PNNL

Abstract

Accurately representing synoptic near-surface temperature variability is crucial to predict weather extremes, yet models remain biased. Existing studies primarily attribute wintertime midlatitude near-surface temperature variability to tropospheric large-scale advection, assuming minimal land influence. However, we find that nudging model’s circulation toward observations yields little improvement in wintertime temperature variance over Northern Hemisphere land, whereas land–atmosphere interactions are more important. We introduce a new scaling framework for temperature variance that incorporates local land-atmosphere feedbacks. Comparing our framework to the mixing length approach—which links temperature variance to the meridional temperature gradient and air parcel displacement (mixing length)—shows that land–atmosphere feedbacks are inherently embedded in the mixing length, a connection previously overlooked. Roles of land-atmosphere feedbacks are evaluated via model experiments with perturbed circulation, land, or both. We find that longwave radiative damping dominates temperature variance responses over meridional temperature gradient when both land and circulation are perturbed.