A Global Climate-Driven Stochastic Drought Model for Risk Assessment 

Drought affects people, agriculture, and businesses across all sectors in every populated continent on Earth, and with climate change, both drought frequency and duration are increasing globally. Losses of $124 Billion to the global economy over the last two decades (1998 - 2017) have been directly attributed to drought. Hence, it is vital to gain an accurate understanding of drought risk in the present and how it may change in the future.  Here, we describe the development of a climate-driven drought model which provides a global view of drought risk for (re)insurers.  First, a historical catalogue (1950-present) consisting of yearly aggregated drought severity and duration footprints is derived by combining a selection of state-of-the-art drought indexes over varied time scales. Second, leveraging this historical catalogue, a large stochastic set of drought footprints is generated via the use of Principle Component Analysis, in which the drought risk is conditioned to the climate state. The model is then deployed on historical climate conditions (ERA5) or alternative and future climate conditions (indicated by the CEMS-LENS multi-member reanalysis model (present-2100)).  These products are critical to inform damage models in the (re)insurance sector, with the model thus far proving useful in predicting subsidence risk in a France-based use case. Showcased results will provide an evaluation of drought risk both in the historical and changing future climate, as well as a newly developed risk score metric based on merged severity and duration information.