Establishing the acute physiological and sleep disruption characteristics of wind farm versus road traffic noise disturbances in sleep: a randomized controlled trial protocol 

Establishing the acute physiological and sleep disruption characteristics of wind farm versus road traffic noise disturbances in sleep: a randomized controlled trial protocol

Introduction

Wind farms continue to gain prominence as a major source of sustainable energy generation in over 90 countries. However, onshore wind farms are also a source of public concern and debate regarding potential adverse effects of wind farm noise (WFN) emissions on nearby communities. Sufficiently loud noise from any source has the potential to disturb sleep, health and well-being through direct sleep disruption effects, and through inherent or acquired noise sensitivity and annoyance effects that could promote insomnia. Predominantly through their large size and complex aero-acoustic effects including blade passage past the turbine tower, wind turbine noise emissions are dominated by low frequencies including infrasound, which is defined as low frequency noise < 20 Hz; below the conventionally accepted lower frequency limit of average human hearing from 20 Hz and 20 kHz. Although individuals with above average low frequency hearing acuity can potentially hear WFN infrasound within a few hundred meters of a wind turbine, WFN infrasound is unlikely to be audible at longer-range distances. Given the importance of hearing for sensing sound, prominent audible WFN features appear much more likely to negatively impact on sleep compared to inaudible infrasound.

The World Health Organization (WHO) concluded that there is overwhelming evidence that exposure to environmental noise has adverse effects on population health. Per annum in Europe, noise pollution is estimated to contribute 1.0–1.6 million healthy life years lost, including 900 000 healthy life years lost through sleep disturbance. Accordingly, to help protect sleep, WHO environmental noise guidelines and allowable noise limits across many jurisdictions suggest that outdoor sound pressure levels (SPLs) at night produced by traffic and other sources, including wind farms, should not exceed an A-weighted equivalent level (LAeq) of 40 dB. Provided the building structure is substantial and windows are closed, outside noise can be attenuated by around 10–15 dB, resulting in indoor noise levels of around 30 dB(A). These recommendations reflect evidence accumulated primarily from road traffic, railway and aircraft effects based on noise level measurements A-weighted to average human hearing curves on the logarithmic dB scale. However, WFN has substantially different acoustic characteristics compared to road, rail, and air traffic noise, including much more predominant low frequency and time-varying noise features which could have different impacts on sleep. Thus, A-weighted noise criteria derived from traffic noise may not be entirely appropriate for WFN, particularly given substantial inter-individual variability in low frequency hearing acuity and more compressed equal-loudness contours at lower frequencies that could potentially render low frequency noise inaudible to some and yet clearly audible and sleep disruptive to others.

Modern WFN is dominated by low frequencies (<200 Hz) at noise propagation distances beyond a few hundred meters most relevant to typical human exposure in neighboring households. Road, rail, and air traffic noise also contain some low frequencies, but are predominantly mid-high frequency (>200 Hz) noise that is substantially more attenuated over distance and by intervening objects compared to low frequency dominated noise. WFN, and particularly prominent aero-acoustic effects from blade-rotation, including dynamically changing lift and potentially stall, contribute to time-varying amplitude modulation (AM). Tonal AM, most likely mechanical in origin, has also been observed in the context of wind farm operation, and the associated low-frequency components have been measured at audible SPLs up to several kilometers from the source. Noise with AM is consistently rated as more annoying compared to noises of equivalent A-weighted SPL without AM. At close distances (< 2 km from the nearest wind turbine), WFN with AM is often described as “swish,” but at greater distances, “rumbling” or “thumping” is a more common description.

WFN characteristics are influenced by many factors such as the number, type and size of turbines; distance from the source; background noise levels which are typically low in rural areas (particularly at night); local topology; wind speed and direction; atmospheric temperature profile (including inversions); turbulence conditions; and the nature and characteristics of intervening structures that impact audibility and perception. Consequently, WFN is highly variable and can be sporadic or persistent, which potentially makes habituation to WFN more difficult compared to more consistent and predictable (e.g. heavy road traffic) or transient (e.g. more sporadic traffic pass-bys) noise sources. Furthermore, unlike traffic noise, which is typically reduced at night when traffic volumes are typically lower, more stable environmental conditions are often favorable to ongoing WFN propagation at night when other background noise and wind-speeds in neighboring lower lying residential locations are usually lowest. Thus, when present, prominent WFN at night has significant potential to disturb sleep of neighboring residents.

Excessive noise from neighborhood, traffic and industrial noise sources is one of the most common public complaints, and an established cause of annoyance, stress, raised blood pressure, sleep disturbance, related health impacts and pharmaceutical use. In terms of sleep macrostructure, nocturnal noise causes sleep fragmentation, resulting in shallower sleep from increased arousals and redistribution of sleep architecture (i.e. increased light sleep (wake and stage 1 sleep) and decreased slow wave sleep and REM sleep. Auditory arousal thresholds measured during N2, deep and REM sleep, in good sleepers and those with sleep onset insomnia do not appear to systematically differ, despite insomniacs reporting being “light sleepers”. This suggests that noise-related sleep disturbance may not be substantially different between individuals, even in those with insomnia. Nevertheless, as has previously been supported by several studies, a vulnerability to stress-induced sleep disturbance clearly has the potential to lead to transient and potentially more chronic conditioned insomnia.

On a microstructural level, cortical electroencephalographic (EEG) response probability and magnitude also depend largely on the type and intensity of a noise stimulus and on sleep depth. In addition to cortical EEG responses, traffic noise is known to trigger autonomic or “sub-cortical” reflex responses in sleep. These reflexes rapidly augment cardiovascular, respiratory and metabolic activity in preparation for behavioral “flight or fight” responses. This includes a blood pressure surge through increased heart rate, and a particularly prominent skin vasoconstriction response readily discernible as attenuation in finger pulse oximeter waveform amplitude. Sensory disturbances producing no visually discernible EEG changes can still produce a clear reflex cardiovascular response. There is also some evidence to support that frequent noise-induced cardiovascular responses without more frequent EEG arousals negatively impacts next-day sleepiness and mood. Thus, it is important to consider that noise-induced sleep disturbances may have important effects on daytime functions.

Although extensive literature regarding environmental noise emitted from road, rail and air exists, important knowledge gaps remain regarding wind farm noise effects on sleep. Data from well-designed studies using objective measures of sleep under carefully controlled noise conditions are particularly scarce. In some of the most comprehensive studies to date, Persson Waye et al. found that amplitude-modulated continuous WFN exposure produced small but statistically significant reductions in self-assessed sleep quality and some aspects of EEG evaluated sleep. Using sleep actigraphy data from around 250 individuals studied over multiple nights, Michaud et al also found some evidence to support small increases in overnight movement time in response to changes in wind farm sound pressure levels. Thus, transient noise events and time-varying features of WFN may be more sleep disruptive than continuous noise.

Adverse effects of WFN exposure have also been attributed to infrasound, but without supporting evidence. A previous study found no discernible EEG changes with overnight exposure to 10 Hz infrasound at 105 dB. A more recent study, using 72 h of simulated wind farm infrasound exposure, also found no evidence to support any discernible effects on sleep. However, no previous study has directly evaluated if wind farm noise, including infrasound, and prominent audible amplitude modulated components, is potentially more sleep disruptive compared to road traffic noise when replayed under carefully controlled laboratory conditions needed to avoid a wide-range of potential confounders.

This paper outlines a study protocol designed to compare the sleep disruption effects of wind farm versus road traffic noise on established sleep, and to examine the impact of wind farm noise exposure during wake and/or sleep on conventional overnight measures of sleep time and quality in individuals with and without prior wind farm noise exposure and noise-related sleep complaints.

The effects of environmental noise on sleep are best tested using direct EEG measures of sleep on cortical activity. During sleep, cortical responses to sensory stimuli are markedly diminished. However, brainstem mechanisms continue to process sensory stimuli, with thalamic “gating” of physiological responses according to stimulus salience and intensity and the depth of sleep itself. Thus, the effects of auditory stimuli on sleep depend on the type and intensity of the noise stimulus and on the depth of sleep during which a noise stimulus occurs. Responses to noise can range from no discernible response in the EEG or any other physiological signal through to full awakening (shifts to faster EEG frequencies > 15 s), but can also include increased micro-arousals (3–15 s shifts toward faster EEG frequencies), reflex cardiovascular responses, and K-complexes in the EEG. On the sleep macrostructure level, preexisting stress and extraneous noises can impair sleep initiation and the return to sleep after waking to reduce total sleep time and sleep efficiency (the percentage of the sleep opportunity occupied by sleep). Thus, carefully controlled laboratory studies of EEG and cardiovascular activation responses to noise exposure during sleep allow for robust evaluation of WFN specific effects on sleep with a reduced risk of confounding from a range of potential biases in real-world noise exposure settings.

Aims and hypotheses

This study sought to clarify the effects of WFN on sleep compared to RTN, an already known disruptor to sleep, and quiet background noise (control). The primary study aims were to compare the dose–response effects of WFN versus RTN on:

1. The probability of EEG-defined micro-arousals and awakenings from sleep (shifts to faster EEG frequencies for ≥ 3 and ≥ 15 s, respectively) under each noise condition on a 20-second noise battery night to assess the acute noise effects.
2. The probability of EEG-defined micro-arousals and awakenings from sleep under each noise condition on a 3-minute noise battery night to assess more sustained noise effects.

The hypotheses for the two primary aims were that:

1. EEG arousal responses are more probable with brief 20-second WFN compared to RTN exposures of equivalent A-weighted SPL.
2. EEG arousals responses, including longer periods of wake, are more probable with more prolonged 3-minute WFN compared to RTN exposures of equivalent A-weighted SPL.

The study was also designed to address the following secondary aims to:

1. Examine the role of wake-related noise exposure prior to sleep onset on objective and subjective measures of sleep disruption and next-day mood, anxiety, sleepiness, and daytime performance, by presenting WFN noise only during wake, only during sleep, and continuously during both wake and sleep throughout separate overnight sleep opportunities.
2. Examine the role of habitual noise exposure history and self-reported noise sensitivity on objective and subjective sleep in four pre-existing populations: individuals living near wind turbines including a group with and a group without noise-related complaints, individuals living in urban areas near road traffic, and individuals living in a quiet rural area.
3. Compare the dose–response effects of sound pressure level and noise type on the probability of EEG (K-complexes and quantitative electroencephalography measures) and cardiovascular activation responses (tachy-brady cardias, finger vasoconstriction and pulse arrival time) using established methods.
4. Examine the dose–response effects of sound pressure level and noise type on daytime listening test outcomes of self-reported annoyance and perceived acceptability for sleep.

Gorica Micic, Branko Zajamsek, Bastien Lechat, Kristy Hansen, Hannah Scott, Barbara Toson, Tessa Liebich, Claire Dunbar, Duc Phuc Nguyen, Felix Decup, Andrew Vakulin, Nicole Lovato, Leon Lack, Colin Hansen, Dorothy Bruck, Ching Li Chai-Coetzer, Jeremy Mercer, Con Doolan and Peter Catcheside
Flinders Health and Medical Research Institute: Sleep Health, College of Medicine and Public Health, College of Science and Engineering, and College of Education, Psychology and Social Work, Flinders University, Australia
Neurosleep, Woolcock Institute of Medical Research, University of Sydney, Australia
School of Mechanical Engineering, University of Adelaide, Australia
Institute for Health and Sport, Victoria University, Australia
Department of Respiratory, Sleep Medicine and Ventilation, Southern Adelaide Local Health Network, SA Health, Australia
School of Mechanical and Manufacturing Engineering, University of New South Wales, Australia

SLEEP Advances, Volume 4, Issue 1, 2023, zpad033, doi:10.1093/sleepadvances/zpad033

Download original document: “Establishing the acute physiological and sleep disruption characteristics of wind farm versus road traffic noise disturbances in sleep: a randomized controlled trial protocol”