Heat exposure has become a global threat to human health and life with increasing temperatures and frequency of extreme heat events. Considering risk as a function of both heat vulnerability and hazard intensity, this study examines whether poor urban dwellers residing in slums are exposed to higher temperature, adding to their vulnerable demographic and health conditions. Instead of being restricted by sampling size of pixels or other land surface zones, this study follows the intrinsic latent patterns of the heat phenomenon to examine the association between small clusters of slums and heat patterns. Remotely sensed land surface temperature (LST) datasets of moderate resolution are employed to derive the morphological features of the temperature patterns in the city of Ahmedabad, India at the local scale. The optimal representations of temperature pattern morphology are learnt automatically from temporally adjacent images without manually choosing model hyper-parameters. The morphological features are then evaluated to identify the local scale temperature pattern at slum locations. Results show that in particular locations with slums are exposed to a locally high temperature. More specifically, larger slums tend to be exposed to a more intense locally high temperature compared to smaller slums. Due to the small size of slums in Ahmedabad, it is hard to conclude whether slums are impacting the locally high temperature, or slums are more likely to be located in poorly built places already with a locally high temperature. This study complements the missing dimension of hazard investigation to heat-related risk analysis of slums. The study developed a workflow of exploring the temperature patterns at the local scale and examination of heat exposure of slums. It extends the conventional city scale urban temperature analysis into local scales and introduces morphological measurements as new parameters to quantify temperature patterns at a more detailed level.
Keywords: Ahmedabad; India; LST; Local scale; MODIS; Morphological patterns; Slums.
Copyright © 2018 Elsevier B.V. All rights reserved.