Abstract: The gray-scale values of band 6 of Shanghai Landsat ETM+ images acquired on 13 March and 3 July in 2001, were converted to pixel brightness temperature to represent the land surface temperature distribution by calculating model of ETM+ TIR brightness temperature. And the overall characteristics, variability, and related influencing factors of urban brightness temperature heat island was generalized. Then regression analyses between Normalized Difference Vegetation Index (NDVI) and brightness temperature on local scale in different landcover types revealed a significant inverse correlation between radiation temperature and vegetation level, and the cooling effect of vegetation on underlaying surface was stronger in summer than in spring. Furthermore, comparisons of the regression lines’ slopes implied that the vegetation’s influence upon surface temperature is conditioned by the underlaying surface structures, so while increasing same vegetation in well-vegetated and sparsely or non-vegetated areas during the summer daytime, the former’s extent of temperature declining is inferior to the latter’s. As suggested, vegetation planning must be adaptive to the local conditions of different land environment.

Key words: brightness temperature, NDVI, ETM+, urban heat island, brightness temperature, NDVI, ETM+