Mike Cassidy

398 karmaJoined Oct 2021Working (6-15 years)


Associate Professor of Volcano and Earth Science at the University of Birmingham. Research affiliate at the Centre for the Study of Existential Risk, and Earth Sciences, University of Oxford. Interested in global catastrophic risks, EA and broad longtermism.


The Risk of Huge Volcanic Eruptions


Hi Stan, thanks for your response. I understand your main thesis now -seems logical provided those ideal circumstances (high global co-operation and normal trade).

VEI 7 eruptions could lead to up to 2-3 degrees of global cooling for ~5-10 years (but more elevated in the northern hemisphere). See here: 

More likely is two VEI 6 eruptions close together, which may provide longer duration cooling of a similar amount ~2 degrees, like in the mid 6th century (Late antique ice age).

The Loughlin chapter didn't account for the incompletness of the geological record like papers published since have done with statistical methods (e.g. Rougier paper I cite in that post), or with ice cores that are better at preserving eruption signatures compared with the geological record. 

In the report a footnote 2 states:
"In mild agricultural shortfalls such as those that may be triggered by crop blight, VEI-7 volcanic eruption or extreme weather, adaptations like redirecting animal feed, rationing and crop relocation would in theory be sufficient to feed everyone".

How did you come to that conclusion? We're only aware of 1 academic study (Puma et al 2015) about food losses from a VEI 7 eruption and it estimates 1-3 billion people without food per year (I think this is a likely an overestimate, and I'm trying to do research to quantify this), so just trying to figure out what you're basing the above statement on, does this take into consideration food price increases and who would be able to pay for food (even if there is technically enough)?

Also note the use recurrence intervals of super eruptions is an order of magnitude off from Loughlin paper which has since been changed (see discussion here: ). Also note, VEI 7 eruptions can sometimes have the same/if not greater climatic impact as super-eruptions, as the magnitude scale is based on the quantity of ash erupted whilst the clmatic impact is based on the amount of sulfur emitted (which can be comparable for VEI 7 and 8 eruptions). I mention these as whilst nuclear war probabilities have huge uncertainties, our recurrence intervals from ice cores of large eruptions are now well constrained, so it might help with the calcs. 

Hi Christopher and Ewelina, thanks for this post, it's nice to see geoscientists in EA and advocating for the skillset that Geoscientists pose and you're right that it already plays a role in many of the EA causes. I may use this these examples to share with my undergrads!

I like the estimation calculation you do to equate the money proportional to planetary defense weighted to its odds. However you seem to use the Ord value of 1 in 8000 per century of a Supereruption. But Toby is talking here of a super eruption the size of Toba eruption which is 5000 km3 in volume. To be registered as a super eruption it's 1000km3 and the estimated probability of this supereruption volume is 1 in 170 per century (Rougier et al 2017). So it would be interesting to re-hash your model with this value instead- I would be interested to see how it compares with our expected value estimation in our paper here: "The financial losses resulting from a large-magnitude eruption are estimated to be in the multi-trillions8, roughly comparable to those of the pandemic. Given the estimated recurrence rate for a magnitude-7 event, this equates to more than US$1 billion per year." 

Another point to make here is the volume of the eruption is not necesarrily proportion to its impact, for the global climate-impacts sulfur content is a better metric, and sometimes there is an asymmetry between eruption volume (or VEI) and climatic impact (Schmidt and Black, 2022). This is where the ice-core record is a good proxy.

I should also add that we have been trying to attain funding from EA sources to conduct research on impacts from large eruptions on society for a few years (e.g. quantifying lives lost and mitigation strategies), but despite the neglectedness, tractability and importance of this risk, the funders have not been interested to date.

To end on a better note, I'd be happy to join your slack group- thanks for creating it.   

P.S. We have a paper on the Ethics of Volcano Geoengineering coming out next month

Just to follow up this post, we want to confirm  that (like many others)  we did not recieve the funds from the FTX foundation for our project before the FTX collapse. 

We want to thank the Clearer Thinking team for all their hard work on this competition (a great example of democratized funding) and for seeing the potential for quantifying the global risk from large magnitude volcanic eruptions. 

If anyone is able to help us find funding to ensure this project goes ahead (or put in touch with people who could help), it would be really apreciated. Thanks all, Mike and Lara

Really nice post- Thanks for writing this! Lot of thoughts- but just a few below for now:

From my experience in Natural hazard research in various countries (Mexico, Indonesia, Carbbean, S America) I can certainly confirm this statement:

One of the major issues with natural disaster prevention is that many believe that the necessary work is already in place, which I don’t think is the case in many countries.

The impacts are also strongly weighted by population exposure and vulnerability, especially for Earthquakes. For example, the magnitude 7 earthquake in Haiti (a strong but not unusual earthquake -occuring ~once a month somewhere in the world), killed 200,000-300,000 people. So targeted preparedness & research in some regions could be significantly more effective. 

Even if there is some organisation/preparedness in some areas/for some hazards, the low frequency, high impact natural disasters often catch orgs/charities out (e.g. we see this with recent Tsunami disasters in Japan and Indonesia), so again action in this area, might be most neglected and therefore effective. 

Thanks for writing this post and your efforts to address these issues! As someone who works in  scientific research I have frustrations about the framework/criteria by which science is funded- so I'm really glad someone connected to effective altruism is looking into this, and I agree that this a really important cause area!

Often the criteria used to judge proposals is based on things like how innovative, novel, timely the science is, if it uses cutting-edge methodology and how suited the candidate is to that study area. Also, as a reviewer of proposals I am asked to judge the proposal's 'excellence'- a quite ambiguous quality- that accounts for significant randomness and bias too.

I think this approach to judge proposals is a limited as it means that a) potentially more important or impactful work will not be funded as it may lack novelty. b) people coming from a problem from a different discipline with differing expertise-yet many scientific breakthroughs come from those from other fields. c) it can lead funding of niche fields of science, to the detriment of neglected areas of greater importance and wider scale tractability. 

Incorporation of the ITN framework would be beneficial, it may help fund more controversial or higher risk science and making decisions based on lotteries, after an initial sift, may also open up the field to more creativity and diversity, and counter some of the other biases that sadly occur. So I really agree with you here.

These issues above relate to the more open calls for proposals and how they are judged, but there are also many schemes one can apply to that are quite narrowly defined and have been decided by relatively few senior academics on a science board. It is not clear to many in the community how these grant calls are are decided, nor how rigorous and unbiased these are, making this process more transparent would be beneficial too. Your ideas based around cost-effectiveness and expected value might also provide more rigor to these decisions.

Anyway- really interested to see what you do with this, and let me know if I can be of help into the UK research council (UKRI) system. I'm currently part of the peer-review college for NERC and can find out more about specific protocols/decision making if useful.

Thanks Ken! Glad you appreciate the importance of this topic.

I'm afraid I've not come across much research about the potential correlation between grand solar minimums and volcanic activity, but let me know where you've come across this and I can look into it.

Thanks for this Toby. I like your suggestion about factoring the risk in this way, and we'll keep you informed about where this all leads. Regarding civilisation collapse & recovery, there's certainly a lot of parallels to abrupt cooling from nuclear and asteroid winters, though the nearer-field hazards (and resulting cascading impacts), may be significantly different. One major uncertainty in this seems to be the location of a super-eruption, which will strongly dictate its effects on society, e.g. similar magnitude super-eruption occurring in the Mediterranean versus New Zealand. So one of the things we hope to look into is identifying the regions & volcanoes where the next super eruptions are most likely to occur. Your book & the longtermism concepts have certainly made me reassess what the most important questions in our field are!

Thanks Jackson for forwarding on these interesting posts and info on this topic, I'm glad others have interest in this issue. I wonder if the topic is slightly underrated because its effects are often manifested as 'long-lived' insidious chronic issues, meaning that judging the cost-effectiveness of different interventions with relatively short randomly controlled trials is much more difficult compared to for instance infectious disease interventions. Perhaps that's where your short-term cognitive effects might be a really useful diagnostic/measure of effectiveness of different interventions- so thanks for drawing my attention to that. 

An effectiveness comparison between vaccination and air filtration/UV light methods as GCBR mitigation is a really good suggestion for future research. I agree that air filtration may prove to be pretty effective, especially  given the time lag to develop & test vaccines too. 

During the writing of this post I spoke to some engineers developing air purifiers for private sector, so if anyone has further interest in this topic in the future - I can put you in touch. 

Thanks Ramiro, I hadn't listened to that 80kh interview-so I'll do that! But yes it could fall into both 'Global health and well being', but also longtermism categories (bio-risk mitigation-wise). 

Also, sorry if this is stupid, but it seems that, unlike CO2, risks from many pollutants (like particulate matters and pathogens) could be significantly mitigated by effective dispersion; so even a normal ventilator could have observable effect on indoor air quality, right? Thus, I wonder if there are / could be any relevant policy recommendations along this line - like for urban design, e.g., "locate potential air pollution emissions by the sea, or spread through areas where they can be dispersed by winds". Does it make any sense?

Yes you're right. I guess there's a difference between ventilation (dispersing the air around, either 'natural' - e.g. windows, or 'mechanical' e.g fans would also do the job) and purification (sucking the air through a filter to scrub out the pollutants/pathogens). I think both go some way to help this problem and just alone ventilation alone is more cost effective (table 1), however ventilation alone may not work in some instances. For instance, if the outside air is poor-quality (dense urban areas/near intense farming), or where there access to outside air is limited (high rise flats/offices where the windows don't open), or for poorly-insulated buildings which are difficult to heat, so fresh air may be limited, it's also hard to judge how much air needs to be replaced, which is where the CO2 and other monitors come in use. 

Regarding the urban design idea, I'm not quite sure how effective this could be, a lot of the sources of air pollution are lots of individual sources, things like vehicle emissions, home heating, agriculture, which would make it hard to disperse by design, but it maybe more feasible with industry. But perhaps policy suggestions that control how much PM2.5/NO2 is emitted by industry might be warranted, which they could control by adding scrubbers/ filters to the outgoing air, if successful that might be applied to the smaller individual scale too.  Thanks for these thoughts!

Load more