City block-based assessment of land cover components’ impacts on the urban thermal environment

Abstract

Rapid urbanization has posed many problems for the urban thermal environment, which has spurred many scientists to focus their studies on the effects of landscape compositions and configurations on land surface temperature (LST) to offer some support to policy-making decisions to mitigate the phenomenon of the urban heat island (UHI) effect. Many prior studies have developed the relationship between land cover types and LST using pixel-by-pixel or grid-by-pixel regression analysis; from the viewpoint of practical usage in urban planning, this study estimates the impact of land cover compositions on LST by taking a city block-based regression analysis approach. Based on Landsat 8 TIRS images over a continuous span of 4 years, the proposed approach is compared to the conventional grid-based method and discussed in terms of its feasibility. We evaluate the relative contributions of different land cover components to the estimated relationship. First, the phenomenon of UHI is found to be an event affecting area rather than a pixel or point location. Second, as more images are employed, a more meaningful relationship between the land cover components and LST can be achieved based on statistical analysis. Third, the estimated effect of land cover composition on LST with different relative weights can be more helpful to allow local governments or other metropolitan areas with similar urban size to make policy decisions in urban management and mitigate the influence of the UHI effect on the human population.