Assessing seasonal forecast skill of dynamical and physical factors and their influence on winter windstorm predictions

Seasonal forecasts gained additional scientific interest over recent years due to the potential to forewarn of extreme events. Previous studies have investigated significant seasonal forecast skill of winter windstorms over Europe. A first analysis showed that the three dominant large-scale modes (NAO, SCA & EA) over Europe play a big role in this predictability and can be used to skilfully forecast the winter windstorm season. But a regression verification analysis with an ANOVA also revealed that the three large-scale modes only explain up to 80% of windstorm frequency and 60% of windstorm intensity variations. Hence, they cannot explain all of the interannual winter windstorm variability. It is known that other atmospheric parameters affect winter windstorms. But are they skilful predicted in the forecast model and how much are they affecting the seasonal forecast skill of windstorm frequency and intensity? This study investigates dynamical and physical factors like the PV, Eady Growth Rate or Rossby Wave measures. Firstly, the seasonal forecast skill of these parameters is investigated within the UK Met Office, GloSea5, for 23 winters between 1993 and 2016. Monthly and seasonal skill analysis of storm relevant parameters show high forecast skill, especially over the southern North Atlantic. The second part of this study utilizes these factors to statistically predict the winter windstorm seasons with a multi-linear regression and we assess which factors have the greatest influence on windstorm predictions.