Scholarlink

<p>Urban heat exposure poses an escalating public health risk in rapidly urbanizing megacities. This study quantified the decadal linkage between impervious surface expansion and land surface temperature (LST) in Shanghai (2013–2023) using multi-temporal Landsat imagery on Google Earth Engine. Impervious surfaces were delineated via an NDVI–NDBI threshold and refined using machine-learning classifiers, while LST was retrieved through a radiative-transfer inversion and validated against MODIS datasets. A strong positive relationship was observed between impervious-surface density and LST across spatial and seasonal dimensions, with core urban areas approximately 4–5&#xa0;°C hotter than surrounding zones. Polynomial regression and geographically weighted regression revealed a nonlinear threshold at ~ 40% impervious cover, beyond which heat accumulation intensified disproportionately. Buffer-based gradient analysis further confirmed an outwardly increasing thermal risk. Incorporating ERA5 meteorological variables enhanced model explanatory power and enabled partial-correlation analyses isolating the effects of surface greenness and land cover. These findings provide quantitative evidence for threshold-driven urban heat mechanisms and a replicable workflow for thermal-exposure mapping and risk mitigation in high-density cities.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Threshold-driven modeling of urban heat exposure under impervious surface expansion: a decadal remote sensing assessment of Shanghai

  • 期刊论文

  • 发表日期：

    2026年1月3日

查看全文
  • Zujun Wang,
  • Xin Sun,
  • Huashan Wang

摘要

Urban heat exposure poses an escalating public health risk in rapidly urbanizing megacities. This study quantified the decadal linkage between impervious surface expansion and land surface temperature (LST) in Shanghai (2013–2023) using multi-temporal Landsat imagery on Google Earth Engine. Impervious surfaces were delineated via an NDVI–NDBI threshold and refined using machine-learning classifiers, while LST was retrieved through a radiative-transfer inversion and validated against MODIS datasets. A strong positive relationship was observed between impervious-surface density and LST across spatial and seasonal dimensions, with core urban areas approximately 4–5 °C hotter than surrounding zones. Polynomial regression and geographically weighted regression revealed a nonlinear threshold at ~ 40% impervious cover, beyond which heat accumulation intensified disproportionately. Buffer-based gradient analysis further confirmed an outwardly increasing thermal risk. Incorporating ERA5 meteorological variables enhanced model explanatory power and enabled partial-correlation analyses isolating the effects of surface greenness and land cover. These findings provide quantitative evidence for threshold-driven urban heat mechanisms and a replicable workflow for thermal-exposure mapping and risk mitigation in high-density cities.