Numerical study of urban expansion and its influence on urban environment using Landsat TM/ETM+ images

Multi-temporal Landsat TM/ETM+ imagery of 1990, 2000, and 2005 were used in this study. And the vegetation abundances, percent impervious surface, normalized difference impervious surface index(NDISI) and brightness temperature were retrieved from each TM/ETM+ dataset. Then the urban expansion, urban heat island and the relationships between LST and other variables which relate to urban environment were investigated. Results indicated that the percentage of urban area in Guangzhou increased significantly, which grew from 18.35% in 1990, to 24.16% in 2000, and further to 33.29% in 2005. But the intensity of urban heat island was not always enlarged during 1990-2005. And the regression analyses showed that, at the pixel-scale, the relationships between LST and other two variables (vegetation abundances and percent impervious surface) were relatively more complicated and could not be described by using the linear regression model. However, LST had a strong positive and negative correlation with percent impervious area and vegetation abundance in the region-scale, respectively. These results may be very useful for moderate- or large-scale ecological modeling and climate modeling.