Pinniped predation on fish farms is a worldwide problem and causes the industry financial losses of up to 10% of the total farm gate value (Nash et al. 2000). This has led to a need for the industry to look at non-lethal measures for controlling seal damage, which include tensioning nets, deployment of a predator net or use of acoustic deterrent devices ( ADDs) (Würsig and Gailey 2002). Acoustic deterrent devices have often been considered a benign method of dealing with the problem. However, the main problems with ADDs appear to be the lack of long-term efficiency and unintended effects on other marine wildlife (Jefferson and Curry 1996). Some studies have found prolonged effectiveness of ADDs in applications where devices are used to protect confined areas (Fjalling et al. 2006; Graham et al. 2009), however, the much more common picture is that animals rapidly habituate i.e. avoidance response wane and predation resumes (Götz and Janik 2010; Mate and Harvey 1987). Several studies that tested currently available ADDs around haulout sites (Jacobs and Terhune 2002), salmon runs ( NMFS 1995) or fish farms (Norberg 1998) found little or no effect on seals and sea lions.
The potential impact of ADDs on other marine wildlife (non-target species), in particular cetaceans is of concern. There is a possibility that long-term exposure to ADDs may damage the hearing system of target and non-target species (Götz and Hastie 2009; Taylor et al. 1997) and that ADDs could cause long-term habitat exclusion of toothed whale (odontocetes). Olesiuk et al. (2002) showed that harbour porpoise (Phocoena phocoena) sightings in the Broughton Archipelago (British Columbia) dropped to 10 % of the expected value at ranges up to 2500 and 3500m from an operating Airmar ADD. In another study, porpoise numbers were found to be significantly lower in an area of up to 1.5km around a deterrent device (Johnston 2002). Long-term habitat exclusion over several years has also been shown in killer whale (Orcinus orca) and Pacific white-sided dolphins (Lagenorhynchus obliquidens) (Morton 2000; Morton and Symonds 2002). One likely reason is that odontocete hearing is about 30-40 dB more sensitive than pinniped hearing ( e.g.Johnson 1967; Kastelein et al. 2002) in the frequency range where most commercial ADDs operate (10-40 kHz).
Sounds produced by current ADDs are not based on biological concepts of aversiveness but aim to transmit loud sound to the target animal. In a previous project funded by the Scottish Government (Janik & Goetz 2008) we developed a deterrent method that used an autonomous, acoustic startle reflex ( ASR) to induce controlled and sustained flight responses in phocid seals but not in odontocetes. The startle reflex is elicited if a stimulus reaches an intensity threshold of 80-90 dB above the hearing threshold within 15ms of its onset (Goetz & Janik 2011). Seals exposed to startle stimuli became more likely to exhibit rapid escape responses in repeated exposure, causing animals to leave the exposure pool and show clear signs of fear conditioning (Götz and Janik 2011). Once sensitized, seals even avoided a known food source and showed prolonged location avoidance even in control periods with no sound playback. Harnessing the ASR is beneficial for several reasons:
1. avoidance responses are limited to the desired area around the device where received levels exceed the startle threshold,
2. the use of isolated, infrequent noise pulses greatly reduces noise pollution and removes the risk of hearing damage,
3. the startle threshold runs roughly parallel to the hearing threshold (Fleshler 1965) allowing a stimulus design that exceeds the threshold for one species but not another if their hearing thresholds are different.
The latter point would also allow a further development of the startle method to deter cetaceans instead of seals, or in combination to deter both animal groups. This might be desirable for short periods of time around marine construction work when construction noise levels could potentially damage marine mammal hearing.
This project tested (a) the effects of startle sounds on seal predation and marine mammal abundance around a test fish farm in comparison to adjacent control farms without operational ADDs, (b) the short-term effectiveness of the startle method on two additional farms when they experienced high seal predation rates and (c) startle thresholds in two bottlenose dolphins to develop startle sounds for echolocating marine mammals. These objectives allowed us to address the questions whether the startle method is effective in the long term in deterring seals from fish farms, whether it is possible to use the startle method aimed at seals in areas frequently used by cetaceans without having an adverse effect on cetaceans, and whether the startle method could be used to deter cetaceans if the design was changed accordingly.