Machine-Assisted Granite Mapping in the Archaean Yilgarn Craton

The Challenge

Granites in the Yilgarn host a range of important mineral commodities, but opportunities for discovery are increasingly limited to areas buried beneath sediment and regolith. Exploration in these regions comes with increasing costs and challenges. Innovative methods to differentiate granitic rock types consistently, objectively, and efficiently beneath cover rocks could reduce exploration risk.

Proposed Solution

By applying machine learning (ML) to integrate large geophysical datasets (gravity, magnetics, and radiometrics), this project will objectively classify lithology across the entire Yilgarn Craton. Supervised and unsupervised ML will be used to classify domains with similar properties, generating consistent and objective bedrock maps, supporting production of a consistent granite map of the entire region. This output will help target early-stage mineral exploration to the areas most prospective for critical and precious minerals (e.g., lithium, REEs and gold).

Proposed Benefits to WA

This project will improve geological mapping efficiency in the key Yilgarn mineral province, lowering exploration risk and increasing discovery rates under cover. These efficiency dividends will help attract private sector investment in WA mineral exploration, reduce environmental impacts of exploration in the state, and position WA as a global leader in exploration technology.

Report DOI

DOI: 10.71342/560531747294