LAND SURFACE TEMPERATURE RETRIEVAL IN WETLANDS USING NORMALIZED DIFFERENCE VEGETATION INDEX-EMISSIVITY ESTIMATION AND ASTER EMISSIVITY PRODUCT

摘要

Land Surface Temperature (LST) has the potential to act as a continuous indicator of the ecological status of wetlands. Accurate emissivity values are required in order to calculate precise LST. We test two emissivity retrieval methods and their influence on LST calculated from a Landsat 7 image of a highly dynamic wetland in Southern Spain. LST calculated using NDVI (Normalized Difference Vegetation Index) threshold estimations and the ASTER emissivity product are compared. The results show differences of around 0-1 K for most land covers, and up to 3 K for areas of bare soil when Landsat and ASTER images have the same acquisition date. Tests using Landsat and ASTER images from different seasons do not show greater differences between both LSTs. This has important implications for automated LST retrieval methods, such as the one planed by the USGS using Landsat and ASTER emissivity products.