Power outages in windstorms: the influence of rainfall preconditioning, wind direction and season

Windstorms are the main cause of large power outages in the UK. Faults to electricity distribution networks during windstorms are predominantly a result of windthrow, the uprooting or breakage of trees by winds that then fall on assets such as overhead lines. The impact of strong winds on windthrow is influenced by a several conditions: trees uproot more easily in saturated soils, they are more vulnerable to strong winds from unusual directions, and they are more susceptible to strong winds in the growing season when their leaves catch the wind. Despite this, risk assessments of impacts, such as power outages, during windstorms generally focus on wind intensity alone. Here, we quantify the influence of contributing variables of windthrow including antecedent rainfall, wind direction of the maximum wind gust, and the season a windstorm occurs in. We demonstrate that including them in a logistic regression model alongside wind speed can improve the predictive skill of the number of electricity faults during windstorms compared to a reference model that only includes wind speed. The analysis uses fault data from the National Fault and Interruption Scheme (NaFIRs) database during the period 2006-2018 in four regions in the UK: South Wales, Southwest England, East Midlands, and West Midlands. Meteorological data is provided by ERA5. Each variable is shown to modulate the impact of strong winds and improve predictive skill, though with some regional variability. The probability of a high fault numbers in a windstorm with winds exceeding 25 m/s can be doubled following high rainfall accumulations and five times higher when strong winds come from a direction that deviates more than 40 degrees south or west from the prevailing south-westerly direction. Furthermore, this probability is doubled in summer months compared to winter. These results can help improve impact forecasting during windstorms and highlight the importance of including these variables in historical and future risk assessments of assets vulnerable to windthrow. Ignoring such contributions may lead to misrepresentation of risk and potential maladaptation, particularly for electricity distribution networks that will undergo a huge transformation as we reduce our dependence on greenhouse gases in the future.