[Paper] Interventions that May Prevent or Mollify Supervolcanic Eruptions

by Denkenberger15th Jan 20185 comments


Natural risksSupervolcanoExistential risk

Supervolcanic eruptions are approximately 10 times as powerful as the eruption that caused the year without a summer in 1816. A supervolcanic eruption would cause local devastation, but the main problem is blocking the sun for years and starvation (with possible loss of civilization without recovery and other far future effects). Indeed, many people think that the genetic bottleneck of humans going to a population of only a few thousand was due to a super volcanic eruption 74,000 years ago. It is widely assumed that there is nothing we can do to prevent or mollify supervolcanic eruptions. However, I came up with over 50 possible interventions. In the paper, we could not get into economics. I have done some initial estimates that indicate that the most promising interventions of adding soil or water on top of the supervolcano to delay an eruption for 100 years would likely be cost-effective only considering the present generation. However, this is in isolation. In reality, the first thing we should do is get prepared with alternate foods that are not dependent on sunlight. This would protect against the majority of the damage associated with a supervolcanic eruption, making prevention of a supervolcanic eruption less cost-effective. Still, since people are already doing research on supervolcanic eruptions, it may make sense to nudge that research towards directions that would reduce global catastrophic/extistential risk.

Here is the full paper and below is the abstract:

A supervolcanic eruption of 10^15 kg could block the sun for years, causing mass starvation or even extinction of some species, including humans. Despite awareness of this problem for several decades, only five interventions have been proposed. In this paper, we increase the number of total possible interventions by more than an order of magnitude. The 64 total interventions involve changing magma characteristics, venting magma, strengthening the cap (rock above the magma), putting more pressure on the magma, stopping an eruption in progress, containing the erupted material, disrupting the plume, or provoking a less intense eruption. We  provide qualitative evaluations of the feasibility and risk of 38 of the more promising interventions. The two most promising interventions involve putting more pressure on the magma and delaying the eruption with water dams or soil over the magma chamber. We perform a technical analysis, accurate to within an order of magnitude, and find that water dams and soil and could statistically delay the eruption for a century with 1 and 15 years of effort, respectively. All actions require essentially untested geoengineering challenges along with economic, political and general public acceptance. Further work is required to refine the science, provide cost estimates, and compare cost effectiveness with interventions focusing on adapting to a supereruption.

5 comments, sorted by Highlighting new comments since Today at 4:45 PM
New Comment

Thanks for this interesting paper. Having looked into this a bit, my impression is that some of the figures on the risk posed by supervolcanoes are too high.

Estimates of the frequency of VEI=8 eruptions vary from 30,000 years to around 130,000 years ( W. Aspinall et al., “Volcano Hazard and Exposure in GFDRR Priority Countries and Risk Mitigation Measures,” Volcano Risk Study 0100806- 00-1-R, 2011, 15; Susan Loughlin et al., Global Volcanic Hazards and Risk (Cambridge University Press, 2015), 97)

If VEI=8 events are as frequent as suggested in your paper (on the order of 10,000 years), it seems extremely unlikely that they would constitute an ex risk: the homo genus would have had to have gone through this 120 times and survived at much lower levels of technical sophistication than today.

Some of the literature estimates the frequency of VEI=9 events at one every 30 million years, with massive uncertainty. (Aspinall et al., “Volcano Hazard and Exposure in GFDRR Priority Countries and Risk Mitigation Measures,” 15.)

Thanks for the feedback. I cited the most recent study that claims to have identified more eruptions than previous studies: Rougier, J., Sparks, R. S. J., Cashman, K. V., & Brown, S. K. (2018). The global magnitude–frequency relationship for large explosive volcanic eruptions. Earth and Planetary Science Letters, 482, 621–629. However, perhaps I should not update so strongly because you are right that other estimates are closer to the order of 100,000 years. That is good to think about what it means in terms of existential risk historically. Survivorship bias should not change things too much. Our circumstance is significantly different now. On the plus side, we have more population, more food storage and better knowledge of what is happening. But on the minus side, a super volcanic eruption could raise tensions such that nuclear war breaks out, which would be even worse.

This is fantastic! Thank you for writing this. I think that far too often people see a problem then say, "it's not tractable because I can't think of anything you can do about it" before they've even given it 10 minutes thought. And often causes require far more than 10 minutes thought to come up with some good potential solutions!

Nice. But you don't take reliability and breakdown concerns in your calculation regarding dump trucks. Especially if they are being run 24/7 in night and bad weather conditions, allow for a nontrivial fraction of them to be broken down and under repair at any given time. This could be an enormous problem for the first few years of action because repair and maintenance services don't exist in such a high concentration to handle so many heavy vehicles in a remote location.

Also, see how many such vehicles are owned by the US National Guard, Seabees, etc. They may not be listed in the FHA's statistics.

Thanks for your careful read. I suggested 160 hours per week operating, which allows for 8 hours a week for fueling, scheduled maintenance, and unscheduled maintenance. This is probably optimistic, but I think my estimates are low by an order of magnitude compared to what we could do with non-dumptrucks (though these would likely require retrofitting).