The temporal dynamics of urban heat islands derived from thermal remote sensing data by local indicator of spatial association in Shenzhen, China

摘要

Urban heat island effect is a hotspot in the study of ecological effects caused by urbanization. Monitoring and modeling urban heat islands (UHIs) has been a key issue of urban climate studies. The method from climatology is not suitable for modeling urban heat islands in multiple centers cities. In this study, we used the thermal infrared data acquired from Landsat TM and ETM+ images in 1986, 1995, 2000 and 2005 in Shenzhen, a rapidly urbanizing area in China. The surface temperatures were retrieved by the Jimenez-Munoz and Sobrino's single-channel method. An approach based on the local indicator of spatial association (LISA) was provided to retrieve UHIs for representing the character of energy aggregation. The urban heat islands were different from the high temperature areas (HTAs) derived by standard derivation method. The characteristics were compared by the differences on magnitude and statistical characteristics of UHIs and HTAs. Temporal dynamics in the spatial patterns were described using the spatial Lorenz curve. Analysis suggested that the mean temperature of UHIs increased from 1986 to 2005, whereas the differences between UHIs and surrounding areas decreased. The spatial patterns of UHIs changed from aggregation to average distribution, then to aggregation.