Estimation of near surface air temperature from MODIS data in the Yangtze River Delta

Near surface air temperature is an important meteorological parameter and is closely related to agriculture production. Comparing with the traditional meteorological observation, satellite remote sensing provides a straightforward and consistent way to obtain air temperature over regional and global scales with more spatially detailed information. In this paper, the temperature-vegetation index method was used to retrieve the air temperature throughout 2005 in the Yangtze River Delta by MODIS land surface temperature and NDVI data. The retrieved air temperatures were validated by the meteorological observed air temperatures. The estimating error was about 2.39℃ with the normal temperature-vegetation index method, but only 72.23% of the samples could be used by this method. After some additional rules were made to broaden the applied range of temperature-vegetation index method, the percentage of valid estimates increased to 80.15% and the estimating error slightly rose to 2.44℃. Further more, the characters and variations of retrieval error were also analyzed. Results show that the retrieval accuracy of temperature-vegetation index method is significantly influenced by the vegetation coverage and land surface heterogeneity in temperature-vegetation index context windows. Estimation errors in high vegetation covered areas are obviously lower than in other areas. The improved temperature-vegetation index method shows good applicability and accuracy in cropland areas during crop growing seasons, which can provide a new approach for acquisition of air temperature of cropland in large-scale.