Why I Resigned from the Editorial Board of Climate over its Akasofu Publication
Posted on 4 September 2013 by Chris Brierley
Guest post from Dr Chris Brierley, University College London
My name is Dr Chris Brierley and I lead the Climate Change MSc program at University College London. I have recently resigned from the editorial board of the journal Climate launched by the publisher MDPI this summer. This action was prompted by the inclusion of a paper entitled “On the present halting of global warming” in their first issue. I do not believe that the paper is of sufficient quality for publication and have decided that I do not want to be associated with a journal with such lapses of judgment.
The scientific method is self-correcting - results influence subsequent thinking that is then tested by experiments. If a theory does not successfully explain the facts, it is either improved upon or it is ignored and a new theory devised. The same process works on the scientific literature, with erroneous papers being either ignored or proved wrong. It can take years for this process to fade papers away though, in which time their mistakes may permeate through to the general public. As this paper also states that humans have little influence on climate change (which is of interest beyond the scientific community), I would like to openly state why I feel this paper should not have reached publication in its present form. I, personally, do not find the paper of sufficient rigour to be considered as scientifically valid. I would like to stress that I level these criticisms at this particular paper - not at the hypothesis being proposed, nor at the author (Prof. Akasofu has deserved reputation as a leading researcher into the aurora).
To summarise the paper, it concludes that the recent hiatus in global warming can be ascribed to natural variability (which it calls the internal “multi-decadal oscillation”) masking the upward trend. There is merit in this suggestion – for example Katsman & van Oldenborgh (2011) compute a chance of 25-30% for natural variability to mask the upper-ocean warming for an 8 year period up to 2020 (note the correction they published about this though). Despite my suspecting the paper’s conclusion about natural variability contributing to the hiatus to be true, I do not feel the evidence provided in the paper comes close to justifying it.
The paper also states that the upward trend is solely a recovery from the Little Ice Age, rather than having a strong anthropogenic component. This assertion was not tested in the paper and would have been falsified if it had been. A response shall be published in the same journal doing just that. The reason I resigned from the editorial board was not this false assertion, but rather the poor application of the scientific method in the paper (i.e. not testing the proposed hypotheses).
There are no agreed criteria to judge the worth of peer-reviewed papers. Instead I will use a lower standard against which to judge the paper - the marking guidelines for our MSc dissertations at University College London. These state that an “outstanding dissertation (90-100%) should approach professional standards of research and could be publishable virtually without revision as a journal paper”. It is so rare that any work falls into that category that the guidelines only explain the criteria for a distinction (a grade above 70%, with my emphasis):
“Originality displayed in construction of main research aims and questions and interpretation of evidence presented. Impressive critical ability and deep understanding of subject area. Substantial original fieldwork or other independent research. High ability in the application of appropriate research techniques and critical commentary on research design and methodology. Extensive reading and thorough understanding of literature consulted. Logical, coherent structure and clear, cogent and persuasive writing style. Excellent presentation with impeccable referencing and bibliography. No or only very minor errors of spelling, punctuation or grammar.”
The publisher deals with the presentation, bibliography and type-setting. MDPI is a professional publisher, so these criteria should and are being met. The papers also has a good writing style, so onto the other factors…
- Extensive reading: The topic of the paper (the recent hiatus in global temperatures) has attracted a lot of attention both from scientists and the media (e.g. Guemas et al., Watanabe et al., and a nice summary in Nature News). Skeptical Science has a fair few posts on it (e.g. here, here and here). Some authors even apportion some of it to natural variability (such as Katsman & van Oldenborgh, 2011), which could have been used to support the discussion of an internal oscillation. This paper only mentions two articles discussing the recent hiatus – and one of those states the heat is just entering the ocean instead. There is no set number of references for a good paper – you just to need to cite the relevant prior work. However, there is quite a lot of relevant work, such as above, that has been missed in this paper, so it does not count as extensive reading. (Although, Watanabe et al. was published after this paper, so is missing legitimately).
- Critical ability: This criterion relates to being able to understand the potential errors introduced by your own and others’ techniques and analyses. In fact, the MSc students must include an Auto-critique in their dissertation specifically to address the applicability of their research approach. It should have been easy to spot those potential errors, as Skeptical Science had already done it from an earlier article over a year before the paper was submitted. However, the paper still contains neither an acknowledgement of these issues, nor any response to them.
- Research techniques: There is little evidence that any of the statements drawn from the time series have actually been statistically tested. This is a very basic research technique, as the mind can ascribe connections where there are none (and the reverse). For example, it is stated that the multi-decadal oscillation in the global time series “is closely related” to the Pacific Decadal Oscillation index. The conventional research techniques to show this would be to look at the statistical significance of the correlation coefficient, to perform regression or to refer to more detailed studies. Rather than just by eyeballing them, as appears to have been done here. This paper therefore does not pass the criteria of “high ability in the application of appropriate research techniques”.
- Substantial Work: In fact, the paper does not give the impression that any of the actual data have been analysed. A great tool for doing simple climate analysis is Climate Explorer, run by KNMI in Holland. For example, I was able to calculate the correlation coefficient between the 10 year smoothed, detrended PDO and 10 year smoothed, detrended temperatures in under two minutes (a screen capture video is here - I didn’t find a significant relationship at the p<0.05 level, but this may not be precisely the comparison you would want perform as detrending the PDO may be inappropriate). If the paper has neglected two minutes of effort to prove one of its fundamental conclusions, you have to wonder whether the analysis that actually has gone into it counts as substantial. This feeling is increased when one notices that every single figure has been either taken from a website or made by others (a Dr Kramm is acknowledged for his efforts in “improving … and providing” two of the figures).
- Originality: Assessing whether a potential paper brings something new to science is a tricky task. It could bring new data, model, method, analysis, ideas or even old data to a new problem. This paper does none of these things. It does not contain new data - even in graphic form as the figures have come from elsewhere. It does not contain a new model - as that was presented in an earlier paper in Natural Science. There is no new method and the analysis appears sadly lacking (see previous two points). The idea of the hiatus being caused by natural variability is not new. It has appeared several times before, such as in Katsman and van Oldenborgh (2011) mentioned above. This paper’s similarity with the author’s previous work means that it has not addressed a new problem either.
One must therefore conclude that this paper does not pass five of the criteria for a distinction at the MSc level, and is therefore logically not of sufficient quality to deserve publication. It appears that neither the three reviewers nor the handling editor of Climate reached the same conclusion. This has made me realize that the journal does not hold the standards that I feel should be strived for in science, leading to my resignation from the editorial board.