RealClimate has an interesting, detailed posting about geo-engineering as a response to climate change, mostly emphasizing the areas that would require more research before it could be seriously considered. Here’s the conclusion paragraph:
The real consensus, as expressed at the National Academy conference and in the AMS statement, is that mitigation needs to be our first and overwhelming response to global warming, and that whether geoengineering can even be considered as an emergency measure in the future should climate change become too dangerous is not now known. Policymakers will only be able to make such decisions after they see results from an intensive research program.
However, there should be useful information even for those whose conclusions are different.
My own concerns are about the unpredictable consequences of apparently straightforward interventions into complex systems: consider introduced species and the unexpected havoc they can wreak. Putting some extra SO2 in the atmosphere doesn’t seem too risky — but I’m sure that letting a couple of rabbits loose in Australia didn’t seem like a big deal either.