How confident are we with the global groundwater recharge estimates under climate change?

Abstract:

Anthropogenic climate change will alter the hydrological cycle, impacting water resources in multiple ways. As groundwater recharge is the sensitive component of the groundwater cycle, understanding processes governing recharge and making a reliable estimate is an important step towards evaluating the climate change effects on groundwater resources. While several attempts have been made to estimate the climate change impact on groundwater recharge, the uncertainties involved in the predictions have not been well addressed. This study aims to quantitatively evaluate the uncertainty in global groundwater recharge prediction using CMIP5 climate data. In this study an empirical model of diffuse pluvial groundwater recharge (0.5⁰x0.5⁰; annual time step upto 2100) is forced with bias corrected climatological outputs from Hadley Centre Global Environmental Model version 2 Earth Systems model - (HadGEM2 -ES); and the Max-Planck-Institute Earth System Model-Low Resolution model (MPI-ESM-LR). Three ensemble members with varying initial conditions of each of the GCMs were used to evaluate the uncertainty due to the internal variability of the GCMs. The results from this study shows that, the global groundwater recharge is expected to increase by 2080 under RCP 4.5 and 8.5, however the rate of increase is highly region specific. Across the globe ~53% of the area is having an increase in recharge, while 38% are expected to have decreased recharge. However, it is apparent from the results that future recharge estimates in most regions show significant variations with RCPs, GCMs and even between ensemble members of the same GCM. Even though, the prediction uncertainty was region specific, globally on an average it varied between 10 to 15 mm/y. The regions with highest percentage change in recharge (either increasing or decreasing) were the ones with highest variation in prediction. This study clearly shows that the prediction uncertainty within different ensemble members are comparable with the uncertainties between RCPs or between GCMs. Thus, neglection of the prediction variance between ensemble members seriously compromises our understanding of the future state of groundwater systems across the globe.