Home Introduction Hydrogeology Simulation of GW Discharge -Equations -SEAWAT -X-Sxn. Model -Regional-Scale Model -Model Limits Conclusions References Appendix 1 Appendix 2 Appendix 3 Appendix 4 Plates PDF Version

Simulation of Ground-Water Discharge to Biscayne Bay

The simulation of ground-water discharge to Biscayne Bay was performed at the local and regional scales. Two local-scale models were developed in cross section to simulate the complex ground-water flow patterns near the coast of Biscayne Bay. With the assumption of steady-state conditions, the cross section models were calibrated using ground-water data collected as part of this study. Results from the cross-sectional models were then used to aid the development of the larger, regional-scale model and provide a detailed spatial representation of the ground-water discharge patterns. The regional-scale model simulates transient ground-water discharge to Biscayne Bay in three dimensions. The regional-scale model was developed and calibrated using field data for an approximate 10-year period from January 1989 to September 1998.

The cross-sectional and regional-scale numerical models were developed using the conceptual hydrologic model shown in figure 15, which is a simplified representation of the actual system. Recharge to the Biscayne aquifer is assumed instantaneous, meaning that the unsaturated zone is not explicitly represented. Another assumption is that the Biscayne aquifer can be simulated with an equivalent porous medium (EPM). Historical observations suggest that isolated, submerged ground-water springs were once found within Biscayne Bay. By using the assumption of an EPM, individual springs and conduits are not explicitly simulated, but rather the properties of the conduit are averaged within model cells. This assumption limits the interpretation of model results at local scales, but is thought to be appropriate when conduits or fractures are much smaller than the scale of the model.

Figure 15. Conceptual hydrologic model used to develop numerical models of ground-water flow. [larger image]

The numerical models were calibrated to field data to ensure that they are reasonable representations of the physical system. Calibration is a subjective process that requires altering the parameters of the numerical model until model results compare with field observations. This process is sometimes referred to as "solving the inverse problem." When an acceptable match between simulated results and field observations is obtained and realistic aquifer parameters are used in the model, it is assumed that the model is a reasonable representation of the physical system. Solving the inverse problem is complicated because more than one set of model parameters can produce an equally well calibrated model. To minimize the potential for this problem, the simplest spatial distribution of model parameters was used to calibrate the numerical models.

During the initial phases of the numerical modeling, several different codes were evaluated to determine their effectiveness in achieving project objectives and simulating the complex variable-density flow patterns observed in the Biscayne aquifer. For example, the HST3D code (Kipp, 1997) accurately simulated flow in cross section; however, because of numerical dispersion, this code did not accurately simulate flow at the regional scale. The SUTRA code (Voss, 1984) also was successfully used for preliminary cross-sectional models, but a three-dimensional version of the code was not available at the time. Eventually a relatively new code, SEAWAT (Guo and Bennett, 1998), was selected for the study because preliminary results indicated that the code: (1) accurately simulated variable-density flow, (2) had acceptable options for minimizing numerical dispersion, and (3) read standard MODFLOW and MT3D data sets, which are easily created. The original version of SEAWAT was modified as part of this study to improve functionality and simulation accuracy (Langevin and Guo, 1999).

Go back to Freshwater-Saltwater Transition Zone | Go ahead to Governing Equations