Groundwater Protection Zone Mapping for Yulin City, China Using Expert and ML Methods

A 2026 study by Chaoqi Yang compares two methodological frameworks for mapping groundwater protection zones in Yulin City, northwest China. The research uses 16 hydrological datasets to construct an Analytic Hierarchy Process-Geographic Information System (AHP-GIS) framework and a machine learning framework combining Principal Component Analysis and Self-Organizing Map (PCA-SOM). The dual-framework approach serves as a prescriptive decision protocol for selecting methods based on management contexts, addressing transparency and objectivity trade-offs.

Use Cases

• Compare expert-driven and data-driven zoning methodologies based on the described AHP-GIS and PCA-SOM frameworks.
• Analyze spatial consistency and divergence in protection zone classification based on the reported 81.10% agreement and 18.90% difference.
• Interpret key drivers of groundwater protection zones using SHAP analysis, as demonstrated for pollution resistance and mining intensity.
• Evaluate trade-offs between model transparency and objectivity for policy implementation, as discussed in the comparative framework.

Strengths

• The dataset is associated with a detailed methodological study comparing two distinct frameworks for a specific application.
• Analysis includes a quantitative comparison showing 81.10% spatial consistency between the two mapping methods.
• SHAP analysis provides interpretable insights, quantifying driver importance (e.g., 12.89% for PCA-SOM key zone drivers).

Limitations

• The primary file is a 2.1 MB DOCX document; the underlying 16 hydrological datasets are not directly accessible or described in detail.
• Row count and column-level documentation for the source data are unknown, limiting suitability assessment for direct ML use.
• The description focuses on the methodological study and results; the actual data's structure, format, and accessibility require manual inspection after download.

Provenance

Source

Chaoqi Yang via figshare.

Collection Method

The study constructed frameworks using 16 hydrological datasets from Yulin City, combining AHP-GIS and PCA-SOM with machine learning methods.

Freshness

Last updated 2026-04-28 05:23:57; freshness should be verified.

Geography

Yulin City, northwest China.