This is a writeup of a shallow investigation, a brief look at an area that we use to decide how to prioritize further research.

In a nutshell

  • What is the problem? Large volcanic eruptions, though rare, could have an extremely negative humanitarian impact.
  • What are possible interventions? Based on current knowledge and technology, large eruptions cannot be prevented, but further research may improve the prediction of eruptions and allow for some mitigation of humanitarian impacts.
  • Who else is working on it? Although scientific organizations fund some basic research on volcanic eruptions, we are not aware of any organizations explicitly working to reduce the humanitarian risk from large eruptions.


Published: June 2013

Why did we look into this area?

  • Large eruptions could have extremely negative impacts on human welfare, mainly through adverse effects on the global food supply. The potential catastrophe is so great that an investment in research could conceptually have high returns.
  • Large eruption risk appears to be more quantifiable than some other global catastrophic risks (e.g., nuclear war).

What is the problem?

In addition to local destruction, large volcanic eruptions could have a significant negative impact on the global food supply.1

Volcanic eruptions are classified based on the volcanic explosivity index (VEI), which measures eruption magnitude on a logarithmic scale.2 Our understanding of the current scientific understanding of the relationship between VEI, frequency of eruption, and likely humanitarian impact is:

  • An eruption of VEI ≥ 7 is likely to occur every few hundred years.3 Such an eruption may have global effects, including cooling the earth’s atmosphere, harming food security, or interfering with plane travel.4 The most recent VEI ≥ 7 eruption was of Mount Tambora in Indonesia in 1815, which is believed to have resulted in crop failures as far away as New England during the following summer.5
  • Previous estimates have suggested that an eruption of VEI ≥ 8 is likely to occur roughly every 30,000 years,6 though more recent (unpublished) research suggests a higher frequency, in the range of an eruption every 10,000-15,000 years.7 Such an eruption is likely to eject ash that could cover millions of square kilometers with deep ash, potentially causing major crop failures.8 A major disruption of this kind could have very destructive second-order effects (such as increasing the likelihood of conflict).
  • Estimates about the frequency of eruptions of VEI ≥ 9 vary by several orders of magnitude: we have seen frequency estimates ranging from roughly every 30,000 years to roughly every 30 million years.9 This disagreement over frequency may stem from an underlying disagreement about the classification of different eruptions,10 but also may be a result of the most recent data being as yet unpublished.11 While a VEI ≥ 9 eruption would be much worse, from a humanitarian perspective, than a VEI ≥ 8 eruption, there seems to be agreement that such an eruption is unlikely to lead to human extinction (especially to the extent that the true frequency of VEI ≥ 9 eruptions is at the higher end of the reported spectrum, which would imply that humanity has survived such eruptions in the past). Incredibly extreme eruptions could conceivably cause human extinction, but such eruptions are both extremely rare and difficult or impossible to prepare for.12

The frequencies cited above are associated with significant uncertainty,13 and we have not vetted them thoroughly.

Although some volcanoes appear to have a fairly stable average frequency of eruptions, there is nonetheless a large amount of variation in the timing of individual eruptions (i.e. information about the average period of a volcano and its last eruption does not necessarily imply much about when its next eruption might be).14

What are possible interventions?

Our understanding is that large eruptions cannot currently be prevented, and that there are not technologies on the near horizon that are likely to change that.15

Accordingly, opportunities to address the humanitarian threat of large volcanoes are likely to focus on research and preparation.

Stephen Sparks, a prominent volcanologist who we spoke with, said the following about potential interventions:16

Possible interventions to address very large volcanic eruptions are:
1. Better modeling of what would happen in the event of a very large eruption, so as to better prepare to for the contingency
2. Improving eruption prediction capacity, for example, by identifying new volcanoes with high potential for large explosive eruptions.
3. Monitoring and characterizing the history and interior structure of these volcanoes.

It is currently possible to use satellites to identify volcanoes showing unrest, which may be sign of imminent eruption. However, it’s not currently possible to be sure that these volcanoes will erupt or how large the magnitude of the eruption will be. It’s difficult to develop methods for determining the magnitude of a potential very large eruption, because very large eruptions happen sufficiently rarely so that data are lacking, and because techniques for imaging the inside of volcanoes are not good enough yet.

If it were possible to determine the amount of magma inside of a volcano near eruption, one could determine those that have the potential for very large eruption. So working on improving imaging techniques could be valuable for predicting very large eruptions.

We also spoke with Sean Brocklebank, an economist and GiveWell supporter. He said:17

Large volcanic eruptions can’t be stopped with current or easily foreseeable technology. Perhaps the best strategy for damage mitigation is to first get better at detecting pending eruptions, and then if we get a few years’ warning on a massive eruption, turn attention toward how to prepare.

Monitoring

While we can’t currently predict medium-size eruptions (of the type that happen every few years), there are reasons to expect that we might be able to detect VEI 8 eruptions years in advance, due to the types of deformations that seem to have occurred prior to such events in the past. See the Nature article, “Decadal to monthly timescales of magma transfer and reservoir growth at a caldera volcano” by Druitt et al, 2012.

Some monitoring is already happening, in many cases as a side effect of earthquake studies. But the data is incomplete and not linked together across the world…

Preparing

Since the biggest damage is to food supply, preparing would likely center around building up a large food supply (perhaps partially by reducing the portion of crops going to livestock), developing a plan for getting the food to the ash-covered areas that can’t produce any food for a year or more, and planning for temporary alternate patterns of cultivation around the world to deal with global temperatures which may be 5 to 15ºC lower for a period of several years.

We do not have a strong sense of how much the relevant research or preparation might cost or what the most likely returns would be. Stephen Sparks estimated that the Global Volcano Model, a project to collect better data about the historical occurrence of volcanic eruptions, likely cost around a million dollars.18

Who else is working on this?

We do not have a clear sense of how much funding is spent on research on large volcanoes, but we understand it to be a relatively small amount, mostly coming from basic research funders like the National Science Foundation.19

We are not aware of any major organizations that are explicitly attempting to work to address the humanitarian risk of large volcanic eruptions, though we have come across two groups of scholars that are working to address this issue.20

Questions for further investigation

Our research in this area has been relatively limited, and many important questions remain unanswered by our investigation.

Among other topics, our further research on this cause might address:

  • How do volcanoes compare in likelihood, severity, and preventability to other catastrophic risks like near-earth objects? Our current understanding is that for a given level of humanitarian harm, volcanoes are a much larger risk than asteroids.21
  • How likely is future research to pay off in information leading to better mitigation?
  • How much is currently spent on research on large volcanic eruptions?

Sources



Source name used in footnotes Link Archived link (for external files)
Aspinall et al. 2011 Source Archive
Brocklebank conversation Source -
Sparks conversation Source -
Self 2006 Source Archive
Global Volcano Model Network Source Archive
Geohazards Community of Practice Source Archive
  • 1.
    • “An eruption of magnitude 7 would emit ash and acid aerosols high into the atmosphere and cool the earth’s atmosphere for several years.
      • This could have negative impacts on the environment, and for example, affect food security. In 1815 there was a magnitude 7 eruption on Mount Tambora in Indonesia, which resulted in crop failure as far away as New England due to frosts in the summer of 1816. This caused a wheat price shock, and resulted in New England farmers migrating to the Midwest and beyond. Since the world today is more interconnected and closer to carrying capacity, a disruption in the food supply could damage human society even more than in the 1800’s.
      • The ash will interfere with airplane travel.
      • There are many other possible environmental effects, and there’s a lot of research to be done on what they would be.”
        Sparks conversation
    • Self 2006 Table 3, p. 2086: “List of phenomena and hazards associated with a hypothetical large explosive eruption”
  • 2.
    • “Volcanic eruptions vary in magnitude, and their magnitudes are represented on a logarithmic scale. The eruptions that have potential to have a globally significant impact are those of magnitude 7 or greater.” Sparks conversation
    • “Magnitude of a volcanic eruption is defined as the Log of mass of magma erupted minus 7. This definition means that magnitude and VEI (Volcanic Explosivity Index, see Figure 3.1) are equivalent to first order.” Aspinall et al. 2011
    • We have treated magnitude and VEI as equivalent below, but the potential for divergence between the two metrics may be a further source of uncertainty.
  • 3. “Magnitude 7 eruptions occur every few hundred years.” Sparks conversation. See also Aspinall et al. 2011 Figure 3.2, pg 15, reporting a return period of VEI > 7 of 490 years.
  • 4.

    “An eruption of magnitude 7 would emit ash and acid aerosols high into the atmosphere and cool the earth’s atmosphere for several years. This could have negative impacts on the environment, and for example, affect food security.

    In 1815 there was a magnitude 7 eruption on Mount Tambora in Indonesia, which resulted in crop failure as far away as New England due to frosts in the summer of 1816. This caused a wheat price shock, and resulted in New England farmers migrating to the Midwest and beyond. Since the world today is more interconnected and closer to carrying capacity, a disruption in the food supply could damage human society even more than in the 1800’s.

    The ash will interfere with airplane travel. There are many other possible environmental effects, and there’s a lot of research to be done on what they would be.” Sparks conversation.

  • 5.

    “In 1815 there was a magnitude 7 eruption on Mount Tambora in Indonesia, which resulted in crop failure as far away as New England due to frosts in the summer of 1816. This caused a wheat price shock, and resulted in New England farmers migrating to the Midwest and beyond. Since the world today is more interconnected and closer to carrying capacity, a disruption in the food supply could damage human society even more than in the 1800’s.” Sparks conversation.

  • 6.
    Aspinall et al. 2011 Figure 3.2, pg 15, reporting a return period of VEI > 8 of 30,000 years.
  • 7.
    “Magnitude 8 eruptions occur every 10k-15k years.
    Magnitude 9 eruptions occur every 30k-40k years.
    Current frequency estimates are based on new data from a 6-year long international effort to compile data, and they are higher than previous estimates. The updated estimates have not yet been published, because the data are so new.” Sparks conversation.
  • 8.

    ” With a VEI 8 volcano, the effects could be globally harmful but human extinction would be unlikely. Ash clouds could cover millions of square kilometers with ash that’s half a meter or a meter in depth, wiping out at least a year’s worth of crops (not including the sunlight blocking effects that would likely reduce crop yields worldwide). The food loss would be the largest potential threat.” Brocklebank conversation.

  • 9.
  • 10. For instance, Aspinall et al. 2011 cite the Toba eruption, approximately 75,000 years ago, as VEI 8 (pg 14, Figure 3.1), while Sparks refers to it as VEI 9: ” The most recent magnitude 9 eruption occurred about 75k years ago at Lake Toba in Sumatra.” Sparks conversation.
  • 11. “Magnitude 9 eruptions occur every 30k-40k years. Current frequency estimates are based on new data from a 6-year long international effort to compile data, and they are higher than previous estimates. The updated estimates have not yet been published, because the data are so new.”
    Sparks conversation.
  • 12.
    • “Eruptions of magnitude 9 are unlikely to cause human extinction. Humans and their ancestors have survived for millions of years despite the occurrence such eruptions…. Because magnitude 9 eruptions are unlikely to cause extinction, and because they’re rarer than magnitude 7 eruptions, which already have substantial disruptive potential, the focus of the study of very large eruptions should be on eruptions of magnitude ~7 rather than magnitude ~9 eruptions.” Sparks conversation.
    • “For a massive flood basalt eruption such as the Deccan Traps (about 65 million years ago, mya) and Siberian Traps (250mya), the damage could be far more extensive. Those eruptions may have led to the extinction of a large portion of all species living at the time. A similar flood basalt eruption today would destroy human civilization and could possibly cause human extinction. Since the events are so rare in earth’s history (on the order of once every billion years), and since a flood basalt eruption would almost be impossible to prepare for, it doesn’t seem very helpful to worry about this type of eruption.” Brocklebank conversation.
  • 13.

    ” The higher the magnitude, the more rare the eruptions are:
    • Magnitude 7 eruptions occur every few hundred years.
    • Magnitude 8 eruptions occur every 10k-15k years
    • Magnitude 9 eruptions occur every 30k-40k years
    Current frequency estimates are based on new data from a 6-year long international effort to compile data, and they are higher than previous estimates. The updated estimates have not yet been published, because the data are so new. The above estimates are probably in the right ballpark, but there’s more work that can and should be done to generate more robust estimates. The uncertainty is the largest for the eruptions of greatest magnitude.” Sparks conversation.

  • 14.
    “While a given volcano may be characterized by stable average eruption frequency, the time between its successive eruptions varies considerably.” Sparks conversation.
  • 15.
    • “The energy released by volcanoes is far too large to prevent them by geoengineering.” Sparks conversation.
    • “Large volcanic eruptions can’t be stopped with current or easily foreseeable technology.” Brocklebank conversation.
  • 16.

    Sparks conversation.

  • 17.
    Brocklebank conversation
  • 18.

    ” Stephen Sparks was involved in an international collaboration with the Smithsonian Institute and colleagues around the world to compile data on past volcanic eruptions. The collaborators have compiled a database on volcanic eruptions from the last 2 million years, which is available at (www.globalvolcanomodel.org/).…

    The project that Stephen Sparks was involved in to create a database of volcanic eruptions from the last 2 million years cost around a million dollars.” Sparks conversation.

  • 19.
    ” There is relatively little funding available for the study of large volcanic eruptions.
    • There is some funding from the National Science Foundation (NSF) and the European Union (EU) for individual academic researchers.
    • The U.S. Geological Survey funds the Yellowstone Volcano Observatory.
    • The Smithsonian Institute is the one organization that is funded for the public good to collect data on volcanic eruptions. Currently the Global Volcano Model network of institutions is collecting data on very large volcanic eruptions. This work principally involves Bristol University (UK), the British Geological Survey, the Geological Survey of Japan and the Smithsonian Institution.
    • The insurance industry is interested in large volcanic eruptions, but has not been funding research in this area, instead it currently prioritizes research on tsunamis as well as floods, earthquakes and tropical cyclones.”

    Sparks conversation.

  • 20.
    • “The GVM project will develop an integrated global database system on volcanic hazards, vulnerability and exposure, make this globally accessible and crucially involve the international volcanological community and users in a partnership to design, develop, analyse and maintain the database system. The GVM project will aim to establish new international metadata standards that will reduce ambiguity in the use of global volcanic datasets. Vulnerability and exposure data will be integrated into the GVM and again new methods of assessment and analysis will be investigated and tested.

      The project also intends to establish methodologies for analysis of the evidence and data to inform risk assessment, to develop complementary volcanic hazards models, and create relevant hazards and risk assessment tools.

      The research will provide the scientific basis for mitigation strategies, responses to ash in the atmosphere for the aviation industry, land-use planning, evacuation plans and management of volcanic emergencies.”
      Global Volcano Model Network

    • “The GHCP brings together groups and individuals involved in various aspects of geohazards, including research, monitoring and risk assessments, mitigation, and adaptation.”
      Geohazards Community of Practice
  • 21.
    • “Very large magnitude volcanic eruptions, asteroid strikes, and major natural disasters in metropolitan areas are the only natural disasters that could have a global impact. In terms of natural hazards events that are likely to have global reach very large explosive eruptions are much more frequent than asteroid strikes.” Sparks conversation.
    • “Comparing the chance of a large volcanic eruption with a given energy release to the chance of an asteroid with the same energy release, the volcanic eruption is about 10 times more likely. You can’t as easily prevent an eruption from happening, but the damages could perhaps be mitigated since most of the damage is loss of food, which can be prepared for with advance warning.” Brocklebank conversation.