Examining the sensitivity of ECMWF IFS weather forecasts to sea‐ice coupling for the summer‐time Arctic and cyclones

Abstract In recent years there has been an advance towards coupled Earth‐system models for weather forecasting. For example, the European Centre for Medium‐Range Weather Forecasts now implements ocean–ice coupling with dynamic sea ice in the Integrated Forecasting System (IFS) at all time ranges. This has the potential to improve weather forecasts in the polar regions, where sea ice influences the overlying atmosphere directly by turbulent exchange, especially in the rapidly warming summer‐time Arctic where thinner and more mobile ice is susceptible to rapid change. In this study, we investigate the sensitivity of IFS (cycle 47r1) weather forecasts to sea‐ice coupling representation in the summer‐time Arctic by comparing three sets of forecasts that are coupled with (i) dynamic sea ice in operational configuration, (ii) static sea ice, and (iii) dynamic sea ice with additional thermodynamic (surface temperature and albedo) coupling. It is found that dynamic sea ice improves predictions of sea ice and the ice edge compared with persistence, especially in the vicinity of Arctic cyclones. The dynamic sea‐ice forecasts exhibit lower near‐surface temperatures (up to 0.5 °C) compared with static sea‐ice forecasts where ice loss has occurred, and differences in near‐surface winds of up to 0.5 , consistent with changing surface roughness over the marginal ice zone. The forecasts with additional thermodynamic coupling have near‐surface temperatures that are up to 1 °C cooler over ice than the operational configuration (correcting a known warm bias), consistent with a more stable boundary layer (BL) and weaker near‐surface winds. The influence of sea‐ice coupling above the BL is small, with differences in cyclone forecasts being smaller than the spread of the operational ensemble. This study highlights the influence of ocean–ice coupling in weather forecasts for the summer‐time Arctic, and the potential gains from improving its representation.