DISCUSSION

Surface-water and ground-water interactions in a constructed wetland were estimated using simple mass balance calculations. Our approach combined hydrologic and chemical data in order that both ground-water discharge and recharge could be determined. Net ground-water flux (discharge-recharge) was reliably estimated from water-budget balance with relatively low uncertainty, but uncertainty in estimating ground-water discharge using a combined water and solute (tracer) mass balance was very high. Time-varying surface-water flows, water levels, and chemical concentrations in this constructed wetland all contributed to the relatively high uncertainty in our estimates of ground-water discharge. In order to improve our confidence in our estimate of ground-water interactions, it was necessary to compare with estimates obtained from other independent approaches including seepage meters and ground-water-flow modeling. Here, we make that comparison and discuss how ground-water interactions are increased by management of surface-water flows at ENR.

Comparison with Independent Measurements

Figure 6. (above) Locations of seepage meters, distribution of vertical hydraulic gradient, and estimation of area-averaged seepage-meter fluxes in ENR [larger image]

Seepage Meters. Even our best estimates of ground-water discharge determined from mass balance approach had relatively high uncertainty. For that reason, we compared results with independent estimates based on measured vertical hydraulic gradients and seepage fluxes determined from seepage meters (Harvey et al. 2000). The seepage meters were constructed from 0.64-cm high density polyethylene sheets molded into conical domes (0.76-m diameter) with a circular ring (0.3-m high) that could be pushed into the peat sediments (Harvey et al. 2000). Between two and four seepage meters were installed at each of twelve sites in ENR (Figure 6). Meters were operated by attaching prefilled bags for periods ranging between 1 hour and 1 week, depending on the rate of change of water volume in the bag. In order to determine area-averaged fluxes by extrapolating to areas without seepage meters, the results from several seepage meters at each site were first averaged. Then those results were combined by grouping and averaging sites that had a similar vertical hydraulic gradient, as determined from wells (Figure 6). Those area-averaged estimates of discharge ranged from 0.3 to 2.4 ha-m/day for four different days that spanned the extremes of differences in surface-water levels between ENR and Water Conservation Area-1 to the east. The four discharge fluxes were regressed against corresponding water-level differences in ENR and WCA-1 (Figure 7a). The best-fit relationship (r² = 0.984) was used to extrapolate the ground-water discharge for the 4-year period of interest (Figure 7b). The average ground-water discharge estimated from the above approach was 0.9 ha-m/day (or 0.06 cm/day). The ground-water recharge was also estimated by seepage-meter measurements in the western part of the ENR where strong downward hydraulic gradients indicate that recharge occurs (Figure 6). The estimated ground-water recharge was 13.0 ha-m/day (or 0.84 cm/day). The seepage-meter-derived estimates were quite similar to our coupled mass balance estimates, which increased our confidence that discharge and recharge are reliably estimated (Table 4).

Figure 7. (above) Independent estimate of ground-water discharge using seepage meters. (a) Regression of seepage-meter estimates of ground-water discharge against water-level differences between eastern ENR and WCA-1 (r2 = 0.984). (b) Ground-water discharge computed for 4-year study period (two-week interval) using regression equation from seepage-meter estimates [click on images above for larger version]

The coupled mass balance and seepage-meter measurements showed good agreement in estimates of both ground-water discharge and recharge. In addition, the net ground-water flux (G_i - G_o) calculated from seepage-meter-derived estimates (-12.1 ha-m/day) was very well matched with the mass balance estimate (12.0 ha-m/day). However, ground-water-flow modeling estimated a net ground-water flux of -8.5 ha-m/day, which is approximately 30% less than both seepage and mass balance estimates (Table 4). We believe that the disagreement was caused by an unrealistically high estimate of ground-water discharge by the ground-water-flow modeling method (Table 4). It is easy to envision a large potential error in the ground-water-flow modeling approach due to difficulty in accurately estimating vertical hydraulic conductivity. Errors in estimating hydraulic conductivity of the wetland peat could be particularly troublesome, since the peat acts as a restricting layer for vertical flow in this system (Harvey et al. 2000).

Are Managers Increasing Ground-water Recharge through Design and Operation of Constructed Wetlands?

Recharge from ENR surface water to the aquifer was the most significant hydrologic interaction with ground water over the 4 years of the study period. Recharge in constructed wetlands is very important because it can potentially transport contaminants, such as nutrients and mercury, into the underlying aquifer. The seepage canal, which is located on the western and northern sides of the ENR, was designed to prevent flooding of adjacent agricultural fields by capturing and returning the recharged ground water to the ENR (Figure 2). The flux of water captured by the seepage canal is significant (22% of pumped inflow from seepage canal). However, the returned inflow from the seepage canal is mostly lower than the estimated ground-water recharge flux (Figure 8a), which indicates that there is a component of recharged ground water that is not captured by the seepage canal. The difference between recharged ground water and recycled seepage canal flow is greatest when the water level in ENR is highest (Figure 8a). In order to determine the control that managers have over the extent of ground-water recharge, our estimate of ground-water recharge was plotted against the rate of surface-water inflow by pumping from the supply canal (Figure 8b). Recharge was positively correlated with the pumping rate of surface water from the supply canal into ENR. This demonstrates that the relatively high surface-water inputs to this constructed wetland have the unintended effect of increasing ground-water recharge.

The effect of ground-water recharge on contaminant mass balance has not been fully investigated. Contaminants such as nutrients and mercury recharged in the wetland could have unexpected transport paths and fates in the aquifer. For example, recharged water that is not captured by the seepage canal could take unexpected pathways and possibly discharge outside of treatment wetland, releasing contaminants to surface water. It is also possible that certain contaminants could be retained or transformed by interaction with aquifer solids within the aquifer. The answers to these questions are certainly important at ENR, where such a large proportion of the water intended for treatment is recharged.

Figure 8. Comparison of estimated ground-water recharge with (a) returned flow from seepage canal (G-250_S) and water level in ENR (ENR202D) and (b) inflow rate of surface water from supply canal into ENR (G-250) [click on images above for larger version]

< Results | Conclusion >