I promised a few folks (@[email protected] @[email protected]) over on this thread an effort post about stratospheric aerosol injection today. My doctorate is in philosophy, but I did my PhD work on the foundations of climate modeling and complex systems theory, and I ended up in a climate lab for two years as a postdoc, where I worked explicitly on this proposal. Long story short, we should not be doing it. Here’s what it is, how it works, and why I think it’s a bad idea.

There’s a ton to say here, so I’m going to just start with a kind of general overview of the proposal and why it worries me. Folks can ask questions if they want to, and I can go into more technical detail or give references on any part of it that folks are interested in. I’ll try to get to any serious comments eventually, but I do have a job and shit. OK, here we go.

Very large volcanic eruptions release a truly mind-boggling amount of volcanic shit into the atmosphere, including compounds known as sulfate aerosols. Sulfate aerosols are small droplets of sulfur-based chemicals that aren’t quite liquid and aren’t quite gaseous. When they get into the upper atmosphere, they can block significant amounts of incoming solar energy, preventing it from getting down to the Earth’s surface where it would otherwise get stuck thanks to the greenhouse effect. This has a major cooling effect: the last big volcanic eruption, 1991’s Mt. Pinatubo, reduced incoming sunlight by 10% and decreased global temperatures by about 0.5 degrees C for two years.

If we wanted to, we could intentionally release similar aerosols into the atmosphere (probably either with high-altitude balloons or literally shooting them into the stratosphere with specialized guns; they need to get way too high up for planes to disperse them). High atmospheric winds would rapidly spread aerosols released near the equator to cover the globe, and with enough aerosols, we could potentially cool the planet off enough to cancel out a quadrupling of GHG concentrations. If successful, such a plan could prevent sea level rises, save the glaciers, and prevent many other extreme weather events that we anticipate would accompany drastic temperature increases. The total cost is projected to be something like $50 billion per year, which is just absolutely insanely cheap by the standards of effective global environmental policy–it’s something on the order of 1% the cost of totally retooling the global economy away from fossil fuels.

OK, so what’s the catch? Well for one thing, global temperature increases aren’t the only consequences of climate change (that’s why we call it “climate change” instead of “global warming”). CO2 in the atmosphere turns to carbolic acid in the oceans, endangering marine life and disrupting ecosystems on land and at sea; aerosol injection (like all SRM plans) would have no effect on ocean acidification. Plus, all that sulfur in the upper atmosphere would start eating away at the ozone layer again, potentially undoing all the progress we’ve made toward closing the ozone hole (remember that?) since banning CFCs two decades ago. Perhaps most worryingly of all, decreased sunlight plus a stronger greenhouse effect would wreak havoc on the water cycle across the globe by decreasing evaporation and changing rainfall patterns. In the years after Pinatubo blew, worldwide rainfall dropped to more than three standard deviations below normal. That’s a gigantic decrease: to use an (imperfect) analogy, a difference of three standard deviations in adult male height in the United States is the difference between someone who is five feet and eight inches tall and someone who is six and a half feet tall. The years after Pinatubo’s eruption saw droughts over much of the world, many of them severe. Moreover, the drought wasn’t equitably distributed: traditionally dry places (like the middle east) tended to get a lot of precipitation immediately after Pinatubo, while traditionally wet places (like SE Asia and the Amazon basin) got much, much less. Both of these are very bad. And that eruption, remember, only cooled the planet by 0.5 degrees: the effect on rainfall is expected to increase with greater aerosol concentrations, and this plan would call for enough to cool the planet by 4 degrees C or more. The effect on agriculture the world over could very well be catastrophic–arguably just as bad (or worse) than the damage done by warming alone.

We may reach a point where the benefits outweigh the costs–particularly if we continue not reducing our GHG emissions–but who’s to say where that point lies? There’s no way to target this plan by region: either we do it to the whole globe, or we don’t do it at all.
Are we comfortable asking farmers on another continent to sacrifice their livelihoods for the sake of other nations’ unwillingness to reduce CO2 emissions? Should we be comfortable with that? Who even is “we” here?

How much agreement among nations is enough to take this plunge? If the citizens of every nation but one decide that the sacrifice is worth it, would we be justified in starting the program even over that single nation’s protests? What if it’s two nations objecting? What if it’s just under half? These are not idle questions–not mere hypotheticals of the type that concern few but moral philosophers inside the seminar room–but real decisions we could be facing before the century’s out, and I’m very worried that we are in no way ready to handle them. Considering the difficulty we have making tough decisions inside individual nations, I can’t even imagine how we’d even begin to deliberate about this as a species.

But the really scary thing about aerosol injection–the thing that really keeps people who work on this up at night–is just how easy it really is, and how difficult to stop it would be once the compound was released. If a Peter Thiel or an Elon Musk (or an Exxon-Mobil) decided to initiate a program like this, it would be almost impossible to stop them if they managed to get started; we’d just have to wait for the aerosols to fall out of the atmosphere, a process that could take years even after the emissions ceased. The same goes for a single nation that decides such a program is in its national interest and elects to go it alone, full speed ahead and damn the torpedoes–or, in this case, damn the rainfall. Would we go to war to stop this from happening? Would we go to war to make this happen, should we decide it’s in our national interest? Either way, the implications are deeply troubling.

Because the residence time–the amount of time a compound hangs around in the atmosphere before it decays or precipitates out–of aerosols is so much shorter than that of greenhouse gases (sulfur dioxide has a residence time on the order of months, as opposed to decades or centuries for most greenhouse gases), once we start this program there’s really no turning around. We’re almost certain that it would reduce the temperature, but then we’ve got a tiger by the tail: either we keep pumping more compounds into the atmosphere forever, or we stop. If we stop, the aerosols–which, remember, are not gasses but rather small droplets of liquid–fall out of the atmosphere entirely within a year or two. Suddenly, all that warming that they’ve been masking is back, only instead of seeing 4+ degrees warming spread over a century, we see it spread over 24-36 months. It’s impossible to overstate how catastrophic that would be, both for human civilization and ecosystems. It would leave the world with no adaptation time. That means we need to keep this technology running with nearly 100% uptime for an indefinite period into the future, or end up worse off than we would have been if we’d just done nothing. Because it’s so cheap and easy to start, it would be really simple for a billionaire (or single nation) to just start doing this in secret without the world noticing until the temperature reduction signal became detectable by the scientific community at large. At that point, we’re locked in and the single-actor geoengineer has effectively taken the world hostage. This fucking sucks for obvious reasons.

There’s more to say here (for instance, there’s some evidence that change to precipitation patterns is at least partially a function of release location, and so who gets impacted in what way depends in part on how and where the injection is performed), but I feel like this is enough of a start. I can dive into more technical detail if people want.