Publication - Research and analysis

Evaluating and Assessing the Relative Effectiveness of Acoustic Deterrent Devices and other Non-Lethal Measures on Marine Mammals

Published: 28 Oct 2014
Part of:
Marine and fisheries

Marine Scotland commissioned a research project aimed at gathering literature and data into the effectiveness of non-lethal measures of deterring marine mammals from a range of activities (e.g. fish farms, renewable developments etc.). This review attempt

Evaluating and Assessing the Relative Effectiveness of Acoustic Deterrent Devices and other Non-Lethal Measures on Marine Mammals
5 Active Acoustic Devices to Reduce Risk of Marine Mammal Collisions with Tidal Turbines

5 Active Acoustic Devices to Reduce Risk of Marine Mammal Collisions with Tidal Turbines

5.1 Background

The renewable energy industry already makes up a significant proportion of Scottish energy production and consumption. The Scottish Government has set the ambitious target of generating the equivalent of 100% of Scotland's gross annual electricity consumption from renewable sources by 2020. Scotland is well supplied with many forms of renewable energy, including hydro (which is already heavily exploited), wind, and wet renewables (wave and tidal). It is recognised that an effective strategy for a low carbon future will require the utilisation of all of these options. Within the portfolio of renewable energy sources, tidal power has the unique advantage of being highly predictable. It has been estimated that Scotland has some 14 GW of tidal potential and has already become a world leader in developing tidal power technologies. A number of tidal generation devices have been undergoing tests for several years at the European Marine Energy Centre ( EMEC) in Orkney and the first commercial arrays are slated to start construction within the next few years in the Sound of Islay and the Inner Sound.

Tidal turbines are driven by some type of rotating blade, often with quite considerable tip speeds. Most prototypes resemble the typical wind-turbine design, with an axis of rotation parallel to that of the water-flow, and these designs may either be ducted or un-ducted. There are also devices such as the TidGen device made by Ocean Renewable Power Company, Maine, USA, which have an axis of rotation perpendicular to the water current. It is possible that these (and other) variations will have relevance if different degrees of collision risk are found, but this cannot be meaningfully considered at present due to lack of information. The potential for larger marine animals, in particular marine mammals, to be in collision with these rotating blades has been raised as potentially serious risk and remains a major environmental concern (Linley et al., 2009) which could slow the pace of development of this important technology.

The current embryonic stage of the industry means that a variety of questions concerning marine mammal interactions are currently unanswered. The most fundamental questions are whether or not a genuine collision risk exists and whether or not this risk is sufficiently great to pose a significant concern.

5.2 Assessment of Risk

The possibility that marine mammals will be hit by rotating turbine blades is seen as one of the most serious acute environmental risks posed by tidal stream electrical generators (Wilson et al., 2007; Wilson and Gordon, 2011). Simple modelling exercises indicate that significant numbers of marine mammals would be struck if they do not take action to avoid tidal turbines (Wilson et al., 2007). We will not know how marine mammals respond to these devices until turbines are deployed in the marine environment and appropriate observational research is conducted to track animal movements and responses in their vicinity. However, given the low water visibility at Scottish tidal rapid sites, it is clear that the main means for detecting these devices at sufficient range will be acoustic.

Odontocetes, which have well developed echolocation abilities, will be able to detect turbines using both active and passive acoustic sensing, while seals and baleen whales will be reliant on passive acoustic cues. A pertinent question then becomes whether marine mammals will be able to detect turbines acoustically at a sufficient range to be able to take appropriate evading action. Tidal turbine devices certainly generate underwater noise. It is likely that much of this will come from the machinery in the gearboxes and generators within the device. Tidal turbine development is still at an experimental stage with a diversity of different device types being trialled. It is probable that they will differ in their underwater noise output and so far, few data on acoustic characteristics are openly available. Background noise levels in strong tidal current areas can be quite high and variable due to the effects of water turbulence and sediments moving in tidal streams. Such tide-induced noise also varies temporally, being strongest when currents are highest, and the distribution of some noise, especially sediment movement noise, varies spatially, often occurring in quite discrete patches which may correlate with sediment patches (Gordon et al., 2011). These high levels of noise could mask turbine sound making these devices more difficult to detect.

KG No. Knowledge Gap
44 The acoustic output of tidal energy devices in all states of operation is unknown.

The only source of empirical data about potential collision risk is the SeaGen installation in Strangford Lough. The SeaGen tidal device in Strangford Lough consists of two 16m diameter turbines, which slide up and down a 3m diameter pile. The blade angle is adjustable so that a maximum tip-speed of around 12m/s can be maintained. Marine mammal observers were stationed on the pile for the first year of installation, allowing collection of data which has since been used to inform the use of active-acoustic monitoring. The turbine was shut down if marine mammals were detected by observers or on active acoustic systems. The minimum range at which a marine mammal was allowed before shutdown was initiated was gradually reduced over this time from 250 m to 50 m, and 15 shutdowns were conducted in the first year. Shoreline surveillance has also been conducted in order to examine any carcasses for evidence of interaction with the turbine, but no collisions have been reported. The requirement to shut down and the emphasis on this type of mitigation monitoring has meant that less has been learned about the extent to which animals detect and avoid operating turbines.

5.3 Potential Use of Acoustic Deterrents

If it transpires that the devices' self-noise is insufficient to alert marine mammals to their presence in the ambient noise field and collision risk does become an issue then it may be necessary to use acoustic alerting devices to make animals aware of the presence of tidal turbines. In a report for the Scottish Government, Wilson and Carter (2013) reviewed this topic, provided some new data on noise levels in some strong tidal current environments and provided some consideration of how acoustic devices might be used to mitigate tidal turbine collision risks. They suggested that to be useful a device should have seven attributes. These were that: (1) The signal must elicit an appropriate response. It should either cause the animal to avoid the immediate vicinity of a device and/or, in the case of small cetaceans, draw attention to it so that active acoustic sensing (echolocation) can be used. However, because devices would likely be used over the lifetime of the turbine arrays, which should be many years, it will be important that they should not cause large scale habitat exclusion, especially from any high energy tidal habitats that may be highly preferred by marine mammals (Gordon et al., 2011; Pierpoint, 2008). (2) The emission schedule of the device need not be continuous but must suit likely approach velocities. An approaching animal, which may be traveling quickly if it is moving with the current, must be able to detect intermittent pulses with sufficient time to be able to take avoiding action. (3) The emission frequencies must be audible to target species. This is a function of both auditory sensitivity at different frequencies of the suite of species involved but also the frequency spectrum and levels of background noise in the area. (4) Amplitudes must be appropriate for the necessary detection ranges and sites. (5) Signals must be directionally resolvable. Animals will only be able to take avoiding action if they can locate the direction that sounds are coming from. Sounds with different characteristics can be more or less readily localised. (6) The warning should be coordinated with the threat. The threat is greatest when the current is strongest and the turbines are turning. There may be a case for only having a device active when turbines are rotating quickly enough to cause injury. (7) The location of the sound sources at a turbine or within an array must facilitate appropriate spatial responses. Scenarios may become more complicated when many devices are deployed in arrays. Then consideration will need to be given to encouraging avoidance of arrays as a whole and avoiding entrapment.

For echolocating animals we suggest that it will be particularly important to consider whether devices affect echolocation behaviour. Animals that are alarmed by a sound that they might perceive as a predatory threat might well cease echolocating to avoid detection and flee. If this occurred it would be counter-productive as the possibility of detecting the turbines using echolocation would be diminished. Wilson and Carter (2013) investigated the likely detection range for two types of existing active acoustic devices: low power anti-bycatch pingers ( section 3.2) and high power aquaculture anti-predator devices ( section 2.2). Their results suggested that, in noisy areas, the former would not be picked up at useful ranges while the latter could cause habitat exclusion for some species and might even carry a risk of causing hearing damage. Wilson and Carter's (2013) detection range modelling may be a little pessimistic in that they assumed a rather wide critical ratio for masking and did not allow for a marine mammal's ability to localise sounds (Branstetter and Mercado, 2006), which may reduce the likelihood of masking. Nevertheless, their general conclusion, that a bespoke sound source may need to be developed for this application is, we believe, a valid one. However, we question whether it is sensible to begin work to develop such a system at this stage when the nature of the collision process and the behaviour surrounding it is unknown, as is whether there is indeed a problem that requires intervention.

There are a variety of questions which need to be addressed in order to make a more informed assessment of the severity of animal-turbine collision risk:

KG No. Knowledge Gap
45 At what range will tidal turbines be detectable above ambient/background noise?
46 To what extent can noise hotspots in tidal areas be predicted based on parameters such as benthic composition? How stable are they spatially and temporally?
47 A greater understanding of the response elicited by existing deterrent devices, how this varies between species and contexts, and how this is likely to change over time.
48 A greater understanding of how marine mammals currently utilise tidal environments, and how this is likely to be affected by new structures and activities and additional sound sources.