Heavy Metals

Models contamination and exposure drivers to map risk hotspots, guiding sampling, remediation, and public-health interventions.

This project uses machine learning to predict heavy metal contamination patterns and related exposure risks for environmental and public health applications. We link measured metal concentrations to covariates such as proximity to sources (mining/industry), land use, terrain and hydrology (transport/accumulation), soil properties affecting mobility, and remote sensing proxies. Models capture nonlinear drivers and interactions, with validation designed to test real spatial generalization. Outputs include risk/hotspot maps and driver summaries to support targeted monitoring and remediation.

Main focus

  • Predict spatial patterns of heavy metals and exposure risk using ML.
  • Identify hotspots and key drivers for targeted intervention.
  • Support efficient sampling, remediation, and health-protection decisions.

Objectives

  1. Compile contamination/exposure datasets and relevant environmental covariates.
  2. Train ML models with spatially realistic validation and uncertainty awareness.
  3. Deliver hotspot maps + driver summaries to guide monitoring and remediation actions.

Graphical abstract

:::: {.row .justify-content-sm-center} ::: {.col-sm-10 .mt-3 .mt-md-0}

::: ::::

::: caption Graphical abstract summarizing the workflow and key results. :::

Papers (selected)

Paper 1 — Catchment-Scale Spatiotemporal Mapping of Heavy Metals

DOI




Paper 2 — Statistical Assessment of Soil Heavy Metal Hotspots

DOI




Paper 3 — Machine Learning Identifies Key Drivers of Mercury Exposure

DOI



References