How Scientists Date Volcanoes: A Guide to the Clocks Behind Eruption Forecasts

Volcanoes kill on their own schedules, and the only way to anticipate those schedules is to know when they erupted before. A sweeping new review published in the E&G Quaternary Science Journal by William McCreary and colleagues at the University of Freiburg, together with collaborators in Australia and Switzerland, takes stock of the numerical dating methods that underpin every eruption chronology of the last 2.6 million years, the Quaternary period. The review is unusual in focusing squarely on volcanism, and its message is sobering: the reliability of volcanic hazard forecasts rests entirely on the quality of dates that many hazard modellers never learn how to scrutinise.

The stakes are far from academic. Cities such as Auckland, Mexico City, Naples and Clermont-Ferrand sit inside or beside volcanic fields where future eruptions are expected but their timing is unknown. Long-term forecasting also shapes decisions as consequential as where to site nuclear waste repositories. Yet the chronological record is startlingly thin. A survey of the southwestern United States found that of more than 2,000 identified Quaternary volcanoes in 37 fields, only about 25 percent have been dated at all, and that figure is comparatively good by global standards. Many published ages come without analytical details or error estimates, or were produced by methods since shown to be flawed.

The review concentrates on dispersed, monogenetic volcanism, in which each small batch of magma, typically less than a cubic kilometre, ascends along a fresh conduit and erupts in a new location over less than a century. Individual volcanoes are short-lived, but a field can remain active for millions of years, and because each new vent opens in a new place, hazard assessment demands spatiotemporal models of the entire field. All the dating methods covered exploit one of the few natural processes that run at constant rates over geological time: radioactive decay. Four of them date volcanic products directly, potassium-argon and argon-argon dating, uranium-series methods, cosmogenic nuclide exposure dating and luminescence dating, while radiocarbon dates organic material killed or bracketed by eruptions.

Argon dating is the workhorse for lava. Potassium-40 decays to the noble gas argon-40 with a half-life of 1,250 million years, and once a potassium-bearing mineral cools below its closure temperature, radiogenic argon becomes trapped in the crystal lattice. The classic K/Ar method measures potassium and argon on separate aliquots, which biases results in heterogeneous samples. The argon-argon variant sidesteps this by converting potassium-39 to argon-39 through neutron irradiation in a reactor, then measuring all argon isotopes at once. Step heating, in which gas is released in increments, reveals whether a sample has remained closed since eruption: an ideal sample yields a plateau age across at least three contiguous steps representing over half the released argon-39. In Argentina’s Payenia Basaltic Province, one basalt flow produced a textbook plateau and an atmospheric argon ratio consistent with the present-day value of 295.5, while a second, apparently very young sample showed a poorly constrained isochron, illustrating how close young basalts push the method’s roughly 1,500-year lower limit.

The method’s power is best shown at the other end of the precision spectrum. Renne and colleagues dated sanidine phenocrysts from the pumice of Vesuvius’s famous 79 CE eruption, documented by Pliny the Younger, and obtained 1.925 plus or minus 0.094 thousand years, agreeing with the historical record despite hints of excess argon. In Australia’s Newer Volcanic Province, argon-argon ages reproduced older K/Ar results but with far better precision, and some ages came out significantly younger, exposing the excess-argon problems the older method could not detect. Even so, the authors caution that ignimbrites, altered lavas and glasses often carry anomalously high atmospheric argon, and that argon-39 recoil during irradiation can distort fine-grained samples.

Radiocarbon dating, by contrast, never dates the eruption itself. It dates organisms, trees burnt by lava, charcoal in pyroclastic flows, or peat layers bracketing a deposit, whose death coincided with or bounds the event. Carbon-14 decays with a half-life of 5,730 years, setting an upper limit near 55,000 years, and accelerator mass spectrometry now delivers sub-percent precision on milligram samples. But the atmosphere’s carbon-14 production has never been constant, so laboratory ages must be calibrated against curves built from tree rings, varved sediments, corals and ice cores. Reservoir effects complicate matters further: magmatic carbon dioxide contains no carbon-14 at all, and modern leaves around the Solfatara crater in Italy’s Campi Flegrei show radiocarbon ages spanning 5,000 years. At Taupo in New Zealand, such magmatic bias can overestimate eruption ages by more than 200 years.

The most famous radiocarbon controversy involves Santorini. Archaeological chronologies place the great Late Bronze Age eruption around 1520 BCE, but more than 100 radiocarbon measurements on short-lived material, plus a dated olive branch from the island itself, point to roughly 1627 to 1600 BCE, a century or more older. Decades of debate have implicated magmatic degassing, calibration-curve anomalies and archaeological assumptions, and the review presents it as a cautionary tale: even a mature method with exquisite precision can yield contested ages when calibration curves are imperfect. Meanwhile, applied work keeps proving its worth. At Vulcano in Italy, radiocarbon combined with tephrostratigraphy revealed at least 19 eruptions between 900 and 1550 CE, one every 35 years, far more frequent than the previously assumed 130-year recurrence. High-density dating of sediments from Lake Motosu similarly refined Mount Fuji’s Holocene history and uncovered previously unknown eruptions.

Uranium-series methods exploit the decay chains of uranium-238, uranium-235 and thorium-232, in which intermediate isotopes fall out of secular equilibrium during melting and crystallisation and then rebalance at known rates. Plotting mineral phases on an isochron yields crystallisation ages, as demonstrated for the Puy de Dôme in France, where a 12.1 plus or minus 1.0 thousand-year age agrees with earlier luminescence and radiocarbon results. Because crystals can linger in a magma chamber long before eruption, these ages are best treated as maxima. The zircon double-dating technique closes that gap by applying both uranium-thorium disequilibrium and helium accumulation to the same crystals, delivering eruption ages from 2 to 1,000 thousand years. At Las Tres Vírgenes in Baja California, this approach produced a corrected eruption age of 36 plus or minus 6 thousand years, overturning a lone radiocarbon date and an ambiguous eighteenth-century historical reference.

Cosmogenic nuclide exposure dating reads a different clock: the buildup of rare isotopes such as helium-3, neon-21, chlorine-36 and beryllium-10 as cosmic-ray secondaries bombard freshly exposed rock surfaces. New lava flows are ideal because they start with no inherited signal, though identifying the original surface and scaling production rates across latitude and altitude remain tricky, especially with few calibration sites in the Southern Hemisphere. Applications range from helium-3 and neon-21 dating of olivine in Argentine basalts, confirming activity as recent as 5,000 years ago, to chlorine-36 dating of dacitic flows on Pico de Orizaba in Mexico, where high chlorine content pushed uncertainties to nearly 20 percent. Luminescence dating, finally, measures trapped charge in quartz and feldspar, reset by heat or light and rebuilt by environmental radiation. It can date xenoliths baked by magma, sediments beneath lava flows, and even fragments shattered by phreatic explosions, as shown for the Eifel maars in Germany, though volcanic minerals are notoriously prone to anomalous fading that protocols like post-infrared infrared stimulated luminescence are designed to avoid.

The review’s overarching advice is blunt: never take a published volcanic age at face value. Each method dates a slightly different moment, cooling, exposure, crystallisation or death, and the most trustworthy chronologies are those reproduced by two or more independent techniques. Detailed reporting of sampling, procedures and stratigraphy, the authors argue, is what allows ages to be re-evaluated decades later. As analytical instruments improve and more eruptions worldwide are dated, the foundation for forecasting the next eruption of an Auckland or a Naples grows firmer, one carefully vetted number at a time.

Subject of Research: Numerical dating methods for establishing eruption chronologies of Quaternary volcanism

Article Title: Review of numerical methods for dating Quaternary volcanism

Article References: McCreary, W., May, V., Mueller, D., Serra, E., & Preusser, F. (2026). Review of numerical methods for dating Quaternary volcanism. E&G Quaternary Science Journal, 75(1), 131-162. https://doi.org/10.5194/egqsj-75-131-2026

Image Credits: AI Generated

DOI: 10.5194/egqsj-75-131-2026

Keywords: volcanism, geochronology, argon-argon dating, radiocarbon dating, uranium-series dating, cosmogenic nuclide exposure dating, luminescence dating, volcanic hazard, Quaternary, tephrochronology, zircon double dating, eruption recurrence