date: Tue, 22 Sep 2009 13:20:56 -0400
from: Jason Smerdon <jsmerdon@ldeo.columbia.edu>
subject: Re: invitation to be a co-author
to: t.osborn@uea.ac.uk

   Hi Tim,

   I have had a chance to go over the paper and include some editorial comments in track
   changes mode of the attached document.  They should all be self explanatory.  Below is a
   list of some specific comments for further consideration.  I of course defer to your
   judgement on whether or not you think the suggestions are worth addressing.  Thanks again
   for including me on this and let me know if you need any further assistance (or
   clarifications on what I write below).

   Many greetings from Spain and Fidel,

   Jason

   ------------------------

   Some Specifics

   1. In Section 2 I have described the uncertainty calculations for the boreholes.  The
   estimates I sent you were +/- 1 SE of the mean trends, but I seem to remember you
   mentioning a 2 SE estimate would be more in keeping with the 95% CI that are typically
   presented for other recons.  I think we had some back and forth with Henry on this, but
   please adjust the text to what you have done for the plot, i.e. used either 1 or 2 SE.

   2.  I have included references where appropriate in the text, but the instrumental
   discussion needs to be clarified a bit.  The Australian data were a collection of
   high-quality data that Henry acquired while down under.  We cited them in Pollack et al.
   2006 with two refs:

   Torok SJ, Nicholls N. (1996) A historical annual temperature dataset for Australia.
   Australian Meteorological Magazine, 45, 251260.

   Della-Marta PM, Collins DA, Braganza K. (2004) Updating Australias high-quality annual
   temperature dataset. Australian Meteorological Magazine, 53, 7593.

   You may want to simply cite Pollack et al. 2006 for these data or the above two refs.

   3. The African data were actually from you when we were preparing the IPCC figures.  I
   assume that they were derived from some version of gridded CRU data for southern Africa,
   but I don't think you sent specifics beyond that.  Perhaps you have it in your records?

   4. In the first paragraph of Section 3 you compare the SH and NH forcing series.  Is it
   worth plotting the difference between these two series as an inset or additional panel?  It
   might be a helpful complement to the descriptions you provide in writing, but I don't feel
   too strongly one way or the other.

   5. Section 3, end of 4th paragraph.  You mention the difference between Ricardo's recon
   using his station data calibration and the new CRUTEM3v record.  You might qualify this a
   bit more.  If the tree recons were calibrated on this new instrumental data they would of
   course match the record better.  It is also worth pointing out how the station data
   estimate matched the old CRUTEM2v grid-box estimate, i.e. are the differences you report
   due to updates in CRUTEM3v or between the station data and the gridded estimates that would
   also have existed in the CRUTEM2v data?

   6. Is it correct that the model uncertainties were estimated from the variability of the
   control run or has something else been done?  If the latter, is it possible to provide
   uncertainty estimates for the NAT500 simulation?  I think this would weaken the case
   somewhat (namely the degree to which the NAT500 and ALL500 sims can be said to be
   different), but it would perhaps be the most honest way of presenting the comparisons.

   7. Section 3, middle of 6th paragraph.  Sentence beginning "There are some periods..."  I
   am not entirely clear what you are saying here.  Are you saying that the proxy recons do
   not include full uncertainty estimates?  This could be clarified.

   8.  I am a little concerned about the borehole referencing and the new CRUTEM3v series.
   You discuss in the last paragraph of Section 3 that the borehole recon sits predominantly
   above the simulation.  But if the new instrumental series is significantly different from
   what we originally used, the borehole recon would need to be rereferenced to the new
   series.  To my eye it looks like the borehole recon is sitting a bit above the trend in the
   instrumental time series and a rereferencing would bring the recon down to where it would
   agree with the simulation more.  If you send me the raw instrumental time series and the
   borehole recons that you plotted, I could check this...

   9. In Table 1 you report correlations between the borehole, model and instrumental series.
    Is this really meaningful, given that the borehole series are just interpolated trends?
   Would it make more sense just to compare the estimated borehole trends and those from the
   instrumental and model data during their period of overlap or say just for the 20th
   century?

   10. In the Intro you mention wether we should include additional refs on uncertainty in
   recons.  You could certainly expand this list with more of the usual suspects (Esper and
   Burger refs would certainly work).  I also like to use some of Mike Evans' work for these
   kinds of things.  This is a good paper that includes a lot of uncertainty discussion:

   [1]http://www.agu.org/pubs/crossref/2002/2000PA000590.shtml

   Figure 10 is particularly relevant for discussions on how uncertainties are timescale
   dependent.  I also like Mike's process-model paper as support for the need to understand
   the multivariate and non-linear impacts in tree-ring responses (and associated
   uncertainties):

   [2]http://www.agu.org/pubs/crossref/2006/2006JG000166.shtml

   Attachment Converted: "c:\eudora\attach\paleodata_model_sh_ver09_jes.doc"

   On Sep 14, 2009, at 5:35 PM, Tim Osborn wrote:

   Hi Jason,
   two weeks or so would be fine... my hint at a relatively quick turnaround
   was to avoid suggestions of major extra work (e.g. "why don't you include
   results from X other GCMs too?").  I'd like to submit by mid October.  Not
   sure where yet; perhaps have a think about journal when you read it in
   detail?
   Unfortunately Keith has been off sick for 3 months, having had to have one
   of his kidneys removed.  He seems to be on the mend, will probably be back
   at work in October.
   Other than that, I think we're all ok in CRU.
   Please say "hi" to Fidel from me.
   Tim
   On Mon, September 14, 2009 4:17 pm, Jason Smerdon wrote:

     Hi Tim,

     Wow.  Thanks for including me on this.  I would be happy to be on the

     paper (I can say that after only a quick scan!).  I am also happy to

     get you comments ASAP.  My only problem right now is that I am leaving

     for Spain on Thursday to spend much of the fall with Fidel G-R in

     Madrid.  I therefore expect that this week and next will be quite

     busy.  It is possible I can have it done earlier (I will have lots of

     time on the plane), but would two weeks be slow for a turn around

     time?  Please feel free to be blunt if you had an earlier date in

     mind...I don't want to slow you up.

     Bottom line: thanks for including me on this and I will work to get

     you comments as fast as possible.

     It was nice to hear from you and I hope all is well with you and

     everyone else at CRU.  Please send my regards to Keith and the others.

     All the best,

     Jason

     P.S.  Did you have a target journal in mind?

     On Sep 14, 2009, at 10:43 AM, Tim Osborn wrote:

     Hi Jason,

     I've now finished drafting a palaeodata-model comparison paper using

     the HadCM3 ALL250/NAT500 runs, with a focus on Southern Hemisphere

     records.  I've used the SH reconstructions that we showed in the

     Palaeoclimate chapter of IPCC AR4.  These included borehole recons

     from Africa and Australia.

     It would be great if you would agree to being a co-author on the

     paper, since it makes extensive use of SH borehole recons, plus I'd

     value your comments on the manuscript as a whole and with some

     specific sections about the borehole records.

     I've taken the liberty of including your name on the complete draft

     that is attached.  Of course if you'd prefer not to be an author,

     that is fine -- just let me know and I'll remove your name in that

     case.

     But hopefully you'll agree :-)

     If so, please send comments, edits etc. to me.

     It's fairly short.  I haven't had any papers as first author since

     2007(!) so I'm under some pressure to submit it soon... hopefully

     you won't suggest any major alterations!

     Best wishes and hope all's well with you,

     Tim<paleodata_model_sh_ver09.doc>Dr Timothy J Osborn, Academic Fellow

     Climatic Research Unit

     School of Environmental Sciences

     University of East Anglia

     Norwich  NR4 7TJ, UK

     e-mail:   [3]t.osborn@uea.ac.uk

     phone:    +44 1603 592089

     fax:      +44 1603 507784

     web:      [4]http://www.cru.uea.ac.uk/~timo/

     sunclock: [5]http://www.cru.uea.ac.uk/~timo/sunclock.htm

     --------------------------------------------------------------------------------------

     Jason E. Smerdon

     Storke-Doherty Lecturer

     Lamont-Doherty Earth Observatory  and

     Department of Earth and Environmental Sciences

     Columbia University

     61 Route 9W, P.O. Box 1000, Palisades, NY 10964

     Phone: (845) 365-8493

     Email: [6]jsmerdon@ldeo.columbia.edu

     Web: [7]http://www.ldeo.columbia.edu/~jsmerdon

     Skype: jason_ldeo

     --------------------------------------------------------------------------------------

   --
   Dr. Tim Osborn
   RCUK Academic Fellow
   Climatic Research Unit
   School of Environmental Sciences
   University of East Anglia
   Norwich  NR4 7TJ, UK
   [8]www.cru.uea.ac.uk/~timo/

