And they don’t come more mega than polar bears. There’s a new report out, ClimateFeedback has blogged it, so I feel duty bound to snark about it.
Skipping rapidly over the press release (note how the “will” of the title is “could” in the text) we come to the reports themselves. And how nice: there is one on model uncertainty. Which even tells you how the models they used were selected: The proposed selection criterion selects models with less than 20% error in their simulations of present-day September sea ice extent. Its probably fair enough, if not very exciting (if you want to feel skeptical look at their fig 9, which is from Stroeve).
But I was more interested in how they match polar bear survival to sea ice. For the southern Beaufort Sea (there seems to be more on this than elsewhere, because this is the US bit?), this turns out to have been done on the basis of capture-recapture studies giving results for “good” ice years 2001-3 and “bad” ice years 2004,5. That doesn’t seem like very many years to me. So clearly they are going to have to treat uncertainty carefully. The abstract sez: Deterministic models yielded estimates of population growth rate lambda under low ICE conditions in 2001-2003, ranging from 1.02 to 1.08. Under high ICE conditions in 2004-2005, estimates of lambda ranged from 0.77 to 0.90. The overall growth rate estimated from a time-invariant model was about 0.997; i.e., a 0.3% decline per year. (nb, ICE in the above is their variable for, confusingly, ice-free days; so high ICE is low ice is bad for bears. nb also ICE is about the ice in the region of the bears, so may bear little relation to overall arctic conditions).
OK, sounds good so far: but then if you look much further down at table 1 you discover that those are just for the estimates by year: eaach of which have substantially higher uncertainties (all of which come from some model which I know nothing of). So the 2001 value, which is 1.059, actually has a 90% confidence range of 0.083 to 1.093. I’m presuming the pop growth has the obvious interpretation: pop(year+1)=pop(year)*lambda, so < 1 means eventual disappearence and > 1 means eventual domination of the earths biomass. 2005 is a bit more constrained at 0.577 to 0.959.
Then the model, of course, assumes that all the variance good-bad is down to ICE. It might be, but with such a small sample of years there could easily be other things going on. So the projection assuming continuing current conditions (-0.3%/y) is so small as to be essentially constant; you need to put in future warming, hence less sea ice, to get declines. They do say The parameter estimates on which the demographic models are based have high levels of uncertainty associated with them but then proceed to wish that away. Another little caveat is that the population dependence on ICE, when projected into the future, goes far outside the range for which the model was calibrated.
There is some about the northern Beaufort Sea too. I was puzzled by the word “Harvest” keeping on popping up, until I realised it meant “killed by people” (there is a bit at the end where they make the astonishing statement “we predict that if… the harvest becomes unsustainable, the population will likely decline”. Only “likely”?). But anyway… sadly that study doesn’t produce comparable population-expansion numbers to try to compare with the previous. This study, despite having a much longer period (1971-2006) can’t such much about sea ice declines impacting the bears because in that region there always is ice, oh dear.
Conclusion: the idea that polar bears might well decline if there is less sea ice seems fairly plausible, but of these two reports one is based on such a small sample of years that it doesn’t really strengthen the “plausible” up very much; and the second doesn’t consider that sort of variation. Exactly where the ice is, is far more important than its overall extent.
[Update: as JOS says in the comments "The lawsuit is part of a bigger campaign to get the government to regulate AGW pollution" and I think thats the bottom line. I don't like this, and I doubt its a good tactic -W]