Assessment of the Spatial Variability of Metal Contaminants Using Digital Mapping.

Area of Science:

• Environmental Science
• Geospatial Analysis
• Soil Science

Background:

• Toxic metal contamination poses significant environmental and health risks.
• Understanding the spatial distribution of toxic metals is crucial for effective management.
• Digital Soil Mapping (DSM) offers a powerful framework for environmental assessment.

Purpose of the Study:

• To investigate the spatial prediction of toxic metals in the Ghorveh Plain using DSM.
• To identify environmental covariates influencing toxic metal distribution.
• To develop accurate predictive models for toxic metal contamination.

Main Methods:

• Utilized digital soil mapping (DSM) methodology.
• Analyzed 150 soil samples for toxic metal concentrations and soil properties.
• Employed genetic algorithms to identify relevant environmental covariates.
• Applied machine learning algorithms (Random Forests, Cubist, Regression Trees) for spatial prediction.

Main Results:

• The Random Forests (RF) model demonstrated the most optimal prediction performance.
• Soil properties and hydrologic factors were identified as primary drivers of toxic metal distribution.
• The RF model, integrated with bootstrapping, generated robust prediction and uncertainty maps.

Conclusions:

• DSM combined with machine learning provides a viable approach for mapping toxic metal contamination.
• The developed models can aid in monitoring and prioritizing remediation efforts in contaminated areas.
• Identifying key environmental covariates enhances understanding of pollution sources and spread.