As one of Tom Schelling’s grad students, I still have the privilege from time to time of seeing him in informal settings, and I almost always carry away something of value. With his permission, I offer a report on his latest thoughts about what is called “geo-engineering” as a response to the impending global warming crisis.
The bottom line: he’s fairly confident that solar radiation management (SRM), in the form of the release of sulfates into the stratosphere, would work to substantially offset the increase in global mean temperature that would otherwise result from the likely trajectory of greenhouse-gas releases; that it would work quickly; that the cost would be manageable; and that the side-effects of the sulfates would be tolerable.
But this might not, in his view, be anything like undiluted good news. SRM does nothing to cure the other impacts of increased carbon dioxide concentrations, and in particular the acidification of seawater. (The Coca-Cola Company was not the first to discover that if you dissolve carbon dioxide in water you get carbonic acid, aka seltzer.) So the danger is that the efficacy of SRM to deal with the warming crisis will slow the effort to control carbon dioxide emissions. As Schelling put it, “the danger is that solar radiation management could work too well and too cheaply.”
Schelling has been applying his own orthogonal-to-everyone-else style of analysis to the global warming issue for nearly two decades.
Before the opponents of restricting greenhouse-gas emissions settled on the tactic of sticking their fingers in their ears and their heads in the sand and simply denying that the problem was real, their favored tactics were to plan for mitigation of consequences (e.g., dike-building) or try to reduce the consequences of GHG release either by recapture (carbon sequestration) or reducing the flux of solar energy so as to control warming even in the face of increased greenhouse effects (solar radiation management). These last two options are in the portfolio called “geo-engineering.”
Schelling, undeterred as usual by the risk of being seen in bad company, urged that we take both mitigation and geo-engineering seriously. He was also open to the possibility that improved modeling would reduce the expected impact of GHG build-up both on global mean temperatures and on the side-effects of warming in the the form of floods, droughts, storms, etc. That, plus doubts about the feasibility of negotiating and enforcing any sort of global deal on GHG emissions, led him to push for what he called “no-regrets” policies: all of the pure-win options such as improving energy efficiency that would pay for themselves even if global warming turned out not to be a major threat.
Now that improved modeling has mostly served to increase the expected climate change as the result of any given level of GHG, Schelling is no longer in the “no-regrets” camp, and puts limited trust in mitigation in the face of the changes that now seem likely unless something major is done to reduce the change in global mean temperature. But he’s still skeptical that GHG emissions can actually be reduced by anything like the required amount, so “geo-engineering” is looking more and more relevant.
Carbon sequestration - capturing carbon dioxide from the atmosphere and taking it out of circulation for long periods - may be feasible, but under any technology now in sight it would be terrifically expensive. (It’s not clear that coal remains economically competitive even with current solar-electric generation if you add the cost of sequestration to the cost of production.)
Solar radiation management, by contrast, is reasonably cheap, with annual costs to offset the estimated 2050 level of GHG running in the low-to-middle eleven figures (tens of billions of dollars per year), which is after all pretty much chump change against a world GDP in the tens of trillions. That’s certainly far lower than any reasonable estimate of the costs of reducing emissions instead.
The current front-runner is sulfate release into the stratosphere. (Jet airplanes already do some of this.) The amount of sulfate in the stratosphere required to offset the temperature effects of the estimated 2050 GHG level would be about 2% of the amount of sulfate currently released each year into the lower atmosphere, so the contribution to acid rain would be negligible. The half-life of stratospheric sulfates is estimated at about a year - long enough to make the policy affordable, short enough to make it reversible.
The result would be a reduction of about 1% in the amount of solar energy reaching the surface; not very large compared to natural solar-cycle variations, though possibly perceptible to the naked eye and doubtless important to some species.
What seems crucially important to me is that SRM could be deployed quickly and have quick effects: the model is the volcanic explosions that occasionally cool the planet by a couple of degrees C for a year or two. That means that, if current models are too optimistic and Inconvenient Truth-like effects impend - rapid melting in Antarctica leading to rapid and potentially self-reinforcing rises in sea level, a shutdown of the North Atlantic Convector leading to massive cooling in Western Europe - SRM could be put in place on an emergency basis.
Again, the problem is that managing global temperature, and the side-effects of global temperature changes, is only a part of the issue. Ocean acidification is the most dramatic, but not the only, consequence of increasing the amount of carbon dioxide in the atmosphere. So insofar as the existence of SRM as an option reduces the urgency with which citizens and political leaders address the problem of carbon dioxide buildup, the result could be a cool planet with fizzy oceans and not much marine life.
Naturally - this is Kleiman speaking not, not Schelling - when the “small government” loons get out of global-warming-denialist mode, they will naturally strongly prefer the SRM option to any sort of carbon controls: regulations, taxes, or cap-and-trade. This will be especially true of the propeller-head techno-optimists who think that every problem has an engineering solution.
But I predict that they mostly won’t get their way. When one side of a debate denies the existence of a problem, it has (deserves) a hard time getting its preferred solutions to the purported non-problem taken seriously. That was the fate of much liberal thinking about crime control; once you’ve announced that “crime” is a made-up problem and that crime control is just a cover for racism, why should anyone pay attention to you when you announce that you know how to control it?
As poetically just as that result would be, and as helpful it would be to the cause of controlling carbon dioxide buildup rather than only its most dramatic bad consequence, it might lead to disaster. Because the fact remains that emissions controls are both expensive and slow-acting. If we ever find ourselves confronted with a real warming crisis, SRM may be not our best option but our only option. It seems to me only prudent to do the R&D now that means we have it available when and if we desperately need it.
As to the objection that SRM would have unknown consequences: Sure it would. So would uncontrolled warming. And so, for that matter, would the substantial-to-drastic increases in worldwide energy prices required to stop continued carbon dioxide buildup. The principle of not taking very large risks unnecessarily - the sane version of the precautionary principle - does not mean that you shouldn’t jump out of the way of an oncoming train until you’ve made a careful estimate of how badly you might get hurt when you land.