Polar lows (PLs) are high-latitude intense maritime mesocyclones that develop near the sea-ice margin or in proximity to snow-covered continents during marine cold air outbreak events. These weather systems, once coined Arctic Hurricanes, are associated with severe weather, such as gale-force winds, large-amplitude oceanic waves, and heavy snow showers or intense blizzard events. Consequently, PLs pose a threat to coastal and island communities, in addition to maritime and air operations in the Arctic. Moreover, PLs may affect deep-water formation in the North Atlantic through strong surface heat fluxes, and thus, are important for the large-scale oceanic circulation. Therefore, the impacts posed by PLs to humans and the broader environment demand a robust understanding of the environmental factors that promote PL formation, and in turn, skillful prediction and projection of PL activity. 

PLs develop rapidly in response to favorable conditions. The synoptic-scale atmospheric configuration determines whether conditions are favorable for PL formation, but since the observational network in the Arctic is sparse, the initial representation of synoptic scale conditions is a major challenge, and thus, even state-of-the-art numerical weather prediction models are sometimes unreliable for PL forecasting beyond daily time scales. Furthermore, due to the coarse grid resolution of climate models (CMs), CMs struggle to resolve the structure and dynamics of PLs and therefore, prediction on subseasonal and longer time scales is also a major challenge. CMs can, however, skillfully predict large-scale climate variables. We utilize this in an alternative approach to PL prediction by constructing an empirical relationship, in an interpretable statistical framework, between key large-scale climate variables and PL genesis frequency. This model not only expands our predictive capabilities to a broad range of time scales, but it also provides a valuable scientific understanding of PLs that illuminates previously undetermined relationships between large-scale environmental variables and PL activity.