Land surface emissivity retrieval from multiple vegetation indices: a comparative study over India

ABSTRACT

Land surface emissivity (LSE) is an important parameter used in the retrieval of land surface temperature (LST) from thermal infrared (TIR) sensors. One of the widely adopted methods for LSE estimation is to treat the LSE of a pixel as a linear combination of soil and vegetation emissivity and use Normalized Difference Vegetation Index (NDVI) for characterizing vegetation cover of the pixel. However, it is unclear if other existing vegetation indices (VI) can be used for estimating LSE using this method. The aim of this study is to compare multiple VIs: NDVI, Enhanced Vegetation Index (EVI), Normalized Difference Water Index (NDWI), Soil Adjusted Vegetation Index (SAVI) and Modified SAVI (MSAVI) for the retrieval of LSE over four sites in India with data from Landsat-7 and Landsat-8 satellites. The LSE estimated from different VIs were compared with the Advanced Spaceborne Thermal Emission Radiometer-Global Emissivity Dataset (ASTER GED) and MODIS 21 LST and emissivity product. It was found that EVI was marginally better performing than NDVI, consistently. However, EVI gave better results only under non- cropped conditions (NDVI < 0.25). Whereas, during cropped stages, NDWI performed better than other VIs for the retrieval of LSE. SAVI and MSAVI resulted in marginally poorer retrievals of LSE in comparison with other VIs and performed inconsistently.