Super-eruptions - Not quite so super?

Grand Prismatic Spring, Yellowstone National Park, Wyoming.

Grand Prismatic Spring, Yellowstone National Park, Wyoming.

17 May 2013

It turns out that one of the deadliest hazards the Earth can throw at us may happen more often than we thought. Darren Mark and Ben Ellis report on how their work in Yellowstone could radically change our understanding of these events, with implications not just for those living nearby but also for the global climate.

The largest explosive volcanic events, known as 'super-eruptions', are one of the greatest geological threats to mankind. Globally, millions of people live in regions that could be devastated by the eruption of a super-volcano - for example, Yellowstone in North America, Campi Flegrei in southern Italy, and Toba in Indonesia. These eruptions can produce hundreds or even thousands of cubic kilometres of magma over days or weeks.

Yet their most widespread effects don't come from locally-devastating pyroclastic flows of superheated gas and rock, but from ash clouds that can circle the globe. Sulphur injected into the stratosphere oxidises to form small droplets of sulphuric acid. These stop sunlight reaching the planet's surface, cooling the climate.

For example, the most recent super-eruption of the Quaternary Period - the one we are in at present - was the eruption of the Young Toba Tuff (YTT), which occurred around 75,000 years ago in what is now Indonesia. It has been suggested as one of the most significant events in the course of human evolution, leading to cataclysmic changes in terrestrial ecosystems and nearly wiping our species out. Yet not all scientists agree. To prove or disprove the theory, we need to know the exact order of events around the super-eruption, as well as precisely how - and how quickly - ecosystems responded.

We can test these relationships with high-precision geochronology. The ash ejected during super-eruptions comprises silica-glass shards and mineral crystals from the fragmented magma, as well as pieces of the volcano itself. We can harvest the different mineral crystals that were growing in the magma before the eruption from the volcanic deposits, and date some of them to reveal the age of the eruption.

High-precision dating techniques are now transforming our view of super-eruptions. These rely on accurately measuring the relative amounts of two different forms of the same element - known as isotopes - in a sample of rock. Some isotopes decay into others at a constant rate, so if we know how much of each was there at the start and can measure what is there now, we can learn how long ago the rocks were created.

These methods are getting more precise all the time. This improvement comes from new technological developments in mass spectrometry, the technique we use to measure minerals' isotopic composition; from refinements to the known rates at which different isotopes decay; and from other changes in our approaches to dating of rocks and minerals. This isn't just a matter of adding another decimal place to a number; it lets us dissect the geological record at the highest level of detail, and accurately sequence the Earth's history.

Little and often?

With these new tools at our disposal, we wanted to test our understanding of super-eruptions by studying one of the largest examples of recent geological times - Yellowstone, a volcano synonymous with the term. The Yellowstone Caldera is well known for three huge eruptions, at around 2·1, 1·3 and 0·6 million years ago.

These episodes were punctuated by long periods of relative peace, during which lava flowed out episodically rather than being hurled explosively into the air. The largest and oldest of the three major explosive events was the Huckleberry Ridge Tuff (HRT), which erupted a volume of rock approximately 2,500 times larger than the recent Eyjafjallajökull eruption in Iceland - a relatively small event that nevertheless caused chaos in the skies across the Atlantic and Europe.

Many ancient 'super-eruptions' may actually have been many separate events.

The HRT has three component parts, known as members A, B and C. They contain a superficially similar mixture of minerals, but they have some subtle yet important differences. Initial mapping in the late 1960s discriminated between the three members on the basis of differences in texture, such as the size and proportions of the crystals, proposing that each erupted from a different place. Having reviewed this literature in detail, we were intrigued by this idea. We wondered - was it possible that each member also erupted at a different time?

We started out by analysing the chemical and isotopic composition of hundreds of crystals of sanidine, quartz, augite and fayalitic olivine from the HRT deposits. Data showed that whereas members A and B were similar, member C was chemically different, suggesting it crystallised under different conditions.

These results added fuel to our fire, and we began a campaign to date each member as precisely as possible. We harvested potassium feldspar from each member, and analysed single crystals using a method known as argon-argon dating at the NERC Argon Isotope Facility. This technique relies on the known decay rate of a naturally occurring isotope of potassium; we measure the relative quantities of this isotope and its decay product to calculate exactly how long ago it was erupted.

Our results showed members A and B emerged at the same time, but member C appeared at least 6,000 years later. Member C accounts for around 12 per cent of the HRT's total volume, and although the eruption of Members A and B is still big enough to count as a super-eruption (estimated at around 2200km3 of rock), the volume of Member C alone, an estimated 290km3, is around 300 times larger than all the material ejected by the 1980 eruption of Mount St Helens.

The study raises the possibility that many ancient 'super-eruptions' may actually have been many separate events that happened across timescales that are short in geological terms, although still very long by everyday standards.

If this is right, it is a paradigm-shifting hypothesis. It implies that although each volcanic event was smaller than we have thought until now, super-eruptions may have happened more often. As well as the hazard potential of more frequent super-eruptions, we have little idea what impact several large eruptions occurring over a short period would have on the global climate, yet this is an extremely important question.

Our research is now focusing on the younger Yellowstone super-eruptions, assessing the super-eruption deposits of Toba, and re-examining Campi Flegrei and Mount Vesuvius, infamous for the destruction of Pompeii in 79AD.

We have found multiple layers of volcanic ash that can be correlated to the YTT, but that are separated by varying amounts of sediment in deep ocean cores. This suggests there may have been multiple eruptions of Toba around 75,000 years ago. Pilot data from all study sites show similarities with our results from Yellowstone, suggesting these other super-eruption deposits are also made up of smaller eruptions over time.

As a result, the most important question we have to resolve is 'how long does it take to generate voluminous super-eruption-sized batches of magma?' This may be the primary control on how quickly one super-volcano eruption can follow another.

With the potential possibility that some super-eruptions could be resolved into smaller, discrete events we wonder whether in times to come, super-eruptions will not be quite so super?

Dr Darren Mark is a post-doctoral research fellow and manager of the NERC Argon Isotope Facility, based at the Scottish Universities Environmental Research Centre. Dr Ben Ellis is a post-doctoral researcher at ETH Zurich.