publications > scientific investigations report > surface-water and ground-water interactions > summary and conclusions

Knowledge of interactions between surface water and ground water is central to an understanding of water budgets, water quality, and ecology in the Everglades, a wetland of national and international significance for which there is presently very little previous information of that kind. Vertical fluxes entering the subsurface (by recharge) or returning to surface water (by discharge) are the principal pathways by which surface water is exchanged with ground water in the underlying peat and sand/limestone aquifer. Managers and restoration planners have few reliable estimates of recharge and discharge in Everglades wetlands, especially in the parts of the wetlands far from the levees. Possibly the most poorly understood aspect of the problem is how the actions of water managers affect interactions between surface water and groundwater in ways that the affect Everglades ecology. One barrier to progress is the logistical constraints on measurements; another is inexperience on what are the best measurements to make to support various needs across several disciplines. This report is a product of a cooperative investigation conducted by the USGS and the South Florida Water Management District (SFWMD) aimed at developing and testing techniques that would provide reliable estimates of recharge and discharge in interior areas of WCA-2A and several other sites in the central Everglades. Goals included testing and comparing several new methods to estimate recharge and discharge, in addition to characterizing spatial and temporal variability of recharge and discharge, and determining the relative importance of several possible controlling factors. The new methods of estimating recharge and discharge included (1) Darcy-flux calculations based on measured vertical gradients in hydraulic head and hydraulic conductivity of peat; (2) modeling vertical transport and decay of the naturally occurring isotopes ²²⁴Ra and ²²³Ra (with half-lives of 4 and 11 days, respectively) through peat; and (3) modeling of transport and decay of naturally occurring and "bomb-pulse" tritium (half-life of 12.4 years) in surface water and ground water.

The Darcy-flux measurement approach was used to estimate ground-water recharge and discharge at 15 sites in the Everglades Nutrient Removal (ENR) Project area and in Water Conservation Area 2A (WCA-2A). This approach required estimates of hydraulic conductivity of peat that were made at 11 of the 15 sites. A simple hydrogeologic simulation was used to assess how levees at the margins of Water Conservation Area 2A have influenced recharge and discharge. Simulations and measurements showed that the highest rates of recharge and discharge (approximately 2 cm per day) occurred within 600 m of levees, as a result of ground-water flow beneath levees. The simulations suggested that recharge and discharge should be orders of magnitude smaller in the interior areas of WCA-2A (> 600 m from levees). However, measurements showed that recharge and discharge were substantially higher than simulations predicted, comparable to fluxes near levees. A 5-year time series (1997-2002) of Darcy-flux estimates indicated that recharge and discharge in the interior wetlands of WCA-2A reversed in direction on weekly, monthly, and annual timescales. Ground-water discharge tended to occur during average to moderately dry conditions when local surface-water levels were decreasing. Recharge tended to occur during moderately wet periods or during very dry periods just as water levels began to increase following precipitation or in response to a "pulse" of surface water released from water-control structures. Discharge also tended to occur at sites in the wetland interior for approximately a week preceding the arrival of the surface-water pulse. It was concluded that ground-water recharge and discharge appear to vary cyclically in the interior wetlands of the central Everglades, and are driven by the differential responses of surface water and ground water to annual, seasonal, and weekly trends in precipitation and operation of water-control structures.

Measurement of environmental solute tracers that retain information about the time that has elapsed since recharge occurred offer another possible solution to estimating recharge and discharge in the Everglades. Activities of short-lived radium isotopes (²²³Ra and ²²⁴Ra ) were measured in pore water of peat at several research sites in WCA-2A. These radium concentration profiles differed from the amount that could be explained without water flow by local production, decay, and exchange with solid phases. The measured disequilibrium is caused by vertical transport of radium with water flowing vertically in the peat along flow paths connecting ground water in the underlying aquifer with surface water. The rate of vertical water flow through wetland sediment was determined from the radium disequilibrium using a combined model of transport, production, decay, and exchange with solid phases. This technique was tested in WCA-2A by quantifying vertical advective velocities at three sites. Vertical water fluxes were determined to be between 0 and 0.5 cm per day for a recharge site, and 1.5 ± 0.4 cm per day for a discharge site located on the upgradient and downgradient sides of the Hillsboro levee, respectively. A site in the interior of WCA-2a experienced both recharge and discharge at an "exchange" rate of approximately 0.9 cm per day. The radium technique should be applicable to any wetland system with different production rates of these isotopes in distinct sedimentary layers or surface water, and is most straightforward in systems where constant ionic strength in pore-water can be assumed, thereby simplifying the modeling of radium exchange.

Average long-term (decadal timescale) recharge and discharge fluxes across the ground surface were estimated in WCA-2A by simulating transport of tritium (³H) and radioactive decay in surface water and ground water. Model parameters included the storage depth of shallow ground water near the top of the aquifer that exchanges with surface water (referred to as "interactive" ground water), average residence time of that ground water, and the associated recharge and discharge fluxes. The residence time of interactive ground water in the simulation was adjusted to achieve the best fit with measured concentrations of tritium concentrations and the measured depth distribution of tritium in the aquifer. Several direct estimates of ground-water residence time determined from tritium/helium-3 isotopic ratios (³H/³He) provided an important check on the results. Recharge and discharge fluxes were computed directly from best estimates of average residence time and depth of shallow, interactive ground water. The results of a model using only tritium data suggested that interactive ground water had an average residence time of 90 years and a storage depth of 3.1 m. Both the residence time and storage depth estimates are expected to be overestimated by the approach that only used tritium data because of the effects of vertical mixing with deeper, older, "tritium-dead" ground water. Analysis of ³H/³He, which is not sensitive to mixing with deep, tritium-dead ground water, indicated an approximate residence time of 25 years for shallow ground water. Results from both the tritium model and ³H/³He analysis were that the long-term average estimates of recharge and discharge in WCA-2A are on the order of 0.01 cm/d.

A comparison between results of the new methods described above revealed order-of-magnitude differences, with fluxes ranging between 0.01 and 3 cm per day. These differences must to some extent reflect real spatial and temporal variability of recharge and discharge in the wetlands. However, an even more important determinant of the order-of-magnitude variability of these estimates is of the inherent limitations of each method. Only a small component is detected by each method of the full distribution of recharge and discharge fluxes that are operative at all spatial and temporal in nature. For example, the comparison of methods demonstrated that recharge and discharge estimates decrease with increasing depth in the subsurface. This is not an artifact of measurements but instead reflects the true nature of surface-water and ground-water interactions in the Everglades. For example, it is only a matter of hours to days from the time that surface water is recharged into shallow flow paths through peat soil before it reemerges as discharge. Meanwhile, the much smaller component of the recharged water that flows more deeply into the sand and limestone aquifer (meters of tens of meters) could be retained for decades, centuries, or even millennia before it is returned to the surface as ground-water discharge. The result is a distribution of flow depths and associated residence times of recharged water in ground water prior to discharge back to surface water.

Another important influence on results is the interaction between true signals of temporal variability of recharge and discharge (especially reversals between these fluxes) and the averaging timescale selected by the investigator. Furthermore, there are inherent aspects of each measurement technique (such as the half-life of an environmental tracer) that influence the sensitivity of each method to a particular timescale of surface-subsurface exchange. For example, modeling transport of the short-lived isotopes of radium in the 1-meter-thick layer of wetland peat provided one of the largest estimates of recharge and discharge (0.9 cm per day), because the approach used relatively short-lived isotopes of radium (4 and 11 day half lives) that are sensitive to the relatively high-frequency (weekly to monthly) reversals in the flux direction caused by precipitation events and surface-water releases from water-control structures. In contrast, recharge and discharge estimates based on modeling transport of the much longer-lived (12 year half- life) isotope tritium were not sensitive to the high-frequency reversals in flux direction in peat, due to tritium's longer half-life and also to fact that tritium was measured in the sand and limestone aquifer where the effects of high-frequency fluctuations in peat pore water are damped out. Consequently, tritium modeling was sensitive only to the relatively small component (0.01 cm per day) of the total exchange fluxes that involve decadal timescale interactions between surface water and ground water.

The South Florida Water Management Model (SFWMM) has been an important tool and has been used extensively by the South Florida Water management District (SFWMD) to design many of the hydrological aspects of the Everglades restoration. Therefore, comparing results of the new methods presented in this report with SFWMM results was important. Because the SFWMM is spatially discretized on a 2-mi by 2-mi square grid, and because recharge and discharge are often estimated from modeling results by averaging on annual or longer timescales, the SFWMM also is subject to scale dependence in its results. Like tritium modeling, longer runs of the SFWMM generally provide results that reflect longer timescale and deeper subsurface interactions between surface water and ground water. For example, a decadal timescale run of the SFWMM (1979 - 1990 "calibration" simulation) produced an estimate of recharge and discharge (0.03 cm per day) that was consistent with tritium modeling (0.01 cm per day). Decreasing the length of a model run for the SFWMM appears to increase its sensitivity to shorter term interactions between surface water and ground water. For example, a shorter (5-year) run of the SFWMM (1991-1995 "verification" simulation) produced an estimate of recharge and discharge that was consistent with Darcy-flux calculations in this investigation (0.1 compared with 0.2 cm per day for the Darcy-flux calculations, respectively).

In summary, measurements of recharge and discharge in the central Everglades are both spatially and temporally scale dependent. As a consequence of the scale dependence, there is no simple measure of recharge and discharge that is generally applicable for all uses. Different methods to estimate recharge and discharge are inherently limited by the technique that is selected as well as by the spatial and temporal averaging of the measurements imposed by the investigator. Consequently, no single method quantifies the full spectrum of shallow and deep recharge and discharge fluxes that are simultaneously active in the Everglades. Each method, however, potentially provides useful information about a particular spatial and temporal subset of the total recharge and discharge fluxes that are occurring. For example, the daily to monthly timescale fluctuations in recharge and discharge that are detected by Darcy-flux calculations and radium modeling are mainly informative about exchange of surface water with peat pore water, whereas the longer-term averaging by tritium modeling and the SFWMM are mainly sensitive to deeper exchange with ground water in the underlying aquifer.