Abstract: In this study,we conducted the downscaling analysis of 10-m wind fields at 6 stations in Da-Xiao-Xing'anling Mountains during spring in 2017.We also used the observations to evaluate the 10-m wind speed and direction simulated with the Weather Research and Forecasting (WRF) model and with the CALMET downscaling model.The correlation coefficients between observed hourly wind speeds and that simulated with the two models reach 0.5-0.7.The prediction accuracy gradually increases with the increase of wind speed.The forecasting deviation of wind speed is relatively large at night and decreases during the daytime.The WRF model predicts the variability of wind speeds better than the CALMET model,and that has a higher correlation with the observations;but for the case of strong winds,the CALMET model performs better than the WRF model.The wind direction simulations of WRF and CALMET models are both in a good agreement with the observations.Wind directions with high prediction accuracy correspond to the prevailing wind directions of each station.Meanwhile,a simple-regression method is used to correct daily mean wind speed.Results indicate that the forecasting accuracy of daily mean wind speed increases by 50% on average,with good prospects in the operational application.