Researchers Develop Efficient New Gold Mapping Method

A novel analytical process combines soil extraction with machine learning to identify gold potential in geological samples.

Updated on Oct. 24, 2026 in Geography

Glass test tubes with soil samples sitting on a dusty, weathered stone surface with geological aggregate nearby.
Researchers have introduced a cost-effective geological mapping method that combines soil extraction and machine learning to improve gold exploration accuracy. AI Illustration. Upload story photo >

Scientists have created a low-cost, reliable method to detect gold in soil by utilizing aqua regia digestion and solid phase extraction with activated carbon. This approach integrates machine learning algorithms, including Random Forest and Gaussian Process Regression, to map potential gold distribution zones.

Why it matters

The study provides a more accessible and precise technique for gold detection in geological surveys. By improving the reliability of soil data analysis, this method could streamline exploration efforts and enhance the accuracy of predictive resource mapping.

The method detects soil gold concentrations between 18.0 ng/g and 53,000 ng/g, with aqueous detection limits reaching 0.01 µg/mL. Researchers confirmed the technique using 40 geological samples, validating performance at a 95 percent confidence level.

The details

The process treats soil samples with aqua regia before using activated carbon to extract the gold. Machine learning models, specifically Ordinary Kriging, Random Forest, and Gaussian Process Regression, then interpret the resulting data to generate predictive maps.

Timeline

  1. October 24, 2026: The research findings were published.

The Big Picture

This development shifts the discipline of geological exploration by offering an alternative to traditional fire assay analysis. By bridging wet chemistry with predictive modeling, the study moves the field toward faster, digital-first resource identification.

This analytical technique could eventually lower the costs of mineral exploration, potentially leading to more efficient identification of gold deposits globally. For the industry, it represents a path toward cheaper, high-accuracy testing that reduces the need for heavy lab equipment.

The takeaway

The research successfully proves that machine learning can be effectively paired with chemical soil extraction to produce highly accurate geological maps. Prospective users in the mining sector can now look toward this method as a viable alternative for processing large volumes of soil samples.

Further reading

Learn more about the latest innovations in Geography.

Source note: This article includes information reported by Nature.