Modeling The Acoustic Noise From A Wave Energy Converter Farm and Its Impact On Marine Mammals at The PacWave South Site, Offshore Newport Oregon
Modeling The Acoustic Noise From A Wave Energy Converter Farm and Its Impact On Marine Mammals at The PacWave South Site, Offshore Newport Oregon
Modeling The Acoustic Noise From A Wave Energy Converter Farm and Its Impact On Marine Mammals at The PacWave South Site, Offshore Newport Oregon
Renewable Energy
journal homepage: www.elsevier.com/locate/renene
Modeling the acoustic noise from a wave energy converter farm and its
impact on marine mammals at the PacWave South site, offshore
Newport Oregon
Jennifer L. Harding a, *, Leiph A. Preston a, Erick Johnson b, Jesse D. Roberts a, Craig A. Jones c,
Kaus Raghukumar c, Erin Hafla b
a
Sandia National Laboratories, Albuquerque, NM, USA
b
Montana State University, Bozeman, MT, USA
c
Integral Consulting, Santa Cruz, CA, USA
A R T I C L E I N F O A B S T R A C T
Keywords: Marine hydrokinetic devices, such as wave energy converters (WECs), can unlock untapped energy from the
Hydroacoustic modeling ocean’s currents and waves. Acoustic impact assessments are required to ensure that the noise these devices
Marine mammal impact generate will not negatively impact marine life, and accurate modeling of noise provides an a priori means to
Wave energy converters
viably perform this assessment. We present a case study of the PacWave South site, a WEC testing site off the
Acoustic impact metrics
coast of Newport, Oregon, demonstrating the use of ParAcousti, an open-source hydroacoustic propagator tool, to
Effective signal level
model noise from an array of 28 WECs in a 3-dimensional (3-D) realistic marine environment. Sound pressure
levels are computed from the modeled 3-D grid of pressure over time, which we use to predict marine mammal
acoustic impact metrics (AIMs). We combine two AIMs, signal to noise ratio and sensation level, into a new
metric, the effective signal level (ESL), which is a function of propagated sound, background noise levels, and
hearing thresholds for marine species and is evaluated across 1/3 octave frequency intervals. The ESL model can
be used to predict and quantify the potential impact of an anthropogenic signal on the health and behavior of a
marine mammal species throughout the 3-D simulation area.
1. Introduction Noise levels in the ocean are a growing concern, and anthropogenic
noise sources, such as shipping traffic, have been found to negatively
Marine Hydrokinetic devices (MHKs) are a promising source of impact marine life [8]. Moreover, the ocean acts as a waveguide [9,10]
renewable energy, converting ocean wave and current kinetics into causing sound to travel great distances and expanding the area of impact
electricity via wave energy converters (WECs) and current energy con for anthropogenic sound sources. While the noise produced by MHK
verters (CECs), respectively. The estimated marine energy resource in devices is less than that of shipping traffic by tens of dB and seismic
the United States that can be extracted by these technologies is 2300 airguns by ~100 dB (e.g. Ref. [10], for instance, there is a lack of direct
TWh/yr, which is 57% of the total generated electricity in the United observations of MHK noise and MHK array noise. It is therefore imper
States in 2019 [1]. Much of this untapped resource is from waves, ative to investigate MHK devices as a source of hydroacoustic noise and
especially on the western United States coast, which account for 1400 determine potential resulting impacts on local marine life by considering
TWh/yr alone [1]. Significant permitting hurdles for the installation of the devices’ unique acoustic signatures, site-specific marine environ
MHK devices exist, including the evaluation of the impact these devices mental properties such as bathymetry and sound speed (e.g. Ref. [9],
have on the surrounding marine life [2,3]. One such impact is the me and the presence of marine species endemic to an area (e.g., Ref. [11].
chanical noise that MHK devices produce, where the frequency content To predict how MHK noise propagates in marine environments, short
and amplitude of the acoustic noise varies both with the specific type of device installation and observation, we rely on predictions made from
and model of the device [4,5] and with environmental factors like sig hydroacoustic modeling. Modeling the acoustic noise produced by MHK
nificant wave height [6,7]. devices allows for the determination of potential marine life impact
* Corresponding author.
E-mail address: jlhardi@sandia.gov (J.L. Harding).
https://doi.org/10.1016/j.renene.2023.04.014
Received 6 December 2022; Received in revised form 10 March 2023; Accepted 4 April 2023
Available online 7 April 2023
0960-1481/© 2023 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
J.L. Harding et al. Renewable Energy 209 (2023) 677–688
Table 1
Descriptions of the inputs to and outputs of ParAcousti.
ParAcousti Definitions Definition Details This Case Study
678
J.L. Harding et al. Renewable Energy 209 (2023) 677–688
at any set of points within the 3-D domain. See section 2.1 for more
details on ParAcousti and its implementation.
Other studies have predicted sound levels due to anthropogenic noise
sources in the ocean [14–18], but many don’t consider the 3-D
complexity of the marine environment, multiple sources (i.e., a WEC
array), nor physically model sound as propagating acoustic waves as we
show here.
679
J.L. Harding et al. Renewable Energy 209 (2023) 677–688
Fig. 4. Cross-sections through the sound speed (a) and density (b) models along the green line in the simulation map in c) show the spatially variable ocean pa
rameters. Sound speed in a) is contoured every 1 m/s, density in b) is contoured every 0.25 kg/m3, and bathymetry is contoured every 5 m (black lines in (c)), with
WEC locations in all panels shown as red and yellow heptagrams. We choose a sandy seafloor with a sound speed of 1700 m/s and density of 1880 kg/m3, shown as a
sand color in a) and b). Cyan triangle shows the projected location in a), b), and c) of the receiver for the pressure time series in d).
680
J.L. Harding et al. Renewable Energy 209 (2023) 677–688
681
J.L. Harding et al. Renewable Energy 209 (2023) 677–688
Fig. 8. Top panels a) and b) show 2-D cross-sections through the 3-D SPL model for a N–S line (red line in Fig. 4c) and an E-W line (green line in Fig. 4c). Middle
panels show the depth-averaged and maximum SPL in the water column and bottom panels show the 1/3 oct. band depth-averaged SPL for the above SPL
cross-sections.
threshold, as described in Fig. 7. Since SNR and SnL are frequency- 3. Results and discussion
dependent, we first filter our pressure traces into 1/3 oct. bands,
compute 3-D SPL for each 1/3 oct. band, and compute SNR and SnL for 3.1. 3D sound pressure level
each band. To compute SNR, we use the observed background noise
measurements from Ref. [23] of the central Oregon coastal area as our The predicted 3-D SPL shows the highest SPLs nearest and above
noise level (Fig. 5), which we then subtract from our 1/3 oct. SPL models each WEC location, reaching 156 dB, and decreasing rapidly in the
to obtain a mean, 1 percentile (P1), and 99th percentile (P99) 3-D SNR lateral directions once outside of the array to ~100 dB–200 m away at
model for each frequency band. To compute SnL, we subtract the near-WEC depths and at ~400 m away at shallower ocean depths. At
hearing thresholds reported in Ref. [11] (Fig. 6) from our 1/3 oct. SPL distances greater than ~200 m outside the WEC array, SPLs decay more
models to obtain 3-D SnL models for each species group at each fre slowly to, on average, 92 dB–2.3 km away near the edge of the simu
quency band. Table 2 shows the species groups considered here, which lation domain. Fig. 8 shows two SPL cross-sections along with the depth-
have been grouped based on common hearing ability after [11]. While averaged and maximum SPL in the water column and depth-averaged 1/
all species groups may not be commonly present at the PacWave site 3 oct. SPL for each cross-section. The depth-averaged SPL is defined as
area, we have included them all for demonstrative purposes. the SPL computed using the arithmetic mean of the root-mean-squared
We introduce a new AIM that combines SNR and SnL into an effective pressure (Prms) in the water column, and the maximum SPL is the SPL
signal level (ESL) to determine if a signal has the potential to be detected computed using the maximum Prms in the water column, noting that as
and is frequency-, location-, and species-dependent. ESL, as outlined in the water column increases and decreases in thickness there are more or
Fig. 7, is set equal to the lesser of the two AIMS SNR and SnL for each 1/3 fewer receivers, respectively.
oct. band. That is, the ESL is equal to SnL if the hearing threshold is The SPL results demonstrate how a particular WEC array geometry in
higher than the background noise level (i.e. the background noise is not a realistic marine environment, where the density and sound speed of
detectable) and equal to the SNR if the background noise level exceeds the water and the bathymetry spatially vary, produces a unique 3-D
the hearing threshold (i.e. the background noise level is detectable). If acoustic environment. An example of bathymetry causing a local
the ESL is greater than zero, then a particular species group at a certain sound heterogeneity can be observed in the SPL model cross-section in
location in a specific 1/3 oct. frequency band can detect the signal. If the Fig. 8b,d where a small canyon likely causes a focusing of sound and a
ESL is less than zero, then the species group cannot detect the signal at local increase of SPL by several dB ~1 km east of the array towards the
the frequency and location. shoreline. This SPL model also shows how the WEC noise quickly
682
J.L. Harding et al. Renewable Energy 209 (2023) 677–688
From the computed 3-D SPL fields we compute 3-D SNR for each 1/3
oct. frequency band and three background noise percentiles as well as 3-
D SnL for each 1/3 oct. frequency band and species group. Fig. 9 shows
an E-W cross-section through the 3-D SNR field at the 100 Hz-centered
frequency band, which is calculated by subtracting the mean, P99, and
P1 background noise level values from the 100 Hz 1/3. oct. SPL. The
depth-averaged SNR plot in Fig. 9d shows that the mean SNR falls below
zero ~1.5 km from the edge of the WEC area. For the P1 background
noise case, representing the quietest conditions, the SNR does not fall
below zero anywhere in the simulation domain, whereas the P99 SNR
representing the loudest background conditions, falls below zero ~100
m outside the WEC array. Fig. 10 shows the same E-W cross-section
through the 3-D SnL field at 100 Hz for each species group, which are
calculated by subtracting each species group’s hearing threshold at 100
Hz from the 100 Hz 1/3 oct. SPL. The LF and PW groups are sensitive to
the 100 Hz signal throughout the cross-section, whereas the HF, VHF,
and SI groups have a depth-averaged SnL below zero for the whole cross-
section area.
Plan view maps of the depth-averaged SNR and SnL around the WEC
array visualize what can be heard by marine species within different 1/3
oct. frequency bands. Fig. 11 shows the depth-averaged SNR in the water
column for the entire simulation domain for three background noise
states at the 100 Hz 1/3 oct. frequency band. For mean background
noise conditions, the 100 Hz depth-averaged SNR drops below zero ~1
km outside of the WEC array. For the P1 quietest background noise
conditions, however, our simulation domain is not large enough to
capture where the 100 Hz depth-averaged SNR drops below zero and is
~10 dB along the edges of the simulation domain. For the P99 loudest
background noise conditions, the 100 Hz depth-averaged SNR is only
above zero ~100 m around each individual WEC. Using maximum SPL
values in the water column (e.g., red line in Fig. 8c and d) to calculate
SNR maps is also possible, which generally increases SNR values by 1–2
dB away from the WEC array, increasing the estimate of maximum
distance at which the signal is above background noise by ~500 m for
the mean background noise case.
The depth-averaged SnL in the 100 Hz 1/3 oct. frequency band for
each species group is shown in Fig. 12. These maps exemplify how the
differing hearing thresholds for each species at the same 1/3 oct. hearing
band affect the resultant SnL. The LF group, for instance, has a lower
hearing threshold and higher 100 Hz depth-averaged SnL values than
Fig. 9. E-W cross-sections through the 1/3 oct. 100 Hz band 3-D SNR fields the SI group, which has a higher hearing threshold and overall lower
using the mean (a), 1 percentile (P1) quietest (b), and 99th percentile (P99) 100 Hz depth-averaged SnL values. The LF and PW groups are the only
noisiest (c) background noise measurements from Ref. [23]. The two species groups for which the 100 Hz depth-averaged SnL does not
depth-averaged SNR for each cross-section in the water column for the above fall below zero within the simulation domain. For the rest of the species
cross-sections are shown in (d). groups, the 100 Hz depth-averaged SnL falls below zero within ~100 m
from each individual WEC. Maps of the depth-average SnL for the LF
decreases away from individual WECs, with maximum and mean SPL species group across the nine 1/3 oct. frequency bands modeled are
sharing similar values at the domain extents, ~93 and ~92 dB, shown in Fig. 13, which demonstrate how the WEC noise source levels
respectively. and LF hearing thresholds vary across frequency bands and contribute to
The variation in the peak SPL nearest each WEC is a result of the unique depth-averaged SnL maps.
receiver grid locations not directly overlapping with the WEC array The ESL is computed by choosing the lesser of the SNR and the SnL
spacing. In Fig. 8c, the depth-averaged and maximum SPL in the water AIMs at each 1/3 oct. band for each species group and background noise
column near the leftmost (southmost) WEC is much higher than the combination and provides one metric that predicts whether a species can
other WECs because there is a receiver much closer (1 m distance) to this detect the WEC array signal in the 3-D simulation domain. All species
WEC than the others. While every WEC is producing the same 1 m level groups are found to have a positive ESL and detect the WEC signal within
of sound of 156 dB, the SPL decreases quicky enough from each indi 40 m distance directly above a WEC device. Fig. 14 shows an example
vidual WEC that the 50 m receiver spacing doesn’t capture these sharp calculation of the maximum ESL in the water column for the LF group at
changes nor the true SPL peaks nearest every WEC. This is a potential the 50 Hz band and P1 quietest background noise conditions. Given that
pitfall where true peaks could be missed during modeled or real data the hearing threshold does not change in space and that we are using the
collection. Potential ways to overcome this pitfall include increasing the same background noise levels throughout the simulation area, the ESL is
receiver grid spacing close to the source(s) or adding additional re always equal to either the SNR or the SnL in the entire domain for this
ceivers at 1 m distances from the source(s). case study. If, however, the background noise levels varied in space,
then the ESL would potentially be equal to the SNR in certain parts of the
simulation domain and SnL in other parts.
683
J.L. Harding et al. Renewable Energy 209 (2023) 677–688
Fig. 10. E-W cross-sections through the 100 Hz 1/3 oct. band SnL fields for each species group (a–g) and the depth-averaged SnL in the water column for those model
cross-sections (h).
Fig. 11. Maps of depth-averaged SNR in the water column in the 100 Hz 1/3 oct. frequency band for mean background noise conditions (left), 1 percentile (P1)
quietest background noise conditions (middle), and 99th percentile (P99) loudest background noise conditions.
The ESL models complement and contextualize the SPL results by certain background noise conditions, and therefore where in space a
showing where in the simulation domain the WEC array signal is marine mammal has the potential to respond to, or change behavior due
detectable to a certain species group in a given frequency band with to, the WEC array noise. 2-D ESL maps can be produced by using the
684
J.L. Harding et al. Renewable Energy 209 (2023) 677–688
depth-averaged SPL, signifying the noise level a marine mammal might background noise conditions. Through modeling, such as shown here,
hear, or by using the maximum SPL, which might be useful for assessing the ESL near each WEC can be computed for each 1/3 oct. frequency
the loudest signal a marine mammal could experience and potentially band, and if just one of the ESL values across the full hearing range of the
respond to. The ESL can then be further evaluated and compared to species is above zero, then the signal is potentially detectable and the
observations of marine mammal behavior in different sound conditions devices could be avoided. If every ESL value across the full hearing range
to link ESL predictions to behavioral risks, such as is presented in for a species is less than zero, then the WEC device sound is undetect
Ref. [11]. able, and with the loss of one sensory input there is a greater potential
For many combinations of species groups, frequency bands, and for physical interaction. Fig. 15 shows an example of the ESL computed
background noise levels, our simulation domain fully captures the areas from a value of SPL at ~30 m depth above the southmost WEC source in
where the ESL is greater than zero, and therefore, where the WEC array the N–S cross-section shown in Fig. 8a for all frequency bands modeled
signal in the water column is detectable and where it is not. Some here and for two species groups at different background noise levels.
combinations of species groups, frequency bands, and background noise Despite not including higher frequencies in this study, the WEC noise at
levels, however, yield ESL models that don’t fall below zero within the this depth can be detected by all species groups considered, even the
simulation domain, and thus our present simulation doesn’t provide VHF group with poorer hearing ability at lower frequencies. At shal
spatial limits on where a species can detect the signal. For example, the lower depths where the SPL above each WEC source is smaller, however,
low-frequency cetacean (LF) group can hear the WEC array noise on the ESL for the VHF group is not positive at any of the 1/3 oct. bands
average in the water column, throughout the entire simulation area for modeled in this case study, but would likely be positive if higher fre
1/3 oct. frequency bands above 25 Hz and in P1 quietest background quencies were included in the WEC source.
noise conditions. For mean and loudest background noise conditions at A limitation of this case study in replicating a realistic scenario stems
frequency bands above 25 Hz, the ESL is equal to the SNR, which falls from the source time function we use, which is associated with a
below zero within the simulation area. For the quietest background particular significant wave height (1.5 m), does not incorporate varia
noise conditions, however, both the SNR and SnL fields above 25 Hz, and tions in operational conditions or signal phase across the WEC array, and
thus the ESL models, maintain positive values at the edges of the is capped at 198 Hz to manage the number of grid cells and reduce the
simulation area. necessary computational resources. The sound that each WEC emits will
Since the required domain size may not be known a priori and larger vary with significant wave height, has higher frequency content up
simulation domains that fully capture the positive ESL area may be wards of 1 kHz, and realistically, will have variability in phase compared
computationally infeasible for other studies, there are simple techniques to other WECs. Moreover, changes in weather can affect significant wave
to potentially address these shortcomings. One such technique is to height as well as background noise levels, especially at higher fre
linearly extrapolate the SPL model out to sufficient distances that cap quencies, which has not been considered in this case study. Further
ture the full positive ESL area, since the SPL has a fairly linear decay modeling needs to be conducted to better quantify these effects on the
away from the WEC array at distances greater than ~500 m (see Fig. 8c soundscape. Other limitations include the stationary nature of the WECs
and d). Another option is to port the pressure time series or SPL values in space, and future modeling efforts can incorporate the spatial motion
from the edges of the simulation domain into another modeling tool to of the acoustic noise source specific to the WECs device type.
simulate pressures or SPLs out to the desired distance from the WEC
array. 4. Conclusions
An important consideration in acoustic impact assessments is the
potential physical harm to marine mammals from interaction with a We presented a case study demonstrating the use of the ParAcousti to
WEC. For this reason, it is important to determine whether the WEC model the broadband acoustic noise generated by an array of 28 WECs at
signal can be detected on the whole for each marine mammal in various the PacWave south site off the coast of Newport, Oregon. Our simulation
685
J.L. Harding et al. Renewable Energy 209 (2023) 677–688
Fig. 13. Maps of depth-averaged SnL in the water column for the LF species group (phocids in water) across the nine 1/3 oct. frequency bands modeled in this study.
Color bar reflects the overall minimum and maximum depth-averaged SNR and SnL for all results (i.e. combinations of species groups, background noise levels, and
frequency bands).
domain utilized realistic sound speed, density, and bathymetry, all of marine species in a 3-D model space, depth-averaged map, or maximum
which affect how sound propagates through a marine environment and map. These values can then be referenced to observations of marine
impact the modeled acoustic environment. From the grid of simulated mammal behavior in response to an ESL, and can be contextualized
pressure over time, we computed 3-D broadband SPL, as well as 1/3 oct. within the marine environment’s importance to marine mammals (e.g.
SPL. Using these 1/3 oct. SPL fields, we computed two marine mammal feeding ground, breeding ground) to assess the signal’s risk to a marine
AIMs, SNR and SnL, to quantify how the simulated soundscape could mammal species. Moreover, the ESL can be used to determine whether a
affect different marine species. We also introduced a new AIM called the marine mammal species is able to detect a signal at all, which controls
ESL, which combines the SNR and SnL into one metric of species-specific whether an avoidance behavior is possible and the risk due to physical
perception potential that considers the background noise level, fre interaction with the device.
quency band, and a species’ hearing ability.
The SPL is highest above each WEC and rapidly decays laterally by
55 dB in ~200 m. At distances greater than 200 m away from the Declaration of competing interest
outermost WECs, SPL decays more slowly by 10 dB over ~1.5 km. We
found that all species groups can hear the simulated WEC noise within The authors declare that they have no known competing financial
40 m distance above an individual WEC, where the LF group experiences interests or personal relationships that could have appeared to influence
the loudest signal with an ESL reaching ~80 dB during quietest back the work reported in this paper.
ground noise conditions.
Following the methodology presented here, ParAcousti can be used Data availability
in acoustic impact assessments to quantify the signal detected by a
Data will be made available on request.
686
J.L. Harding et al. Renewable Energy 209 (2023) 677–688
Acknowledgements do so, for United States Government purposes. The DOE will provide
public access to these results of federally sponsored research in accor
The authors would like to thank Brandon Southall for his input and dance with the DOE Public Access Plan https://www.energy.gov/dow
guidance. This research was funded by the U.S. Department of Energy’s nloads/doe-public-access-plan.
Water Power Technologies Office. This paper describes objective tech
nical results and analysis. Any subjective views or opinions that might be References
expressed in the paper do not necessarily represent the views of the U.S.
Department of Energy or the United States Government. This article has [1] L. Kilcher, M. Fogarty, M. Lawson, Marine Energy in the United States: an
Overview of Opportunities, National Renewable Energy Laboratory, Golden, CO,
been authored by an employee of National Technology & Engineering 2021. NREL/TP-5700-78773, https://www.nrel.gov/docs/fy21osti/78773.pdf.
Solutions of Sandia, LLC under Contract No. DE-NA0003525 with the U. [2] R.S. O’Neil, G.J. Staines, M.C. Freeman, Marine Hydrokinetics Regulatory
S. Department of Energy (DOE). The employee owns all right, title and Processes Literature Review, Pacific Northwest National Laboratory, 2019, https://
doi.org/10.2172/1591886. PNNL-28608,.
interest in and to the article and is solely responsible for its contents. The [3] Pacific Energy Ventures, LLC, Handbook of Marine Hydrokinetic Regulatory
United States Government retains and the publisher, by accepting the Processes, United States Department of Energy, Water Power Technologies Office,
article for publication, acknowledges that the United States Government 2020. DOE-EE-1793.
[4] C. Frid, E. Andonegi, J. Depestele, A. Judd, D. Rihan, S.I. Rogers, E. Kenchington,
retains a non-exclusive, paid-up, irrevocable, world-wide license to The environmental interactions of tidal and wave energy generation devices,
publish or reproduce the published form of this article or allow others to
687
J.L. Harding et al. Renewable Energy 209 (2023) 677–688
Environ. Impact Assess. Rev. 32 (1) (2012) 133–139, https://doi.org/10.1016/j. [22] NOAA National Centers for Environmental Information, Central Oregon Coastal
eiar.2011.06.002. Digital Elevation Model NAVD 88, 2015. Retrieved from: https://www.ncei.noaa.
[5] O. Langhamer, K. Haikonen, J. Sundberg, Wave power—sustainable energy or gov/metadata/geoportal/rest/metadata/item/gov.noaa.ngdc.mgg.dem:11500/h
environmentally costly? A review with special emphasis on linear wave energy tml#.
converters, Renew. Sustain. Energy Rev. 14 (4) (2010) 1329–1335, https://doi. [23] J.H. Haxel, R.P. Dziak, H. Matsumoto, Observations of shallow water marine
org/10.1016/j.rser.2009.11.016. ambient sound: the low frequency underwater soundscape of the central Oregon
[6] G. Buscaino, G. Mattiazzo, G. Sannino, E. Papale, G. Bracco, R. Grammauta, coast, J. Acoust. Soc. Am. 133 (5) (2013) 2586–2596, https://doi.org/10.1121/
S. Mazzola, Acoustic impact of a wave energy converter in Mediterranean shallow 1.4796132.
waters, Sci. Rep. 9 (1) (2019) 9586, https://doi.org/10.1038/s41598-019-45926- [24] D.J. Hasselman, D.R. Barclay, R. Cavagnaro, C. Chandler, E. Cotter, D.M. Gillespie,
1. G.J. Staines, Chapter 10: Environmental Monitoring Technologies and Techniques
[7] K. Haikonen, J. Sundberg, M. Leijon, Hydroacoustic measurements of the noise for Detecting Interactions of Marine Animals with Turbines. In 2020 State Of the
radiated from wave energy converters in the Lysekil project and Project WESA, in: Science Report, Pacific Northwest National Lab, 2020.
1st International Conference and Exhibition on Underwater Acoustics, 2013, [25] A.B. Gill, I. Gloyne-Philips, J. Kimber, P. Sigray, Marine renewable energy,
pp. 199–209. UA2013 (Corfu, Greece, 23rd-28th June 2013). electromagnetic (EM) fields and EM-sensitive animals, Mar Renew Energy Technol
[8] C.M. Duarte, L. Chapuis, S.P. Collin, D.P. Costa, R.P. Devassy, V.M. Eguiluz, Environ Interact (2014) 61–79.
F. Juanes, The soundscape of the Anthropocene ocean, Science 371 (6529) (2021), [26] J. Tougaard, Underwater noise from a wave energy converter is unlikely to affect
https://doi.org/10.1126/science.aba4658. marine mammals, PLoS One 10 (7) (2015), e0132391, https://doi.org/10.1371/
[9] W. Kuperman, P. Roux, Underwater Acoustics, Springer Handbook of Acoustics, journal.pone.0132391.
2007, pp. 149–204. [27] C. Bassett, J. Thomson, B. Polagye, K. Rhinefrank, Underwater noise measurements
[10] National Research Council, ocean noise and marine mammals, in: Ocean Noise and of a 1/7th scale wave energy converter, in: Paper Presented at the OCEANS’11
Marine Mammals, The National Academy Press, Washington, D.C., 2003. MTS/IEEE KONA, 2011 (Waikoloa, HI, USA).
[11] B.L. Southall, D.P. Nowacek, A.E. Bowles, V. Senigaglia, L. Bejder, P.L. Tyack, [28] J.J. Finneran, Noise-induced hearing loss in marine mammals: a review of
Marine mammal noise exposure criteria: assessing the severity of marine mammal temporary threshold shift studies from 1996 to 2015, J. Acoust. Soc. Am. 138 (3)
behavioral responses to human noise, Aquat. Mamm. 47 (5) (2021) 421–464, (2015) 1702–1726, https://doi.org/10.1121/1.4927418.
https://doi.org/10.1578/AM.47.5.2021.421. [29] B.L. Southall, J.J. Finneran, C. Reichmuth, P.E. Nachtigall, D.R. Ketten, A.
[12] L.A. Preston, Paracousti User Guide. Sandia Report SAND20169291, 2016. E. Bowles, P.L. Tyack, Marine mammal noise exposure criteria: updated scientific
Albuquerque, NM, USA. recommendations for residual hearing effects, Aquat. Mamm. 45 (2) (2019)
[13] K.E. Buenau, L. Garavelli, L.G. Hemery, G. García Medina, A review of modeling 125–232, https://doi.org/10.1578/am.45.2.2019.125.
approaches for understanding and monitoring the environmental effects of marine [30] B.L. Southall, Evolutions in marine mammal noise exposure criteria, Acoust. Today
renewable energy, J. Mar. Sci. Eng. 10 (1) (2022), https://doi.org/10.3390/ 17 (2) (2021), https://doi.org/10.1121/at.2021.17.2.52.
jmse10010094. [31] W.W.L. Au, M.C. Hastings, Principles of Marine Bioacoustics, vol. 510, Springer,
[14] E. Hafla, E. Johnson, C.N. Johnson, L. Preston, D. Aldridge, J.D. Roberts, Modeling New York, 2008.
underwater noise propagation from marine hydrokinetic power devices through a [32] C.W. Clark, W.T. Ellison, B.L. Southall, L. Hatch, S.M. Van Parijs, A. Frankel,
time-domain, velocity-pressure solution, J. Acoust. Soc. Am. 143 (6) (2018) 3242, D. Ponirakis, Acoustic masking in marine ecosystems: intuitions, analysis, and
https://doi.org/10.1121/1.5039839. implication, Mar. Ecol. Prog. Ser. 395 (2009) 201–222, https://doi.org/10.3354/
[15] C. Erbe, D.M. Farmer, Zones of impact around icebreakers affecting beluga whales meps08402.
in the Beaufort Sea, J. Acoust. Soc. Am. 108 (3 Pt 1) (2000) 1332–1340, https:// [33] W.T. Ellison, B.L. Southall, C.W. Clark, A.S. Frankel, A new context-based approach
doi.org/10.1121/1.1288938. to assess marine mammal behavioral responses to anthropogenic sounds, Conserv.
[16] O.W. Ikpekha, F. Soberon, S. Daniels, Modelling the propagation of underwater Biol. 26 (1) (2012) 21–28, https://doi.org/10.1111/j.1523-1739.2011.01803.x.
acoustic signals of a marine energy device using finite element method, Rev Energy [34] K. Raghukumar, S. McWilliams, G. Chang, J. Roberts, C. Jones, Wave energy
Power Qual J (2014) 97–102, https://doi.org/10.24084/repqj12.246. converter arrays: optimizing power production while minimizing environmental
[17] K. Rossington, T. Benson, P. Lepper, D. Jones, Eco-hydro-acoustic modeling and its effects, in: Paper Presented at the Offshore Technology Conference, 2019.
use as an EIA tool, Mar. Pollut. Bull. 75 (1–2) (2013) 235–243, https://doi.org/ [35] E.P. Chassignet, H.E. Hurlburt, O.M. Smedstad, G.R. Halliwell, P.J. Hogan, A.
10.1016/j.marpolbul.2013.07.024. J. Wallcraft, R. Bleck, The HYCOM (HYbrid Coordinate Ocean Model) data
[18] V.L.G. Todd, L.D. Williamson, J. Jiang, S.E. Cox, I.B. Todd, M. Ruffert, Prediction of assimilative system, J. Mar. Syst. 65 (1–4) (2007) 60–83, https://doi.org/10.1016/
marine mammal auditory-impact risk from Acoustic Deterrent Devices used in j.jmarsys.2005.09.016.
Scottish aquaculture, Mar. Pollut. Bull. 165 (2021), 112171, https://doi.org/ [36] K.V. Mackenzie, Nine-term equation for sound speed in the oceans, J. Acoust. Soc.
10.1016/j.marpolbul.2021.112171. Am. 70 (1981) 807–812, https://doi.org/10.1121/1.386920.
[19] International Electrotechnical Commission, Marine Energy - Wave, Tidal and Other [37] S.K. Henkel, L.A. Gilbane, Using benthic macrofaunal assemblages to define habitat
Water Current Converters - Part 101: Wave Energy Resource Assessment and types on the northeast pacific sedimentary shelf and slope, Estuar. Coast Shelf Sci.
Characterization, 2015 (Retrieved from). 246 (2020), https://doi.org/10.1016/j.ecss.2020.107056.
[20] W.-C. Wu, T. Wang, Z. Yang, G. García-Medina, Development and validation of a [38] National Marine Fisheries Service, 2018 Revision to: Technical Guidance for
high-resolution regional wave hindcast model for U.S. West Coast wave resource Assessing the Effects of Anthropogenic Sound on Marine Mammal Hearing (Version
characterization, Renew. Energy 152 (2020) 736–753, https://doi.org/10.1016/j. 2.0): Underwater thresholds for onset of permanent and temporary threshold shifts,
renene.2020.01.077. NOAA Technical Memorandum, 2018. NMFS-OPR-59.
[21] B. Robertson, G. Dunkle, J. Gadasi, G. Garcia-Medina, Z. Yang, Holistic marine [39] S.B. Martin, C. Morris, K. Broker, C. O’Neill, Sound exposure level as a metric for
energy resource assessments: a wave and offshore wind perspective of metocean analyzing and managing underwater soundscapes, J. Acoust. Soc. Am. 146 (1)
conditions, Renew. Energy 170 (2021) 286–301, https://doi.org/10.1016/j. (2019) 135, https://doi.org/10.1121/1.5113578.
renene.2021.01.136.
688