Future flood and flood-on-drought peaks and volumes in the Northern Apennines through bias-correction and rainfall-runoff transformation of hourly climate forcings

Study region: Panaro basin, Northern Apennines, Italy. Study focus: Climate change threatens the resilience of flood protection systems, demanding accurate prediction of flood peaks and volumes and their expected future changes. Particularly, accurate prediction of flood volumes is of paramount importance for operational purposes, such as flood hazard mapping or detention basins. Furthermore, abrupt shifts from extended dry periods to floods pose significant challenges to regional early warning systems. Predicting future flood events is particularly challenging over small and mid-sized catchments, characterised by response times of a few hours, where both long time-series of past hydrometeorological observations and future projections at fine time scale are needed. The work proposes a methodology for the reliable simulation of past and future flood occurrence and magnitude, in terms of flood peaks and volumes. It relies on continuous streamflow simulation through i) hourly bias-correction of regional climate projections, ii) rainfall-runoff modelling and iii) a peak-over-threshold approach for flood characterisation. Moreover, a novel combination of metrics for analysing the evolution of compound flood-on-drought events is tested. New hydrological insights: The hourly bias-correction approach effectively adjusts the ensemble of regional climate models, allowing accurate reproduction of the flood regime. Simulations project a decrease in the frequency of flood events, but higher flood peaks and volumes are expected. In contrast, future compound drought and flood events are projected to become more frequent.