A Big Chill
http://dotearth.blogs.nytimes.com/2008/1…
More on Whether a Big Chill Is Nigh
By Andrew C. Revkin
…I was on the road yesterday and had no time to collate earth scientists’ reactions to the Nature paper positing that the world … is poised to enter a quasi-permanent big chill…
Richard B. Alley, Pennsylvania State University…
The glaciological community has for decades harbored the widespread belief that the thermal evolution of the ice sheet, and the effect of this evolution on ice flow, are central in the ice-age cycling (not all communities agree, but there is plenty of literature on this from the land-ice crowd), so use of a temperature-independent rheology for the ice leaves out one favored explanation for termination of extensive glaciation.
Interesting line of attack: “this is a new idea, and it does not match the favorite old idea. Ah-ha!”
Carl Wunsch of the Massachusetts Institute of Technology has fired quite a broadside:
Surely this isn’t science in any conventional sense. Taking a toy model and using it to make a “prediction” about something nearly a million years in the future, is a form of science fiction … it isn’t science. The prediction itself is untestable – except a million years from now, and the model “tests” that quoted are carefully chosen to be those things that the model has been tuned to get “right,”
So, if a model that is only testable over a very long span of time is not science, then we have just removed much of astrophysics, evolution, all of climatology, and plate tectonics from the realm of science, have we not?
Also, I love the scare quotes. The model does not make predictions – it makes “predictions”.
Events that confirm with the model would not be right – they would be “right”.
Etc.
If I make a four-box model of the world economy, and predict the US stock market level 500 years from now, who would pay any attention?
Given that the climate record of ice ages goes back tens of thousands of years, and the cycle of ice ages is measured over thousands of years, I’m supposed to believe that it’s inherently stupid to make a model that operates on the same time intervals?
It doesn’t deserve the light of day except as the somewhat interesting mathematical behavior of a grossly over-simplified set of differential equations.
Oh noz! Someone is making a model that is simplified from the real world.
Let me repeat that – a model that is simplified!
Hang them!
Why should anyone take it seriously? The wider credibility of the science is ultimately undermined by such exercises.
Unlike, say, the party line on global warming…
Anyway, a rebuttal:
Dr. Crowley has responded to Dr. Wunsch (I’ll end the up-front debate here and any further comments by scientists can come in the comment string below):
These are pretty harsh words and they almost sound convincing. Except that we do make an attempt to validate the model with respect to how it has performed against the best estimate of global ice volume we have – Shackleton’s 2000 record (I might add that I am not sure of any other modeling study that has tried to validate itself against the Shackleton record). And it is this very same model run that is run past the present.
We also point out that similar processes appear to have gone on in the deeper geological past with respect to Antarctica. We again point out that observations clearly indicate an increase in variability towards the present. Does one somehow expect that just to stop? If so, why, Carl?
With respect to the dismissal of the model, we point out that we have separately applied the model to the great Permo-Carboniferous glaciation on Pangaea and to the Snowball Earth problem (although some people do not like one of the solutions it produces). Are not these tests of the models, or a separate test we have done with respect to noise forcing of the model (referred to in paper)? This is actually a fairly well tested model (Dick Peltier has done separate work on the Pleistocene – again referred to in our paper).
Sorry Carl, this is science – you might not like it, but it is science.
TC
Oh, snap!
