The Influence of Summer Deep Soil Temperature on Early Winter Snow Conditions in Eurasia in the NCEP CFSv2 Simulation

Abstract The National Centers for Environmental Prediction (NCEP) Climate Forecast System version 2 (CFSv2) has a large cold bias in the model's deep soil temperature during summer. This study explores the potential triggering effect of that bias on excessive Eurasian snow cover in early winter. Snow cover appears erroneously early in the fall, especially in western Eurasia, in long simulations with CFSv2. The seasonal transition may be too early because the model land surface temperature (LST) reaches its freezing point earlier than observed, so that new snow cannot melt. This process initiates snow‐albedo feedback too early. The early cooling of LST is partially influenced by a seasonal resurfacing of the cold bias in the deep soil layer. From winter to early spring, a cold bias prevails in LST and upper soil temperature as snow cover remains. During this season, the temperature in the deep soil is generally warmer than in the upper soil and has relatively little bias. From spring to summer, the cold bias in the upper soil becomes smaller as it warms up in response to solar heating. On the other hand, the deep soil temperature has a noticeably smaller seasonal change than observed, resulting in a severe cold bias during summer. As the solar radiation declines quickly in early fall, the cold deep soil temperature causes additional cooling in the upper soil layer and helps to bring LST to the freezing point early in the western Eurasia, which leads to enhanced bias in the snow properties.