： Land surface temperature (LST) is an important parameter at the 23 land-atmosphere interface. The Collection 6 (C6) MODIS LST products are publicly 24 available. Three refinements were performed over bare soil surfaces in the C6 MODIS 25 LST products when compared with the Collection 5 (C5) MODIS LST products. To 26 facilitate the use of the LST products in a wide range of applications, it is necessary to 27 comprehensively evaluate the accuracies of the C6 MODIS LST products. In this 28 study, we validated the C6 MODIS LST products using the temperature-based method 29 2 over various land cover types, including grassland, cropland, cropland/natural 30 vegetation mosaic, Gobi, sandy dune, and desert steppe. In situ measurements were 31 collected from sites under different atmospheric and surface conditions, including six 32 SURFRAD sites in the United States, two KIT sites in Portugal and Namibia, and four 33 HiWATER sites in China. In general, the accuracies of the C6 MODIS LST products 34 at night are better than those during daytime. The daytime RMSE varies from 35 approximately 1.5 K to 5.6 K, whereas the night-time RMSE is less than 2 K at all 36 sites except for the HiWATER SSW site. Furthermore, the accuracies of the C6 37 MODIS LST products were compared with those of the C5 MODIS LST products 38 over bare soil surfaces. The C6 MODIS LST products are in excellent agreement with 39 the in situ LST measurements at the KIT Gobabeb site, with biases of 0.36 K during 40 the day and 0.24 K at night, and RMSEs of 1.5 K during daytime and 0.74 K during 41 night-time. However, there are no improvements in the accuracies of the C6 MODIS 42 LST products when compared with the C5 MODIS LST products due to further 43 overestimation of emissivities at the four HiWATER sites. 44 45 Key words: Land surface temperature, MODIS, temperature-based validation method, 46 split-window algorithm, in situ measurements. 47 48 1. Introduction 49 Land surface temperature (LST) is an important climate variable, which is related 50 to surface energy and water balance. It is also a key parameter for various studies 51 including hydrology, climatology, environment, and ecology (Anderson et al., 2008; 52
