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, which 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 the confidence in future projections used by stakeholders and policymakers. Traditionally, ETCs are studied using techniques such as case studies and manual synoptic typing. However, these approaches are time consuming, require subjective analysis and do not necessarily identify the coincident large-scale meteorological patterns (LSMP). Here, self-organizing maps (SOMs) are applied as an automated machine-learning approach to characterize the LSMP and associated impacts 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 -- an open-source Lagrangian feature detection package -- and are integrated with SOMs to classify storm properties associated with each pattern. Of note, we define simulated sleet, freezing rain, and snow by implementing an ensemble precipitation typing procedure typically applied to numerical weather prediction output. We composite these precipitation quantities (and relevant thermodynamic variables) for each SOM pattern to define synoptic setups critical to accurately simulating frozen extremes. We then apply the reanalysis-derived SOM as a reference in order to evaluate the skill of CMIP6 historical experiments in simulating the LSMP and ETC hazards over NEUS.