Pompeii’s eruption let geologists set their atomic clock to 0.4% accuracy

Pliny the Younger’s eyewitness account of Vesuvius gave geologists a date they could trust to the day, and a team led by Paul Renne of the Berkeley Geochronology Center has now used it to tune argon-argon dating to 0.4% accuracy. The test case was the August 24, 79 CE eruption that buried Pompeii and its neighbors. The paper, titled “Pliny the Younger advances geochronology,” appeared in Science Advances, volume 12, issue 39, and Berkeley published its account of the work on September 25, 2026.

The result matters because argon-argon dating is normally checked against other rocks, which carry their own uncertainties, and because the recent end of the timescale, a few thousand years back, is exactly where precision is hardest to demonstrate. A volcanic eruption with a date written down by a witness is a rare exception, and it let the team grade the clock against an answer already known.

Eight sanidine crystals from Oplontis

The samples came from Oplontis, a Roman town buried alongside Pompeii. According to UC Berkeley’s own account of the work, the pumice was collected in 1998 and analyzed only recently, and it came from the earliest phase of the eruption, when potassium-rich magma emerged first. PopSci’s coverage of the paper reports that the age rests on eight sanidine crystals, a potassium-bearing feldspar that holds argon well.

The method counts argon-40, the product of potassium-40 decay. Heat from an eruption drives argon out of the crystals, so the clock restarts at the eruption, and the argon that accumulates afterward measures the time elapsed.

PopSci adds that the improved clock bears on radiocarbon dating too, which is used for material younger than about 55,000 years. The Debrief reports that the calibration also helps line up methods such as argon-argon, uranium-lead and carbon-14 against each other, and that a tighter timeline helps judge cause and effect between geologic events, for instance a volcanic episode and a mass extinction.

1,938 years against 1,946

Berkeley reports that the team’s age came out at “1,938 ±13 years prior to when the minerals were analyzed in 2025.” Pliny’s account puts the eruption 1,946 years before the same 2025 date. The eight-year gap between the two is the 0.4% accuracy, and the 13-year spread is the 0.7% precision. The same release credits the improvement to the better Oplontis pumice.

Those two numbers describe different things. Accuracy asks how close the measured age sits to the true one, and precision asks how tightly the measurements cluster around their own average. The 0.4% figure applies to this calibration against a documented eruption, as the team tested it. It is not a promise that every age the method produces will land within 0.4% of the truth.

The Debrief’s account of the study adds that the team used Pliny’s letter to fix the date of August 24, 79 CE, then used that benchmark to recalibrate the method, which it tested against eight samples of known age.

Half-life, carbon-14 and three volcanic cities

The calibration also tightens a constant that dating depends on. The paper’s summary of results gives a potassium-40 half-life of 12.044 ± 0.088 billion years, a value the study describes as nearly twice as precise as the best previous direct experimental measurement. The release lists Renne’s collaborators as Caroline Hasler, William Cassata, Jack Carter, Anthony Fuentes and Andrew Tholt, with Andrea Marzoli of the University of Padua; funding came from the National Science Foundation, the Ann and Gordon Getty Foundation and the Berkeley Geochronology Center.

Renne founded the center in 1994 and directs it, and he is also a professor in residence in UC Berkeley’s Earth and Planetary Science Department, according to the center’s profile. His stated case for the work is practical: in his words, building the eruptive history of a volcano over relatively recent time depends on precision and accuracy both.

Berkeley’s release names Naples, Mexico City and Yogyakarta as places where volcanic history bears on risk, and it says a better-calibrated clock can also be used to cross-check carbon-14 and uranium-lead dating. IFLScience, in its report on the paper, notes that the technique can now be pointed at prehistoric events with no written record at all.

The Vesuvius test is the one case where the answer was already written down, and the clock matched it within eight years. Berkeley’s release stresses the reach of the result: the same potassium-40 clock runs from eruptions of a few thousand years ago to rocks billions of years old, which is why a single Roman summer matters to work on far older events: the half-life figure is shared across all of it, and the Vesuvius calibration is what tightened that figure.

This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.

More from Morning Overview