LAUSR

As the backland of the Qinghai-Tibet Plateau, the river source region is highly sensitive to changes in global climate. Air temperature estimation using remote sensing satellite provides a new way of conducting studies in the field of climate change study. A geographically weighted regression model was applied to estimate synchronic air temperature from 2001 to 2015 using Moderate-Resolution Imaging Spectroradiometry (MODIS) data. The results were R2 = 0.913 and RMSE = 2.47 °C, which confirmed the feasibility of the estimation. The spatial distribution and variation characteristics of the average annual and seasonal air temperature were analyzed. The findings are as follows: (1) the distribution of average annual air temperature has significant terrain characteristics. The reduction in average annual air temperature along the elevation of the region is 0.19 °C/km, whereas the reduction in the average annual air temperature along the latitude is 0.04 °C/degree. (2) The average annual air temperature increase in the region is 0.37 °C/decade. The average air temperature increase could be arranged in the following decreasing order: Yangtze River Basin > Mekong River Basin > Nujiang River Basin > Yarlung Zangbo River Basin > Yellow River Basin. The fastest, namely, Yangtze River Basin, is 0.47 °C/decade. (3) The average air temperature rise in spring, summer, and winter generally increases with higher altitude. The average annual air temperature in different types of lands following a decreasing order is as follows: wetland > construction land > bare land glacier > shrub grassland > arable land > forest land > water body and that of the fastest one, wetland, is 0.13 °C/year.