Abstract
The Urban Heat Island (UHI) effect, a phenomenon in which urban regions exhibit higher temperatures than the surrounding rural areas, is creating growing environmental and public health challenges in the context of global climate change. This study aims to evaluate the potential of satellite remote sensing and machine learning techniques to quantify the intensity of the UHI effect across 20 urban-rural pairs in cities around the world. A Random Forest model was trained on variables derived from Landsat 8 imagery, such as NDVI, NDBI, and albedo along with urban morphology and topographic features to predict UHI intensity. The model achieved a mean absolute error of less than 0.15 degrees Celsius, accurately capturing thermal patterns across regions and revealing both expected and counterintuitive heat dynamics. The results validate the efficacy of remote sensing-based UHI assessment, particularly in data-scarce regions, and highlight the role of land cover, vegetation, and urban morphology in influencing the local microclimate. This approach offers a scalable methodology for urban heat monitoring and can be used to support climatically sustainable urban planning strategies.