A physically based algorithm for retrieving land surface temperature under cloudy conditions from AMSR2 passive microwave measurements. Issue 5 (19th March 2019)

• Record Type: Journal Article
• Authors: Huang, Cheng; Duan, Si-Bo; Jiang, Xiao-Guang; Han, Xiao-Jing; Leng, Pei; Gao, Mao-Fang; Li, Zhao-Liang
• Is Part Of: International journal of remote sensing. Volume 40:Issue 5/6(2019)
• Page Start: 1828
• Page End: 1843
• Publication Date: 2019-03-19
• Subjects: Remote sensing -- Periodicals; Télédétection -- Périodiques; 621.3678
• DOI: 10.1080/01431161.2018.1508920
• Languages: English
• ISSNs: 0143-1161

Abstract

ABSTRACT: Satellite remote sensing provides a unique way to measure land surface temperature (LST) at regional and global scales. Algorithms using thermal infrared (TIR) data provide a reliable way to retrieve LST. However, they are limited to clear-sky conditions due to their inability to penetrate clouds. As an alternative for LST retrieval, passive microwave data are much less affected by clouds and water vapour than TIR data. In this study, we presented an improved physically based algorithm for the retrieval of LST under cloudy atmospheric conditions from Advanced Microwave Scanning Radiometer 2 (AMSR2) brightness temperature measurements at 18.7 and 23.8 GHz vertically polarized channels based on the assumption that the emissivity relationship between the two adjacent frequencies is linear. The performance of the algorithm was firstly evaluated using simulation data, with a root mean square error (RMSE) of approximately 2.1 K. Moreover, the RMSE value reduces with precipitable water vapour (PWV) increasing. This algorithm was further applied to AMSR2 measurements. The retrieved cloudy LST was compared with ground-based air temperature over China in 2016. The bias varies from approximately 2 K to 4 K and the RMSE from approximately 4 K to 6 K during daytime and night-time. To eliminate the systematic bias between the retrieved LST and the ground-based air temperature, a linear adjustment was performed to the retrieved LST during daytime and nighttime, respectively. The accuracies for the adjusted LST are nearly the same during daytime and night-time, with an RMSE of approximately 3.6 K. The combination of this physically based LST retrieval algorithm with TIR LST algorithm is attractive for generating an all-weather LST product at global scale.