Wednesday, August 30, 2006

Eight ways in which Annan and Hargreaves are wrong.....

Ok, that should be possibly. Annan and Hargreaves have used Baysean methods to set limits on climate variability using modeling results. Their estimates are only as good as the models and the observations that are available. Comes A. Barrie Pittock, from CSIRO in EOS to list a number of reasons why scientists might be underestimating climate change

  1. "Global dimming is decreasing", a large rapid decrease in a negative forcing.
  2. "Permafrost melting is widespread", increasing greenhouse gas warming in the Lord knows what ways
  3. "Biomass feedbacks are kicking in" in ways that constitute a positive feedback
  4. "Artic sea ice is retreating rapidy", a positive feedback via albedo changes
  5. "The northern and southern annular modes have become more positive"
  6. "Rapid disintegration of ice shelves around the Antarctic Peninsul, surface melting of the Greenland Ice Sheet and acceleration of outlet glaciers"
  7. "Tropical cyclones may be more intense"
  8. "A significant slowing of the Atlantic meridional overturning circulation" and other changes in the North Atlantic
To the extent that we are entering an area of the climate parameter space that we have not previously visited (not in the observations) or where phenomina not in or relatively minor in the models become important (not treated or properly treated) Bayesean statistics can be little help. Moreover, the approach has a weakness that it considers all inputs on a nuetral basis, whereas an expert panel might be able to figure out that for certain conditions only one of the inputs is important.

This is not to say that Annan and Hargreaves are wrong, just that it can only be as complete as its inputs, and to the extent that we know that the inputs are lacking, we can question the answer they offer.


coby said...

Well, to be fair to A&H there conclusion was not about climate variability but climate sensitivity to 2xCO2, AFAIUI. As such, the GHG feedbacks you mention are not releant to their conclusions.

That said, clearly the more important question is how much, all told, is the climate going to change? I never thought that their paper was in any way reassuring.

James Annan said...


As Coby points out, this stuff is all fairly tangential to our work. As for "reassuring", well it is surely a bit more reassuring than thinking that climate sensitivity might be 6C or higher, which is an idea that some have proposed (and indeed still seem to be desperately clinging to, against all the evidence). But I wouldn't want anyone to think this means that I don't believe all the standard stuff about climate change.

EliRabett said...

I absolutely believe Annan & Hargreaves conclusion about the PREDICTIONS of climate sensitivity.

What I worry about is what is NOT in the models, and from what I can gather, the net sum of large negative outcomes is a lot more worrisome than the net sum of large positive outcomes.

I think James has nailed it with respect to current theory. I think that policy has to worry about not only what we know with a large degree of certainty, but the possible omissions. Increasingly I think these are biological, particularly that global climate changes will affect local biology which will exacerbate the local condition, which, well you get the idea.

Annan & Hargreaves are proponents of Baysean probability analysis, another sort is the Delphic panel. I think you would get a different answer from that sort of analysis, would it be better? Who knows?

Hank Roberts said...

I do wish we had more biologists involved in this whole discussion, in its many forums -- and all the other areas of study that might impinge on where we're going.

The old toast -- "here's to pure mathematics, may it never be of any use to anybody" -- was idealistic in its day, but by the time Tom Lehrer was writing lyrics, the idealism came back as "once the rockets go up, who cares where they come down, that's not my department, ...."

I think James has done a good job of clarifying the question why "climate sensitivity" doesn't mean "how sensitive or reactive or touchy the climate is going to be in the next few decades" -- it's about results in the long run after we are all dead when things settle down.

But I wish there were a better explanation of the whole rate of change question. I suppose that's akin to wishing everyone learned and understood calculus.

Now that's scary.

Anonymous said...

It seems to me that arguing about climate sensitivity is largely a distraction.

If nothing is done to stop CO2 emissions, it will not matter whether the sensitivity is 3C or 6C for a doubling of CO2. Either way, the 6C increase (and greater) will eventually occur. It will just take longer.

It's time to stop debating the science and to start DOING something since even an increase of 3C could have disastrous consequences.