Luo et al. (2025) Surface Soil Moisture Retrieval over Winter Wheat Fields Based on Fused Multispectral and L-Band MiniSAR Data

Identification

• Journal: Water
• Year: 2025
• Date: 2025-11-22
• Authors: Ziyi Luo, Xianyu Zhang, Yonghui Wang, Chengcai Zhang, Mingliang Jiang, Xingxing Zhu
• DOI: 10.3390/w17233345

Research Groups

• School of Water Conservancy and Transportation, Zhengzhou University, Zhengzhou, China.
• Department of Remote Sensing Engineering, Henan College of Surveying and Mapping, Zhengzhou, China.
• Henan Academy of Geology, Zhengzhou, China.
• Institute of Farmland Irrigation, Chinese Academy of Agricultural Sciences, Xinxiang, China.

Short Summary

This study proposes a high-accuracy surface soil moisture (SSM) retrieval method for winter wheat fields by fusing Sentinel-2 satellite and UAV multispectral data with L-band MiniSAR observations. The results demonstrate that multi-platform data fusion combined with machine learning significantly outperforms single-satellite approaches, particularly at shallow soil depths.

Objective

• To evaluate the synergistic potential of fused multispectral imagery and L-band MiniSAR data for retrieving SSM in winter wheat fields using a modified Water Cloud Model (WCM) and machine learning algorithms.

Study Configuration

• Spatial Scale: Field-scale study conducted in Xunxian, Hebi City, Henan Province, China (flat terrain, calcareous fluvo-aquic soil).
• Temporal Scale: Field campaign conducted from 26 to 29 April 2024, during the jointing stage of winter wheat.

Methodology and Data

• Models used: Modified Water Cloud Model (WCM) incorporating the Dimidiate Pixel Model (DPM) for vegetation coverage correction; Machine learning algorithms: Random Forest (RF), Extreme Gradient Boosting (XGBoost), and Extreme Learning Machine (ELM).
• Data sources:
  • Satellite: Sentinel-2 multispectral imagery (10 m resolution).
  • UAV: RedEdge-MX multispectral camera (0.08 m resolution) and L-band MiniSAR system (0.6 m resolution) providing four polarization modes (VV, VH, HV, HH).
  • Ground Truth: 40 soil sampling points at depths of 0–10 cm and 0–20 cm, and 24 vegetation water content (VWC) sampling points.
• Data Processing: Pixel-level fusion of multispectral data resampled to a unified 1 m spatial resolution.

Main Results

• Fusion Performance: Fused multispectral data (Sentinel-2 + UAV) consistently improved retrieval accuracy compared to using Sentinel-2 data alone across all machine learning models.
• Depth-specific Accuracy:
  • 0–10 cm depth: The best performance was achieved using fused data with VV polarization and the RF model ($R^2 = 0.85$, $RMSE = 1.51\%$, $MAE = 0.95\%$).
  • 0–20 cm depth: The best performance was achieved using fused data with VV polarization and the XGBoost model ($R^2 = 0.67$, $RMSE = 2.61\%$, $MAE = 1.98\%$).
• Algorithm Sensitivity: ELM showed the most significant improvement after data fusion, with $R^2$ increases of up to 0.40 and $RMSE$ reductions of up to 18.24%.
• Polarization Effects: Co-polarization (VV and HH) generally outperformed cross-polarization (VH and HV), with VV being the most effective for SSM retrieval in this study.

Contributions

• Demonstrates the effectiveness of integrating L-band MiniSAR (which offers greater penetration than C-band) with fused multi-platform optical data for agricultural monitoring.
• Provides a modified WCM framework that better accounts for vegetation coverage effects in dense crop canopies.
• Establishes a high-precision, scalable methodology for field-scale soil moisture mapping that leverages the complementary strengths of satellite and UAV platforms.

Funding

• National Natural Science Foundation of China (Reference code: 52579025).
• Natural Science Foundation of Henan (Reference code: 222300420539).
• CMA-Henan Key Laboratory of Agrometeorological Support and Applied Technique (Reference code: AMF202409).

Citation

@article{Luo2025Surface,
  author = {Luo, Ziyi and Zhang, Xianyu and Wang, Yonghui and Zhang, Chengcai and Jiang, Mingliang and Zhu, Xingxing},
  title = {Surface Soil Moisture Retrieval over Winter Wheat Fields Based on Fused Multispectral and L-Band MiniSAR Data},
  journal = {Water},
  year = {2025},
  doi = {10.3390/w17233345},
  url = {https://doi.org/10.3390/w17233345}
}

Generated by BiblioAssistant using gemini-3-flash-preview (Google API)