Methods for tracking fine scale underwater movements of marine mammals around marine tidal devices

5 Discussion

5.1 Fine scale movement technologies

We have reviewed the applicability of three generic types of technology to investigate fine scale behaviour in the immediate vicinity of an underwater turbine. These are: animal-borne tags that collect and relay ashore individual's movement data, static active sonar and static passive acoustic arrays.

Our target specification for such a system is: a range extending to 100m of a turbine, a temporal precision of 1s and a spatial precision of 1m. This specification is not immutable. As we summarise below, however, a synergy of techniques should approach our target specifications. However the only way to determine system performance is to test it under field conditions.

5.1.1 GPS/ GSM tags

GPS/ GSM tags have the advantage of describing movement and behaviour where ever tagged seals go - and not just in the vicinity of the turbine. But seals caught and tagged near a turbine may not subsequently visit the turbine area. This may be in part a function of individual variability and small sample size. At the Sound of Islay a study of harbour seals indicated that 30% of the locally tagged individuals remained in the Sound and close to the area of the proposed turbine array. This figure is used in Table 1 to illustrate the 'plausible sample size'. We therefore suggest a sample size > 20 seals is required to provide sufficient statistical power to infer population behaviour. Grey seals tend to move greater distances (McConnell et al. 1999) and thus have a greater risk emigrating from the tagging area.

However GSM/ GPS tags on their own do not provide data of sufficient spatial and temporal resolution. The incorporation of dead reckoning to provide detailed underwater tracks depends upon a detailed knowledge of the current dynamics. Moreover, the incorporation of dead reckoning into GPS/ GSM tag will require about a year of development and testing.

5.1.2 Passive acoustic monitoring

A passive acoustic monitoring ( PAM) array fixed in the vicinity of a turbine can acoustically detect and track vocalisations of odontocetes and seals fitted with acoustic pingers at the required temporal and spatial resolution. A substantial engineering and financial effort is required to establish a local hydrophone and data processing. Thus it is sensible to use this investment to also track seals tagged with individually-coded acoustic pingers. Whilst there still is a risk that locally tagged seals may not visit the vicinity of a turbine, the relative cheapness of acoustic pingers means that a greater number of seals can be tagged. Vocalising baleen whales may also be tracked, but the uncertainty in their vocalising patterns means that there would be an unknown rate of false negative detections.

5.1.3 Active sonar

Active sonar is akin to underwater acoustic radar. It has been proven to detect and track marine mammals in the vicinity of underwater turbines at sufficient spatial and temporal scales. It is good at detecting both seals and cetacea and viewing the raw data usually allows distinction between a seal and an odontocete. However, this discrimination is vastly improved if used in combination with the species specific (or to individual specific if acoustically tagged seals are available) PAM systems. Active sonar is the only technology that will detect and track baleen whales.

5.1.4 Marine mammal classes

We consider three classes of marine mammals: 1. seals, 2. toothed whales (odontocetes) and 3. baleen whales (see Table 1). UK seals and baleen whales may occasionally vocalise, but these events are sufficiently rare or unpredictable that they cannot be used reliably to detect and track animals with a PAM system. Thus the only technique applicable to baleen whales is active sonar. Odontocetes, however, frequently vocalise and thus may be readily tracked in the vicinity of a PAM array. Seals (unlike UK cetacea) have the advantage of being readily caught and fitted with active tags - including acoustic pingers which can be tracked with PAM arrays.

5.1.5 Suggested system configuration

To balance the strengths and weaknesses of the above technologies we suggest that the following generic configuration represents the best probability of achieving the overall objective:

• Establish of a static PAM array around one or more turbine to track vocalising odontocetes.
• Tag local seals with acoustic pingers so that they can also be detected and tracked by the PAM array.
• Establish one or more active sonars on the turbine to detect and track all marine mammal species (including baleen whales).

5.2 Impact detection technologies

None of the technologies discussed above is likely to provide data of sufficient quantity and quality to confidently distinguish a collision from a near miss. Yet impact detection is essential to interpret fine scale movements. For example, which of the close interactions recorded by the tracking technology result in an impact?

Whilst video surveillance has distinct limitations (water turbidity, bio-fouling and not functional during darkness) it does provide a potentially powerful detection capability when conditions allow. Recent developments in low light level camera technology might extend usefulness well into the twilight hours. Mechanical detection, using inbuilt accelerometry and strain sensor on the turbine blades, is not limited to favourable environmental windows. Also, it would be a cheap system to employ. However its potential (or otherwise) has yet to de demonstrated.

5.2.1 Roadmap for evaluating impact detection systems

To detect direct impact we suggest pilot testing both systems:

• Establish two video cameras pointing up- and down-stream of the turbine. Evaluate whether the automatic detection software systems discussed in the PAM array section could be incorporated. The video surveillance system could be triggered by detection events from the active sonar and PAM systems.
• Test whether Mechanical Detection using the turbine's sensors will detect the strike of a carcass. This test should also be combined with video surveillance. If successful then trigger logging of the turbine sensors with PAM or active sonar detection event.

5.3 High energy sites

Areas of high tidal energy offer both challenges and opportunities to study the fine scale behaviour of marine mammals. On the negative side they are areas of turbulent flow with unpredictable, though perhaps important, local eddies. The flow produces high levels of ambient noise which can interfere with techniques that use sound - either actively or passively. The high current and consequent higher erosion rates from suspended material increases the practical difficulties in establishing and maintaining hardware underwater.

However there are positive aspects. The establishment of a turbine means that the surrounding area will have been well surveyed, although perhaps not sufficiently for our purposes (see below). It is also likely that both power and high bandwidth communication channels will be locally available. In addition the actual structure of the turbine device may be made available for the attachment of hardware.

The information derived from the technologies considered here is only part of that needed to interpret fine scale behaviour and detection of impact. Generally we highlight the need for the larger scale distribution, movements and behaviour of the target species to be investigated. The background biology is essential. Similarly information about the local physical environment is required. This includes: detailed current and bathymetry, turbine construction and operation schedules, and turbine noise generation levels and local sound propagation models.