Abstract
As nations seek to reduce their reliance on energy from fossil fuels, developers are turning to the marine environment as a potential source of low carbon energy. This is leading to new types of anthropogenic activity and the industrialisation of marine areas. One newly developing marine renewable sector is tidal energy and in the UK and elsewhere there is interest in utilising strong tidal currents to power underwater turbines. The majority of these devices extract energy using large, un-protected, freely rotating blades, which pose a potential collision hazard for larger marine animals. The tips of some proposed blades will reach speeds of 12.5 ms-1 and could potentially injure or kill fish, birds and marine mammals (Wilson et al. 2007). The risk posed is poorly understood, because little information exists on how animals utilise tidal habitats and there is a lack of any understanding of how marine mammals will react to turbine structures once deployed (Frid et al. 2012; Benjamins et al. 2015).
Appendix 1: Constructing the KE Buoy
Appendix 2: Software Instructions
Several pieces of biological information are required for each species of concern to be able to predict collision risk. These include: the density of animals in the area of interest, their depth distribution and underwater movements and the extent to which they can detect and avoid the devices.
Methods for determining density are well established for most species, although some difficulties are introduced in tidal areas due to rough sea states, small site sizes and difficulties of surveying in fast moving currents (Gordon et al. 2011). The response of animals to turbines can only be effectively measured once the devices are deployed and operational, (although some experiments investigating behavioural responses to noise have been attempted pre turbine deployment (NERC/Defra RESPONSE project NE/J004251/1). The natural behaviour of marine mammals in tidal habitats, including their depth distribution and underwater movement, is very poorly understood. Traditionally these somewhat unusual areas have been poorly studied, in part because they are such physically challenging areas in which to work. This report focuses on the development of a practical and cost effective method to determine the underwater behaviour of harbour porpoises (Phocoena phocoena L.) and other echolocating cetaceans in tidal rapid habitats.
Harbour porpoises are the most commonly encountered marine mammal species in Northern European shelf waters (Hammond et al. 2013). They are listed under Annex II and IV of the EU Habitats Directive, which requires Member States to assess potential conservation threats, including industrial developments. Harbour porpoises are shy, elusive and difficult to detect visually in all but the calmest of sea states. However, like many toothed whales, they are highly vocal animals, producing characteristic very high frequency (130 kHz) echolocation clicks nearly continuously, both to sense their surroundings and hunt for prey (Dubrovskii et al. 1971; Mohl & Andersen 1973; Villadsgaard et al. 2007; Linnenschmidt et al. 2013). Transient sounds, such as harbour porpoise clicks, can be detected using hydrophones and high speed data acquisition systems (for a porpoise sampling at >300 kS/s) in conjunction with appropriate signal processing software (Madsen & Wahlberg 2007). The use of such methods to study animals is referred to as passive acoustic monitoring (PAM). PAM can complement and is often more effective than visual methods for detecting and studying harbour porpoises and crucially, for the application described here, PAM can be used to detect and localise the position of these cryptic animals in three dimensions underwater.
PAM methods have been employed to determine presence and density of harbour porpoises for decades (Chappell et al., 1996; Gillespie and Chappell, 2002). Multiple hydrophone elements (hydrophone arrays) can be used to determine the position of animals underwater by measuring the time delay in receiving the same signal between different elements. Compact linear stereo hydrophone arrays are often towed behind survey vessels and can be used to calculate a distance to vocalising animals (Leaper et al. 2000). These can then be combined with concurrent visual observations allowing the absolute density of animals within an area to be determined (Leaper & Gordon 2012). However, towed stereo arrays do not provide sufficiently accurate data to determine animal movements in detail. Much larger arrays with a greater number of more widely spaced hydrophone elements can provide instantaneous 3D locations of vocalising animals. Such ‘large aperture’ arrays, which are generally deployed on the seabed e.g. (Roy et al. 2010; Wiggins et al. 2012) or from drifting vessels or buoys e.g. (Watkins & Schevill 1971; Heerfordt et al. 2007; Wahlberg 2002; Hastie et al. 2006; Miller & Dawson 2009), can provide more accurate information on animal locations, and in some cases , have been used to reconstruct three dimensional movements of animals underwater. Deploying such arrays in a tidal environment is difficult. Strong currents mean any seabed devices require large weights to remain stationary and acoustic sensitivity can be hampered by significant flow noise over hydrophones. Drifting systems are also problematic as the dynamically changing environment within tidal races systems can result in hydrophones moving in an unpredictable manner.
Since 2010, the Sea Mammal Research Unit (SMRU) has been developing large aperture hydrophone arrays and the associated software to track fine scale underwater movements of harbour porpoises and other echolocating cetaceans in tidal rapids, with an emphasis on providing information on animal diving behaviour for use in EIA (Environmental Impact Assessment) and collision risk assessment (Macaulay et al. 2015). This tracking system consists of a 30-45 m freely hanging vertical array and a small (~0.5 m) tetrahedral cluster of hydrophones deployed from a drifting vessel. The hydrophone array can drift through tidal rapids and determine the georeferenced positions of animals underwater, providing detailed information on underwater behaviour and, crucially, on depth distribution to help assess collision risk with tidal turbines.
Years of study and extensive calibration trials have shown this system is a powerful and cost effective methodology to assess harbour porpoise behaviour in tidal rapids. However, the requirement for a substantial vessel to drift through a tidal race means it is both expensive, sometimes dangerous to deploy in tidal rapids and requires a sizeable field team. Thus the final design of hydrophone array that emerged after many years of development was somewhat cumbersome, making it difficult for non-specialised research groups to utilise these methods. However, recent developments in digital acquisition and processing tools and further design refinements would enable us to recreate the functionality of the full system in a small, affordable buoy based format, which, in conjunction with streamlined, user friendly, open source software would make the technology more widely available. The development of such a system forms the basis for this NERC Knowledge Exchange (KE) project.
In 2013, SMRU applied for and was awarded a knowledge exchange contract to package the drifting hydrophone array into an easy-to-use autonomous buoy capable of collecting the same quality of data. The advantages of such a system are numerous. A buoy is safer as it does not require a vessel to drift through tidal rapids. As such it can be deployed in rougher weather and is easier to use at night. It is also much smaller than a boat based vertical array and so can be deployed from a RHIB (Rigid Hull Inflatable Boat) or other small vessel, significantly reducing costs. Perhaps the greatest advantage is that an autonomous buoy is far easier to use and requires a much smaller specialist team to be present. Thus data collection is both less expensive and should be achievable by most marine environmental consultancies.
Our remit with the KE project was, to use readily available and off the shelf components in developing a cost effective accessible system supported by user friendly software. The main deliverable was to be detailed instructions on how to assemble the necessary hardware, including details of suppliers for components, the software and instructions on how to use it (these are provided in the two Appendices to this report).
This report focuses on an overview of the design of and methods to construct a Porpoise Locating Array Buoy (the PLABuoy). The buoy consists of a small waterproof enclosure (barrel) containing a communication and recording system attached to a 30m long flexible hydrophone array with a weight on the end to keep it steady in the water column. The weight and barrel are easily lifted by a human and the whole system can be deployed and recovered from a small outboard-powered vessel.
Open source software has been created to run on the PLABuoy embedded computer to enable acoustic data collection, wireless interaction with a tablet or computer whilst deployed and for post processing the collected data. Algorithms and new features have been added to PAMGuard (Gillespie et al. 2009) , a widely used, open-source, passive acoustic software suite, in order to facilitate data analysis. Details on how to set up the recording computer and electronics are discussed (in Appendix 2) along with setting up the development environment and downloading source code, allowing new features to be added or existing features to be improved.
Tidal races are harsh environments and equipment must reflect this. The report therefore also details how to cost effectively build robust hydrophones which can survive rough treatment during deployment and recovery from a small vessel. Another consequence of tidal races is the potential for flexible vertical arrays to bend substantially underwater, due to differential currents or wind against tide effects. The use of IMU (Inertial Measurement Unit) sensors to track this movement, allowing the positions of hydrophones to be determined, is also described.
Finally, trials of the first prototype PLABuoy took place off Anglesey in Wales (April 2015). These demonstrated that the system was robust, practical to deploy and capable of providing accurate locations and tracks of calibration sound sources as well as harbour porpoises underwater. Results of these trials are presented.