Impacts of Urbanization on Land Surface Temperature and Vegetation Cover in a Semi-Arid City: the Case of Maragheh

Abstract

In semi-arid regions such as Iran, climate change and rapid urbanization are transforming urban landscapes, intensifying heat stress, and posing challenges for sustainable planning. This study analyzes the spatiotemporal evolution of land use and land cover (LULC), vegetation loss, and land surface temperature (LST) in Maragheh city, Iran, between 1990 and 2023. Using multi-temporal Landsat imagery, Random Forest classification, and high-resolution heat mapping, the main patterns of urban expansion, vegetation loss, and thermal change are identified. The results show a significant increase in urbanized areas and a parallel decline in vegetation, leading to a marked increase in LST and the expansion of urban heat spots. These findings underscore the urgent need to integrate green infrastructure and climate-adaptive designs into urban development strategies in semi-arid cities. Future urban development strategies in semi-arid regions must integrate climate-adaptive design and long-term territorial monitoring using remote sensing tools.Graphical Abstract The workflow begins with multi-temporal Landsat imagery (1990-2023), which provided consistent datasets for monitoring spatiotemporal changes in land cover and thermal conditions. From these images, a set of spectral indices was derived, including the Normalized Difference Vegetation Index (NDVI) and Proportion of Vegetation (PV) to quantify vegetation health and cover, as well as the Normalized Difference Built-up Index (NDBI) and Urban Index (UI) to characterize the intensity of urban expansion. Subsequently, a Random Forest classification algorithm was applied to generate accurate Land Use/Land Cover (LULC) maps, enabling the identification of transformations across vegetation, agricultural, bare soil, and urban categories. In parallel, thermal information was retrieved through the Landsat thermal bands and enhanced via the TsHARP algorithm, resulting in high-resolution Land Surface Temperature (LST) maps. Based on these outputs, Urban Hotspots (UHS) were delineated using a statistical threshold approach (mu + 2 sigma), isolating areas with anomalously high surface temperatures. The graphical abstract illustrates the integration of these methods and highlights the principal findings: (i) urban cover in Maragheh increased by 85% between 1990 and 2023, primarily replacing agricultural and vegetated land; (ii) vegetation declined by 46%, with a parallel 3.93 degrees C rise in mean LST; and (iii) UHS expanded from only 4% of the study area in 1990 to more than 45% in 2023. The schematic underscores the strong correlations observed: vegetation-related indices (NDVI, PV) show a cooling effect, while urbanization indices (NDBI, UI) are positively associated with LST increase. These results emphasize the urgent need for sustainable urban planning and the incorporation of green infrastructure to mitigate urban overheating and preserve ecological resilience in semi-arid cities.