Usefulness of Soil Moisture and Actual Evapotranspiration Data for Constraining Potential Groundwater Recharge in Semiarid Regions

Abstract

AbstractUnsaturated flow modeling is widely used to estimate potential groundwater recharge. In semiarid regions, modeled recharge is usually highly uncertain due to the uncertainty in model parameters and also because recharge is often a small fraction of overall rainfall and actual evapotranspiration (AET). This study attempted to assess the usefulness of soil moisture and AET data on reducing the recharge uncertainty using an experimental site in semiarid southeastern Australia as an example. A biophysically based model capable of simulating unsaturated flow, energy balance, and plant growth was employed to conduct recharge modeling. Soil moisture measured at five different depths of the vadose zone and AET derived from an eddy covariance tower were used to constrain the recharge modeling. The best‐calibrated model shows reasonable agreement between modeled and simulated soil moisture and AET time series, with three‐year average recharge of 50.9 mm. The model uncertainty analysis using Monte Carlo simulations indicates that recharge estimated from the unsaturated flow modeling could be highly uncertain (0–213 mm/year). The uncertainty analysis shows that the soil moisture was more useful than AET in constraining groundwater recharge, particularly the measurements at the top of both soil horizons. The combined use of both the soil moisture and AET led to the most constrained recharge uncertainty (14.8–87.3 mm/year). This study suggests that for practical applications, soil moisture data from at least the top of each soil horizon and AET data (either remotely sensed or directly measured) should be used together to constrain recharge modeling.