Surface Radiance Temperature Variability Analysis through Multi-Level Cell Matrix (MLCM) and Statistical Techniques

Abstract

The global inclination towards the urbanization enlightens the evolving concern for study on variation of the urban climate due to the heat island effect and global warming. Surface radiance temperature (SRT) is undoubtedly one of the most significant parameters in urban. The work presents the thermal image processing techniques to retrieve the SRT from the LandSat ETM+ data. The spatial variability was analyzed to illustrate their indigenous spatial distribution over urban-rural (Rurban) areas contributing to heat island. The results apparently indicate the increasing spatial thermal configuration along Rurban areas and vice versa. The percentage of urban landscape patches with high, sub high and low densities indicates their inherent characteristics. High temperature type was the dominant class in urban core, while sub medium temperature and low temperature type was the main heat landscape type in rural area. In the urban fringe, temperature variability showed very complex result from the various land use. Statistical tools such as spatial coefficient of variation, the correlogram and Semivariograms are introduced to characterize the spatial variability for each of the directional attribute. These tools also helped in detecting and quantifying the major to minor scales of spatial variability. The study also validates that the approach of integration of spatial variability with statistics for a reliable tool to monitor the thermal dynamics. The present paper addresses the role of urban behavior in the study of thermal variability for understanding energy demand.