A surface soil temperature retrieval algorithm based on AMSR-E multi-frequency brightness temperatures

ABSTRACT

In this study, a multi-frequency statistical algorithm is proposed for retrieving surface soil temperature () from AMSR-E (Advanced Microwave Scanning Radiometer for Earth Observation System) brightness temperature () observations. The algorithm was developed based on a regression analysis between from all AMSR-E bands and the corresponding in situ observed by a soil moisture and temperature network in the central Tibetan Plateau. The new algorithm effectively utilizes information from the different bands provided by AMSR-E, lessening the influence of soil moisture, vegetation, and water vapour. Further validations were conducted based on seven soil moisture and temperature observation networks distributed globally. The results showed that the new multi-frequency algorithm can produce values with a mean bias of less than 2 K and a root mean square error of less than 3 K for different vegetation-covered areas of the globe. Compared with a widely used single-band inversion algorithm, the new multi-frequency algorithm has better accuracy in estimating and does not suffer from considerable overestimation or underestimation across these networks, indicating good transferability. This algorithm could contribute to research relating to the land energy balance by providing consistent and independent long-term estimates of daily global . Nevertheless, the new algorithm has limited ability to retrieve of frozen soil, given that AMSR-E values are affected by the deep soil temperature after a soil is frozen.

Acknowledgements

This work was supported by the National Basic Research Program of China: [Grant Number 2015CB953703] and the National Natural Science Foundation of China: [Grant Numbers 91537210 and 41371328]. The brightness temperature data of AMSR-E were provided by the National Snow and Ice Data Center. The in situ data were obtained from ISMN. The computation of this work is supported by Tsinghua National Laboratory for Information Science and Technology.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the National Natural Science Foundation of China: [Grant Numbers 91537210 and 41371328] and National Basic Research Program of China: [Grant Number 2015CB953703].