Skip to main content
U.S. flag

An official website of the United States government

20-34. Machine learning to enhance mineral resource knowledge for the United States


Closing Date: January 6, 2022

This Research Opportunity will be filled depending on the availability of funds. All application materials must be submitted through USAJobs by 11:59 pm, US Eastern Standard Time, on the closing date.



Global issues, technological advances, population increases, and increasing affluence are driving increasing needs for a range of Earth resources, from critical minerals to aggregates, that are necessary to meet supply shortfalls for the energy, industrial, and defense sectors of the Nation’s economy. Our knowledge-based society relies on data-driven Earth science to deliver fit for purpose solutions to an astonishing range of questions and problems at scales that range from nanometers to continental. Some of the data that are needed to solve these questions and problems must be newly acquired knowledge, but in many cases, answers rely, at least in part, on information in existing databases (e.g., federal and state survey databases). By automating data extraction from many databases and applying modern analytical approaches we will not only find new insights from the data, but will also improve our ability to predict the location of new mineral deposits. This critical information is needed to make informed decisions on utilization of the Nation’s mineral endowment.

The U.S. Geological Survey (USGS) is the Nation's largest natural resource and civilian mapping agency, delivering reliable and unbiassed scientific information to its customers and affiliates. A central part of the USGS’s mission is to deliver actionable intelligence at scales and timeframes relevant to decision makers.  There is a pressing need to develop improved methods to deliver rapid and authoritative information on both undiscovered and known mineral deposits. We seek a postdoctoral Mendenhall Fellow to conduct research using modern machine learning approaches to enhance delivery of mineral resource knowledge for the United States.

Machine learning is the use of computer algorithms that improve automatically through experience and using data. In Earth Science, there have been recent, successful applications of these methods for classification and prediction, such as the use of random forests, deep forests, gradient boosting machines, self-organizing maps, and other mineral prospectivity mapping techniques (e.g., Xiong and others, 2018; Occhipinti and others, 2020; Zuo, 2020, and references therein). These models characteristically integrate geological, geochemical, geophysical, remote sensing, and drilling data, and can be applied at a range of scales.

The Mendenhall postdoc would join the USGS mineral deposit database project (USMIN), which is building comprehensive 21st century geospatial databases that are the most authoritative sources of information about important mines and mineral deposits in the United States and its territories. The principal research opportunities lie in using machine-learning methods in novel ways to combine various datasets for classification, change detection, clustering, or forecasting/prediction to advance a broad range of Federal, State, non-governmental organizations, and private sector stakeholder priorities. This research opportunity requires a candidate with 1) experience in mineralogy and/or economic geology, 2) experience in data mining and data processing, 3) experience programming with the intent on developing analysis and predictive tools (Python or R is preferred), and 4) GIS skills particularly with ESRI software. In addition to the above, experience with big data and machine-learning analytics is a plus.

The project would explore and analyze existing data (e.g., USMIN, MINDAT, Mineral Resource Data System – MRDS, and others) to assess mineral deposit locations and historic records, identify key data gaps, and work with other researchers to identify the most important mineral deposit types for potential domestic critical mineral resources. In addition to performing research, the candidate would document their workflow and processes and, with guidance from the program, create custom tools, GUI based programs, scripts, and middleware that would systematize the workflows and processes required to identify, query, extract, categorize, analyze, and export the results from existing mineral databases.

We welcome submissions from scientists with a broad understanding of geology, data science, and mineral deposits. Some suggested research areas based on their relevance to advance the priority areas highlighted in the FY 2022 President's Budget are listed below.

Mine waste: The FY 2022 President's Budget recognizes the need for increased science in support of reclamation of abandoned mines on federal and non-federal lands. This effort is complex and the USGS must respond to stakeholders’ needs in a rapid and authoritative manner. Machine learning could be used to help identify abandoned mines and to help prioritize their remediation based on their potential physical and environmental hazards.

Waste as a resource: Recovery of associate (secondary or tertiary) minerals from mine waste represents a possible source of critical minerals to help meet the increasing needs of the U.S. Evaluation of whether critical minerals can be economically extracted from mine waste depends on accurately characterizing the mass of the mine waste, and what ore-bearing minerals remain in the waste. Automation of waste volume calculations and detection of discrepancies between reported production and measured waste using machine learning is another potential avenue for research, as is integrating other data such as remote sensing and lidar to explore for, characterize, and quantify waste materials.

Critical minerals: USMIN has published databases for 11 of the 35 minerals or mineral groups included in the federal critical minerals list, and databases for the remaining critical minerals will be published before December 2024. The applicant is invited to explore new and/or improved methods to capture critical minerals data more quickly and effectively from the many disparate primary sources where those data reside, which would produce tremendous efficiency gains.

Carbon sequestration: Most carbon capture and storage projects inject CO2 into sedimentary basins and require an impermeable cap-rock seal to prevent the CO2 from migrating to the surface. However, the most stable, long-term storage mechanism for atmospheric CO2 is the formation of carbonate minerals such as calcite, dolomite and magnesite, and this can be achieved either ex situ, as part of an industrial process, or in situ, by injection into geological formations where the elements required for carbonate-mineral formation are present. There is an opportunity to collaborate with state surveys and industry to use modelling techniques to quantify the mineral carbonation potential of different rocks in active mining areas, to help evaluate whether these processes could be economic.

Geophysical and Remote Sensing Data: The USGS is acquiring vast amounts of airborne magnetic, radiometric, and hyperspectral data over areas permissive for hosting undiscovered mineral resources as well as areas that have undergone historic mining. The use of known locations (e.g., historic mining locations) as training and testing points for supervised learning, has shown great progress in creating predictive maps. These results could be used to improve algorithms to map mine waste piles as a potential resource.

Interested applicants are strongly encouraged to contact the Research Advisor(s) early in the application process to discuss project ideas.


Occhipinti, S., Metelka, V., Lindsay, M., Aitken, A., Pirajno, F., and Tyler, I., 2020, The evolution from plate margin to intraplate mineral systems in the Capricorn Orogen, links to prospectivity: Ore Geology Reviews, v. 127, p. 103811.

Xiong, Y., Zuo, R., and Carranza, E.J.M., 2018, Mapping mineral prospectivity through big data analytics and a deep learning algorithm: Ore Geology Reviews, v. 102, p. 811-817.

Zuo, R., 2020, Geodata science-based mineral prospectivity mapping: A review: Natural Resources Research, v. 29, p. 3415-3424.

Proposed Duty Station: Lakewood, Colorado

Areas of PhD: Machine learning, computer science, data science, mathematics, statistics, geology, geochemistry, geophysics, remote sensing, geographic/cartographic sciences, hydrologic sciences, environmental geochemistry, or related fields (candidates holding a Ph.D. in other disciplines, but with extensive knowledge and skills relevant to the Research Opportunity may be considered).

Qualifications: Applicants must meet the qualifications for one of the following: Research Cartographer, Research Chemist, Research Geologist, Research Geophysicist, Research Hydrologist, Research Physical Scientist.

(This type of research is performed by those who have backgrounds for the occupations stated above. However, other titles may be applicable depending on the applicant's background, education, and research proposal. The final classification of the position will be made by the Human Resources specialist.)

Human Resources Office Contact: Sinar Santillano Oliveros, 303-236-9585,