Performance of Dynamic Downscaling for Extreme Weather Event in Eastern Mongolia: Case Study of Severe Windstorm on 26 May 2008

Abstract

Global weather forecasts do not have sufficient performance to predict the local severe weather events that are accompanied with cyclones and cold fronts due to their coarse horizontal resolution. This study investigated the performance of dynamical downscaling (DD) using mesoscale model to simulate the severe windstorm in eastern Mongolia which occurred on 26-27 May 2008.
Our results revealed that the DD experiments were successful in capturing the general features of the windstorm in terms of wind and temperature patterns. The timing and amplitude of drastic changes in the simulated temperature and wind speed were very similar to that observed than that obtained from the global atmospheric data, suggesting that DD is capable of predicting extreme wind storm events in Mongolia. Analyses on the nested domains indicate that the DD has crucial impact on the performance for simulating severe storm even with a moderate resolution (27 km), and further nesting (9 and 3 km) plays a role to improve it. Furthermore, the maximum wind speed approaches the observed value more closely as the horizontal resolution increases, although it still underestimates the observed wind speed even in the 3 km mesh domain. On the other hand, the abrupt temperature change is captured well even in the low-resolution domain, suggesting a difference in necessary horizontal resolution for temperature change and maximum wind speed.