The VC list of 25 super eruptions – continued
In our quest for major eruptions, we are continuing our journey around the world, moving north from Indonesia.
Japan’s southernmost main island is volcanically highly active. Past explosions have left large calderas, separated into two groups. In the centre of Kyushu is Aso, and it lies along a NE-SW graben. In the south are Kakuto, Aira, Ata and (submerged) Kikai, along the Kagoshima graben. Geologically, the island is caught in a fight between three plates (Eurasia, the Philippine plate, and further to the east the Pacific plate; the last two are subducting) and it is rotating under the pressure. This rotation has caused faults and grabens to form, and the subducting Philippine plate, which is quite young and warm, has been feeding magma into these zones for about 1 million years.
10. Aso (600 km3)
The most powerful eruptions of Kyushu have come from Aso. It also made the VC NDVP list. The caldera measures 25 by 18 km across. There have been four significant eruptions here, conveniently numbered as Aso-1 (265 thousand years ago, 50 km3), Aso-2 (140 thousand years ago, 50 km3), Aso-3 (125 thousand years ago, 100 km3, and Aso-4 (89 thousand years ago, 600 km3). (Erupted volumes, not DRE.) The last two volumes may have been underestimated. Aso-4 was by far the largest of the four eruptions. It left ash 15 cm deep at a distance of 1700 km from the explosion; pyroclastic flows covered the northern and central parts of Kyushu and even reached Japan’s main island, Honshu. It produced mixed magmas, partly rhyolitic and partly dacitic. It appears that mafic magma was injected into the magma chamber shortly before the eruption, at a depth of 14 km. Post caldera eruptions became basaltic.
11. Aira (463 km3)
This is located at the southern end of Kyushu. It lies at the top end of Kagoshima Bay, which is formed by a rifting back arc. The 17 by 23 km caldera and was produced 29 thousand years ago. The well-known Sakurajima volcano grew up on the rim of this caldera, a common event after caldera-forming eruptions. At the time this was a shallow water basin, separated from the bay by a ridge. (The bay was at the same level as now but sea level was 100 meter lower.) This was a double eruption, coming from pre-Sakurajima and from the central caldera. The eruption had an initial Plinian phase with pyroclastic flows, already reaching over 100 km3, which erupted from a vent near Sakurajima. The central vent exploded afterwards, erupting 300 km3; The second event created a much larger pyroclastic flows. Within a radius of 70 km, 40% of the land area was covered in 10 meter of pyroclastics. Compared to many other eruptions on the list, this was a highly explosive event.
12. Kucharo (175 km3)
Moving away from Kyushu to the north across Japan, we find this large caldera in Hokkaido which measures 20 by 26 km. There have been 8 caldera-forming eruptions here between 400 thousand years ago and 40 thousand years ago. The largest one of these occurred 120 thousand years ago, and it is estimated to have had an erupted volume of 175 km3 – a bit small for our list but large by any other measure! However, the caldera is very large and this suggests the possibility that the erupted volume has been underestimated. The eruption came from three separate vents, two purely rhyolitic and one with an additional mafic component. It began with a Plininan eruption but the main phase that followed had no strong Plinian phase. As in other cases discussed here, it is suggested that the eruption was triggered by injection of mafic magma into a silicic magma chamber.
Russia’s far east
This area is highly volcanic but remains under-studied. There is a fair knowledge on holocene eruptions but little is known about older activity. Kamchatka has the highest number of calderas per length along the volcanic arc in the world. Many show evidence of multiple eruptions. This is a high-risk area for large eruptions.
13. Opala IV (250 km3)
There have been four major eruptions from this volcanic field at the southern tip of the Kamchatka peninsula. The most recent, conveniently called ‘Opala IV’, was the largest. It is dated to 40 thousand years ago. The caldera is 12 by 14 km. The current Opala volcano is located at the northern edge of this caldera. The volcanoes here are basaltic to andesitic, however the ejecta of the Opala IV eruption are rhyolitic.The size of this eruption is poorly constrained, as neither the height of the stratovolcano nor the depth of the young caldera are known. Ignimbrites from this eruption are found throughout southern Kamchatka, over an area of 1800 km2. The ash was blow towards the NNE, where the ash was 25 cm thick at a distance of 300 km. The size of the eruption is uncertain, but it is believed to be the largest eruption in Kamchatka for 50 thousand years.
14. Pauzhetka caldera (400 km3)
Going back further in time, we find this caldera, also on the southern tip of Kamchatka peninsula. It measures 27 by 18 km across and is classed as a volcano-tectonic depression. The centre of the caldera shows a resurgent dome; around this dome are areas of hydrothermal activity. The eruption that formed the caldera 420 thousand years ago covered much of the southern part of Kamchatka in the so-called Golygin ignimbrite. The volume is derived both from the ignimbrite and from the volume of the depression. Two thirds of the ejecta was deposited near the crater and the rest was blown southeast towards and across the Pacific Ocean.
After the Kuriles we arrive at the Aleutian arc and Alaska. This area has had frequent and very significant eruptions, but not quite large enough for us. Moving on, we find the western continental US. And it is a whopper.
The Snake River – Yellowstone volcanic province
The western US has had enormous eruptions starting more than 30 million years ago. The older calderas run from Nevada to New Mexico. The more recent ones cluster around the Snake River, along a hotspot track. This hotspot has made history. It began with the Columbia flood basalt around 18 million years ago. After it moved further in-land, it left a series of calderas, culminating in three major eruptions at its most recent location, Yellowstone. Yellowstone is the largest known caldera on Earth, with a diameter of 85 by 45 km. There were three major eruptions here, two of which qualify for our list. The other one, Mesa Falls, erupted 700 km3 which makes it a high-VEI-7, impressive and destructive but smaller than the other two. Yellowstone produces both basaltic and rhyolitic eruptions. The large explosions produce rhyolite.
15. Lava Creek (1000km3)
This was the most recent of the three major eruptions of Yellowstone. It is dated to 630 thousand years and spread ash across much of the US. The deposits are 200 meters deep around the caldera. Recent research has shown that the eruption consisted of up to four distinct events, where two of the events were separated by enough time that the earlier ejecta had cooled. In the case of the Katmai eruption (1912), such cooling took two decades. The time interval between the events of the Lava Creek may have been as long as 170 years, sufficient for it to be considered two separate eruptions. Analysis of crystals show that the magma reservoir received fresh heat in the period of 1000 to 10,000 years before the eruption, suggesting that the eruption was triggered by an influx of new magma into a cooling reservoir.
16. Huckleberry Ridge (2500 km3)
This was the first large Yellowstone event. The eruption occurred 2.1 million years ago and it blanketed half the US. The eruption occurred in three pulses. The third pulse may have been a separate event as it was fed from a different magma source. Recent research shows evidence for reworking and cooling of the deposits between the phases, and suggest that the eruption may have been episodic, over years to decades. (Elliott Swallow, in a 2019 paper, raises the question how close together in time eruptions need to be to count as one event. For current eruptions, a year may be too long, while for ancient eruptions, events a century apart may still be viewed as one event.) The reported volume includes all three pulses. If the last one is excluded, the total drops to ‘only’ 2000 km3 ejecta.
17. Kilgore (1800 km3)
Before Yellowstone, the hot spot lingered at the Heise volcanic field, along the Snake river valley. Eruptions here occurred between 6.6 and 4 million years ago. There were a number of basaltic lava flows, and four rhyolitic explosions left ignimbrite deposits. The last of these four is Kilgore, dated to 4.5 million years ago and it is believed to be the largest: the tuff is found across Idaho, Montana and Wyoming. The thickest ignimbrites are a few hundred meters deep. There is no known caldera in the Heise field: the area was covered in deep lava flows some time after the major eruption. The lack of knowledge on the caldera means that the eruption volume is uncertain. The number given here is conservative, but on the other hand assumes that it was generated in a single eruption.
California to Colorado
The Snake river volcanics are hot spot related. But in fact the western US has had similarly-sized eruptions also in places without hot spot activity. The focus of activity was around Nevada, around 30 million years ago. This includes the Wah Wah Springs eruption which is often included in VEI charts. This particular eruption I did not include because of age, and because of a lack of independent confirmation of the size and nature of the eruption. That is not a great loss, as there are other similarly-sized candidates in the region.
18. La Garita (7500 km3)
This eruption is considerably older than others we have looked at, but is included because of its size. It is sometimes claimed as the largest eruption (VEI) known. It is the source of the Fish River ash, and has been dated to 28.1 million years ago. The caldera is located in the San Juan Mountains in southwestern Colorado: it is one of a cluster of calderas within the San Juan volcanic field. A volcanic shield once covered much of the southern Rocky mountains but it has eroded a lot. The field lies at the head of the developing Rio Grande rift, a region of extension. The caldera measures 35 by 75 km and it formed in three steps during the La Garita eruption. The eruption was dacitic, and produced an ash flow 50 to 300 meters thick, as far as 100 km from the caldera.
The volume is often quoted as 5000 km3 (DRE). This is based on estimates by Lipman, who assumes that the outflow extends to 75 km at an average thickness of 100 meter, and adding 1 km depth of deposits inside the caldera. If I instead assume that the thickness linearly declined with distance, and adding the caldera in-fill, I find 3000 km3 DRE, or 7500 km3 ‘VEI ejecta’. This illustrates the uncertainties that are implicit in deriving volumes for ancient eruptions, and the need to be cautious about the most extreme values. Even with the lower number it is one of the largest eruptions on our list, but it may not qualify for VEI 9!
19. Monotony tuff (7500 km3)
But La Garita did not stand alone. The Monotony tuff layer is found over much of Nevada and is part of the same series of Nevada ignimbrite eruptions 30 million years ago. The region has many similarities to the more recent activity in the Altoplano. At the time this was a high plateau with thick crust (60 km), above a subduction layer. It is sometimes called the Great Basin altiplano. The monotony tuff is dated to 27 million years ago. It is identified with a caldera which may be as large as 75 km (gravity measurements suggests it may be a composite) and the composition of the tuff is remarkably uniform over the full range. The thickness reaches 500 meters in places. The volume is uncertain but I have used one of the more conservative estimates (3000 km3 DRE). The uniformity makes it likely that this came from a single eruption. It rivals, and possibly exceeds, La Garita.
20. Long Valley (600 km3)
This eruption is separate from either the Snake River or Nevada events. The caldera is in eastern California, along the Sierra Nevada. It measures 17 by 32 km. The eruption that formed the caldera happened 760 thousand years ago. It formed the so-called Bishop tuff, erupted in a 90-hour long, Plinian event with an eruption rate which increased with time. The eruption may have migrated along the ring fault. The eruption was rhyolitic. However, the same volcanic system has also produced basaltic eruptions: this mix is quite a common feature of these large eruptions. The magma chamber had been present for over a hundred thousand years, in a cool, crystal-rich state, but had been reheated into an eruptible state shortly before (perhaps by a thousand years) before the event.
21. Atitlán (1200 km3)
This famous Guatemalan location has been active for 14 million years: Atitlán may stake a claim to being the oldest volcano on the planet which is still erupting. It is very unusual for one region to show volcanic activity over such a long period. There have been 3 eruption cycles leading to caldera collapses. During the first eruptive cycle between 14 and 11 million years ago, what is commonly called the María Tecún Tuffs was ejected, with a total ejected volume of 2800 cubic kilometres. It created a caldera 15 by 25 km. The tuff forms five separate layers which were erupted at different times: two of the layers show a magnetic pole reversal compared to the others. The largest of the layers (MTT II), 11.6 million years ago, produced ash covering most of the United States, and created a solidified tuff sheet up to 700 metres thick. Using road cuts, drill samples from mines in Mexico, and distal ash finds in Pensacola, Florida, it was calculated to have been around 1200 cubic kilometres.
After the eruption a new cycle of volcanic growth and caldera formation began. The Atitlán II Caldera formed during the San Jorge ash flows 10 – 8 million years ago, a series of VEI-6 eruptions. One million years ago the current Atitlán III Caldera formed during the 350 cubic kilometres Los Chocoyos eruption. It confirms the pattern that major calderas are repeat offenders.
The most famous major eruption in Europe is that of Santorini (not Vesuvius which was locally devastating but globally not as significant). But this volcano is also a repeat offender: it has had four caldera-forming eruptions, 180,000 years, 70,000 years, 21,000 years and 3600 years BP. The final eruption destroyed a local civilization, caused a devastating tsunami, and deposited up to 80 meters ash, blown mainly towards Turkey. The four eruptions were probably of similar size. However, these eruptions were around 100 km3, too small for our list in spite of their fame. Europe lacks VEI-7 eruptions. There is however one European eruption which can be added to our list.
22. Campei Flegrei (250 km3)
The largest eruption known in Europe is Naples’ Campi Flegrei event, 39 thousand years ago. It deposited the Campanian ignimbrite across Europe, Russia and North Africa. There was an initial phase with a volcanic column at least 40 km high, followed by collapse and creation of the caldera, a rejuvenated volcanic column, and widespread ignimbrite deposition extending 1500 km from the eruption point.
The caldera has a maximum diameter of 13 km, not quite in the same league as the very largest eruptions. However, it devastated a continent (admittedly Europe is not the largest of continents). 250 km3 of ash was ejected. On the opposite side of the Bay of Naples, 35 km away, the pyroclastic deposits are 40 meters thick. The eruption was followed by a second event 10 thousand years later which formed the Masseria del Monte Tuff: this ejected 16 km3 (DRE). There was a third event of similar size 15 thousand years ago.
The African rift
The volcanic history of Africa is not well known although volcanoes are present along the African Rift. Ethiopia in particular may be the proud owner of the most underestimated volcanic risk in the world. Here, over 10 million people leave within 30 km of a volcano that has erupted within the holocene. And it has some impressive, but little studied, scars.
23. Awassa (1000 km3)
Awassa is the largest of these scars, an eroded caldera which measures 40 by 30 km, partly overlapping with the more recent Corbetti caldera. It is near the Africa rift valley. The caldera is intensely cultivated. The deepest part is occupied by Lake Awassa, the smallest of the Rift valley lakes. Nearby Aluto had a major explosive eruption 310 thousand years ago but this was dwarfed by Awassa. Vogripa lists it as 1000 km3. Along the western wall of the caldera, a major fault exposes more than 250 m of tephra. The main eruption is dated to 1.27 million years ago.
24. Munesa crystal tuff (1100 km3)
This a tuff layer between 250 and 400 meters thick, prominent along both sides of the rift valley in Ethiopia, around Guraghe. The originating caldera is missing; it may have been buried underneath lava flows within the rift valley. The tuff is dated to 3.5 million years ago. The tuff was capped by a lava flow.
25. Ngorongoro Crater (500 km3)
Eruptions can be powerful by volume and energy. But they can also be powerful in what they create. The best example of that is Ngorongoro Crater on Tanzania’s Serengeti. There are three other calderas nearby, all natural wonders, but Ngorongoro tops them all. It is reputedly the largest complete caldera in the world, and home to the best of Africa’s wildlife. It is 18 km across, and is the remnant of a shield volcano. The volcano was active between 2.3 and 2.0 million years ago, when it was within the African rift: the area has since been uplifted. Very little is know about the climactic eruption: it is likely to have produced the Naabi ignimbrite, a 20-meter thick layer in Olduvai Gorge, but the size (and how much was rifted lava) is unknown. Based on caldera size, the erupted volume may have been up to 500 km3.
This list of 25 began with the desolate magnificence of the Altiplano, and ended with the most evocative caldera on the planet. What can we learn from this list? The list is itself incomplete, and every reader will know of one or more eruptions that could or should have been included. A VEI 7 eruption may happen perhaps twice per millennium. Over the time span of several million years covered in this list (ignoring a few very ancient ones), there may have been 5000 VEI-7 eruptions. A list of 25 just doesn’t get close.
Accepting that this list is incomplete and badly biassed, we can still point at a few interesting aspects.
Very few of the listed eruptions are in the tropics. That is unlikely to be a real effect. Toba shows that the tropics can do anything that the rest of the world can. Why is Central America only in our list through an ancient eruption? What happened to the tropical eruptions? The most likely cause is that the high rate of erosion removes the signs of ancient calderas. The best cases of old eruptions are found at dry mid-latitudes, especially in the high Andes where rain is a rare event and scars survive almost forever. Elsewhere, the book of volcanic history is erased year by year.
Many of the calderas in this list are repeat offenders. A location that managed a VEI 8 has a high probability of doing it again, albeit typically after a wait of 100 thousand to a million years. The waiting time is in part because of large amount of magma that needs replacing. At the top end, for a high supply rate of 0.01 km3/yr, it takes 100 thousand years to provide the 1000 km3 of magma – this is enough for a large VEI 8. This is the minimum time. Taupo is an example of such a fast system. Other areas take longer to build up to a VEI 8. Several conclusions can be drawn:
(1) Magma supply rates of super eruptions are no higher than that of other volcanoes. High magma supply gives frequent eruptions – not larger ones. When an eruption happens depends on the magma pressure versus the strength of the chamber roof. A faster magma supply rate just causes an earlier eruption.
(2) The cyclic eruptions show that super-eruptions do not completely destroy the magma chamber(s), nor the feeding conduit. This is easy to understand for trapdoor calderas, and makes sense for caldera collapse, but is more difficult for massively explosive eruption. Indeed, there is limited evidence for major explosions: ejecta from Plinian eruptions are often minor or even absent in these eruptions. The main eruption mechanism may be fountaining along the caldera rim.
(3) The magma indeed stays in the chambers for a very long time. The evidence for this comes from the large fraction of crystals in the ignimbrite. The magma chambers appear to be fairly cool, around 700-900 C, but able to keep this temperature for a million years. Some of the ignimbrites show evidence for episodic heating, including shortly before the eruption. In those cases the super eruption may be triggered by intrusion of fresh, hot magma into the massive chamber, either remobilising the stale rhyolite, or re-melting the crystal mush.
We can expand on this. The locations of the super eruptions tend to show two different eruption types: smaller mafic eruptions, and massive silicic ones. The magma evolves to silicic during the long wait in the chamber. The magma originates as mantle melt, which over time incorporates crustal melt. The magma chamber may have supported a flood basalt eruption, where again it takes a million years before the silicic super eruptions may start (remember that flood basalt eruptions themselves were excluded from this list.) But many eruptions show evidence of mixed magma, where the eruption is finally triggered by injection of new, hot (mafic) magma into the mushy reservoir. There are several cases where eruptions come from several magma chambers, possibly with some intervening time. The reheating affects several chambers in the region, not just one. This heating of an entire region may help explain the exceptional size of the calderas.
The eruption rates deserve a further look. The super eruptions appear not to be particularly long lasting. They are not like Eldgja or Laki, lasting months to a year: the duration of the overall event is limited. Toba is a good example: this eruption appears to have lasted little more than a week. But this requires exceptionally high eruption rate. For Toba, the average rate was of the order of 10 km3 (DRE) per hour. Instantaneous rates during eruption pulses may have been much higher. These rates are far too high to maintain a tall convective column: the weight of the ash collapses the column of air. This limits the height the ejecta can reach. Some of the ash reaches altitudes of 40 km, but the majority stays well below this. The collapsing columns may in fact be the cause of the huge ignimbrite sheets. Super-eruptions come in a variety of shapes and sizes, but they should not be seen as scaled-up versions of Krakatau.
What causes super eruptions? The distribution across the Earth shows that the largest such eruptions happen where oceanic crust subducts under very thick continental crust. This happens under the Altiplano where the crust has thickened to 70 km. It is also seen at Yellowstone, and in the past under Nevada. The oceanic crust provides the melt, and the thick crust provides the extremely strong lid which can hold the pressure for a million years. The oceanic crust is an essential ingredient, as shown by the fact that the equally thick crust of Tibet does not show super eruptions – its crust comes from a collision of two continents. As an aside, this also indicates that the hot spot of Yellowstone does not on its own explain the super eruptions there. The subduction and thickness of the crust are also important ingredients.
The second type of location where super eruptions can occur are in areas of extension within continental crust. This is the case in Taupo, Kyushu, and Ethiopia, all of which are developing rifts. Naples, although smaller, may also fit this type. Toba is also located in an area of some extension, and is located on the Sunda continental plate. This also explains why Sumatra has these eruptions, but islands east of Java do not, as they are not part of the Sunda plate.
The listed eruptions in Russia do not fit either description. Alaska could be included in this. These eruptions are smaller (VEI 7), more recent (i.e., more frequent), and are in regions of very fast subduction. It appears that this is the pattern of volcanic arc: frequent, large, but not quite super. Perhaps the magma chambers cannot grow to sufficient size in the absence of sufficient continental crust.
Our search for the elusive super eruptions has thrown up some interesting findings. These events repeat in the same location over a few hundred thousand to a million years. They favour locations were there are at least two out of three conditions: subduction, excessively thick crust, and rifting. These conditions allows a magma chamber to mature for a million years. Of the ones we selected, only Yellowstone coincides with a hot spot, so this is not an essential ingredient. The hot spots of Iceland and Hawaii did not feature in the list. Finally, even though we called these ‘super’ eruptions, the events tend not to be strongly explosive. This is probably related to the lack of explosive power: the chamber roof does not collapse, and neither the magma chamber nor the conduit are damaged in the eruption. The sheer size and eruption rate may prohibit strong explosions. This makes super eruptions survivable (at a distance, at least). Some newspapers might be disappointed.
At the end of our journey around the world in 25 eruptions, we return to the story of Mikhail Goldstein. There is a link between volcanic war zones and music. In 2010, the violinist Eric Silberger was traveling to take part in the Menuhin competition in Norway. He never arrived. His flight was cancelled following the eruption of Eyjafjallajökull, and instead of playing a competition in Norway, he ended up entertaining stranded passengers in Newark airport in New York. In 2014, he finally managed to settle the score, when he became the first violinist to play inside a volcano. Silberger performed Paganini’s 24th Caprice in A minor, inside a cave in Thrihnukagigur, a dormant volcano near Reykjavík. It has the reputation of being an exceptionally hard piece to play: it must have seemed appropriate for the location.
Reports say he gave a fiery performance. Whether it was a super performance is not recorded. But just like Goldstein silenced one of the worst battles of the second world war, Thrihnukagigur has not erupted since. In the battle between music and volcanoes, the score is a draw.
Albert, August 2019