Passive acoustic monitoring helps protect wildlife
Can a new generation of passive sonar systems help man and animal coexist in the oceans? E&T finds out.
The increasing exploitation of the oceans for power generation, shipping and mineral extraction is bad news for wildlife. Human offshore activity can cause underwater noise that affects marine mammals such as seals, whales, and dolphins, as well as disturbing the fish that they feed upon.
In order to protect sea life, many offshore regions have laws that require dedicated visual observers to watch for marine life and signal when human activities, such as exploration, construction or offshore power generation, should be slowed down or even stopped in order to avoid disturbing it. This is costly for the companies involved, and doesn't work well in poor weather or at night. The answer may be to listen more carefully to the local wildlife.
Aquatic mammals, or cetaceans, produce a wide variety of sounds. These range from the low-frequency (15 to 2000Hz) calls of baleen whales, to the impulsive broadband 'clicks' produced by sperm whales (peak frequencies of around 6 to 16kHz) and dolphins (peak frequencies of around 10 to 130kHz). Many dolphin species also produce tonal 'whistles' (at frequencies of 2 to 80kHz), while harbour porpoises produce high-frequency (peaking at around 130kHz) echolocation clicks.
Passive acoustic monitoring
Scientists now want to use passive acoustic monitoring (PAM) systems to listen for these vocalisations, and so ensure the creatures' safety while also eliminating unnecessary delays to sea-borne work.
In theory, PAM can determine whether marine mammals are nearby, even when they cannot be seen. There are certain instances where the distinct clicking noise patterns of a marine mammal, such as a sperm whale, can be clearly detected.
"Whales, especially sperm whales, are very vocal marine mammals, and periods of silence are usually short and most often occur when these animals are at the surface and may be detected using visual observers," explains Dr Chuck Monnett, a marine ecologist in the US Depatment of Interior's Minerals Management Services (MMS), who concentrates on Alaska's outer continental shelf. "However, sperm whales are at the greatest risk of potential injury from seismic airguns [used in exploration] when they are submerged and under the airgun array.
"PAM appears to be very effective at detecting submerged and diving sperm whales, and some other marine mammal species, when they are not detectable by visual observation. The MMS strongly encourages operators to participate in an experimental programme by including PAM as part of the protected species observer programme."
The underlying principles of passive acoustic sensing and the analysis of the resultant data in conjunction with temperature and salinity information are not unique to PAM. But PAM is the first deep-sea instrumentation to enable the study of noise-making marine animals (especially fish) over a wide area. PAM also offers versatility by combining advanced sensing, packaging, and data-processing design features with technology adapted from proven marine instrumentation systems.
Many agencies, such as the MMS in the US and the Joint Nature Conservation Committee in the UK strongly recommend the use of PAM, but do not yet require it. However, the MMS says that if human activity is to be started at night or in poor visibility, "ramp-up cannot begin ... unless ... Experimental Passive Acoustic Monitoring is used."
PAMGUARD: PAM goes open-source
The technology for PAM is still at an early stage of development. There are several commercial systems, which, at their most basic, are used by an operator or software to listen for underwater noises.
The system uses a hydrophone that can be attached to a floating buoy, hung over the side of a ship or towed in an array. Hydrophones are simple to use and rugged, and so offer better monitoring than can be achieved by a human observer in poor visibility or at night. They do however still lack the ability to detect the direction or range of a signal source - and surveyors may need to use a variety of hydrophones, depending on the species being monitored.
While PAM software exists, the source code is not freely available for others to expand and improve. This means that: the underlying assumptions - and therefore the margins for error - are not readily understood; the code evolves slowly; and source-code improvements are at the mercy of the time and resources of the few developers responsible.
Military and some commercial organisations keep their detection, classification and localisation technologies in-house and protected. But for PAM to evolve rapidly, what is needed is an environment that raises the technique's profile and creates a means of tapping into the intellectual resources of the whole research community.
Industry and marine environmentalists are aware of this need, and so an open-source program called PAMGUARD has been developed. The project was set up to provide a standard software infrastructure for the acoustic detection, localisation and classification of marine mammals, and for research into their abundance, distribution and behaviour.
"In its current stage of development it provides a powerful, flexible and easy-to-use program for real-time acoustic detection and localisation of cetacean vocalisations that combines the functionality of several previous software products and, in many cases, extends them," says Douglas Gillespie of the Sea Mammal Research Unit at the University of St Andrews. "Thus, it is well positioned to provide the standard tool for PAM during mitigation operations and towed hydrophone surveys."
As PAM systems become more complex, the amount of data that needs to be analysed is increasing dramatically. Even a relatively simple PAM configuration will handle many gigabytes of data per hour and this data may take many forms. It could include raw audio data coming from some kind of input device, such as a sound card or audio file, ancillary data such as GPS positions and hydrophone depth information, and the output from the various detection and localisation modules.
In the loosely coupled programming framework of PAMGUARD, each module that needs data from another subscribes to its output and is notified when new data become available. Having multiple modules subscribe to the same data increases program efficiency since, for example, the same spectrographic data can easily be used for displays and as the input to a detector.
In some cases, such as with raw audio, the arrival of new data will be regular and frequent. In others, such as the output of a detector searching for a particular sound, it may be intermittent. Some data may be used and discarded immediately. Other data may need to be stored for a considerable time. So the data managers within each module have to query regularly each subscribing module to determine for how long data are required before it can be discarded.
The future for PAM
Determining when, and if, a marine mammal has been disturbed is not easy. Does a change in the direction in which a seal is swimming indicate that it has been disturbed, or that is has just changed its mind? And will the ocean's mammals soon learn to live with the increasing encroachment of human activity into their domain?
The best that everyone can hope for from technology is that an accurate PAM system can be developed that reliably locates marine mammals, so that humans can continue developing their activities at sea without disturbing those creatures for whom it is home.
You can hear some of the sub-sea sounds at the Communications: telecoms, fixed, mobile and satellite web page.