The occurrence of cold spells over different regions of North America has been linked to windy extremes over western Europe. These events – termed pan-Atlantic extremes – are mediated by an anomalous state of the North Atlantic storm track. While it is known that the occurrence of European windstorms is modulated by the state of the storm track, the relative contribution of the North American cold spells to European wind extremes is not easy to quantify. In this study, cold spells over two regions of North America are clustered with respect to the evolution of the large-scale circulation over the North Atlantic. The contribution of cold spells to the European wind extremes is then ascertained using circulation analogs, so that different states of the North Atlantic storm track can be compared for days with and without cold spells. Consistent with previous work, two main pathways emerge from the analysis, called “zonal” and “wavy” for simplicity. For a wavy pathway, North American cold spell occurrence is associated with more frequent European wind extremes than expected from the state of the North Atlantic storm track. For the other pathways, on the other hand, the anomalous state of the storm track was able to account alone for the more frequent wind extremes than climatology observed across Europe, with no or little ascertainable contribution from the cold spells. This analysis clarifies the causality of wintertime pan-Atlantic extremes and how these link to different atmospheric dynamical pathways.

To better understand the impacts of climate change on Europe, it is important to understand changes in the wintertime large-scale circulation. The framework of weather regimes provides a powerful tool for studying the highly nonlinear Euro-Atlantic circulation, but exactly how these regimes will be altered by anthropogenic climate change is still imperfectly understood. Using the recently developed approach of geopotential-jet regimes, applied to an ensemble of state-of-the-art CMIP6 models, we show that the centres of action of anticyclonic regimes are not projected to change substantially by the end of century, even under an extreme warming scenario. Instead, the regimes are expected to become less persistent, making long-lived blocking events less likely. We show that these two key elements of the regime response can be captured in a simple Lorenz-like model subjected to parameter variations, emphasising the conceptual link between observed atmospheric regimes and the regimes identified in basic mathematical systems.