Why Can’t We Predict When a Volcano Will Stop Erupting?

Why Can’t We Predict When a Volcano Will Stop Erupting?
Smoke rises from a volcano on the Canary Island of La Palma, Spain on Friday Oct. 1, 2021. (Photo: Daniel Roca, AP)

On September 19, Cumbre Vieja — a volcanic ridge covered in the rocky scars of past paroxysms — erupted for the first time in 50 years. Every single day thereafter, molten rock cooking at nearly 1,093 degrees Celsius has poured from multiple fissures and vents, frequently accompanied by explosions propelling glassy ash skyward. Around 2,500 buildings, many of them homes, have been destroyed as the lava has meandered to the sea, where it continues to build an onyx delta shrouded in a haze of ash, steam, and acid.

Naturally, the 7,500 people who have had to flee the area want to know when this tempestuous act of magmatic madness will come to an end. Unfortunately, volcanologists don’t have an answer to that all-important query. A similar level of uncertainty arises whenever a volcano erupts, whether it’s spewing lava and ash near people or far from them.

Volcanologists are starting to get pretty good at forecasting the start of an eruption. If a volcano has been thoroughly monitored for some time, then scientists get to know what its usual activity is like: the quakes it makes when magma or hydrothermal fluids move about; the way it breathes in and out and changes its shapes; the sorts of gases it belches out. If one or several of those parameters begin to significantly shift, then it may suggest something sinister is approaching.

Lava flows from a volcano on the Canary Island of La Palma, Spain on Friday Oct. 1, 2021. (Photo: Daniel Roca, AP) Lava flows from a volcano on the Canary Island of La Palma, Spain on Friday Oct. 1, 2021. (Photo: Daniel Roca, AP)

When solid observations are combined with a decent knowledge of the volcano’s eruptive history, volcanologists can at the very least warn those nearby that an eruption in the next few days or weeks is more probable than it was previously. But “the end, it’s very difficult to forecast,” said Maurizio Ripepe, a geophysicist at the University of Florence in Italy. Why?

Confusingly, there is no agreement on what defines an eruption’s termination. That may sound silly — surely, it’s when there’s no more lava shooting out of the volcano, right? — but volcanoes don’t care about how us puny surface-dwellers measure time. These titanic lithic entities exist and operate on timescales far exceeding the ones we are used to.

Take Hawaii’s Kīlauea: Lava started coming out of part of that colossal volcano in 1983, and although its eruptive activity ebbed and flowed, with some brief pauses in lava production, it only really stopped in August 2018 — 35 years after that extraordinary eruption sequence began. Then, in December 2020, lava once again poured into Kīlauea’s summit crater. Were those 18 months a pause, or a (temporary) end to the volcano’s eruptive shenanigans?

To a volcano, the distinction between the two is meaningless. Sam Mitchell, a volcanologist at the University of Bristol, said that what matters to those living around these volcanoes is the answer to a slightly different question: When has a volcano stopped erupting long enough for it to be safe to return? When can a community start to rebuild? In other words, what stops a volcano from erupting on a human timescale?

A plume rises near active fissures in the crater of Hawaii's Kīlauea volcano on Monday, Dec. 21, 2020. (Photo: M. Patrick/U.S. Geological Survey, AP) A plume rises near active fissures in the crater of Hawaii’s Kīlauea volcano on Monday, Dec. 21, 2020. (Photo: M. Patrick/U.S. Geological Survey, AP)

Knowing how much magma is available to erupt would be fantastic. Scientists can’t yet see magma shifting about below the surface, but they can at least infer how much is there. The amount the ground inflates is a proxy for how much pressurised magma is vying to burst through, and the seismic soundtrack made by magma breaking through rock can be used to track its movement. If that inflation turns into deflation, then it may be a sign that the magma supply is running dry or is at the very least losing the pressure required to punch through, said Pablo González, a physical volcanologist at the Spanish National Research Council in Tenerife.

But this sort of monitoring only really tells you what’s happening on a very shallow level. “We don’t see the deeper system,” said Ripepe — and that’s a problem.

Magma pools close to the surface when it’s about to erupt, but that magma supply is fed by molten rock coming from more abyssal caches, which are themselves cooked up by the actions of seriously deep-seated tectonic processes. Even if the amount of magma present at shallow depths is determined, technological limitations mean that the volcano’s plumbing system is often rarely visualized. That means it’s essentially impossible to know whether the volcano is being continuously supplied with extra magma or if its stores are soon to expire.

Lava flows from a volcano on the Canary Island of La Palma, Spain on Saturday Oct. 2, 2021.  (Photo: Daniel Roca, AP) Lava flows from a volcano on the Canary Island of La Palma, Spain on Saturday Oct. 2, 2021. (Photo: Daniel Roca, AP)

“We have no way at the moment to understand how much magma will come up from the deeper system,” said Ripepe. And without knowing that, it is difficult to forecast when an eruption will run out of fuel.

It isn’t even as straightforward as fully evacuating a volcano’s reserves. Magma can ascend but often won’t erupt. Often, it ponds in a reservoir or a labyrinthine network of possible escape routes, either remaining partially molten or freezing solid. “We never know how much of the magma that’s being put into the system will erupt,” said Mitchell.

Just look at Kīlauea. In 2018, from May to August, it performed the grand finale of its 35-year-long eruption by expunging 320,000 Olympic-size swimming pools’ worth of lava, destroying 700 homes in the process. On August 1, it looked like there was nothing stopping it. But by August 4, all eruptive activity had come to an end. And subsequent studies found that no more than 33% of the magma reservoir fuelling that eruption had been drained.

Why did it stop if it had so much more left to give? No one’s quite sure, but it underscored the fact that volcanoes don’t stop erupting solely because they’ve run out of molten matter.

A handful of volcanoes are uncomplicated. They have a certain amount of eruptible magma, made clear by (among other things) how inflated the ground has become, and after a flamboyant, prolific start to proceedings, the amount of lava coming out per second begins to drop off. If that drop off remains continuous, you could extrapolate it and guess when the eruption is going to end. That’s what happened with the 2014-2015 eruption of the Bárðarbunga volcanic system in Iceland. The lava, which flowed from a simple fissure, went from a cascade to a drizzle, and its disappearance in February 2015 was forecast by volcanologists.

“The best indication of an ending may come from the evolution for the eruption rate,” said Mike Poland, the scientist-in-charge at the U.S. Geological Survey’s Yellowstone Volcano Observatory. “But there are some examples where that doesn’t work at all, and the best one I can think of is Kīlauea 2018.”

That summer, the volcano’s summit progressively collapsed as lava gushed from its lower eastern flank in distinct pulses. Scientists soon realised that the pulses were being controlled by those collapses: like a piston, the falling roof was putting pressure on the magma, effectively squeezing it out of the volcano’s side. The eruption stopped when the collapses ceased. There was no Bárðarbunga-like decline from a torrent of lava to a trickle, nothing that could be used to see into the volcano’s future.

Smoke rises from the lava eruption on Holuhraun, northwest of the Dyngjujoekull glacier in Iceland, Monday, Sept. 1, 2014.  (Photo: Eggert Johannesson, AP) Smoke rises from the lava eruption on Holuhraun, northwest of the Dyngjujoekull glacier in Iceland, Monday, Sept. 1, 2014. (Photo: Eggert Johannesson, AP)

The amount and type of gas eructing from a volcano may help volcanologists work out when an eruption will end. They already clue them in to an eruption’s beginnings. When magma rises to a depth of about 29 km, carbon dioxide bubbles out of it and flees to the surface. If instruments can detect this gas, scientists are alerted to the existence of magma that may be trying to erupt. Similarly, if instruments detect a drop-off in this gas (and others), it may suggest the eruption is losing momentum: Without trapped gas, a magma isn’t buoyant, and it will struggle to erupt. But this relationship is rarely that clear-cut. And certain gases, like carbon dioxide, are difficult to measure when they already exist in the atmosphere in abundance.

At this point, attempting to forecast an eruption’s death may seem largely futile. But there is some hope.

History provides hints. Based on a multitude of past outbursts, eruptions of fluid lava on La Palma, much like the eruption happening right now, “tend to last up to five months,” said González. The shortest of the bunch are over in just a few days or weeks. “This eruption is heading to the two-month mark.”

Will it end before the five-month mark? “Who knows,” said González. Maybe it will set a new record. Perhaps it will stop next week. It all depends on the supply of eruptible magma, which remains invisible. But the more a volcano’s eruptive history is known, either by looking at ancient geologic deposits or by documenting eruptions in real time, the clearer a volcano’s “average” eruption duration will become.

Like forecasting an eruption’s opening salvo, working out when an eruption will end depends on tracking a range of things — the volcano’s seismicity, its deformation, its gas output, and so on — and determining what combination of signals points toward a time of tranquility. One way such efforts could be accelerated is through machine learning. A rudimentary artificial intelligence, fed data from countless past eruptions, may be able to spot these patterns far quicker than a human ever could.

If so, scientists could let it spy on an eruption in real time and predict when it will end. “That’s kind of where we take volcanology next,” said Mitchell.

Lava flows from a volcano on the Canary Island of La Palma, Spain on Friday Oct. 1, 2021. (Photo: Daniel Roca, AP) Lava flows from a volcano on the Canary Island of La Palma, Spain on Friday Oct. 1, 2021. (Photo: Daniel Roca, AP)

Incredible advances are being made in volcanology all the time, so a future in which the entirety of an eruption’s evolution, from start to finish, can be known from the first murmurs is easy to envisage. But we must be patient. “Volcanology is a young science,” said Arianna Soldati, a volcanologist at North Carolina State University.

Careful monitoring efforts, and allowing scientists to work closely with emergency services, saves people from a volcano’s wrath. Thousands of homes may have been destroyed on La Palma, but so far, nobody has died. But it will be some time before volcanologists will be able to make confident forecasts of an eruption’s end, especially if it’s one that’s causing enormous social upheaval. “We can only hope that it will end one day,” said Ripepe.

There is one upside to Cumbre Vieja’s ongoing eruption: Scientists can use it to better understand how volcanoes work, including what causes them to erupt and what makes them stop. One day, perhaps, forecasting both will become routine. “And with every eruption that happens,” said Mitchell, “we inch a little bit closer” to that goal.

Robin George Andrews is a volcanologist turned freelance science journalist. His first book, Super Volcanoes: What They Reveal about Earth and the Worlds Beyond, is now available in bookstores everywhere.