Passive Acoustic Monitoring Systems

Ecologic has been a pioneer in promoting the use of passive acoustic monitoring systems to detect cetaceans during mitigation and population surveys.  PAM can dramatically increase the efficiency and consistency of detections for certain species.  Generally we recommend its use in addition to visual monitoring and users should be aware of both the  strengths and shortcomings of existing systems.

PAM systems consist of both hardware and software.  

Hardware

Ecologic can provide various customised acoustic monitoring systems integrating components from a variety of providers with items built in-house.  An example of an acoustic monitoring system is provided below.

We work closely with Magrec to specify acoustic equipment suitable for marine mammal users and we have an arrangement to provide this equipment.  For more information on the Magrec range click here.

The towed hydrophone streamers are based on a well-established design that has proven highly cost –effective, robust and reliable.  It allows flexibility in terms of inclusion of different hydrophones and other sensors and can, if necessary, be disassembled and repaired in the field.   Seismic PAM hydrophones operate in an environment which is extremely noisy and in which the risk of hydrophone loss or damage is significant.  This argues for utilising simpler and less expensive systems built with a view to being repaired in the field.  Systems used during seismic operations typically have a 400m tow cable and 50m deck cable.

Filtering and amplification equipment is custom built to our specifications by Magrec.  Important features include strong low frequency filtering, and optional electrical isolation to avoid electrical noise problems.

High frequency click detection (particularly useful for porpoises) is achieved with the use of custom built three channel envelope tracing hardware provided by New Leap.

Audio and low ultrasonic bands (up to 44kHz) are digitised using a USB sound card. An appropriate digital acquisition card is used if the high frequency detection equipment is used.   Detection and data recording is all achieved on a standard PC.

Systems like this have been used widely by Ecologic and others from a wide variety of platforms ranging from sailing yachts to ocean-going ice breakers and in waters from the tropics to the Antarctic.  Seismic survey mitigation monitoring has been conducted from both guard vessels and the main seismic survey vessel itself.  Operation from the seismic vessel has proven most straightforward and would be favoured in most situations in the future.

  Monitoring from a guard vessel on the Atlantic Frontier

Systems have been provided for a number of organisations including NOAA, MMS (for SWSS and SWAMP), TAMU in USA and QinetiQ, SMRU and Guardline in UK.  Mitigation and surveys have been completed for CONOCO, Coflexip, NOAA, Alterra.

Software

The system we recommend for mitigation work relies heavily on a suite of acoustic detection and data collection programs that were developed as part of the Shell/Birmingham project by Oliver Chappell and Douglas Gillespie.  Most of these are based on routines initiated by IFAW and IFAW continues to support the core programs which can be downloaded from http://www.ifaw.org/page.asp?unitid=660.  Although the different detection modules can be run independently, when run within the “Logger 2000” data collection program they constitute a single integrated system providing real time information about a wide variety of signals on which management and mitigation decisions can be made.  Logger 2000 also provides an auditable record of all detections and can make continuous recordings or make sample recordings of detected signals (from at time buffer) or recordings at regular intervals. 

This suite of programs which have been  in use/development/test by a wide community of users for over eight years have proven highly reliable.  They are straightforward and user-friendly, having been designed for field biologists to use; they work well in their standard configurations but offer extensive options for customisation and they are available for free.  We also feel that the fact that these program were written by programmers that were working on seismic surveys, indeed some of the coding was done at sea during surveys, is a significant factor in their relevance to this application.  All detector programs calculate and display a bearing to detections.  “Rainbow Click” detects audio band clicks from sperm whales, and other odontocetes, including pilot whales and dolphins.  It is particularly useful with sperm whales and outputs data to a plotting program in Logger that will plot the locations of whales in small groups by target motion analysis.  “Whistle” detects tonal signals, such as baleen whale moans and the whistles of dolphins and pilot whales.  “Porpoise” detects and distinguishes the ultrasonic vocalisations of porpoises (narrow band ultrasonic pulses are produced by all phocoenids and cephalorynchus dolphins species).  Another freely available program, Dave Mellinger’s “Ishmael”, has been found to be a very useful support for monitors as it provides an easily interpreted scrolling spectrogram display and allows signals that have already been heard to be recalled and explored.

Screen shot from the Whistle detector program

Strengths and Shortcomings

Like all PAM systems this configuration has general strengths and also shortcomings.  Shortcomings are essentially a function of the environment in which they must operate, in particular the high noise levels, and the nature of the signals – namely the vocal behaviour of different species of cetacean.  The system has proven very effective in detecting small odontocetes and sperm whales, increasing detection reliability by an order of magnitude during trials conducted off UK.  PAM  probably makes the biggest contribution in detection of sperm whales principally because they can be detected at significant ranges, they are consistently vocal for a large proportion of the time and because “Rainbow Click” is so well developed as a sperm whale detection and analysis tool.  Smaller odontocetes such as dolphins, killer whales, pilot whales and dolphins can be detected at useful ranges but they often move so quickly that target motion analysis is typically not effective.  The effective range for harbour porpoise ~400m is usually within most mitigation ranges so that any reliable detection should lead to action.  Beaked whales produce promising signals (moderately high frequency and narrow band), but we have insufficient knowledge of their vocal behaviour to determine how PAM can best be used with this species.

The value of PAM in mitigating the effects of seismic on baleen whales has yet to be fully explored.  These species generally vocalise at low frequencies which will be vulnerable to masking by vessel and flow noise.  Further, although some baleen whale vocalisations are very powerful, they appear to be less consistently vocal than most odontocetes.  Many of their vocalisations appear to be breeding calls and may be produced seasonally and  be produced solely or predominantly by males.  It may also be the case that some baleen whales are less vocal in on-shelf than they are in oceanic waters.

While the system was initially developed and deployed from a guard vessel recent deployments from a seismic vessel indicate that there are few problems and many advantages to working on the seismic survey vessel itself.

Like other existing PAM equipment the system we use does not yet provide an instantaneous location for detections, and while reliable locations for vocalising sperm whales can be determined in some cases we do not believe these techniques can be applied to many other species.  Other methods of determining range, including statistical analysis of received levels and analysis of surface and bottom reflections could be explored.  Mitigation protocols could also be adapted to make better use of the acoustic cues that PAM systems can provide.  Deployment of larger two or three dimensional arrays could provide real time location ability.  This would require new software but this development would carry little technical risk and could incorporate many of the detection algorithms from within current programs.  Deployment of large arrays from seismic vessels should be relatively straightforward as these vessels are already adapted for this, and our current marine mammal monitoring hydrophone streamers could be readily adapted to facilitate this.

Send mail to Web@ecologicuk.co.uk with questions or comments about this web site.
Last modified: August 18, 2004