This site uses cookies. By continuing to use this site you agree to our use of cookies. To find out more, see our Privacy and Cookies policy.
Skip to the content

[IOP] A community website from IOP Publishing

environmentalresearchweb blog

Can volcanoes trigger ice ages?

The idea that a super-enormous volcanic eruption — or hypereruption — would alter the climate dramatically has been around for a long time. It fits the facts about the biggest historical eruptions we know of, and also our understanding both of how volcanoes work and how the atmosphere works. But could the drama extend to tipping the climate from an interglacial state to a full-blown ice age?

The answer, as has long been believed, is still No, according to Alan Robock and colleagues in a paper published last year. They added several new kinds of potential cooling mechanism to two climate models, and were unable to trigger an ice age.

When a volcano goes off, it is always unpleasant for those in the immediate neighbourhood. The climatologist’s concern, however, is with the broader consequences. A violent enough eruption can loft its products into the stratosphere, where they can persist long enough to spread around the world.

The main culprit is sulphur dioxide, SO2. It reacts with water vapour to form a haze of sulphuric acid droplets. The droplets increase the scattering of incoming solar radiation, making the atmosphere more reflective and cooling the Earth slightly. The more SO2, the more cooling.

The snag is that the haze doesn’t last. The atmospheric effects of Pinatubo in 1991, the largest eruption of recent times, were detectable for a few years at most.

Krakatau in 1883 was bigger than Pinatubo. Tambora in 1815 was even bigger, and still stands as the largest eruption in the historical record. If we turn to the geological record, the largest eruption we know of is that of Toba in Sumatra, in about 72,000 BC. Toba yielded a quantity of stratospheric SO2 hundreds of times that of Pinatubo, which was about 20 megatonnes.

Robock and colleagues injected 300 “Pinatubos” of SO2 into the baseline run of their models, but also tried amounts as great as 900 Pinatubos. With a dynamic vegetation module, they explored the feedback on global temperatures of widespread death of vegetation due to the volcanic cooling. The feedback was not very impressive. Precipitation dropped markedly, but cooling reached about 10 degrees at most, and recovery was nearly complete after about a decade. Coupling the climate model to an interactive model of atmospheric chemistry, they found that the SO2 reaction products persist for longer and produce greater total cooling — as much as 18 degrees — but still no permanent, ice-age-like change in the climatic state. The cooling was partly offset by warming influences, such as more water vapour and ozone in the stratosphere, and more methane in the troposphere. All of these are greenhouse gases.

One thing that bothers me about the Robock study, which is a step forward, is that it still may not cover all the bases. For example the model runs may not have been long enough to pick up delayed responses of the ocean to reduced inputs of heat during the cooling episode. And the climate models were unable to follow the behaviour of the other sluggish players in the drama, the glaciers themselves.

On the other hand, look at what actually happened. In an older paper, Zielinski and co-authors found a signal from Toba in an ice core drilled in Greenland: about six years of strongly enhanced deposition of sulphate, followed by a 1,000-year long “stadial”. Stadials, identified by looking at ratios of the isotopes of oxygen, are relatively short cool episodes within ice ages. However Toba was preceded by 2,000 years of more moderate cooling, which suggests that the stadial proper might have happened anyway. What is more, the oxygen isotopes repeat a very similar pattern in the 2-3,000 years after the end of the “Toba stadial”: rapid warming, moderate cooling, rapid cooling, with no evidence for volcanism at all. In fact, these two excursions look rather like Dansgaard-Oeschger events.

So we have a plausible but not compelling link between our only known hypereruption and a limited amount of long-term cooling. If a Toba happened tomorrow, it might presage a short stadial, but not a long one, and anyway stadials ought not to be at the top of your list of things to worry about. But on the purely intellectual side, the effort to understand Toba nevertheless bears on an important question. How hard do we have to hit the climate system before it really gets upset, or, putting it another way, what does “tipping point” mean?


This entry was posted in In from the cold and tagged , . Bookmark the permalink.
View all posts by this author 


  1. Vinny Burgoo

    Still no comments? It’ll have to be me, then.
    Your final question is indeed important. If an abrupt 10 Kelvin shift can’t knock climates beyond a tipping point, what chance a 2 or 3 or 6 Kelvin shift over 100 years?

  2. Graham Cogley

    Dear Vinny – Good question. As far as facing the 21st century is concerned, it is probably relevant that a hypereruption produces a cooling, not a warming. That is, a transition to more abrupt rates of change is likely to require quite different conditions, depending on what direction you are going in.
    I am not sure I like the concept of “tipping point”, though. For example, at present the best assessment is that the West Antarctic Ice Sheet is unlikely to collapse in response to 21st-century forcing. But we still have to face up not only to the impact of 2K or 3K of future warming but to all the impacts of past warming that haven’t hit us yet. So we need at least one more concept besides tipping points, namely the idea of committed change.

  3. Odd that the little ice age isn’t mentioned. The confidence in the models continues to impress me though – despite an inability to generate ice ages, the models must still be correct, and it is the forcing which needs to be adjusted.

Leave a comment

Your e-mail address will not be published.


  • Comments should be relevant to the article and not be used to promote your own work, products or services.
  • Please keep your comments brief (we recommend a maximum of 250 words).
  • We reserve the right to remove excessively long, inappropriate or offensive entries.

Show/hide formatting guidelines

Tag Description Example Output
<a> Hyperlink <a href="">google</a> google
<abbr> Abbreviation <abbr title="World Health Organisation" >WHO</abbr> WHO
<acronym> Acronym <acronym title="as soon as possible">ASAP</acronym> ASAP
<b> Bold <b>Some text</b> Some text
<blockquote> Quoted from another source <blockquote cite="">IOP</blockquote>
<cite> Cite <cite>Diagram 1</cite> Diagram 1
<del> Deleted text From this line<del datetime="2012-12-17"> this text was deleted</del> From this line this text was deleted
<em> Emphasized text In this line<em> this text was emphasised</em> In this line this text was emphasised
<i> Italic <i>Some text</i> Some text
<q> Quotation WWF goal is to build a future <q cite="">
where people live in harmony with nature and animals</q>
WWF goal is to build a future
where people live in harmony with nature and animals
<strike> Strike text <strike>Some text</strike> Some text
<strong> Stronger emphasis of text <strong>Some text</strong> Some text