T5 Innovations in Geological Mapping

Sunday, 22 September 2024
8:00 AM - 12:00 PM
206B

CRITICAL MINERAL ASSESSMENTS WITH AI SUPPORT (CRITICALMAAS): A COLLABORATIVE EFFORT TO MODERNIZE USGS MINERAL RESOURCE ASSESSMENTS

Paper number: T5-5
Program: Technical Sessions
Time: 9:00 AM – 9:15 AM

Authors

  • Joshua Rosera Presenting Author
    Geology, Energy & Minerals Science Center, Reston, VA
  • Graham Lederer
    Geology, Energy & Minerals Science Center, Reston, VA
  • Margaret Goldman
    Geology, Geophysics, and Geochemistry Science Center, Denver, CO
  • Garth Graham
    Geology, Geophysics, and Geochemistry Science Center, Denver, CO

Abstract

The U.S. Geological Survey (USGS) is tasked with assessing undiscovered resources for 50 critical mineral materials, including those needed for the energy transition. The existing methodology for mineral resource assessments relies on manual processes of data gathering, preparation, and spatial analysis. Although these methods are well established, they are too time-intensive to deliver actionable information to stakeholders, and their reliance on manual processes poses challenges for reproducibility. To further compound the challenge, many critical minerals were not the focus of previous assessments conducted over the last 40 years, and information about their ore grade, tonnage, or by-product concentrations is not currently compiled into analysis-ready formats.

The USGS has partnered with the Defense Advanced Research Projects Agency (DARPA) and the Advanced Research Projects Agency – Energy (ARPA-E) on a collaborative program to accelerate the time-consuming aspects of mineral resource assessments and make the results more reproducible. The goals are approached through human-centered artificial intelligence engineering and automation. The program is divided into four technical areas (TAs): 1) automation to georeference map images and extract feature points, lines, and polygons; 2) data engineering of mineral occurrence information, including compilations of deposit grade and tonnage data; 3) supervised and unsupervised mineral prospectivity modeling; and 4) human-in-the-loop tools for data discovery, enhancement, model feedback, and storage. Preliminary results from TA1 demonstrate improvement in automated georeferencing and feature extraction relative to similar models developed in DARPA’s 2022 challenge. A knowledge-graph developed by TA2 can be queried to build preliminary grade and tonnage datasets and to extract training data for mineral prospectivity modeling (TA3). Experiments integrating data from TA1 and TA2 generated mineral prospectivity maps for magmatic Ni-Co and Mississippi Valley-type Zn-Pb±Ga±Ge deposits in the United States. Highly prospective areas align well with findings from prior studies and can be used as a basis to rapidly generate initial permissive tracts for assessment teams. These tools are an advancement towards more efficient and reproducible assessments needed to guide land use and policy decisions.

Geological Society of America Abstracts with Programs. Vol. 56, No. 5

doi: 10.1130/abs/2024AM-403292

© Copyright 2024 The Geological Society of America (GSA), all rights reserved. Permission is hereby granted to the author(s) of this abstract to reproduce and distribute it freely, for noncommercial purposes. Permission is hereby granted to any individual scientist to download a single copy of this electronic file and reproduce up to 20 paper copies for noncommercial purposes advancing science and education, including classroom use, providing all reproductions include the complete content shown here, including the author information. All other forms of reproduction and/or transmittal are prohibited without written permission from GSA Copyright Permissions.