Accuracy Evaluation of Two Remotely Sensed Land Surface Temperature Products in High Vegetation Areas

Land surface temperature (LST) plays an important role in the balance of water and energy between land surface and atmosphere. The accuracy of remotely sensed LST products can be checked by using ground measured temperature data. In this study, 0 cm ground surface temperature (GST) and 2 m air temperature (AT) data from China Meteorological Administration Land Data Assimilation System (CLDAS-V2.0) were used to verify the accuracies of AQUA/Moderate-resolution Imaging Spectroradiometer (AQUA/MODIS) and FengYun-3D/MEdium Resolution Spectral Imager (FY-3D/MERSI) LST products in high vegetation covered areas. The results showed that the higher the canopy height, the larger the difference between remotely sensed LST and 0 cm GST. The bias between remotely sensed LST and 0 cm GST changing from -7 K in evergreen broadleaf forest area to -3.7 K in urban and built-up area. For AQUA/MODIS, the unbiased Root Mean Square Error (ubRMSE) are 2.77-3.39 K compared with 2 m AT and 2.69-3.87 K compared with 0 cm GST. However, for FY-3D/MERSI LST, the ubRMSE range from 3.42-4.01 K compared with 2m AT and 3.39-4.15 K compared with 0 cm GST. The precision of AQUA/MODIS products is better than that of FY-3D/MERSI LST products in high vegetation covered areas.