Improving the downscaled springtime temperature in Central Asia through assimilating meteorological and snow cover observations

Abstract

Central Asia is vulnerable to climate change due to its scarce water resources and fragile ecosystems. However, the limited number of meteorological observations in the region restrict the study of its climate, hydrology and ecology. In order to improve the downscaled springtime temperature in Central Asia, this study explored the impact of atmospheric and snow data assimilation on climate simulations in Central Asia based on the Weather Research and Forecast (WRF) model and the WRF Data Assimilation (DA) system. The results based on climate simulations in Central Asia during the spring of 2017 show that the WRF model surface temperature simulation has a significant cold bias in Central Asia due to underestimation of snow melt. By assimilating conventional meteorological observations, the cold bias in Central Asia was reduced. This improvement is the result of both the direct effect of the analysis increment, and feedback effects from snow and atmosphere. In addition to the assimilation of atmospheric data, snow melt in Central Asia was better simulated through the assimilation of Japan Aerospace Exploration Agency (JAXA) Satellite Monitoring for Environmental Studies (JASMES) snow cover data. This further reduced the cold bias of the springtime temperature in Central Asia. Compared with an experiment that only assimilated atmospheric observations, the experiment that assimilated both snow and atmospheric data reduced the increase in temperature from the analysis and simulated a warmer land surface. This resulted in more sensible heat flux from the surface to the atmosphere and stronger sublimation and evaporation and thus improved the simulation of soil moisture.