BRENNEN

Industries / Mining and materials / Mining and quarrying / Salt and mineral extraction

Finding deposits

Finding deposits is the search for buried beds of salt minerals and for brines, underground water heavy with dissolved salts, and the work of measuring how much of each mineral they hold.

5

Milestones achieved, from 1937 to 2025.

1

To watch for: announced, not yet confirmed outside.

2

Open deliverables. Nobody has shipped them.

The road

AchievedWatch forOpen

Achieved

  1. 1937

    Trona in a gas well

    In October 1937 Mountain Fuel Supply Company took core samples from its John Hay #1 well in southwest Wyoming while drilling for oil and gas. The core held thick zones of trona, a mineral refined into soda ash for glass, and other salt minerals, some of which, such as shortite, were identified for the first time. The US Geological Survey published a reference to the minerals in 1940.

    main.wsgs.wyo.gov

    Credit

    Mountain Fuel Supply Company; the US Geological Survey

  2. 1943

    Potash in an oil well

    Potash, a family of potassium salts used as fertilizer, was recognized in 1943 in core from a hole that Norcanols Oil & Gas, a subsidiary of Imperial Oil, drilled near Radville in southeastern Saskatchewan. In 1946 another oil well near Unity, in west-central Saskatchewan, brought up potash core grading 21.6% K2O (potassium oxide, the unit potash is sold by) at a depth of 3,466 feet. The Encyclopedia of Saskatchewan dates the first find to 1942.

    LessonA rich core is not yet a mine, and the first shaft sunk at Unity was lost to water (see Rock salt and potash mines).

    magazine.cim.orgesask.uregina.ca

    Credit

    Norcanols Oil & Gas (Imperial Oil)

  3. 2012

    Rules for counting brine

    Lithium brine sits in the pores of buried sand and salt, and it moves and mixes as it is pumped. In 2011 Jon Houston and four co-authors argued in Economic Geology that the codes for reporting mineral resources, written for solid ore, needed changes for brine: the estimate depends on how much brine the rock will give up when pumped, which is harder to measure than total porosity, and recovery is low. On 1 November 2012 the Canadian Institute of Mining, Metallurgy and Petroleum (CIM) published best-practice guidelines for reporting lithium brine resources and reserves.

    LessonA deposit that flows has to be measured by what a well can pump out, not only by what a sample holds.

    doi.orgmrmr.cim.org

    Credit

    J. Houston, A. Butcher, P. Ehren, K. Evans and L. Godfrey (paper); CIM's sub-committee on lithium brines (guidelines)

  4. 2024

    Lithium mapped by machine learning

    The US Geological Survey (USGS) and the Arkansas Office of the State Geologist trained a random forest model, a machine-learning method that combines many decision trees, on old and newly collected brine samples plus geology and temperature data. The model predicts lithium concentrations across southern Arkansas, including where no samples exist. With the rock's porosity, the team estimates 5.1 to 19 million metric tons of lithium in brines of the Smackover Formation, equal to 35 to 136% of the current US lithium resource estimate. This is lithium in place, not what can be recovered.

    LessonBrine samples from the 1960s with more than 500 mg/L of lithium could not be verified, and new sampling near the same locations did not find such values, so old analyses need re-sampling before anyone builds on them.

    usgs.govpubs.usgs.gov

    Credit

    Katherine Knierim and colleagues, USGS, with the Arkansas Office of the State Geologist (Science Advances, September 2024)

  5. 2025

    Lithium held in the rock

    A team including Lawrence Berkeley National Laboratory measured lithium in brine and rock from the Salton Sea geothermal field in California, using a scientific drill core (California State 2-14) and commercial wells. They give the brine reservoir's inventory as 4 to 18 million metric tons of lithium carbonate equivalent (the weight of lithium carbonate the lithium would make), and report an even larger amount held in the minerals of the reservoir rock, with the highest contents in chlorite, a mica-like mineral.

    osti.gov

    Credit

    J. Humphreys, M. Brounce, M. A. McKibben, P. Dobson, N. Planavsky and B. Kalderon-Asael (Economic Geology, 2025), funded by the US Department of Energy

Watch for

  1. A second Smackover lithium field

    Smackover Lithium, a partnership of Standard Lithium (55%) and Equinor (45%), reports an indicated resource (sampled closely enough to plan with) of 1.77 million metric tons of lithium carbonate equivalent at an average 562 mg/L of lithium, and an inferred resource (more loosely sampled) of 2.13 million metric tons at 572 mg/L, for its Franklin project in East Texas. The upper-zone grades are the average of samples from three exploration wells taken in 2025 and 2026. The qualified persons who signed the estimate are independent of the company under Canada's disclosure rule NI 43-101. A preliminary economic study models up to 70,000 t a year of lithium carbonate from the early 2030s.

    Achieved whenbrine from the project is pumped at production rates and the lithium output is reported, so the grades can be checked against the samples.

    standardlithium.com

    Claimed by

    Standard Lithium and Equinor, September 2026

Open

  1. 01

    Score the lithium map

    A public test of the USGS lithium map for the Smackover Formation. Run the survey's published model unchanged, then compare its predicted lithium at each well with lithium measured in brine samples taken after the model was built. Use wells inside southern Arkansas where they exist, and wells across the state lines in East Texas and northern Louisiana as a test outside the model's area.

    The brief
    Credit · unclaimed
  2. 02

    Check old brine readings

    An open screening tool, and a flagged copy of the brine lithium records for one basin in the USGS produced-waters database. Each record gets a flag for each known way a reading goes wrong: a new well sampled before its drilling fluids were flushed out, a well treated with fresh water or chemicals, a nearby water-injection well in the same zone, water condensed out of gas, fluid from the wrong zone, or a lab value that does not fit the sample's salt content.

    The brief
    Credit · unclaimed

Open 01 · the brief

Score the lithium map

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A public test of the USGS lithium map for the Smackover Formation. Run the survey's published model unchanged, then compare its predicted lithium at each well with lithium measured in brine samples taken after the model was built. Use wells inside southern Arkansas where they exist, and wells across the state lines in East Texas and northern Louisiana as a test outside the model's area.

Why it is still open

The USGS says its predictions, made on a grid of about 1.5 square miles, are "unsuitable for quantifying lithium at specific locations". A 2025 conference abstract by Benjamin Rostron (University of Alberta) and Gavin Jensen (Saskatchewan Energy and Mines) warns that inaccurate lithium values "are being further incorporated into source studies, exploration models, and resource estimates". Searches for this brief found no published comparison of the map with later samples.

pubs.usgs.govgsameetings.secure-platform.com

Built with what exists

The USGS data release for the map (DOI 10.5066/P9QPRYZN), which holds the input files, Python and R scripts and prediction grids; lithium values from company technical reports filed on SEDAR+ in Canada and EDGAR in the US; QGIS; Python.

  1. First step

    Rerun the data release scripts until they reproduce the published map, then list every Smackover brine lithium value published since September 2024 with its well location.

  2. Done looks like

    At least 20 brine samples taken after 2022, each listed with its source, the model's low, median and high prediction for that spot, and the measured value, with the share that falls inside the predicted range and the median error in mg/L.

Open 02 · the brief

Check old brine readings

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

An open screening tool, and a flagged copy of the brine lithium records for one basin in the USGS produced-waters database. Each record gets a flag for each known way a reading goes wrong: a new well sampled before its drilling fluids were flushed out, a well treated with fresh water or chemicals, a nearby water-injection well in the same zone, water condensed out of gas, fluid from the wrong zone, or a lab value that does not fit the sample's salt content.

Why it is still open

Rostron and Jensen list these well-side problems and the difficulty of measuring tens of mg/L of lithium in brine with more than 300,000 mg/L of dissolved solids. They say the errors can give values "significantly higher, or lower, than actual" and that numerous examples exist in the current literature. The USGS could not verify 1960s Smackover readings above 500 mg/L.

gsameetings.secure-platform.compubs.usgs.gov

Built with what exists

The USGS National Produced Waters Geochemical Database (version 3.0, December 2023); well records from state oil and gas regulators; Python; a language model to read scanned source reports, with a person checking each result (see the Note "Seven more flagged").

  1. First step

    Download version 3.0 of the database, pull every Smackover record with a lithium value, and plot lithium against total dissolved solids to find the outliers.

  2. Done looks like

    Every lithium record above 50 mg/L in one basin flagged with its reasons, and the flags tested on at least 30 old records that were later re-sampled, reporting how many of the records that failed re-sampling the flags caught.