Decadal circulation variability in storm track regions: a comparison of reanalysis and model data

The Northern Hemisphere jet streams exhibit variability on decadal to multidecadal timescales, potentially providing an important source of climate predictability. Compared to reanalysis data, however, global climate models underestimate this low-frequency variability.

While the origin of this missing variability remains unclear, a possible cause might be insufficient ocean-atmosphere coupling. Increasing horizontal model resolution to capture mesoscale frontal processes can enhance this coupling and, consequently, decadal atmospheric circulation variability associated with decadal sea-surface temperature variability.

To characterize decadal circulation variability, we analyze patterns of low-frequency variability of Northern Hemisphere wintertime sea-level pressure and 700 hPa zonal wind using low-frequency component analysis. By comparing ERA5 reanalysis to CESM1 simulations at both a standard 1° and a higher 0.25° atmospheric resolution, we aim to (1) compare model and reanalysis, and (2) investigate the gain of increasing model resolution.

We find that the ratio of low-frequency to total variance in the Northern Hemisphere storm track regions is considerably higher in the reanalysis than the model data for all leading modes of variability. Our analysis also reveals differences in the composition of modes of low-frequency variability between reanalysis and the two model resolutions.

This research advances understanding of how global climate models represent low-frequency circulation variability at different resolutions, which is crucial to improve decadal predictability and climate projections.