Spatiotemporal Evolution and Driving Factors of County-Scale Land Surface Temperature on the Northern Slope of the Tianshan Mountains, China, 2005–2024

Against the background of global climate change, understanding land surface temperature (LST) variations and their driving mechanisms in arid oasis regions is essential for regional thermal environment management. This study investigated daytime and nighttime LST on the northern slope of the Tianshan Mountains during 2005–2024 using MODIS MOD11A2 and multi-source environmental data. The Theil–Sen estimator, Mann–Kendall test with Benjamini–Hochberg false discovery rate correction, spatial autocorrelation analysis, and GeoDetector were applied to examine temporal trends, spatial differentiation, and driving factors at the county scale. The results revealed a pronounced mountain–oasis–desert thermal gradient and clear day–night asymmetry, with oasis areas generally exhibiting daytime cooling and nighttime heat retention. At the regional scale, neither annual nor seasonal daytime or nighttime trends remained statistically significant after FDR correction. At the county scale, no annual trend was significant, while significant seasonal changes were mainly concentrated in autumn, with nighttime warming occurring in more counties than daytime warming. Spatial autocorrelation showed marked diurnal and seasonal differences, with significant global spatial dependence occurring only under selected annual and seasonal conditions, while significant local spatial associations were limited to a small number of counties and were dominated by High–High clusters. GeoDetector identified topography as the most stable control of LST spatial differentiation, while vegetation, hydrothermal conditions, and anthropogenic factors showed stronger seasonal dependence. Factor interactions generally enhanced explanatory power, and sensitivity analysis confirmed stable dominant-factor patterns under different discretization schemes.