date: Wed Jun 25 14:57:44 2003 from: Keith Briffa subject: RE: Regarding paper submitted to The Holocene to: "Isaksson, Elisabeth" Dear Elisabeth I am pasting below two reviews of your paper. I have had a third reviewer look briefly at the paper and at these reviews ( particularly because of the delay on my part ) and the third reviewer agreed that the other reviews were reasonable. You will see that all of them agree that the paper should not be published in its present form . The recent added delay has been while I then went through the paper and all reviews carefully myself to give what I believe is an objective opinion of my own . I too feel that I can not justify acceptance in the present form but I note that neither of the original reviewers recommend rejection . The normal procedure at this stage would though be a polite rejection on the grounds of pressure of space and the apparent requirement for significant new work. I certainly will not recommend this course of action and instead request that you and your co-authors look over these opinions and let me know whether you think it possible to deal adequately with them. I would be happy to consider a revision . One referee has indicated their willingness to review a re-submission though one has indicated that they are not prepared to look at the paper again. Therefore I would have to go to a third reviewer but I would be happy to approach one from among a number you might care to suggest. I would have to provide this new reviewer with the original manuscript and referees' comments also. I believe the reviewers are trying to be constructive but it seems they share doubts about the way in which you are presenting the evidence as though the differing forcings acting on the two ice core series are well known whereas the separation of their effects are not easily achieved. In this respect I have to agree. I am also confused about the influence of sea ice or temperature (or distance from the drilling site) . Also are you advocating the use of one set of core data only in future wider studies? Can the data be interpreted as an independent winter temperature record in this regard? This would be very valueable. I would ask that you consider the reviews and let me know how you wish to progress matters. I realise that this will be a disappointing reply , not helped by the unacceptable delay in receiving it . This was my fault alone and I am very sorry for it. Keith P.S. As I am typing this I was just wondering also about the similarity of the two ice core series over the 19 - 20th centuries . Given the differences in them prior to this period , does this fact represent a signal of anthropogenic warming or some such of itself? Perhaps the two series could be differenced to identify some local as opposed to far off climate signal . Probably nonsense but thought i would mention it anyway. Here are the reviews ------------------------------------------------------------------------------------------- ------------------------------------------------------- REVIEWER 1 ------------------------------------------------------------------------------------------- -------------------------------------------------------- CLIMATE AND SEA VARIABILITY AROUND SVALBARD ISAKSSON ET AL. This paper addresses the interpretation of isotope data from ice cores from Svalbard. This is an important study as more recent evidence suggests that the Arctic marginal seas in this sector are an important indicator of climate change and an extended record of climate variability for the region would be very useful. This paper considers some of the issues that have led to the past neglect of such proxy records. Unfortunately, I do not consider this paper warrants publication in its current state. Some basic statistical lapses limit the confidence that can be placed on the correlation results and the paper contains a considerable amount of speculation. Indeed, there is not much fresh analysis in the paper and much recounting of the results of previously published work coupled with unsupported speculation. The authors may well be correct in the views they take but without harder evidence few others will be convinced. Having said that, I do believe that a major revision backed by additional research could result in a publishable paper and have, therefore, made a number of constructive comments to this end. The scale of the proposed revisions mean that the paper would need to be re-refereed. The suggestions below are indicative of the kind of revision that I regard as necessary and should not be taken as a comprehensive list. General comments 1. At present, bar graphics, this paper only contains one new analysis, the correlation study between the various data series. The authors should carry through this approach by remedying the statistical deficiencies noted below and bringing in additional climate data to support the interpretation. For example, there is a classic case of post hoc rationalisation on page 8 where a lack of correlation is explained away. Further work is needed to test this explanation. Without this, the authors are building castles on sand. 2. There is a lengthy paper on the links between the Iceland sea ice record and climate parameters which addresses many of the points raised on page 7 (Kelly, P.M., Goodess, C.M. and Cherry, B.S.G. (1987) Journal of Geophysical Research, 92 (C10), 10835-10843.) Many other references on this subject exist in the climatic or glaciological literature and examination would most likely resolve some of the unresolved issues here. The interactions between sea ice, air temperature and ocean temperature are well known, despite the authors claim otherwise. 3. The structure of the paper needs attention as there is a tendency to return to points already made or issues already discussed: cf. final paragraph on page 8 and following would be better placed earlier when links between the core data and other variables are considered. Again, this discussion is very speculative. Statistical methods 1. The use of running means is not advisable due to phase distortions that can occur at the interannual level (WMO Technical Note 79). All filtering should use binomial filters or the equivalent. 2. Figure 3 alone does not support the interpretation that the data are surprisingly similar both in trends and amplitude. And this claim is in fact contradicted on page 6, final paragraph. Perhaps there is similarity with the eye of a believer but the only real match is the warmth of the 20^th century compared with the previous period. But even this differs in character between the two cores with a step change in one and a more gradual trend in the other. This kind of statement must be supported by statistics. Similarly, I dont see the cold period around 1780 in the upper record. There are two short warmish period before and after the period stated but that is not the same thing! The discussion on pages 5-7 is full of similar claims with no statistical support. For example, the difference in the nature of the recent trends is noted on page 6 but then explained away by a speculative mechanism which is proposed but not tested. This is post hoc rationalisation and must be removed. 3. As far as I can tell the significance testing of the correlation coefficients has not made allowance for autocorrelation in the smoothed series. This is a serious error as autocorrelation can seriously inflate such statistics. Moreover, the possibility that the one significant result has occurred by chance must be discussed. ------------------------------------------------------------------------------------------- -------------------------------------------------------- REVIEWER 2 ------------------------------------------------------------------------------------------- -------------------------------------------------------- This paper reveals much about the difficulty of interpreting oxygen isotope data in the simplistic way that many non ice specialists would wish. It demonstrates how a complicated mix of environmental variables affect the detailed numbers that are extracted from a single ice core and it clearly shows the pitfalls that the likes of Mann and his co-workers and imitators will experience in using these types of records to present global temperature series. For this reason I would like to see this paper published, but I am unhappy with the way the authors present this evidence as though through rose-colored glasses. They conclude that their study shows that Svalbard ice cores provide important information on both local and regional climate variability in the Arctic, despite their relatively low altitude and periodic melt. My interpretation of what they show would be that they show how it is not possible, on the basis of a single core, to know how to interpret changing isotopic values separately in terms of changing temperature and precipitation seasonality, surface ocean circulation (through its effect on sea ice), atmospheric circulation, and variable, very local conditions, such as wind. The authors indulge in much hand-waving to explain away differences in the two ice core records, but make little attempt to test the theories using instrumental data. There are temperature records going back as far as the eighteenth century and mean sea level pressure maps going to the late nineteenth century. Surely these can be used to explore whether different parts of Svalbard (and the different isotope records) relate to temperatures and different circulation characteristics, say before and after the 1920s or during the twentieth century? Similarly, there is a lack of quantification when series are being compared. It does not inform us much to cite evidence of severe ice in the Baltic between 1880 and 1896 or cold in Uppsala in 1862-1871 and 1825-1884 that seems to coincide with cold at Lomonasovfonna between 1870 and 1890, if the longer term consistency or lack of consistency is not quantified. The reference to "changing" atmospheric conditions between 1880-1910, implied to be a cause of one of the major differences in the two ice core records, only confuses the reader, if this is not developed further by looking to see how atmospheric circulation across Svalbard changed. One page 7, it is stated that there is "correspondence" between the sea-ice extent record and the blocked Austonna record. No figures are given and it seems that only the post 1850 trends coincide. Correlations are given for smoothed isotope records and various annual temperature records, but the only one that appears significant is for Jan Mayen. The significance for this in Table 2 is almost certainly over-stated because no allowance has been made for coincident autocorrelation in the records, and no mention is made of the likely dependence of this result on trend. Not enough quantitative evidence is included in the comparisons or discussions and the justification for comparing individual summer, winter, or annual series should be explicit. This also goes for the comparisons with proxy data. I would have liked to see much more on the comparison with the NAO (see work of Lisa Barlow and Jim White) and the removal of the unconvincing discussion and vague reference to expected ENSO associations. There are long reconstructions of NAO by several authors, some based on accumulation or isotope data in Greenland. These could be compared over hundreds of years with these isotope data. The paleodata comparisons are sketchy and do not seem well organised or systematic; again the discussion or conclusions are vague. What is needed here is some insight into how the Svalbard data would be expected to agree or disagree with the other records and a clearer discussion of the extent to which this is true or not. What is the logic for comparing summer-responsive tree rings and why pick out a northern Greenland ice core and not others? There likely are more records available from the Russian Arctic (ice core in Novia Zemlyia and other tree records). The comparison with glacier data seems inappropriate and where is the comparison with earlier Svalbard ice core records (even at lower resolution)? This section would benefit from some plots to show the various series as well as more real correlations. Other points I would mention briefly, and that the authors should discuss, are: different resolution of data through time - possible attenuation of seasonal or annual or decadal records; clear statement of dating uncertainties in their records; discussion of identifying specific seasonal records; consider model based evidence of sea ice controlling factors - there have been numerous studies (Walsh or Hibler?). I would like to see this evidence published, but it needs to be presented in a clearer, more considered way and the similarities and differences between the records on Svalbard and further off must be quantified better and explained more logically. I recommend this be reconsidered after major modification. I do not wish to review it if it is. ------------------------------------------------------------------------------------------- --------------------------------------------------------- At 11:27 AM 6/24/03 +0200, you wrote: Dear Keith, It is now a whole year (!) since I submitted the paper "Climate and sea ice variability around Svalbard-inferences from two ice core d18O records" and I still have not got any response about it. Could you please tell me where things stand now? Regards, Elisabeth Isaksson *********************************************** Dr. Elisabeth Isaksson Norwegian Polar Institute The Polar Environmental Centre N-9296 Tromsų Norway Ph. +47-77 75 05 15 Fax. +47-77 75 05 01 e-mail:elli@npolar.no [1]http://www.npolar.no/ [2]http://www.miljo.no ******************************************** -----Original Message----- From: Keith Briffa [[3]mailto:k.briffa@uea.ac.uk] Sent: 2. mai 2003 10:00 To: Isaksson, Elisabeth Subject: RE: Regarding paper submitted to The Holocene Elisabeth I am currently waiting on a new reviewer , having given up on one earlier one (after some nagging I caused some offence !) . Another review I had was very cursory and ambivalent (and so not much use to me ). I am promised another by the time I return from two short trips next week and so will forward a more detailed response then. Sorry about the delay, partly caused by some medical problems and I am now trying to work through a large backlog. Keith At 11:22 AM 4/23/03 +0200, you wrote: >Dear Keith, >Do you have any news on our submitted paper "Climate and sea ice >variability around Svalbard-inferences from two ice core d18O records"? I >would appreciate to hear were it is as at this point. > >Regards, >Elisabeth Isaksson > >*********************************************** >Dr. Elisabeth Isaksson >Norwegian Polar Institute >The Polar Environmental Centre >N-9296 Tromsų >Norway > >Ph. +47-77 75 05 15 >Fax. +47-77 75 05 01 >e-mail:elli@npolar.no >[4]http://www.npolar.no/ >[5]http://www.miljo.no >******************************************** > > > >-----Original Message----- >From: Keith Briffa [[6]mailto:k.briffa@uea.ac.uk] >Sent: 10. desember 2002 12:33 >To: Isaksson, Elisabeth >Subject: Re: Regarding paper submitted to The Holocene > > >Elizabeth >the problem has been referees. I sent it out to two , one of which sent >back a very cursory "seems ok to me " response by email and the other who >consistently has not responded to requests for an update ( though >admittedly left too long before pushing). The paper was therefore only >recently sent to another two referees who were asked to respond quickly . I >will hassle these more strongly. It is very likely , unless some real >problem is found, that this will be published - but I need at least one >positive response from one of the new referees. I am sorry for this delay , >but I have to say that it is getting more difficult to get referees to >respond ( increasing workloads). I will let you know as soon as I hear >more. Please contact me again in the new year if you don't hear. The >journal is also receiving many papers now and we are having to become much >harsher in selecting papers so as to try and keep a balance in >area/techniques/proxies etc. We are keen to publish papers in your field so >I ask you to patient for a little longer and we will try to make up some of >the delay in the next stage. Thanks > >Keith > >At 10:34 AM 11/27/02 +0100, you wrote: > >Dear Dr Briffa, > >I submitted the paper "Climate and sea ice variability around Svalbard- > >inferences from two ice core d18O records" to The Holocene in the end of > >June and I wonder where in the process it is now. Thanks in advance! > > > >Regards, > >Elisabeth Isaksson > > > > > > > >*********************************************** > >Dr. Elisabeth Isaksson > >Norwegian Polar Institute > >The Polar Environmental Centre > >N-9296 Tromsų > >Norway > > > >Ph. +47-77 75 05 15 > >Fax. +47-77 75 05 01 > >e-mail:elli@npolar.no > >[7]http://www.npolar.no/ > >[8]http://www.miljo.no > >******************************************** > >-- >Professor Keith Briffa, >Climatic Research Unit >University of East Anglia >Norwich, NR4 7TJ, U.K. > >Phone: +44-1603-593909 >Fax: +44-1603-507784 > >[9]http://www.cru.uea.ac.uk/cru/people/briffa/ -- Professor Keith Briffa, Climatic Research Unit University of East Anglia Norwich, NR4 7TJ, U.K. Phone: +44-1603-593909 Fax: +44-1603-507784 [10]http://www.cru.uea.ac.uk/cru/people/briffa/ -- Professor Keith Briffa, Climatic Research Unit University of East Anglia Norwich, NR4 7TJ, U.K. Phone: +44-1603-593909 Fax: +44-1603-507784 [11]http://www.cru.uea.ac.uk/cru/people/briffa[12]/