Investigation on snow-soil-atmosphere interaction and impact on the climate signals through a pacemaker experiment of SWE

Snow mass in the Northern Hemisphere has been recently changed due to acceleration of snow melting under global warming reported by IPCC AR5, suggesting that the negative variation of snow cover possibly heats atmospheric conditions near the ground surface. Recent studies mentioned that a modeling simulation using constrained snow physical properties would be needed to quantify the effect of snow condition changing on atmosphere (van den Hurk et al., 2016; Henderson et al; 2018). In this study, we investigated the effect of snow mass change on surface air temperature by atmospheric general circulation model using snow nudging technique. In the nudging technique, forecasted snow mass is nudged toward reanalyzed snow by a relaxation coefficient. We incorporated the nudging scheme into atmospheric general circulation model MIROC6 (Tatebe et al., 2019) and conducted two experiments with the and without the nudging (Hist-snow and Control, respectively). Hist-snow and Control are long-term simulations from 1901 to 2010. Supplied snow mass for nudging was obtained from land offline simulation (Reanalysis) using MIROC6 with reanalysis data GSWP3-FD (van den Hurk et al., 2016). The correlation coefficients between Hist-snow and Reanalysis for inter-annual changes of snow cover fraction and surface air temperature during spring season (MAM) in Asia region were 0.90 and 0.64, respectively. The correlations were increased than those for Control (0.46 and 0.51 for snow cover fraction and surface air temperature). This is probably due to that Hist-snow can properly reproduced land surface albedo in the model. Inter-annual changes in other regions and other seasons showed similar increase of the correlation coefficient by the nudging. In addition, the root mean square deviations between Hist-snow and Reanalysis for inter-annual changes of surface air temperature were generally reduced than those of Control from spring to winter over north hemisphere. For example, the RMSD during MAM in Asia region for Hist-snow and Control was 0.71 and 0.82 K, respectively. Although further study is necessary to quantify attribution of snowpack to atmospheric conditions, result suggests that high reproducibility of snow physical properties in climate model lead to more realistic pace of inter-annual trends in surface air temperature.