SIMULATION OF BASIN-SCALE CHANGES IN GROUNDWATER QUALITY ASSOCIATED WITH CHANGES IN GROUNDWATER WITHDRAWALS, SALT LAKE VALLEY, UTAH

Changes in groundwater withdrawals can affect the mixing of water from different sources within an aquifer and may affect the suitability of the aquifer as a drinking-water source and other uses. Under natural conditions in the basin-fill aquifer in Salt Lake Valley, Utah, a plume of water containing less than 500 mg/L dissolved solids is surrounded by groundwater of much higher dissolved solids. The source of the low dissolved solids water is mainly precipitation from areas underlain by quartzite and sandstone conglomerate in the Wasatch Range. Groundwater withdrawals have increased in some areas in the basin, and as a result, the distribution of dissolved solids in groundwater has changed because of the mixing of water with different dissolved solids concentrations.

In this study, groundwater flow and solute transport models are used to simulate variations in recharge and groundwater withdrawal and their effect on the distribution of dissolved solids in the basin. The simulations help to clarify the factors associated with historical changes in water quality and are used to predict future water-quality changes given scenarios of future recharge and pumping. Flow and transport models were calibrated using measurements of head and tritium concentrations. The atmospheric tritium peak coincides with the time of increased groundwater withdrawal, and the migration of the tritium front encompasses the area where most of the increased groundwater withdrawal has taken place. The model was calibrated using tracer data that span the time and area in which water-quality changes are expected; this reduces the uncertainty of the predictions made using the model. Using the calibrated model, scenarios of changing source locations and concentrations of dissolved solids allow prediction of basin-scale changes in water quality. Results show a good match to observed dissolved-solids trends.

Uncertainty in the model predictions is being assessed by considering model and parameter uncertainty. A calibration-constrained Monte Carlo analysis will be done for two models, one with a detailed geostatistical realization of the basin and the other using a dual-porosity domain. Simulated concentration changes over time and their variances will be compared.