> It seems credible as a way to slow global warming, but there will be side effects (as there has to be with such significant modification of our environment) that are hard to predict, and probably some of those side effects could be harmful.
I think it's certain that some of the side-effects will be harmful, and it's largely a question of magnitude. Because sulfur dioxide has a limited lifetime in the atmosphere, however, it can be tested safely(1) at scale. From the article:
The data from the Mount Pinatubo eruption tells us that sulfur dioxide in the stratosphere can cause cooling comparable to the global warming we've seen so far. But it tells us something else as well: The effects of stratospheric sulfur dioxide have a limited lifetime, on the order of just a couple of years. This is bad in the sense that to be useful, the injection program has to be sustained. But it is good -- very, very good -- in the sense that it reduces risk. The program could be scaled up, the impact could be measured, and, if the side-effects are too serious, it can be scaled back down again very rapidly. This greatly reduces the risk of the undertaking.
(1) The side-effects shouldn't be any more harmful that what we've already lived through with the Mount Pinatubo eruption.