 Despite numerous attempts to reduce systematic errors in forecasting near-surface air temperature sat at high latitudes, the problem remains significant. This study reveals a statistical pattern of warm sat biases under cold weather conditions and cold sat biases under warm weather conditions with the largest errors found in shallow atmospheric boundary layers, ABLs. The modeled excessive ABL thickness in northern Eurasia is attributed to the problem. The study used the operational model SLAV of the Russian hydro-meteorological center and evaluated two turbulent schemes against observations from 73 meteorological stations. The Toucan scheme did not improve the damped response of temperature and reduced agility of temperature fluctuations in two thick ABLs, and the sat errors in shallow ABLs persist in the new scheme. The study emphasizes the need for a closer look at the ABL thickness in NWP models. This article was offered by Igor Isor, Mikhail Tolstik, Rostislav Fadeev, and others. We are article.tv, links in the description below.