Tiny and contentious “crystal clocks” in volcanic rocks could aid early warning for eruptions

Mount Taranaki is a sleeping giant. The volcano, New Zealand’s second-highest peak, rises some 2,500 meters from foggy lowlands to a snowcapped summit. It last oozed hot rocks and gases in the mid-19th century, and it’s anyone’s guess when it might erupt again. But the very rock of this place, the ridges and boulders that beckon 20,000 visitors a year, could hold clues about the timing of Taranaki’s next eruption.

Tiny crystals embedded in volcanic mountains, including Taranaki in New Zealand, pictured here, could hold secrets to magma’s behavior underground before eruptions. Image credit: Shane Cronin (University of Auckland, Auckland, New Zealand).

That’s thanks to tiny crystals embedded in the volcanic mountainside. Alpinists anchoring their ropes into Taranaki’s sheer walls can see glassy-white flecks of crystallized minerals in the gray lava rock. Under a microscope, the flecks look like tiny tree rings, with concentric layers that grew while the rock was deep underground and still molten.

Over the past few decades, geologists have learned how to read these rings. By studying how long the crystals take to form, volcanologists are calibrating so-called “crystal clocks” that reveal how long magma can linger underground before erupting. That might tell climbers how much time they have between the first grumblings of a waking volcano and a massive eruption.

Last year, a study of crystal clocks from Mount Taranaki suggested that warning time could be terrifyingly short, perhaps less than a week, says volcanologist Fidel Costa at France’s Université Paris Cité, part of the team behind the study (1). That prediction is helping GNS Science—New Zealand’s equivalent of the US Geological Survey—develop an improved eruption model for Taranaki.

Yet, there is intense debate about the accuracy of crystal clocks. Various laboratories have found that the same clocks can apparently “tick” at dramatically different rates, raising uncertainties about how reliable they might be as a forecasting tool. Researchers are making a concerted effort to refine the technique and improve its precision. “We want an improved tool, but there will never be a universal tool,” cautions Phil Shane at the University of Auckland in New Zealand, who co-authored the recent Taranaki study. “We should be more upfront about all the uncertainties.”

Under a microscope, this plagioclase crystal in a rock erupted from the Taranaki volcano appears to have something like tree rings. In fact, the growth bands record progressive changes in the magma as it grows. In this case, the white bands are rich in calcium and magnesium, and they formed when pulses of new magma entered the system beneath the volcano. Image credit: Phil Shane (University of Auckland, Auckland, New Zealand).

Tick Tock

In the textbook diagram of a volcanic eruption, liquid magma spurts up and out the top of a mountain. The reality is less straightforward, explains Costa, as he pulls out a drawing of a volcano, sitting above a deep magma reservoir (2). As the magma moves up through the crust, it can release fumes, cause earthquakes, and offer other warning signs that an eruption might be coming. However, there are two hotly debated possibilities for the way the magma rises.

One possibility, Costa explains, is a direct ascent with molten magma shooting up from a deep reservoir to the surface in a matter of days. In that case, even a slight grumble from a sleeping mountain could trigger an evacuation. The other possibility is that magma moves some distance toward the surface and then stalls, still deep underground, where it could sit for tens of thousands of years before someday reactivating and erupting.

Magmas can stall because they’re a mix of solid and liquid rock, as well as gas, which cools as it rises toward the surface, notes Shane. If enough of the magma cools below about 720 °C, it will solidify and remain stuck underground. It stays that way until a future injection of hot magma from below reheats and melts it, buoying it upward to erupt. Given that stop–start scenario, a mountain’s grumbling might not be an immediate cause for concern.

From a safety point of view, a bubbling vat of magma that could blow at any time is much more hazardous than a plug of mostly solid rock, says volcanologist Kari Cooper at the University of California, Davis. To understand what’s happening underground, Cooper led a seminal 2014 study that analyzed tiny crystal clocks trapped in the rock of past eruptions, to infer how the magma had behaved before reaching the surface (3).

She began by crushing volcanic rock samples from the two most recent eruptions of Oregon’s Mount Hood and sifting out crystallized minerals that formed underground as the magma journeyed to the surface. These crystals included plagioclase, which can be radiometrically dated by measuring how much uranium in the mineral has decayed into thorium and radium. Plagioclase is one of several minerals, including zircon, that are used for radiometric dating. These tough minerals do not remelt or chemically erode and offer a durable record of when crystals started forming in the magma. “It starts the clock,” says Bradley Pitcher, an igneous petrologist at Columbia University in New York.

Cooper used a second type of crystal clock to work out how long the magma had been in a hot and liquid state that erupts relatively easily. This second clock can use some of the same minerals—for example, plagioclase—or other types of crystals, including pyroxene, olivine, and quartz. By studying the extent of diffusion between the layers of these crystals, researchers can estimate how long the crystals experience temperatures hot enough to erupt.

By the time the temperature drops to about 750 °C, so many crystals have formed that the magma becomes more than 60% solid, taking on the gooey consistency of asphalt, Pitcher says. Hence, the age differences between these crystals and the radiometric dates reveal how long the magma was in a less-dangerous, mostly solid state.

The oldest plagioclase crystals in Cooper’s 2014 work were at least 20,000 years old. And the most time they’d spent at high temperature was on the order of hundreds of years. Hence, the highest percentage of time that the crystals could have spent at those high temperatures was 12% of the magma’s total lifetime, suggesting that the magma was locked into a solid form under the mountain for most of the 20,000 years. The notion that magma spends the majority of its life in a solid form, trapped in reservoirs under volcanoes, became known as “cold storage” about a decade ago, Pitcher says. “This was one of the big breakthroughs in how we think of volcanoes.”

These eight vertical tube furnaces can be heated to mimic temperature conditions underground before eruptions. Image credit: Daniele Cherniak (State University of New York at Albany, Albany, NY; furnaces housed in E. Bruce Watson’s experimental labs at the Rensselaer Polytechnic Institute in Troy, NY).

Trial by Fire

Using this second type of crystal clock depends on the lab-based work of researchers such as Daniele Cherniak, an experimental geochemist. She and her colleagues measure key chemical parameters in various types of crystals in the lab, which other geologists can use to estimate how long similar crystals took to form in nature.

At her lab bench at the State University of New York at Albany, Cherniak polishes a slab of volcanic quartz crystal that has many microscopic layers. Its core precipitated first, followed by a series of tree-ring-like layers that accreted after each new injection of hot magma rose from below to reheat the old rock.

In theory, each crystal layer should be chemically unique, reflecting the mix of magmas from which it formed, and a cross-section of an entire crystal should have crisp, distinct rings. Real-world crystals, however, don’t have such clear boundaries. That’s because atoms can migrate between neighboring layers over long periods, making the chemical compositions of the layers a little more homogenized. Older crystals experience more of this boundary-blurring effect.

Researchers can re-create this process in the laboratory, by measuring the rate of atomic diffusion inside crystals and observing the depth and concentration of uptake. That offers a standardized way to date how long wild crystals took to form, based on the amount of mixing between their layers.

To calculate those standard diffusion rates, Cherniak uses polished pure crystal, such as quartz, which she sandwiches against another material—titanium, for example—in a furnace as hot as 1,500 °C. “We’re looking to measure uptake of titanium in the quartz after heating the samples for extended time,” she says. The process mimics an accelerated version of what’s happening underground.

Cherniak next removes the material from the furnace and uses a helium-ion beam to measure the concentration of titanium absorbed by the crystal. Helium ions in the beam bounce off the titanium, silicon, and oxygen atoms in the quartz crystal with different energies. The result is a spectrum of helium scatter that reveals the concentrations of all the elements at specific points in the crystal, Cherniak explains. The helium ions that bounce off titanium—a considerably heavier atom than silicon and oxygen—have the most kinetic energy, for example.

Because Cherniak also knows the temperature of the furnace and the duration of the experiment, she can work backward to calculate a generalizable diffusion rate of the compound through the crystal. Experimental geochemists publish these diffusion rates as standards. Other researchers then apply the standards to extrapolate the age of real-world crystals based on the extent of various compounds’ diffusion through the mineral.

Dating Struggles

A few big assumptions in these types of calculations, however, give some researchers pause. First, says Mike Jollands, an experimental petrologist at the Gemological Institute of America in New York, many real-world crystals may not start with the crisp delineations between their layers assumed by diffusion equations. Second, diffusion rates are temperature-dependent—and geologists’ estimates of the temperatures of real-world magma are subject to big margins of error.

It’s so difficult to reliably reproduce conditions below Earth’s surface, including high pressure and temperature, that various experimental labs have found significant disparities in their experimentally measured diffusion rates. In a controversial 2020 study, for example, Jollands et al. (4) reported that the diffusion rate of titanium through quartz was more than 1,000 times slower than Cherniak had reported in 2007 (5). Replication is an issue for the field, and when studies disagree, “sometimes we get to stalemates,” Jollands says. “In this field, it’s very, very rare to replicate something,” he says, which ultimately shook his trust in the applicability of crystal clocks as reliable tools to predict eruptions. Since leaving academia, Jollands says being removed from the controversy “takes a lot of stress out of my life.”

Jollands does not think the whole concept should be tossed out, but he says that benchtop experiments must be replicated to confirm standardized diffusion rates. Proceeding with caution is especially important now, he says, as academic studies move into real-world use as forecasting tools.

Take Shane, Costa, and colleagues’ recent study of crystal clocks from Mount Taranaki (1). The researchers looked at plagioclase crystals from Taranaki’s past eruptions and found the crystals had two layers: a magnesium-poor core and a thin, magnesium-rich outer rim. The magnesium gradient between them was “so steep, it was like jumping off a cliff,” Shane says. The stark compositional difference between the crystal core and outer layer suggests that there was little time between the formation of the outer layer and the eruption, Costa concluded. Based on the history of other sudden eruptions, “we might see some seismicity a week before,” Costa says.

Taranaki is a popular tourist destination, situated in an agricultural area with rich oil and gas reserves. New Zealand monitors seismicity and gas release beneath the mountain and models what might happen if Taranaki woke up. In February 2024, Shane participated in a workshop with GNS Science to develop an eruption model for Taranaki and gave a presentation on their crystal clock study. His conclusion: “The magmas rise fast, and there could be little warning.”

Given the variables involved in crystal clock calculations, however, they shouldn’t be a standalone forecasting tool, Shane says. Many of the variables involved in these calculations are probably in the ballpark, but could be refined, he says. “Like all science, the crystal clocks have lots of opportunities for investigation and improvement,” Shane says. “The key to understanding phenomena like volcanoes or hurricanes or earthquakes is having a wide toolbox,” he notes, but “there will never be a universal tool.”

Shane is participating in ongoing meetings with GNS Science to refine the model, though he can’t share many details yet. The goal is to educate the public on what to expect if the mountain comes back to life. He says that improving the spatial precision of spectrometry methods could lead to better crystal-clock predictions. Most commonly available spectrometry tools can only offer precise composition measurements down to tens of nanometers. Finer resolution would allow researchers to measure diffusion rates closer to real-world temperatures—around 750 °C, for instance—helping to make eruption forecasts more reliable by revealing how quickly magma can change. “The closer we can look at things spatially,” Shane says, “the more accurate we may become.”