Stochastic analyses to identify wellfield withdrawal effects on surface-water and groundwater in Miami-Dade County, Florida

Several stochastic analyses were conducted in Miami-Dade County, Florida, to evaluate the effects of wellfield withdrawal on aquifer water levels, canal stage, and canal flow. Multiyear data for withdrawals at four water-supply wellfields, water levels at the S-121 canal control structure and groundwater head at a nearby monitoring well were used to determine the interrelation between wellfield withdrawals and water levels in the canal and aquifer. A spectral analysis was performed first on the wellfield withdrawals, showing similar patterns of fluctuations, but no well-defined seasonality. In order to compare water-level response with withdrawals at each wellfield, the intercorrelation effects between wellfields was removed through a ‘causal chain’ approach where the inter-wellfield correlation is used to isolate the wellfield/water-level correlation.Most computed correlations have magnitudes less than 5 percent, but with statistical significance above 90 percent. Results indicate that withdrawals from the wellfields most distant from the canal had no significant correlation to the canal levels. However the highest correlation was not at the wellfield closest to the canal, but at the two wellfields at the intermediate distance that have higher withdrawal rates. The hydraulic interconnectivity of the canal with the rest of the canal network, covering the study area, allows the canal equalizes with all connected canals. This explains why proximity to a particular canal location does not appear to be as important a factor as the withdrawal rate. Groundwater levels are more highly correlated to a wellfield on the same side of the canal, and to pumping wells in the same wellfield on the same side of the canal. This indicates that canals are an effective barrier and source/sink for the groundwater. Further nonlinear correlation analysis indicates that high withdrawal rates disproportionally affect water levels and are the predominant effect on the canal.

► We analyzed the effects of wellfield withdrawal on aquifer and canal water levels.
► A spectral analysis showed similar patterns of wellfield withdrawals.
► A ‘causal chain’ approach removed the intercorrelation effects between wellfields.
► Wellfield withdrawal rates affect canal levels more than the wellfield's proximity.
► Canals behave as an effective barrier and source/sink for the groundwater.