Evaluating Northern Hemisphere Atmospheric Blocking in post-CMIP6 Earth System Models

Atmospheric blocking events are long-lasting quasi-stationary high-pressure systems that persist for several days to weeks. They have been strongly linked to temperature extremes, such as winter cold spells and summer heatwaves, which can have substantial socio-economic impacts across various regions. State-of-the-art global climate models simulations of atmospheric blocking still underrepresent blocking frequency and persistence. Biases in the mean state, together with biases in sea surface temperatures, coarse horizontal resolution, and the choice of blocking detection method, have been highlighted as key contributors to such discrepancies. Recent studies highlight that increasing horizontal resolution can improve atmospheric blocking representation, with improvements being region-, season- and detection method-dependent. In this study, we aim to evaluate atmospheric blocking representation in the most advanced existing Earth System Models across the northern hemisphere. Blocking events are detected using an Eulerian blocking index based on the reversal of the meridional gradient of daily geopotential height at 500 hPa. We analyze a range of post-CMIP6 simulations from the Optimal high resolution Earth System Models (OptimESM) project, and compare blocking activity against the ERA5 dataset for the 1979-2014 period. Preliminary results indicate that, compared to ERA5, the historical simulations generally underestimate Scandinavian blocking frequency throughout the year, with underestimation also occurring in Greenland for all seasons except boreal spring. For Central Atlantic, models tend to overestimate blocking during winter and spring, but underestimate it in summer and autumn. Blocking frequency in the North Pacific is overestimated in winter and spring and underestimated in summer, while showing the closest agreement with ERA5 in autumn. Furthermore, trends in atmospheric blocking frequency for the 1979-2014 period show substantial regional and seasonal variability when comparing ERA5 reanalysis with model simulations, while also emphasizing the ongoing disagreement between reanalysis and model simulations.