USGS
South Florida Information Access
SOFIA home
Help
Projects
by Title
by Investigator
by Region
by Topic
by Program
Results
Publications
Meetings
South Florida Restoration Science Forum
Synthesis
Information
Personnel
About SOFIA
USGS Science Strategy
DOI Greater Everglades Science Plan
Education
Upcoming Events
Data
Data Exchange
Metadata
publications > paper > simulation of integrated surface-water/ground-water flow and salinity... > discussion

Simulation of integrated surface-water/ground-water flow and salinity for a coastal wetland and adjacent estuary

5. Discussion

Home
Governing Equations
Numerical Implement.
Application of Model
>Discussion
Summary & Conclusions
Acknowledgments & References
Tables, Figures, Eqs.
PDF Version
Prior to performing simulations with the integrated model, the surface-water and ground-water models were independently developed and calibrated to the extent possible. For the initial surface-water model, exchange with ground water was considered negligible (Swain et al., 2004). Ground-water model development was performed after the surface-water model was developed, and thus simulated surface-water stages and salinities were applied as boundary conditions over the aquifer surface. This stepwise approach had two advantages. First, it was relatively easy to identify and correct input and runtime errors for the individual models before they were integrated. Second, computer runtimes for the ground-water model were only a couple of hours, whereas the integrated model required over 30 h to run. The shorter computer runtimes were particularly useful during calibration of the ground-water model to aquifer salinity. Because of the highly transmissive nature of the Biscayne aquifer and a relatively stable freshwater/saltwater interface in southern Florida (Sonenshein, 1997), aquifer salinities were assumed to be in equilibrium with current water levels and hydrologic stresses. Thus, an additional level of confidence in the ground-water model was established when it could be shown that after the model reached dynamic equilibrium (through repeated simulations), the simulated freshwater/saltwater interface was in the observed location. Only minor salinity adjustments at the ground-water boundaries were required as part of this calibration process as hydrodynamic dispersion was not active for the simulations.

Limitations were periodically encountered using the explicit, time-lagged approach to couple the surface-water and ground-water models. For some sensitivity simulations with very large leakage rates, convergence could not be achieved during solution of the ground-water flow equation. Evaluation of the convergence problems indicated that very large leakage rates caused numerical oscillations in the implicit solution. Ground-water heads measured in the field can respond quickly to hydrologic stresses. For the Everglades application of the integrated model, however, large leakage rates may persist throughout the day in the model, because of the 1-day length of the stress period in SEAWAT, whereas actual leakage rates would decrease as ground-water heads respond more quickly. A decrease in the length of the ground-water stress period could probably have improved convergence. These convergence problems were encountered only in a few instances, and thus the day lag, which is computationally many times faster than using an hourly lag or fully implicit solution, was a necessity for this particular application. Future efforts using the integrated model should follow the example of Fairbanks et al. (2001) and focus on determining the relation between accuracy and efficiency for different coupling approaches and timestep lengths.

The Buttonwood Embankment clearly is an important physiographic feature in the Taylor Slough area. Model results and field observations suggest that freshwater flow into Florida Bay occurs primarily through the coastal creeks, rather than as overtopping of the embankment. This flow pattern has allowed field investigations to quantify with a high level of certainty the flow exchanges between the coastal wetlands and Florida Bay (Hittle, 2000; Hittle et al., 2001). Confidence in the predictive capability of the integrated model is due largely to the accuracy and long-term record length of creek discharge data. Many coastal wetlands in other locations, however, are not separated from the adjacent marine water body by an embankment or barrier to overland flow, and thus measurement of freshwater outflows is not as straightforward.

< Previous: Application of Model | Next: Summary & Conclusions >



| Disclaimer | Privacy Statement | Accessibility |

U.S. Department of the Interior, U.S. Geological Survey
This page is: http://sofia.usgs.gov/publications/papers/swgwflow_salinity/discussion.html
Comments and suggestions? Contact: Heather Henkel - Webmaster
Last updated: 19 October, 2005 @ 11:34 AM(KP)