Pink crystal spodumene.Credit: Robert Lavinsky
New machine learning models can predict the location of minerals on Earth, and possibly on other planets, by exploiting patterns of mineral associations. This progress is of great value to science and industry, who are continuously exploring deposits to unravel the history of the planet and mine resources for practical uses such as rechargeable batteries.
A team led by Shaunna Morrison and Anirudh Prabhu aims to develop methods to identify specific mineral occurrences, a goal traditionally considered to be both a science and an art. This process has often been dependent on personal experience and sound luck.
The team created[{” attribute=””>machine learning model that uses data from the Mineral Evolution Database, which includes 295,583 mineral localities of 5,478 mineral species, to predict previously unknown mineral occurrences based on association rules.
The authors tested their model by exploring the Tecopa basin in the Mojave Desert, a well-known Mars analog environment. The model was also able to predict the locations of geologically important minerals, including uraninite alteration, rutherfordine, andersonite, and schröckingerite, bayleyite, and zippeite.
In addition, the model located promising areas for critical rare earth elements and lithium minerals, including monazite-(Ce), and allanite-(Ce), and spodumene. Mineral association analysis can be a powerful predictive tool for mineralogists, petrologists, economic geologists, and planetary scientists, according to the authors.
Reference: “Predicting new mineral occurrences and planetary analog environments via mineral association analysis” by Shaunna M Morrison, Anirudh Prabhu, Ahmed Eleish, Robert M Hazen, Joshua J Golden, Robert T Downs, Samuel Perry, Peter C Burns, Jolyon Ralph and Peter Fox, 16 May 2023, PNAS Nexus.
DOI: 10.1093/pnasnexus/pgad110
