Urban Land Surface Temperature Retrieval From High Spatial Resolution Thermal Infrared Image Using a Modified Split-Window Algorithm

The urban canopy multiple scattering thermal radiative transfer (UCM-RT) model, incorporating the effects of urban geometry and adjacent thermal radiation from neighboring pixels, depicts the process of thermal radiation transfer on the urban surface and therefore provided new opportunity to develop new retrieval algorithms for urban land surface temperature (ULST). This article aims at developing an urban split-window (USW) algorithm for deriving ULST from high spatial resolution thermal infrared (TIR) data from the visible and infrared multispectral sensor (VIMS) onboard Chinese GaoFen-5 (GF-5) satellite. The VIMS provides four-channel TIR image with a spatial resolution of 40 m. The coefficients of the USW algorithm were obtained based on several subranges of atmospheric column water vapors (cwv), emissivity, and sky view factors (SVFs) under various land surface conditions, by removing the geometry, adjacent, and atmospheric effects. Methods of estimating urban pixel emissivity and cwv in urban areas were also conducted. The sensitive analysis of instrument noise and uncertainty of cwv, pixel emissivity, and SVFs demonstrated the reliability of the USW algorithm in ULST retrieval. The accuracy evaluation shows that the root-mean-square errors of the ULST results are less than 0.7 K in theory. Compared with the conventional SW algorithms and publicly released LST products, the USW algorithm obtained better results in estimating ULST, especially in high-density building areas. Finally, the USW algorithm is expected to be beneficial to the application of multiple TIR sensor (TIRS) data, for example, the newly launched GF-5 No. 2 satellite images.