Noise in salmon farming
In modern aquaculture, both at sea and on land, farmed salmon live in environments characterized by significant human-made noise. This noise may potentially have a negative impact on the most farmed animal in Norway. However, the extent and effects is not fully understood. Our aim was to describe the typical sounds within salmon aquaculture, develop methods to investigate long-term effects of noise, test habituation to noise and summarize present knowledge within the field.
Methods and setup
Firstly, sound was measured over representative periods a t 10 different farms divided among open or closed sea cages and land-based systems. Secondly, we exposed salmon experimentally to known stressful low-frequency sound (10 Hz) for 5 minutes daily , to better understand the responses . Behaviour and physiology, together with m olecular markers in the brain were to be identif ied for stress assessment . Thirdly, salmon smolts were experimentally exposed to "fish-farming sounds" in a predictable or unpredictable setup in freshwater tanks, followed by unpredictable sounds in sea cages to study both short- and long-term effects on stress and habituation. Finally, a review was written.
Results and discussion
The sound farmed salmon experience ranges from background levels around 95 dB re 1 μPa to typical levels approaching 134 dB, with very noisy periods up to 157 dB. These levels rang e within levels measured in other studies but provide further details. Closed sea cages were noisiest, likely caused by pumps continuously supplying water. Land-based systems were somewhat less noisy and open cages quietest , especially at night, though noisy operations and boats were present during the day. On three occasions, extremely high sounds exceeding 175 dB were measured, either when an underwater door was slammed shut or when blastings occurred nearby. The frequency range of the noise fell within the audible range for salmon. Overall, farmed salmon live in a noisy world, with occasionally very high sounds.
Low frequencies (<15 Hz) are particularly stressful for salmonids, and salmon exposed to 142-169 dB re 1 μPa immediately responded with flight reactions , elevated plasma cortisol levels, and reduced neuronal activity in the hypothalamus. Over time, fewer fish reacted to the daily repetitive, familiar sound, and the cortisol physiological stress response diminished. S almon thus showed behavioural and physiological habituation. However, after a month and 30 exposures, the salmon still had increased serotonergic activity (indicating chronic stress in the brain) and suppressed expression of genes in the hypothalamus that are normally positively associated with growth and reproduction.
Freshwater parr- smolts were experimentally exposed to predictable, unpredictable or control tank sounds, followed by exposure to unpredictable sounds in sea cages. No group differences in the freshwater phase were present, but similar behavioural reactions and no physiological responses in cortisol or neurochemical signalling of serotonin or dopamine were seen . When individuals were further acutely stressed , normal physiological responses suggested that none of the sound environments were more stressful during the freshwater phase. In sea cages t he most interesting (and unexpected) finding was that control fish reacted behaviourally the least, while fish that experienced sound in freshwater responded relatively strongly with a disorganized swimming pattern , more so in predictable compared to the unpredictable group. Cortisol responses in seawater were absent when exposed to boat noise or similar, but normal when additional acute stress was given . The neurochemical signalling responses showed that after the acute stress test, both the control and unpredictable groups responded with increased serotonergic activity in specific areas of the brain, while fish exposed to predictable sound did not. Elevated cortisol levels without a corresponding increase in serotonin levels may be harmful by impairing brain health (reduced neuronal plasticity and imbalance in the body’s energy budget). Behavioural and neurochemical responses suggest that fish exposed to high levels of predictable sound during the freshwater phase become more stressed when exposed to unpredictable sound in the seawater phase.
The review reveals that most research has been measured in sound pressure (dB re 1 μPa), while particle movement (ms-2 ), the most important sound component to which salmon respond, is scarcely studied. Most studies are on physoclistous fish, having hearing thresholds far below the physostomous salmon. Salmonids can detect frequencies from 1 to 1000 Hz, but with lowest thresholds between 100 to 500 Hz, frequencies that overlap noise in aquaculture. Cumulative exposures to sound levels exceeding 177 re 1 μPa2 s caused moderate physiological damage, while levels exceeding 183 re 1 μPa2 s caused lethal damage in chinook salmon. Acute sounds typically trigger a behavioural flight or stress response, where salmon try to avoid the noisy area or swim in disrupted directions and speed. Long-term studies exposing salmon to ≤ 149 re 1 μPa triggered behaviour that disappeared over time, while physiology and welfare were unchanged (except for the study mentioned above with 142-169 dB re 1 μPa). A few studies report that infrasound (e.g., 5 Hz) is particularly frightening for salmon, supporting the long-term study where daily exposure to 10 Hz resulted in poorer welfare. Habituation to sound is common, both in terms of behaviour and physiology.
The combined works compile a list of recommendations for noise in salmon aquaculture. Noise measurements within the audible range for salmon should be considered against “established” threshold values for known negative consequences on behaviour, physiology, and physical damage. Unnecessary high noise should be avoided, and unpredictable noise minimized. Fish farmers should actively manage the use of noise sources near the fish and consider whether habituation is possible. Blasting work nearby or other acute high sounds can lead to undesirable behavioural changes or injuries and should be avoided. Low noise levels in freshwater may result in more resilient salmon in seawater. Mitigations could include identifying noise sources with frequencies < 500 Hz and dampening the sound by separation or physical attenuation. A good example of noise reduction is to soften the closing of hatches and avoid them slamming shut with a bang.
We must better understand the long-term consequences of different sound environments in the freshwater phase (high vs. low noise). We have already observed that very loud sounds affect the brain in a way that may hinder growth or impair brain health, though the long-term consequences have not been studied.
Normal sounds in aquaculture (<135 dB re 1 μPa ) are not expected to have significant negative consequences for the salmon, but short noisy periods (<157 dB re 1 μPa ) could have negative effects depending on their duration, while the acute sounds observed (>175 dB re 1 μPa ) could lead to hearing loss and physical damage. The knowledge compilation highlights research needs for more precise threshold values of how frequencies and durations affect salmon behaviour, physiology and welfare. Advisory knowledge for blasting is insufficient . Standardized measurement methods should be developed to enable fish farmers to measure relevant noise that could affect salmon, choose mitigation measures, and assess their effects. D ata on particle acceleration and its consequences, the most important sound component for salmon, is lacking.