Central Asia is a region with primarily arid and semiarid climate and highly heterogeneous landscape.Recently,various ecological models have been used to understand how local ecosystems have responded to and interacted with changes in surface temperature and precipitation and their tolerance to changes in climate and environmental condition.These models suffer from the limitation of available high resolution and quality climate data,however.To ease this limitation,the Weather Research and Forecasting(WRF)model is implemented in Central Asia in this study after repeated testing and evaluations to optimize the model.The validated WRF for the region is further used to downscale the ERA-Interim reanalysis data to a new regional dataset of 20km×20km resolution.For validation,the downscaled surface temperature and precipitation of seven years(1980-1986)are compared to in situ observations and further contrasted with the comparison of the ERA-Interim data and the observation.The duration is selected to take advantage of the abundant available station observations.Results show high performance of the WRF model in describing surface temperature and precipitation and substantial improvements of the downscaled data particularly in describing the variance of those variables,compared with the driving reanalysis data.The validated WRF,which can be used to downscale past as well as future climate,and the downscaled high-resolution and quality dataset present both a tool and a useful dataset for ecological and climate research in Central Asia.