An Analysis of Thermal Characteristics Associated Land Cover Using Landsat Satellite Data

摘要

Direct applications of remote sensing thermal infrared (TIR) data in landscape ecological research are rare due to limitations in the sensors, calibration, and difficulty in interpretatiom Currently there are lot of methodologies for examining the relationship between land surface temperatures (LST) derived from TIR data and land use/land cover extracted from remote sensing images. In this study, Landsat TM images of Pearl River Dleta, Guangdong province, acquired in 1990 and 2008, were used to get the 6 land use types of built-up land, arable land, water, fishery, and shrub. Next, pixel-based LST was retrieved for 1990 and 2008 using the thermal band of Landsat TM image. An examination of the relationship between the land use/land cover (LULC) and LST maps was finally conducted to comprehension tbeir interactions. Results indicated that Landsat image was effective for quantifying the heat environment and providing reliable measurements of land use change. LST was found to be positively correlated with impervious surface but negatively correlated with vegetated land. Each temperature zone was associated with a dominant LULC category.