Impacts of Urban Expansion on the Land Surface Thermal Environment Based on GeoAI: A Case Study of Changchun

Urban expansion converts natural surfaces into impervious land and markedly alters the land surface thermal environment. Taking Changchun (including Gongzhuling) as a case study, this research retrieves summer land surface temperature (LST) and computes NDVI, NDBI and MNDWI from Landsat 8/9 imagery on the Google Earth Engine platform, integrating GLC_FCS30D land cover together with elevation, slope and night-time lights to analyze the coupling between urban expansion and the surface urban heat island (SUHI) from 2013 to 2023. The results show that the impervious (built-up) area of Changchun increased from 1941 km² in 2013 to 2029 km² in 2022, a net growth of about 88 km² (roughly 4.5%). Summer LST over built-up land is markedly higher than over vegetation, averaging about 36–38 °C versus about 31 °C and yielding an SUHI intensity of roughly 5–7 °C; this indicates that urban expansion, by replacing cooler vegetated and cultivated surfaces with strongly warming impervious land, is the main mechanism of local surface warming. A Random Forest model is used to simulate the spatial distribution of LST (test-set R² = 0.50, RMSE = 1.9 °C), and variable importance identifies the normalized difference built-up index (NDBI, i.e. built-up land) as the dominant driver of surface warming. Urban heat-risk zones derived from the predicted LST surface show that high-risk areas are mainly concentrated in the central urban districts and the built-up areas of the surrounding county seats. The findings provide remote-sensing monitoring and intelligent simulation support for urban-expansion management, heat-island mitigation and ecological resilience enhancement in cold-region cities.