The Perfect Is the Enemy of the Good

Mark Thoma thinks that geoengineering is too risky as a solution to global warming. Quoting an article in Scientific America he writes,

Geophysicist Kate Ricke of Carnegie Mellon University and her colleagues show that one of the more feasible geoengineering methods—injecting reflective particles into the atmosphere to mimic the world-cooling effects of a volcanic eruption—will have effects that vary from place to place. So, for example, India might be rendered too cold (and wet) by a level of particle injection that's just right for its neighbor China while setting the levels to India's liking would toast the Middle Kingdom.

What's worse, the computer models that show that such injections might work in the short term also show that they will change global weather patterns by making part of the atmosphere more stable—and therefore less likely to promote storms. That means less rainfall to go around—and these side effects become worse with time. ...

He goes on to conclude

Engineers used to show up in comments and tell me that, unlike economists, they know how to build systems in ways that prevent the chance of catastrophic collapse like we had in the financial system. ...Even if the models said this will work without any worrisome side effects or geographic differences, the stakes are too high to use that result as an excuse to delay action on the global warming problem. We need to start solving this problem now...

Considering conventional wisdom's solution is ridiculously expensive and politically infeasible, avoiding a viable and inexpensive strategy even if they models say they are fine is embarrassingly sloppy. I'll certainly agree that the climate change models are better than the economists by a long shot (climatologists don't have to worry about the cloud's expectations or rainfall's irrational exuberance) but let's be careful about putting too much faith in them. And even if the models are accurate and this strategy will have adverse side effects (lower rainfall being a big one) that doesn't mean it's not the best option. Cutting global warming the traditional way generally runs in the tens to hundreds of billions of dollars a year. Volcano simulation (from what I read in Superfreakonomics tallies in the tens of millions of dollars a year. That's a lot of extra money to subsidize irrigation.

But what really irks me is that this geoengineering plan could be set up and executed in a few years and we would notice the effects, both positive and negative, in less than a year. Then, if the models were correct, we could just shut it off and, in actually trying something, gain so much insight about fixing the problem that the quality of our discussion of climate change would increase ten fold. And the environmental and economic damage would be minimumal, not to mention that global warming would be (slightly) curtailed.