Effects of low frequency sound such as emitted by offshore wind turbines on fish were investigated in an annular tank with 10 m diameter filled with sea water 1.26 m deep. The tank was divided in neighbouring quarters by sound barriers, so that sound pressure differences of 32 to 52 dB were achieved. The experimental fish were free to move around in the tank and therefore could avoid high sound levels. Spontaneously, the fish preferred one of the quarters. Therefore the experimental sound was produced in that quarter by an underwater loudspeaker to test whether the fish would leave that quarter during sound production.
For the investigations, cod and plaice were chosen as two important North and Baltic Sea species with differently pronounced hearing abilities. In both species, juveniles and adults were tested. With every group of fish, 9 to 10 experiments were done, consisting of 24 hours continuous pure tone production followed by 5 days of recovery. Five frequencies between 25 and 250 Hz and sound levels of 130 and 140 dB re 1 μPa were chosen.
The 24 hour periods before, during, and after sound production were evaluated by registering the fish number, their distribution and their behaviour in the tank quarter containing the sound source, using continuous video surveillance with overhead cameras.
15 juvenile (32 - 53 cm TL) and 13 adult (56 - 72 cm TL) cod were tested. Activity was higher in daytime, which was more pronounced in juvenile cod. Except for 250 Hz, in most experiments significantly less fish were observed in the preferred quarter during sound production, than in the periods before and after, but the quarter was not left completely. Reactions were most pronounced at 60 Hz and 90 Hz.
From these results escape in some measure of cod from sound in the vicinity of offshore wind farms would be expected. The one day duration of the sound does not allow conclusions, however, whether the escape would be permanent.
20 juvenile (24 - 32 cm TL) and 20 adult (26 - 43 cm TL) plaice were tested. Without sound, the preference for one tank quarter was very strong in adults and less pronounced in juveniles. The fish showed a diel rhythm with higher activity in daytime, which was more obvious in adults than in juveniles. During sound production reactions were inconsistent in that both avoidance or attraction were found. Dependence on frequency or sound level was not found, nor was complete avoidance of the preferred quarter.
When specimens were added to the tank while the sound was being played in the preferred quarter settlement was delayed.
The results indicate that plaice detected the sound but the type of reaction was variable. They suggest that permanent avoidance of offshore wind farm areas would not be expected.
Behavioural reactions could be observed at sound levels of less than 30 dB above the detection threshold. The results are discussed in connection with different thresholds presented by other authors indicating the urgent need for further research to define reliable reaction thresholds.