An Investigation of the Relationship between Land Cover Ratio and Urban Heat Island

This paper investigates the relationship between land cover ratio and UHI in Guangzhou of China. Firstly, the hourly mean temperature was computed from the hourly automatic weather station (AWS) data in October, 2004. Secondly, the CBERS remote sensing image acquired in October 10, 2004 was classified using support vector machine (SVM). Thirdly, the classification outcome was overlapped with a round buffer with 1.5KM radius centered on the AWS, then the land cover ratio, Edge Density (ED) and Mean Fractal Dimension (MFRACT) of the buffer were computed out. Finally, correlation coefficient between hourly mean temperature and land cover ratio, ED and MFRACT was calculated. It concludes that heat island intensity is heavier during the night-time than daytime. Stations which have higher vegetation ratio and higher ED have lower heat island effect. On the contrary, stations which have the higher impervious ratio and lower ED have more serious heat island effect.