Multi-scale Relationships Between Impervious Surface, Vegetation, Water And Urban Heat Island

In this Thesis, Landsat ETM+imageries were used to detect the spatial pattern of urban heat island (UHI) and impervious surface (IS) in shanghai, to explore the relationship between IS, vegetation coverage, water surface and UHI at multi-scale support with remote sensing and GIS. Some suggestions of how to mitigate UHI were proposed. The findings are as followed:(1) UHI in Shanghai mainly distributes in urban central area and industrial districts in summer and winter. The area of UHI accounts for 51% in summer and 45% in winter. Large water body, for example, the Huangpu River is heat island in winter but cold island in summer. Farmland and urban green is cold island in summer.(2) The method estimating impervious surface area from fractional vegetation cover which was proposed by Carlson and Traci Arthur (2000) is applicable to remote sensing images acquired in summer and the accuracy is pretty good, but not to images acquired in winter.(3) The percentage of impervious surface area (%ISA) within the area of the outer-ring way in Shanghai is very high, especially in central urban area and industrial area, the %ISA is more than 90%. Most of the area within the inner-ring way is high-density urban area. The percentage of low-density area within inner-ring is obviously less than that between inner and outer ring. But the percentage of non-urban is almost equal to that of medium-density.(4) The land surface temperature (LST) is positively correlated to %ISA (R2=0.87, p< 0.01), but negatively to NDVI and MNDWI. Multiple linear regression between LST and %ISA, NDVI and MNDWI indicated ISA could increase LST, but vegetation and water body decrease LST, the vegetation is better than water body.(5) The density of urban area can influence the thermoregulation effect of vegetation and water body. The thermoregulation effect increases with urban density for vegetation, but decreases for water body. In lower density area, the effect of water is better than that of vegetation, but weaker than that of vegetation in the medium-and high-density area. (6) The UHI mitigation effect of vegetation and water body is influenced by scale, i.e., extent. The cooling effect of vegetation decreases with extent, but for water body, it increases with extent. The LST is positively correlated to %ISA, but no significant scale effect.(7) To effectively mitigate the urban heat island effect, I suggested that municipality and or policy-maker should consider increasing the water body coverage in lower density urban area, but increase vegetation proportion in medium and high density urban area to mitigate UHI. And also consider paving pervious surface materials instead of impervious surface.