I'm coming around: Due to, in no small measure, Eno's consistent and excellent posts concerning the real science behind global warming and other environmental problems, I have been trying to follow a more nuanced and studied approach to these issues when I read of them. Recently, the usually excellent journal, Nature, published a study titled "Extinction risk from climate change". The thrust is that within 50 years, up to 1.25MM species may be extinct from our warming trends. Alarming stuff, and when it's published in Nauture, you at least have to take notice, and cannot immediately dismiss it as alarmist fluff.

Well, fortunately we have Easterblogg who takes the article to task. First, Gregg notes that the entire study is based on computer modeling. And, when you base something entirely on computer modeling...well, you've seen the Matrix movies haven't you? You can make the computer say whatever you want. More accurately, depending on your inputs and assumptions, the computer is bound by its own operating logic to produce results in accordance with same. Or as Gregg puts it:

Computer models are also notorious for becoming more unreliable the farther out they project, as estimates get multiplied by estimates, and then the result is treated as specific. This is a 50-year projection, and everything beyond the first few years should be treated as meaningless statistically, given that tiny alterations in initial assumptions can lead to huge swings at the end of a 50-year simulation. Nature is a refereed journal, but it appears that all the peer-reviewers did was check to make sure the results presented corresponded to what happened when the computer models were run. There does not appear to have been any peer-review of whether the underlying assumptions make sense.

Now, you'd at least think, that despite whatever the computer is saying, it would have to be modeled based on empirical data. Like say maybe you looked at the last time the planet went through a warming spell, and try to determine through fossil records and the like as to how the species handled the change, right? Ummm, I guess that's not required either. Gregg again:

Estimates of global warming vary quite widely, as they too are driven by computer model, but the Intergovernmental Panel on Climate Change (IPCC), the most-cited source of global-warming projections, now expects somewhere around three to six degrees Fahrenheit of warming by 2100. (IPCC estimates are all over the map; set that aside for the moment.) Assuming that happens--many estimates are lower--we'd expect one or two degrees of warming by 2050. European temperatures rose naturally by one or two degrees at the end of the "Little Ice Age" of the fourteenth through nineteenth centuries. This rise did not cause a mass extinction in the region; in fact, it appears to have caused few or no extinctions. Why would the same level of temperature increase suddenly trigger a mass extinction now?

Anyway, read the post, it's quite convincing and from a source that you wouldn't ordinarily expect.