Abstract: Surface temperature is a crucial parameter in the studies of urban environmental monitoring, geothermal anomalies, global climate change, and other topics. The full-spectrum spectral imager on board Gaofen-5 (GF-5) satellite has a high spatial resolution imaging capability of 40 m in the thermal infrared band, which can offer detailed information on the spatial distribution and variation of surface temperature. In this paper, two split-window algorithms, two-channel and four-channel, are used to invert GF-5 data to obtain the surface temperature of Tianjin, China, in March 2019, and the accuracy of the inversion results is evaluated using ASTER surface temperature product at the same time. The results show that the optimal root-mean-square error of the two-channel split-window algorithm for surface temperature is 1.19 K, while that of the four-channel split-window algorithm is 2.31 K. And the comparison of different channel combinations shows that the B9/B10 combination of the two-channel split-window algorithm has the highest inversion accuracy. In addition, to verify the application capability of GF-5 high-resolution thermal infrared data in monitoring the heat island effect, a comparative analysis of GF-5 and ASTER surface temperature and heat island effect is conducted in the region surrounding Yadian Reservoir of Tianjin. The results show that the high-resolution surface temperature data of GF-5 can play an important role in urban thermal environment monitoring and urban planning.

Key words: Gaofen-5, thermal infrared, land surface temperature, split-window algorithm, urban heat island effect