Exposure experiments with cod and continuous sound

In the SpawnSeis project, which started in 2018, we aimed at studying the effects of seismic airgun shooting on wild, free ranging, spawning cod. However, in recent years, the project has expanded to include other sources of anthropogenic sounds, as well as other periods than just the spawning period. Cod behaviour is studied using an acoustic telemetry grid in Austevoll, Western Norway. This set up has proven to be a very good experimental site for studying the effect of anthropogenic sounds on spawning cod (McQueen et al. 2022; 2023; 2024) as well as giving new insight into basic biology and spawning behaviour of wild cod (Skjæråsen et al. 2024)

In autumn 2018, a total of 36 acoustic receivers were placed in two telemetry grids, on two separate cod spawning grounds. These act as one exposure area (30 receivers) and a smaller reference area (6 receivers) (Figure 1). Single telemetry receivers were also placed in northern and southern exit routes from the larger telemetry grid, as well as on three nearby spawning sites to document potential use of multiple sites during a season. Additionally, a curtain of three telemetry receivers has been deployed to cover the western exit route from the study area, and several single receivers were deployed north of the grid to increase detection coverage in this area. Over the years, the reference area has been removed, and the receivers here used to replace lost ones in the main grid. Due to occasional losses of receivers, the number and position of receivers deployed has varied throughout the study period. In particular, the single receivers at locations F, G and H are gone, and there have been a reduced number of receivers at locations D and E in recent years.

About 50 cod have been tagged annually with acoustic transmitter tags. Tagging was performed in January in 2019, 2020, 2021, 2022 and 2024, in December 2022, in April and June 2025, and in February 2026. In most years, mainly mature, spawning fish were tagged, but also a few immature individuals. The tag batteries last for about 2 years, so fish from different tagging cohorts may be detected in the grid at the same time.

The movement of these fish can be tracked, vertically and horizontally. This allows us to study potential changes in behaviour such as vertical and horizontal avoidance, changes in home range and activity level in response to sound exposures. Two experiments with seismic air guns (Sivle et al. 2021) one with a marine vibrator (Sivle et al. 2022) and one with a sparker source (Sivle et al., 2024) have been conducted in this bay and results have shown that air gun sound exposure levels (SEL) up to 145 dB re 1 uPa²s does not cause cod to move away from the area during the spawning period (Mcqueen et al., 2022), and changes in behaviour onsite are rather minor (McQueen et al. 2023). The marine vibrator exposure also did not cause fish to leave the test site, but changes in swimming depth and activity levels were observed (McQueen et al., 2024). The sparker source did not result in any changes in behaviour (McQueen et al., 2025)

Low frequency noise from passing vessels may affect both behaviour and vocalization of fish (e.g. Stanley et al. 2017, van der Knaap et al. 2022) and may be a problem in particular in coastal areas with much traffic as well as along shipping lanes, and in areas used for various offshore industry such as oil rigs and wind farms. Offshore wind farms themselves also produce low frequency, continuous noise and tones that may potentially disturb fish. Therefore, it is important to obtain a better understanding of how such noise affect fish abundance, behaviour and vocalization.

The abovementioned experiments with airgun and a marine vibrator were conducted on adult, spawning fish in the spawning period. Those fish may be highly “fixed” on their ongoing spawning activity, with the potential cost of a reaction being missed spawning opportunities. Therefore, a reaction may be more easily triggered at other times of the year, e.g. during the feeding period in summer. Disturbances during feeding activity also have the potential to lead to population-level consequences (Soudijn et al. 2020), so are also important to understand. In 2023, 2024, 2025 and 2026 we have therefore conducted experiments as identical as possible during winter spawning (2024 and 2026) and summer feeding periods (2023 and 2025).

Four identical experiments were conducted in Bakkasund, June 2023, February 2024, July 2025, and February 2026. Each experiment was conducted with 3x3 blocks of 1 hour. Each block consisted of 3 treatments;

1) Vessel noise (vessel at anchor, with engine running)

2) Tone (playing tones at 100 Hz, hence representing a frequency likely to be found at an operating wind farm) and

3) Silent (vessel at anchor, all engines off).

2.1 Operation area

The experiment was conducted in Bakkasund, a cod spawning area in Austevoll in western Norway.

Placement of the source vessel was in the middle of the bay. This position was chosen to give the best possible exposure range and coverage of the tagged fish inside the bay (see and detailed explanation in Appendix 1).

Hydrophones were placed in three different locations, spread out in the bay to record the received sound level at various locations where fish are distributed. Hydrophone locations were chosen by based on locations of where fish were frequently detected in years prior to the first experiment (Figure 2). For consistency, the same locations were used for all experiments.

2.2  Vessel

The experiment in summer 2023 was conducted with the vessel “MV Sjåsund”, operated by Sjåsund Marine (Figure 3a). The vessel is a catamaran with length of 15 m and with a crew of 3. The vessel operates e.g. at fish farms and is representative of medium sizes vessels normally found in the area.

The experiment in winter 2024/2026 and summer 2025 was conducted with the research vessel “Hans Brattstrøm”, operated by the Institute of Marine Research (Figure 3b). This vessel is 24 m with a crew of 2 and of similar size of fishing and recreational vessels in the area.

2.3 Equipment

Sound source and Hydrophones

A low-frequency sound projector/underwater loudspeaker was used to emit a low-frequency (100 Hz) tone. The system consisted of a PC used as a signal generator, a low-frequency submersible acoustic projector (C-Bass M72-110, supplied by GeoSpectrum Technology Inc. (GTI), Canada), a power supply, and an M620 power amplifier supplied by GTI (Figure 4).

The emitted tone centred at 100 Hz, representing a frequency range likely emitted by offshore wind turbines and for consistency with other experiments with fish larvae (Cresci et al. 2023).

This is also within the range of the best performance of the speaker (Figure 5).

The sound source was lowered from the stern of the vessel at a depth of 5 m. In 2023 it was deployed directly from the stern, while in 2024, 25 and 26 it was deployed from the ships crane 4 m distance from the stern. Source level at close range was monitored by a hydrophone (Ocean Sonics icListen SC35-ETH) lowered to the same depth as the source. The position of the icListen changed somewhat both between years and blocks within a year, due to practical reasons and optimalizations to avoid placement next to other sound sources at the vessel (water outlets etc). The detailed set up for source and icListen hydrophone are shown in Figure 6 for 2023, 2024 and 2025 and in Figure 7 for 2026.The Ici Listen hydrophone For Block 1 (and all of 2025), the hydrophone was too close to the aggregate water outlet (aggregate used in the silent periods). On the port side, engine operation caused water outlet issues, and for Blocks 2–4, the hydrophone was deployed from the bow, which moved strongly with the waves. The port side was eventually judged best and used for the remaining blocks.

Sound exposure in the bay was recorded with hydrophones of type SoundTrap ST300HF from Ocean Instruments, rigged as shown in Figure 8 to avoid flow noise in years 2023, 24 and 25. In 2026 the rigging was different due to deployment challenges, and the rigging used in 2026 are shown i Figure 9.

Placement of the hydrophones in the bay were based on known aggregations of fish (shown in Figure 2). The specific placements in all years are shown in Figure 11 .

CTD

A Conductivity-Temperature-Depth (CTD) instrument was used to make a profile of the temperature of the water column. This allows to make a sound speed profile in the area as well as locate the depth of sound channels. A SAIV SD204 (Figure 10) sonde was used in summer 2023, while the onboard CTD instrument, at Hans Brattstrøm, a RBRmaestro³, was used in winter 2024, 2026 and summer 2025.

2.4 Exposure design

The exposure design followed the design used for the airguns, marine vibrator and sparker experiments by having a block design with each block consisting of three treatments: the sound source, a vessel control and a silent control. In the previous experiments however, the vessel control has been used to separate the effect of the sound source from that of the vessel, since the sources were towed from a moving vessel hence exposing the fish to both the source and the vessel sound. In the current experiment however, with the vessel at anchor, the engine could be turned off during the sound source exposure. Further, in the previous experiments, the vessel and source were moving at the outside of the bay, hence the vessel noise was relatively weak inside the bay compared to the sound from those sound sources. In this experiment, with a source of lower source level, and the source level placed in the middle of the bay, the vessel sound was therefore considered a source itself; representing a low frequency, continous broadband noise source. The sound source emitted a single frequency (100 Hz) tone, allowing comparison by the difference in response of the fish to these sound types.

The three treatments were:

• 1h of active transmission with 100 Hz tone from the C-BASS speaker, all engines off (tone)

• 1h with the vessel running all engines (vessel)

• 1h without the engines or speaker (silent)

The order of these treatments within the blocks was randomized prior to starting. A schedule of 10 blocks was made prior to each of the experiments. However, in 2023 the source was damaged and in 2024 the weather limited the number of available days. Hence, in these years only 6 blocks were conducted. In 2025 and 2026 10 blocks were conducted.

2.5 Fish telemetry data

The data from the fish telemetry receivers were downloaded in the periods between experiments, usually once or twice per year. The fish tags transmit a unique ID. The telemetry receivers store the information transmitted by the tags when the fish is within range. The detection logs on the receivers therefore give an indication of fish presence near the receiver over time. The detection dataset is filtered to exclude likely false detections (single detections within a day).

As the tags also contain pressure and accelerometer sensors, they additionally transmit information on the swimming depth and activity levels of the fish.

Due to the dense grid of receivers inside the Bakkasund bay (the test site), the same tag transmission can be recorded on more than one receiver. This allows for estimating the position of the tagged fish, using hyperbolic positioning. This is done using a specialised software (Fathom Position, Innovasea). The positions are estimated together with an estimate of error, and the most uncertain positions (with the highest error estimates) are filtered out of the dataset.

In summer 2023, three days with a total of 6 blocks were conducted from June 14^th to June 16^th. In winter 2024, three days with a total of 6 blocks were conducted from Feb 27^th to March 1^st. In summer 2025, 4 days with a total of 10 blocks were conducted, from June 30^th to July 3^rd. In winter 2026 4 days with a total of 10 blocks were conducted, from February 16^th to 19^th.

3.1 Experiments

In summer 2023, during the tone treatment in block 6, the source had got a crack, and was leaking, thus not performing optimally, and could not be used any further. Therefore, only 6 blocks were conducted in total (Table 1 ).

In winter 2024, one of the four days with the vessel (Feb 29^th) the weather was not allowing us to operate. Hence, 6 blocks were conducted in total (Table 2).

In summer 2025, a total of 10 blocks were conducted as planned, shown in Table 3.

In winter 2026, a total of 10 blocks was conducted, shown in Table 4.

We aimed to have the hydrophones and source vessel at predetermined locations (Figure 2). The vessel was at anchor but was drifting somewhat due to wind. However, to not disrupt the blocks by engine noise, we decided not to move the vessel between runs. Vessel position was noted at start of each run. Positions of source vessel for the three years/seasons are shown as filled circles in colours representing the various years in Figure 11 .

3.2  Sound monitoring

Sound from the 100 Hz tone as well as from the vessel was monitored at three different sites in the bay with sound trap hydrophones, and at close range with an icListen hydrophone. Deployments and details of these are shown in Table 3 .

Hydrophone recordings

The tone, with characteristics as measured from the centre bay hydrophone shown in Figure 12, was a continuous sound over 1 hour (Figure 12 a), had its the main energy at 100 Hz, but some harmonics, mainly at 200 and 300 Hz (Figure 12 b, c).

Recordings from all three hydrophones show that the 100 Hz tone was clearly audible at all three hydrophone locations, here shown with examples from 2023 (Figure 13) and 2024 (Figure 14), hence could likely be detected by cod throughout the entire bay. Vessel noise is however not clearly distinguished from the silent periods.

For all exposure blocks, Sound Exposure Level (SEL) was calculated for all three treatment types (Tone, Silent, Vessel) at each of the hydrophone locations.

The sound pressure was recorded at a sampling rate of 48 kHz for the SoundTrap and 16 kHz (2023-2025) or 64 kHz (2026) for the icListen. Pressure data were band-pass filtered between 5 and 1000 Hz using a third-order zero-phase Butterworth filter. Ten-second sound exposure levels, SEL (ISO 18405:2017), were then calculated by time-integrating the squared pressure over 10 s windows and using a 10log ₁₀ transformation. SEL was computed every 10 s with a 9 s overlap between successive windows.

Sound Exposure Level (SEL, 10 sec) for all three treatments at the four hydrophones are shown for the different years in Figure 15 (2023), Figure 16 (2024), Figure 17 (2025) and Figure 18 (2026).

As seen from figures 15-18, the tone sound was clearly higher than the ambient/silent and vessel noise at the centre and inner positions, naturally since these were closest to the source. At the side bay and outer positions, the tone sound was above the vessel noise in all year except 2026, where the sound level seem to have been somewhat reduced compared to the previous years. Vessel noise is clearly above the ambient (silent) noise in center and inner positions, while at outer and side positions more or less falls into the ambient noise.

For all blocks, the median SEL (over 10 sec) was calculated. As an example, the results from the tone exposure for all blocks in 2026 measured at the centre position is shown in Figure 19, and as can be seen, there are variations between blocks. The median SEL (10) s was estimated for each block by taking the median of the linear sound exposure for all the 10 second periods for each block and then converting back to dB re 1 uPa²s. The center/vessel position measurements for 2026 were especially variable due to the movement of the hydrophone between the blocks (see Figure 9). Table 4 shows the highest (max) and lowest (min) of these values for each hydrophone position for each year. Sound exposure levels at the center bay, hence closest to the source varies from 159. 8 to 174.6 dB re 1µPa²s. Variations occur both between blocks every year, as well as between years. This may to some extent be due to propagation differences, but most likely most of it are caused by slight differences in the geometry; the vessel was not at the exact same spot at all times (drifted slightly at anchor, and not exactly at the same start position every time) as well as the hydrophone positions were slightly different between years (Table 3, Figure 11). The lowest values were recorded in the side bay position, ranging from 102.1 6 dB re 1µPa²s, which is more or less similar as the ambient noise, and up to 125.7 dB re 1µPa²s at the most (Table 4). For all years except 2023, the source seem to produce a tone with some cyclic variability in sound pressure level, e.g. very clearly seen in Figure 16 . We are not sure about the reason for this. This may be due to the directivity of the source and its motion in the waves.

3.3  Hydrographic observations

In summer 2023, CTD measurements were taken at the source vessel positions every day, several times. Additionally, a CTD was taken at all the hydrophone positions during day 3 of the experiment (June 16^th). Positions and times for the CTD casts are shown in Table 5 and Figure 20 .

In winter 2024, CTD casts were taken at each of the hydrophone positions after deployment and retirement, as well as at the source vessel position every day. CTD positions are shown in Table 5 and Figure 20 .

In summer 2025, CTD casts were taken at each of the hydrophone positions after deployment and retirement, as well as at the source vessel position every day. CTD positions are shown in Table 5 and Figure 20 .

In winter 2026, CTD casts were taken at each of the hydrophone positions after deployment and retirement, as well as at the source vessel position every day. CTD positions are shown in Table 5 and Figure 20 .

Results of the CTD casts are shown in Figure 21 (2023), Figure 22 (2024), Figure 23 (2025) and Figure 23 (2026).

As can be seen from the profiles, there is a clear difference between the profiles in summer and winter. In winter, sound speed increase with depth, and are highest towards to bottom, while in the summer there is a decrease with increasing depths and sound speed are highest close to the surface. This is due to temperature differences in the profiles for the various seasons. No clear sound speed channels can be seen in any of the years.

3.4  Behavioural observations

Fish presence

The number of tagged fish detected inside the test site (inside the bay, Bakkasund), varied between experiments, with 10 fish detected in the test site during the summer 2023 exposure, 26 detected during the winter 2024 exposure, 15 during the summer 2025 exposure and 33 during the winter 2026 exposure. This includes fish that were detected continuously during the exposures, and fish that were only detected intermittently or occasionally. If a fish stops being detected on receivers in the test site during the exposure, this may indicate that the fish left the area. In summer 2023 there was 1 fish that possibly left the bay during the exposure period, in winter 2024 none of the detected fish left the bay during the exposure, in summer 2025 1 fish appeared to leave the bay early during the exposure and return at the end of the exposure period, and in winter 2026 three fish potentially left the site during the exposure. During all exposures, several of the tagged cod were also detected at the gate areas outside of the main test site (Figures 25-26).

Comparing the number of fish detected within the grid per day with previous years without exposures does not reveal any obvious decrease in fish presence during the exposures, which could be indicative of fish leaving the area due to the exposure (Figure 27-28).

Depth and activity

There were quite some individual variations in swimming depth and acceleration values both during summer (example for 2023 shown in Figure 29) and winter (example for 2024 shown in Figure 30). For example, some fish have very clear diel vertical migrations (e.g. fish ID 1542169 in Figure 29), while other do not (e.g. fish ID 1400439 in Figure 29). For both swimming depth and accelaration, more in-depth analyses will be conducted at 2 levels; 1) Before-During After (BDA), comparing the full exposure period (all days with exposure) to similar periods of days before and after (Table 6) to explore more long term effects and 2) Treatment level analysis, to see how these change in response to the various treatments (vessel-silent-Tone). Figures 31-34 show the depth and acceleration of all fish combined for both the BDA and Treatment level periods. It is difficult to detect differences between treatments or experimental phases from simply plotting the data, as there is a large amount of individual variation. Statistical analysis will be conducted to disentangle potential differences related to the exposure from other sources of variation. Note that a change of settings of the tags used in 2024 and 2025 means that acceleration values of up to 4.9 m s ^-2 could be recorded in those years, but in the other years the upper limit that could be recorded by the tags was 3.5 ms ^-2 .

Fish positions

Fish were distributed throughout the bay during the first two experimental periods, in 2023 and 2024 (Figures 33-34). The positions of the fish have not yet been estimated for the last two experiments, in 2025 and 2026. Potential changes in space use and horizontal movements will be investigated using the position data, similar as has been done by McQueen et al. 2023.

Cresci A, Zhang G, Durif CMF et al. Atlantic cod (Gadus morhua) larvae are attracted by low-frequency noise simulating that of operating offshore wind farms. Commun Biol 2023; 6 (1):art. 1. https://doi.org/10.1038/s42003-023-04728-y.

McQueen K, Meager JJ, Nyqvist D et al. Spawning Atlantic cod (Gadus morhua L.) exposed to noise from seismic airguns do not abandon their spawning site. ICES Journal of Marine Science 2022;79(10). https://doi.org/10.1093/icesjms/fsac203.

McQueen K, Skjæraasen JE, Nyqvist D et al. Behavioural responses of wild, spawning Atlantic cod ( Gadus morhua L.) to seismic airgun exposure. ICES Journal of Marine Science 2023;80(4):1052–65. https://doi.org/10.1093/icesjms/fsad032.

McQueen Kate, Sivle LD, Forland TN et al. Continuous sound from a marine vibrator causes behavioural responses of free-ranging, spawning Atlantic cod (Gadus morhua). Environmental Pollution Jan. 2024:123322. https://doi.org/10.1016/j.envpol.2024.123322.

McQueen Kate, Sivle LD, Khodabandeloo B et al. Free-ranging Atlantic cod did not change their behaviour in response to a sparker seismic sound source. Marine Environmental Research 2025;210:107254. https://doi.org/10.1016/j.marenvres.2025.107254.

Sivle L, McQueen K, Khodabandeloo B. Survey report for testing a sparker source on cod behaviour. (SPAWNSEIS SPARKER) Toktrapport 2024-16. ISSN: 1503-6294. Havforskningsinsituttet.

Sivle LD, Forland TN, De Jong K et al. pawnSeis MV Exposure Experiment - Survey Report , Toktrapport 2023-4. ISSN: 1503-6294. Havforskningsinsituttet: Havforskningsinsituttet, 2022. https://www.hi.no/hi/nettrapporter/toktrapport-en-2023-4.

Sivle LD, Handegard NO, Kvadsheim PH et al. Cruise report_SpawnSeis_2021 , Toktrapport. Toktrapport Nr.9-2021. ISSN:1503-6294: Havforskningsinsituttet, 2021. https://www.hi.no/hi/publikasjoner/toktrapporter/2021/spawnseis-seismic-exposure-experiment-on-free-ranging-spawning-cod-nr.-9-2021.

Skjæraasen JE, Olsen EM, McQueen K et al. Sex-specific vertical movements of spawning atlantic cod in coastal habitats inferred from acoustic telemetry. Sci Rep 2024;14(1):23242. https://doi.org/10.1038/s41598-024-74896-2.

Soudijn FH, Kooten T van, Slabbekoorn H et al. Population-level effects of acoustic disturbance in Atlantic cod: a size-structured analysis based on energy budgets. Proceedings of the Royal Society B 2020;287(1929):20200490.

Stanley JA, Van Parijs SM, Hatch LT. Underwater sound from vessel traffic reduces the effective communication range in Atlantic cod and haddock. Scientific Reports 2017;7(1). https://doi.org/10.1038/s41598-017-14743-9.

van Der Knaap, I., Ashe, E., Hannay, D., Bergman, A. G., Nielsen, K. A., Lo, C. F., and Williams, R. 2022. Behavioural responses of wild Pacific salmon and herring to boat noise. Marine Pollution Bulletin, 174: 113257

 A1.1 Considerations on placement of sound projector

Where to place the hydrophones and sound projector depended on different factors; where the tagged fish are likely to be during the exposure period, where it is logistically and weatherwise best to operate, and where it is likely to get good and representative recordings.

For the winter experiment, it occurred relatively close after the tagging, and at the same time as previous experiments in the spawning season, and therefore the cod were assumably numerous and well spread over the area.

During summer however, a part of the tagged fish may have disappeared, and we did not have good insight into their distribution. We therefore did some mapping of the tagged fish from previous years in May/June. Based these previous years, it seemed likely that between 9 -13 fish would be present in Bakkasund in June (data from 2019-2021). Fish positions in June during the years 2019-2021 are shown in figure A1.1. The data from June 2022 was not available until after the summer 2023 exposure experiment, and therefore the data from end of May 2022, was used, shown in figure A1.2.

Both in 2020 and 2021, as well as in May 2022, the area around receiver number 7 seem to be an area occupied by tagged fish. This is also an area in the middle of the bay, ensuring ensonifying a good part of the bay.

The source has a sound source level of about 180 dB re 1µPa. With a very simple spherical spreading model (transmission loss = 20log(R)), the sound level will be 160 dB re 1µPa after 10 m, and around 145 dB re 1µPa after 35 m. Cod communicate with a sound level at approximate 110 dB re 1µPa (Stanley et al. 2017). With these simple calculations, the exposure will be louder than the communication up to a distance of for at least a km, likely about 1.5 km.

The source will be placed close to receiver 7. This is a place where at least in 2020/21/22 there was quite a lot of fish, as well as being a position central in the bay, where the sound will be spread and ensonify the most central parts of the bay to levels exceeding the communication level. A basic plot of the area where the sound should be loader than the communication level is shown in figure A1.3.

Based on recordings of tagged fish present in Bakkasund in June previous years, their depth distribution varies between 9 and 18 m swimming depth. It is therefore important to ensure that this depth is actually ensonified properly.

Further, it is important to get below the thermocline to ensure that the sound is projected into the full depth range of the cod.

Figure A1.4 show the sound speed profiles measured in June 2021, showing a thermocline between 10 and 20 m.

The source will therefore be deployed at a depth of 5 m.

A1.2 Considerations on placement of hydrophones

During the previous exposure experiments with air guns and marine vibrator, hydrophones have been positioned in the inner, middle and outer part of the bay (see McQueen et al., 2022). However, since we are now emitting the sound from a different position, these placements may not be optimal. In the current experiment, we aimed at placing the hydrophones in areas were fish are likely to be found, to measure the sound levels experienced by the fish. However, it is also of interest to get the best possible picture of sound propagation in the bay, and the hydrophones will therefore be placed at two of the positions used in the air gun/MV experiments (inner and outer) and the third hydrophone will be placed in one of the sites with most fish likely present (side bay). These are the positions shown in Figure 2.

The depth of all three hydrophone positions are approximately 50 m.

Distance from the source to the closest (inner) hydrophone will be approximately 250 m, thus representing a radius of relatively loud sound, representing the sound audible to fish closest to the source will hear. The furthermost (outer) hydrophone will be at a distance of about 1300 m, hence at the range of where the sound may almost reach the level of the background noise.

Details of the sound exposure

The speaker sounds are not capable of playing back all the frequencies from an operating wind farm, and therefore we have chosen not to make a playback of recordings hereof. Ideally, we were aiming to play various tones likely to be found at a wind farm but do to the speaker having best output around 100 Hz, and dropping off at both sides, other frequencies will have quite a lot lower output end hence reach over less distance. Therefore, we decided to limit the tone to 100 Hz. This may also make the connection to the work done on fish larvae (Cressci et al. 2023).

References

Cresci A, Zhang G, Durif CMF et al. Atlantic cod (Gadus morhua) larvae are attracted by low-frequency noise simulating that of operating offshore wind farms. Commun Biol 2023; 6 (1):art. 1. https://doi.org/10.1038/s42003-023-04728-y.

McQueen K, Meager JJ, Nyqvist D et al. Spawning Atlantic cod (Gadus morhua L.) exposed to noise from seismic airguns do not abandon their spawning site. ICES Journal of Marine Science 2022;79(10). https://doi.org/10.1093/icesjms/fsac203.

A2.1 Sound trap settings

In previous experiments, the sound traps have been used mainly to record cod vocalisations. Gain setting have therefore been at “high”. However, in the current experiment, they will mainly be used to record the level of the vessel and tone to get the received SEL at various sites. Therefore, the inner bay hydrophone (closest to source) is set to “low” to avoid clipping, while the side bay and outer bay is set to high as previously. Screendumps for all settings for the various years are shown in Figures A1.5-7.

A2.1 Settings of IcListen hydrophone

In 2023, 2024 and 2025, sample rate is set to 16kHz, continous sampling and file length 30 min. In 2026, sample rate was 64 kHz, and file length was 5 min.

Setting of sound source

Source is connected to the power amplifier, which again is connected to a PC running the sound file.

Power amplifier er set on max power and PC audio is set on 50 % (to ensure not overload). Make sure no clipping of signals (shown by light diods in amplifier).

Sound file is played using program Audiacity.