Machine-learning crystal size distribution for volcanic stratigraphy correlation
2024

Machine Learning for Analyzing Volcanic Rock Sizes

Sample size: 65 publication Evidence: high

Author Information

Author(s): Martin Jutzeler, Rebecca J. Carey, Yasin Dagasan, Andrew McNeill, Cas Ray A. F.

Primary Institution: University of Tasmania

Hypothesis

Can machine learning effectively identify crystal size distributions in volcanic rocks for stratigraphic correlation?

Conclusion

The study demonstrates that machine learning can reliably analyze crystal size distributions in volcanic rocks, aiding in their stratigraphic reconstruction.

Supporting Evidence

  • The machine learning method allows for rapid and unbiased analysis of crystal size distributions.
  • Results from the crystal size distribution analysis were validated by bulk-rock chemical analyses.
  • The study identified six main dacite bodies in the northern sector and three in the southern sector based on crystal size distributions.

Takeaway

This study shows how computers can help scientists quickly measure the sizes of tiny crystals in volcanic rocks to understand their history better.

Methodology

The study used machine learning algorithms to analyze crystal size distributions from photographs of volcanic rock samples.

Potential Biases

Manual quality control is recommended to mitigate potential misidentification by the machine learning model.

Limitations

The method's effectiveness decreases with highly altered rocks where phenocryst textures are lost.

Digital Object Identifier (DOI)

10.1038/s41598-024-82847-0

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication