Industries / Mining and materials / Mining and quarrying / Salt and mineral extraction
Water and land
This work tracks what pumping brine, dissolving salt and mining soluble rock do to rivers, lakes, the ground and the people nearby, and sets limits on it.
5
Milestones achieved, from 1980s to 2022.
2
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1980s
Sinkholes on the Dead Sea
Sinkholes began to open along the Dead Sea coast in Israel and Jordan in the early 1980s. The Geological Survey of Israel counted more than 2,500 on the western shore by 2009, forming at 200 to 380 a year since 2003, and traced them to fresh groundwater dissolving a buried salt layer as the sea fell about 1 m a year. It says the fall of more than 30 m since the early 1930s reflects the diversion of Jordan River water and the large evaporation ponds of the mineral industries in Jordan and Israel. Along the ponds, where the water level has not fallen since the early 1970s, about 3 sinkholes form a year. The survey has published maps of where sinkholes are likely since 2004.
LessonWhere a water level falls over buried salt, fresh water can reach the salt and the ground above it can collapse.
Credit
Meir Abelson, Ran Gabay, Eyal Shalev and Yossi Yechieli, Geological Survey of Israel
2006
Flooded mine under a city
Potash and rock salt dissolve in water, so water that gets into a mine can widen its own path. In June 1986 water came through the roof of the Berezniki-3 potash mine in Russia's Upper Kama (Verkhnekamsky) deposit and ended in a large collapse. In October 2006 water broke into the Berezniki-1 mine, which lies almost entirely under the city of Berezniki, population 150,000; the mine flooded and closed, and at least five sinkholes formed, the largest 440 m by 320 m. Satellite radar measured the ground sinking by up to 17 cm a year from 2016 to 2020, checked against ground leveling surveys.
LessonA 2023 review of the deposit rates hydraulic backfilling, pumping a slurry into mined-out rooms, as the most effective protection against flooding and subsidence.
Credit
Vjacheslav Andrejchuk (study of the 1986 collapse); Sergey Samsonov and Alexandr Baryakh (radar measurements)
2021
Salt limits for the Werra
K+S's potash plants on the Werra River in central Germany release salty wastewater into the river. K+S says chloride reached 40,000 milligrams per liter (mg/L) in the 1970s and 1980s from plants in the former East Germany, and that a dry separation method, ESTA, which sorts crushed salt by electric charge instead of dissolving it, helped cut its wastewater from 20 million m³ to 7 million m³ between 1997 and 2015. In December 2021 the Kassel Regional Council allowed discharges until the end of 2027, with the chloride limit at the Gerstungen gauge falling from 2,400 mg/L in 2021 to 2,000 mg/L in 2022 and 1,700 mg/L for 2024 to 2027, and a cap of 5 million m³ of wastewater a year, down from 6.7 million. K+S stopped injecting salty wastewater into deep rock at the end of 2021.
Credit
Kassel Regional Council; K+S
2022
A regulator cuts brine pumping
In November 2016 Chile's environmental regulator, the Superintendencia del Medio Ambiente (SMA), opened a case against SQM Salar for six infractions at the Salar de Atacama, mainly pumping more brine than its permit allowed. A first compliance program worth about US$25 million was approved in January 2019. After a ruling by the First Environmental Tribunal in Antofagasta, the SMA approved a new program in August 2022, with 52 actions costing more than 46 billion Chilean pesos. It cuts brine pumping in steps to at most 822 liters per second from 2027, a little under half the rate the permit allowed, and fresh-water use to 120 liters per second, half the permitted flow. SQM has published its pumping and monitoring data online since June 2019.
Credit
Superintendencia del Medio Ambiente; First Environmental Tribunal of Antofagasta; complainants including the Atacameño communities of Peine and Camar
2022
Great Salt Lake record low
Companies on the Great Salt Lake in Utah take nearly 2 million tons of minerals a year, including salt, sulfate of potash and magnesium, from more than 85,000 acres of diked evaporation ponds, by the Utah Division of Water Resources' account. In November 2022 the lake fell to 4,188.5 ft at the South Arm causeway gauge, below the 1963 record of 4,191.35 ft. Science News reported in April 2023 that mineral extraction accounts for 9% of the water consumed, against about three-quarters for irrigation.
Credit
Utah Division of Water Resources; US Geological Survey, which has kept records of the lake since 1847
Watch for
AI reads radar for sinkholes
A deep-learning model, trained on the areas of sinking ground that the Geological Survey of Israel drew by hand on satellite radar images, picks out the ground movement that comes before sinkholes on the Dead Sea's western shore. The authors call their work the first comprehensive evaluation of such a model on this kind of radar data.
Achieved whenthe survey uses the model in its routine monitoring, or an independent group tests it on new radar images and publishes its hit and miss rates.
Claimed by
Gali Dekel, Ran Novitsky Nof, Ron Sarafian and Yinon Rudich, in the journal Remote Sensing, January 2026
Lithium without more lake water
EnergyX says a plant it plans with Compass Minerals at the Great Salt Lake will take out lithium "without incremental withdrawal from the Great Salt Lake," drawing first on brine already in Compass's ponds for salt, potash and magnesium chloride. The first phase targets up to 10,000 t a year and the whole plant 30,000 t a year.
Achieved whenthe plant runs and state water-use records show no added withdrawal from the lake for it.
Claimed by
EnergyX, in a memorandum of understanding with Compass Minerals, May 2026
Open
01
Test the lagoon finding openly
An open, repeatable test of whether brine pumping shrinks the protected lagoons of the Salar de Atacama in Chile, carried forward to 2026: monthly lagoon areas measured from satellite images, set against the pumping, fresh-water and well records SQM publishes, with code and data public.
The briefCredit · unclaimed02
One water count per salar
A public table of the water used by every lithium brine operation in production, counted the same way for each: fresh water drawn and brine pumped per tonne of lithium carbonate, with each figure linked to the impact report, permit or company report it comes from.
The briefCredit · unclaimed
Open 01 · the brief
Test the lagoon finding openly
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
An open, repeatable test of whether brine pumping shrinks the protected lagoons of the Salar de Atacama in Chile, carried forward to 2026: monthly lagoon areas measured from satellite images, set against the pumping, fresh-water and well records SQM publishes, with code and data public.
Why it is still open
A 2022 study by Juan Ignacio Guzmán and colleagues measured the main lagoon systems from 1986 to 2018 and found that brine extraction did not have a negative effect on their area. A 2025 study by Andreas Link and colleagues says changes in lagoon areas and in the mixing of brine and fresh water "are not yet comprehensively captured by existing models." Searches for this brief found no open code or data that would let others repeat the 2022 test with newer records.
Built with what exists
SQM's public monitoring site, which gives histories for 325 monitoring points, including lagoon surface, brine extraction and fresh-water intake; Sentinel-2 and Landsat images through Google Earth Engine; Python; MODFLOW 6, the US Geological Survey's free groundwater model.
First step
Download SQM's published lagoon-surface records for one lagoon and measure the same lagoon from Sentinel-2 images for one year, to see whether the two agree.
Done looks like
Monthly areas for the main lagoon systems from 2016 to 2026; the 2022 study's yearly change in total lagoon area for 1986 to 2018 (between a 0.03% loss and a 0.01% gain a year) reproduced from public images; and a model fitted on data up to 2022 that predicts 2023 to 2026 lagoon areas with its error stated, all with code and data public.
Open 02 · the brief
One water count per salar
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A public table of the water used by every lithium brine operation in production, counted the same way for each: fresh water drawn and brine pumped per tonne of lithium carbonate, with each figure linked to the impact report, permit or company report it comes from.
Why it is still open
A 2025 study by Walter Fernando Díaz Paz and colleagues counted water for two Argentine operations only, Olaroz and Fénix. From their environmental impact reports for 2021 it found a water footprint of 51.0 m³ and 135.5 m³ per tonne of battery-grade lithium carbonate, plus 537.4 m³ and 319.6 m³ of brine. The authors found discrepancies in the information the reports give, and write that baseline data are "often either unavailable or not freely accessible." Figures in the literature they review run from about 5 m³ of fresh water per tonne to around 2,000 m³ in total, depending on what is counted.
Built with what exists
Environmental impact reports of the kind the 2025 study used; Chile's environmental assessment system (SEIA); company sustainability reports; the water footprint method of Hoekstra and colleagues, which the 2025 study followed; a large language model such as Claude to pull figures from long reports, with a person checking each one (see the Note "Seven more flagged"); a spreadsheet.
First step
Get the impact reports for Olaroz and Fénix, reproduce the 2025 study's figures, then add a third operation.
Done looks like
At least 10 brine operations in at least three countries, each with fresh water and brine per tonne for the same year and the same boundary, every value linked to its page in a source, and the 2025 study's figures for Olaroz and Fénix reproduced within 10%.