Predictability assessment of cold–wet–windy pan-Atlantic extremes

The occurrence of cold spells in different North American regions has been related to an anomalous state of the North Atlantic storm track, resulting in concurrent wet and windy extremes in Western Europe. Two dynamical pathways have been identified as potential origins of these pan-Atlantic compound extremes. In the first "wavy" pathway, the extremes are triggered by a Rossby wave train propagating from the Pacific towards the Atlantic, associated with a pronounced Alaskan ridge. The second "zonal" pathway is characterized by a high to the west of Greenland that favours the southward displacement of a trough over the eastern USA and an upper-level trough over southwestern Europe. The aim of this study is to assess the predictability of these two extreme pathways, using a combination of ERA5 data and ECMWF ensemble reforecasts. We first compute and analyse dynamical systems indicators of local dimension and persistence, which act as proxies for the intrinsic predictability of atmospheric configurations. We next analyse the reforecast data at different lead times and compute skill scores for the two pathways. We find that the zonal pathway is more predictable and persistent than the wavy pathway compared to their respective flow analogs -- namely other similar atmospheric configurations. However, both pathways are less predictable than the average of all extended winter days. This underscores the considerable challenge posed by the forecast of pan-Atlantic compound extremes.