Airborne Observations of Thermal Anisotropy from Urban Residential Neighbourhoods in Salt Lake City

Most surface covers show a directional dependence (anisotropy) of remotely sensed brightness temperature, meaning that the same area can appear to have different surface temperatures when viewed from different directions. This is important as more interest is placed in extracting urban surface temperatures, particularly from satellite imagery. Urban environments in particular have been found to produce large anisotropy due to the three-dimensional surface temperature distribution of the urban surface. Past work, however, has mainly focused on built up parts of a city, and less so on treed residential areas. This research focuses on characterizing the anisotropy of urban residential areas with varying amounts of tree canopy coverage. Three sites in Salt Lake City, Utah USA were sampled over the course of three days in July 2018. Sites 1 and 2 are characterized by tall trees (1.5 times building height) and plan area tree cover values of 18% and 22% respectively. Site 2 is a more open neighbourhood with a building plan area fraction of approximately 15% compared to 26% in site 1. Site 3 has both shorter (0.5 times building height) and sparser (15% plan area tree cover) trees. Measurements were obtained by imaging three study sites of approximately 500m x 500m with a thermal imager (FLIR T650sc with a spectral resolution of 7.5 to 14 µm and a 45° x 34° lens) from a helicopter with nadir and two off-nadir angles of 45° and 25° sampled for each flight. A Heitronics KT15.81 (8 - 14 µm) thermal radiometer was used to provide an independent evaluation of the thermal anisotropy. An Aaronia GPS Logger was connected to the thermal imager mount to measure the camera orientation. Three flights were flown over each study site to capture anisotropy associated with three different solar zenith angles of approximately 50°, 20°, and 64°. The large field of view permits subdividing the images to represent three separate off-nadir angles each associated with a FOV of 20°. A total of 14 off-nadir view angles for each of the 8 cardinal and intercardinal view directions is obtained over 9 flights. Atmospheric corrections using the single channel method were performed using MODTRAN6 with atmospheric profile data obtained from locally launched radiosondes and a ground-based microwave radiometer. Simultaneous measurements of surface temperature obtained from fixed and mobile traverses are used to evaluate the corrections. For each study site we present polar plots of the directional brightness temperature, from which the maximum anisotropy and difference from nadir temperature are derived along with an assessment of the spatial and temporal variability of the anisotropy. These observations will provide an important evaluation data set for models of urban thermal anisotropy.