ON THE INTERPRETATION OF RECHARGE ESTIMATES BASED ON STEADY-STATE MODEL CALIBRATIONS

Groundwater recharge is often estimated through the calibration of groundwater flow models. In many circumstances, hydrogeologists have no choice but to use steady-state models to simulate groundwater flow. When calibrating such a model to synoptic water-level data, errors are naturally introduced to inferred recharge values because data on temporal variability may be insufficient to justify transient simulations or to calibrate to average water-table elevations. We examine the nature of these calibration errors by considering some simple mathematical and numerical calculations. These calculations demonstrate that calibrating a steady-state groundwater flow model to maximum or minimum water levels yields estimates of steady-state recharge that have the same value as the transient recharge at the time the water levels are measured. These steady-state recharge values, however, are less than the maximum transient recharge rates and greater than minimum transient recharge rates. Because minimum and maximum water levels occur infrequently, most models will be calibrated to water-level data collected during times of rising and falling water levels. Calibrating a steady-state groundwater flow model during rising water levels will produce a recharge rate that underestimates transient recharge; conversely, calibrating a model during falling water levels overestimates transient recharge. We also calculate the minimum monitoring time required to produce an average water level by modifying the dimensionless parameter of Townley (1995), which relates the aquifer storage, half-length, and transmissivity to transient forcing periods. Our calculations suggest that this parameter may be used not only under ideal, periodic conditions, but also in highly-transient systems that involve fluctuations at time scales ranging from hours (storm events) to months (seasonal fluctuations) to years (interannual fluctuations such as ENSO).