Multi-model assessment of sub-seasonal predictive skill for year-round Atlantic-European weather regimes

The prediction skill of sub-seasonal forecast models is evaluated for seven year-round weather regimes in the Atlantic-European region. Reforecasts based on models from three prediction centers are considered and verified against weather regimes obtained from ERA-Interim reanalysis. Results show that predicting weather regimes as a proxy for the large-scale circulation outperforms the prediction of raw geopotential height. Greenland blocking tends to have the longest year-round skill horizon for all three models, especially in winter. On the other hand, the skill is lowest for the European blocking regime for all three models, followed by the Scandinavian blocking regime. Furthermore, all models struggle to forecast flow situations that cannot be assigned to a weather regime (so-called no regime), in comparison with weather regimes. Related to this, variability in the occurrence of no regime, which is most frequent in the transition seasons, partly explains the predictability gap between transition seasons and winter and summer. We also show that models have difficulties in discriminating between related regimes. This can lead to misassignments in the predicted regime during flow situations in which related regimes manifest. Finally, we document the changes in skill between model versions, showing important improvements for the ECMWF and NCEP models. This study is the first multi-model assessment of year-round weather regimes in the Atlantic-European domain. It advances our understanding of the predictive skill for weather regimes, reveals strengths and weaknesses of each model, and thus increases our confidence in the forecasts and their usefulness for decision-making.

Automated summary

This study evaluates the prediction skill of sub-seasonal forecast models for seven year-round weather regimes in the Atlantic-European region. The models from three prediction centers are tested against weather regimes obtained from ERA-Interim reanalysis. Results show that predicting weather regimes outperforms raw geopotential height and Greenland blocking has the highest skill, while European and Scandinavian blocking have the lowest skill. The study also highlights the challenges in forecasting no regime situations and the improvements in model versions.