The battle of Stalingrad was among the bloodiest and most horrific of the second world war. During the cold winter of December 1942, the Commissar of Stalingrad organised open-air party meetings for local artists and musicians to encourage the exhausted and hungry soldiers. One of those musicians was the violinist Mikhail Goldstein. On New Year’s Eve, Goldstein went out to the trenches and found scenes of horror. He played Russian folk themes for the soldiers, but eventually played Bach, even though that was a German composer. The music was played over the loudspeakers as the shooting ceased. When he was finished, silence reigned, until a loudspeaker from the German side sounded, saying in Russian, ‘Play more Bach. We won’t shoot.’ Goldstein continued to play. After an hour and a half of playing his exquisite music in the trenches, he returned to the Commissar and the war resumed.
This anecdote has absolutely nothing to do with volcanoes, apart from volcanoes and Stalingrad having moved close to each other in the dictionary, since the renaming to Volgograd. However, it shows the power that music has over us – but only the best music. The story wouldn’t have worked so well with an average musician and composer. Isn’t this the same with volcanoes? A small eruption from a minor volcano attracts a few comments on our blog. In contrast, Leilani dominated our news for months. Size matters.
Newspapers love records. Super-eruptions feature regularly in both the tabloids and the popular science magazines, ignoring the minor fact that we have not seen one in historical times. Articles compete in finding the biggest eruption of all times. VC of course would like to shine a light in the extreme darkness of such super-eruptions. With so many eruptions competing to be the biggest, can we find the truth? The goal of this post is to identify 25 major eruptions. That should be enough to keep anyone happy, and should keep volcano warfare at bay for many months. We would also like to know what makes these extreme eruptions tick: what causes them, and how do they proceed?
Measuring the power of a volcano is not a trivial task. What counts? Is it the lava it produces? Krakatau produced none. The decibels of the explosion? Krakatau outperformed Tambora on this one. The amount of explosive ejecta (tephra/ash)? Tephra is fluffy and takes up much more volume than the rock that produced it. The height of the plume? This is driven in part by the explosion but mostly by convective air currents. The number of casualties? Most volcano-related casualties come from tsunamis and flank collapses, rather than directly from the eruption. Impact on worldwide weather? This depends on the sulfur fraction in the ejecta which can vary widely even between eruptions of the same size. You run into problems no matter how you define it.
Eruptions are measured on the VEI scale, where the ‘E’ stands for ‘explosive’. This scale grades eruptions according to the volume of tephra they produce: tephra is pulverized rock, and so this relates to the energy expanded in breaking up the mountain. This is called the ‘erupted volume’. It is two and a half times the volume of the original rock or lava, called the ‘dense rock equivalent (DRE): the amount is inflated compared to the hole that it leaves in the ground. A major eruption is VEI 7, and this corresponds to an erupted volume of more than 100 km3. To visualize this, over an area of 100 by 100 km, it would give an ash layer 10 meters thick. A VEI 8 eruption produces more than 1000 km3 – now the ash over this same area is 100 meters thick. Krakatau was VEI 6. Effusive eruptions produce lava, sometimes in copious amounts, but this is not what the scale was designed to measure.
The aim of this post is to make a list of 25 powerful ‘super’-eruptions. Flood basalts and large igneous provinces are excluded: eruptions on the list should be listed on the VEI scale. A super-eruption is often defined as VEI 8, but some volcanic regions don’t reach that and produce no more than the occasional VEI 7: for equality & diversity sake we did include some VEI 7’s. There is no claim that these are the ‘largest eruptions ever’ (how could you know?) but they show what volcanoes can do – given a few million years.
Time matters. The further back in time we go, the less we know. Within historical times, the largest eruptions have been Tambora (1815), Kuwae (1453), Rinjani (1258), Paekdu (946) and Ilopango (540), where the identification of Ilopango is probable but not certain, and Kuwae is also not confirmed as culprit. The estimated erupted volumes of these eruptions (in terms of VEI) are 160 km3, 100 km3, 100 km3, 110 km3, and 84 km3, respectively. The volumes for the older eruptions are uncertain by perhaps 50%. That gives 4 to 5 VEI-7 eruptions over the past 2000 years. There may have been more, as the volcanoes of the major eruptions of 537 and 1809 have not yet been identified, and a major eruption from Taupo in the first millennium (perhaps 150 km3) should be added but it has an uncertain year.
To find much larger eruptions, and to locate the elusive VEI 8, we need to go back much further in time. Obviously, if our records are incomplete even for the past 2000 years, it must be much worse when going back a few million years. That begs an important question: how do we identify -and quantify- such ancient eruptions? There are two main methods.
Large explosions leave large holes, and the largest can be identified even after millions of years. The word ‘caldera’ comes from Spanish; it derives from cooking pot. (It exists in English as well, as cauldron.) Calderas and craters describe similar structures, but a crater is smaller, perhaps of order 1 km, while calderas can be 10 to 100 kilometers in diameter. (Meteor impacts, on the other hand, continue to be called ‘craters’ even if very large. There is little consistency in geology.)
Calderas form through explosions or through collapse, or by a combination of these. If aided by a graben and/or fault line, they can become elongated: these are called volcano-tectonic depressions.
Calderas give a good indication of the eruption size, as the missing volume should equate to the amount of rock or lava (DRE) ejected in the eruption. However, there are some provisos. We don’t know whether there was a large edifice on top of the caldera which would need to be added to the missing volume. Infall, resurgence and sedimentation can have raised the floor of the caldera. Collapse of the sides, or later eruptions, can have enlarged the caldera. Erosion can remove all traces of a caldera over time. The biggest proviso of all has largely been resolved, however, and that is that some calderas are so large that you can’t see them. Satellite imaging has shown this, for it picked up caldera so large that from within, the rim would be over the horizon. The Yellowstone caldera is a case in point.
Indeed, some calderas are unbelievably large. The Toba caldera measures 100 by 29 km, and the Yellowstone caldera is 72 by 55 km. Mons Olympus on Mars beats both of these. Its caldera is 85 by 60 km, and up to 3 kilometers deep. It consists of five nested calderas of which the largest measures 60 by 60 km.
The caldera area is a rough indicator of erupted volume. The plot shows the relation between caldera area and the erupted volume. Based on this, any caldera more than 300 km2 in area is worth a second look.
The second sign post of an ancient, major eruption is found in the ejecta. The ash and tephra covers large areas, and in the case of major eruptions, to great depth. These are the ignimbrite sheets, and they are the best tracers of eruptions more than 1 million years old, when the crater may long have been disappeared. They are not perfect, though. Often, it is not clear which is the originating volcano. A thick ignimbrite sheet may not come from a single eruption, or even a single volcano, and conversely, ignimbrite sheets at different locations may (or may not) come from the same eruption. Only a small part of the sheet may be exposed. And ignimbrite sheets are themselves not immune to erosion. They have survived best in deserts, and thus the oldest major ignimbrite eruption are from a few regions only, especially in the dry Andes.
The bottom line here is two-fold: the old ignimbrite record is very incomplete, and the listed volumes may combine more than one event and thus be overstated.
It is not easy to derive the size of an eruption that occurred millions of years ago. Scientists are often reluctant to state a size of an eruption. When they do bite the bullet, in general the area covered by the ignimbrites is determined and the thickness is measured at several places. The total volume of this is obtained by multiplying the average thickness with the area. The volume is converted to dense rock equivalent (DRE) by using a typical density of ignimbrite, around 2 gm/cm3 versus the density of rock (2.5 gm/cm3): this reduces the volume by 25%. The amount of material deposited back into the caldera has to be added to this. This depends on the depth of the original caldera which is unknown. As an assumption, caldera in-fill is assumed to be equal to the ignimbrite volume. The same amount is often added for ashes, so that the (DRE) volume becomes three times the ignimbrite. There are many uncertainties. The ignimbrite will only be visible at the surface over a small part of the range. The assumption of average thickness will significantly overestimate the volume if the thickness declines with distance. The ash amount is effectively unknown. For mildly-explosive eruptions, the amount deposited inside the caldera may be several times larger than that outside. Finally, if the caldera did not collapse as a block but acted as a trapdoor which collapsed on one side with the other side acting as a hinge, the caldera volume is halved. The published volumes may easily be off by a factor of two. For old eruptions, it may also be uncertain whether the ignimbrite sheet came from a single event and from a single source.
A super-eruption is a devastating event. Luckily, two aspects help to reduce the world-wide impacts. The first is that the eruption may involve fissuring along the caldera fault, with impressive fountains reaching huge heights, but without a single massive explosion. That will reduce the height of the ejecta, and limit (somewhat) the spread. Second, if eruption rates become too large (more than 1010 kg/s), the eruption column collapses under its own weight, again limiting the height it reaches. For Toba, the eruption rate may have reached in excess of 1011 kg/s. The two effects may have combined to explain the lack of evidence for a significant world-wide impact from Toba. But within 1000 km or so of the event, the effect would remain enormous. It would be equivalent to an ecological reset. I imagine the mouse-like kiwi bird wandering through an ash covered landscape, wondering where all the predators have gone.
And now it is time to present our list of 25. We had to be selective: we tried to cover a range of locations on Earth, so smaller eruptions (if VEI 7 can be called small) were included if they were huge for the local area, and major eruptions were left out if the area already had several events included. Preference was given to more recent eruptions – though no historical events made the list. Many of the listed eruptions occurred within the last few million years. There had to be an explosive (non-basaltic) element to the eruption: flood basalt eruptions were not included.
The list is organised by area. The local area is briefly described, followed by a list of the selected eruptions, identified by name and by volume. The parentheses give the erupted volume, which is what is used on the VEI scale. Note that in the scientific literature, eruption volumes are more often given in DRE: we converted this by using a factor of 2.5. Each eruption is briefly described, with the DRE volume where appropriate.
The VC list of 25 major eruptions
The Altiplano Puna Volcanic Complex
The APVC is located in the central Andes, between 21 and 24 degrees south, and shared between Bolivia, Chile and Argentina. It is a bone-dry, 4-km high plateau which covers an area of 70,000 km2. The plateau is punctuated by volcanoes reaching 6km in height, including the highest volcano in the world. The region has an exceptionally thick crust (70km). The entire plateau is volcanic: 15,000 km3 of magma was erupted here between 11 and 2 million years ago. Seven eruptions are believed to have reached VEI 8. There are four major calderas over 40 km in size, all of which had multi-cycle eruptions. The APVC is among the most impressive volcanic areas of the past 10 million years. The activity here peaked 4 million years ago and began to diminish 2 million years ago. The region remains active but recent eruptions are in the tens of km3 range, not hundreds. However, if pulses occur every 2 million years, it may be too early to write the region off.
1. Pastos Grandes (2000 km3)
This composite caldera is the northernmost of the Altiplano calderas. It measures 32 by 52 km. The caldera contains a salt pan and some lakes, where the brine is being mined for lithium. There have been three or four major eruptions within the caldera. The most recent of these occurred 2.9 million years ago: it deposited the Pastos Grandes ignimbrite over an area of 6000 km2, consisting of dacite bordering on rhyolite. The thickness of the ignimbrite sheet is 20-90 meters. The total volume has recently been re-measured as 800 km3 DRE, a bit lower than older determinations. Three quarters of the ignimbrite ended up inside its own caldera, either because of a low eruption column or because of early collapse of the caldera. As common in the Altoplano, there are no plinian deposits.
2. Guacha (3000 km3)
This has been the location of two major eruptions. The first of these caused the Guacha ignimbrite, This eruption is estimated at 1300 km3 DRE, and happened 5.6 million years ago; it formed a caldera estimated at 40 by 60 km. The second eruption deposited the Tara ignimbrite, occurred 3.5 million years ago and produced 800 km3 DRE. The caldera is a trap-door system, hinged on the western side.
The inflating volcano of Uturuncu is located on the northern edge of the caldera, and Cerro Chajnantor (site of a number of major astronomical observatories) is on the southwestern edge. There is a resurgent dome in the centre. It is common for large calderas to have a non-volcanic central resurgent dome, with newly active volcanoes located along the caldera rim.
3. La Pacana (6000 km3)
Of the calderas in the APVC, La Pacana is the largest, at 65 by 35 km. It has had two major eruptions, possibly in quick succession, where the first was rhyolitic and the second was dacitic. The second eruption was the largest. It is dated to 4 million years ago and is the origin of the so-called Atana ignimbrite. The eruption has been associated with a ring fracture. There were four different pulses within the eruption. The size given here comes from a recent re-evaluation: older estimates are 2-3 times less.
4. Cerro Galan (1500 km3)
This caldera is also in the central volcanic zone of the Andes, but is further south than the Altoplano and is not part of APVC. It lies on the eastern side of the Andes, in Argentina. The ignimbrite sheet of its eruption extends as far as 40 km from the caldera, with a thickness up to 200 meters. The eruption is dated to 2 million years ago. The caldera measures 16 by 27 km and is of a trapdoor type. The volume of the eruption is estimated as 630 km3 DRE.
The Taupo Volcanic Zone
The Taupo Volcanic Zone, on The North Island of New Zealand, extends over a length of approximately 350 km. It runs from Mount Ruapehu in the southwest to Whakatane volcano (submarine) in the northeast. This is a rifting zone which is widening by about 1 cm per year. There are a number of calderas, belonging to six separate volcanic centres: Rotorua, Okataina, Maroa, Taupo, Tongariro and Mangakino. Over the past 1.6 million years, there have been at least 25 caldera-forming eruptions. The most recent eruption was Tarawera in 1886. Taupo is the site of the most recent VEI 8 eruption on Earth. If the APVC was the dominant mover and shaker several million years ago, the TVZ has a strong case of being the current boss.
The TVZ is notable for several things. The central part of the TVZ erupts silicic magma, while at either end the magma is andesitic. Only the central part produces the huge caldera eruptions. Several different magma chambers can be active during the same eruption. Finally, there is a tendency for large eruptions here to occur in pairs, separated by a few thousand years or less: possibly, the changing stress field from one magma chamber collapse triggers a nearby chamber to activate.
There have also been numerous smaller eruptions, where ‘smaller’ does not mean ‘small’. The largest of those has been carbon-dated to between 230 and 500 AD. (The ice core record suggest 433 as a plausible date.) The pyroclastic flows of the so-called Taupo eruption reached a distance of 80 km and the plume reached 50 km height. It ejected around 120-150 km3, and was a low VEI 7. But it was dwarfed by what the TVZ does on a good day.
5. Oruanui (1170 km3)
This VEI-8 eruption occurred 26,500 years ago, in nearly the same place as the Taupo eruption mentioned above. It formed a caldera 35 km across (later further enlarged by the Taupo eruption). It ejected an estimated 530 km3 DRE. The ignimbrite was up to 200 meters deep. During th eruption, there were at least 10 separate pulses over two phases, with a hiatus of weeks to months between the first and second phase. The eruption intensified over time, and the later pulses were affected by magma-water interaction. (It may be relevant that Taupo is in a high rainfall area.) The 8th eruption was the most energetic with pyroclastic flows reaching 90 km distance. The 10th and last eruption created the most volume. The entire North Island was covered under 20 to 200 cm of ash. As far as we know, this is the most recent VEI 8 eruption on Earth.
6. Whakamura (2500 km3)
The origin of this ignimbrite sheet lies north of Lake Taupo. It was erupted between 320 and 340 thousand years ago. There were two eruptions, separated by 10 thousand years. The first phase involved 1500 km3 DRE, making it the largest of the eruptions of the TVZ. The second phase involved 500 km3, still enough to qualify as VEI-8 . The magma of the main eruption had entered the magma chamber some 300 years before the explosion: the eruption was apparently triggered by intrusion of mafic magma into an existing silicic magma chamber.
7. Kidnappers (2500 km3)
This ignimbrite sheet rivals Whakamura in size. It is dated to 1.01 million years ago and is widespread in New Zealand, both on land and off-shore, reaching almost 200 km from the source. The DRE volume is estimated as 1200 km3. The Kidnappers eruption involved three distinct magma chambers which erupted in the same event. Shortly after this event, the Rocky Hill ignimbrite (200 km3) was expelled, in a separate eruption but from the same location.
Indonesia tops the list of eruptions per year, and is also responsible for at least two VEI-7 eruptions in the past millennium. Extrapolation would suggest that the place should be littered with the scars from VEI 8 eruptions. But although there are a few cases, they are not nearly as common as might be expected. Overall there is remarkable little evidence for such eruptions. One reason may be that the tropical climate plays havoc with history; erosion from the tropical rains can quickly remove any sign of ejecta blankets. The continuous volcanic activity may also cover up evidence of old craters. There is little doubt that the ancient record of major eruptions here is very incomplete. But it also appears that it you cannot just extrapolate from number of VEI-7 eruptions to VEI 8.
8. Toba (7000 km3)
Even a casual look at the map of Sumatra shows this caldera. At 100 by 35 km, it beats the caldera of Yellowstone by length, though not by area. The eruption occurred 74 thousand years ago. The deposits inside the crater are 600 meters thick; the ignimbrite sheet covers much of Sumatra and the ash reached India and possibly Africa. The eruption volume is estimated at 2800 km3 DRE; one study finds that it may have been as large as 4000 km3. The rhyolite was ejected along the rim of the caldera, possibly as a near-continuous fissure. The eruption was short-lived, with estimates ranging from nine days to as little as 15 hours. The eruption column was up to 40km high. As is common in large calderas, there have been several major eruptions from this location: the previous eruptions are dated to 1.2 million, 840 thousand and 500 thousand years ago.
9. Tondano (est. 1000 km3)
This is the largest caldera on Sulawesi, at 20 by 30 km. It is part of the volcanic arc that forms the north arm of Sulawesi. A younger, smaller caldera is located within it, and Lake Tondano occupies part of the old caldera. The main eruption occurred 900 thousand years ago. It deposited 200 meters of rhyo-dacite pumice, nowadays called the Tondano tuff. There is no published volume for the eruption, and the number above is a guess based on the thickness and spread of the tuff.
Albert, August 2019