Geology topics

Margaret A. Goldman

Publications and source records attributed to Margaret A. Goldman.

5 recordsLinked to original sources

USGS2025

Bayesian mapping of regionally grouped, sparse, univariate earth science data

Some earth science data are naturally grouped by region, and it is often desirable to map these data by region. However, if there are only a few samples within each region, then the map should be smoothed in an appropriate way to mitigate the problems that arise from having only a few samples. A smoothing algorithm based on a Bayesian hierarchical model is developed and presented in this report. This algorithm has several features that make it especially suitable for mapping earth science data: it can account for measurements that are censored, it can process multiple datasets with different measurement errors and different censoring thresholds, and it can calculate the uncertainty in any statistic that is mapped. The algorithm is demonstrated by mapping gold concentrations that are measured in streambed sediments in the Taylor Mountains quadrangle in southwestern Alaska.

Alaska↗

USGS2025

Extracting data from maps: Lessons learned from the artificial intelligence for critical mineral assessment competition

The U.S. Geological Survey (USGS), Defense Advanced Projects Research Agency (DARPA), NASA Jet Propulsion Laboratory (JPL), and MITRE ran a 12-week machine learning competition aimed at accelerating development of AI tools for critical mineral assessments. The Artificial Intelligence for Critical Mineral Assessment Competition solicited innovative solutions for two challenges: 1) automated georeferencing of historical maps, and 2) automated feature extraction from historical maps. Competitors used a new dataset of historical map images to train, validate, and evaluate their models. Automated georeferencing pipelines attained a median root-mean square error of 1.1 km. Prompt-based extraction (i.e., with user input) of polygons, polylines, and points from geologic maps yielded median F1-scores of 0.77, 0.56, 0.35, respectively. Geologic maps pose numerous challenges for AI workflows because they vary significantly. However, despite its short duration, the competition yielded promising results that have since spurred further innovation in this area and led to the development of new AI tools to semi-automate key, time-consuming parts of the assessment workflow.

Applied Computing and Geosciences↗

USGS2023

Tungsten resources of the northern Rocky Mountains, Montana and Idaho— A synthesis and quantitative assessment of skarn-hosted resources

Mineral resource assessments performed by the U.S. Geological Survey provide a synthesis of available information about the location of known and suspected mineral deposits. This study focuses on skarn-hosted tungsten resources in the northern Rocky Mountain region of east-central Idaho and western Montana which have seen moderate tungsten trioxide production in the past from a variety of mineralization styles including skarn, vein and replacement, and wolframite-quartz veins. The area’s geology is dominated by large Cretaceous and Tertiary plutons that are emplaced into a belt of Mesoproterozoic to Permian sedimentary rock and affected by tectonism related to the Sevier and later Laramide orogenies. Known tungsten skarn mineral sites are associated with contacts between Cretaceous plutons and calcareous and argillaceous sedimentary or metasedimentary rocks, including two skarn deposits in Montana (Calvert and Browns Lake) that are consistent with an updated grade and tonnage model. This study (1) delineates permissive tracts where undiscovered tungsten skarn deposits may occur within 1 kilometer of the surface; (2) presents a tungsten mineral site dataset from a variety of public sources; (3) evaluates currently available geochemical, geophysical, and radiometric age data in support of tract delineation; (4) provides probabilistic estimates of the amount of tungsten and tungsten-mineralized rock that could be contained in undiscovered deposits within one major tract; (5) estimates the value of total undiscovered deposits using economic filter analysis; and (6) provides a synthesis of metallogenic controls on regional tungsten skarn and granitoid-related mineral deposits. Two permissive tracts were delineated: the Great Falls tectonic zone (GFTZ)-Cretaceous tract, for which a quantitative assessment was performed, and the Bitterroot tract, which was assessed in a qualitative manner. The quantitative three-part assessment, conducted in August 2019, indicates that undiscovered tungsten resources might exist in skarn-type deposits within the study area. Using a negative binomial function, a mean of 4 undiscovered deposits was calculated from panel estimates. Simulation results that combine an updated grade and tonnage model with estimates of undiscovered deposits include the amounts of ore and contained tungsten trioxide at different levels of uncertainty. A mean of 250,000 metric tons and median of 200,000 metric tons contained tungsten trioxide was calculated for the undiscovered deposits within the GFTZ-Cretaceous tract. The value of undiscovered deposits was estimated using a new economic filter that considers factors such as mine type, deposit depth, deposit geometry, metallurgical recovery rate, cutoff grade, and tract area. A review of the regional Archean to Paleogene geology suggests that ore metal (copper, molybdenum, and tungsten) variations in intrusion-related deposits of Montana and Idaho may be controlled by a number of factors including the age and composition of underlying basement terranes, depth of emplacement, pluton chemistry and degree of fractionation, redox conditions, and aqueous fluid-melt partition coefficients.

Idaho, Montana↗

USGS2023

Automated georeferencing and feature extraction of geologic maps and mineral sites

The predictive power of mineral prospectivity analysis depends on high quality, spatially accurate, analysis-ready datasets. Of paramount importance are geologic maps and mineral site data, but the state of readiness for utilizing these datasets remains sub-optimal for advanced computational techniques. As the U.S. Geological Survey (USGS) fulfils its mission to map the distribution of critical mineral commodities, non-georeferenced maps held within historical collections represent rich sources of input data. Through a series of machine learning challenges organized by the Defense Advanced Research Projects Agency (DARPA) in collaboration with the USGS, significant progress has been made in accelerating data ingestion, processing, and preparation tasks that enable mineral prospectivity mapping and mineral resource assessment workflows. Specifically, two tasks that previously required time-intensive human effort, 1) georeferencing map images, and 2) legend-based feature extraction from map images, are discussed.

Conference Paper↗

USGS2020

User guide to the bayesian modeling of non-stationary, univariate, spatial data using R language package BMNUS

Bayesian modeling of non-stationary, univariate, spatial data is performed using the R-language package BMNUS. A unique advantage of this package is that it can map the mean, standard deviation, quantiles, and probability of exceeding a specified value. The package includes several R-language classes that prepare the data for the modeling, help select suitable model parameters, and help analyze the results. This user guide describes the BMNUS package and presents step-by-step instructions to model data that accompany the package.

Techniques and Methods↗