Like many of you reading this, I am a diver. On land, I work with one of the leading companies that are developing Artificial Intelligence (AI) algorithms for analyzing industrial sounds. In theory, we can monitor any machine or production line that exists using sound, and in the process, maximize output, lower operating costs, optimize maintenance, and reduce the impact of industrial production on the environment.
You might wonder where I’m going with this but I promise there is a connection: As divers know, sound is plentiful in the ocean and it is certainly not exclusive to just the charismatic marine mammals. Entire underwater environments such as coral reefs, have their own distinctive landscapes of sounds, or “soundscapes”. The fish and other creatures living on coral reefs make a vast range of sounds which can tell us a lot about both their health as well as the health of the reef.
In essence reefs are huge machines humming along happily like a well-oiled piece of equipment when healthy, or screeching and screaming when under threat of a breakdown.
Monitoring the oceans is a huge feat, ranging from satellites in space, spot-checks of water chemistry, hitching trackers to animals, to manual labour-intensive visual surveys. All of these methods come with limitations that make it difficult to assess the ocean’s health; imagine heading out on a dive to conduct a fish survey, knowing that half the creatures are nocturnal, camouflaged, or just simply tucked under a head of coral that you won’t see. Given the threats our oceans - and reefs in particular – are facing, accurate and rapid measurements are paramount in our fight to protect them.
Today’s technology can pick up sounds and patterns that are undetectable to the human ear and hold a wealth of data. A growing number of studies have confirmed relationships between the soundscapes of marine habitats and their health conditions. Surveying reefs using bioacoustics can therefore provide new understandings that may explain both ecological and behavioural processes. A big bonus here is that sound travels incredibly well underwater, so the data collected is much less prone to human error or bias. While there is still a lot left to learn to about this field, bioacoustics are likely to play an important role in understanding our marine world.
What if we took it one step further though and applied AI and machine learning to these soundscape analyses in our efforts to capture much more accurate and faster assessments of reef conditions?
Although coral reefs are busy producing a cacophony of tangled sounds - not dissimilar to a production line with hundreds of different machines - AI methods can now distinguish between the overall sounds of healthy and unhealthy reef components, the same way as we can distinguish between a well-operating piece of equipment and machines that are on the edge of a break-down. Imagine receiving data in real time using non-invasive tactics to see if our conservation efforts are working in some parts of the reefs compared with others?
We at Scubavox are fascinated by the idea of using AI with bioacoustics systems around the globe to monitor the health of our reefs. We try to do a lot of talking on behalf of the reefs, but perhaps it is time we stopped to listen to what the reefs have to say themselves?