Assessing impacts of future potential climate change scenarios on aquifer recharge in continental Spain

摘要

Climate change will modify the availability of groundwater resources in the future. Thus the evaluation of average aquifer recharge from precipitation, and its uncertainty, becomes a key subject in determining suitable countrywide water policies. The confident prediction of renewable groundwater resources requires an accurate evaluation of aquifer recharge over time and space, especially in large territories with varied conditions for aquifer recharge such as continental Spain. This study asses impacts of future potential climatic change scenarios on distributed net aquifer recharge (NAR) from precipitation over continental Spain. For this, the used method (1) generates future time series of climatic variables (precipitation, temperature) spatially distributed over the territory for potential aquifer recharge (PAR), and (2) simulates them within previously calibrated spatial PAR or NAR recharge models from the available historical information to provide distributed PAR or NAR time series. The information employed comes from the Spain02 project for the historical climatic data, from Alcala and Custodio (2014, 2015) for the historical spatial NAR, and from the CORDEX EU project (2013) regional climate models (RCMs) simulations for the future climate scenarios. A distributed empirical precipitation-recharge model is defined by using a regular 10 km x 10 km grid, and assuming that precipitation (P) and temperature (T) are the most important climatic variables determining PAR, while their spatiotemporal variabilities determine the impacts of future potential climatic scenarios on renewable groundwater resources. Potential plausible pictures of future climate scenarios are defined by combining information coming from different RCMs and General Circulation models (GCMs), downscaling techniques, and ensemble hypothesis. These scenarios were simulated within the used precipitation-recharge model to estimate impacts on NAR. The results show that global mean NAR decreases by 1