Spatial Autocorrelation Analysis of Land Surface Temperature Distribution Influenced by Built Areas in the North Central Province of Sri Lanka

Urbanization and population growth have led to increased built-up cover, impervious surfaces, and a decline in vegetated spaces. This has exacerbated the zoning of high-temperature clusters in populated areas. While satellite remote sensing techniques have demonstrated a positive empirical association between impervious surface cover and high-temperature zones, their spatial correlation has yet to be understood. This study assesses the spatial autocorrelation of impervious surface cover and high-temperature zoning in the North Central Province of Sri Lanka using a SAR-based Built-Up Index (BUI) coupled with Sentinel-2 optical imagery indices. Using normalized difference water index (NDWI) and normalized built-up area index (NDBI), land use was visually analyzed and classified. MODIS land surface temperature (LST) maps were employed to estimate surface temperatures. The classified images and LST maps were re-sampled and analyzed using the bivariate Moran’s I index to assess the spatial autocorrelation between BUI and LST values. The analysis reveals spatial anomalies in LST spread near the outskirts of urban areas, water bodies, and green patches, suggesting potential spatial propagation patterns. Zones with high spatial correlation between BUI and LST indicate potential future expansion areas of high-temperature clusters, emphasizing the need for targeted urban planning interventions to mitigate the urban heat islands.