Thursday, May 3, 2012


date: Thu, 28 May 2009 16:11:03 -0700
from: Darrell Kaufman <>
subject: Re: question
to: "" <>

Yes, I am beginning to think that the model has somewhat lower
sensitivity than seen in the proxy data, as suggested by the reviewer.
And, yes, the model data are based on 50-year means whereas the proxy
data used 100-year means. We did this to keep the n the same for the
two regressions (model = 1000 years; data = 2000 yr). I suppose I
could use 50-yr means from the proxies and just randomly select half
of the data for comparison. Before we go down that road, our NCAR co-
authors are going to get back to me with more information on the model.

This just in from Dave:
I don't want to jump to the conclusion that the model is not sensitive
enough. After all, for the T31 simulation (the resolution used for the
mid-Holocene orbital run) t is ~3 degrees colder in the Arctic in the
model's pre-Industrial climate compared to present day (1990); this is
almost 3X the global average change. The Arctic is even colder when
increasing the resolution to T42 (more realistic sea ice). It is
possible that orbital forcing using T42 resolution would have higher
sensitivity - that is one caveat we could make.

Let me know if you have other suggestions.

On May 28, 2009, at 3:45 PM, wrote:

> Darrell
> I am intriqued about how the model can get the one very anomalous warm
> point that does likely influence the regression result for simulated
> data
> against model insolation. Have you looked into this? It is likely
> fortuitous that the regressions in Fig 4 give the same slope,
> because of
> this. Also , what would be the difference if insolation was
> calculated in
> the same way for both a and b ? Am I right in remembering that they
> were
> not in fact the same?
> Keith


No comments:

Post a Comment