Ambient temperature modelling from surface characteristics and associating urban morphology with thermal discomfort

Urban heat island assessment is of paramount importance when monitoring microclimate changes, increased heat stress, mortality and energy consumption. Simply analysing land surface temperature patterns for human comfort and health assessment is often inadequate. In this study, we attempt to resolve this inadequacy with ambient temperature modelling from multiple surface characteristics for a tropical megacity. A 336 datapoint-based multilinear regression model was formulated to predict the maximum summer air temperature with a mean absolute error of 0.694°C. Specific locations of critical thermal environments were demarcated with Getis-Ord statistic. Based on our results, extreme hotspots covered 10.53 per cent of the city on a daily basis at 99 per cent confidence level. The application of an outdoor thermal comfort index highlighted the existence of very strong heat stress zones over 14.67 per cent of the study area. The air temperature hotspots were verified with intra-urban variation in carbon sequestration. This environmental parameter was used to strengthen the observed results since, carbon sequestration directly links urbanization with degradation of thermal environments. Further, five urban morphological parameters were analysed to conclude that building density and height were the most significant urban design factors leading to increased air temperature. Simultaneous local climate zone mapping depicted that heat islands were dominated by mid and high-rise built-up areas, of which 68.44 km² of that area is suitable for vertical and roof gardening. Another 33.29 km of road stretch was delineated within low-rise built-up areas with scope for green landscaping to improve urban sustainability.