Subseasonal Predictions of Polar Low Activity Using a Hybrid Statistical‐Dynamical Approach

The subseasonal prediction of polar low (PL) activity is explored using a hybrid statistical‐dynamical approach. A previously developed PL genesis potential index is paired with ECMWF reforecasts and forecasts to predict regional statistics of PL activity across the sub‐Arctic. Regional PL activity is skillfully predicted in all regions at forecast ranges of up to a month. Additionally, the predictability limit of this hybrid framework (estimated using reanalysis data) is found to be highest over the Nordic Seas, Irminger Sea, Labrador Sea, and Bering Sea. We find that climate modes can strongly influence subseasonal prediction skill and are a potential source of predictability. Overall, our results highlight a promising prospect for the subseasonal prediction of PL activity. Plain Language Summary Polar lows (PLs) are intense mesocyclones forming over high‐latitude oceanic regions that pose hazardous risks to coastal communities and marine and air operations. This motivates the skillful prediction of PLs, which remains a major challenge at lead times beyond several days. We show that skillful forecasts of PL activity over a forecast range of weeks to a month in advance can be achieved by using a statistical modeling approach in combination with numerical model forecasts. Prediction skill varies by region, and is shown to be sensitive to major modes of variability in the large‐scale atmospheric circulation. Key Points At weekly to monthly forecast ranges, regional polar low (PL) activity is skillfully predicted using a PL genesis potential index Major climate modes can strongly influence regional prediction skill and can be used to identify forecasts of opportunity