Post-processing climate projections of precipitation for the Po river basin: will Italy's North become water-constrained?

Surface and groundwater resource availability depends on precipitation patterns. Climatic change may alter not only future annual totals of precipitation but also its temporal distribution. In regions depending strongly on snow accumulation for steady water supply, this can lead to water constraints. We process climatic projections of precipitation from 19 models of the Climate Model Inter-comparison Project 5 for the Po river, Italy. The study area hosts Italy's most important lakes and reservoirs and is inhabited by 19 million people. The river basin is also known for its productive areas of irrigated agriculture. We apply a Bayesian processor of uncertainty, which we calibrate on a comprehensive set of high-resolution gridded observations. The processor outputs predictive densities of precipitation for selected prognostic time windows. These densities can be used in conjunction with an utility function to estimate potential losses and/or evaluate the benefits of mitigating actions. For the study area, annual precipitation will not change notably in the future for both an optimistic and a pessimistic scenario. The temporal distribution of precipitation will become affected. These potential changes result in considerable strain on storage capacity and water flows needed to satisfy irrigation demand as well as hydroelectric and thermal energy production.