Integrating a heatscape index and a Patch CA model to predict land surface temperature under multiple scenarios of landscape composition and configuration

Abstract

Effectively predicting and mitigating urban heat island (UHI) through regulation of urban landscape patterns remains a global challenge. This paper introduces a new framework for predicting land surface temperatures (LST). We introduce a heatscape index (SCSS) to establish a relationship between landscape patterns and LST within spatial grids. Furthermore, the framework employs an urban CA model, equipped with a sophisticated patch generation engine, to forecast urban landscapes under multiple scenarios. LST under these scenarios is then calculated based on the established regression relationship. Application of the framework in Wuhan yields valuable insights: (1) The regression coefficient (R2 = 0.85) between SCSS and LST indicates that changes in SCSS can well explain the spatial heterogeneity in LST, especially in areas with evident UHI effects. (2) The simulated and retrieved LST for 2015 shows high spatial consistency, achieving an R2 value of 0.81, indicating the model's credibility. (3) Scenario simulations demonstrate that rapid and compact urban land expansion would exacerbate the UHI effect, leading to extreme high-temperature areas expanding towards suburban regions and significant temperature increases in urban fringe areas. The proposed framework provides powerful tools for policymakers and urban planners to mitigate the UHI effect and enhance urban sustainability.