Comparison Between Remotely-Sensed and Numerically-Simulated High-Resolution Surface Temperatures at an Urban Commercial Site

Abstract

This study compares high-resolution urban surface temperatures (Ts) obtained through airborne-remote sensing and from a heat balance simulation model to clarify the characteristics of detailed Ts distributions within an urban district and confirmed the capabilities and limitations of the simulation model for urban Ts assessment. The remote sensing observations were conducted in the daytime of summer and winter. For the heat balance simulation, detailed three-dimensional (3D) geometry and component material data were collected by using a previously developed modeling method, materials database, and field surveys. 3D Ts distributions were simulated for the summer and winter days. Both remote sensing and numerical simulation showed the effects of surface geometry and solar reflectance, and also the time-integrating effects of the heat capacity of materials. The point-to-point comparison of results between the numerically simulated Ts and remotely sensed Tr revealed that the Ts was close to the Tr at many points for both seasons. The Tr values were slightly smaller than those of Ts for these points owing to the effects of surface emissivity. It is confirmed that the heat balance simulation model can effectively simulate Ts distribution through the comparison with the airborne remote sensing data. The factors causing the difference between Ts and Tr were assessed as mixels in the remote sensing image, reproductivity of materials in the model, actual microclimate and anthropogenic heat emission, and the limitation of the heat balance simulation that used a one-dimensional (1D) heat conduction equation.