The system process studies of the southern Everglades area have been integrated into a two-dimensional numerical model of surface-water flow through the wetlands and into Florida Bay. The model chosen for this task is the Surface-Water Integrated Flow and Transport in Two Dimensions (SWIFT2D) code. The SWIFT2D model numerically solves finite-difference forms of the vertically integrated mass and momentum conservation equations in conjunction with heat, salt, and constituent flux transport equations. An equation of state for salt balance is included in the model to account for pressure gradients that result from density variations. Thus, the hydrodynamic and transport computations are directly coupled. Information on ground-water and surface-water elevations, rainfall, evapotranspiration, frictional resistance terms, boundary water levels and flows, wetland flow velocities, ground-water inflows, and wind-frictional terms are all derived from the process studies and incorporated into the model. The model serves to synthesize the results of the process studies into an integrated dynamic representation of the hydrologic regime.

For initial calibration and testing, a 2-month (July-August 1997) simulation period was used. This period coincided with field-measured flow velocities in the wetland and at model boundaries. The field data that are most significant for model comparison are the continuous discharge measurements made at five coastal creeks. The topographic high along the coast, referred to as the Buttonwood Embankment, acts as a major hydraulic control, and the creek flows are a good indication of the behavior of the system. Because field research of controlling system processes yields most of the input data, the calibration parameters are kept to a minimum which allows for more confidence in simulation results.

The comparison of simulated creek flows with and without the effects of wind highlighted the importance of wind-forcing effects on water movement. Frequent flow reversals (inland direction) measured in the creeks could not be simulated without incorporating wind into the model. Using field-measured wind speed and direction in the model, these reversals were well simulated. However, wind-driven water-level fluctuations in Florida Bay are not the primary driving force in flow reversals. Simulating the measured water-level fluctuations in the bay without wind-forcing effects on the inland wetlands does not represent the flow reversals.

With the exception of the deeper parts of Taylor Slough, surface-water flow in the wetlands can be induced to move in any direction by the wind. Flow direction varies so rapidly that model-generated, daily flow-velocity averages cannot be compared with field measurements taken over a period of several minutes. However, model-generated and field unit discharge values (average velocity times depth) that are simultaneous coincide well. Both simulated and field-measured data record episodes when southward flow in the deeper parts of Taylor Slough occurs simultaneously with wind-driven flow to the north in adjacent shallow wetlands.

The flat, slow-flowing wetlands of the Everglades have very low gradients; the largest water-level gradients and the major processes which control flow occur at the coast. Water-level gradients tend to be on the order of 0.15 foot per mile in the inland wetlands, but simulated coastal gradients are, at times, 0.5 foot per mile. This observation is supported by specific coastal creek water-level measurements. The topographic high (Buttonwood Embankment) found along much of the coast is the most important flow control in the system, usually confining flow to the coastal creeks.

During wetter periods, the salinity of Joe Bay and the other subembayments is negligible, and a sharp salinity gradient exists offshore. Despite periodic wind-induced flow reversals, saline water is flushed from inland areas during wet periods. During dry periods, salinity in the Seven Palms Lake area and Joe Bay can be significant.

A 7-month period (August 1996 to February 1997) was simulated to establish coastal boundary conditions to aid U.S. Army Corps of Engineers hydrologic models. For the most part, this extended simulation period used the same input parameters as the shorter simulation. One important difference is the replacement of the northern flow boundary at Taylor Slough Bridge with a water-level boundary, largely due to a lack of flow data for the 7-month period.

The 7-month simulation allows quantification of the long-term effects of wind on total flow at the coast. Although the short-term effects can be dramatic, the net effect on cumulative flow volumes required a separate analysis. At the largest outflow point, Trout Creek, simulated flow volumes were 18 percent lower when the effects of wind are considered. Measured and simulated flow volumes with the effects of wind were within 6 percent of each other, so the wind differences can be considered significant. At lower outflow points, such as Taylor Creek, wind may not have notable effects on flow; differences were less than 2 percent. Trout Creek has a deeper water area upstream, Joe Bay, within which the seiche probably accounts for the higher sensitivity of flow to wind.

An analysis of the short- and long-term simulations suggests that the upper part of Taylor Slough flows southwesterly, but the water generally proceeds southeasterly from the lower slough toward the Joe Bay area. The West Lake/Seven Palms Lake area, which appears to be in-line with the southwesterly orientation of Taylor Slough, actually receives less water from the slough than Joe Bay. This is due to relatively high topography north of the Lakes area and a higher hydraulic gradient occurring between Taylor Slough and Joe Bay. The West Lake/Seven Palms Lake surface-water flows are nearly cut off from Taylor Slough during dryer periods, as has been confirmed by measured flows at the coastal creeks.

The SICS model is a useful tool for understanding the effects of processes controlling the coastal hydrology of Florida Bay and for defining boundary conditions of other models. This model is currently being expanded to include adjacent areas and to further refine the model process.

(This abstract was taken from the Proceedings of the South Florida Restoration Science Forum Open File Report)