Bridging Information From Paleo-Hydrological and Climate Model Ensembles to Assess Long Term Hydrological Drought Hazard

Characterizing the evolution of drought frequency and severity under anthropogenic global warming remains a key challenge because of the mismatch between the length of instrumental records and the long-term variability of drought features. To address this gap, we propose a modeling framework that combines river flow observations, paleo-hydrological reconstructions, and climate model simulations. Such diversity of climate information, that is bridged in a flexible approach, allows evaluating the hazard of hydrological droughts for any large catchment globally. By focusing on the specific case of Alpine regions and analyzing the information contained in an ensemble for the period 1100-2100, we show that, compared to the past nine centuries, the mean annual flow in the Po River (Italy's main water course) may decrease by about 10% during the 21st century, while the mean drought duration and severity are likely to increase by approximately 11% and 12%, respectively. Future drought conditions are likely to match, or even exceed, the driest period of the Medieval Climate Anomaly under different emissions scenarios. This indicates unprecedented drought conditions in Alpine regions in the coming decades, thus calling for an increased preparedness in managing water resources under climate change.