Seasonal Forecast Skill of Intensity and Frequency of Extratropical Winter Windstorms, their Relation to the NAO and Dynamical Forcing Factors

Abstract:

Skilful seasonal forecasts of extreme events are of highest interest for multiple actors in society, industry, and not least the general public. Extratropical cyclones and windstorms belong to the most damaging events in the mid-latitudes. Consequently, a useful prediction capacity of these severe events, including a clear understanding of factors contributing to predictability would be of great benefit.

The latest seasonal forecast model of the UK Met Office, GloSea5 GC2, is used for 23 winter season (DJF) and validated against ECMWF reanalysis ERA5. Severe, damage-prone windstorms are tracked with an impact-based algorithm using 10m wind speed (Leckebusch et al., 2008). The storm intensity is defined by an event-based objective measure per footprint, the Storm Severity Index (SSI), based on the cubic exceedance of wind speed over the local 98^th percentile.

This presentation focuses on two important storm measures, the frequency (further to Befort et al., 2019) and for the first time the event-based accumulated intensity per season. For storm frequency, the latest GloSea5 version reveals comparable skill to previous versions (Befort et al., 2019) confirming the principle positive trend in skill for predicting extreme events in highly relevant regions. GloSea5 GC2 is showing significant forecast skill for track densities at the end of the North Atlantic and Pacific storm tracks, respectively. The new investigation on event-based storm severity reveals also positive and significant forecast skill over similar regions of the Northern Hemisphere as for the storm frequency, downstream of both the Atlantic and Pacific storm tracks, thus affecting crucial economic regions in Western Central Europe.

A skilful prediction of the North Atlantic Oscillations (NAO) is also present in the latest forecasts. Further on, the forecast skill of the two storm measures is related to that of the NAO for the North Atlantic European region. This study also investigates the influence of dynamical factors (e.g. baroclinicity, surface heat flux, etc.) on storm predictions.