Quantifying the influences of land surface parameters on LST variations based on GeoDetector model in Syr Darya Basin, Central Asia

Spatiotemporal variability in the land surface energy flux is largely and comprehensively affected by many factors, including land surface temperature (LST), land coverage, soil characteristics, terrain conditions, etc. In contrast with other climatic zones, arid and semiarid lands have fragile ecological environments that are more sensitive to land surface energy flux changes. In this study, we used MODerate Resolution Imaging Spectroradiometer (MODIS) LST products (2001-2015) for air temperature comparisons. Then, we investigated the spatiotemporal variation in LSTs in the Syr Darya Basin (SDB) during 2001-2015. More specifically, a new statistical model known as GeoDetector was adopted to analyze the driving factors controlling the spatiotemporal variation in LSTs. The result shows that the MODIS LST can provide a good estimation of air temperature, especially at night. The LST change rate can be considered as an important indicator of climate change (rapid warming at high altitudes) and human activities (increased water consumption of crop). Based on the GeoDetector model, we determined that the elevation explained more of the LST distribution (84-90%) and spatiotemporal variation (22-26%) than any other land surface parameters. The combination of albedo and the other explanatory variables can significantly increase the explanatory power of each single factor, especially with elevation.