An Eulerian evaluation of intense low-pressure systems over North America in CMIP6 and a regional climate model

Low-pressure systems, such as extratropical cyclones (ETCs), are a key component of mid-latitude climate variability and frequent drivers of extreme weather in North America. This study evaluates 27 CMIP6 global climate models (GCMs) and the Canadian Regional Climate Model (CRCM6), driven by ERA5 reanalysis or three CMIP6 GCMs, in reproducing surface pressure anomalies and near-surface humidity during intense cyclonic events over 1980-2014. We introduce a novel Eulerian approach that identifies the strongest low-pressure systems at each grid point and characterizes their local temporal evolution using a three-parameter decomposition: [Formula: see text] intensity, mean [Formula: see text], and [Formula: see text]. This decomposition effectively captures key regional differences and provides a new approach to assess model biases in simulating local dynamical and thermodynamical conditions associated with intense systems. Results indicate that GCMs share systematic biases in characterizing local environmental changes, though the origins vary regionally. Model resolution influences error magnitude but is less critical over continental areas. In high track density regions or areas affected by explosive cyclones, higher resolution reduces pressure anomaly errors, while near-surface humidity shows no clear resolution-related pattern. CRCM6 simulations exhibit the smallest errors domain-wide and perform consistently regardless of driving data. They show particular added value north of [Formula: see text]N and over oceans, especially where GCM errors are large, although some substantial errors persist in certain regions. This study underscores the challenges in simulating intense extratropical and tropical cyclones, highlights the importance of accurately representing synoptic and regional surface conditions, and provides a framework for evaluating model performance, informing improved dynamical downscaling and climate projections.

The online version contains supplementary material available at 10.1007/s00382-026-08260-7.