Sensitivity of Groundwater Recharge Estimations to Climate Change and Frequency Distribution of Precipitation Events: Insights from a Lysimeter Modelling Study

An adequate quantification of groundwater recharge is required for sustainable water resource management and robust hydrological model predictions. However, estimation of the future temporal evolution of groundwater recharge rates at annual, seasonal or daily time scales remains a challenging task due to the strong linkage of infiltration processes with predicted changes of the frequency distribution of precipitation. While many studies have addressed recharge processes under climate change scenarios, only limited work has been carried out systematically focusing on the impact of precipitation variability.

Thus, we simulated groundwater recharge in the Swiss pre-Alpine Rietholzbach research catchment based on a water balance model, that was calibrated to daily direct recharge measured at a weighting lysimeter. Three approaches, employing different degrees of complexity, were utilized to obtain future climatic forcing functions. First, a relatively simple delta change factors approach for three stationary time periods was applied. For this scenario 10 different climate model chains were used to determine the predictive uncertainty accompanied with the different GCM (Global Circulation Model) x RCM (Regional Climate Model) combinations. Second, these delta change values were combined with a stochastic weather generator, which allows for a more realistic simulation of climatic variability compared to the simple delta change downscaling approach. Further, additional uncertainty, linked to natural climate variability in the climatic forcing functions, can be determined more efficiently. Finally, the frequency distribution of precipitation events was artificially adjusted to account for a larger amount of extreme events, thus simulating what is expected for the future.

The first simulations indicate that annual recharge is less affected by changes of the frequency distribution of precipitation, however there was a marked impact on the seasonal signal. All approaches demonstrated that the uncertainty of projected recharge rates were relatively large. Preliminary results of the third modeling approach indicate that the impact of extreme precipitation events seems to be underestimated in many climate change impact studies.