Monday, February 26, 2007

Interesting...

The EGU sessions are up. I've got a couple of posters in CL20 (probabilistic climate prediction), and was interested to note the following:

From Chris Forest's abstract:
"The estimated 90% range of climate sensitivity is 1.9 to 5.0 K (including expert prior)."
From the abstract of his GRL paper on the same topic, published last year:
"The estimated 90% range of climate sensitivity is 2.1 to 8.9 K."
That was based on a uniform prior, which as little as 12 months ago was such an automatic choice that it literally went without saying. In fact the GRL paper contained results with both uniform and expert priors, but it was only the former that featured in the abstract.

Looks like a step in the right direction to me...

5 comments:

EliRabett said...

Told Jah:)

C W Magee said...

Why would a uniform prior give a higher lower limit?

James Annan said...

LL,

It's simply that Chris's expert prior assigns a rather higher probability to S < 2 than his uniform prior does.

Anonymous said...

Bit of back and forth In Nature today on climate sensitivity:

http://www.nature.com/nature/journal/v446/n7131/full/nature05707.html

James Annan said...

Thanks, although I can't see it until I get back to work tomorrow...