Approach

Ground-water recharge and discharge in the Everglades usually has been estimated by surface-water mass balance modeling (MacVicar and others, 1994; Choi and Harvey, 2000). The precision of those estimates is linked closely to the density and quality of surface-water flow and stage measurements (Choi and Harvey, 2000). This investigation builds on former studies through analysis of ground-water hydraulic data, peat hydraulic properties, and through direct measurements of vertical fluxes between ground water and surface water using seepage meters. Those additional measurements allowed characterization of spatial and temporal variability of recharge and discharge at the study sites.

The purpose of this approach was broader than just improving the accuracy of previous estimates of recharge and discharge. As mentioned previously, hydraulic data from ground water have the advantage of pinpointing locations of recharge and discharge. By determining the spatial distribution of recharge and discharge, it was possible to test hypotheses about the factors that are most important in controlling recharge and discharge in the Everglades. In particular, most previous work assumed that ground-water flow beneath levees dominates discharge and recharge in the Everglades. The present study quantified those processes in interior areas of the wetlands to test that assumption. Results were expected to be used in guiding future improvements of regional models such as SFWMM.

Detailed methods were published previously in Harvey and others (2000). The only information given here is that which is immediately relevant to understanding the results presented in this section. Harvey and others (2000) should be consulted for all further information concerning survey accuracies, well and seepage-meter design and emplacement, slug tests in peat, and measurement of surface-water and ground-water levels.

Regional horizontal hydraulic gradients were computed for ENR or WCA-2A ground water in a given aquifer layer by calculating the inclination and direction of a plane that matched most closely the measured heads. By convention, a negative gradient indicated flow in a direction from areas of higher to lower hydraulic head. The average computed directions of horizontal flow are reported in table 13. Those flow directions were compared with the orientation of three research transects (one in ENR and two in WCA-2A) that had been established previously (fig. 1, also see figs. 2 and 3). Research transects in ENR and WCA-2A were chosen because the orientation of transects was thought to be similar to average regional directions of ground-water flow. Horizontal hydraulic gradients along the research transects were computed by dividing the difference in head in two wells screened in the same layer by the distance between the wells. Because the general orientation of transects did compare well with the computed regional gradients in table 13, it was possible to make further computations of horizontal gradients using only the transect data. The advantage of computing gradients from transect data, instead of relying on regional horizontal gradients alone, was the ability to compare results between specific areas of interest, such as areas close to and far from levees.

Vertical hydraulic gradients also were computed using data from all of the study sites in ENR and WCA-2A where two or more wells were screened at different depths. The computation used the difference in head between surface water and shallow wells divided by the peat thickness. The sign convention resulted in a positive vertical gradient when recharge is occurring (downward flow from surface water to ground water) and a negative gradient when discharge occurs (upward flow from ground water to surface water).

Hydraulic conductivity (K) of peat was estimated in the peat by slug tests in piezometers. Slug-test results were compared with an independent estimate based on seepage-meter measurements that were combined with measured vertical hydraulic gradients to calculate K using Darcy’s law. Detailed methods and complete results are provided in Harvey and others (2000).

Recharge and discharge fluxes through peat were estimated using seepage meters or, alternatively, Darcy-flux computations. For reasons explained later, only seepage-meter results are reported for ENR sites and only Darcy-flux calculations are reported for WCA-2A sites. The seepage-meter approach, meter design, procedures for installation and use, major assumptions, and precision and limit of detection, all are described in Harvey and others (2000).

In contrast to results from ENR, seepage-meter measurements at WCA-2A were judged as unreliable. The major evidence for rejecting seepage-meter results at WCA-2A was that the direction of fluxes measured on particular sampling dates often did not agree with the direction indicated by hydraulic gradient data. For example, on the westernmost transect in WCA-2A (F1-F4-U3), seepage-meter measurements often indicated recharge at the same time that hydraulic gradient data suggested discharge (Harvey and others, 2000).

The seepage-meter measurements at WCA-2A may have been compromised by problems that were specific to WCA-2A. For example, errors were caused in part by the higher hydraulic conductivity of peat at WCA-2A (table 9), which may have been high enough to cause measurement errors due to stretching or compression of seepage bags because of wind or moving vegetation. It should be noted that seepage-meter data collected at single sites in WCA-2B and WCA-3A were judged as reliable. This reliability was because seepage measurements from triplicate meters at WCA-2BS and WCA-3B15 tended to be similar and close to or slightly above detection limits.

The alternative to seepage meters was calculation of vertical fluxes through peat using Darcy’s law. Briefly, the Darcy-flux approach used a measured vertical hydraulic gradient across the peat and a vertical hydraulic conductivity for peat to compute a vertical water flux between ground water and surface water. The vertical gradient was estimated as the difference between surface-water stage and water elevation in the shallowest monitoring well (generally the "GW4" well located approximately 10 ft below the peat surface), divided by the thickness of the peat at that location.