Shi et al. (2025) Improved soil moisture mapping using an integrated cyclic modeling and bias correction approach

Identification

• Journal: Remote Sensing Applications Society and Environment
• Year: 2025
• Date: 2025-10-01
• Authors: Yajie Shi, Wei Dai, Guangsheng Chen, Xi Zhang, Nan Li, Weijun Fu
• DOI: 10.1016/j.rsase.2025.101741

Research Groups

• National Key Laboratory for Development and Utilization of Forest Food Resources, Zhejiang A&F University, Hangzhou, China
• College of Environment and Resources, College of Carbon Neutrality, Zhejiang A&F University, Hangzhou, Zhejiang, China
• Louisiana State University-Agricultural Center, Bossier City, LA, USA
• Department of Biosystems Engineering and Soil Science, University of Tennessee, Knoxville, TN, USA
• Environmental Sciences, University of California, Riverside, CA, USA

Short Summary

This study developed an integrated cyclic modeling and bias correction approach using an XGBoost model to downscale soil moisture, producing a 500 m resolution product with significantly improved accuracy compared to single-shot modeling.

Objective

• To develop an improved approach for downscaling soil moisture (SM) products to a higher spatial resolution (500 m) by integrating cyclic modeling and bias correction, addressing limitations of existing low-resolution SM products and enhancing downscaling accuracy.

Study Configuration

• Spatial Scale: Surface (0–5 cm) soil moisture product at 500 m spatial resolution.
• Temporal Scale: Daily (1 day) temporal resolution.

Methodology and Data

• Models used: Extreme Gradient Boosting (XGBoost) regression model.
• Data sources: International Soil Moisture Network (ISMN) stations (quality controlled and bias corrected), Soil Moisture Active Passive Level 4 (SMAP-L4) SM data, dynamic environmental variables.

Main Results

• The XGBoost model effectively described the relationship between soil moisture and environmental variables, achieving a correlation coefficient (R) of 0.98 and a root mean square error (RMSE) of 0.007 m³/m³.
• The generated 500 m soil moisture data showed good comparability with SMAP-L4 SM data, with 83.2 % of 1996 points having R > 0.6.
• The integrated cyclic modeling and bias correction approach improved downscaling accuracy, with R, RMSE, and mean absolute error (MAE) improving by 6.5 %, 9.3 %, and 9.6 %, respectively, over single-shot modeling.

Contributions

• Developed a novel integrated cyclic modeling and bias correction approach for soil moisture downscaling.
• Demonstrated improved downscaling accuracy by recycling qualified stations and combining predicted bias with dynamic environmental variables.
• Produced a 500 m/day surface soil moisture product that can supplement regional soil moisture databases.

Funding

• Not explicitly stated in the provided text.

Citation

@article{Shi2025Improved,
  author = {Shi, Yajie and Dai, Wei and Chen, Guangsheng and Zhang, Xi and Li, Nan and Fu, Weijun},
  title = {Improved soil moisture mapping using an integrated cyclic modeling and bias correction approach},
  journal = {Remote Sensing Applications Society and Environment},
  year = {2025},
  doi = {10.1016/j.rsase.2025.101741},
  url = {https://doi.org/10.1016/j.rsase.2025.101741}
}