Extending the lead time of skilful seasonal predictions of European Winter Storms

European winter windstorms (EUWS) are the primary meteorological hazards in Europe leading to significant socioeconomic impacts. Recently, seasonal prediction systems have been shown to skilfully predict the seasonal frequency of EUWS with 1-month lead time. Here, we introduce an approach to extend skilful prediction of the seasonal activity of EUWS to a 2-month lead time by using ocean subsurface predictions from pre-November initialisations. This is done by identifying preconditioning signals of EUWS activity in the ocean subsurface heat content in September and October, and developing an ocean-based proxy forecast for EUWS activity. Comparing our approach with an ensemble sub-selection approach, known to be able to improve skill of long-range predictions, we find our approach outperforming the sub-selection approach, leading to skilful predictions from early September and October. We discuss a possible mechanism via reemerging heat anomalies in early winter influencing the lower-level baroclinicity with cold sea surface temperatures in the north-Atlantic. Our findings are especially relevant for practitioners (e.g., financial institutions) utilising seasonal forecasts to increase use of skilful early pre-season predictions.