Super-eruptions are rare. That is probably a good thing, even for Tallis. Too many things can go wrong for us in a supereruption. If you need a reminder, a super-eruption is often defined as a VEI-8, erupting more than 1000 km³ of ‘stuff’, leaving a crater 20 kilometers or more in diameter. Hector has written about volcanoes with super-potential: ten volcanoes with super-eruption potential part i/ and part ii . An overview of volcanoes that have done these things in the past can be found in Power of the past part i and part ii. Enjoy!

These are not eruptions like we have seen in modern times. The two largest eruptions of the past 2000 years are Tambora in 1815, and Taupo around 200 AD. They ejected around 150 km³, decent VEI-7’s but nowhere near the magic VEI-8 border. The most recent VEI-8 is the Oruanui eruption (1170 km³), which happened 26,500 years ago. It left a caldera 35 km across. If you now wonder where on Earth this was, think New Zealand. It was almost the same place as the Taupo eruption of 1800 years ago. All of the North Island was covered in tephra between 20 cm and 2 meters deep, and nearer the eruption as much as 200 meters. New Zealand seems such a nice country – but of all the nations in the world, it is the most likely to attack us with an oversized eruption. (No worries, we have our man on location tasked with keeping an ear to the ground.)

How would a super-eruption affect us? First to know is that they are not always like the run-of-the-mill VEI-7 events. They are big but not necessarily as explosive as smaller eruptions. The Oruanui eruption was a series of perphaps 10 events which together accounted for the ejecta. Eruption rates must have been very large, but you can have too much of a good thing: too much tephra suppresses the rising convection that carries dust to the stratosphere: the weight of the tephra is too much and the rising column collapses, instead forming pyroclastic flows. Even the enormity of the Toba eruption deposited most of its ash in the region between Indonesia and India, and left eastern Indonesia much less affected. Fountains will still reached enormous, almost Jesperian, heights, of course.

The local area, say within a few hundred kilometers, would be uninhabitable for a long time. Sulphur would kill all life in an even wider region. There could be other toxins: fluorine and mercury poissoning are possible over large areas. And we would expect a massive impact on climate, with global cooling far in excess of the year-without-summer of Tambora. (But not in the US, which apparently has decided that it is immune to climate.) No, a VEI-8 would not be fun. It is a good thing they are so rare.

Toba, the largest eruption of the past million year, happened 73,700 years ago (give or take 600 years). It has been blamed for the ice age and for a genetic bottleneck in humanity. However, in both cases the evidence is insufficient for a conviction. The ice age was quite happily progressing for 100,000 years, and clearly did not need Toba. (Similarly, the suggestion that the Laacher See caused the Younger Dryas did not survive when improved dating showed the innocence of the eruption.) And people re-occupied India very quickly after Toba, while if the world population had really been down to a few thousand, repopulation would have taken far longer. We know that large eruptions can cause years of cooling, but predictions of much longer-lasting cooling may be mainly due to bad press. On the other hand, the link between extreme warming and flood basalts is well established. These effusive eruptions dwarf even a VEI-8 and emit large amounts of CO2. Hence the silent spring of the Permian mass extinction.

A recent paper has studied the climate impact of a lesser known, relatively recent super-eruption. It was located not in distant New Zealand, but around the corner, in Guatemala.

The Los Chocoyos eruption

Lake Atitlan is one of the (many) volcanic wonders of Guatemala. The lake is elongated, measuring 18 by 8 km. It is flanked by three volcanoes, of which the namesake Volcan Atitlan is the youngest and most active: it formed in the holocene. Toliman is older and dormant, and San Pedro is extinct. Remove these volcanoes, and a more circular lake appears. This lake formed in a major eruption. It is called the  Los Chocoyos eruption (literally ‘ the Parakeets’). The eruption ejected 1200 km³, a solid VEI-8 and almost 10 times larger than any eruption of the past 2000 years. At a time when every 3^rd full Moon is called a ‘supermoon’, we can surely call this a super-eruption.

So when did it happen? There were smaller eruptions from the location 54,000 and 130,000 years ago: the ejecta of the large event are sandwiched in between. That still leaves a lot of time. The eruption has been dated to 84,000 years ago but with significant uncertainties. In 2021, a paper appeared which used radioactive dating of zircon crystals in the ejecta. These give the date when the magma crystalized, presumably close to the time of the eruption. They date the eruption to 75,000 years ago. That is within a few thousand years of Toba! Ift would mean that the world had two super-eruptions in the amount of time in which we have had ‘just’ two Tambora-sized events. It was a different world in those days.

Is there corroborating evidence? That brings us to the recent paper on this eruption, by Helen Innes et al., which discusses the dating and the worldwide climate impacts.

Much of the 1200 km³ of tephra ended up in the ocean. This is not surprising for an eruption in the narrow Central American isthmus. Such sediment is not easy to date. However, during the ice ages there were changes in the composition of the ocean water. In colder periods, the fraction of the 18O isotope increases in water, because the heavier isotope (compared to the normal isotope of 16O) requires more energy to evaporate. More of the lighter 16O goes into the atmosphere. There is quite a lot of water in oceans, so this has only a small effect the composition of sea water. But the water in the atmosphere is strongly affected, and this shows up in ice record. The snow in such periods is low in 18O. The 18O fraction in the ice cores is well determined, and we know how and when it changed during the ice age. And although the effect on ocean water is smaller, it is not zero and it shows up in sediment on the sea floor. Even though rain retuns to the ocean, during an ice age this return is delayed because so much of the rain and snow becomes locked up in the glaciers. So the sea water changes as the glaciers wax and wane. The sediments contain the remnants of the plankton, and pores contain the ancient sea water in which the plankton lived. The sediments form a layer in the ocean floor, and the 18O of the sea water of the time can be measured around that tephra layer.

Comparing these to the known levels of 18O over time, gives a best fit of 80,500 years ago, with an uncertainty that may reach 2000 years. The dates fall midway between the older and newer dates of the eruption.

The authors combined this with a date based on ice cores. There is ice in both Antarctica and Greenland of this age, and they have been used to find sulphate peaks. There is such a large peak dated to 79,500 years ago, with an uncertainty of 3500 years. Now sulphate in itself does not identify the specific eruption. But if there is tephra in the ice core, then there is a better chance of relating it to a specific volcano. Tephra travels much less far than sulphate: it drops out of the atmosphere too soon. Only the largest eruptions can deposit tephra around the world: finding tephra in ice cores is rare. However, Atitlan was a very large eruption.

The team therefore went through the ice to find tephra fragments. It is hard work. They found 69 fragments. The next task was chemical analysis, to find the exact chemical composition. Four of the fragments were rhyolitic with high silicate, as found in Atitlan, 3 from Greenland and 1 from Antarctica. The analysis is not detailed enough to definitively prove an origin from Atitlan, but it is a plausible candidate. But it is notable that this is only the second eruption with tephra found at both poles, after the 1256 Rinjani eruption.

The combination of the large sulphate spike around the time derived from the sediment suggests this is the correct date, and the tephra fragments strengten the conclusion. The team finally dates the Parakeets eruption to 78,700+-3,600 years ago, combining the three available dates. 

That means the two super-eruptions were 5000 years apart, not 2000. Still, it shows how probabilites work. The fact that something big has just happened does not change the probability for the next one. Double dipping does happen in nature.

Sulphate

So how bad was the Parakeets eruption? The local human population was still 60,000 years away so was not in danger. But an eruption this size affects the entire world. Sulphate is particularly worrying, since up in the stratosphere it reflects sunlight, and this cools the Earth. We have seen this with Pinatubo which masked global warming for several years. Different eruptions can produce very different amounts of sulphate: it depends on the magma and local rocks. An amount of 5 Megatons (Mt) of sulphur is sufficient to produce notable effects. In scientific papers it is normally quoted in units of Tg (Teragram), where conveniently 1 Mt is the same as 1 Tg. Pinatubo produced around this amount (5 Tg). Tambora produced some 60 Tg. This is not easy to measure, by the way. With a modern eruption we can use satellites to measure the atmosphere, but Tambora was a bit too early for that. It can be extrapolated from the amount found in the ice cores but this depends on how far away the eruption was and how effective it spread. It can also be scaled from the amount found in the local tephra. But estimates can differ by factor of 2 or more: for Tambora, they vary between 30 and 120 Tg.

For Toba, a well-studied super-eruption, the suphur amount is estimated as 200 Tg. Innes etal estimate the sulphur from Atitlan as 226+-48 Tg. This half the amount estimated from tephra (520 Tg of Sulphur). The difference gives an indication of the real uncertainty.

How does this compare to more recent eruptions? Sigl et al (2022) compiled a list based on ice cores, shown in the figure below. There have been eruptions up to 100 Tg. Samalas (Rinjani) stands out in the past millennnium: even though it may have been a bit smaller eruption than Tambora, it produced mor sulphur. Taupo doesn’t seem to have produced any (which seems a bit surprising). Further back, Crater Lake was a big one. But none were as large as either Atitlan or Toba.

The Toba eruption was twice as as Atitlan. But it produced a similar amount of sulphate in the ice cores. VEI is not everything.

Climate impact

With these estimates, the amount of sulphur produced by Atitlan was about 4 to 8 times larger than that of Tambora, while the tephra volume was about 8 times larger. These numbers appear reasonable. But what would such a massive amount of sulphur do to the climate? After Tambora, people noted that the Sun lacked power and it was possible to look at the Sun (something I would not recommend at any time!). With 4 times as much sulphate, the skies must have looked white and the Sun barely strong enough for decent shadows! The whole world would be like Manchester. (Herodian already note that “The atmosphere in the country is always gloomy.” He would not have appreciated the post-Parakeets skies.)

Climate models indicate that with this much sulphate, global temperatures would drop by 6 C, with recovery taking 20 to 30 years. The lower temperatures should lead to a significant increase in sea ice. Such a brief event is not easy to detect in the most ancient ice records, but may just about be detectable. So was it?

The figure below shows a close-up of the 18O levels in the ice cores. )The precise dates have the previously mentioned uncertainty of a few thousand years: that is why they are sloughy different from the adopted date for the Los Chocoyos eruption mentioned above.) The blue and black lines show the Atitlan layer in the ice core. The Antartic (blue) and Greenland (black) layers don’t quite match up but this reflects the uncertainty in relative dating of the two. There is indeed a dip in 18O fraction in the ice core at the time of the sulphate layer, evidence for global cooling. And it was short-lived: over decades, the temperatures recovered to what they had been before.

How does that compare to Toba? The figure below shows a larger section, with the Chocoyes eruption on the left and Toba on the right, just after 74 ka (‘kilo-annum’, if you wonder). It was at a time of significant cooling which lasted for almost 2000 years. Even within the ice age, temperatures could fluctuate from one century to the next. Climate is not intrinsically stable: we should not rely on it and are fools to play with it. But that is beside the point.

So did Toba change the climate in a way that Atitlan did not? That is often claimed. But it is not so clear from the record. We see what happened but need to deduce the causes. Around the date of the Toba eruption, there is an initial peak at 74ka (only in black, Greenland), followed by a decline. There is a brief deeper dip at 73.6 ka, which is especially notable in the blue (Antarctic) ice core.

Looking harder, the long cooling started arund 74.2 ka. If the deeper dip is due to Toba, then that cooling is not related to the eruption. The decades-long dip could be eruptive but the 2000-year phase is probably not.

In both Toba and Atitlan, there was a cooling, severe enough to show up in the ice cores and in the ocean fauna. So yes, super-eruptions are bad news. If one were to happen now, we would be deep in shit trouble. But they do not have lasting effects. Over decades the climate returns to what is was before, or at least where it would have been without the eruption.

Bt if that is the case, why do flood basalts have much longer-lasting effects on the climate? That is because of how they affect the climate. Super-eruptions produce sulphate, and this drops out of the atmosphere over years. Flood basalt produce copious CO2 and this lasts much longer in the atmosphere. Sulphate cools – CO2 warms. Flood basalt cause hot-house climates which last for tens of thousands of years, and which have been linked to several mass extinctions. Super-eruptions are bad, but flood basalts are devastating.

From past to future

Should we be concerned about the next super-eruptions? That is a ‘yes’: if one were to happen, there would be significant trouble and a lot of people would not survive. There have been at least three super-eruptions in the past 100,000 years. That suggests one per 30,000 years, so over a human life time, a chance of 1 in 300. NASA would call that ‘safe to fly’ or ‘human rated’ (which is based on a chance of loss of spacecraft and crew of 1 in 500). (The space shuttle reached 1 in 50, explaining why we no longer fly it.) So yes, we should be concerned and do research into predicting them, but we do not need to be overly worried. The chance of a VEI-7 in a life time is about 1 in 3, so we should really be prepared for that. I don’t think we are.

How about a flood basalt? Those happen rougly once per 10 million years, and change the climate for a much longer time than super-eruptions do. One day we may have geo-engineering ability on a scale big enough to do something about them. We have time to think and dream, to come up with solutions for the future.

But far more urgent is that other type of flood basalt. We are producing CO2 at a rate even flood basalts did not manage. The climate effects of this are playing out in front of our eyes. We are currently on track for 3.5 C of global warming this century. We are our very own super-eruption – we are engaged in an uncontrolled experiment in geo-engineering; the tendency has moved to denial. Climate change denial is a crime with our children as victims. But denial never lasts: the next major weather crisis, like the burning of Australia a few years ago, will change the political atmosphere again. And there are positives: technology to create the energy transition is progressing fast in spite of the current political climate, driven mainly, perhaps, by the lack of access to reliable energy sources for China and Europe. And used Teslas are becoming cheap as dirt. We can still determine our own future.

We can’t control volcanoes. But we can learn from them.

Albert, March 2025

Further Reading

Ice core evidence for the Los Chocoyos supereruption disputes millennial-scale climate impact

Ten volcanoes with super-eruption potential: Part I”

Ten volcanoes with super-eruption potential: Part II

Power of the past: a compilation of 25 super eruptions and

Power of the past: a compilation of 25 super eruptions – continued.