date: Wed, 30 Sep 2009 16:53:55 -0400
from: jarmour@nas.edu
subject: Final Decision made for PNAS MS#2009-09401
to: p.jones@uea.ac.uk

   MIME-Version: 1.0 Content-Transfer-Encoding: binary Content-Type: multipart/alternative;
   boundary="_----------=_12543440352625311" X-Mailer: MIME::Lite 3.024 (F2.74; T1.27; A2.04;
   B3.07; Q3.07) Date: Wed, 30 Sep 2009 16:53:55 -0400 Message-Id: <62125434403547@ejpweb15>
   September 30, 2009
   Title: "Improving Spatial Temperature Estimation of Global Climate Change - The Case of
   China"
   Tracking #: 2009-09401
   Authors:
   Jin-Feng wang (Chinese Academy of Sciences)
   Mao-Gui Hu (Chinese Academy of Sciences)
   George Christakos (San Diego State University)
   Cheng-Sheng Jiang (Chinese Academy of Sciences)
   Yan-Sha Guo (Chinese Academy of Sciences)
   Ai-Hua Ma (Beijing Normal University)
   Dear Dr. Jones,
   Thank you for reviewing the manuscript by wang et al. [MS# 2009-09401]. We appreciate your
   generosity in contributing your time, expertise and judgment to maintaining the standards
   of PNAS in particular and science in general. On the basis of your opinion and that of the
   other reviewer, Michael Mann has decided to Reject the manuscript.
   Sincerely yours,
   Josiah Armour
   PNAS Editorial Office
   (p) 202.334.2679
   (f) 202.334.2739
   (e) jarmour@nas.edu
   -----------------------------
   PDF for Review (or manuscript file when applicable): [1]Merged PDF
   Resubmission Cover Letter (when available):
   Reviewer #1's Review for 2009-09401:
   Suitable Quality?:
   No
   Sufficient General Interest?:
   No
   Conclusions Justified?:
   No
   Clearly Written?:
   No
   Procedures Described?:
   No
   Supplemental Material Warranted?:
   Not Applicable
   Willingness to Re-review?:
   No
   Comments (Required):
   This is a very short but disjointed paper that I found very hard to comprehend. It appears
   to propose another method to optimize temperature interpolation for area averaging, but I
   couldn't understand what is proposed or why it is unique. Also the paper sections appear to
   be out of order. This problem has been analyzed by many researchers over the years and from
   what I can gather this really doesn't provide an improved method, at least not by the
   evidence in the paper.
   Reviewer #2's Review for 2009-09401:
   Suitable Quality?:
   No
   Sufficient General Interest?:
   No
   Conclusions Justified?:
   No
   Clearly Written?:
   No
   Procedures Described?:
   No
   Supplemental Material Warranted?:
   No
   Willingness to Re-review?:
   Yes
   Comments (Required):
   Review of Wang et al
   General
   This paper looks on a very quick read as if it has already been submitted to another
   journal (e.g. Nature as there is a short Methods section at the end). I don't think it is
   appropriate for PNAS, as the authors (who come from the climate impacts field) seem unaware
   of much of the climatic literature. Even the papers that are referenced have not been that
   well read. I can fully understand why the paper was rejected earlier and it should be
   rejected again.
   If a paper like this makes the claims that it does, then it should have much more detail
   with the results fully justified. I'd recommend the authors write up a paper with about ten
   times the detail. A simple way of showing that their unintelligible method is better would
   be to give full details and show that it is working by leaving out some data. With 637
   sites in the Chinese network it would be quite easy to do this.
   As the paper is poor, I'm only going to make a few points.
   1. Much of the abstract is a set of unjustified statements. Efforts are being made to use
   more data, but it should also be realised that much of the additional data are correlated
   to data already available. Just because more stations are being used doesn't mean the
   results are any more accurate. To assess accuracy there has to an understanding of the
   number of spatial degrees of freedom - which is discussed in reference 25. More station
   data doesn't always mean greater accuracy.
   2. It is unclear what the standard deviation of the annually averaged temperature
   predictions is, as well as having any idea what the standard error deviation is. These two
   terms need to be defined with formulae. In particular, I've no idea what the second is -
   that plotted in the right-hand plots in Figures 2 and 3. Also I cannot understand why the
   data are plotted only from 1996 as Chinese temperature data in this network extend back to
   the 1950s.
   3. None of the explanations in refs 2-15 is the correct one for the global temperature
   series. Many are impacts, so can't be the cause. I'd suggest the authors read Ch 9 of the
   2007 IPCC Report - WG1.
   4. The authors use the term predictions a few times. None of the observed station
   temperature data are predictions. Meaningful predictions are something quite separate from
   the accuracy of the global temperature record.
   5. The authors should look at Ch 3 of the 2007 IPCC Report (WG1) to see what the
   inhomogeneities are that affected land and marine temperature records. Ref 18 appears to
   look at the future. What is needed to be known is changes that have occurred with site
   moves and local and environmental changes.
   6. In the analysis of temperature data, two aspects need to be realised - absolute
   temperatures and temperature anomalies. It has been shown in countless papers that these
   are best analyzed separately. I suspect you have got the result you have because you are
   including the absolute temperatures in your analysis.
   7. There are numerous techniques for spatial interpolation and averaging. As I've stated
   earlier, to show that yours is better requires a full and complete analysis in a
   specialized journal.
   8. Which Chinese dataset of 637 stations is this? Ref 21 gives some details about different
   Chinese networks. It is unclear what your one is based upon. Has it been homogenized for
   example?
   9. Your proposed title for your technique 'Nonhomogeneity' is unfortunate. I'd suggest you
   find another name and also read Peterson et al (1998) for some of the methods that have
   been used.
   10. Finally, as stated earlier, the method as given in the final section is unintelligible.
   Papers like this need to be 10 times larger with complete details. This is just too brief
   to be able to understand. It is not clear whether you are producing a gridded dataset
   across China, or just a new Chinese average?
   References not in the reviewed paper
   Peterson, T.C., T.R. Karl, P.F. Jamason, R. Knight and D.R. Easterling, The first
   difference method: Maximizing station density for the calculation of long-term global
   temperature change. J. Geophys. Res., 103, 25967-25974, 1988.

