Assessing urban landscape dynamics and its relations to changing surface thermal character and prospects: a...

Machine learning (ML) models are leading analytical techniques that provide insights into exploring urban landscape dynamics over time. The ML at present plays a crucial role in comprehending city growth and directing spatial processes of urban planning. The present study investigates the influences of LULC dynamics in the context of urban thermal variations based on modeling future land surface temperature (LST), particularly during summer and winter seasons, using advanced ML algorithms and earth observation satellite images (Landsat 8 and Landsat 5). The study shows that the built-up area in Durgapur Municipal Corporation (DMC) has increased by 39% during 1991-2021 by contracting agricultural land (32.20 km²) which is followed by green space (16.36 km²) and barren land (9.56 km²). The succeeding growth of seasonal LST (increase of about 0.23 °C and 0.17 °C in mean LST during summer and winter season, respectively) over DMC indicated that a warming trend of urbanization on the surface was reported during 1991-2021.The rapid urban expansion and landscape spatial heterogeneity in the study area significantly affect urban land surface temperature. The biophysical aspects are analyzed and explained in the study area to uncover the nature of urbanization in tropical areas. Due to high heat capacity, evaporative and rising latent heat flux of the study area, the urban water bodies and vegetation have a moderate effect on the urban heat island phenomenon. As a result, an increase in urban green space and avoidance of non-impermeable surface seems to be prospective strategies to minimize the adverse impacts of ULST. Therefore, further research findings would assist scholars, legislators, and urban planners in redirecting and further developing sustainable measures for mitigating detrimental effects of urban heat island.