Understanding the relationship between extremes of wind and inland flooding in the UK

Windstorms and flooding pose a significant socio-economic threat to the United Kingdom andcan cause significant financial loss. For example, the great October storm of 1987 damaged whole elements of the national electricity grid in the west of the UK. Storms can also be associated with heavy precipitation, for example, extensive inland flooding was caused by a series of slow-moving storms in the case of the winter floods of 2013/14 in the South East of England. The UK Met Office and Environment Agency estimated the financial loss attributable to the 1987 and 2013/14 events at €6.4bn and €1.5bn respectively. The question of correlations between windstorm and flood events remains open, for example the risk of a 1987-scale event "colluding" with the economically adverse meteorology of the 2013/14 season being poorly unquantified. If wind and flood risk is correlated then insurers are under-estimating both capital requirements and risk policy price, exposing them to very substantial liabilities.

Here, a collaborative project between academics and insurers has been undertaken to improve our understanding of the spatial-temporal distribution of risk from extreme, compounded windstorm and inland flood events in the UK. Statistical analysis of different data sets (~40 years of winter ERA5 reanalysis daily maximum winds, as well as observational precipitation and river flow gauge data) reveals wind and inland flooding are modestly correlated across the UK. In addition, we find substantially more compound events than expected by chance, some of which can be linked to named UK storms.

In terms of the large-scale atmospheric drivers, there appears to be no particular preferred path for the storms associated with compound wind and flood events. However, we find that compound events appear to be moderated by the amount of rainfall in the days preceding a windstorm, rather than the overall storminess of any given year. Further, we investigate the relationship in very extreme (200-year return period) windstorms and precipitation from the 1000-years of high-resolution HiGEM climate simulations.