Vesuvius in AD 79 now anchors a sharper clock for Earth’s past

Paul Renne’s laboratory at the Berkeley Geochronology Center took eight crystals blasted out of Mount Vesuvius in AD 79 and measured their age to a precision of 0.7 percent and an accuracy of 0.4 percent. The eruption, which buried Pompeii, is among the few ancient geological events with a date written down by a witness, and that makes it a rare test of a dating method meant for rocks billions of years old.

The work was published September 25, 2026 in Science Advances, with colleagues from UC Berkeley and the University of Padua in Italy, and its co-authors include Caroline Hasler, Bill Cassata, Jack Carter and Andy Tholt. Its aim is a better ruler for volcanic and geological history, not a better date for Pompeii.

Eight Sanidine Samples From Oplontis

The team used argon-argon dating, which tracks potassium-40 decaying into argon-40 inside volcanic crystals. Gizmodo’s report explains the logic: in hot magma, argon escapes; once the rock is erupted and cooled, argon builds up, and measuring it gives the time since eruption. The samples were sanidine crystals, a potassium-rich feldspar, from eruption deposits.

The Brighter Side reported that the crystals came from pumice at Oplontis, and that the measurements comprised 153 incremental heating steps drawn from about 700 milligrams of mineral. The result: 1,938 years, give or take 13, before the 2025 analysis, which translates to 87 CE plus or minus 13 years. That interval overlaps 79 CE, and Gizmodo describes the match as accurate to within a decade.

Pliny’s August 24 Date as the Fixed Point

The reason this works is that Vesuvius has an independent answer. The Debrief reported that the team used Pliny the Younger’s eyewitness account to establish August 24, 79 CE as the eruption date, then calibrated the argon-argon method against it. Graduate student Caroline Hasler handled the historical side, including analysis of coins found at the site. Some historians favor a later date in autumn, partly because of a coin from Pompeii; Gizmodo reports that Hasler’s analysis indicated the coin was likely produced before September, which supports the August timeline.

Renne told Gizmodo the aim was to test how close the method could get on something geologically very recent; in geochronology, he noted, 79 CE is “like yesterday.” The point is that a method used on rocks millions or billions of years old is rarely tested against a known calendar date, because nothing that old has one. Vesuvius supplies a rare overlap, a rock young enough to have a written date and old enough that the potassium-40 clock has ticked a measurable amount, so any systematic error in the method shows up as a gap between the argon age and the calendar.

Potassium-40 Half-Life and Linked Clocks

The calibration feeds back into the physics. According to The Debrief, the team revised the half-life of potassium-40 so that it is twice as precise as the original nuclear-physics value. The Brighter Side describes it as an improved decay constant, the quantity that anchors every argon-argon age. A better constant sharpens every age computed with it, including those of volcanic layers far older than Vesuvius.

The payoff reaches beyond argon. The Brighter Side says the researchers envision linking uranium-lead dating for deep time, argon-argon dating for volcanic deposits and radiocarbon dating for recent archaeological records, using volcanic layers that contain both datable minerals and charred organic material. A shared anchor in such layers can tie the three systems to one timeline. Renne, quoted there, said that when the aim is to reconstruct the eruptive history of a volcano in relatively recent time, precision and accuracy really count. The Debrief lists the improved dating among the tools for assessing volcanic risk around cities including Naples, Yogyakarta and Mexico City.

Renne has done this before, with a coarser result. A 1997 UC Berkeley release reported that argon-argon dating placed the Vesuvius eruption about 1,925 years before then, only seven years off the historical record, and quoted Renne saying the first attempt landed within five percent of the date, with hopes of cutting the error to one percent or less. The Berkeleyan’s account put the statistical error at plus or minus 94 years. The 2026 result carries an uncertainty of 13 years, roughly seven times tighter on the same volcano, and the work in 1997 and in 2026 both came out of Renne’s group and both tested the method on the same event, which makes the comparison direct.

The Berkeley Geochronology Center, which describes itself as the only institution wholly dedicated to deciphering the history of Earth and nearby planets, is funded by private donations and NSF grants. What has been published so far is a calibration, and the 0.7 percent figure belongs to eight Vesuvius samples, not to every rock the method will be asked to date.

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

More from Morning Overview