All that should be enough to dismiss it, but I might as well go on:
- This wasn't a random survey - all the members were asked to respond and so the ones who did were self-selected.
- The survey report doesn't say what response rate it had in percentage terms.
- The survey could be completed online or faxed in, and the report doesn't describe any security methods (a past online survey of "climatologists" by Dennis Bray had been crashed by denialists who gained password access). Providing identifying information is listed as "optional."
- The question on the cause of climate change is inadequate: it asks to choose between primarily natural factors, primarily human factors, or both factors together. The third option is a different question. They should've asked if it was primarily natural, primarily human, or if the science wasn't clear enough to judge which cause is primary.
- Amusingly, only 20% trust the climate models, but a solid majority think climate change is a significant risk to public safety. How do they know that if the models stink? I guess they trust the theory, or just take a ruler to the instrumental record and extend the upward-sloping line out a few decades. Whatever. (Might be evidence against the idea that the poll was crashed, though.)
UPDATE: 52,000 members, 1,000 self-selected to respond back.