The Upper Floridan aquifer (UFA) is the principal water source for industrial and public supply in Glynn County, Georgia. Wells in active pumping centers that tap the UFA for industries near the city of Brunswick have created an upward hydraulic-head gradient in the Floridan aquifer system, which has allowed high chloride (saline) groundwater from the Fernandina permeable zone of the Lower Floridan aquifer (LFA) to migrate upward into freshwater zones. Chloride concentrations of more than 250 milligrams per liter—the State and Federal secondary drinking-water standard—have been measured in a 2-square-mile area near downtown Brunswick.

An existing regional U.S. Geological Survey modular finite-difference groundwater-flow model (MODFLOW-2000) was modified using greater horizontal and vertical resolution to enable more detailed simulation of the effects of pumping in the vicinity of chloride contamination. Modifications to the regional model consisted of (1) limiting grid size to a maximum of 500 feet (ft) per side in the vicinity of the chloride plume; (2) representing the upper and lower Brunswick aquifers with distinct model layers; (3) similarly, representing upper and lower water-bearing zones of the UFA with distinct model layers in Glynn and Camden Counties, Ga.; and (4) establishing new hydraulic-property geographic zones in the UFA within Glynn County. The revised groundwater-flow model was calibrated to steady-state conditions that were assumed to exist during 2000 and 2004. The calibration and framework of the revised groundwater-flow model were documented in a separate report. For the current study, steady-state conditions were calibrated using October 2015 pumping rates in the Brunswick/Glynn County area as a 2015 Base Case. The 2015 Base Case simulation was used as the basis to evaluate seven groundwater-management scenarios in the Brunswick/Glynn County area.

Seven groundwater management-scenarios were developed on the basis of short- and long-term groundwater-use projections for the UFA in the Brunswick/Glynn County area. Scenarios A and B simulated additional pumping in the upper water-bearing zone (UWBZ) of the UFA at existing public-supply wells located near a chloride plume and planned public-supply wells to be constructed north of downtown Brunswick. Scenario C simulated a shutdown at Brunswick Cellulose Inc. and Pinova Inc. and the resulting deactivation of nine production wells, with a combined total pumping of 31.3 million gallons per day (Mgal/d) for the 2015 Base Case simulation. Scenario D (three scenarios) simulated 12.5, 25, and 50 percent (designated Scenarios D1, D2, and D3) of the total pumping of 31.3 Mgal/d at Brunswick Cellulose and Pinova. The objective of Scenario D was to determine pumping rates that may reverse groundwater-flow directions toward the Brunswick Cellulose well field and potentially allow groundwater with higher chloride concentration to migrate toward nearby public-supply wells. Scenario E simulated an additional pumping of 5 Mgal/d from the UWBZ of the UFA at a recently constructed production well within the Brunswick Cellulose well field.

Backward particle-tracking (MODPATH) analysis in public-supply wells located just outside the chloride plume to the north shows that predominant groundwater-flow directions are from the northeast toward the Brunswick Cellulose well field. The analysis covered 20- and 50-year periods for the 2015 Base Case and Scenario C simulations with 100 percent of backtracked particles remaining in the UWBZ and lower water-bearing zone of the UFA. Groundwater-flow directions are characterized by some vertical movement and dominant horizontal movement away from the chloride plume in the northern Brunswick area. For the 2015 Base Case simulation, the mean rate of particle movement ranged from 268 to 413 feet per year. For the Scenario C simulation, the mean rate of particle movement ranged from 89 to 182 feet per year with 50 percent of particles migrating from the chloride plume area. The rate of particle movement is influenced most by the horizontal hydraulic-head gradient in the UWBZ of the UFA.

The revised groundwater-flow model is subject to the limitations documented in the original model. In addition, the values used for the specified-head boundaries in the Floridan aquifer system for the 2004 calibrated model were based on the sparse data available and were not changed for the 2015 update to the model. These model boundaries control 80 percent of the inflows and about 60 percent of the outflows. Composite-scaled sensitivities of the model parameters indicate the revised model is most sensitive to pumping rates, followed by the horizontal hydraulic conductivity in the UFA for zones along coastal Georgia.