Detecting cyclones with seasonal forecasts? Development of a novel standardised Windstorm Index for the forecasting and impact-oriented analysis of extreme wind events

This preliminary study introduces the Standardised Windstorm Index (SWI), a novel tool designed to quantify the impact of extreme wind events in different geographical regions. The SWI is developed by first fitting the Weibull distribution to daily maximum wind speed data, followed by an inverse normal transformation to obtain a standardised index. This method enhances the accuracy of extreme wind event detection compared to conventional standardisation techniques. Using seasonal forecasts from the SEAS5 system, the SWI demonstrates its ability to effectively detect tropical cyclones and windstorms in the Southern African Development Community (SADC) region, showing an improvement of more than 20% in the accuracy metric compared to raw standardised SEAS5 data. However, it is important to note that this improvement is primarily driven by better identification of non-events rather than an increase in cyclone detection sensitivity, as discussed in the main text. This study also acknowledges some limitations, including assumptions in the extreme event detection procedure, which may not fully capture the variability and uncertainty within seasonal forecasts.Moreover, the use of ERA5 for the bias correction of SEAS5 wind speed data may introduce inaccuracies in the input data used for calculating the SWI, due to the scarcity of observations assimilated in ERA5 within the SADC area. Future work will focus on refining these methods, extending the geographical and temporal scope to improve its robustness and applicability. Although preliminary, our results emphasise the potential of the SWI as a valuable tool for improving the predictive skills of seasonal forecasts and supporting proactive efforts for climate risk management and adaptation strategies.