Surface Urban Heat Island in Shanghai, China: Examining the Relationship between Land Surface Temperature and Impervious Surface Fractions Derived from Landsat Etm+ Imagery

This paper investigates the relationship between the surface urban heat islands (SUHI) and the percent impervious surface area (%ISA) in Shanghai, China. The %ISA was characterized from a Landsat-7 ETM+ multispectral dataset using the Linear Mixture Spectral Analysis (LMSA). Several critical steps being taken to derive %ISA were discussed, including atmospheric and geometric correction, water feature masking, endmember selection through the maximum noise fraction transformation, spectral unmixing for endmember fractions, and accuracy assessment. The resultant %ISA was qualitatively evaluated by visually comparing its spatial patterns to the landuse pattern of Shanghai. The spatial variability of land surface temperature (LST) to %ISA was evaluated and compared to the variability between LST and NDVI, a conventional factor for SUHI prediction. The results indicated a strong and significant correlation between LST and %ISA for the spectral data involved in the study and virtually no relationship at all between LST and NDVI. The strong LST to %ISA relationship suggested that it was exactly %ISA that accounted for a large share of the urban heat island problem in Shanghai. * Corresponding author.