A hybrid physical / statistical model for predicting the probability of very extreme rainfall

Water more valuable than oil? This may become true as water demand increases rapidly due to population growth. The urgent need to have accurate estimates of water availability was one of the motivations for Rachel White’s PhD. She worked with a widely used regional climate model to simulate rainfall and runoff, which control water availability. She focused on the case of the Olifants Basin in South Africa, an area under severe water stress. White noticed strong discrepancies between the runoff calculated by the model and the runoff observed. By implementing a new method to describe water penetration into the soil and its evaporation, she obtained results much closer to her observations. White's work might now be used for studies of water availability around the world, which are essential for enabling nations and NGOs to design water management solutions.
I am using the WRF regional climate model to study the relationship between rainfall and runoff at the river basin scale (4km resolution), in particular studying the seasonal change in the runoff coefficient (ratio of runoff to rainfall). We aim to model the probability of extreme rainfall and extreme runoff events under possible future climate change conditions. Extreme runoff events are likely to lead to flooding and we distinguish between surface runoff events, affecting the local area where rainfall occurs, and groundwater runoff extremes which can affect an area away from the rainfall event.