Ground truthing global-scale model estimates of groundwater recharge across Africa

Groundwater is an essential resource for natural and human systems throughout the world and the rates at which aquifers are recharged constrain sustainable levels of consumption. However, recharge estimates from global-scale models regularly disagree with each other and are rarely compared to ground-based estimates. We compare long-term mean annual recharge and recharge ratio (annual recharge/annual precipitation) estimates from eight global models with over 100 ground-based estimates in Africa. We find model estimates of annual recharge and recharge ratio disagree significantly across most of Africa. Furthermore, similarity to ground-based estimates between models also varies considerably and inconsistently throughout the different landscapes of Africa. Models typically showed both positive and negative biases in most landscapes, which made it challenging to pinpoint how recharge prediction by global-scale models can be improved. However, global-scale models which reflected stronger climatic controls on their recharge estimates compared more favourably to ground-based estimates. Given this significant uncertainty in recharge estimates from current global-scale models, we stress that groundwater recharge prediction across Africa, for both research investigations and operational management, should not rely upon estimates from a single model but instead consider the distribution of estimates from different models. Our work will be of particular interest to decision makers and researchers who consider using such recharge outputs to make groundwater governance decisions or investigate groundwater security especially under the potential impact of climate change.