Urban heat island dynamics in Rawalpindi: a 30-year remote sensing analysis and future projections

Urban Heat Islands (UHIs) pose a significant environmental challenge in rapidly urbanizing regions, influencing climate resilience, energy consumption, and public health. This study presents a comprehensive spatiotemporal analysis of UHI dynamics in Rawalpindi, Pakistan. Time-series analysis over 30 years (1990–2020) using Landsat satellite data processed through Google Earth Engine (GEE). Unlike previous studies, this research integrates a hybrid methodology. Random Forest classification for high-precision Land Use Land Cover (LULC) mapping correlates with the Statistical Mono-Window algorithm for accurate Land Surface Temperature (LST) estimation. The findings reveal a 22.2% increase in built-up areas. That impacts the past three decades, correlating with a 3.04 °C rise in LST, at an annual increment of 0.12 °C. This study applies the Cellular Automata-Artificial Neural Network (CA-ANN) model to project future UHI intensification in Rawalpindi. Results aim to forecast a 13.8% increase in built-up areas by 2030, which is expected to exacerbate thermal stress and energy demand. According to the results, the built-up area has grown highly erratic, mainly at the expense of vegetated land. A statistically significant correlation (r = 0.43) between LST and urban expansion underscores the urgency of adopting targeted mitigation strategies. Our research has some limitations; one of the important ones was that we could not assess multiple potential socio-economic drivers of urban growth because of a scarcity of spatial data. The results highlight the necessity for sustainable urban planning approaches, including integrating vegetation and heat-mitigating urban infrastructure to reduce thermal stress and enhance climate resilience. © 2025 Elsevier B.V., All rights reserved.