Soil moisture downscaling using multiple modes of the DISPATCH algorithm in a semi-humid/humid region

Authors: ZHENG, J. - Hohai University; LU, H. - Hohai University; Crow, Wade; ZHAO, T. - Chinese Academy Of Sciences; MERLIN, O. - Collaborator; RODRIGUEZ-FERNANDEZ, N. - Collaborator; SHI, J. - Chinese Academy Of Sciences; ZHU, Y. - Hohai University; SU, J. - Chinese Academy Of Sciences; KANG, J. - Universiti Teknologi Mara (UITM); WANG, X. - Hohai University; GOU, Q. - Hohai University

Submitted to: International Journal of Applied Earth Observation and Geoinformation
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 8/30/2021
Publication Date: 9/29/2021
Citation: Zheng, J., Lu, H., Crow, W.T., Zhao, T., Merlin, O., Rodriguez-Fernandez, N., Shi, J., Zhu, Y., Su, J., Kang, J., Wang, X., Gou, Q. 2021. Soil moisture downscaling using multiple modes of the DISPATCH algorithm in a semi-humid/humid region. International Journal of Applied Earth Observation and Geoinformation. 104:102530. https://doi.org/10.1016/j.jag.2021.102530.
DOI: https://doi.org/10.1016/j.jag.2021.102530

Interpretive Summary: Soil moisture estimates obtained from satellite-based sensors are potentially of value for a broad range of agricultural applications. However, their relatively coarse spatial resolution (>30 km) generally prevents them from resolving individual farms and fields. While downscaling techniques have been developed to sharpen the resolution of satellite-derived soil moisture products, these techniques are generally more effective in arid/semi-arid climates than in more humid agricultural regions. This paper examines the potential for applying an existing soil moisture downscaling technique to a sub-humid agricultural region of China. In doing so, it identifies changes to the downscaling algorithm that are required for it to perform adequately in an agricultural landscape. The results of this study will eventually be used to track the availability of root-zone soil moisture at fine spatial scales and better mitigate the impact of soil moisture extremes on local agricultural productivity.

Technical Abstract: Remotely sensed soil moisture (SM) with the enhanced spatial resolution (several kilometers or even tens of meters) is of value for important local applications like irrigation scheduling and farm-scale water management. The Disaggregation based on Physical and Theoretical scale Change (DISPATCH) algorithm is an established evaporation-based tool for passive-SM downscaling. However, its performance in areas where evapotranspiration is generally energy-limited remains uncertain. Here, DISPATCH is applied to coarse-scale L band passive-SM to evaluate its ability to produce a 1-km downscaled SM product in the semi-humid/humid Huai River Basin (HRB). The influence of different calibration modes of linear SEE (soil evaporative efficiency) model on disaggregation results and LST (land surface temperature) inputs at different temporal resolutions are also investigated. The performances of DISPATCH were evaluated using a spatio temporal statistical analysis against available in-situ SM. Globally, DISPATCH performs poorly in HRB with low correlation and least-square regression slopes (significantly different to unity) between downscaled-SM and in-situ observations. This relatively poor performance is attributed to low background SM spatial variability and weak moisture-evaporation coupling within the HRB. However, in spite of those global poor performances, an improvement in the SM spatio-temporal representation was observed under summer-dry conditions over the HRB region. The uncertainty of DISPATCH estimates can be reduced by the multi-date calibration of linear SEE model within the HRB. Likewise, downscaled SM based on merging multiple LST products for three consecutive days have better temporal coverage and lower uncertainty than comparable products only using daily LST retrievals.