Analysis of water levels during aquifer storage and recovery in a municipal well field

Aquifer storage and recovery (ASR) is an effective water management strategy in semi-arid environments characterized by high temporal variability in surface water supplies. In this study we analyze data from a municipal well field in the Denver Basin (Colorado, USA) that has been used for ASR during the past 15 years. We apply a novel analytical model to compute water levels at individual injection/extraction wells. The model includes terms for well-loss effects and recoverable (non-pumping background) water levels. Calibration to historical data is accomplished using multiple optimization steps. First, aquifer transmissive and storage properties are determined by derivative analysis (minimizing the difference between observed and computed temporal water level derivatives), which allows for accurate estimation of aquifer properties without knowledge of static water levels. Subsequent optimization steps are used to determine well-loss coefficients and recoverable water levels. The well-loss effect differs in magnitude depending on the direction of pumping; for similar rates, injection produces much greater head change (at the well) compared to extraction. The calibrated model serves as a management tool that can be used to plan ASR operations.