Abstract/Summary

While floods are among the most damaging natural hazards, they have helped shape human developments over the last millennia, fostering scientific understanding and technological advances in an attempt for their mitigation. We now more skilfully predict floods at increasing lead times, through probabilistic hydro-meteorological forecasting. But we are now facing new challenges. Have we reached the limits of predictability with seasonal streamflow forecasting? This thesis contributed to the implementation and design of operational seasonal streamflow outlooks, as part of the European and the Global Flood Awareness Systems. Openly available, they give users an overview of potential streamflow changes on sub-seasonal to seasonal timescales. The analysis of both systems highlighted current limits in seasonal predictability and the importance of initial hydrological conditions and the land surface memory. To tackle these limits of predictability, a sensitivity analysis was developed to guide developments for tangible future seasonal streamflow forecast improvements. Are technical and scientific advances increasing faster than the rate at which forecasts are usable for decision-making? As shown by the application of serious games and research interviews at the Environment Agency (to guide a successful transition to probabilistic forecasts for flood early warning in England), science (e.g. forecast skill) is not necessarily a limiting factor for decision-making. Improved communication between scientists and decision-makers, aimed for instance at understanding the complex landscape in which decision-makers operate, is key to a successful adoption of the latest science in practice. Art can help bridge the communication gap, and this thesis culminated in an art exhibition, ‘Gambling with floods?’, at The Museum of English Rural Life (Reading, UK) from 1 to 15 November 2019, created to reach a wide audience. Overall, this thesis has shown that a closer interaction between decision-makers, scientists and artists is urgently needed for a co-leadership on improving science for society.