High resolution ensemble flood forecasting

Abstract: The number and impacts of pluvial floods are likely to increase with the growth of our cities and as extreme weather is anticipated to intensify with climate change. Im- proved preparedness is needed which may be attained owing to recent development of high-resolution hydro-meteorological observations and forecasts as well as geographical data. This paper investigates the capacity of the HYPE model for rainfall-runoff mod- elling and ensemble forecasting at hourly resolution. The analysis includes evaluation and application of several new high-resolution data sources: radar-based precipitation (HIPRAD), urban land-use data (EEA Urban Atlas) and high-resolution ensemble forecasts (MEPS). These components are finally integrated in a forecasting prototype for a catchment in southern Sweden. The results showed that HYPE, forced with HIPRAD and with land-use from Urban Atlas, performed well with a long-term Nash- Sutcliffe Efficiency > 0.8 at hourly level. Analysis of selected pluvial-type high-flow events close to an urban area indicated a good representation of fast runoff. The application of MEPS forecasts has been demonstrated for a few single events with promising results. Overall it is concluded that the 1-hour forecasts provide added value compared with the 1-d step and that an increased resolution in time and space is important to accurately forecast pluvial-type events.