cc: alverson@pages.unibe.ch, christian.pfister@hist.unibe.ch, cullen@ldeo.columbia.edu, drdendro@ldeo.columbia.edu, druidrd@ldeo.columbia.edu, gyalistras@giub.unibe.ch, juerg@giub.unibe.ch, k.briffa@uea.ac.uk, mann@multiproxy.evsc.virginia.edu, p.jones@uea.ac.uk, schmutz@giub.unibe.ch, stocker@climate.unibe.ch, xoplaki@giub.unibe.ch
date: Wed, 26 Apr 2000 10:03:01 -0400
from: Ed Cook <drdendro@ldeo.columbia.edu>
subject: Re: NAO discussion
to: Heinz Wanner <wanner@giub.unibe.ch>

Hi Heinz and others,

Thank you for your comments.  The problem of non-stationarity in the NAO
teleconnection with climate on both sides of the North Atlantic is of
obvious concern.  It would be nice to know if there is a "core region" of
stability in the teleconnection, similar to what appears to be the case
between ENSO and climate in the southwestern US and northern Mexico.

However, assuming that the NAO teleconnection over western Europe and
eastern US is more or less stationary, it still remains that there is an
east-west pressure gradient across the North Atlantic in the Azores High
region, with Lisbon/Gibraltar representing the eastern/European end of that
gradient. This is the reason why I think that the comparisons made in the
Schmutz et al. paper are biased towards the result that they found. If you
only use European-based data to extend the NAO index back in time, it is
certain that this index will be better correlated with the
Iceland-Gibraltar NAO index than with any other estimate based on data that
includes North American proxies. I am not saying that the tree-ring proxies
are perfect, or as accurate as we would like. But the conclusions of the
paper are unfairly negative because they implicitly assume that the Jones
Iceland-Gibraltar NAO index is a regionally unbiased reflection of the
"true" NAO back in time (whatever that is). I do not think that this has
been shown to be the case. So, I do not agree with your conclusion that
your "index can be used for long-term testing of proxy-based indices on the
monthly to decadal timescales." To do so implicitly assumes that your L
index is unbiased.  Since both the Iceland-Gibraltar and L estimates of the
NAO are unlikely to be regionally unbiased (my opinion anyway), using
either index to test proxy reconstructions of the NAO based in part on
North American tree rings may not produce accurate results. This is my main
point. Show me that the L-index is a regionally unbiased expression of the
NAO and I will agree with your conclusion.

Cheers,

Ed

>Dear colleagues,
>
>after several discussions I had with other colleagues on the same topic
>within the last four weeks it was very interesting for me to follow the
>mail discussion concerning NAO reconstructions between Ed Cook, Mike Mann
>and Christoph Schmutz. I will not waste your time and just express my view
>very shortly:
>
>(1) There is no doubt that the dynamical structure behind teleconnections
>like NAO, AO and AAO, ENSO, PNA, EU, etc.
>is highly non-stationary / intermittent in space and time. Looking at the
>NAO/AO dynamics, this fact has already been recognized by European
>scientists like Exner in the early years of this century.
>
>(2) It is therefore difficult to find the best available two-point indices
>in the modern time (based on instrumental or early instrumental data).
>
>(3) The intermittence of the climate system with its strongly variable
>forcing conditions requires high-quality reconstructions of these indices
>(with a sound consideration of the responsible mechanisms!).
>
>(4) Even very nice publications exist, we have to study in more detail how
>far the different sources (data from natural archives AND documentary data)
>- if available - are suited to reconstruct the best possible SLP, SST and
>SAT patterns (e.g., temperature versus precipitation reconstructions,
>single high quality point informations versus dense networks with lower
>quality indices, winter versus summer phenomena), and where. These data
>have to be used to get to significant composite reconstructions which also
>allow to reduce the noise.
>
>I hope we can discuss these problems in upcoming conferences and workshops
>soon.
>
>With very kind regards,   Heinz
>
>
>*********************************
>Prof. Heinz Wanner
>Institute of Geography
>Climatology and Meteorology
>Hallerstrasse 12
>CH-3012 Bern (Switzerland)
>Phone   +41 (0)31 631 88 85
>Fax     +41 (0)31 631 85 11
>e-mail  wanner@giub.unibe.ch
>http://www.giub.unibe.ch/klimet/
>*********************************



