JoVE - Journal of Visualized Experiments
JoVE Visualize
Contact Us

Machine learning-based prediction of deep soil metal(loid) contamination in industrial areas: Role of surface environmental factors

  • 0School of Metallurgy and Environment, Central South University, Changsha, 410083, China; School of Resource & Environment and Safety Engineering, Hunan University of Science and Technology, Xiangtan, 411201, China; Hunan Province Key Laboratory of Coal Resources Clean Utilization and Mine Environment Protection, Hunan University of Science and Technology, Xiangtan, 411201, China.
Environmental Pollution +

|

November 1, 2025

|

November 1, 2025

View abstract on PubMed

Summary

This summary is machine-generated.

This study uses a random forest (RF) model to predict deep soil metal(loid) contamination using surface environmental data. This approach reduces the need for costly deep soil surveys, aiding in efficient remediation strategies.

Area Of Science

  • Environmental Science
  • Geochemistry
  • Soil Science

Background

  • Predicting soil contamination is vital for effective remediation and risk management.
  • In-situ contamination surveys are expensive and time-consuming.
  • Deep soil contamination poses significant environmental and health risks.

Purpose Of The Study

  • To develop a random forest (RF) model for predicting deep soil metal(loid) contamination.
  • To utilize readily available surface environmental factors for prediction.
  • To assess the model's accuracy and applicability in smelting areas.

Main Methods

  • Application of a random forest (RF) model.
  • Utilizing surface soil physicochemical properties (SOM, Fe, S, metal(loid)s, clay content).
  • Incorporating anthropogenic factors (pollution sources, land use, land cover).

Main Results

  • The RF model accurately predicted the distribution of As, Cd, Cu, Pb, and Zn in deep soil (R² values 0.757–0.897).
  • Deep soil metal(loid) content negatively correlated with surface SOM, Fe, and S, and positively with surface metal(loid)s and clay.
  • Prediction errors decreased with higher surface soil metal(loid) content and were insensitive to prediction depth.

Conclusions

  • The RF-based approach provides a cost-effective and low-carbon solution for deep soil contamination assessment.
  • This method significantly reduces the need for extensive drilling and energy consumption.
  • Enables targeted remediation efforts and improved risk prevention for contaminated sites.

Keywords: