Comparison Analyses of Equal Interval Method and Mean-standard Deviation Method Used to Delimitate Urban Heat Island

Abstract

Fuzhou was chosen as our study area.Using Temperature/Emissivity separation algorithm,land surface temperatures(LST) were retrieved from ASTER(Advanced Spaceborne Thermal Emission and Reflectance Radiometer) data set which was collected on April 6th,2004.Two methods named equal interval method and mean-standard deviation method were employed to divide land surface temperatures into four,five and six classes respectively.Synchronously,the domain of urban heat island(UHI) was determined according to its different definitions.Then multi-angle comparisons covering the structure,spatial distribution and the tiny undulation of the UHI delimitated by different methods were applied to choose the better methods.The result indicated that the percentage of the UHI area delimitated by both methods was fluctuant with the class number,while for the intensity of the UHI,mean-standard deviation method was less sensitive to the number of thermal field than that of equal interval method.Furthermore,the former was better than the latter in describing UHI spatial distribution and details.Based on standard deviation of land surface temperature against its mean value,the mean-standard deviation method,to some extent,likely minimized the problems that resulted from the inconsistent time periods of collecting remote sensing data in studying the multi-temporal urban heat island.However,our research was only a preliminary case study,more details and uncertainties about these methods and results needed to be further studied.