Chapman’s “17 Reviews” gave me a chance to look again at the Pedersen, Pedersen and van den Berg studies, along with looking for the first time at the Shepherd study. Until recently these were the only peer-reviewed journal-published studies that looked at the health effects of wind turbines. More recently the Nissenbaum study was finally published and thus becomes the 5th such study. I figured I was on a roll so might as well make it complete.
Every adult resident within 1.5 km of a turbine in 2 different locations (Mars Hill and Vinalhaven, Maine) was offered a questionnaire. 23 out of 33 and 15 out of 32 returned them. Additionally, an equivalent number of residents living from 3.3 to 6.6 km from the nearest turbine also filled in identical questionnaires. These questionnaires focused mainly on sleep differences between the groups, using 2 validated sleep-related questionnaires and 1 validated general health questionnaire, along with some additional questions.
The basic results were presented in Table 3. To make it easier to read I’ve separated the numbers into two charts. The PSQI and ESS are sleep indicators, with higher numbers indicating worse, with the normal dividing lines between “good” and “bad” of 5 and 10 respectively. The SF36 MCS is a mental component score, with higher being better, while the SF36 PCS is a physical component score. The “P-value” column gives the odds that these differences are due to chance. Typically, a P-value of less than 0.05 (there’s less than a 5% chance) is considered statistically significant, and smaller numbers indicate a more definitive difference.
The first chart compares the turbine group with the control group as a whole.Looking at these numbers there’s not much doubt that the turbine group’s sleep was significantly worse than the control group’s. Their overall mental health was also worse, while their physical health didn’t differ.
These findings are consistent with all of the other 4 studies. They all found annoyance and indications of sleep disturbance and Nissembaum confirms that in the most rigorous study to date.
Not content with demonstrating significant sleep (and thus health) differences between the groups, he then splits each of the groups in half and comes up with the following results.This is even more remarkable. The effects of the wind turbines seem to extend out even into the control group, located over 3.3 km away from the nearest turbine. Nissenbaum creates 3 best-fit graphs (Figures 1, 2 and 3) to show this pictorially, creating dose-response curves, although he stops short of providing either a safe setback distance or an estimate of the portion of a population that would suffer at any particular distance.
While we can certainly quibble about certain aspects of this study (as we could about any study) it is hard to dismiss its overall finding – that being a wind turbine neighbor is bad for your sleep, among other things.
This study gathered enough attention that CanWEA and AWEA figured they had to respond to it. They paid some familiar names, Knopper and Ollsen, to rebut it, which they did under their company’s name: Intrinsik. They previously produced 1 of the 17 reviews that Chapman thought so highly of, so giving them the business was a safe bet – they’d already dismissed Nissembaum’s preliminary findings.
Their main points are:
- Sound Levels. Nissenbaum’s conclusion was the sound levels were responsible, but he “…gave such little consideration to collection of actual sound data measurements…”. Rather, he used distance.
- Sleep Scores. Both the turbine group and the control group were on average “poor sleepers” (their PSQI score was above 5).
- Regression Lines. Nissenbaum seems to have confused the meaning of their regressions’ P-value.
- More Sleep Scores. Neither the turbine group and the control on average suffered “clinically relevant” sleepiness in the day time (their ESS score was above 10).
- MCS differences are not related to noise itself. Since Nissembaum’s categories were distance-based, he hasn’t proved that noise was the culprit.
- Visual and attitude are the problem, not noise.
At the risk of getting even more tedious, let’s take these in turn.
1. While Nissenbaum did relate measured noise levels to distances at both locations, it is true that he never measured noise levels at the respondents’ homes. But then, Pedersen and van den Berg didn’t either, and Knopper accepts their results. Complaining about it now is at best inconsistent. The basic problem, for all researchers, is the cost. Nissenbaum et al apparently did this study on their own nickle, unlike Knopper and Ollsen who were paid by the industry for their response.
2. Both groups were on average poor sleepers. SO WHAT? The turbine poor sleepers were still much worse off than the control poor sleepers. And the fact that both groups were poor sleepers may go a long way in explaining why there wasn’t a significant difference in the number of poor sleepers. But there is another explanation, one that is entirely consistent with the data – maybe the sleepers in the control group, who were initially presumed to be entirely free of the turbines’ influence, really weren’t. Australian researchers (among others) have been receiving complaints beyond the control group’s distances.
3. I suspect Nissenbaum et al, especially Aramini, know exactly what the significance of a regression’s P-value is. It shows that the slopes they show in figures 1, 2 and 3 have almost no chance of being zero. And recall that a slope of zero is what the industry is claiming – that wind turbines do not affect health, of which good sleep is a non-disputed part.
4. Neither group was, on average, excessively sleepy during the daytime. As with #2 above, SO WHAT? The turbine group did suffer more daytime sleepiness (with all the risks that entails), and Knopper wants to quibble about how the average didn’t shift into clinical relevance territory? This is grasping at straws.
5. The entire discussion Knopper introduces about using distance vs. using noise is a perfect example of knowingly sowing confusion in an effort to continue the “he said-she said” distraction. It is unarguable that noise reduces with distance. If something else reduces with distance that could even plausibly cause the groups to differ so much, what would Knopper suggest it is? Grasping at straws, he might reply something about visual angle. And what would visual angle have to do with sleep, which is almost entirely a night-time activity? Nope, sorry.
And even if Nissenbaum had calculated noise levels it wouldn’t have, from any practical aspect, made the study any more accurate. There’s a long history of those calculations being significantly in error. That is especially true for turbines in configurations like those at Mars Hill and Vinalhaven. So why bother? Distance is unarguable. If Nissenbaum was seeking to determine i.e. acceptable noise levels then this type of criticism might be valid. But he wasn’t.
6. Visual effects and attitude are another of the ways Knopper seeks to confuse the issue, even stating “While this may be true, visual cue and attitude by themselves have been shown to be stronger drivers of psychological responses than a wind-turbine specific variable like sound itself (e.g., Pedersen 2004).” Really? Pedersen said that? Read the complete unedited conclusion to Pedersen’s 2004 study and judge for yourself.
A significant dose–response relationship between calculated A-weighted SPL from wind turbines and noise annoyance was found. The prevalence of noise annoyance was higher than what was expected from the calculated dose. It is possible that the presence of intrusive sound characteristics and/or attitudinal visual impacts have an influence on noise annoyance. Further studies are needed, including a larger number of respondents especially at the upper end of the dose curve, before firm conclusions could be drawn. To explore attitude with regard to visual impact, some of these studies should be performed in areas of different topography where the turbines are less visible. There is also a need to further explore the influence of individual and contextual parameters.
So Knopper translates “influence on noise annoyance” to “stronger drivers” than “sound itself”. Orwell would be proud.
Nissenbaum is almost unarguably the most rigorous study yet undertaken regarding wind turbines and health effects. It used several validated surveys, multiple locations, multiple groups, and measured sleep effects – which are an undisputed factor for health. Of course it can be quibbled with – any study can. But on balance, its conclusions are rock-solid: wind turbines make for unhealthy neighbors.