Quality of research on wind energy published in the Bulletin of Science, Technology and Society

In the past two years, an obscure allegedly “peer reviewed” research journal, the Bulletin of Science, Technology & Society, has published a series of papers by authors who have anti-windfarm track records.

The Bulletin has appeared erratically over the past few years. It was indexed between 1981-1995 by the Web of Science, the international scientific indexing platform which “covers over 10,000 of the highest impact journals worldwide, including Open Access journals and over 110,000 conference proceedings.”

But after 1995 it was dropped from the list of journals being indexed, generally a sign that indexing services regard a journal as having fallen below an acceptable scientific standard. In the 14 years it was indexed, a citation search conduced on 21 September 2012 shows that it published 961 papers that have been cited for a grand total of just 345 times–an average of 0.36 per paper –almost a homeopathic strength citation rate. Web of Science shows it has published only seven papers which have been cited 7 or more times, with the most cited paper in its history having been cited just 15 times. PubMed, the indexing service of the U.S. National Library of Medicine, also does not index the journal.

Nonetheless, anti-windfarm websites have jubilantly described the journal as a “leading scientific peer-reviewed journal” and the issue as “groundbreaking”. In summary, this is a journal which cannot be described as low ranking in scientific research publishing. It is more accurately described as “unranking”.

All the papers in the journal on wind farms are only accessible by pay-per-view, but to give an idea of the quality, here’s an example.

Krogh CME. Industrial wind turbine development and loss of social justice. Bulletin of Science, Technology and Society 2011;31(4):321-333.
This paper contains no “methods” section, so it fails to conform to the most basic requirement of scientific reporting: that it contains details of how the research reported was undertaken. This is a fundamental requirement because without it, readers have no way of assessing the adequacy and rigor of any investigation, and whether any results reported and conclusions drawn are justified or not. Instead, the author–a retired pharmacist who PubMed shows published one paper in 1985–explains that she “began investigating reports of adverse health effects made by individuals living in the environs” of wind turbines in Ontario, Canada for “more than two years”. That’s it. And this is called peer reviewed “research”.

Instead of describing any research, the author has written a paper which mixes up statements somehow apparently made to her by de-identified informants about negative effects of exposure to turbines with similar examples from other parts of the world, from websites and submission to enquiries by wind farm opponents. We are told nothing about the process by which her informants were interviewed, the questions they were asked, how they were selected and whether her “study” was approved by any institutional research ethics committee. This is not a paper that would make first base as an example of serious scientific investigation about wind farms and health. Its findings contain not a single example of any informant reporting anything but adverse effects of exposure to wind farms, when it is widely acknowledged that a large majority of those so exposed report no adverse effects nor complain about the turbines and even like them!

In an attempt to understand the process of peer review that had been followed, in August 2011, I wrote to the editor of the Bulletin, asking the following questions:

1. Were you approached by those participating in [a 2010 anti-wind farm meeting held in Ontario] to publish these papers? Or did the initiative come from you?

2. Did you personally edit this issue or were guest editors used? If so, can you please describe how they were selected?

3. Was there a charge made to the authors to publish their papers together like this?

4. It is plain that all the papers are openly negative about wind farms, which is curious given that there is a large body of research that demonstrates a very different picture. Did you put out a call for submissions or approach researchers working in this area to submit manuscripts?

5. Did you approach any authors who did not have affiliations with the anti-wind farm movement?

6. Were all the papers peer reviewed?

7. Did the authors propose their own reviewers and were these the reviewers used?

8. Can signed or de-indentified copies of these reviews be made available to others on request?

Over several testy e-mail replies, the editor made the following comments:

“A third party mediated between the organizers of the symposium and myself. We are dealing with a very difficult situation in which there is no balanced approach to begin with. Deep pockets have controlled the research agenda and professional people with impeccable credentials did what they did in this case out of there (sic) own pocket. As far as refereeing is concerned, never has any issue been so over-refereed by people with impeccable credentials in anticipation of the kinds of concerns you voice.

“I can assure you that this Bulletin is not a front for any special interest group and that I would not have dreamt of publishing this issue had it not been for the questionable conduct of the windfarm industry and government officials. The issue attempts to create a little bit of balance, and show that there are legitimate other voices coming from people with impeccable credentials who are not funded because of their views.”

No copies of reviews or reviewers’ names were provided. A researcher from the University of Adelaide has subsequently written directly to the authors and received no reviews back. The principle of open peer reviewing is widely discussed in research publishing and while requests by others to see reviews are unusual, refusal to be transparent can only promote suspicion about the process, particularly when the quality of the papers is considered.

In August 2011, the Bulletin published another paper on wind farms, also involving author Krough. The abstract is below.
Wind Turbine Acoustic Investigation: Infrasound and Low-Frequency Noise—A Case Study
1. Stephen E. Ambrose seaa@myfairpoint.net
2. Robert W. Rand
3. Carmen M. E. Krogh

Abstract: Wind turbines produce sound that is capable of disturbing local residents and is reported to cause annoyance, sleep disturbance, and other health-related impacts. An acoustical study was conducted to investigate the presence of infrasonic and low-frequency noise emissions from wind turbines located in Falmouth, Massachusetts, USA. During the study, the investigating acousticians experienced adverse health effects consistent with those reported by some Falmouth residents. The authors conclude that wind turbine acoustic energy was found to be greater than or uniquely distinguishable from the ambient background levels and capable of exceeding human detection thresholds. The authors emphasize the need for epidemiological and laboratory research by health professionals and acousticians concerned with public health and well-being to develop effective and precautionary setback distances for industrial wind turbines that protect residents from wind turbine sound.

I am unable to post the whole paper because this would violate copyright conditions. But below is a critique of the paper by Tom Evans, an acoustician from Resonate Acoustics in South Australia.

Measurement of noise levels below 10 Hz
– Some of the equipment used is not suitable for the purposes for which it appears it may have been used. For example, the SV22 microphone is quoted by the manufacturer as only being suitable for measurements between 10 Hz and 20 kHz, yet it appears they may have corrected the response at frequencies below 10 Hz to obtain low frequency measurements (far from ideal). They do not note which equipment was used to take which measurements, so the effect of these problems is not known.
– They used an integration period of <128 ms which means that they are measuring less than one cycle of frequencies of 8 Hz and lower. Note that ISO 7196 (which defines the G-weighting for measurements of infrasound) states that measurements should have an integration period of 10 seconds, although you can use an integration period of no lower than 1 second where noise levels above 1 Hz are dominant.
– They have used a very coarse octave band method to determine G-weighted noise levels, which does not meet the requirements of ISO 7196. This introduces a potential error of 12 dB in their G-weighted measurements.
– Given the above, there are serious doubts about the accuracy of their measurements below 10 Hz even though many of their conclusions are related to this frequency range (such as amplification of noise levels within the house).

Measurement procedure
– No attempt is made to determine the accuracy of their outdoor measurement setup when measuring low frequency noise and infrasound levels. I would suspect that the results would be heavily influenced by wind across the microphone.
– No serious attempt is made to describe the measurement process within the house – were windows/doors open or closed? Were rooms occupied? What about other sources of the measured levels within the house such as the refrigerator?
– They did not determine the actual sound level output from the wind turbines and the contribution of the turbines to the measured levels at the house. Low frequency/infrasound levels have only been compared for two scenarios – turbines off (low wind speed) and turbines near rated power (much higher wind speed). Would suspect that the difference in wind speed is causing the difference in measured levels, rather than difference in noise output from the turbines. A significant influence of wind noise on the measurements would also explain the questionable result regarding the attenuation of noise with distance from the source. It seems convenient for their findings that only A–weighted noise measurements were taken when the turbine power output was low (150 – 350 kW).

Power generating windmills

– All of the reported complaints from residents appear to be about audible sound–‘fluctuating sound, the endless swish and thumps’. However, the study has not looked at these concerns in any real depth and has focussed on infrasound from the outset, suggesting that this was always the aim.
– They reference a 60 dB(G) OHC detection threshold for infrasound based on Salt’s research and said that measured levels exceeded that (although only marginally indoors). This criterion has no basis for the assessment of infrasound, and even Salt acknowledges that it does not indicate a level that is annoying or harmful to humans. Infrasound levels regularly exceed 60 dB(G) in different environments (see Sonus study)–it cannot be considered an acceptable criterion for assessment.
– We have regularly measured higher G-weighted noise levels than are reported in this study at a number of sites (including residences) located nowhere near wind farms–no adverse responses were noted at these locations.
– There is no assessment of low frequency noise or infrasound against commonly accepted standards for these types of noise–assume this is either a blind spot or has not been reported as the levels are compliant.
– There is no assessment of the low frequency tones against standards for tonality. These tones could potentially be a source of annoyance (if they are caused by the wind turbines, which has not clearly been identified), but have been largely ignored other than to group them in with “infrasound”.

Simon Chapman, Prof. of Pub. Health, U. of Sydney, http://www.awea.org/blog/