Ecohacking, or not?

(Photograph: Alberto Garcia/Corbis)

Would pouring sulfates into the atmosphere -- as the volcano above is doing -- be a good thing for humans to attempt as a solution to global warming?

• Maybe, says this article from today's edition of The Guardian. 

• Maybe not, says this blog entry from Jamais Cascio, CRN's Director of Impacts Analysis.

Geoengineering -- or 'ecohacking' as The Guardian dubs it -- is an idea gaining more and more attention these days. So, will such radical ideas ever be attempted?

We think it's a safe bet to predict that the following things will happen, in roughly this order:

1. New and stronger evidence of atmospheric global warming will be found each year.
2. Clearer results of warming in the form of damaging climate change will occur with increasing frequency.
3. Computer models for determining the probable course of events will become more sophisticated and more accurate.
4. Those models and the evidence from 1 and 2 will show that only the most drastic action (meaning huge cuts in CO2 emissions over a short period of time) could head off the worst case scenarios.
5. Political support for the required drastic action will be in short supply and/or slow in coming.
6. Detailed proposals for geoengineering will gain a fresh hearing as the public, government, and industry all look for a quick fix that may cost a lot but will not require so much lifestyle change. 

Assuming that our crystal ball is functioning properly today, and that the scenario outlined above plays out, what will be the result?

We're quite worried that some attempts at ecohacking will be undertaken -- either by governments or by well-meaning vigilante billionaires -- but that the system dynamics, secondary impacts, and possible side-effects of whatever methods they choose may not be fully understood. The earth's climate is massively complex, with long, slow functions that ebb and flow, while simultaneously being subject to sudden, rapid cascades of change. If we try to mess with all that unadvisedly, we may be very sorry.

Mike Treder

Comments

[Hal]

At what point in that chain of events would you anticipate new technologies arising that will drastically change the state of play with regard to the costs of removing CO2 from the atmosphere, or otherwise mitigating the increase in CO2?

[Tom Craver]

"systemic feedback, secondary impacts, and possible side-effects of whatever methods they choose may not be fully understood. The earth's climate is massively complex, with long, slow functions that ebb and flow, while simultaneously being subject to sudden, rapid cascades of change"

Suppose for the past 40 years we'd been slowly slipping into an ice age due to long lasting human particulate pollution of the upper atmosphere. Suppose the common sense proposed solution was "drastically cut emmissions of particulate pollution and let the atmosphere slowly clean itself" - but that hadn't been done and in fact we'd increased the rate of particulate pollution. And now (in this scenario) Arctic ice is staying frozen year round where it used to thaw, and European ski resorts are being closed due to dangerous avalanche conditions from huge and frequent snowfalls.

And suppose someone was proposing that we consider, as an emergency measure, experimenting with increasing CO2 concentrations to counter the cooling.

Should we decide that climate is too complex and hard to predict, so we couldn't possibly project the impact of increasing CO2 concentration? But isn't that what we're trusting AGW climate models to tell us about the uncontrolled geo-engineering experiment we're running now by rapidly increasing CO2 concentrations?

That argument can be taken either way of course.

Maybe we shouldn't be so trusting of computer climate models. They are supposed to be based on climate science principles, but they still have to be "trained" to match past trends before being used to project forward. That doesn't speak well of our understanding of climate science, or of models based on that!

Alternatively, if we're sure of our AGW models, shouldn't we trust them to tell us whether to take a chance on a particular geo-engineering scheme?

[Mike Treder, CRN]

That's an important question, Hal, but it's very difficult to answer. First, you'd need to have a reasonably sure idea of when those game-changing new technologies can be expected to arrive. And second, you'd have to be able to confidently assign a timeframe for each item in the chain of events.

One problem is that there are so many unpredictable variables involved, such as the potential for rapid acceleration in warming from feedback loops (permafrost melts, albedo changes, etc.), which is why I chose to list the steps involved without trying to date them.

Then throw in some possibilities like global economic booms and busts, wars, revolutions, you name it, that might distract the attention of leaders and the public away from the climate change issue, and you can see that predicting the timing of political responses is also hard.

So, if we don't reach my step #6 above until, say, 2025, by then there might be some new technologies that could make a big difference. However, given the dynamic complexity of the Earth's ecosystem, it might be very hard to know what to do without making things worse. Plus, the longer we wait, the more chance there is of getting stuck in a runaway spiral of feedback loops.

If, on the other hand, events develop quickly and a mandate for action arises as early as perhaps 2015, then it's less likely that we'll be prepared with brave new technologies to correct the problem.

Bottom line is, the more I learn about all this, the less optimistic I become...

[Jamais Cascio]

Actually, Tom, the current generation of comprehensive "general circulation" (atmosphere-ocean) models don't need to be "trained," and do an excellent job of post-dicting past climate. They're not perfect, by any means, but they're significantly more powerful than the versions from even five years ago -- smaller regional "cells," more oceanic layers, more variables to tweak.

The bad news is that, with each successive improvement of our climate models, the news gets worse.

You ask (reasonably) if, since we trust these models to tell us what's happening with AGW, why don't we trust them to tell us what would happen with geoengineering?

We do, to an extent -- models are the driver Robock's piece, for example. But our climate models aren't all we have for understanding what's going on: we also have our observations. With geo, all we have are the models. "So, let's test them and have some observations," you might respond. And we should, but we need to be damn sure that even the smaller test levels won't have unexpected side-effects.

Here's an analogy that helps to illustrate this:

You have a big, transparent tub of swirling water. For awhile now, your annoying older sibling has been adding in some paint, and the water has been gradually getting murkier. The paint eventually precipitates out, and that's good, but the rate at which the paint is going in still exceeds the rate at which it precipitates out. You're arguing with your older sibling to stop it, or you're telling Mom, but with only marginal results.

You get the bright idea to add in something that might make the paint precipitate faster (or turn transparent, or something). Paint thinner, perhaps. You *think* it should work, since you're getting pretty good at spotting the patterns and understanding the flow of the paint in the swirling water. But here's the rub: once you add it in, you can't get it out again. You can't go back to the status quo. If it doesn't work the way you think it should, or even makes things worse, you're hosed.

Okay, over-simplified analogy, but it speaks to the central point: as the climate is a complex, chaotic system, making new changes to the system can be irreversible, and have outcomes that are very hard to predict, even with good models.

All of this may sound like I'm saying that we should never undertake geoengineering, under any circumstances. I'm not; I think it's best left as a last-ditch, oh crap we're boned, option, but research now to figure out the potential side-effects and unintended consequences is absolutely critical.

[Tom Craver]

Jamais -

But are the equations of the new models derived solely from now-better-understood "first principles"?

Or have model makers simply incorporated equations with parameters tuned on the older "training" models - without solid understanding of why those parameters work, or whether they'll keep working for the future (i.e. whether they're really able to predict climate).

If so, I wouldn't trust them to predict the impact of geo-engineering projects, other than maybe those approaches which reduce CO2.