Urban heat island (UHI) has multiple negative impacts on cities from heat related morbidities to excessive energy demands by buildings. Hence, better understanding of the contributing factors on its formation ensues ene efficient mitigation of these adverse effects. Previous studies demonstrate a significant correlation between the vertical elevation of urban morphology and UHI, however, topological parameters are barely considered. This study aims to improve the surface UHI prediction by integrating the impact of land surface elevation using a novel parameter of terrain factor (TF), which is integrated with other morphological parameters to develop an artificial neural network (ANN) model to predict the spatial distribution of UHI indicated by land surface temperature (LST). Morphological parameters were derived from two case studies representing areas with high terrain variation and relative flat terrain in Illinois, USA. The developed model was utilized to predict the LST for parts of the city, not initially included in the training process. Integration of TF significantly improves the LST predictions for high terrain variation areas, as the average root mean square error decreased from 1.26 to 0.90(circle)C and R-2 increased from 0.74 to 0.81. In conclusion, TF has significant impact on the surface UHI in areas with a significant surface relief variation.

The impact of the urban heat island (UHI) on health of urban residents, energy consumption, and urban air quality has become increasingly significant in cities. Despite of multiple recent studies on developing UHI predictive models, the impact of vertical morphological parameters on the UHI formation is barely investigated. The current work investigates the extent of vertical morphological components impact on the UHI intensity and proposes a novel vertical urban morphological parameter defined as the elevation factor. Using remote-sensing satellite images and a GIS based approach, both horizontal and vertical morphological components of San-Francisco, as the case study city, was investigated for 20 representative built up areas. High-resolution geometrical data of these areas were then correlated with their land surface temperatures (LST). Linear regression analysis and ANOVA WA revealed that the model with the newly proposed elevation factor has a significant impact on representing of the UHI behaviour with a coefficient of determination factor (R-2) of 0.68. The model was then successfully validated with the UHI prediction of 24 randomly selected points from each of the five new areas. Furthermore, the impact of elevation morphology on the UHI intensity was investigated as the elevation factor was seen to provide a stronger correlation between urban morphology and UHI spatial distribution.