The use of acoustic devices to warn marine mammals of tidal-stream energy devices

Introduction - Is there a collision issue?

The incidence of marine mammals colliding with man-made objects is well-known and has received considerable attention in recent years. Collisions with fishing gear and entanglement is a now well quantified global problem (Read et al. 2006) and has elicited substantial research and mitigation efforts (Kraus et al. 1997). Similarly, impacts between marine mammals and vessels ranging from small boats to ships is sufficiently wide-scale and frequent to merit attention from global bodies such as the International Whaling Commission ( IWC) and the International Maritime Organisation ( IMO 2008). Investigations of these and other physical interactions have shown that in addition to the most obvious scenarios (animals swimming into nets or being struck by moving propellers) other less obvious contacts also occur. These include examples such as whales ensnaring themselves in marker buoy lines (rather than actual fish-capture equipment) and fatal collisions with the bows of ships rather than their propellers (Laist et al. 2001).

It is therefore a possibility that marine mammals might collide with marine renewable energy devices, with tidal-stream energy devices presenting the most obvious scenario (with potential parallels to birds and bats being struck by wind turbines, Barclay et al. 2007). The rotational motion of turbine blades (vertical and horizontal) coupled with the relatively rapid passage of water past devices, presents a conceivable collision scenario for species manoeuvring in the same water masses (Wilson et al. 2007). This possibility has been raised as an area requiring attention by a variety of fora when considering the potential environmental impacts of marine renewable energy devices (Linley et al. 2009).

At this time, it is unknown whether injurious mammal-turbine collisions will actually occur with sufficient regularity to be a significant concern. This is primarily because of the embryonic status of the industry with a limited number of full-scale deployments, the limited range of installation sites, the diversity of species so far exposed and their own spatial responses to the devices. However, with the imminent deployment of multiple full scale devices in several parts of the British Isles (The Crown Estate 2010) the issue of collision and its frequency or absence is likely to become clearer in coming years.

To consider the issue of collision in advance of the construction of commercial scale arrays, some basic interaction modelling has been carried out (Wilson et al., 2007). This suggested the spatial co-incidence is sufficient in that if marine mammals do not take appropriate swimming manoeuvres then physical interactions may be common enough to be of concern. However, with marine mammals being both mobile and agile, it is likely that some degree of avoidance (long-range) or evasion (close-range) will occur. This is only possible however if animals are able to accurately perceive the location of the turbines as they approach. For marine mammals coming from upstream (i.e. on a relatively rapid closing course), two of their primary senses: vision and mechanoreception (Dehnhardt 2002) are likely to be useful only at short ranges (tens of meters or less) such that they would only be relevant for close-range evasion. It is therefore more likely that they will gain their information on the presence, motion and three dimensional underwater extent of turbines from their auditory senses (Carter 2007; 2013). For odontocete cetaceans (i.e. toothed whales, dolphins and porpoises), echolocation (if used at the time) is likely to provide information out to greater than 100 m (Goodson et al. 1988; Au 1993) though it is unclear how/if echolocating animals might perceive and respond to long but thin and sweeping structures such as rotor blades using this directed and intermittent sense. It is otherwise likely that information on the presence of submerged energy devices will come from the machinery's own acoustic emissions (rotors, gearing, flow noise etc.). As mysticete cetaceans (baleen whales) and pinnipeds (seals) do not use active sonar (Dehnhardt 2002) their long-range awareness of turbines is likely to come entirely from auditory detection of device noise alone.

Modelling work carried out to calculate how far coastal marine mammals are likely to be able to detect tidal turbines (Carter 2007), showed that detection distances (and times) were highly variable depending upon the interplay of four variables in particular. These were:

1) the frequency specific acoustic output of devices,

2) the ambient noise at the site,

3) the site-specific propagation characteristics and,

4) the acoustic sensitivity of the species of concern (see Figure 1).

Other variables such as signal to noise detection thresholds were also important but less critical.

Figure 1. Schematic of the key parameters required for acoustic detection of a tidal-turbine. Sound emanating from a turbine (left) propagates away from the device while experiencing transmission loss (black line slope) and at some point drops below the site's ambient noise floor (green line). An approaching marine mammal (with its own frequency specific audiogram, thin wavy line), will only be able to acoustically detect the turbine when the turbine noise exceeds a threshold relative to ambient noise. This provides a minimum detection distance and based on the swimming speed and water flow rate provides an indication of the amount of time that the animal has to respond. This scenario assumes that the ambient noise floor exceeds the minimum hearing thresholds of the species of interest which is reasonable in tidally-energetic sites.

Of the four main variables, device noise is currently known only for a few (primarily prototype) devices, but is likely to become more readily available in the immediate future as commercial-scale turbines are deployed, particularly at test sites. Though there is little direct information on propagation in tidal-stream sites, there are sufficient published data to make useful estimates. Animal audiograms are also available for only some species but provide reasonable information on what odontocetes and seals in Scottish waters should be able to perceive. Information on baleen whales hearing is very limited (Richardson et al. 1995) and therefore they are not considered further here. The final variable, ambient noise, is relatively well known for the open sea (Richardson et al. 1995) but there is little tidal-site relevant information available. However, it can expected that both the lateral and vertical water motion and sediment transport will produce levels of ambient sound greater (or considerably greater) than in the more commonly studied environments.

Using available information (and appropriate assumption envelopes) on the acoustics of tidal sites, propagation, devices and animal audiograms coupled with known swimming speeds, Carter (2007), in a modelling study, calculated a series of distance-to-turbine detection scenarios. Because of the uncertainties associated with the input data, there was considerable variation in the outcomes but for many likely scenarios, the detection distances were short (500 m or less) and these became even shorter (100 to 10 m or less) when ambient noise levels were elevated by factors such as surface waves equivalent to Beaufort sea state 6 or greater.

Though this work was preliminary, it illuminates the possibility that while acoustic detection will be a primary cue for approaching marine mammals, under many of the likely circumstances, detections ranges will be too short for animals to effectively avoid close-range encounters with tidal turbines. One obvious solution to this detection issue would be to add further acoustic cues to turbines for them to be more detectable to marine mammals in their vicinity (Figure 2). This report therefore better explores:

1) whether or not there is actually a need for acoustic warning devices,

2) whether acoustic warning devices developed for other applications (e.g., seal scarers or bycatch reducing pingers) might prove valuable in a tidal-stream context,

3) what risks there may be of using such devices and

4) what characteristics purpose-built versions would need.

Figure 2. A development on Figure 1, where additional acoustic cue(s) are added to a turbine to increase the distance (and time) over which approaching acoustically sensitive animals can detect and potentially respond to the renewable energy device.