120-18 What’s the Point? Quantifying the Spatial Accuracy of Mineral Site Locations to Support High Standard Critical Mineral Assessments

Abstract:

Understanding the quantity and distribution of domestic critical mineral resources is an important step in ensuring future supply, informed U.S. Government policymaking, and international relations. Through the Earth Mapping Resources Initiative (Earth MRI) and other efforts, the U.S. Geological Survey (USGS) and state partners are collecting and compiling new high-resolution geophysical, geochemical, spectral, and geological data to support efforts to map and quantify critical mineral resources. These data can be used to construct mineral prospectivity models to aid USGS assessments of critical minerals. However, training and validating mineral prospectivity models requires data on known mines, deposits, and occurrences – collectively called mineral sites.  Most mineral site data are held in legacy datasets, such as the USGS Mineral Resource Data System (MRDS). Although it is generally understood that the locations of mineral sites in MRDS are often approximate, no study has attempted to quantify its spatial uncertainty. This study addresses that gap to better evaluate the level of effort, and justification, required to improve and verify mineral site data for critical mineral assessments.

We analyzed mineral site information from two independent mineral resource assessments: 1) pegmatite-related mineral sites in the Black Hills, South Dakota, and 2) a national dataset of graphite mineral sites. Both studies deduplicated MRDS records and improved locations for mineral sites using historical documents, topographic map-derived data, and modern digital elevation models. We compared the improved mineral site data to legacy coordinates to quantify spatial accuracy. Preliminary results showed that in the Black Hills (374 sites) the spatial deviation in the legacy locations had a median of 250 m, and a first-quartile (Q1) and third-quartile (Q3) of 100 m and 475 m, respectively. The national graphite dataset (516 sites) had similar distances of 350 m (median), 200 m (Q1), and 625 m (Q3). For comparison, mineral prospectivity mapping uses modern high-resolution datasets such as 1-meter lidar, 50-meter geophysical data, and 1:100,000-scale geologic mapping with ca. 50-meter resolution. These results demonstrate that less than 25% (i.e., Q1) of the studied mineral sites are accurate to within 100 m. Map scale and data resolution influence how mineral prospectivity models are fit and utilized. Improving mineral site locations can enhance model performance because it ensures that the target variables are aligned with high-resolution datasets.

Geological Society of America Abstracts with Programs. Vol. 58, No. 5, 2026

doi: 10.1130/abs/2026AM-18816

© Copyright 2026 The Geological Society of America (GSA), all rights reserved.