date: Tue Jul 19 15:38:31 2005
from: Keith Briffa <k.briffaatXYZxyz.ac.uk>
subject: Re: thoughts and Figure for MWP box
to: Tom Crowley <tcrowleyatXYZxyze.edu>, Jonathan Overpeck <jtoatXYZxyzrizona.edu>
Tom et al
thanks for remarks - in response to Tom's questions
At 18:23 18/07/2005, Tom Crowley wrote:
a few comments -
1) are you trying to choose between my way of presenting things and your way - ie, w
2) with your data, do they all go through from beginning to end?
pretty much - and have been standardised over the maximum period for each (not necessarily
the best way?)
3) why include chesapeake, which is likely a salinity record?
Because Moberg used it in their latest reconstruction - I agree that I would not use it
because of the dubious temperature signal (salinity effect and no local replication) and
poor dating control (and I do not like the way the Moberg method effectively over weights
the low-frequency predictor series in their analysis).
4) some of your data are from virtually the same site - Mangazeja and yamal are both w.
siberia - I composited data available from multiple sites to produce one time series,
which is equally counted against the other regions, which might (greenland, w.U.S., e.
Asia) or might not have multiple records in them
Just to reiterate - I understood after the group chat with Susan S. in Beijing , that we
were being asked to try to produce a "cloud" diagram including as many of "original"
predictor series ,from all the reconstructions, to see if it provided an "obvious" picture
of the unprecedented warming over the last millennium or so. Tim and I are in no way trying
t produce a different Figure for the sake of producing a different Figure . In practice
this is hard to do (because some records are sensible "local" composites already, and how
far do you go in showing all input data? The problem of what and how to composite is tricky
- and no obviously "correct" way is apparent.
Having said this , Tom's way is fine with me (provided the composites are robust) and we
get general agreement. Am happy to go with Tom's Figure , or version that incorporates as
many records as possible - but as we have said - without the composite or temperature
scaled add ons.
5) I am not sure whether it is wise to add me to the CA list, just because the reviewer
is supposed to be impartial and a CA loses that appearance of impartiality if he has now
been included as a CA - may want to check with Susan S. on this one to be sure - still
happy to provide advice
My own position on this is that you are an "unofficial" referee, who has (and still is)
making a significant contribution - I see no conflict
6) I am happy to go in either direction - include or not include my figure - all I need
are specific directions as to what to do, as CLAs you people need to decide, and then
just tell me what or what not to do
Agree - CLAs please rule on the individual record/composite question - I am very happy to
go with Tom's Figure. We did ours because we were asked to.
7) I am a little unhappy with the emphasis on hemispheric warmth - lets face it, almost
all of the long records are from 30-90N - the question is: how representative is 30-90N
to the rest of the world? for the 20th c. one can do correlations with the instrumental
record, but co2 has almost certainly increased the correlation scale beyond what it was
Absolutely agree , and hope this comes over in text (and bullets) - if not needs
strengthening (note David R's comments).
you could correlate with quelcaya - not sure how many other records there are that are
annual resolution - in the tropics I have produced a tropical composite (corals +
Quelc.) but it only goes back to ~1780 - corals just don't live v long - in that
interval at least the agreement is satisfactory with the mid latitude reconstruction but
there is only 100 years extra of independent information beyond the instrumental
We have gone round in circles over this , but understand consensus to be that Quelc. not a
clean temperature record. Agree corals would be better longer (the new coral-based
reconstruction by Rob Wilson et al
goes back to 1700 and shows unprecedented tropical warming . Along with the text from Julie
we can not go much further, but the importance of extending the tropical (and SH records
needs to be very clear)
.THIS MAY NEED TO BE ADDRESSEDAS A GENERAL ISSUE SOMEWHERE (SHORTLY) IN YOUR DOC
Really hope it is already - but advise if you think not
Thanks for this - lets take lead from J and E now (also can you advise on state of play
with the Hegerl et al manuscript?)
Jonathan Overpeck wrote:
Hi Keith, Eystein and Tom: See below (BOLD) for my comments. Thanks for moving this
forward and making sure we do it right (i.e., without any bias, or perception of bias).
Dear Peck, Eystein and Tom
At this point we thought it was important to review where we think we are with the MWP
First, we have no objection to a Figure . Our only concerns have been that we should
1/... be clear what we wish this Figure to illustrate (in the specific context of the
MWP box) - note that this is very different from trying to produce a Figure in such a
way as to bias what it says (I am not suggesting that we are, but we have to guard
against any later charge that we did this). We say this because there are intonations in
some of Peck's previous messages that he wishes to "nail" the MWP - i.e. this could be
interpreted as trying to say there was no such thing, and
SORRY TO SCARE YOU. I **ABSOLUTELY** AGREE THAT WE MUST AVOID ANY BIAS OR PERCEPTION OF
BIAS. MY COMMENT ON "NAILING" WAS MADE TO MEAN THAT ININFORMED PEOPLE KEEPING COMING
BACK TO THE MWP, AND DESCRIBING IT FOR WHAT I BELIEVE IT WASN'T. OUR JOB IS TO MAKE IT
CLEAR WHAT IT WAS WITHIN THE LIMITS OF THE DATA. IF THE DATA ARE NOT CLEAR, THEN WE HAVE
TO BE NOT CLEAR. THAT SAID, I THINK TOM'S FIGURE CAPTURED WHAT I HAVE SENSED IS THE MWP
FOR A LONG TIME, AND BASED ON OTHER SOURCES OF INFO - INCLUDING KEITH'S PROSE. THE IDEA
OF A FIGURE, IS THAT FIGURES CAN BE MORE COMPELLING AND CONNECT BETTER THAN TEXT. ALSO,
THERE ARE MANY WAYS TO LOOK AT THE MWP, AND AS LONG AS WE DON'T INTRODUCE BIAS OR
ANYTHING ELSE THAT WILL DILUTE THE MESSAGE IN THE END, THE IDEA IS TO SHOW THE MWP IN
MORE WAYS THAN TWO (THAT IS, THE EXISTING FIGS IN THE TEXT THAT KEITH AND TIM MADE).
2/ ...agree that we have done this in the best way.
The truth is that there IS a period of relative warmth around the end of the 1st and
start of the 2nd millennium C.E. , but that there are much fewer data to base this
conclusion on (and hence the uncertainty around even our multiple calibrated multi-proxy
reconstructions are wide). The geographical spread of data also impart a northern (and
land) bias in our early proxy data.
NEED TO BE CLEAR ABOUT THIS BIAS IN THE CAPTION AND BOX TEXT
My understanding of Tom's rationale with the Figure is that we should show how, because
the timing of maximum pre-20th century warmth is different in different records, the
magnitude of the warmest period (for the Hemisphere , or globe, as a whole) is less than
the recently observed warmth.
YES, BUT IN A WAY THAT SAYS "LOOK, HERE ARE THE ACTUAL REGIONAL CURVES - CHECK IT OUT
FOR YOURSELF" INSTEAD OF JUST SAYING (IN A SCIENTIFICALLY MORE STANDARD MANNER - HERE
ARE THE VARIOUS, MOST ROBUST, LARGE AREA RECONSTRUCTIONS. IN MY MIND, THE LATTER
(KEITH/TIM FIGS IN THE MAIN TEXT) WILL BE THE MOST APPEALING/CONVINCING TO PALEOCLIMATE
SCIENTISTS, BUT TOM'S MIGHT HELP THERE, AND CERTAINLY WITH NON-PALEO SCIENTISTS AND
POLICY FOLKS. MIGHT HELP... IF IT DOESN'T NOTHING LOST, BUT IF IT COULD HURT CONVEYING
UNDERSTANDING, THEN ITS BAD TO USE THE NEW FIGURE.
The reconstructions we plot in Chapter 6 already express the mean Hemispheric warmth
(after various selection and scaling of data), and so the additional information that
the MWP box figure should show must relate to the scatter of the proxy data. There seems
to be a consensus that this is best done by showing individual records , and we are
happy to agree.
What we worry very much about, however, is that we should not produce a Figure that then
conflicts with the picture of proxy evidence for Hemispheric mean warmth as a
whole,shown in the main Chapter Figure. By showing a composite (as Tom has done) and
scaling against another (30-90degrees N) temperature record - this is just what is done.
ABSOLUTELY RIGHT - CAN'T HAVE CONFLICT.
As we promised, Tim has produced a similar Figure, using the same series plus a few
extras, but omitting the composite mean and the scaling against instrumental
temperatures. The idea was to include as many of the original input series (to the
various reconstructions) as we could - though avoiding conflicting use of different
versions of the same data. The precise selection of records will have to be agreed and,
presumably, based on some clear, objective criteria that we would need to justify (this
will not be straight forward). This, along with Tom's plot (forwarded by Peck) is in the
We would like to get your opinion now, and especially Tom's, on the points regarding the
composite and scaling. We would be in favour of just showing the series - but do they
make the point (and emphasise the message of the text in the box)? Or does the scatter
of the various series as plotted, dilute the message about the strength of 20th century
mean warming (note the apparently greater scatter in the 20th century in our figure than
in Tom's)? Can you all chip in here please.
WHAT ABOUT THE IDEA THAT WE ONLY SHOW THE SERIES FOR THE MWP, SINCE THE COMPARISON TO
THE 20TH CENTURY IS DONE WELL (AND BEST?) IN THE TEXT FIGS (WHICH I'M ATTACHING JUST IN
CASE TOM DOESN'T HAVE, ALONG WITH THE TEXT - IF YOU HAVE TIME, TOM, PLEASE READ COMMENT
ON ANYTHING YOU WISH, BUT CERTAINLY THE LAST 2000 YEARS BIT - ASSUME YOU'LL BE DOING
THIS AT THE REVIEW STAGE ANYHOW...)
ANOTHER THING THAT IS A REAL ISSUE IS SHOWING SOME OF THE TREE-RING DATA FOR THE PERIOD
AFTER 1950. BASED ON THE LITERATURE, WE KNOW THESE ARE BIASED - RIGHT? SO SHOULD WE SAY
THAT'S THE REASON THEY ARE NOT SHOWN? OF COURSE, IF WE ONLY PLOT THE FIG FROM CA 800 TO
1400 AD, IT WOULD DO WHAT WE WANT, FOCUS ON THE MWP ONLY - THE TOPIC OF THE BOX - AND
SHOW THAT THERE WERE NOT ANY PERIODS WHEN ALL THE RECORDS ALL SHOWED WARMTH - I.E., OF
THE KIND WE'RE EXPERIENCING NOW.
TWO CENTS WORTH
Keith and Tim
P.S. We agreed in Beijing that we should definitely ask Tom to be a CA .
TRUE - BUT HAS ANYONE CONFIRMED W/ TOM. TOM, YOU OK W/ THIS?
THANKS - A GREAT DISCUSSION, AND LETS SAY THE JURY IS STILL OUT ON THIS FIGURE UNTIL WE
ALL ARE COMFORTABLE WITH WHAT IT LOOKS LIKE IN THE END.
Professor Keith Briffa,
Climatic Research Unit
University of East Anglia
Norwich, NR4 7TJ, U.K.
Attachment converted: Macintosh HD:mwpbox_figures.pdf (PDF /�IC�) (0008A8AE)
Professor Keith Briffa,
Climatic Research Unit
University of East Anglia
Norwich, NR4 7TJ, U.K.