Connecting Large‐Scale Meteorological Patterns to Extratropical Cyclones in CMIP6 Climate Models Using Self‐Organizing Maps.

Extratropical cyclones (ETCs) are responsible for the majority of cool‐season extreme events in the northeastern United States (NEUS), often leading to high‐impact weather conditions that can have wide‐ranging socioeconomic impacts. Evaluating the ability of climate models to adequately simulate ETC dynamics is essential for improving model performance and increasing confidence in future projections used by stakeholders and policymakers. ETCs are traditionally studied using techniques such as case studies and synoptic typing, however, these approaches can be time‐consuming, require subjective analysis, and do not necessarily identify the coincident large‐scale meteorological patterns (LSMPs). Here, we apply self‐organizing maps (SOMs) as an automated machine‐learning approach to characterize the LSMPs and associated frequency and intensity of discrete ETC events over NEUS. The dominant patterns of geopotential height variability are identified through SOM analysis of five reanalysis products during the last four decades. ETC events are tracked using TempestExtremes and are integrated with SOMs to classify the accumulated cyclone activity (ACA) associated with each pattern. We then evaluate the skill of CMIP6 historical experiments in simulating the LSMPs and ETC events identified in the SOM. Our results identify a robust bias toward more zonal patterns, with models struggling to reproduce the more amplified patterns typically associated with the highest cyclone activity. While model resolution has some impact on simulation credibility, model configuration appears to be more important in LSMP representation. The vast majority of CMIP6 models produce too few ETCs, although model errors are distributed around historical reanalyses when ACA is normalized by storm frequency. Plain Language Summary: Winter storms can have devastating socioeconomic impacts across the United States. Climate models are used to understand and predict these winter storms and how they might change in the future, therefore helping to inform climate policy and emergency services. It is essential that we evaluate these climate models so that their performance and accuracy can be optimized. One way of evaluating climate models is to assess their ability in reproducing the large‐scale atmospheric conditions which occur during extreme events. Here we apply a cyclone tracking algorithm and self‐organizing maps, a machine‐learning approach, to automate the process of identifying key patterns associated with historical winter storm activity. We then use this approach to assess if CMIP6 climate models are able to reproduce the same patterns including the winter storm frequency and intensity. Our results show that, regardless of resolution, most CMIP6 models struggle to simulate the more extreme patterns and tend to favor weaker patterns that are not as conducive to cyclone formation. As a result, most models simulate too few winter storms, however, the error in storm intensity is more varied, with some models producing stronger storms and others producing weaker storms on average. Key Points: Despite resolution, CMIP6 models struggle to reproduce amplified synoptic patterns over the US, with a robust bias toward zonal patternsMost CMIP6 models produce too few cyclones over the northeastern US, while model errors are more varied when considering cyclone intensityIncreased resolution increases the number and intensity of simulated extratropical cyclones but does not improve model representation of synoptic patterns [ABSTRACT FROM AUTHOR]