An eruption that ejects more than 1000 km3 of material (ash, pumice, rock…) is considered a super-eruption, a VEI-8. These represent the greatest volcanic events that have taken place during human existence. Such apocalyptic phenomena attract a lot of attention, from scientists, volcanoholics and doomsayers.
The term supervolcano has become increasingly popular but also increasingly misused. Why is the low VEI-7 Campi Flegrei being called a supervolcano? The term has the problem that it seems to imply that a so-called “supervolcano” will blow gigantic again and as pointed out by some scientists this is not true. We have volcanoes that have done super-eruptions but won’t do them again and others that have not, but will at some point in the future. So regardless of a volcano having done one of these, how do we know what actual potential a volcano can have of producing them?
Problem, as always with volcanology, is that things happen deep below where we cannot see them. A meteorologist can watch a thunderstorm being born, but the birth of a volcanic eruption takes place within the crust, often many kilometres down. Indirect geophysical and geochemical methods are all we have to study these processes. Knowledge on how volcanic caldera systems grow and collapse has been evolving recently. In light of these developments I am here to answer a personal question that some of you may have asked yourselves. Where will the next VEI-8 super-eruption of the planet take place? And since he who asks a question cannot avoid the answer (particularly if you ask it compulsively), I shall give my personal view based on the new paradigms that are emerging in volcanology.
The rollback model
Before I can answer that question we should know how caldera systems grow, in general. They do not show up randomly over the world but actually burst into clusters of systems that erupt over a short period on time. Such upticks of calderas, ignimbrites (giant pyroclastic flows), and lava domes are called silicic flare-ups. Understanding them is the first step. The vast majority of flare-ups occur in subduction zones and are related to the same process, this being the transition from flat slab subduction to steep slab subduction through slab roll-back which results in the steepening of the subduction angle. These may be, to you, a lot of new terms, but important terms when talking calderas.
Sierra Madre Occidental with more than 400,000 km3 is the mother of all flare-ups (if we ignore Australia) and can be used as a model. Western United States and México used to be in a period of flat-slab subduction called the Laramide Orogeny/Magmatism during which volcanic activity shifted away from the trench and was relatively low. The Laramide slab detached around 50 million years ago triggering a chain of events. At 45-30 million years the subducting slab underwent rollback towards the trench. Mainly andesitic volcanism and some ignimbrites retreated trenchward together with the slab and preceded the spectacular silicic flare-up at 30-20 Ma that covered much of western México in up to a kilometre of ignimbrites. This happened together with extension that eventually culminated in rifting of the Gulf of California at 18 Ma.
Let’s look more closely at this cascading sequence of events, because it seems to repeat in a remarkably similar way in most flare-ups. Where does all the magma erupted in ignimbrites and lava domes come from? The first element is water and it comes from the subducting slab. During flat-slab subduction volcanism often stops completely but water keeps seeping from the oceanic subducted crust and sediments into the overlying lithosphere mantle and continental crust, the crust becomes hydrated which lowers the melting temperature of the rock and makes it easier to turn it into magma later on.
A slab detachment usually leads to the flat-slab plunging into the mantle backwards towards the trench. This is called rollback. The growing mantle wedge creates upwelling of the hot asthenosphere, which provides heat, and decompresses as it ascends. The overlying volcanic arc is stretched which creates widespread extension. Melting results from water, heat and decompression; this is the perfect cocktail. It is interesting to note that for the Sierra Madre Occidental the proximity to the East Pacific Rise seems to play as a factor. North America is moving over the East Pacific Rise, this perhaps promotes the mantle being hotter or helps create extension. The result is that basaltic melt intrudes into the lower crust and thickens it while the lithospheric mantle melts or delaminates and is replaced by the hotter asthenospheric mantle (<1300°C). The crust in turn melts as well. This generates enormous batholiths reaching up to hundreds of kilometres across where clusters of large calderas develop, and hence flare-ups and super-eruptions. Widespread extension provides pathways for the magma to go up and can climax into the formation of a proper focused rift. The most intense flare-ups seem to be those in which the extension succeeds to progress into a rift, like the Taupo Volcanic Zone, Sierra Madre Occidental, and if we go further back in time, Australia. [caption id="attachment_11068" align="alignnone" width="681"] The subduction angle is varied. Also note the 2 prominent flat slab areas in the Andes. These are the Peruvian and Pampean flat-slabs and are areas where volcanic activity has died off. At some point they will steepen and unleash flare-ups. By Gavin P. Hayes, USGS.[/caption]
Earth’s distribution of large caldera systems is not even. Most tend to be concentrated in those areas where the transition from flat slab to steep slab has taken place. An example is the Central Andes. The Altiplano flat slab started to disappear around 30 Ma in southern Peru. The resulting flare-up is already over but was followed by steepening of the Altiplano flat slab further south at 12 Ma. This second episode created the remarkable Altiplano-Puna batholith and related calderas, among others. The flare-up is still active although in apparent decline.
The next largest active flare-up is taking place in New Zealand, here the Taupo Volcanic Zone is related to a rollback affecting the entire Tonga-Kermadec arc that has resulted in a 3000 km long back-arc rift (the Lau Basin). The Taupo Rift is just its southern continental end. At its northern end the Lau Basin is spreading at 15 cm/year as the Tonga Arc moves over the Pacific Plate in the fastest subduction of the planet (24 cm/year). However only the continental crust of New Zealand has resulted in significant ignimbrite activity, here the volcanic arc migrated east, towards the trench, over the last 12 million years, presumably in a transition from flat slab to steep slab subduction, though I haven’t found this mentioned anywhere. Andesitic volcanism was followed by ignimbrites and calderas starting at 1.6 Ma concentrated over a 120 km segment of rift. Despite the relatively small area and short duration, the Central Taupo Rift has already produced a volume of around 20,000 km3. It can be considered to be at its peak strength.
While these are the 2 major flare-ups of our times there are also a number of smaller ones active throughout the world. Southern Kyushu, Japan, is another good example, where a brief steepening of the subduction at 5-3 Ma has resulted in a trenchward migration of the volcanic arc and extension. A chain of calderas (Kikai, Ata, Aira and Kakuto) line up with the Kagoshima Graben.
Back to the Andes, but further south, we find the Payenia flat slab, which stopped being flat some time 5-2 Ma ago. This one is quite remarkable because it has resulted in the formation of a mini flood basalt. Over the past 2 million years 8400 km3 of basalt have been erupted in the back-arc through shield volcanoes and volcanic fields while large calderas and ignimbrites erupted through a series of grabens in the volcanic arc. The basaltic volcanism is thought to be a result of the asthenosphere upwelling in response to the slab rollback and a shallow mantle plume has been imaged below Payún Matrú shield volcano. The rollback process seems complete except perhaps in the northern parts of the slab. Here the Calabozos and Laguna del Maule large caldera systems remain very active. The latter is currently inflating at more than 20 cm/year, faster than any other silicic system on Earth.
Continental rifts and hotspots
Almost all major silicic volcanic provinces can be related to the steepening of subduction and ensuing extension/rifting. This probably applies, I suspect, to areas of intense ignimbrite activity in Hokkaido (Japan) and the Eastern Volcanic Zone of Kamchatka, although I couldn’t find scientific literature on this subject. However there are other settings in which calderas form. The key element as we will later see is melting of large volumes of the crust and the formation of batholiths.
The Yellowstone hotspot has produced at least 11 super-eruptions throughout its history. Here the main element is heat. Yellowstone shows that a powerful hotspot can manage to melt the crust and generate super-eruptions. Then why is it such a rare case? The reason probably is that Yellowstone is the only strong deep-seated mantle plume located under continental crust, the others are located in the ocean where they form shield volcanoes instead.
A third situation that can result in large calderas is continental rifting. This is shown by the East African Rift where many VEI-7 sized calderas and at least 1 VEI-8, Awasa Caldera, occur. In East Africa melting is due to decompression, but also to an anomalously hot mantle related to the Afar Plume, which dramatically showed up 30 million years ago kick-starting rifting.
And then we find the Tibesti Mountains which form a cluster of sizable calderas (VEI-7) in the middle of the Sahara, far from rifts, hotspots or subduction zones. So how do we explain this oddball? My guess is probably extension and asthenosphere mantle upwelling which often results in basaltic volcanic fields but in extreme cases may evolve into something more exciting than that.
A final interesting kind of silicic flare-ups that I see are those which happen in small localized rifts within subduction zones. These include the Macolod Corridor in the Philippines, the Managua Graben in Nicaragua, the San Salvador Graben in El Salvador and the Bay of Naples in Italy. They have some things in common that may be coincidental or not. Each has pair of one or two silicic caldera systems next to a mafic caldera system: Laguna de Bay-Taal, Apoyeque-Masaya, Ilopango-San Salvador and Campi Flegrei-Vesuvius. Each happens to have a city of more than 2 million inhabitants in it, which makes all 4 of these volcanic areas among the most hazardous on Earth.
So we have seen the main settings and locations where large calderas form but now it is time to dissect a particular system and see how it works in the inside.
Both observations and models are showing that the old magma chamber idea falls short to capture the complexity of a volcano’s internal plumbing. Instead it turns out that volcanic systems are formed by extensive networks of melt lenses and areas of crystal mush that extend from the mantle to the surface in most cases. What is a crystal mush? This term refers to rock that is only partially molten, with less than 50 % of melt, and the rest is solid crystals. You can picture it like a sponge of crystals that is soaked with melt. Most of the volume stored within a volcano is probably in this state, and mush is not eruptible. The actual magma is ponding in layers or lenses. Multiple such layers can exist from the MOHO all the way to the surface and containing magma of different chemistries.
Different kinds of volcanoes have different internal structures. A spreading ridge and a large caldera are going to have completely different insides! A spreading ridge for example has a mush of basaltic composition (gabbro) that extends vertically only the few kilometres from the MOHO to the brittle crust. The mush will be narrow but long, extending below the axis of the spreading ridge for up to hundreds of kilometres. Eruptible magmas will form only very small lenses at the top of the mush that feed dikes and small eruptions. On the other hand a large caldera system will have mush bodies that extend vertically over tens of kilometres, horizontally tens or even hundreds of kilometres, both long and wide, and which can host many magma lenses of various chemical compositions, with some containing very large volumes. We can use the Altiplano-Puna of the Andes as the model for a large caldera plumbing system.
The Altiplano-Puna system has all the rights to be called a batholith, the largest of intrusions. It connects about 10 large caldera systems and many tens of stratovolcanoes down to a large single mush body that is about 350×180 km horizontally. Geophysical and geochemical observations reveal that the inner structure is layered like a cake, with 3 main levels of storage.
The first is formed by melting of the mantle and its intrusion and ponding in the lower crust, which forms a mush zone at 85-40 km depth. The scientific term for these zones is MASH, which stands for Melting, Assimilation, Storage and Homogenization. This first level is basaltic, mafic and ultramafic in composition. The second level is the famous Altiplano-Puna Magma Body (APMB) which is a 500,000 km3 body of andesite with up to 25 % melt, a roof at a depth of 15 km bsl and a temperature of 1000 °C. The active stratovolcanoes feed from this body. The third level is formed by a series of shallow reservoirs that sort of branch upwards from the AMPB. These are located 3-8 km deep and contain dacites and rhyolites that fed past super-eruptions, ignimbrites and lava domes. The shallow reservoirs were active mostly 8-4 million years ago during a series of activity pulses, however mantle melting seems to have declined afterwards and these reservoirs are mostly frozen. So much for the Altiplano-Puna!
Models show that these systems probably develop from the bottom up, and the eruptive history of many large caldera systems, like Taupo, seems to support this, because andesitic volcanism precedes rhyolitic, or in other words a deeper mush zone that feeds andesite eruptions forms before shallow reservoirs that feed the rhyolite eruptions. Models predict that the lower crust needs to melt first in order to maintain a higher heat flux into the upper crust that would allow reservoirs to form there. So it seems reasonable to believe that caldera systems in general grow upwards from the mantle.
How magma travels upwards is also important. Magma can intrude through cracking, the formation of sills and dykes. Another way is porous flow where melt travels up through the mush using the spaces in between crystals, but this is very slow and only works over long timescales. More intriguing are overturning events in which an amount of melt builds up in a certain level and becomes unstable due to buoyancy so that it moves up to an overlying level through a diapir/channel/dyke. These events could explain the evidence for magma chambers being generated very quickly before an eruption. Overturning events could deliver large intrusions of melt proportional to the size of the body/MASH that the melt originates from, this could rapidly assemble melt at the top of the system, increase temperature, turn rock and mush into new melt and create overpressure due to decompression of gasses.
The size of the shallow reservoir is what determines if a volcano is capable of producing VEI-7 or VEI-8 super-eruptions. There needs to be a sufficient volume of eruptible dacite or rhyolite magma built up. The calderas that result from super-eruptions are very wide but with a downdrop that is not too different from smaller calderas. This shows reservoirs are relatively flat, sill-like, and that to build up enough volume they must be laterally extensive. For example stratovolcanoes have mush systems that are vertically extensive because they reach to the surface, but are often very narrow so they tend collapse into VEI-6 to low VEI-7 calderas. Instead the largest calderas would be expected to form from volcanic fields and broad stratovolcano complexes that are wider in area.
Caldera systems are in fact closely related to volcanic fields, this is why they themselves are often called volcanic fields. Melting of the lower crust or uppermost mantle over a wide area would reflect at the surface as a basaltic volcanic field, Mayotte is an example of such a system. Here a dyke in 2018 intruded directly from the MOHO (mantle) and resulted in the formation of a new vent, a monogenetic volcano that only erupts once. Melting of the upper crust instead results in the formation of andesitic, dacitic or rhyolitic volcanic fields which are much more rare. Los Humeros in México is an example of a system that seems to have evolved from an andesite VF, to a rhyolite VF and later collapsing into a large caldera.
Andesitic volcanic fields are a relatively rare occurrence in the world and may show the areas where magma bodies similar to the Altiplano-Puna are forming which gives an idea where new flare-ups are emerging. BY FAR the most spectacular ones are located in Mexico.
In 1943 a new volcano emerged from a cornfield in the State of Michoacán, México, this volcano would come to be known as Paricutín. Early this year, 2020, a volcano-tectonic swarm took place in this area which has been later been interpreted as a sill intrusion, reminding how active this location is. If we go back to 1250 AD a 9 km3 monogenetic shield (El Metate) formed close to the current location of Paricutin. This is just one part of the Trans-Mexican Volcanic Belt that is everywhere covered in volcanic fields of dominantly andesite lava. In some areas monogenetic shields cover the entire landscape so tightly that they overap on each other, what does this mean? It means that the upper crust is melting almost everywhere along the 1000 km of the Trans-Mexican Belt.
If we look at Mexico it has all the ingredients. The lithospheric mantle has almost disappeared (considered to have melted) below the active volcanic arc so that hotter than 1300 degrees asthenosphere lies directly below the crust. Rollback of a flat slab is taking place and will still take some millions of years to be completed. Volcanoes are running southward as the slab retreats. Chains of stratovolcanoes such as Tlaloc-Telapon-Iztacciuatl-Popocatepetl, are being created as the stratovolcano chases the sweet spot of water release at which the slab sinks to 110 km; this leaves curious mini hotspot-like trails. Additionally México is under widespread extension with grabens popping all over the place, and there is the threat of the East Pacific Rise jumping into México like it did at the Gulf of California, perhaps bringing a Sierra Madre Occidental all over again!
Another promising area is the western United States. Rollback has been taking place over the past 10 million years in Oregon and Washington with a westward migration in volcanism. Some sizable volcanic fields dot the area with monogenetic shields. The nearby East Pacific Rise could try to jump eastward onto the continent and long Iceland-like fissures around Three Sisters hint at rifting. Newberry, Shasta, Medicine Lake, Lassen and Three Sisters all look like potential VEI-7 volcanoes to me and each could pull off a Crater Lake. The hot mantle and widespread extension also give birth to silicic systems at unexpected locations, like Coso or Clear Lake. So this is another emerging flare-up.
The very largest
By this point I must have mentioned nearly all flare-ups of the world. Those that were, those that are and even those that will come. It is from these areas that many VEI-7 eruptions will originate in the future, but also the even greater almighty super-eruptions will come from them. But which volcano shall be the next VEI-8. This question will be answered in upcoming articles!
Research and further reading
Mike Poland wants to end the supervolcano word:
Rollback and back-arc basin opening recreated in an animated model:
Sierra Madre Occidental was followed by opening of the Gulf of California:
Andes likes to alternate flat slab with steep slab:
Meet the Payenian flare-up and flood basalt:
Mushes, magmas, and even more mushes:
Bottom-up growth of magmatic systems:
The extraordinary history of the Trans-Mexican Belt:
Rotating the Cascades: