Assessing the impact of climate change on Combined Sewer Overflows based on small time step future rainfall timeseries and long-term continuous sewer network modelling.

The evolution of the climate in the future will probably lead to an increase in extreme rainfall events, particularly in the Mediterranean regions. This change in rainfall patterns will have impacts on combined sewer systems operation with a possible increase of spilled flows, leading to an increase of untreated water volumes released to the receiving water. Due to the impact of overflows on the water cycle, local authorities managing combined sewer systems are wondering about the extent of these changes and the possibility of taking it into account in stormwater management structure design. To do this, rainfall data with a fine time step are required to better master the shape of the hyetographs that are crucial to get a relevant rainfall/runoff relationship in an urban environment. However, there are currently no simulations of future rainfall series available at a time step compatible with the needs in urban drainage field. In this work, future rainfall time series with a fine time step are elaborated with the aim to be used in urban hydrology. The proposed approach is based on simulations results from five regional climate models in the framework of the Euro-Cordex program. It consists in a spatial downscaling step followed by a temporal disaggregation. The rainfall time series obtained are then used as input for a calibrated and validated hydrological model to investigate the evolution of annual CSO volumes and frequencies by 2100. The results show an increase of annual spilled volumes between 13% and 52% according to the considered climate model. This increase will most likely be a problem regarding compliance of sewer networks in line with the water framework directive, particularly the current French regulations. No clear trends were observed on the CSO frequencies. If there is a consensus for all the carried-out simulations to conclude that the CSO volumes will increase, we must remember that actual regional climate models suffer from limited spatial and temporal resolution and don't explicitly solve convection processes. Due to this point uncertainty concerning the evolution rate remains important particularly for intense rainfall episodes. New generations of climate models are needed to accurately predict intense episodes.