Zhenghao et al. (2026) Integrated remote sensing retrieval of surface and root-zone soil moisture through physical mechanisms-guided machine learning

Identification

• Journal: Journal of Hydrology
• Year: 2026
• Date: 2026-09-09
• Authors: Li Zhenghao, Yuan Qiangqiang, YUE Linwei, Yang Qianqian, Huanfeng Shen, Liangpei Zhang
• DOI: 10.1016/j.jhydrol.2026.136393

Research Groups

• School of Geodesy and Geomatics, Wuhan University, China
• Key Laboratory of Geospace Environment and Geodesy, Ministry of Education, Wuhan University, China
• Department of Geography, Faculty of Social Sciences, Hong Kong Baptist University, Hong Kong Special Administrative Region of China
• State Key Laboratory of Information Engineering, Survey Mapping and Remote Sensing, Wuhan University, China

Short Summary

This study proposes an integrated retrieval framework for surface and root-zone soil moisture that combines physical mechanisms with machine learning models, demonstrating strong generalization and spatial extrapolation capabilities.

Objective

• Investigate the feasibility of integrating physical mechanisms with machine learning models to retrieve surface and root-zone soil moisture from remote sensing data

Study Configuration

• Spatial Scale: Continental scale in China
• Temporal Scale: Long-term (e.g., annual, seasonal)

Methodology and Data

• Models used:
  • Radiative transfer model based on brightness temperature data
  • Machine learning models for optimizing physical model structure and parameters
  • Physical mechanism-guided machine learning retrieval model
• Data sources:
  • Satellite remote sensing data (brightness temperature)
  • In-situ soil moisture measurements

Main Results

• The integrated retrieval framework demonstrated strong generalization and spatial extrapolation capabilities in jointly retrieving surface and root-zone soil moisture.
• High-accuracy soil moisture retrieval was achieved in areas with limited monitoring sites.

Contributions

• This study provides a novel approach to integrating physical mechanisms with machine learning models for large-scale soil moisture retrieval, addressing the limitations of existing methods.
• The developed framework enables high-accuracy soil moisture retrieval in areas with limited monitoring sites, providing robust data support for hydrological and climate-related research.

Funding

• This research was funded by the National Natural Science Foundation of China (Grant No. 42161140701)
• The State Key Laboratory of Information Engineering, Survey Mapping and Remote Sensing, Wuhan University

Citation

@article{Zhenghao2026Integrated,
  author = {Zhenghao, Li and Qiangqiang, Yuan and Linwei, YUE and Qianqian, Yang and Shen, Huanfeng and Zhang, Liangpei},
  title = {Integrated remote sensing retrieval of surface and root-zone soil moisture through physical mechanisms-guided machine learning},
  journal = {Journal of Hydrology},
  year = {2026},
  doi = {10.1016/j.jhydrol.2026.136393},
  url = {https://doi.org/10.1016/j.jhydrol.2026.136393}
}