Long-term trends in European extreme floods from 1950 to 2020

Hydrological extremes are non-stationary, displaying long-term trends and natural oscillations. These changes in extremes can be driven by multiple factors including climatic (climate variability, climate change) and socio-economic (land use changes, water management changes) factors. In this work, we analyse extreme river flows in Europe for the period 1950-2020. We aim to identify long-term trends in extreme floods and estimate the contribution of the two aforementioned factors in these trends. The assessment is performed with modelled streamflow data generated with the spatially distributed physically based model LISFLOOD. We force the model with bias-corrected and statistically downscaled climate data derived from the ERA5-Land climate reanalysis and the EMO-5 dataset. We also created a variant of the climate dataset with the global warming effect removed statistically from the data. Return periods of extreme flood events are estimated through a non-stationary extreme value analysis for each river point with an upstream area greater than 100 km². To disentangle the influence of the different factors driving changes in extreme flows, the hydrological model is run under various scenarios: (i) historical (historical climate and dynamic socio-economic) (ii) static society (historical climate and static socio-economic), (iii) counterfactual climate (historical climate without global warming and dynamic socio-economic). Available preliminary results enable presenting long-term space-time dynamics of European floods.