Bulletin of Science, Technology & Society
The online version of this article can be found at:
2012 32: 128 originally published online 17 August 2012Bulletin of Science Technology & Society
Stephen E. Ambrose, Robert W. Rand and Carmen M. E. Krogh
A Case Study−−Wind Turbine Acoustic Investigation: Infrasound and Low-Frequency Noise
On behalf of:
National Association for Science, Technology & Society
can be found at:Bulletin of Science, Technology & SocietyAdditional services and information for
What is This?
- Aug 17, 2012OnlineFirst Version of Record
- Sep 10, 2012Version of Record >>
at Chongqing University on October 11, 2013bst.sagepub.comDownloaded from at Chongqing University on October 11, 2013bst.sagepub.comDownloaded from at Chongqing University on October 11, 2013bst.sagepub.comDownloaded from at Chongqing University on October 11, 2013bst.sagepub.comDownloaded from at Chongqing University on October 11, 2013bst.sagepub.comDownloaded from at Chongqing University on October 11, 2013bst.sagepub.comDownloaded from at Chongqing University on October 11, 2013bst.sagepub.comDownloaded from at Chongqing University on October 11, 2013bst.sagepub.comDownloaded from at Chongqing University on October 11, 2013bst.sagepub.comDownloaded from at Chongqing University on October 11, 2013bst.sagepub.comDownloaded from at Chongqing University on October 11, 2013bst.sagepub.comDownloaded from at Chongqing University on October 11, 2013bst.sagepub.comDownloaded from at Chongqing University on October 11, 2013bst.sagepub.comDownloaded from at Chongqing University on October 11, 2013bst.sagepub.comDownloaded from at Chongqing University on October 11, 2013bst.sagepub.comDownloaded from
Bulletin of Science, Technology & Society
32(2) 128 –141
© 2012 SAGE Publications
Reprints and permission:
Industrial wind turbines (IWTs) are being situated near
human habitation in increasing numbers. In some communi-
ties individuals who are exposed to wind turbines report
experiencing negative impacts including adverse health
effects. Falmouth Massachusetts, USA is a community
located in a quiet rural environment where there were
reports of negative health effects from locating IWTs too
close to residences. Some Falmouth residents have identi-
fied wind turbine noise as a cause of negative effects.
During a noise study investigating acousticians experi-
enced adverse health symptoms similar to those described by
residents living at the study location and near other IWT
sites. The onset of adverse health effects was unexpected and
persisted for some time after leaving the study area.
This case study provides wind turbine noise measure-
ments and other technical data and describes the symptoms
experienced by the investigators and explores the plausibility
that wind turbine low-frequency energy could contribute to
reported adverse health effects.
Falmouth, Massachusetts, U.S. Wind Turbines
Falmouth, Massachusetts recently installed three IWTs
(Vestas, V82, 1.65 MW); two owned by the town located at
the municipal wastewater treatment plant (WIND1 and
WIND2) and one privately owned at a nearby industrial park
(NOTUS). This area has a limited amount of daytime business
activity and only a distant highway with low traffic volumes
at night. The area is representative of a quiet rural environ-
ment with widely spaced houses. WIND1 and NOTUS are
installed with the nearest residences approximately 400 m
(1,300 feet) and 520 m (1,700 feet), respectively.
The WIND1 and NOTUS IWTs were installed over sev-
eral months, with WIND1 being the first to come on line in
March 2010. A short time later, neighbors began to complain
about excessive noise coming from WIND1. Later that year,
NOTUS began operation and similar complaints came in
from other neighbors. Complaints continued for months and
neighbors were reporting that they could not adjust to the
fluctuating sound, the endless swish and thumps. They found
the noise to be intrusive and disruptive to normal at home
activities. WIND2 was not operating during this study.
These fluctuating audible sounds or amplitude modulations
are the routine characteristic of IWTs and can be disturbing
in of Science, Technology & SocietyAmbrose et al.
S.E. Ambrose & Associates, Windham, ME, USA
Rand Acoustics, Brunswick, MA, USA
Killaloe, Ontario, Canada
Stephen E. Ambrose, S.E. Ambrose & Associates, 15 Great Falls Road,
Windham, ME 04062, USA
Wind Turbine Acoustic
Investigation: Infrasound and
Low-Frequency Noise—A Case Study
Stephen E. Ambrose
, Robert W. Rand
, and Carmen M. E. Krogh
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.
wind turbines, infrasound, low-frequency noise, physiological symptoms, adverse health effects
Ambrose et al. 129
and stressful to exposed individuals (G. Leventhall, 2006).
During moderate wind speeds the IWT noise was clearly audi-
ble outdoors and for some, indoors. At times the noise included
an audible low-frequency tone that came and went. Neighbors
commented that the wind turbine noise was more noticeable
indoors and it interfered with their relaxation and sleep.
The town responded to the numerous and persistent com-
plaints by requiring postoperational noise surveys to deter-
mine if there were justifications for complaints. Neighbors
responded by hiring legal counsel and had independent noise
measurements performed and evaluated for adverse impacts.
Most measurements were conducted by experienced acousti-
cians. The primary acoustic quantifier measured was the
average A-weighted sound level (dBA). The sound levels
generally ranged from the mid-30s to mid-40s dBA. Some
noise-level variations were due to differences for time of
day, wind speed, and wind direction (upwind or downwind).
Measured sound levels were fairly consistent from each sur-
vey provider. However, the acoustic reports had little effect
on complaint resolution.
Falmouth Health Complaints
After WIND1 and NOTUS IWT started up, neighbor com-
plaints included adverse health symptoms. They had days
where they were unable to enjoy the previous peace and
tranquility while at home, unable to relax, felt tense, and felt
a strong desire to be someplace else. They noticed some
relief when outdoors. The lessening of adverse effects when
outdoors and the indoor worsening are consistent with the
findings of low-frequency noise (LFN) effects exposure
(Burt, 1996). Typically, the indoor A-weighted sound level is
lower than the outdoor, especially when indoor human activ-
ity is at a minimum. The house exterior walls provide more
middle- to high-frequency band attenuation than for the low
and very low bands. Therefore, the average A-weighted
sound level by itself may not be a useful measurement indica-
tor for determining the potential for IWT complaints.
Some complainants described having significant diffi-
culties living in their home with reports of experiencing
headaches, ear pressure, dizziness, nausea, apprehension,
confusion, mental fatigue, lassitude (inability to concen-
trate, lethargy). These were worse when IWTs were operat-
ing during moderate to strong winds. A few neighbors
moved their bedrooms into the basement in an attempt to get
a good night’s sleep. Others were forced to leave their home
to sleep farther away at a family or friend’s house or even in
a motel. These symptoms (DeGagne & Lapka, 2008; Schust,
2004) and behavior patterns (H. G. Leventhall, 2004) are
consistent with LFN exposure suggesting that IWT low-
frequency energy may be a factor.
The purpose of the study was to confirm or deny the pres-
ence of infrasound (very-low-frequency noise, acoustic
waves, or pressure pulsations less than 20 Hz) and LFN
emissions (20-200 Hz) created by an IWT. The combination
of infrasound and low-frequency noise is defined as ILFN.
If ILFN was present the study was to determine: (a) if it was
greater than or uniquely distinguishable from the ambient
background levels and (b) if it exceeded human detection
thresholds. It was not the intention of this study to determine
the precise mechanism that linked the IWT to the physiolog-
ical or psychological symptoms being reported by residents.
The scope of this study was conducted at one home that is
representative of many other households that have com-
plained about noise and adverse health effects. The investi-
gators assessed differences between outdoor and indoor
Acoustic Measurements and
Acoustic measurements were made with precision sound
measurement instruments and dual-channel computer-based
signal analyzer software. These instruments were capable of
measuring very-low-frequency energy, as low as 1 Hz.
Frequency response was flat (within 1 dB) to 2 Hz and 6 Hz
for the two primary measurement channels. Prior to com-
puter analysis, the microphone and preamplifier frequency
response were corrected to flat (1-6 Hz) using manufacturer
data sheets. Instruments are itemized in Table 1.
Each sound-level measurement system was indepen-
dently field-calibrated (end-to-end) prior to and verified
after the survey measurements with an acoustic sound-level
calibrator (Brüel & Kjær, Type 4230 or Larson Davis
CAL200), generating a 1,000 Hz tone with 94 dB sound
pressure level (SPL) reference 20 µPa root mean square
(RMS). Sound-level meters and acoustic calibrators had cur-
rent laboratory calibration certificates traceable to National
Institute of Standards and Technology.
The ANSI (American National Standards Institute) filter
characteristics of Type 1 instrumentation have a long impulse
response time at low frequencies. At 1 Hz, the ANSI 1/3
octave band impulse response is close to 5 seconds. Thus,
ANSI filters do not capture the fast peak pressure changes
occurring in the low and infrasonic frequencies (Bray &
To observe fast peak pressure changes, signal analysis
was improved by using an external digital filter in series with
the digital recording playback output, and then analyzing the
digital data with a fast Fourier transform (FFT) signal ana-
lyzer with short time length (<128 milliseconds).
Field testing was conducted in general accordance with
applicable ANSI Standards, ANSI S12.18-1994 (“Procedures
for Outdoor Measurement of Sound Pressure Level,”
Method 1) and S12.9-1993/Part 3 (“Procedures for Short-
Term Measurements with an Observer Present”). Indoor-
outdoor simultaneous measurements were made using two
microphones to determine the outside-to-inside level reduc-
tion (OILR) for the exterior walls and roof. The OILR
130 Bulletin of Science, Technology & Society 32(2)
measurements were performed in accordance with ASTM
E966-02 (ASTM, 2010). The indoor microphone was fitted
with a 4-inch windscreen and mounted on a microphone
stand in the master bedroom at a location where the reported
adverse symptoms were more pronounced. The outdoor
microphone was fitted with a 4-inch windscreen and placed
inside a RODE Blimp for improved wind and shock mount
protection. The entire outdoor system was mounted on a tri-
pod, positioned 5 feet above the ground, and located away
from house and trees. Wind speeds were light at the outdoor
microphone position. In addition to noise measurements,
weather, temperature, and wind speed data were collected.
The A- and C-weighting, octave band, and FFT analysis
were performed with SpectraPLUS software in real time and
recording mode on-site. The recorded data were analyzed off-
site using the postprocessing features. G-weighted sound lev-
els were computed using FFT settings for octave band analysis
of the G-filtered 4, 8, 16, and 31.5 Hz octave bands using the
G-weighting corrections which are the average value for the
one-third octave bands comprising each full octave band (ISO
7196:1995, “Acoustics–Frequency Weighting”). While coarse
in approach, the method was determined to be a usable trade-
off between analysis time, accuracy, and computational
requirements. It should be noted that the dBG levels obtained
using the ANSI octave band filtering would not capture the
highest peak pressure changes, so data reported are considered
to understate the peak dBG levels.
The A-, C-, G-weighting and unweighted (dBL) filter
functions are shown in Figure 1.
The A- and C-weighting filters discount frequency-level
contributions below 1,000 Hz and 20 Hz, respectively. The
G-weighting was created for evaluating infrasound, peak-
ing at 20 Hz with rapid declines above and below which
follow the recognized hearing response to pure sine waves,
with a slope of 12 dB per octave. Unweighted (or dBL;
dashed line) has flat frequency response over the entire
The survey was started in the late afternoon of April 17,
2011 (Day 1) and concluded in the morning of April 19,
2011 (Day 3). The weather conditions were representative
of pleasant warm, windy spring days with cool, calmer
Outdoor measurements were made when weather condi-
tions were favorable for measurements (ground-level winds
≤ 9 mph [miles per hour] and no precipitation). Observed
weather conditions and the nearest publicly accessible met
tower are presented in the appendix.
Wind Turbine Operations
In the spring of 2011, Falmouth imposed a maximum wind
speed restriction on their WIND1 in an effort to mitigate
neighbors’ complaints. WIND1 operation was modified to
curtail power generation whenever the hub-height wind
speeds exceeded 10 m/s. The town did not curtail NOTUS
even though it was close to neighbors. The manufacturer has
a setting to trip units off when the hub-height wind speed
exceeds 32 m/s.
Figure 1. Weighting functions
Source. Adapted with permission from figure located at http://oto2.wustl
Table 1. Instrument List
Instrument Manufacturer Model
Microphone Brüel & Kjær 4165
Preamplifier Larson Davis 2221
Microphone GRAS 40AN
Preamplifier Larson Davis 902
Sound level meter Larson Davis 824
Calibrator Brüel & Kjær 4230
Audio interface Sound Devices USBPre2
Recorder M-Audio Microtrack II
Software Pioneer Hill SpectraPLUS 5.0
Microphone Svantek SV22
Preamplifier Svantek SV12L
Sound level meter Svantek 949
Calibrator Larson Davis CAL200
Audio interface ROGA DAQ2
Recorder TEAC DR100
Octave band (Hz) 4 8 16 31.5
dBG correction (dB) −16 −4 +7.7 −4
Ambrose et al. 131
Observations and Comments
Day 1: Hub-height wind speeds were from the west at
20 to 25 m/s, gusts exceeding 30 m/s (66 mph, gale
force aloft). Surface winds were light from the south-
east, contrary to upper level westerly winds. At night,
the hub-height wind speed slowly decreased to light,
whereas the surface wind speed decreased to nearly
Outdoor noise measurements were first made on arrival at
the study house. The NOTUS turbine was clearly audible
(520 m distant) and WIND1 (1,220 m distant) was off.
Within 20 minutes of setting up work stations inside the
study house, the investigators started to experience a loss of
well-being and continued to worsen with time. They had dif-
ficulty performing routine survey and measurement tasks:
connecting instruments, assessing for proper operation, and
calibration. They experienced inability to stay focused using
a computer or track survey scope of work.
After repeated efforts, it was determined that reliable
indoor measurements were not possible because of debilita-
tion. No meaningful measurements were acquired at ML-1
during the first evening when winds were strong.
Near midnight the wind speed started to decrease, prompt-
ing an effort to leave the house to attempt outdoor noise mea-
surements nearer NOTUS. These measurements are
discussed in more detail in the “Sound Level Versus Distance
Day 2: Light pre-dawn hub-height wind speed slowly
increased during the morning to above 18 m/s and
continued throughout the day and decreased to light in
the early evening. During the early night the wind
speed remained light.
NOTUS noise was dominating with outdoor and indoor lev-
els in the low 40s and 20s dBA, respectively. Spectral, one
third, and full octave band sound levels were viewed with
computer-based frequency analysis software for several hours
during the day. Infrasound and low frequencies were of special
interest and these had the highest unweighted SPLs. Outdoor–
indoor (OILR) measurements were conducted. Digital record-
ings were made for a postprocessing at a later date.
Day 3: After midnight the wind speed increased to
strong and decreased to light at sunrise.
Normal workday sounds from nearby commercial activity
were intermittently audible. There were faint noises from die-
sel equipment operating at a nearby sandpit, light traffic on
Rte. 28, 1,700 m (5,600 feet) away and an occasional vehicle
on the nearest road, 300 m (1,000 feet) away). NOTUS was
stopped and WIND1 was inaudible but operating in light
winds as observed by ILFN modulations detectable on ana-
lyzer. This presented an opportunity to obtain digital record-
ings with WIND1 operating alone in light winds at ML-1.
The wind died and the survey was concluded mid-morning.
Sound Level Versus Distance Measurement
Sound-level measurements and recordings were made at
four distances to show the noise level decrease with increas-
ing distance and the distance for blending into the back-
ground acoustic environment. This technique can be called
“level versus distance,” “walk-away,” or “stepped distance.”
Measurements with digital recordings were made at three
locations trending north-northeast away from NOTUS
(MLA, B, and C at 80, 250, & 410 m (260, 830, 1,340 feet),
respectively) in the Falmouth Technology Park, as shown in
Figure 2. Measurements were ceased when it started to rain
after 1:30 a.m. The fourth location (ML-1) was to the south-
east at the survey residence (at 520 m or 1,700 feet). NOTUS
noise was dominant at all measurement locations.
IWT power outputs were obtained from the NOTUS and
WIND1 websites. Figure 3 shows the power output and
Table 2 was created to correlate the NOTUS IWT power
output, measured dBA, dBG, and dBL data at ML-1 and
adverse health effects experienced by the investigators at ML-1
during the operating conditions of the NOTUS wind turbine.
Figure 2. NOTUS measurement locations
132 Bulletin of Science, Technology & Society 32(2)
Figure 4 was created by combining Table 2 with Figure 3
to show the relationship of NOTUS power output, wind
speed, and health states experienced by the investigators.
WIND1 was configured with an operational cap at 10 m/s
and was off during the higher wind speeds. The investigators
were most noticeably affected when the IWT power output
was highest, with wind speeds more than 10 m/s at hub
height for NOTUS while at the study location (at 520 m).
Figure 4 also shows the hours when the investigators were
not as severely affected. Symptoms moderated during the first
night when IWT power output dropped when nighttime noise
measurements were made near NOTUS, and later while sleep-
ing. When the power output increased (with wind speed greater
than 10 m/s) during the following morning, symptoms returned,
yet slowly went away (with increased distance from the IWT)
after leaving the area for breakfast. On returning to the study
house (at 520 m) the symptoms quickly set in again and
remained strong until late afternoon when IWT power output
dropped with lower wind speeds. The investigators left for an
evening meal and symptoms moderated somewhat, yet, even
with the increased distance from the IWT, the symptoms contin-
ued strongly enough to suppress appetite and affect thinking.
When the investigators went to bed they had fitful sleep with
numerous awakenings. Concurrently, IWT power output
increased during the night, with average hub-height wind speeds
fluctuating above and below 10 m/s during the early morning
hours. In the morning, winds decreased to light, with NOTUS
stopped and WIND1 turning in the distance (at 1,220 m).
Onsite Analysis Conducted on Day 2
A representative outdoor noise spectrum (RMS) was plotted
with the outer hair cells (OHCs) and inner hair cells (IHCs)
dBG thresholds, as shown in Figure 5A. The graph shows
that the NOTUS 22.9-Hz tone exceeds the OHC threshold
of 45 dB at 22.9 Hz. The 129-Hz tone exceeded the IHC
threshold and was confirmed as audible outdoors (see, OHC
and IHC) in the “Discussion” section).
The simultaneously measured indoor noise spectrum
(RMS) is shown in Figure 5B. The graph shows that the
NOTUS 22.9-Hz tone again exceeds the OHC threshold. The
129-Hz tone was less audible than outdoors. The spectrum
was amplitude modulated and the averaged spectrum does
not reveal the peak sound levels which may have exceeded
the audibility threshold.
Time-History Tone Analysis
NOTUS noise levels and frequency content noticeably fluctu-
ated with time. It would be appropriate to analyze these varia-
tions versus time focusing on the 22.9-Hz tone because it was
shown to be detectable by the OHC. A 20 to 24 Hz 10th order
digital bandpass filter was inserted between the digital record-
ing output and the analysis input channel for SpectraPLUS
software set to acquire FFT frames at 23-millisecond intervals
using Hamming weighting. These furnished the band-limited
tonal energy at 22.9 Hz free of ANSI filter response times.
Figure 6 shows the indoor time history of 22.9-Hertz
amplitude variations above and below the OHC threshold of
45 dB. This graph shows amplitudes as high as 60 dB, which
is 10 dB higher than the 50 dB average. The total fluctuation,
maximum to minimum exceeds 50 dB.
This graph shows that the OHC is receiving pressure
events nearly every 43 milliseconds at least 50% of the time
during the measurement. The 22.9-Hz tone was not audible
because it was not strong enough to exceed the IHC thresh-
old (approximately 72 dB at 22.9 Hz).
Time-History dBG Analysis
Indoor and outdoor recordings with NOTUS operating were
made on the afternoon of Day 2 and with NOTUS not oper-
ating due to very light wind on the morning of Day 3. This
enabled time-history plots showing the dBG differences
between NOTUS “ON” and NOTUS “OFF” for both
indoors and outdoors as shown in Figure 7A and B. These
data illustrate amplitude modulations exceeding 60 dBG.
They were acquired through ANSI filter octave bands cor-
rected to dBG. Because of ANSI filter impulse response
times, they do not capture the highest peak pressure levels.
Indoors, the NOTUS “ON” dBG levels were about 20 dB
higher than when “OFF.” Outdoors, the NOTUS “ON” ver-
sus “OFF” dBG difference was about 10 dB.
Sound Level Versus Distance Measurement
Outdoor sound levels decrease at about 6 dB per doubling
of distance (6 dB/dd) as depicted by the inverse square law
for acoustic frequencies. Sound level versus distance mea-
surements were plotted using a semilog scale for distance.
This graphing method typically shows the drop of sound
level as a straight line as the distance increases.
Figure 3. Wind turbine wind speed and power output
Ambrose et al. 133
The “stepped distance” data combined with the data at
ML-1 show that the NOTUS noise level decreases with dis-
tance uniformly, as shown in Figure 8.
Two trend lines are included; the lower dashed line
shows the dBA sound levels decreasing at a predictable
6 dB per distance doubling (6 dB/dd). The dBA trend line is
faired through a wind speed of 8 m/s per the NOTUS speci-
fication wind speed. The upper dashed line is for unweighted
sound levels, which was controlled by frequencies below
20 Hz. The unweighted sound levels decrease at about 3 dB/
dd, which is representative of cylindrical spreading.
Noise levels at the study house showed that the indoor
levels were more than 20 dBA quieter than outdoors.
However, the unweighted dBL levels were several dB
higher indoors than outdoors, indicating that the house was
providing reinforcement (amplifying) of the very low
House Noise Reduction
Measurements were made with the NOTUS “ON” with hub-
height wind speeds averaging about 20 m/s. One-minute
duration transfer function analysis measured the difference
between outside and inside noise levels. The difference is
shown by narrow band frequency (FFT) in Figure 9A, and
by full-octave bands in Figure 9B.
The two graphs show the OILR by the two exterior master-
bedroom walls and roof. Negative values indicate attenuation
Table 2. NOTUS Operations, ML-1 Sound Levels, and Adverse Health Effects
Hub wind speed
(kW) Location dBA dBG dBL Symptoms experienced
Day 1: 25, gusts: 35 1,600-1,700 Indoors n/a n/a n/a Nausea, dizziness, irritability, headache,
loss of appetite, inability to
concentrate, need to leave, anxiety
Outdoors n/a n/a n/a Felt miserable, performed tasks at a
Night 1: 0-9 150-350 Indoors 18-20 n/a n/a Slept with little difficulty
Day 2: 20, gusts: 30 1,350-1,500 Indoors 18-24 51-64, pulsations 62-74, pulsations Dizzy, no appetite, headache, felt
miserable; performed tasks at a
reduced pace. Desire to leave
Outdoors 41-46 54-65, pulsations 60-69, pulsations Dizzy, headache, no appetite. Slow.
Preferred being outdoors or away
Night 2: 150-350 Indoors 18-20 n/a n/a Slept fitfully, woke up
Day 3: calm to 6 OFF Indoors 18-20 39-44, random 50-61, random Improvement in health. Fatigue and
desire to leave
Outdoors 32-38 49-54, random 57-61, random Improvement in health. Fatigue and
desire to leave
Figure 4. Survey operations at ML-1
Figure 5. (A) Outdoor and (B) indoor NOTUS sound levels
(averaged) versus outer hair cell (OHC) and inner hair cell (IHC)
134 Bulletin of Science, Technology & Society 32(2)
and positive values show amplification. The graphs show
high-frequency attenuation of 20 dB or more, about 15 dB in
the 31.5-Hz octave band, and about 10 dB in the 8- and 16-Hz
octave bands. The very low-frequency bands show amplifica-
tion of about 3 and 8 dB in the 4- and 2-Hz bands,
Because of the house structure dramatically influencing
interior very-low-frequency levels, the meter measurement
units were changed from the log scale (dB) to a linear Pascal
to expand the “y”-axis scale. The outdoor and indoor octave
band Pascal levels are shown in Figure 10A and B, respec-
tively. These are averaged levels and do not illustrate the
dynamic amplitude modulation.
The difference between indoors and outdoors time his-
tory is shown in Figure 11. The outdoors graph shows the
influence of higher frequencies that are not present indoors
due to structure attenuation. Dynamic amplitude modula-
tion is clearly visible.
The comment “It’s like living inside a drum” has been made
by many neighbors living near IWT sites. These comments
suggest that IWT low-frequency energy is being acousti-
cally coupled into the interior space. Coherence analysis
was used to determine the relationship between outdoor and
indoor acoustic signals. Coherence values approaching 1.0
have a strong correlation and when less than 0.7 there is
significantly less correlation. Figure 12 presents the coher-
ence analysis results with the strong correlation, 0.7 to 1.0
Figure 6. 22.9-Hz tone and OHC threshold
Note. OHC = outer hair cell; RMS = root mean square; SPL = sound pres-
Figure 7. (A) Indoor and (B) outdoor dBG levels
Figure 8. NOTUS root mean square (RMS) sound level versus
Figure 9. Outside-to-inside level reduction: (A) fast Fourier
transform and (B) octave band
Ambrose et al. 135
The highlight banding shows which frequencies inside
the house are judged to be directly coupled to the outside
energy. High coherence was evident for the very low infra-
sonic frequencies and at 22.9 and 129 Hz.
Dynamic Amplitude Modulation
Wind turbine noise has a unique sound characteristic that
distinguishes it from other man-made and environmental
noise due to the strong dynamic amplitude modulation
caused by the blades. Overall dBA, dBC, and dBL acoustic
signatures were graphed as level versus time, as shown in
Figure 13. The amplitude modulation was occasionally
audible as indicated in the dBA time history. The dBL time
history has higher amplitude modulations than dBA and
dBC because there is no filter reduction for lower frequen-
cies and, the strong amplitude modulations occurring at the
blade pass frequency are revealed.
A comparison of the overall dBL indoors versus outdoors
shows that the indoors levels are about 2 to 8 dB higher than
outdoors, as shown in Figure 14. This graph also shows that
the amplitude modulation increased in range indoors with
rise and fall exceeding 10 dB per second.
The increase in the dBL levels and amplitude modulation
indoors is consistent with and supports neighbors’ comments
that it is worse indoors than outdoors.
NOTUS “ON” and “OFF”
Outdoor measurements with NOTUS “ON” show stronger
pulsation fluctuations than when NOTUS is “OFF,” as
shown in Figure 15.
Pressure Pulsation Exposure
and Dose Response
It is generally accepted that human response and cumulative
effect to intrusive noise exposure increases with number of
peak noise events and peak level. This is consistent with the
gradual onset over some 20 minutes of adverse health effects
experienced by the investigators at ML-1 on the first day and
the repeated onset of symptoms when returning to ML-1 dur-
ing the survey.
For total unweighted sound exposure, the investigators
were exposed to dynamically modulated pressure pulsations
Figure 10. (A) Outdoor and (B) indoor sound pressure in Pascals
Figure 11. Pressure fluctuation time history in Pascal
Figure 12. Coherence, outdoors to indoors
Figure 13. Outdoors sound levels: NOTUS “ON” (April 18, 2011)
136 Bulletin of Science, Technology & Society 32(2)
every 1.4 seconds (NOTUS blade pass rate) at the study
house (Figure 15). After being indoors for 15 minutes, the
pulsations totaled 642 peak pressure events. Every hour
there were 2,570 pressure events. When the physiological
effects were worst (at 5 hours exposure) the total exposure
was 12,800 blade-pass peak pressure events. The time-
history data suggest that over 50% of the peak pressure
impacts exceeded the 60 dBG physiological OHC threshold
(see OHCs and IHCs in the “Discussion” section).
The occurrence of pressure events at 22.9 Hz (Figure 6)
is much higher. The acoustic pressure at 22.9 Hz dropped
well below OHC threshold and then peaked over OHC
threshold but not over the IHC threshold, at a rate of more
than 82,000 per hour and more than 400,000 in 5 hours. If
50% of the 22.9-Hz pressure levels were detected by the
OHC that would result in more than 200,000 stimulations to
the OHCs in a 5-hour period.
Human Detection Thresholds
Sound pressure is the small alternating deviation above
and below atmospheric pressure due to the propagated
wave of compression and rarefaction. The unit for sound
pressure is the Pascal (symbol: Pa). SPL or sound level is
a logarithmic measure of the effective sound pressure of a
sound relative to a reference value. It is measured in deci-
bels (dB) above a standard reference level. The commonly
used “zero” reference sound pressure in air is 20 µPa
RMS, which is usually considered the median threshold of
human hearing (at 1 kHz). Some 16% of the population is
about 6 dB more sensitive than the median, and some 2%
is 12 dB more sensitive. The percentage of people who are
more sensitive who choose to live in quieter rural areas is
unknown. That is, those living in quiet areas may have
sensitivity shifted toward lower thresholds and self-select
Frequency is measured by the number of waves per sec-
ond or Hertz (Hz). The average range of hearing is 20 to
20,000 Hz with the greatest sensitivity in 1,000 to 4,000 Hz
range. At the most sensitive frequency around 4 kHz, the
amplitude of motion of the eardrum is about 10 to 9 cm,
which is only about 1/10 the diameter of a hydrogen atom.
Thus, the ear is very sensitive, detecting signals in the range
of atomic motion.
Outer Hair Cells and Inner Hair Cells
There are two types of hair cells in the cochlea where sound
pressure is converted to nerve impulses; the IHCs and the
OHCs. The IHCs are fluid connected and velocity sensitive,
responding to minute changes in the acoustic pressure
variations based on frequency, with sensitivity decreasing at
a rate of −6 dB per downward octave. IHCs detect audible
sounds and they are insensitive to low-frequency and infra-
sonic acoustic energy. In contrast, the OHCs are mechani-
cally connected, or DC-coupled, to movements of the
sensory structure and respond to infrasound stimuli at mod-
erate levels, as much as 40 dB below IHC thresholds. The
approximate threshold for physiological response by OHCs
to infrasound is 60 dBG.
Figure 14. Acoustic pressure fluctuation time history (indoors
versus outdoors; April 18, 2011, 3:22 p.m.)
Figure 15. NOTUS “ON” and “OFF” sound pressure levels
Figure 16. Human audibility curves
Ambrose et al. 137
Figure 16 shows the IHC and OHC responses compared
with ISO 2003 and Møller and Pedersen (2011) audibility
measurements. Adapted with permission, from figure
located at http://oto2.wustl.edu/cochlea/romesalt.pdf
OHC responses to infrasound are maximal when ambient
sound levels are low. Furthermore, low-frequency sounds
produce a biological amplitude modulation of nerve fiber
responses to higher frequency stimuli. This is different from
the amplitude modulation of sounds detected by a sound-
level meter (Salt & Lichtenhan, 2011).
Adverse Health Effects
A 2011 Ontario Review Tribunal Decision found that wind
turbines can harm humans if placed too close to residents
This case has successfully shown that the debate should
not be simplified to one about whether wind turbines
can cause harm to humans. The evidence presented to
the Tribunal demonstrates that they can, if facilities are
placed too close to residents. The debate has now
evolved to one of degree. (Erickson v. Director, 2011)
Some individuals exposed to wind turbines report experienc-
ing adverse health effects which include physiological and psy-
chological symptoms as well as negative impacts on quality of
life (Harry, 2007; Krogh, Gillis, Kouwen, & Aramini, 2011;
Nissenbaum, Aramini, & Hanning, 2011; Phipps, Amati,
McCoard, & Fisher, 2007; Shepherd, McBride, Welch, Dirks,
& Hill, 2011; Thorne, 2011). In some cases the adverse effects
are severe enough that some individuals have elected to aban-
don their homes. In other cases, homes of individuals reporting
health effects have been purchased by the wind energy devel-
oper (Krogh, 2011). The World Health Organization’s (1948)
definition of health includes physical, mental, and social well-
being. Adverse impacts associated with IWTs fall within the
WHO definition of health.
Pierpont (2009) describes symptoms reported by individ-
uals living near wind turbines. Symptoms include “sleep dis-
turbance, headache, tinnitus, ear pressure, dizziness, vertigo,
nausea, visual blurring, tachycardia, irritability, problems
with concentration and memory, and panic episodes associ-
ated with sensations of internal pulsation or quivering when
awake or asleep.” G. Leventhall (2009) states,
I am happy to accept these symptoms, as they have been
known to me for many years as the symptoms of extreme
psychological stress from environmental noise, particu-
larly low frequency noise . . . what Pierpont describes is
effects of annoyance by noise–a stress effect . . .
An expert panel review commissioned by the American
Wind Energy Association and Canadian Wind Energy
Association stated that these symptoms are not new and have
been published previously in the context of “annoyance” to
environmental sounds and are an example of the “well-
known stress effects of exposure to noise” associated with
noise annoyance (Colby et al., 2009).
Wind turbine sound is perceived to be more annoying
than other equally loud sources of noise (Pedersen, Bakker,
Bouma, & van den Berg, 2009). Higher levels of annoyance
may be partly explained by wind turbine noise amplitude
modulation, lack of night time abatement, and visual
impacts. Wind turbine tonal and audible low-frequency
sound are also plausible causes of wind turbine noise annoy-
ance (Møller & Pedersen, 2011) and reported health effects
(Minnesota Department of Health, 2009) and, may play an
important part in the cause for adverse community reaction
to large IWTs installed close to residences in quiet areas.
Complaints associated with wind turbine low-frequency
noise are often more prevalent indoors than outdoors.
Recently there have been recommendations to address the
impacts of wind turbine low-frequency noise (Howe
Gastmeier Chapnik Limited, 2010; The Social and Economic
Impact of Rural Wind Farms, 2011).
Wind turbine noise standards and most regulations are
based on the averaged A-weighting metric which suppresses
the amplitude of low-frequency noise predictions in model-
ing and application submittals. Averaged A-weighted sound-
level measurements are unsatisfactory when individuals are
annoyed by low-frequency sound and amplitude modulation
(H. G. Leventhall, 2004; Richarz, Richarz, & Gambino,
2011). The A-weighting filter severely attenuates low-
frequency signals (the primary frequency range of most
community noise complaints) and essentially eliminates
acoustic signals below 20 Hz where “infrasound” is located
in the acoustic frequency spectrum.
Low-frequency vibration and its effects on humans are
not well understood and sensitivity to such vibration result-
ing from wind turbine noise is highly variable among
humans (National Research Council, 2007). Whether expo-
sure to wind turbine infrasound can contribute to adverse
effects in humans is a subject of considerable debate. There
are aspects of infrasound from wind turbines that are not
unanimously accepted by all technical and medical practi-
tioners (Howe Gastmeier Chapnik Limited, 2010). Some
discount wind turbine infrasound as a concern on the basis
that levels are below the hearing threshold (Colby et al.,
2009; G. Leventhall, 2006). It is noted that other noise
sources can generate infrasonic energy, such as surf and
thunderstorms. However, wind turbine low-frequency
energy presents a recurring and/or unpredictable pressure
signature, with audibility or delectability occurring over a
much longer period of time than other environmental
sources of low-frequency energy.
An audible or detectable acoustic or pressure signature is
valuable for subsequent monitoring of system design and
138 Bulletin of Science, Technology & Society 32(2)
correlating with complaints and exploring the plausibility
that wind turbine low-level low-frequency energy could con-
tribute to reported adverse health effects.
Infrasonic thresholds for human perception have been
found to be lower than those previously estimated based on
traditional sinusoidal hearing tests. There is evidence indi-
cating that vestibular system does respond to sound we can-
not hear (Salt & Hullar, 2010). Infrasound is understood by
acousticians to refer inaudible acoustic energy for frequen-
cies less than 20 Hz. There is increasing evidence that the
OHC can detect nonsinusoidal pressure fluctuations at
lower amplitudes than the IHC. Current research estimates
that sound levels of 60 dBG for frequencies from 5 to 50 Hz
can stimulate the OHC for the human ear (Salt &
Cochlear microphonic responses to infrasound recorded
in endolymph of the third turn of the guinea pig cochlea are
suppressed by the presence of higher frequency sounds. This
suggests that the physiologic response to infrasound may be
maximal when heard under quiet conditions, such as that
may occur in a quiet bedroom in the vicinity of a wind tur-
bine (Salt & Lichtenhan, 2011).
Sleep disturbance is one of the most common adverse
health effects reported by neighbors living near IWTs
(Hanning & Evans, 2012; Minnesota Department of Health,
2009). The investigators experienced sleep disturbance,
especially during the second night when hub-height wind
speeds were greater than 10 m/s. The indoor sound level was
low at around 20 dBA and was below levels typically recom-
mended to minimize sleep disturbance.
Sleep disturbance during this study was experienced by
the investigators and reported by the home owners. A first
assessment of the analyzed noise level data appears to show
a stronger correlation with the 60-dBG threshold than it does
with dBA-weighted sound levels. Recorded noise level anal-
ysis shows that NOTUS produces a strong 0.7-Hz blade-pass
modulation and a strong 22.9-Hz tone sufficient to be
detected by the OHC but remain inaudible.
Noise and Pressure Pulsations
This study revealed dynamically modulated low-frequency
and infrasonic energy produced by NOTUS. The acoustic
energy from NOTUS was found to be greater than and
uniquely distinguishable from the ambient background lev-
els without NOTUS operating. NOTUS produced dynamic
infrasonic modulations that were not present when the wind
turbine was off. NOTUS “ON” produced tonal energy at
22.9 and 129 Hz, which were found to be strongly coupled
to the study house interior. Amplitude modulations below
10 Hz were amplified indoors, suggesting a whole house
acoustic cavity response.
The dBG levels indoors were dynamically modulated at
the blade-pass rate and tonal frequencies and exceeded the
vestibular physiological threshold guideline of 60 dBG.
Adverse Health Effects
A dose-response relationship to peak pressure events
detected by the OHC is supported by the gradual onset of
adverse health symptoms while near the IWT. At SPLs asso-
ciated with worsened health symptoms, NOTUS produced
low-frequency pressure pulsations that could be detected by
the ears’ OHCs but not by the IHCs. Health effects moder-
ated when dBG levels fell well below the 60-dBG guideline
when the wind turbine was OFF, or when well away (several
miles) from NOTUS.
The rapid onset of adverse health effects during the
study confirms that wind turbines can harm humans if
placed too close to residents. During the study, investiga-
tors without a preexisting sleep deprivation condition, not
tied to the location nor invested in the property, experi-
enced similar adverse health effects described and testified
to by residents living near the wind turbines. Sound mea-
surements acquired during the study indicate that
A-weighted sound levels did not correlate to adverse health
effects experienced. Adverse health effects experienced by
investigators were more severe indoors where dBA levels
were approximately 20 dBA lower than outdoors levels.
The dBL (unweighted) and dBG (infrasonic-weighting)
levels were higher and more strongly amplitude-modulated
indoors compared to outdoors. The increase in amplitude
modulation indoors was consistent with the stronger
adverse health effects experienced indoors.
Wind turbine audible sound is perceived to be more
annoying than equally loud transportation or other industrial
noise and can be expected to contribute to stress-related
health effects. Symptoms reported by some individuals
exposed to IWTs can include sleep disturbance, headache,
tinnitus, ear pressure, dizziness, vertigo, nausea, visual blur-
ring, tachycardia, irritability, problems with concentration
and memory, and panic episodes associated with sensations
of internal pulsation or quivering when awake or asleep.
This acoustic study suggests that health effects reported by
residents living near wind turbines may not be exclusively
related to audible sounds. Inaudible amplitude modulated
acoustic energy can be detected by the inner ear and can affect
humans more at low ambient sound levels, consistent with
complaints of worse conditions indoors than out near IWTs.
The study results emphasize the need for epidemiological and
laboratory research by health professionals and acousticians
concerned with public health and well-being. These findings
underscore the need for more effective and precautionary set-
back distances for IWTs. It appears prudent to include a mar-
gin of safety sufficient to prevent inaudible low-frequency
wind turbine noise from adversely affecting humans.
Ambrose et al. 139
April 17, 2011
April 18, 2011
Figure 17. Day 1: Changeable weather with wind speeds 25 to 30 m/s at the hub, gusting to more than 35 m/s. Wind direction west–
southwest. Barometer “low” and variable. Sunny and partly cloudy. Temperature 45°F to 50°F
Figure 18. Day 2: Sunny with wind speeds 15 to 20 m/s at the hub, gusting to 25 to 30 m/s. Wind direction west–southwest. Barometer
“low” and rising during the day. Temperature 45°F to 50°F
140 Bulletin of Science, Technology & Society 32(2)
Authors Stephen E. Ambrose and Robert W. Rand would like to
acknowledge the residents of Falmouth who welcomed them into
their homes, extended their hospitality, communicated their experi-
ences, and provided their time and assistance.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to
the research, authorship, and/or publication of this article.
The author(s) disclosed receipt of the following financial support
for the research, authorship, and/or publication of this article: The
data collected in the sound study were partially funded by a local
resident through a grant for the purpose of conducting an indepen-
ANSI S12.18-1994 [R2004]. American National Standard proce-
dures for outdoor measurement of sound pressure level. Retrieved
ANSI/ASA S12.9-1993/Part 3 [R2008]. American National
Standard quantities and procedures for description and
measurement of environmental sound. Part 3: Short-term mea-
surements with an observer present. Retrieved from http://
ASTM (2010). Standard guide for field measurements of airborne
sound insulation of building facades and facade elements.
Definition: outdoor-indoor level reduction, OILR—in a speci-
fied frequency band, the difference between the time-averaged
exterior sound pressure and the space-time average sound pres-
sure in a room of a building (ASTM Designation: E 966–02).
Retrieved from http://www.techstreet.com/standards/astm/
Bray, W., & James, R. (2011, July). Dynamic measurements of
wind turbine acoustic signals, employing sound quality engi-
neering methods considering the time and frequency sensitivi-
ties of human perception. Paper presented at Noise-Con 2011.
Retrieved from http://www.windaction.org/documents/33315
Burt, T. (1996). Sick building syndrome: Acoustical aspects. Indoor
and Built Environment, 5, 44-59. Retrieved from http://ibe.
Colby, W. D., Dobie, R., Leventhall, G., Lipscomb, D. M.,
McCunney, R. J., Seilo, M. T., & Søndergaard, B. (2009, December).
Wind turbine sound and health effects: An expert panel review.
Washington, DC: American Wind Energy Association and Cana-
dian Wind Energy Association. Retrieved from http://www.canwea.
DeGagne, D. C., & Lapka, S. D. (2008). Incorporating low fre-
quency noise legislation for the energy industry in Alberta,
Canada. Journal of Low Frequency Noise, Vibration and Active
Control, 27, 105-120.
Erickson v. Director. (2011). Ministry of the Environment, Environ-
mental Decision Case Nos. 10-121 and 10-122. Retrieved from
April 19, 2011
Figure 19. Day 3: Winds stopped in morning and the field study concluded at 9:30 a.m.
Ambrose et al. 141
Hanning, C., & Evans, A. (2012). Wind turbine noise. British Medi-
cal Journal, 344, e1527.
Harry, A. (2007, February). Wind turbines, noise and health. Retrieved
Howe Gastmeier Chapnik Limited. (2010, December 10). Low
frequency noise and infrasound associated with wind turbine
generator systems: A literature review (Rfp No. Oss-078696).
Mississauga, Ontario, Canada: Ministry of the Environment.
ISO (1995). Acoustics—Frequency-weighting characteristic for
infrasound measurements (ISO 7196:1995). Retrieved from
Krogh, C. (2011). Industrial wind turbine development and loss of
social justice? Bulletin of Science, Technology & Society, 31,
Krogh, C., Gillis, L., Kouwen, N., & Aramini, J. (2011). Wind-
VOiCe, a self-reporting survey: Adverse health effects, indus-
trial wind turbines, and the need for vigilance monitoring.
Bulletin of Science, Technology & Society, 31, 334-345.
Leventhall, G. (2006). Infrasound from wind turbines: Fact, fiction
or deception. Canadian Acoustics, 34, 29-36.
Leventhall, G. (2009). Wind turbine syndrome: An appraisal. Hear-
ing before the Public Service Commission of Wisconsin.
Leventhall, H. G. (2004). Low frequency noise and annoyance.
Noise & Health, 6, 59-72.
Minnesota Department of Health. (2009, May). Public health
impacts of wind turbines. Retrieved from http://www.health.
Møller, H., & Pedersen, C. S. (2011). Low-frequency noise from
large wind turbines. Journal of the Acoustical Society of
America, 129, 3727-3744.
National Research Council. (2007). Environmental impacts of wind
energy projects. Washington, DC: National Academies Press.
Nissenbaum, M., Aramini, J., & Hanning, C. (2011, July). Adverse
health effects of industrial wind turbines: A preliminary report.
Paper presented at the 10th International Congress on Noise
as a Public Health Problem (ICBEN) 2011, London, England.
Retrieved from http://www.windvigilance.com/about-adverse-
Pedersen, E., Bakker, R., Bouma, J., & van den Berg, F. (2009).
Response to noise from modern wind farms in the Netherlands.
Journal of the Acoustical Society of America, 126, 634-643.
Phipps, R., Amati, M., McCoard, S., & Fisher, R. (2007). Visual
and noise effects reported by residents living close to Manawatu
wind farms: Preliminary survey results. Retrieved from http://
Pierpont, N. (2009). Wind turbine syndrome: A report on a natural
experiment. Santa Fe, NM: K-Selected Books.
Richarz, W., Richarz, H., & Gambino, T. (2011). Correlating very
low frequency sound pulse to audible wind turbine sound,
Aercoustice Engineering Limited, Ontario, Canada. Fourth
International Meeting on Wind Turbine Noise, Rome, Italy.
Salt, A. N., & Hullar, T. E. (2010). Responses of the ear to low
frequency sounds, infrasound and wind turbines. Hearing
Research, 268, 12-21. doi:10.1016/j.heares.2010.06.007
Salt, A. N., & Kaltenbach, J. A. (2011). Infrasound from wind tur-
bines could affect humans. Bulletin of Science, Technology &
Society, 31, 296-302. doi:10.1177/0270467611412555
Salt, A. N., & Lichtenhan, J. (2011, April 12-14). Responses of the
inner ear to infrasound. Paper presented at the Fourth Interna-
tional Meeting on Wind Turbine Noise, Rome, Italy.
Schust, M. (2004). Effects of low frequency noise up to 100 Hz.
Noise & Health, 6, 73-85.
Shepherd, D., McBride, D., Welch, D., Dirks, K. N., & Hill, E. M.
(2011). Evaluating the impact of wind turbine noise on health-
related quality of life. Noise & Health, 13, 333-339.
The social and economic impact of rural wind farms. (2011,
May 17). Public hearing before the Senate Community
Affairs References Committee, Commonwealth of Australia.
Retrieved from http://www.aph.gov.au/senate/committee/
Thorne, B. (2011). The problems with noise numbers for wind farm
noise assessment. Bulletin of Science, Technology & Society,
World Health Organization. (1948, April 7). Preamble to the con-
stitution of the World Health Organization as adopted by the
International Health Conference, New York, 19-22 June, 1946;
signed on 22 July 1946 by the representatives of 61 States (Offi-
cial records of the World Health Organization, no. 2, p. 100).
Stephen E. Ambrose has more than 35 years of experience in
industrial noise control. Board Certified and Member INCE since
1978, he runs a small business providing cost-effective environ-
mental noise consulting services for industrial and commercial
businesses, municipal and state governments, and private
Robert W. Rand has more than 30 years of experience in industrial
noise control, environmental sound and general acoustics. A
Member INCE since 1993, he runs a small business providing
consulting, investigator, and design services in acoustics.
Carmen M. E. Krogh, BScPharm, provided research and refer-
ence support. She is a retired pharmacist with more than 40 years
of experience in health. She has held senior executive positions at
a major teaching hospital, a professional association, and Health
Canada. She was former Director of Publications and Editor-in-
Chief of the Compendium of Pharmaceutical and Specialties
(CPS), the book used in Canada by physicians, nurses, and other
health professions for prescribing information on medication.