Volcanoes erupt all the time. It may seem a quiet time to us but that is because most eruptions are small and low impact, and stay below the radar. An excellent daily overview can be found on http://lechaudrondevulcain.com/blog-spotlight-two-column/ As I write this, it lists on-going eruptions at Sinabung, Etna, Stromboli, and Sabancaya. Volcanodiscovery also lists Nishinoshima, Suwanosejima, Semeru, and Kerinci as being active. Volcanoes with long-lasting lava lakes may be added to this: Erebus, Nyiragongo, Erta Ale, Masaya, Yasur, Mount Michael. This is a normal level of activity for our planet. Only Sinabung has really made the news in the past few days. It had a decent size explosion with ash falls in local villages and the government recommended that the people should wear masks. (Where have I heard that before?)
Large eruptions are much less frequent. The largest effusive eruptions of the last 20 years have been Holuhraun and Nishinoshima, the first one the largest by eruption rate, the second one (probably) larger by volume. The largest explosive eruption was Puyehue-Cordon Caulle in 2011, a borderline VEI-5. On average, there is one VEI-5 eruption per decade but the last few decades have been below par. Before the 2011 boom, the last such eruption was Mount Hudson in 1991. Admittedly this was less than two months after Pinatubo, so the current lull came after lightning hitting twice within months. Chance is like that. VEI-6 events occur on average once every 40 years or so. In the last 200 years, we have seen Krakatau, Santa Maria, Novarupta, Quizapu, and Pinatubo.
And VEI-7’s are even harder to find. In the last 2000 years, we have suffered Tambora, possibly the 1452 eruption (cautiously attributed to Kuwae), Rinjani, Baekdu, Ilopongo (recently upgraded to VEI-7) and Taupo. VEI-7’s happen about once every 300 to 400 years. Only 8 eruptions are classified as VEI-7 between 2000 and 10000 years ago: either the last millennium has been exceptionally active, or much more likely), we are missing another 20-25 VEI-7 eruptions in the holocene. It is not easy to measure the VEI scale of a prehistoric eruption.
VEI-8 eruptions have been covered elsewhere. They are so rare that we don’t need to worry about them for now, and they also are very different type of eruptions. A VEI-8 is more than just a bigger bang.
Where were those VEI 7’s of the holocene? Two were in Indonesia (the most recent ones, in fact), 1 in New Zealand, 1 in Korea, 1 in Japan, 1 in Europe, 1 in the US, 1 in Central America, 1 in South America – you get the idea: they are fairly equally distributed across the world. The Aleutian arc and Kamchatka are probably the most productive. But with perhaps two-thirds of the missing, it is hard to be sure. Indonesia and PNG may have had far more than they are credited with. The scars even of a VEI-7 can be healed surprisingly quickly.
Looking through the list, it is striking how unfamiliar the names of the VEI-6 and VEI-7 volcanoes are. The list of currently active volcanoes at the start of the post contains many names that readers here will readily recognize. But the largest eruptions have names that you would not expect to see in the daily headlines. Krakatau is of course famous, but that is because of its eruption and the publicity it received. Tambora was bigger and more deadly, but its eruption happened before the era of daily world news and it remains far less well known.
And before their cataclysmic eruption, these mountains were even less known. Krakatau was not recognized as potentially active until the eruption began. Tambora was not even recognized as a volcano until 5 years before its eruption when it first showed some sign of activity. This was in spite of it being one of the tallest mountains in Indonesia – but we do not know the exact height as no one was interested in it at the time! Tambora’s last known eruption before the big one in 1815 is dated to 740AD (+-150 years). Pinatubo was just a hill, until the USGS raised the alarm weeks before the explosion. Santa Maria had been inactive for at least 500 years, i.e. there was no known activity in historic times. The first indication of something building was nine months before the eruption with an earthquake swarm. Novarupta was a flank eruption from Katmai; Katmai had no known eruptions before the big one, and was only known to occasionally ‘smoke’ in the decades before. For older eruptions we lack historical records, but we know that Vesuvius was not known to be volcanic before its VEI-5 eruption (although some writers did suspect it to be): it had been quiet for at least three centuries. Santorini was inhabited before its big event, suggesting it too had been inactive.
The link between explosions and dormancy makes sense from an explosive point of view. Young, fresh magma tends to be effusive. Over time it becomes stale and more explosive. The ash from big volcanic explosions tend to be white from silicates, while the lava from effusive eruptions tends to be black. Magma turns white with age. A slow cooling by a magma chamber is a danger. Like the Beirut explosive, the substance of volcanoes becomes more explosive as it ages.
In some cases, there may have been a seismic component to the eruption build-up as well. Vesuvius had a major earthquake 17 years before its eruption. A report of ‘tainted air’ at Pompei after the earthquake could be due to gas emissions from Vesuvius after the earthquake although this is speculative. Krakatau suffered a major earthquake five years before it went boom, and seismic activity in the area increased after that time. Pinatubo suffered a major (M7.7) earthquake a year befor its eruption, and steam emissions were first reported shortly after this. Santa Maria suffered a major earthquake (M7.5) 6 months before the eruption, although it should be noted that seismicity at the volcano had begun already 3 months earlier. The VEI-6 eruption of Huaynaputina happened a year after a major earthquake. Archeology at Santorini indicates damage from an earthquake some time before the eruption.
Does this mean that the earthquakes caused the eruption? That is hard to prove. These large earthquakes are tectonic, not volcanic, and they were not caused by the volcano. Large earthquakes are also not that uncommon. The events seem to be normal for the region, to happen perhaps once a century or so, but close enough to the eruption that people noticed it.
Can we make it sound plausible? How can an earthquake cause an eruption? Charles Darwin was the first to suggest a relation, when he noted that four volcanoes erupted in Chile shortly after the major earthquake (M8+) that he experienced there. Not all of these four eruptions may have been related to the earthquake though. An earthquake can’t cause eruption directly. Even a large earthquake does not destabilize the magma chamber by shaking, nor does it squeeze the magma upward. What can happen instead is that the earthquake causes fracturing at the surface: this can cause a flank collapse, and release magma that is already close to the surface. Slopes of volcanoes can be at the limit of their stability, steep and with poor cohesion. An earthquake can do notable damage.
A second way an earthquake can cause an eruption is by unlocking a fault. A locked fault does not give way easily, and will not allow magma to intrude. But unlock it, and the stress reduces dramatically. It may now be possible for magma to get in, and open up a pathway that allows it to migrate up.
The typical time between a large earthquake and a large eruption seems to be a year to a decade. (This assumes that the two events are related!) That is not enough time to fuel a VEI-6 (or even VEI-5 – it might just do a VEI-4). Magma supply rates are normally around 0.001 to 0.01 km3/yr. Instead, rising magma can re-heat an existing reservoir, one that has been cooling for some centuries. The magma will have turned to a slush, filled with crystals. The new heat re-melts the silicic mush, and much like microwaved honey, it becomes mobile. This happened in Yellowstone, where its last eruption 70,000 years ago(!) came 10 months after reheating of the magma.
Reheating is a common cause of large eruptions. The VEI-5 Quizapu eruption of 1932 (on the flank of Cerro Azul) came after a reheating episode. The funny thing here was that the same thing had happened at the same volcano in 1846, but the earlier eruption was effusive while the second one was explosive. Other than that, the eruptions were the same size (5 km3). Why this difference? In both cases a dacite, cooled magma chamber was recharged by andesite at 1100 C. In 1846, this led to an increase of the magma temperature by 50C, which made the magma less viscous. The gas in the magma chamber escaped, and the remaining fluid magma happily fed an effusive eruption. In 1932, a smaller amount of andesite entered the magma chamber, and the temperature increased only by 20C. But now the gas had already gone and this changed the response by the magma to the heating. When gas bubbles take up a significant part of the magma volume (say 10%), the heat that causes the magma to expand compresses the bubbles. The bubbles take up much of the excess pressure. Without gas bubbles, the magma can’t expand because the volume is already full. This causes a rapid rise of the pressure, something that didn’t happen in 1846. The larger pressure (from a smaller recharge event) led to a strong explosion.
The following videos shows how an explosion deep underwater can cause a much stronger shock wave than one near the surface. This is because of the same effect: a liquid such as water (or magma) is not easily compressible, and it gives very strong pressure shocks.
To summarize what we found: large eruptions are more likely to happen in volcanoes that have been dormant for some centuries, letting the magma chamber become cool and stale. When new magma recharges the magma chamber and reheats it, the pressure can shoot up because there is no gas to soak up the pressure. The explosion can now happen within months to a decade. An earlier eruption can make the next one worse by degassing the magma further. A large tectonic earthquake in the region can trigger a recharge event, and can set up the volcano for the big stage. (It should be noted this is still a a rare event. Most large earthquakes near most dormant volcanoes don’t do this.) As an aside, the fact that the eruption is triggered by reheating suggests that there may be little inflation until close to the eruption itself. Don’t expect an extended warning!
The big question is now whether we can see this coming. Phrased a little better: if you were to look at us in the year 3000, which 3-4 volcanoes would have been replaced by new, large calderas?
A VEI-7 evacuates a lot of volume. That means that we should be looking at decent-sized mountains which have this volume. That won’t catch all, since some smaller mountains may have a surprisingly large amount of magma underneath but you have to start somewhere. Tambora was a stratovolcano over 4km high, and was among the highest peaks of Indonesia. Rinjani also must have been a large mountain. VEI-6/5 progenitors, on the other hand, may be less striking. Pinatubo and El Chichon were not that notable. We want a volcano that is listed as dormant. A good sign of this is the presence of erosion; a smooth stratovolcano like Fuji suggest a high level of activity, even if the most recent eruption may have been centuries ago, and that is not what we want. A weakness in the edifice may ‘help’. Large eruptions are often associated with flank collapses, and these are more likely to happen in (overly) steep volcanoes. The region in which our dormant volcano is located should still be volcanically active: the UK Lake District is an impressive old volcano, but it does not qualify fur us. Finally, other large calderas in the region are a good sign, as the potential for VEI-7’s may be build into the fabric of the local volcanoes.
Let’s give an example. Kamchatka has a history of large eruptions. Are any mountains there looking risky? Well, yes. Bezymianny is often mentioned, as it suddenly became active in 1955, after a long repose, and has been doing decent explosions almost like clockwork since. But it is a bit small. There is in fact a better candidate right next to it.
Kamen is the second tallest volcano in Kamchatka (4579 m), is dormant, steep, has erosion features and a flank collapse scar that is a little over 1000 years old. The flank collapse left a slope of up to 70 degrees. The other slopes are 30-45 degrees. This is a seriously steep volcano. Kamen is flanked by two highly active volcanoes, Bezymianny and Klyucheskaya Sopka. The latter is a beautiful Fuji look-alike, 180 meters taller than Kamen (making it the tallest volcano in Kamchatka), and active every few years – the last eruption ended only in July. And this world-class volcano (Klyucheskaya Sopka) is only 7000 years old. With such neighbours, what is the risk of another magma recharge in Kamen? 1000 years from now, will Kamen still be there or will Kamchatka have gained another lake?
When we asked the same question a few years ago, cbus came back with some suggestions. I will re-issue his candidates here, for you to consider.
1. Kronotsky – Kamchatka
Kronotsky is not quite as tall as Kamen but still reaches a respectable 3528 m. It is steep, with slopes of around 35 degrees. It seems to be in a long period of repose, with the only reported activity of the last century a small phreatic explosion in 1923. The summit crater is reportedly plugged with a volcanic neck (whatever that means). The edifice is likely weakened by glacial dissection, and it sits in a region where every other nearby volcano has gone caldera in a large manner. As such, if it wakes up, it very well may go boom in a dramatic fashion, possibly due to a large slope failure, leading into a caldera eruption.
2. Ulawun – New Britain
Ulawun follows a similar trend: a tall conical edifice in a region of many caldera large eruptors. It is located about 130 km from Rabaul and at 2334 meters is the highest mountain in New Britain. Similar to Kronotsky, there is evidence of slumping and existing edifice weaknesses at Ulawun. In fact the risk analysis for the volcano specifically identifies slope failure as a concern. It states that Ulawun is 400 meters higher than most of the volcanoes in the Bismarck and may be at the limit of structural stability. If it is anything like it’s neighbours (which is fairly likely), it could be approaching the end of its life-cycle. This would result in a caldera eruption, where it would then start to rebuild in a new cycle afterward. It is an active volcano, with a significant eruption as recent as 2019.
3. Agua – Guatemala
Not much is known of this Amatitlan somma volcano, but it’s enormous (3760 meters tall), steep, in a region where many massive flank collapses occur, and in an extremely active volcanic region even though there are no known eruption from Agua itself. The magma supply may have simply been diverted to another nearby volcano, but it also could just be plugged up after growing so large, which may not be a good thing.
4. Almolonga – Guatemala
Almolonga breaks the mold from the previous candidates. It is not the type of volcano that builds a large edifice, and instead is a dome complex that does not have one central vent, and has some very sticky and viscous magma. It features a ring-dike configuration of dacitic and rhyolitic lava domes. In a lot of ways, Almolonga looks much like a resurgent dome complex of a large caldera, but the only caldera it has is a small 3km caldera which may be more of a crater than a true caldera where the magma chamber was destroyed. Another worrisome part of Almolonga is that it sits in a very similar tectonic environment that produces other large calderas – one with both extension as well as subduction magmatism. Amatitlan and Atitlan both sit to the south of it, and I wouldn’t be surprised if it were very similar to those volcanoes.
5. Laguna Del Maule – Chile
Del Maule has been inflating significantly for quite some time now. It covers a large area, and quite a few publications have noted that it could be a large eruptor if it decides to go. Luckily, it’s in the middle of nowhere, so not so many people will be affected. Against it, the very fact that it inflates may indicate that the magma is behaving fine and that the next eruption will be a normal (for this volcano) VEI-4. Who knows?
6. Novarupta – Alaska (USA)
Yes, Novarupta had a huge eruption just over 100 years back. With that said, I have a suspicion that this could be more of a Crater lake type situation, where we see a large VEI-6 that is more of the precursor than the end-result. The alignment of the Novarupta vent in the middle of 5-6 volcanoes suggests that the primary Novarupta magma chamber is very large, and located centrally away from the small Katmai Caldera. This could mean that there is still a lot more down there that could re-erupt with a decent bit of recharging. Certainly not out of the realm of possibility within 1000 years.
7. Ata Caldera – Japan
Most people here know about Sakurajima and Aira Caldera, but Ata may be the one more worth worrying about right now. It is located at the head of Kagoshima Bay, 40 kilometers from Aira. Ata is just as large, and unlike Aira, has not been releasing pressure since 885 A.D. Prior to this date, the somma KaimonDake would have frequent VEI-4 eruptions, but everything in this area has gone quiescent since then. That to me is somewhat worrying. Ata also seems to love creating large phreatoplinian eruptions that form maar caldera structures, which would be a disaster if this were to happen.
This completes cbus’ list. But what do you think? Which volcano keeps you awake at night, either in fear or in expectation? In 1000 years time, which volcano will unexpectedly be missing, and will have been replaced by a large caldera? Let us know!
And if you like this, you may be interested in the NDVP list as well, available at https://www.volcanocafe.org/the-new-decade-volcano-program-ndvp/