Urban Heat Islands Modeling and Analysis

A key issue in thermal remote sensing of urban areas is how to use pixel‐based land surface temperature (LST) measurements to characterize and quantify urban heat islands (UHIs) observed at various spatial scales. This chapter introduces methods for characterizing and modeling UHIs using remotely sensed LST data. It also introduces a kernel convolution modeling method for two‐dimensional LST imagery to characterize and model the UHI in Indianapolis, USA, as a Gaussian process. The main contribution of this method lies in that UHIs can be examined as a scale‐dependent process by changing the smoothing kernel parameter. Furthermore, the chapter introduces an object‐based image analysis procedure to extract hot spots from LST maps with a case study of Athens, Greece. The chapter utilizes spatial and thermal attributes associated with these objects (hot spots) for the analyses of the intensity, position, and spatial extent of UHIs.