Autonomous recording for marine ecology

Cetacean species, including whales, dolphins and porpoises, are considered indicators of the health of marine ecosystems around the world. While a number are known to be endangered, a lack of data means that the population size and conservation status of many species are impossible to estimate. These animals are vulnerable to the effects of human activities and the noise they cause.

In Brazil, researchers carry out underwater acoustic monitoring to assess the ecological impact of industrial activities on the coast. As well as quantifying human‑generated noise, this type of study is very useful for scientists studying cetaceans, because the efficient transmission of sound in water means the tones and clicks they produce can be detected hundreds of kilometres away. However, commercial underwater recorders are expensive and inflexible, with proprietary software and hardware that is difficult or impossible to modify. Earlier this summer, though, a team from the University of São Paulo in Brazil published a paper about the flexible, low-cost autonomous recorder they have built, based on a Raspberry Pi, in open-access journal PLOS ONE. This is what it looks like:

The hydrophone – an underwater microphone – is on the top, protected by the cage that you can see, which is made of stainless steel; when deployed, all of the other components are inside the 50cm PVC case on the right. With walls approximately 9.5mm thick, this enclosure successfully withstands pressures of up to 10 bar, equivalent to those experienced almost 100m underwater, in pressure‑chamber testing.

Output from the hydrophone is passed via a signal-conditioning board and then a USB audio codec including an analogue-to-digital converter before being processed and stored by the Raspberry Pi. There’s a battery pack of five ordinary D-size Duracell batteries, with room in the enclosure to add four more such packs in parallel, and a power management module including a real-time clock.

So that the device doesn’t have to consume power during transport to the deployment location, the power management unit incorporates a Hall effect latch, controlled by a magnet on the outside of the enclosure, to connect or disconnect the batteries via a relay. Once the unit has been deployed, the real-time clock can control the relay to power the Raspberry Pi on or off at scheduled times. For their tests, the team used a 128GB SD card, one of the largest compatible cards they could find, although the limiting factor for autonomous functioning of the recorder proved to be power rather than data storage capacity.

The team deployed their autonomous recorders to locations on the eastern and southeastern coast of Brazil for field-testing, and they all performed satisfactorily, monitoring marine traffic and whale and dolphin populations. From the results of their tests they estimate that with the maximum number of five battery packs, the devices could provide almost two weeks of continuous recording, or over four months of recording at one hour per day. They used a Raspberry Pi Model A; the Model A+, smaller and with even lower power consumption, would eke out the power for longer.

The recorder has various settings that users can alter to optimise for different mission requirements: the scheduling of recording times and the nature of any automatic post-processing can be adjusted, as can the recording sample rate (the whistles and clicks of dolphins are best captured at a higher sample rate than low-frequency whale vocalisations, for example). At an estimated cost of US $500, it should be an attractive option for research groups faced with the alternative of splashing out six times as much on a less customisable commercial device.

It’s very good indeed to see Raspberry Pi used to build low-cost open hardware for research and study. The last time I poked around the web looking for open labware, there were some encouraging examples, but they were a little thin on the ground; now the most slapdash of searches returns a clutch of exciting results, from OpenTrons’ crowdfunded liquid-handling robot to a <$100 fluorescence microscope via my personal did-they-really-make-that? favourite, 3D-printed Raman spectrometer ramanPi. Setting up a research group or a teaching lab in a few years’ time might be a very different thing to what scientists have been used to.