To: Keith Briffa <email@example.com>
Subject: Bullet debate number 1
Date: Wed, 15 Feb 2006 11:46:17 -0700
Cc: Eystein Jansen <firstname.lastname@example.org>, email@example.com
Hi Keith (and Eystein - we need your opinion) - thanks for the quick
response. I think it easier (imagining the mess of email that could
result) if we focus on one bullet/email. So I'll start w/ the first,
and hope that Eystein can also weigh in.
With regard to the first one below, I agree that we can leave
statistics out of it. Good point.
But, I think we must at least address Susan's concern. To do
otherwise would be counterproductive. She makes sense. I think your
MWP results is quite appropriate - they were published in Science,
and in my reading of the paper, you are convincing. If it's in the
chapter, it makes sense to draw on it for the exec summary. Please
defend more convincingly, or suggest an alternative way to deal with
Susan's concern - what is the significance (not statistical) of this
one record being warmer? We need to say it.
If you really want to leave as is, please write your response in a
way that I can forward to Susan - we can't ignore he comment in this
case, because other (me, at least) think it makes sense. So we have
to convince her too - this is big stuff for the AR4, and will be in
the TS/SPM. We can't be as vague as the current bullet is.
And as for the MWP box fig, I think it should be as you suggest -
combine the existing fig w/ the new one from Tim and your paper. I
think Tim might already be working on it?
Sorry to be a tough guy, but this bullet needs to be more clear.
>do not think you will like what I say here , but I am going to give
>straight answers to your questions.
>The new draft says enough in the text now about "far-less-accurately
>dated" and "low-resolution proxy records that can not be rigorously
>calibrated" in relation to this paper (Moberg et al.) . It is not
>appropriate to single the one series out for specific criticism in
>the summary . The use of the word "only" implies we do not believe
>it. Mike Mann's suggestion begs a lot of questions about what
>constitutes "significantly warmer". You need to have a Null
>Hypothesis to test . If you mean would the estimates in Moberg and
>the other reconstructions (during medieval time) show significantly
>different means using a t-test - then of course not , but this tells
>us nothing other than they are not likely samples from totally
>different populations - an almost impossible test to pass given the
>wide uncertainties on all reconstructions . Incidentally, we do not
>have formal (calibration ) uncertainties for Moberg anyway (just
>boot-strapped uncertainty on the average low-frequency curve).
>I think the vagueness is necessary - "suggests slightly" and is appropriate.
>I would not call out The results of Tim and my paper either. It is
>just an aside in the Medieval box at present , perhaps with a Figure
>to accompany the original if you agree, but without more text in the
>Chapter , which I do not consider appropriate, it should not be
>highlighted as a bullet.
Jonathan T. Overpeck
Director, Institute for the Study of Planet Earth
Professor, Department of Geosciences
Professor, Department of Atmospheric Sciences
Mail and Fedex Address:
Institute for the Study of Planet Earth
715 N. Park Ave. 2nd Floor
University of Arizona
Tucson, AZ 85721
direct tel: +1 520 622-9065
fax: +1 520 792-8795