On Geoengineering

Originally published at Oliver Morton's blog.

I recently had the great pleasure of attending this year's Breakthrough Dialogue at Cavallo Point, an event at which the Breakthrough Institute brought together kindred spirits of disparate views to hash out some of the many issues that that Institute takes an interest in. On the basis of this Economist special report I was invited to talk about nuclear power, but in the many fruitful interstices of the meeting found myself talking about geoengineering quite a lot, because this is the sort of crowd where that sort of discussion makes sense, and because I am working on a book on the subject.

Towards the end of the meeting, a friend mentioned to me that perhaps I should be more careful in such conversations - people seemed to be getting the wrong idea about what I believed. This may be the case - I can't really vouch for what message people were picking up, and I'll admit that I sometimes run off at the mouth and that jet lag when drink has been taken doesn't always help matters.

That said, I think there is a danger to being too careful in talking about geoengineering. If all the people who know about geoengineering are meticulous in the care that they take in talking about it, they will create no new misapprehensions - but they may do little to dispel old misapprehensions, and they may pass up the opportunity to carve out for geoengineering a more central place in our ongoing discussion on climate. I think it deserves that place; if I didn't I wouldn't be writing a book about it.

But while there may be good reason to be expansive in one's talk, there's no good reason for being careless, or even sloppy, in one's reasoning. I have tried to be pretty careful in published stuff in the past, such as this 2007 piece in Nature and this 2010 piece in Prospect. Some time in the future I hope to provide all the clarity and nuance one could wish for in the book. But for the time being, here are a few key points in my current thinking, expressed with what I hope is appropriate care.

Geoengineering and the two degree limit. I am not a great fan of the two-degree limit; among other things I think it's arbitrary, that it sends an unnecessarily negative message about adaptation, and that it is unrealistic. To get a 50:50 chance of not going over the two degree limit through mitigation alone requires a level of decarbonisation far beyond what is currently on the table, let alone what has been achieved historically. That said, I realise that the two-degree limit is a pretty dominant framing of the problem. So I think people using this framing, or drawn into debates that use this framing, should feel compelled by intellectual honesty to talk about geoengineering in this context.

There are many unknowns when it comes to geoengineering that uses stratospheric aerosols to cut down the level of incoming sunlight a little. The implications of the predictable effects on the hydrological cycle are unknown. So is the degree of ozone depletion entailed. The direct effects on human health are unquantified. No-one knows how to set up a form of governance for such efforts that would be both enduring and equitable. On top of the known unknowns there doubtless unknown unknowns. But there is also what seems to be, to the world at large, an "unknown known": on the basis of current research one can say with high confidence that a stratospheric geoengineering programme scaled to the task would be able to keep the current increase in global temperature below two degrees over quite a long period and across quite a wide range of emissions scenarios.

I think there is a strong argument for going further and saying that this is currently the only plausible intervention that would deliver the much discussed 50% chance of staying under two degrees. I know of very few people who know a lot about energy systems and believe that those currently deployed will be replaced quickly and thoroughly enough to achieve those odds through decarbonisation alone. On occasion I might be quite forcible in saying that those few people are wrong. But that, I accept, is to some extent a matter of opinion about what is technically possible and politically and economically desirable.

This, though, is not a matter of opinion: To say that a two-degree limit can only be achieved by radical decarbonisation, for example by deploying huge fleets of new types of nuclear reactor, is simply not true unless you rule out geoengineering ex ante. If you are doing that, then you should say you are doing that, and explain why. If you are not doing that, and you are still sticking with the two-degree framing, then you should explain what limits you would impose on the contribution geoengineering can make. If you persist in giving the two-degree limit a big role in your analysis but make no mention of geoengineering you are either ignorant or deceitful.

Or we are simply short of space, I can imagine journalistic colleagues replying. Adding geoengineering to the discussion simply to then rule it out just makes things longer, and more complicated for the public, god bless 'em. I appreciate these problems. I have faced them myself. And I have decided that I am going to fight them harder from now on. Leaving geoengineering out of the story simply because it isn't easy to fit it in does a disservice to the truth. It would be justifiable if one thought geoengineering so outlandish as to be truly marginal. If I thought that I would not be continuing to try and write a book about it. If you think geoengineering needs to stay marginalised you should make the case for that marginalisation, rather than just acquiescing in it.

Supporting geoengineering research v supporting geoengineering deployment v supporting geoengineering design. Almost everyone writing about geoengineering supports further research on the subject. This is not that surprising, for three reasons: 1) it is clearly a rational position to take on a possible partial solution to a global environmental problem; 2) most people writing about geoengineering have done research on the subject, and thus have a revealed interest in such research 3) being against research is a hard position to take in the culture we inhabit. Almost no-one writing about geoengineering supports immediate deployment. This again is not surprising. Though the risks of geoengineering are, I think, overplayed, they are real and require more study, and the political issues have not begun to be discussed realistically.

(An aside: how does this reticence square with what I was just saying about having to talk about geoengineering seriously if one talked about the two-degree limit? In two ways. One is that one does not need to start geoengineering now in order to keep to the two degree limit. It would still be possible a few decades down the line. The second is that I am not myself wedded to the two-degree limit as a hard boundary on what people do or don't do.)

What very few people take a position on is geoengineering design. Geoengineering is mostly studied by natural scientists, and their studies tend to naturalise it: to treat it as a thing in the world to be examined. For practical purposes, the majority of the technical literature about stratospheric geoengineering treats it as a loading of stratospheric aerosols (or simply a lowering of the solar constant) sufficient to counteract the warming effect of a doubling of atmospheric carbon dioxide, or something pretty similar. Starting from such a definition it is then possible to explore other aspects of geoengineering as if it were a thing in the world - what does this thing do to the hydrological cycle, what are its regional effects, what happens when it stops, and so on. There has been a lot of interesting modelling research done in this way.

But though this approach produces useful knowledge, it also induces an unwarranted specificity. It equates geoengineering as a whole too closely with a specific test case. This has a specific drawback and a general drawback. The specific drawback is that a geoengineering intervention that entirely cancels out the warming effect of a doubling of carbon dioxide is a very large geoengineering intervention - far larger than an intervention designed, say, to shave warming down from three degrees to two. That latter sort of intervention seems much more plausible than the former.

The more general drawback is that geoengineering is not a natural phenomenon with costs and benefits to be measured as well as possible. It is, as the name makes clear. something that needs to be designed. Geoengineering research should at least in part be aimed at designing instrumentalities and strategies that would optimise the benefits and risks, rather than just assessing the benefits and risks of one possible instantiation. Here's a plausible mission statement for such a programme that I sketched out a while back for my own purposes:

To explore the development of a well characterised, reversible technology, the use of which promises to curb a profound harm caused by greenhouse warming while not risking comparable harm of some other sort, and -within the context of a continuing transition to a carbon-neutral economy -- to work towards deploying such a technology in a safe, timely, transparent and equitable way, if such a course is possible.

I think few of the natural scientists, or for that matter social scientists, currently working on geoengineering would be willing to sign up to such a commitment. They are happy characterising a particular version of geoengineering, but would not be happy trying to design the best version possible and of trying to find a way round potential show-stoppers. That would feel too political, and beyond their remit. I hope over time to convince at least some that this reticence is worth overcoming.

(A note on urgency. I include the transition to a carbon-neutral economy in that mission statement in all sincerity. But I think it is disingenuous not to accept that if feasible geoengineering interventions were to be designed, the urgency driving that transition would diminish. In the real world efforts towards mitigation would likely slacken if plausible and acceptable geoengineering were on offer. Avoiding over-reliance on geoengineering would thus be a real problem. That said, whatever urgency is currently felt is clearly insufficient to the task of radical decarbonisation, and if the geophysical situation were to change in such a way as to increase the feeling of urgency that would in itself present problems if geoengineering is not possible, since mitigation has an effect on climate only over decades.)

Preferring geoengineering to nuclear power. This may have been the issue on which I expressed myself most incautiously at the Breakthrough Dialogue. I am not deeply opposed to nuclear fission as a source of power; I think the radiological risk is overblown, that the waste can be dealt with, and that proliferation risks can be moderated, though not expunged. I do however have a fairly well developed sense of the drawbacks that apply to nuclear power as it is practised in the real world, rather than as it exists in the imaginations and on the drawing boards of enthusiasts, and I think that overcoming those drawbacks, if it proves possible, will be hard and fairly slow. On the basis of this I think nuclear power is unlikely to play a large role in a quick and deep decarbonisation, as some of its advocates would wish. And I think a massive commitment to trying to get such a benefit from nuclear power would run risks in terms of massive costs and the transfer of a great deal of power to nuclear establishments many of which have not historically proved themselves transparent, trustworthy or easily regulated.

On the basis of my perception of the risks and benefits of a massive nuclear-fission build-out (two to three additional terawatts in a similar number of decades is the sort of scale I took some people to be talking about) and my perception of the risks of a modest (<0.5W/m^2) stratospheric geoengineering programme, I think I would prefer a world with the latter to the former. But I accept that, by comparing a reality-informed view of what a massive nuclear power programme might entail with a less constrained view of the possibilities of geoengineering I may be erring for the birds in the bush over the bird in hand. It is certainly true that one can imagine instantiations of geoengineering that are as politically unpalatable as giving a lot more heft to the nuclear establishments of the world. I think people could do better. Other opinions are available.

Being serious. This is really the underlying point. At a recent meeting Rob Socolow suggested that we should divide the world into people who do or don't think the risks of climate change are an urgent matter and people who do or don't think decarbonisation is difficult (pdf). A lot of the green movement is in the do/don't quadrant - do take climate change seriously, don't think getting rid of fossil fuels is all that difficult ("just needs political will," dontcha know). People opposed to current or intensified action on climate (sceptics, lukewarmers, status-quo-ers, whatever) are in the don't/do quadrant - don't see climate change as a serious risk, do think decarbonisation is difficult, or at least costly.

Like Rob, I am in the do/do quadrant. I do think climate change poses serious risks, and I do think decarbonisation is difficult. That is why I think it is worth taking the possibility of geoengineering seriously enough to see how well it might be done.

An award-winning science journalist and author of Eating the Sun: How Plants Power the Planet, Oliver Morton is briefings editor at The Economist.