Automation and standardisation of a passive acoustic monitoring (PAM) system

Lead Research Organisation: University of Bristol
Department Name: Biological Sciences

Abstract

The objective will be achieved by creating a software package with a user-friendly computer interface to display real-time sound spectrograms and power spectra, as well as applying DCL algorithms to record marine mammal sounds, with the potential to extend the device to monitor and quantify anthropogenic sound emission.
 
Description This project explored potential modifications of a military marine passive acoustic system (SSQ 906G Low frequency analysis and recording (LOFAR) Sonobuoy), produced by the partner organisation Ultra Electronics Sonar Systems, for wider commercial use. Furthermore, automated systems to detect and classify marine mammals by their emitted sounds were explored to simplify detection and classification of marine mammals for environmental surveys around marine renewable sites.

To identify modifications needed for more widely applications of Ultra Electronics Sonobuoys, current and potential future requirements of passive acoustic monitoring systems were evaluated using a questionnaire. Analysis of returned questionnaires revealed three main usage trends: 1) Passive acoustic monitoring devices are used in a wide range of disciplines for various purposes (research, monitoring programs, industry, policy making, etc). To meet multiple requirements, passive acoustic monitoring devices thus need a high dynamic range to record quiet to very loud sounds, they need to include a wide frequency range to record very low frequencies (man-made sounds, fish and marine mammals) to very high frequencies (marine mammal ultrasonic clicks), the capability for long-term data recording on multiple channels and large data storage capacities. 2) The majority of users work in environmental-related research or in environmental monitoring projects. Thus, reusable and retrievable devices are preferred over disposable ones to minimise littering the seafloor. 3) To date, the main focus has been on marine mammal monitoring and research and thus, acoustic data collection has been restricted to sound pressure. However, research and monitoring programs on soundscapes and sounds emitted by animals other than marine mammals are increasing. As a consequence, sonobuoys will require multichannel systems for simultaneous recordings of several acoustic measurements, such as particle motion (the aspect of sound perceived by most fish and invertebrates) in addition to sound pressure at different depths and sound frequency ranges.

Comparison with features of Ultra Electronics Sonobuoys revealed four modifications necessary for wider commercial use: 1) the currently disposable Sonobuoys need to be designed reusable and retrievable, 2) data need internet, instead of current radio transmission techniques to enable recording of acoustic data outside human hearing range, 3) multiple connectors need to be included for collection of simultaneous acoustic recordings at different depths of different hydrophone types and accelerometers to record particle motion. Overall, this evaluation showed that creation of a product including all essential modifications, as well as novel features for future applications to assure competitiveness with alternative products will be difficult to achieve cost-effectively.

Automated detection and classification programs were evaluated as a tool to facilitate marine mammal monitoring tasks to conduct marine mammal surveys during construction and operation of marine renewable energy devices. The detection performance of marine mammal calls of two detection and classification programs was compared: PAMGUARD v1.12.05 BETA (Gillespie et al. 2009) and Marine Mammal Acoustic Detector (MMAD; trial version 3-2 WAV; Kaon Ltd., Guildford, UK). Overall, both detection programs fulfilled their purpose to detect different types of marine mammal calls. Dolphin whistles were generally less likely to be detected because their relatively complex structure and frequency range are more susceptible to distortion. Both programs produced false negative and false positive detections. Moreover, optimisation of detection parameters for several marine mammal call types proved difficult. For both programs, it is important to verify detections by checking the spectrograms before taking action or analyse data. While MMAD provides an easier user interface than PAMGUARD, users can not control, modify or add detection algorithms, but MMAD developers offer algorithm modifications tailored to specific user needs. Thus, it will depend on the end user's requirements, budget and expertise whether PAMGUARD or MMAD will be the program of choice. In relation to applications with Ultra Electronics Sonobuoys, two ways to proceed were identified: 1) using MMAD in collaboration with developers to modify algorithms to specific end-user's needs, 2) using PAMGUARD with a simplified in-house user interface allowing direct user control over parameter settings. In both cases, findings indicate that trained staff is necessary to adjust, optimise and survey detection parameters.
Exploitation Route Information gained in questionnaires from users of passive acoustic monitoring devices of different disciplines (research, environmental monitoring, policy makers, marine renewable device developers, as well as oil and gas industry stakeholders who survey sound emission by their activities) about current and future trends of technical requirements provide useful guidance for developers of passive acoustic monitoring devices outside the partner organisation Ultra Electronics Sonar Systems.

My assessment of Ultra Electronics Sonar Systems Sonobuoys and marine mammal detection programs is used by the partner organisation as a recommendation and guideline on whether and which modifications will be needed to meet requirements for wider commercial use of their product.
Sectors Aerospace, Defence and Marine,Environment,Manufacturing, including Industrial Biotechology
 
Description My assessment of modifications of Ultra Electronic passive acoustic devices necessary (SSQ 906G LOFAR Sonobuoys) to meet current and future user requirements for wider commercial applications have been used by the industrial partner to evaluate costs and benefits of further development of their Sonobuoys. Since several companies already included these modifications in their devices and are currently developing the integration of more functions, such as particle motion measurements in addition to sound pressure, my partner company Ultra Electronics Sonar Systems decided not to develop their sonobuoys further for the time being, with respect to the modification costs and the relatively small market for passive acoustic monitoring devices.
First Year Of Impact 2013
Sector Manufacturing, including Industrial Biotechology
Impact Types Economic
 
Description Ultra Electronics Weymouth 
Organisation Ultra Electronics
Country United Kingdom 
Sector Private 
PI Contribution My expertise as a behavioural ecologist and bioacoustician complemented my business partner's acoustic and technical expertise to identify potential modifications needed for a wider commercial use of a passive acoustic monitoring device (SSQ 906G LOFAR Sonobuoy) manufactured by the industrial partner. Furthermore, my expertise served to identify relevant sounds to evaluate the reliability of the detection and classification algorithms of marine mammal sounds used in this project. These detection programs were evaluated for potential use with the passive acoustic devices to simplify marine mammal monitoring procedures during installation and operation of marine renewable energy devices.
Collaborator Contribution My business partner Dr Peter Dobbins, principal scientist at Ultra Electronics Sonar Systems, contributed to the project with his acoustic and technical expertise. Furthermore, he trained me in sonobuoy handling. He supported me throughout the entire four months internship from May to August 2013 and gave me a unique insight into the different stages involved in developing and modifying a device for wider commercial use.
Impact - Internship report - Contribution to two conference papers
Start Year 2013