Subsurface seepage across the eastern boundary of the Everglades in southern Florida is controlled, in part, by a levee and canal system, which was constructed to maintain sheetflow conditions within the Everglades and prevent flooding in urban areas. Because of water-level differences between these areas, a relatively steep hydraulic gradient exists across the levee system, which, coupled with a highly permeable aquifer, causes considerable subsurface flows. Recognizing the potential magnitude and importance of subsurface seepage across the levees, the U.S. Geological Survey (USGS) in conjunction with other Federal, State, and local agencies and Indian tribes, has included the quantification of subsurface seepage as one of the objectives of the USGS South Florida Ecosystem Program.

The purpose of a recent study conducted by the USGS, as part of the South Florida Ecosystem Program, was to develop a fully integrated model in the vicinity of Levee 31N capable of simulating three-dimensional ground-water flow in the Biscayne aquifer combined with surface-water routing capabilities for the L-31N Canal. The modeling results were utilized to quantify seepage rates below Levee 31N and to develop an algorithm suitable for providing real-time estimates of seepage through and below the L-31N canal. The study site encompasses a 110-square kilometer area along Levee 31N and includes part of the East Everglades and the western extent of suburban Miami-Dade County. The eastern and western portions of the site are separated by Levee 31N and its adjacent canal (L-31N). The northeastern part of Everglades National Park (ENP), including the eastern edge of Northeast Shark River Slough, is on the western part of the study site. The eastern part of the study site includes lakes created by rock-mining activities, three municipal pumping wells (combined pumping rate of 0.66 cubic meter per second), and urban development. Immediately underlying the land surface is the Biscayne aquifer, which is characterized by hydraulic conductivities that generally exceed 3,000 meters per day and transmissivities of about 90,000 square meters per day. Throughout most of the site, the ground-water table is at a depth of 1 to 2 meters below land surface, and the L-31N canal intersects the upper part of the aquifer, permitting a direct hydraulic connection.

Model development for the study included gathering data for input and calibration. Supplementary data collected as part of this study included direct measurements of seepage within ENP, lithologic information obtained from two new geologic cores, and data collected from the operation of newly installed monitoring wells. Results from seepage meter tests indicate that vertical seepage rates are dependent upon vertical and horizontal hydraulic gradients and increase within the Everglades nearer to the levee. The geologic cores collected in the Everglades, along with data collected from other sources, reveal the presence of two semiconfining layers below the study site. These layers were incorporated into the model.

The numerical model utilized for simulation purposes was a modified version of MODBRANCH. MODBRANCH couples a quasi-three-dimensional ground-water flow model, MODFLOW, with a one-dimensional surface-water routing model, BRANCH. Modifications to the MODBRANCH code focused on changing the original relation used to model leakage between the surface-water and ground-water systems to one referred to herein as the ìreach-transmissivityî relation. The performance of the modified version of the code has been verified using relatively simple runs with known solutions as well as by comparison to results from the original version of MODBRANCH for more complex scenarios. The numerical model results were used to develop an applied real-time seepage algorithm, which is an algebraic relation based theoretically on the reach transmissivity equation incorporated into the modified MODBRANCH routine. Leakage estimates using the algorithm require only aquifer heads, and the algorithm can be used with real-time monitoring data to provide estimates for seepage along Levee 31N.

Results of the study are of significance to water managers and those involved in Everglades restoration efforts. For example, seepage estimates provided by the South Florida Water Management Model (SFWMM), which is used by local water managers to determine the effect of canal operations on water levels throughout southern Florida, can be checked with the results of this study. The current seepage algorithm used by the SFWMM is a regression relation that was empirically derived through comparisons with a two-dimensional ground-water flow model. The new algorithm developed through this study is physically based.

Results of this study also are necessary for developing a water balance for the northeastern part of ENP, an area which is targeted for future restoration efforts. Water deliveries to the area will likely change in the near future to affect a change in water levels. The model developed through this study can be used to estimate the impact of water-level changes on subsurface seepage rates below Levee 31N, and water deliveries to the area can be adjusted accordingly with respect to seepage loses. Additionally, subsurface outflows from ENP replenish a part of the Biscayne aquifer which underlies urban areas. During dry periods, the aquifer is at risk of saltwater intrusion near the coast, induced by limited inflows and municipal well-field pumpage. During wet periods, urban areas are prone to flooding which can be aggravated by excessive subsurface flows from the Everglades eastward. Maintaining acceptable water levels in urban areas requires quantifying and ultimately controlling subsurface flows below the levee system. An accurate estimate of the transfer of water across the Everglades boundary is, therefore, crucial for water management in urban areas. The model also can be used to evaluate the potential effects of proposed expansion of rock-mining activities and municipal well-field pumpage immediately east of Levee 31N on seepage rates below Levee 31N and, ultimately, the water balance of the East Everglades.

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