On Sunday, February 19, 2012 10:26 AM, Brian Blais wrote:
<snip>
d) the models are tuned to this temperature profile. so, if the proxy
data are wrong, or the temperature reconstructions from them are wrong,
then the models will be wrong. It won't matter if all the models agree,
or that they agree with the proxy data (to which they are tuned).
</snip>
What do you mean by the "models are tuned to this temperature profile"?
The models (that I am aware of) are based upon the physics, not the
prior temperature records. Certainly, prior temperature records can be
used to check and "fine tune" sub-grid scale processes, but that doesn't
necessarily mean that such fine-tuning, even if based on erroneous data,
invalidates the models. I myself have found erroneous fine-tuning in
the models and, when corrected, had negligible impact on the
simulations. I'm just wondering if perhaps I am misinterpreting what
models and proxy data you are referring to.
By the way, except for the unusual warming in 1998 and the unusual
cooling in 1992-1993, the models from 1988 seem to have predicted the
warming pretty well so far (as far as I can tell from Mann's 5th slide
in <http://www.meteo.psu.edu/~mann/Mann/lectures/GMUOct11.ppt>).
In fact, I have yet to see a prediction based on scientific processes
(i.e., given the rise in CO2 and other physical processes) that has come
as close to the observations as the predictions made by the climate
models. Science is about making testable predictions. For those who
question whether the global warming is unprecedented and/or whether
humans are responsible for most of it, I hope they create a model (based
on physical processes) that will make testable predictions of its own.
That way we can make reasonable comparisons.