Dynamic Model
for
|
 |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Background Model Features Theory Interface Calibration Testing Discussion Details DocumentsSTA Calibrations Reservoir Calibrations DMSTA2 Enhancements BackgroundFour Stormwater Treatment Areas (STA's, schematics) are being operated to reduce phosphorus loads in agricultural runoff & releases from Lake Okeechobee prior to discharge into the Everglades Water Conservation Areas. Three additional STA's are being designed under the Everglades Construction Project. The treatment objective of these systems is to achieve a long-term, flow-weighted-mean outflow concentration of 50 ppb or less While the treatment concept does not include management to encourage specific vegetation types, STA's are likely to be dominated by cattails & other rooted macrophytes. The ultimate goal of the program is to achieve water quality standards by 2007. Discharge concentrations well below 50 ppb will be necessary to achieve this goal. Ultimate treatment requirements will reflect numeric phosphorus criteria currently being developed by the Florida Department of Environmental Protection. Several technologies are being investigated to accomplish further reductions in phosphorus. A target outflow concentration of 10 ppb is assumed in evaluating these technologies. Ultimately, treatment areas will be integrated into other hydrologic & water-quality components of the long-term restoration plan for the Everglades (CERP). Existing data indicate that it will be difficult to achieve outflow concentrations less than 20-30 ppb with macrophyte-based STA's. The advanced treatment technologies being investigated include various combinations of physical, chemical, & biological processes. Among these is the concept of designing & managing the treatment areas to promote growth of periphyton and/or submersed aquatic vegetation (SAV) relative to macrophytes. The acronym "PSTA" has been coined to represent a treatment area dominated by periphyton. The STA's were sized using a steady-state model calibrated to soil & water-column phosphorus data from Water Conservation Area 2A. This site describes the development of an enhanced model (DMSTA) that provides a framework for integrating experimental & field-scale monitoring data and for designing the next generation of treatment areas. SFWMD describes ongoing research projects providing data for model calibration. Most of the research is being conducted in mesocosms, test cells, & field-scale cells located in the Everglades Nutrient Removal Project, a 3700-acre treatment area that has been operating since August 1994. Model FeaturesDMSTA simulates daily water & mass balances in a user-defined series of wetland treatment cells, each with specified morphometry, hydraulics, and phosphorus cycling parameters. Up to six treatment cells can be linked in series and/or parallel to reflect compartmentalization & management to promote specific vegetation types. Each cell is further divided into a series of continuous stirred tank reactors (CSTR's) to reflect residence time distribution. Water-balance terms for each cell include inflow, bypass, rainfall, evapo-transportation, outflow, seepage in, and seepage out. Parameter estimates for the phosphorus cycling model have been developed for various vegetation types. The model is coded in Visual Basic for Applications; the user interface is an Excel workbook. Factors Considered by DMSTA, but not by the Steady-State STA Design Model
Input Data Requirements
Model CalibrationDMSTA's phosphorus cycling model contains three primary parameters that require calibration to each vegetation type (C1, C0, K). Two parameters (C1 , C0 ) are calibrated to paired biomass P & water column P data from several systems. The third parameter (K) is calibrated to outflow concentration time series from prototypes that are most representative of each vegetation community (emergent macrophytes, submergent macrophytes (SAV), and periphyton). A fourth, secondary parameter (ZX) reflects the apparent dependence of phosphorus uptake on water depths. Based upon comparisons of observed and predicted concentrations over a range of depths and datasets, this factor has been found to be appropriate for emergent and SAV communities, but not for periphyton. Model calibrations for each vegetation type have been subsequently tested against data from other experimental & field-scale platforms.
The steady-state solution of the model predicts that biomass P storage is linearly related to concentration. The parameters C1 and C0 are related to the slope and intercept of that relationship. The parameter C1 (~22 ppb for each vegetation type) essentially determines the scale of the biomass P storage term. Predicted outflow concentrations from treatment cells are insensitive to this parameter. The parameter C0 (4 - 12 ppb) determines the lower concentration limit in a steady-state system. It also largely controls the turnover rate of the phosphorus storage term. As C0 decreases, turnover rates increase and uptake rates respond more rapidly to variations in inflow loads. C0 estimates are sensitive to storage measurements in the low concentration range. Such data are relatively scarce and uncertain because of low storage turnover rates (requiring long experiments to achieve steady-state biomass levels), heterogeneity of plant communities, and biomass P sampling difficulties. For these reasons, C0 estimates derived from storage/concentration data have greater uncertainty, as compared C1 estimates. Based upon model testing results, the initial C0 estimate derived from storage/concentration data (4 ppb) has been increased to 12 ppb in the case of SAV systems. This was necessary in order to provide a satisfactory fit of outflow concentration data from experimental platforms over ranges of concentration and loadings expected to be encountered in full-scale treatment systems, as described below. This adjustment was not necessary in the case of emergent and periphyton datasets.
The third parameter of the P-cycling model (K) reflects the net first-order removal rate (settling velocity) at steady state. Together, the parameters K and C0 determine the outflow concentration in a steady-state system and correspond to parameters of the first-order K/C* model frequently used in designing treatment wetlands for phosphorus and other water quality components (Kadlec & Knight, 1996). Given estimates of C1 and C0, K values have been calibrated to outflow concentration time series data from platforms representative of each vegetation type. Least-squares K estimates have been derived for each platform using the residual sum of squares of log-transformed outflow concentration data as an objective function and the Marquardt non-linear parameter estimation algorithm (Bard, 1974). To provide a period for biomass grow-in and to decrease sensitivity to assumed initial conditions, the model has been run for periods ranging from 0.5 to 8 years before the data are used for calibration.
A
single prototype dataset has been been selected for each vegetation
type. Selection criteria include characteristics of
vegetation, length & quality of dataset, and spatial scale (larger
preferred). The prototypes for SAV and PSTA correspond to those
selected by the respective
research teams for their independent modeling
efforts. The calibration procedure has yielded 5 parameter
sets for potential use in designing treatment areas:
Steady-state solutions predict equilibrium P storage, turnover rate, and phosphorus uptake as a function of water-column concentration for systems with constant flow, loading, & depth for each calibration set. Two parameter sets have been calibrated to submergent vegetation communities (SAV & SAV_C4 ) using data from Cell-4 of the Everglades Nutrient Removal Project. The Cell-4 period of record extends from February 1995 through September 2001. Data prior to June 1996 are thought to reflect startup conditions for the vegetation community and have been used for model initialization. Data after October 2000 are considered less reliable because of extreme drought and missing concentration data from a new outflow structure (G309) associated with STA-1W operation. The SAV calibration (K = 129 m/yr, C0 = 12 ppb) is based upon the June 1996 - October 2000 period for Cell-4, as well as data from other SAV platforms. The need to consider other platforms is based upon the fact that calibrations appear to be dependent phosphorus and/or hydraulic loads which, in the case of Cell-4, generally exceed ranges expected to occur in the polishing cells of optimized STA's. The SAV_C4 calibration (K = 90 m/yr, C0 = 4 ppb) is based upon data from the "optimal performance period" of Cell-4 (January 1998 - December 1999) and ignores data from other SAV platforms. Inflow volumes and loads were relatively stable in 1998-1999 and outflow concentrations were relatively low (flow-weighted mean = 14 ppb vs. 22 pbb for the entire record). This calibration has been provided at the request of SFWMD for use in basin feasibility studies. The specification of three distinct vegetation types is a simplification because all of the calibration platforms, like full-scale STA's, contain mixed communities. Most of the SAV and PSTA platforms were managed to exclude or minimize cattails and contained mixtures of submergent plants and periphyton. The NEWS (Nonemergent Wetland System) calibration is a combination of the SAV and PSTA calibrations that attempts to mimic the transition from submergent vegetation dominance to periphyton dominance as average concentrations and biomass P storage decrease. The transition is assumed to occur over a concentration range of 20 to 10 ppb, similar to that observed along the phosphorus and vegetation gradients in Water Conservation Area-2A. The SAV calibration alone predicts that uptake and storage of P would approach zero as water column concentrations approach 12 ppb. The NEWS calibration prevents this "dead end" by allowing for a conversion to periphyton communities. From a practical perspective, NEWS is intended to mimic a system that is managed to exclude emergent macrophytes (by control of water levels or herbicides) and is not specific with respect to the actual community composition. The extent to which the actual SAV or periphyton species composition impact phosphorus removal would be reflected in the model error distributions. One important limitation, however, is that certain SAV types that may have relatively low uptake rates (such as hydrilla) are not represented in the datasets. Data from approximately 90 platforms have been compiled for use in testing the calibrations to prototype datasets. The platforms include mesocosms, field-scale experiments, full-scale treatment areas, and natural wetlands. After eliminating datasets that are not statistically independent (e.g., 2-phase experiments, cells in series), have mixed vegetation communities, or are derived from experiments with known artifacts, data from 68 platforms are available for model testing. Potential artifacts are associated with the short duration of many of these experiments. These artifacts include insufficient time for full development of the vegetation, releases from initial substrates (typically peat or shellrock), and seasonal factors. To investigate effects of duration and eliminate possible seasonal effects, a subset of 31 platforms with at least 1 year of data (after a startup period of at least 3 months, typically > 6 months) has also been identified. Testing results indicate, however, that model performance is similar for the short and long duration experiments. The following approaches have been taken in testing the model calibrations using the independent datasets:
Detailed results can be accessed through links in the table below. Using the recommended parameter sets for each vegetation category (EMERG for emergents, NEWS for periphyton & submergents), the model explains 91% of the variance in the observed flow-weighted mean outflow concentrations from all 68 platforms with a residual standard error of 15%. Corresponding statistics for geometric mean outflow concentrations are 89% and 17%, respectively. Observed and predicted concentrations for all vegetation types combined are plotted in the attached figure. Results are similar for platforms with at least 1 year duration. With some exceptions (see below), residuals are reasonably independent of platform characteristics (e.g., area, depth, velocity, inflow concentration, experiment duration) and season, both on weekly and average time scales. With further screening of apparent outliers (ENRP Cell-1 & macrophyte dryout experiments), the explained variance increases to 93% and residual standard errors decrease to 13%. These datasets can also be discounted based upon independent criteria. Cell-1 is influenced by seepage inflows from an adjacent wetland (Loxahatchee National Wildlife Refuge), has relatively high topographic relief (contrary to the flat-bottom assumption built into the model), was shut down for a period of ~1 year during the calibration period to allow construction of test cells, and has relatively uncertain inflow loading estimates because of variations in flow and concentration across the 10 inflow culverts (G252A-J), two of which were sampled. The macrophyte dryout experiments (4 emergent and 4 SAV datasets) were designed to test the effects of dryout on phosphorus releases from peat in mesososms with and without emergent vegetation. While the unplanted tanks evolved into SAV communities (Chara), the tanks were not seeded with or otherwise managed to promote target SAV species (Ceratophyllum, Najas), unlike the other SAV mesocosms. These experiments were run at low hydraulic and phosphorus loads which would require relatively long startup periods to allow for vegetation establishment and reduce the effects of phosphorus release from the initial peat substrate. Releases from the initial substrate may account for the fact that the phosphorus stored in the vegetation at the end of the experiments was much higher than predicted by the model both in the planted and unplanted tanks (see datasets labelled MD_EMERG.. and MD_SAV..). In addition, the measured biomass P exceeded the cumulative phosphorus loads to the mesocosms over the entire experimental period. Testing results are discussed below for each parameter set. Observed outflow concentrations for SAV platforms are compared with model predictions using the SAV_C4, SAV, & NEWS calibrations in the attached figures (time series, platform means). Model testing results are generally less satisfactory for the SAV platforms, as compared with the emergent and periphyton platforms. This is partially a consequence of the fact that, as compared with the others, the SAV platforms include much higher ranges of phosphorus and hydraulic loads which in may cases exceed those expected in full-scale treatment areas. In addition, SAV performance may depend upon the species of vegetation and/or the species of phosphorus (SRP vs. organic P), as discussed below. The most important pattern identified in testing the model against SAV datasets is a negative correlation between residuals with three inter-related variables (phosphorus load, hydraulic load, and predicted outflow concentration) in the SAV datasets. This pattern is particularly strong in the case of the SAV_C4 calibration. Residuals are negatively correlated with predicted concentration (R2 = 0.87 & 0.88 for flow-weighted & geometric means), and the average residual in the low concentration range is ~0.4 (40% under-prediction). By increasing C0 from 4 to 12 ppb and using the entire Cell-4 period of record, the correlations are eliminated with the SAV calibration, but the scatter is wide in the low concentration range. The correlations are also present in the NEWS calibration, but are less strong (R2 = 0.22-0.37) and the average bias in the low concentration range is < 10% (vs. ~40% for the SAV_C4 calibration). The same pattern is manifested in positive correlations between K values calibrated separately to each SAV dataset and phosphorus or hydraulic loads. Design phosphorus loads to existing STA's (1W, 2, 34, & 5) range from 1.0 to 1.2 g/m2-yr and hydraulic loads, from 1.7-3.3 cm/d. For a typical STA (1.1 g/m2-yr & 2.5 cm/d) and "optimized" design involving 3 hydraulic compartments (cells) in series, the hydraulic load to each cell would be 7.5 cm/day and the P load to the inflow cell would be 3.2 g/m2-yr. Assuming an inflow phosphorus concentration of 25 ppb, the P load to the last (polishing) cell would be 0.7 g/m2-yr. Many of the SAV datasets (including the Cell-4 prototype) exceed the range of P and hydraulic load expected for a typical optimized STA. Extrapolation of performance (K values) from high to low loading ranges typical of polishing cells would be dangerous, given the apparent positive correlations between K & load and between model residuals & load or predicted concentration. Extrapolating over a range of other system parameters (e.g., area, depth, velocity, etc.) is of less concern because model residuals appear to be generally uncorrelated with these factors. Calibrated K values in nine SAV platforms with phosphorus and hydraulic loads that are typical of STA polishing cells range from 30 to 60 m/yr, as compared with 90 m/yr for SAV_C4. Because concentration, hydraulic load, and P loads are strongly inter-correlated, it is not clear which of these factor(s) is causally linked to K values. The design consequences of extrapolating the calibrations would be less severe if the causal factor is concentration, as opposed to hydraulic or P load, because concentrations in Cell-4 are within the design range, at least down to a concentration of 14 ppb. The correlation between K and concentration or load is eliminated in the SAV calibration (C0 = 12 ppb), but again the scatter is wide in the low concentration range The K vs. load correlation for SAV may reflect factors that are not considered directly in the model but are correlated with total P load or concentration and influence P uptake in SAV communities, such as calcium or SRP levels. As shown in the attached figure, both SRP concentrations and the ratio of SRP/Total P are strongly correlated with Total P concentrations at long-term monitoring stations and research sites in the ENRP. Within Cell-4, Total P levels decrease from 55 ppb to 22 ppb, while SRP levels decrease from 18 ppb to 6 ppb and the ratio of SRP/Total P decreases from 0.35 to 0.25. As the SRP/Total P ratio decreases, the K value for Total P would be expected to decrease because SRP has been shown to be removed more rapidly than non-SRP (~Organic P) in these systems (DBEL Draft Final Report, March 2002). Lower K values measured in SAV test cells and mesocosms at the south end of the ENRP may reflect low inflow SRP/Total P ratios at that locations (0.19 - 0.21), as compared with Cell-4 inflows to (0.35). Revising the model structure to include two phosphorus compartments or a 2nd order phosphorus uptake relationship (as observed for reservoirs & detention ponds) would tend to reduce the K vs. load/concentration dependence. Alternatively, the correlations might reflect improved flow distribution (less short circuiting) at higher hydraulic loads (i.e. Kadlec's parallel path analysis). Additional research and analysis may support enhancement of the model to account for these factors. Despite their small spatial scale, the substrate experiments conducted by DBEL at the ENRP South Test Site provide the longest datasets for testing DMSTA calibrations to SAV communities in the low concentration and loading range. Three substrates (peat, sand, and limerock) were compared in mesocosms operated at an average hydraulic load of ~7 cm/day, inflow concentration of ~25 ppb, and phosphorus load of ~0.7 g/m2-yr. These parameters are reasonably representative of expected designs for STA polishing cells. Calibrated K values (50, 38, and 40 m/yr, respectively) are considerably below the SAV_C4 value (K = 90 m/yr). Attached figures compare observed and predicted time series for each substrate and each of three parameter sets (SAV_C4, SAV, and NEWS). The NEWS calibration seems to work best for each platform. The experiments were not replicated, so there is no statistical basis for concluding that K values depend on substrate.
The EMERG & NEWS calibrations provide the best overall fit of data from all platforms, as well as platforms with at least one year of data. The NEWS calibration is recommended for designing treatment cells with submergent communities. While the SAV calibration may also be used, NEWS has the advantages that it does not reach a "dead end" at low P levels and provides a better fit of data from the SAV substrate experiments, the longest operating SAV platforms in the low concentration range.
The SAV_C4 calibration is not recommended for designing treatment cells with outflow concentrations below 25 ppb, based upon the following observations:
The above recommendations hinge upon potentially important limitations of available datasets (e.g., small spatial scale, short duration, and/or relatively steady inflows) and limitations of the model, as reflected by apparent correlations between K and loads/concentrations for SAV communities. Despite limitations in the datasets, there is no basis for rejecting them outright, especially considering the fact that the apparent differences between Cell-4 and the 11 other SAV platforms in the low concentration and loading ranges can be at least partially explained by diifferences in inflow SRP/Total P ratio, a factor considered important by SAV researchers (DBEL Draft Final Report, March 2002).
Discussion - Periphyton Calibrations
Model testing results are generally satisfactory for periphyton platforms using the PSTA or NEWS calibrations. Observed & predicted outflow concentrations are compared in the attached figures (NEWS time series & platform means; PSTA time series & platform means). Both calibrations are consistent with data from the periphyton/sawgrass marsh south of the C111 canal, the largest periphyton platform. Consistent biases in model residuals are not apparent over a wide range of conditions (weekly residuals, platform means) when the calibrations are applied to independent datasets. The performance of PSTA test cells subsequent to the period used for model calibration/testing has been consistent with or better than that predicted by the PSTA or NEWS calibrations.
Periphyton cells are most likely to be placed near the outflow treatment areas (polishing cells). Flow-weighted-mean concentrations for periphyton datasets ranged from 7 to 25 ppb (inflow) and 6 to 20 ppb (outflow). Most of the datasets are in phosphorus loading ranges typical of polishing cells. There is some indication of a positive correlation between K and load/concentration, although data are limited in the high loading range. If such a correlation exists, extrapolating the calibrations to higher load/concentration ranges may produce conservative results (i.e. over-predicted outflow concentrations).
The performance of SAV & PSTA platforms converge at low concentrations and loads, particularly in shallow systems. Calibrated K values are 16, 24, and 90 m/yr for the EMERG, PSTA, and SAV_C4 calibrations (C0 = 4 ppb). The average K for SAV datasets in the lower loading range typical of STA polishing cells and most of the PSTA datasets is ~40 m/yr. These comparisons reflect relative performance at water column depths at or above 60 cm. At shallower depths, the predicted performance of PSTA improves, relative to the other communities because uptake rates are proportional to depth in the emergent and SAV calibrations, but not in the PSTA calibration. This feature, inferred by residuals analysis of the datasets, may reflect areal vs. volumetric uptake mechanisms (planar characteristic of periphyton mats vs. vertically-distributed characteristic of vegetation & periphyton in SAV communities) and/or hydraulic phenomena (increased short-circuiting at shallow depths). At a water depth of 30 cm, depth-adjusted K values would be 8, 24, 45, and 20 m/yr, respectively. Thus, in shallow systems with phosphorus and hydraulic loads typical of STA polishing cells, K values for PSTA (24 m/yr) and SAV communities (20 m/yr) would be similar. This may reflect the fact that these communities would tend to have the same composition under these conditions, as embodied in the NEWS concept.
The potential effect of substrate (peat vs. shellrock or limerock) is an important issue with respect to PSTA systems because of the high potential cost of substrate preparation in full-scale treatment areas. The PSTA datasets include 11 experiments with shellrock substrates (including the prototype used for calibration) and 7 experiments with peat substrates. South Test Cell 13 (treatments PSTC_1_PE and PSTC_4_PE) had a peat substrate and rapidly evolved into a cattail community. No attempt was made to control cattail growth, which may have been promoted by the high phosphorus content of the peat initially placed in the cell. To reflect the actual vegetation characteristics (vs. focus of research project), these datasets have been placed in the emergent category. Performance was similar to that observed for emergent platforms. Emergent growth was controlled to some degree in the remaining 5 PSTA experiments that were conducted in mesocosms. These have been included in the datasets used for testing the PSTA and NEWS calibrations (see datasets labeled "PSTA_??_PE"). These calibrations under-predict the average flow-weighted and geometric mean outflow concentrations in these cells by 12% and 9%, respectively (significantly different from zero at p < .10). The K values calibrated to peat mesocosms (mean = 18 m/yr, std dev = 5 m/yr) are below those calibrated to shellrock systems (mean = 23 m/yr, std dev = 5 m/yr), (p = 0.07). Thus, there is some indication that model predictions may be biased for PSTA communities on peat substrates, but the bias is small. It is also possible that the effect of peat substrate would decrease after longer periods as phosphorus release from antecedent soils diminishes, provided that cattail growth is controlled.
The NEWS calibration is recommended for periphyton communities. The PSTA calibration can be used in cells with inflow concentrations less than 25 ppb managed to promote periphyton. Most of the supporting platforms had shellrock substrates. Assuming that macrophyte invasion can be controlled, available data indicate that PSTA or NEWS calibrations may under-predict cell outflow concentrations by ~10% on peat substrates. If macrophytes dominate the population, the performance is expected to similar to that predicted using the EMERG calibration.
Macrophytes accounted for ~50% of the measured phosphorus storage in the PSTA prototype (ENRP South Test Cell 8). The performance of PSTA South Test Cell 3 (shellrock) has been slightly better than that predicted by the NEWS calibration in recent months (May 2001 - March 2002), despite development of a significant macrophyte population. Apparently, an appreciable macrophyte biomass can develop in the PSTA community before performance is degraded below that predicted by the PSTA or NEWS calibrations. It is possible that performance is degraded when emergent vegetation is sufficiently dense to exclude periphyton by some mechanism (e.g., shading).
Discussion - Emergent Calibration Model testing results are generally satisfactory for the emergent (EMERG calibration). Observed & predicted outflow concentrations for emergent platforms are compared in the attached figures (time series, platform means). The calibration is consistent with the data from the northern portion of WCA-2A, the largest platform and the prototype used for designing the macrophyte-based STA's. Consistent biases in model residuals are not apparent over a wide range of conditions (weekly residuals, platform means) when the calibrations are applied to independent datasets. Most of the emergent datasets are in relatively low phosphorus loading ranges that are typical of STA footprints as a whole (~1 g/m2-yr). The performance of 3 out of 4 emergent test cells subsequent to the period used for model calibration/testing has been consistent with that predicted by the EMERG calibration. Given that higher uptake rates of the periphyton and SAV communities, emergent vegetation is more likely to be placed in the inflow region than at the outflow region of treatment areas. Flow-weighted-mean concentrations for the emergent datasets ranged from 20-154 ppb (inflow) and 12-49 ppb (outflow). Projected mean inflow concentrations exceed 154 ppb in some basins (e.g., 167 ppb for STA-5). Using the model in these basins would require extrapolation of the calibrations. Potential implications could be evaluated by testing the model against datasets with higher inflow concentrations.
Summary of Recommended Calibrations
The EMERG & NEWS calibrations provide the best overall fit of data from all platforms, as well as platforms with at least one year of data.
The EMERG calibration is recommended for use in treatment cells that are not managed to promote a specific vegetation type. Depending on antecedent soil conditions, startup procedures, and water level regimes, it is likely that these will evolve into emergent (i.e.. cattail) communities.
The NEWS calibration is recommended for use in treatment cells that are managed to promote non-emergent vegetation (SAV and/or periphyton). The separate SAV or PSTA calibrations also give unbiased predictions within the ranges of conditions represented in the calibration & testing platforms.
The software does not forecast the success or failure of management efforts to produce a specific vegetation type. In assigning a parameter set to a particular cell, the user is making the assumption that the cell can be successfully managed to produce the corresponding vegetation type, as well as species that are typical of those identified in the platforms studied by the research teams.
Model predictions are most reliable when the system variables are within the range of the data used for calibration and testing. The software produces warning messages when the system variables are outside of the data ranges for each parameter set.
Further testing and refinement of the model will be based upon data from ongoing field-scale experiments and monitoring of full-scale STA's.
Detailed Results of Model Calibration & Testing
Preliminary Results - Field-Scale PSTA Cells
Supporting Documents & LinksPotential DMSTA Applications to Marsh Impact & Recovery Assessment under Long-Term Plan Design Basis for Stormwater Treatment Areas Everglades Construction Project Preliminary Application of DMSTA to C-44 Project, prepared for Camp Dresser & McKee, May 2004. Calibration to Lakes, Reservoirs & Wetlands North of Lake Okeechobee, prepared for SFWMD by Wetland Solutions & W. Walker, 1993. |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Kadlec, R. H. & W. W. Walker, “Dynamic STA Design Model - Design Model Structure for Treatment Wetland Systems at Low Phosphorus Concentrations for Everglades Protection”, presented at the Joint Conference on the Science and Restoration of the Greater Everglades and Florida Bay Ecosystem, Palm Harbor, Florida, April 15, 2003. Kadlec & Walker, Technology Review of Periphyton Stormwater Treatment, Draft, November 2003. DMSTA - Features &
Potential CERP Applications, Workshop for CERP/Recover, SFWMD,
October 10, 2003. |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| http://www.wwwalker.net/dmsta/index.htm 03/07/05 | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
