cc: m.hulme@uea.ac.uk
date: Fri, 01 Oct 1999 10:18:41 +0100
from: Phil Jones <p.jones@uea.ac.uk>
subject: Paper 980G by Luterbacher et al
to: b.d.mcgregor@bham.ac.uk


 Dear Glenn,
    I've tried to ring you to talk about this paper that IJC has rejected
 on 21/9/99. My interest is that I'm one of the et al's. I've talked to
 Mike here about it and he suggested I contacted you about the decision.
 Your decision has been made on the basis of one reviewer only. I know from
 Jurg Luterbacher that you've had difficulty getting a second reviewer to
 respond.
    We (me, Mike and the Swiss) think we have a real grievance and are
 questioning the decision. Fortunately Mike didn't attend the last IJC
 meeting otherwise I might have been able to find out who this well
 respected referee is. I know I shouldn't have asked Mike but I was so
 annoyed with the two main comments that I asked him !
    His two points are basically wrong !

 1) 'Patterns during the 20th century are applicable to earlier epochs'.
 This assumption applies to all paleo reconstruction papers ever written.
 OK, it is an assumption called the 'Principal of Uniformitarianism' and
 we could have stated it clearer, but it is one that has been made by
 countless thousands before us. If it is not valid we might as well give up.
 The method used in the paper is not the same as infilling SST fields to get
 complete fields, a technique that I would question ( this technique is
 usually used to get complete fields to drive GCMs). Our paper uses real
 data for the past and attempts through a calibration/verification exercise
 to derive circulation patterns for earlier periods. The only real
 requirement for the technique to work properly is that all the long
 time series used in the reconstructions are homogeneous.

 2) 'Changes in climate forcing through time invalidate such statistical
 relationships'. Here the reviewer's concern is completely wrong as he/she
 intimates because 'year to year variability is much greater than long-term
 climate change'. It is not just greater is at least one order of magnitude
 greater, maybe more. The regression-type relationships derived in the
 paper are almost entirely based on high-frequency relationships. Longer
 timescale variability is relatively small in this regard. Longer timescale
 change over the 1675-present is not that great anyway (the 1730s for
 example are only slightly cooler than the 1980s). 

 Global average sea level pressure can't change but patterns will change in
 different periods and it is these that influence the surface temperature
 and precipitation patterns. The drawback of using surface temperature
 and precipitation patterns to reconstruct circulation indices is that you
 can't then go and look at changes in circulation/surface climate
 relationships as these have been used in the reconstructions. However,
 using methods like this produces far more reliable reconstructions than
 proxy data (trees, ice cores etc) as we can derive indices on a monthly
 basis and not just for a season or year. (By the way proxy data 
 reconstructions of circulation indices suffer the same problems as
 it is temperature and precipitation and not the circulation that alters
 tree ring widths/densities and ice core composition). With the earlier
 paper from the same EU project (Jones et al, IJC 19, 347-364), pressure
 reconstructions were derived from station pressure data, so here from
 1780 we will be able to look at the change in the strength of circulation/
 surface climate relationships through time (I'm doing this in another
 paper).

 Finally, and I didn't mean to get into all this detail, but it seemed to
 just flow, the reconstruction technique does use gradients and not just
 the station data. The canonical correlation technique relates PC-based
 patterns of real cirulation data to PC-based patterns of station temperature
 and precipitation and one pressure site at Paris. The regression
 relationships are based on the patterns, hence gradients are used
 implicitly.

 In conclusion, I'm hoping you will reverse your decision and allow us to
 resubmit a slightly revised, and reduced in size ( it was too long) paper
 and send it out to reviewers who will respond in a reasonable time frame.
 The problems raised by the reviewer are no problems and we can easily
 address them. They don't invalidate the results.

 Cheers
 Phil

 
Prof. Phil Jones
Climatic Research Unit        Telephone +44 (0) 1603 592090 
School of Environmental Sciences    Fax +44 (0) 1603 507784 
University of East Anglia                      
Norwich                          Email    p.jones@uea.ac.uk 
NR4 7TJ
UK

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

    
