Weather Patterns and Antecedent Conditions Driving Extreme Floods in UK Benchmark Catchments

Abstract: Extreme fluvial floods pose severe socioeconomic and environmental risks across the UK. This paper addresses the critical need to identify the most influential features driving extreme flood events, including atmospheric circulation patterns, and land-surface antecedent conditions, through the integration of datasets from ERA5-Land, CAMELS-GB, and the Met Office Weather Patterns (MO30). Understanding the interplay between atmospheric circulation patterns and antecedent conditions as drivers of flood extremes remains a significant research gap. This paper addresses this gap through employing machine learning techniques (random forest models) to assess the relative importance of daily synoptic scale weather patterns, large scale weather regimes and antecedent land-surface conditions as predictor variables for the target variable of extreme flood magnitudes within the UK's most 'natural' catchments (UKBN2). Findings reveal cyclonic types with deep lows, very windy types, the North Atlantic Oscillation positive phase (NAO+) and southwesterlies as key drivers of the top 1% flood magnitudes. Our analysis also reveals further regional and seasonal variations in the dominance of these drivers. These insights highlight the necessity for further investigation on how driver relationships with extreme floods vary spatially, temporally, and under future climate changes. 
 

Biography: Hello, my name is Emma, and I am a DPhil researcher in Atmospheric Physics at the University of Oxford. I am interested in the atmospheric and land-surface drivers of extreme fluvial catchment floods in the UK. My research takes an integrated, interdisciplinary approach in understanding the relative importance of flood drivers and how this varies across time and space. I primarily use machine learning methods to answer my research questions. Please feel free to reach out to me via email emma.ford@hertford.ox.ac.uk. I look forward to meeting you!