Unveiling hidden heavy metal hotspots in mining landscapes using integrated hyperspectral remote sensing

Area of Science:

• Environmental Science
• Geoscience
• Remote Sensing

Background:

• Regional ecological risk assessments often use interpolated toxic heavy metal (THM) data, leading to spatial uncertainty and masking fine-scale contamination.
• Sparse field sampling and interpolation errors limit the accuracy of traditional heavy metal contamination mapping.

Purpose of the Study:

• To evaluate an integrated ground-satellite hyperspectral framework for mapping THMs and ecological risks in early-stage mining areas.
• To develop a robust inversion workflow for enhancing the accuracy of THM estimation using hyperspectral data.

Main Methods:

• Developed a hyperspectral inversion workflow combining spectral transformations, band optimization, and machine learning algorithms.
• Calibrated satellite spectra with ground-based measurements to improve estimation accuracy for copper (Cu), nickel (Ni), and chromium (Cr).
• Generated high-resolution, spatially continuous maps of pollution indices (I_geo, INI, E_i, RI).

Main Results:

• Estimation accuracy for Cu, Ni, and Cr improved by 31%, 10%, and 34%, respectively, after calibration.
• Identified distinct spatial patterns, with Cr showing severe enrichment (I_geo = 3.24-3.53) and Cu/Ni showing mild-moderate enrichment (I_geo = 0.40-2.01).
• Detected localized high-risk hotspots beyond mining footprints, indicating the influence of natural transport mechanisms.

Conclusions:

• Hyperspectral inversion offers superior capability in capturing spatial nuances of soil contamination compared to traditional monitoring.
• The framework provides high-precision data for early-stage risk management and boundary-spanning pollution control in mining regions.
• Identified off-site contamination zones, highlighting the need for integrated monitoring approaches.