. The annual average storage of in Figure 9, utilizes the land area that was mum size constraint to the withdrawal pump 1,550 acre-feet is 6 percent of the maximum determined to be available for siting reser- capacity for FEC-01. To meet the FEC phos- storage of 25,900 acre-feet at reservoir full voirs and STAs. ISTOK-02 consists of one phorus removal requirements, a large pump conditions. The relatively short reservoir reservoir and three STAs with sizes indicated size is needed to capture large loadings that water residence time of 2.4 days resulted in in Table 5. ISTOK-02 consists of a 6,480-acre follow high rainfall events. little difference in flow weighted total phos- offline reservoir on C-41A Canal with return RESOPT 3.0/STA simulations were con- phorus concentrations in reservoir influent of discharge back to the C-41A Canal. MM2 ducted to optimize the PAA system. FEC-01 and discharge. The average HLR to the STA North STA is also located on the C-41A simulations consisted of selection of the was 1.24 cm/day, only a small fraction of the Canal, while MM2 South and MM3 STAs capacity of the pump for withdrawal from maximum HLR. withdraw from the C-40 Canal and the C-41 Fisheating Creek to the reservoir, specifica- The system design allows the treatment Canal, respectively (Figure 9). tion of minimum stream flow in the creek system to capture large flow and loading The task of optimizing the ISTOK 02 below which no withdrawal was allowed, events, while average loadings are limited. design and achieving the target phosphorus selecting a maximum STA hydraulic loading The average areal phosphorus loading to the removal goal of 60 metric ton per year was rate (HLR), and selecting a minimum reser- STA was 8.26 kg/ha-year, and the STA was formidable for two reasons: voir depth below which no reservoir release capable of removing 86 percent of this load. First, there was a discussed discrepancy would occur. The minimum stream flow was The average phosphorus concentration in the between the location of phosphorus loading the 15th percentile low flow (9.7 cfs), maxi- effluent of the FEC 01 STA was 23.4 ug/L. within the Istokpoga Planning Area and the mum STA HLR was 20 cm/day, and mini- A cautionary note is that consistently size of available land areas. Analysis of flows mum reservoir depth was 1.0 foot. achieving effluent phosphorus concentra- and loadings from the major canals indicated The main degree of freedom in FEC-01 tions of less than 25 ug/L may be a challenge that Harney Pond (C-41) Canal accounted modeling was selection of the maximum for the macrophyte-dominated STAs that Continued on page 74 withdrawal pump capacity from Fisheating were specified in LOWP Average Total Creek. RESOPT 3.0 model results generated a design. In subsequent phases, Average Discharge Receiving Phosphorus Transport the planning process must Canal time series of reservoir discharge and total Water 1000 acre- % Mton / year % phosphorus concentration that were used consider alternative structur- ft./year Harney Pond Lake 203.2 45.2 36.3 51.8 directly as input to the DMSTA model for the al or operational adjust- (C-41) Okeechobee

ments to the alternatives to Indian Prairie Lake STA. 92.4 20.5 15.4 22.0 Simulations were conducted across a minimize, or eliminate the (C-40) Okeechobee Kissimmee C-41A 138.2 30.7 15.2 21.7 range of withdrawal pump capacities (Figure need to reach such low efflu- River Canals on West Side of Kissimmee 6). A withdrawal pump capacity of 3,500 cfs ent concentrations in order Kissimmee River South of 16.1 3.6 3.2 4.6 C-41A Canal River was required to achieve the target annual to meet the overall target for TOTAL 449.9 100.0 70.1 100.0 average total phosphorus removal of 50 met- load reduction. ric tons per year. Total phosphorus removals The effect of FEC-01 on Table 4: Annual Average Volume and Phosphorus increased as maximum pump capacity phosphorus fluxes in the Loading to Lake Okeechobee (Future without project). increased and were dominated by the STA. Fisheating Creek Planning The relatively limited reservoir removal pre- Area is shown in Figure 8.

72 • JUNE 2005 • FLORIDA WATER RESOURCES JOURNAL

Average Annual Total Mton/year Phosphorus Transport

C-41A Canal 0.4

Canals South of C-41A Canal 1.5

Indian Prairie Canal (C-40) 0.7

Harney Pond Canal (C-41) 0.0

MM2 STA North 0.0

MM2 STA South 7.9

MM3 STA 0.0

Table 7: Phosphorus loading components for ISTOK-02.

planning process must consider structural or operational adjustments to this plan to mini- mize or avoid the need to achieve these low Figure 9: Configuration of ISTOK-02 Planning Area Alternative. effluent concentrations. Alternative designs were conceptualized, developed, and simulated, resulting finally in the ISTOK-02 configuration shown in Figure Continued from 72 Canal was only 7,509 acres. 9. The basis for development and analysis of for 51.8 percent of phosphorus loading, while The larger available land areas were alternative designs was to evaluate individual Indian Prairie and C-41A Canals accounted located to the north and west. From an earli- system components using an approach that for about 22 percent of phosphorus loading er analysis, it was determined that the phos- consisted of apportioning the 70.1 minus 60, each. While phosphorus loading was highest phorus reduction ability of the 7,509-acre or 10.1 metric tons per year “allowable load- to the south and west, the available land area available land area adjacent to C-41 Canal ing” among the individual contributing com- directly adjacent to Harney Pond (C-41) was insufficient to enable ISTOK-02 to ponents. achieve the overall The contributors to the “allowable load- load reduction objec- ing” were discredited into canals and man- tive of 60.0 MTY. agement measure discharges. Canal loadings Modeled Area (acre) The second chal- consisted of any phosphorus that was not Management lenge to achieving 60.0 Measure withdrawn to a management measure or that Reservoir STA Total Total MM MTY removal was was returned from a reservoir. Management phosphorus concentra- measure contributions to the “allowable 1 6,480 0 6,480 6,480 tion. The flow weighted loading” consisted of management measure phosphorus concentra- discharges. 2 North 0 7,500 7,500 tion for the loadings Several design features were specifically 23,500 from the Istokpoga/ 2 South 0 16,000 16,000 developed to achieve the phosphorus Indian Prairie Planning removal target and are incorporated in 3 0 7,509 7,509 7,509 Area was 126 ug/L. ISTOK-02: Achieving 86-percent S MM3 was operated as a “high rate” STA to Table 5: ISTOK-02 Configuration. reduction resulted in achieve high areal removal rates. an estimated flow S MM3 effluent was routed to MM3 South weighted “effluent” STA for further treatment of the high efflu- MM2 North MM2 South concentration of 18 MM3 STA ent concentrations from MM3 STA. STA STA ug/L. S Water from Harney Pond (C-41) Canal Maximum Hydraulic Loading Rate Achieving these 12.1 15.1 16.1 Water to that was not withdrawn to MM3 (cm/day) low average phospho- STA was also routed to MM2 STA South.

Average Hydraulic Loading Rate rus levels is a consid- 1.2 1.9 2.2 S MM2 North STA discharge was routed to (cm/day) erable challenge for C-40 Canal, where it could be further treat- reservoir/STA systems Average Phosphorus Loading Rate ed in MM2 STA South. 3.6 4.2 11.5 (kg/ha-year) using macrophyte- Numerous iterative simulations were dominated STAs. As Average Phosphorus Loading Rate used to balance the interacting design deci- 2.9 2.9 8.6 (kg/ha-year) with the analysis for sions of reservoir pump size and discharge Fisheating Creek Average Effluent Total Phosphorus rate, size of pumps to the three STAs, and STA 14.1 15.2 35.6 Concentration (ug/L) alternatives, subse- size. Phosphorus removal from MM1 reser- quent phases of the voir as a function of withdrawal pump capac- Table 6: ISTOK-02 STA Operating Characteristics.

74 • JUNE 2005 • FLORIDA WATER RESOURCES JOURNAL

to close to the 8-foot maxi- phorus than MM2 South due to the higher mum working depth. areal removal rates that are predicted to occur 6.0 The reservoir dried out in with higher influent concentrations. s u r

o only one year, while maxi- Continued on page 76 h ) p p r

s mum capacity was reached at

a 5.5 o e h y

/ least once in 70 percent of P n l o

a Plan Area (acres) t t the years. Average annual o M 5.0 ( T Reservoir 6,480 l l storage ranged from 7,000 to a a

v 650 cfs Withdrawal Pump

u STAs 31,009 o n n over 50,000 acre-feet and n m 750 cfs Withdrawal Pump Total 37,409

e 4.5 A averaged 31,500 acre-feet. R e 850 cfs Withdrawal Pump

g Reservoir a

r Average phosphorus removal

e Average Storage (acre-ft.) 31500 v 4.0 ranged from less than one A 0 25 50 75 100 125 150 175 200 Average Water Retention Time (days) 112 half to 12.3 MTY,and reflects Flow Weighted Phosphorus the difference in water-col- Concentrations (ug/L) Maximum Discharge Rate (cfs) Influent 84.7 umn phosphorus mass from Discharge 47.9 beginning of year to end. Figure 10: Effect of discharge rate on phosphorus removal Stormwater Treatment Areas in MM1 reservoir. Operating characteristics Hydraulic Loading Rates (cm/day) for ISTOK-02 STAs are sum- Maximum 12.2 – 16.1 marized in Table 6. A total of Average 1.2 – 2.2 ity and discharge rate is shown in Figure 10. 37,409 acres are dedicated to reservoirs and Total P Loading Rates including 2.9 – 8.6 The figure indicates limited increases in atmospheric deposition (kg/ha-yr) STAs. The reservoir provides an average stor- Total P Removal Rates (kg/ha-yr) 14.1 – 35.6 annual phosphorus removal as withdrawal age of 31,500 acre-feet, and the 112-day aver- Average Annual Total Phosphorus Removal pump size increases from 650 to 850 cfs, but (Mton/year) age residence time provided 5.6 MTY phos- Reservoir 5.6 highest removal for a 100 cfs discharge rate phorus removal. STA MM2 North 8.7 for all withdrawal pump capacities. The average areal phosphorus removal STA MM2 South 19.2 The final ISTOK-02 design included a rate of MM1 reservoir is 2.1 kg/ha-year, or STA MM3 26.1 750 cfs withdrawal pump capacity and 100 well below the more efficient STAs. The three Total 59.6 cfs discharge rate. MM2 reservoir achieved an Average Annual Average Total Phosphorus STAs account for over 90 percent of the phos- Loading (Mton/year) annual average phosphorus removal of 5.6 phorus removal, but there are substantial dif- Without PAA 70.1 MTY. Performance features of MM1 reser- ferences in STA areal removal rates. Although With ISTOK-02 PAA 10.5 voir are shown in Figures 11 through 13. less than half as large as MM2 STA South, Reduction 59.6 Average depths ranged from just over 1 foot MM3 STA removes substantially more phos- Table 8: ISTOK-02 Summary.

FLORIDA WATER RESOURCES JOURNAL • JUNE 2005 • 75 Continued from 75 Annual Depth Maximum, Average, Minimum The average phosphorus concentrations in the effluent of the 9 ISTOK-02 STAs were 14.2, 15.2, and 35.6 to ug/L, respectively, in MM2 STA North, MM2 STA South, and MM3 STA. MM2 STA North and 8 MM2 STA South were in a “low loading” range, while the higher load- 7 ing to MM3 STA resulted in higher effluent phosphorus concentra- 6 tions. ) ) . t t A cautionary note is that consistently achieving effluent phos-

f 5 ( h t t phorus concentrations of less than 25 ug/L may be a challenge for the p e e 4

D macrophyte-dominated STAs that were specified in LOW design. 3 Alternatives will have to be considered to address this issue in subse-

2 quent phases of the planning process.

1 Conclusions & Recommendations 0 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 1 2 3 4 0 7 9 0 5 6 4 8 9 3 4 8 7 3 6 7 0 5 6 9 0 5 8 9 8 6 0 9 9 9 9 9 9 7 7 7 8 8 8 8 8 8 8 8 8 8 9 9 9 7 7 7 7 7 7 6 6 6 6 9 7 Model simulations were extremely valuable to sizing reservoir 9 9 9 9 0 9 9 9 9 9 0 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 and STA systems in the Lake Okeechobee Watershed. Simulations Year provided a rational basis for the locations and sizing of reservoirs, Figure 11: MM1 reservoir average depth. STAs, pump stations, and preliminary operations. Model simulations predicted that targeted reductions of total phosphorus into Lake Okeechobee could be met with reservoir and STA systems in both Fisheating Creek and Lake Istokpoga/Indian Prairie Planning Areas. Several key recommendations emerged from the system modeling. Average Annual Storage (acre-ft.) S Seasonal effects to reservoir and STA will likely be quite significant, 60000 and seasonally based operating rules should be developed. S Areal removal rates of macrophyte dominated STAs decrease signif- 50000 icantly as influent phosphorus concentrations decline below 50 ) ) . . ug/L. The result is large increases in required STA area for relative- t f - e e

r 40000 ly small increases in annual mass removal. c a a ( Alternative STA assemblages should be explored for the Lake

e S g

a 30000 r r Istokpoga/Indian Prairie Canal Planning Area, including sub- o t

S mersed aquatic vegetation (SAV) systems, periphyton assisted STAs, e

g 20000 a and STA systems in series. r e v

A S Reservoir/STA interactions should be systematically explored 10000 through a program of modeling and field data collection. S A comprehensive modeling framework is needed to integrate mul- 0 7 8 9 5 6 7 8 9 0 1 2 3 4 5 6 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 tiple reservoir/STA systems to predict and optimize overall phos- e 6 7 7 7 7 7 7 7 7 7 7 8 8 8 8 8 8 8 8 8 9 9 9 9 9 9 9 9 9 9 0 6 6 6 6 8 g 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 0 a 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 r phorus removal and water storage in the Lake Okeechobee e v A Year Watersheds.

Figure 12: MM1 average reservoir storage. References • HDR Engineering, Inc. (2005). Development of Alternative Plans, Part 1 – Storage and Water Quality, Section 3 Optimization of Planning Area Alternatives, Lake Okeechobee Watershed Project, May 2005. • Keller, C. and R. Knight (2004). Constraints on Hydraulic Loading Rates Annual TP Reduction (Mton/year) = to Stormwater Treatment Areas. Memorandum to HDR from Wetlands TP In Stream + TP In Precipitation - TP Out Release - TP Out Overflow Negative values reflect net added storage in control volume Solutions, Inc. Gainesville, FL, April 14, 2004. Control volume = reservoir water column excluding sediments • Lijklema, L., R. Aaderlink, G. Bloom, and E. van Duin (1994). Sediment 13 transport in shallow lakes: Two case studies related to eutrophication. In 12 Transport and Transformation of Contaminants Near the Sediment-Water

) 11 r

a Interface, J. DePinto, W. Lick, and J. Paul, eds. CRC Press, Inc. Boca

e 10 y / Raton, Fla. n 9 o t • Sheng, P. and W. Lick (1979). The transport and resuspension of sedi-

M 8 (

n 7 ments in a shallow lake. J. of Geophys. Res., 84: 1809-1825. o i t

c 6

u • Smith, D., H. Zarbock, and L. Hornung (2004) Water and Phosphorus

d 5 e Budget Modeling of Large-Scale Water Management Facilities in the R 4 P

T 3 Lake Okeechobee Watershed, Presented at Florida Stormwater l a

u 2 Association Meeting, Orlando, FL, December 2004. n n 1 A • Soil and Water Engineering Technology, Inc. (2004). Lake Okeechobee 0

r Watershed WAM Modeling Project. http://www.swet.com/ 1 2 3 5 6 7 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 7 8 9 0 4 8 9 5 6 e 7 7 7 7 9 6 7 7 7 7 7 7 8 8 8 8 8 8 8 8 8 8 9 9 9 9 9 9 9 9 9 0 6 6 6 6 v 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 0 9 9 A 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 • US Army Coastal Engineering Research Center (2002). Shore Protection Year Manual Volume 1, Fort Belvoir, VA. • Walker, W. and R. Kadlec (2004). Dynamic Model for Stormwater Figure 13: MM1 reservoir average phosphorus reduction. Treatment Areas. http://wwwalker.net/dmsta

76 • JUNE 2005 • FLORIDA WATER RESOURCES JOURNAL