Praxis et al. (2025) Development of a Machine Learning-Based Soil Moisture Data Gap-Filling Model

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Crisis and Emergency Management Theory and Praxis
• Year: 2025
• Date: 2025-12-30
• Authors: Crisis and Emergency Management: Theory and Praxis, Tae Hyung Kim Tae Hyung Kim, Hyeong Yoon So Hyeong Yoon So, Se Jeong Lee Se Jeong Lee, Hyeon-Cheol Yoon Hyeon-Cheol Yoon
• DOI: 10.14251/crisisonomy.2025.21.12.105

Research Groups

Not specified in the provided text. The study focuses on monitoring systems in Haenam and Yesan.

Short Summary

This study developed and evaluated machine learning models to gap-fill missing soil moisture data from monitoring systems in Haenam and Yesan. The E-dataset XGBoost model, utilizing lagged, accumulated, and time-series precipitation features, demonstrated superior performance for gap-filling 0.1- and 0.2-meter soil moisture data.

Objective

• To gap-fill missing soil moisture data from monitoring systems in Haenam and Yesan using machine learning models trained with precipitation-based features.

Study Configuration

• Spatial Scale: Monitoring sites within the Haenam and Yesan regions.
• Temporal Scale: Continuous monitoring data (specific temporal resolution not provided).

Methodology and Data

• Models used: XGBoost (XGB) algorithm.
• Data sources: Soil moisture monitoring systems (Haenam, Yesan), corrected precipitation data.

Main Results

• Training D and E datasets, which incorporated lagged, accumulated, and time-series precipitation features, combined with the XGB algorithm, showed superior performance in gap-filling.
• The E-dataset XGB model consistently outperformed other combinations and is deemed suitable for gap-filling 0.1- and 0.2-meter soil moisture data in the Haenam and Yesan regions.

Contributions

• Development and validation of a robust machine learning approach (E-dataset XGB model) for gap-filling missing soil moisture data at 0.1- and 0.2-meter depths using precipitation-based features.
• Provides a practical solution to enhance the usability of soil moisture monitoring data in regions prone to observation gaps.

Funding

Not specified in the provided text.

Citation

@article{Praxis2025Development,
  author = {Praxis, Crisis and Emergency Management: Theory and and Kim, Tae Hyung Kim Tae Hyung and So, Hyeong Yoon So Hyeong Yoon and Lee, Se Jeong Lee Se Jeong and Yoon, Hyeon-Cheol Yoon Hyeon-Cheol},
  title = {Development of a Machine Learning-Based Soil Moisture Data Gap-Filling Model},
  journal = {Crisis and Emergency Management Theory and Praxis},
  year = {2025},
  doi = {10.14251/crisisonomy.2025.21.12.105},
  url = {https://doi.org/10.14251/crisisonomy.2025.21.12.105}
}