The finite-difference model SEAWAT was used to model the flow of submarine groundwater discharge (SGD) into the Indian River Lagoon, an estuary in central Florida. The anisotropy ratio was spatially calibrated in the entire domain using four methods: (1) estimation of the root mean square error (RMSE), (2) estimation of the Nash-Sutcliffe efficiency (NSE) index, (3) two-sided t-test, and (4) visual comparisons of measured and model-predicted freshwater equipotential lines below the entire transect. The measured freshwater equipotential lines were obtained by the inverse distance weighted interpolation (IDW) of measured heads at 17 points. Calibration by the RMSE calculation was found to be meaningless since model predictions under or over predicted the annual SGD by 43 to 211% even when RMSE values were small (0.048 m to 0.085 m). The NSE statistic was not reliable when comparisons were made at the 17 measured points. However, adding 18 more points based on the IDW greatly improved the NSE calibration. The results provided by the two-sided t-test agreed well with the visual comparison. It is recommended that the IDW be conducted to obtain a complete distribution of the freshwater equipotential lines prior to conducting statistical analyses. The IDW not only improves the statistical analyses but also provides another calibration tool through visual comparison. Both the two-sided t-test and the visual comparison indicated that good statistical results were obtained when the predominant anisotropy ratio was 100 and values of 10,000 or higher were used close to the lagoon bed.
展开▼
