subject: Receipt of review for Nature 2008-04-04235
Dear Professor Jones
This email is to acknowledge receipt of your review for the manuscript by Mr Stine and co-authors, entitled "Changes in the Phase of the Annual Cycle of Surface Temperature". Thank you for your help in this matter.
A copy of your review is attached below for your reference.
For Dr Karl Ziemelis
Remarks to the Editor:
The GCM aspct of the paper is way out of date. There are many newer models which are not flux corrected. They also need to emphasize some aspects of the work, so need to have explained the rest in a specialist paper.
Remarks to the Author:
Review of Stine et al. on changes in the annual cycle of surface temperatures
There are a lot of interesting and thought-provoking ideas in this paper, but there seems far too much detail for this to be considered as a Nature paper. After an initial read through, my first thoughts were that this would seem much more appropriate in an extensive paper in say J. Climate or JGR. There are many useful aspects of basic climatology that could be elaborated upon given slightly more space. After finishing this review and more consideration, I am still of the view that for this to be a useful contribution it needs more space for the issues to be addressed. With all the detail in one paper, it would then be possible to write a shorter paper (for Nature) where the more interesting parts could be discussed. So my recommendation is to reject the paper in this form.
I noted a number of thoughts during the review:
1. I can't see the relevance of the first sentence as two completely different timescales are under consideration. The annual cycle is large locally, but less so for global average. The glacial-interglacial cycle is for the global average: it too would be much larger locally.
2. A number of papers look at seasonal temperature changes. The annual is just a convenient way of distilling everything down to one time series plot. The global datasets are gridded versions of the available data, so seasonal and spatial patterns can be shown and generally are (see the latest IPCC Report, in Trenberth et al., 2007).
3. Thomson (1995) found a number of interesting features of global temperature and also the CET series. One was a problem with the CET data before the calendar change in the UK in 1752. The CET data could be useful to look at the natural variability of phase changes in a long record, albeit for one location on the land. There are other long-term records for a few sites in Europe and North America that go back to the late-18th and early-19th centuries respectively. A long daily series for Kansas will be published soon - see the University of Arkansas web site (tree-ring group).
4. The climate model used by Wallace and Osborn (2002) was HadCM2, which did incorporate flux adjustments. This model is fairly old now. A useful addition to the paper would be to look at the full range of the AR4 models, focussing on the many that now do not need flux adjustments. All the models used in Ch 10 of AR4 did not use flux adjustments, so it is important to see if these much newer models do things differently. My guess would be that they would, but that there would be quite a difference between them. Detailed study of their changes in phase and amplitude might prove a useful measure of reliability assessment. There is also another set of climate model runs that could be looked at as well - these are those given the time history of observed SST changes over the 20th century. A comparison of these - with and without additional external forcing (from the sun, volcanoes, greenhouse gases and sulphate aerosols) would likely elucidate some useful information. Do the chan
in the phase of the annual cycle come from the changes in SST, or do they require the changes in external forcing?
5. The authors point out the obvious aspects of earlier springs with a warming of the climate, and this needs to be clarified with the definition based on the 1/yr harmonic used here. Could be easily illustrated with the CET series.
6. The HadCRUT3 dataset is a combination of land temperature anomalies with SST anomalies over the ocean. The absolute climatology is of air temperatures over the land and marine air temperatures over the ocean. There is no need to use a land mask, as the two components can be used separately (CRUTEM3 and HadSST2). So there is no need for a different ocean mask from ref 27 to be used. There may also be an issue with these gridded data due to fewer constituent station series over land in earlier decades. There has been an attempt to allow for this over land with the CRUTEM3v version.
7. It is nice to see the effective spatial degrees of freedom being considered. It would be useful in a longer paper to compare the results from ref 13 with those detailed in Jones et al. (1997).