Stochastic modelling of future extreme discharges as a tool for water resources engineering

In water resources engineering, extreme discharge events are often used for designing hydraulic structures. Although different approaches exist, extracting these events from long-term discharge records is recommended. Yet, given climate change, it is expected that extreme events will occur at a larger pace. To mitigate potential future disasters, one should assess the severity of future events and use the latter in design studies. In this project, a framework allowing for the generation of precipitation and evaporation time series will be developed that accounts for the climatic conditions as predicted by regional climate models. However, as the latter models show a bias when modelling the current climate, this framework should correct these deviations in both current and future climate. Based on the long time series of precipitation and evaporation modelled using the framework and by invoking a hydrological model, discharge records can be obtained from which extreme events can be selected. This then leads to a library of extreme discharge events with corresponding return periods that can be used for designing hydraulic structures based on expected future extremes. Through this approach, the advantage of working with very long time series to ensure a good representation of extreme events is combined with the computational need for designing with short time series that only include the relevant extreme events, making it interesting for operational water resources management.