Exploring the connection between morphological characteristic of built-up areas and surface heat islands based on MSPA

Abstract

Spatial pattern of built-up areas can exert an important influence on surface urban heat islands (SUHIs). While previous studies relied heavily on landscape indicators to characterize land use spatial pattern, such quantitative indicators cannot correspond to distinct spatial positions. By comparison, morphological spatial pattern analysis is an appealing alternative for measuring land use pattern as it can spatially visualize various land use elements from the perspective of mathematical morphology. Therefore, this research moves a step forward in understanding the spatio-temporal connection between the morphological characteristic of built-up areas and SUHI intensity. Taking a largely-urbanized city as an example, we first analyzed the SUHI intensity and morphological characteristic of built-up areas according to multi-temporal remote sensing data. Second, we used the Pearson correlation test and random forest regression to explore the linear and non-linear connections between SUHI intensity and various influence factors, respectively. Compared with landscape pattern, land use morphological characteristic plays a more important role in urban thermal environments. Both the correlation test and random forest regression suggested that the proportion of “core” was a dominant influence factor of the SUHI. The SUHI intensity in the study area was positively connected to the proportions of core, edge, and bridge of built-up areas, but negatively connected to the proportion of islet. Our results can help urban planners emphasize the influence of morphological characteristic in an earlier phase of thermal environment design. This assessment workflow can also be reused for various climatic factors in other cities because of its flexibility.