Changes in mean evapotranspiration dominate groundwater recharge in semi-arid regions

Groundwater is one of the most essential natural resources and is affected by climate variability. However, our understanding of the effects of climate on groundwater recharge (R), particularly in dry regions, is limited. Future climate projections suggest changes in many statistical characteristics of the potential evapotranspiration (Ep) and the rainfall that dictate the R. To better understand the relationship between climate statistics and R, we separately considered changes to the mean, standard deviation, and extreme statistics of the Ep and the precipitation (P). We simulated the R under different climate conditions in multiple semi-arid and arid locations worldwide. Obviously, lower precipitation is expected to result in lower groundwater recharge and vice versa. However, the relationship between R and P is non-linear. Examining the ratio R/P is useful for revealing the underlying relation between R and P; therefore, we focus on this ratio. We find that changes in the average Ep have the most significant impact on R/P. Interestingly, we find that changes in the extreme Ep statistics have much weaker effects on R/P than changes in extreme P statistics. Contradictory results of previous studies and predictions of future groundwater recharge may be explained by the differences in the projected climate statistics.