Whale skeletons lie along the coast of Fuerteventura, Canary Islands (Spain), as a brutal reminder of the harmful effects of military sonar. It is believed that the sonar of ships and submarines is one of the factors that contribute to the beaching of whales, confusing the animals’ own sonar and causing them to head for beaches.
But this technology harmful to whales may soon have a competitor.
Lori Adornato, project manager at the US military research agency Darpa, believes we can detect submarines if we pay more attention to natural sounds, rather than just firing sonar pulses.
“We currently treat all these natural sounds as background noise or interference, and we’re trying to change that,” says Adornato. “Why not make use of these sounds and see if we can find a signal?”
His project, called Persistent Living Aquatic Sensors (Pals), “listens” to marine animals as a way to detect underwater threats.
Current, air-launched sonar buoys — developed by the military to detect underwater enemy activity — only work for a few hours in a small area, due to limited battery life. The Pals system will be able to cover a wide region for months.
It will be able to provide near-constant monitoring of shorelines and underwater channels. Adornato says that species that inhabit reefs and reliably stay in place are likely to be the best sentinels. “You need to be sure that your body will always be there”, says the specialist.
The Pals project is funding several teams to investigate different lines of research, using very different species that live on the reefs.
Laurent Cherubin, from the Atlantic University of Florida in the United States, coordinates a team researching the black grouper, a common fish in American waters, which can weigh up to 300 kg and is known for making loud sounds.
“It’s a loud sound, but at a low frequency,” explains Cherubin. “Black groupers are territorial and often make these sounds when they spot an intruder on their territory.”
The sounds produced by this fish can be detected from 800 meters away, but the objective is not always to ward off intruders and predators. Black groupers also make sounds for mating, delimiting territory and other purposes, which still remain a mystery.
The team is now focused on warning calls, something like trying to hear the bark of a guard dog against intruders, according to Cherubin. Differentiating these calls from the rest is not easy, so the team created machine learning algorithms for this task. It took thousands of recordings before the algorithms were able to distinguish and classify different grouper calls.
The algorithm can then be turned into software, which is driven on a small but powerful processor installed in an underwater microphone, or hydrophone. A set of hydrophones can cover a reef, listening to grouper calls and tracking them as the cause for the calls moves from one grouper territory to another.
Analyzing conversations between fish may seem bizarre, but the work of Project Pals at defense systems provider Raytheon is much more like traditional anti-submarine sonar—in fact, with one important difference.
“We are trying to detect the echoes created when the shrimp noises are reflected by the vehicles”, says scientist Alison Laferriere, from Raytheon. “It’s very similar to a traditional sonar system that detects echoes of sound generated by its source.”
In other words, the system works like other normal sonar, but uses shrimp-produced noise instead of artificial ones.
The pistol shrimp, also called the snapping shrimp, is known to emit the loudest sound produced by any living creature on Earth — 210 decibels (by way of comparison, the takeoff of a jet plane generates 120 decibels).
This invertebrate produces its characteristic snap by closing its pincers so quickly that it generates enough sonic energy to stun prey.
Pistol shrimp also communicate with each other. A group of pistol shrimp emits a constant roar that Laferriere compares to the sound of frying bacon.
“The signal created by a pistol shrimp has a very short duration and an incredibly wide amplitude,” says Laferriere. “A single crackle is much quieter than a traditional sonar source, but there can be thousands of crackles being triggered per minute.”
Laferriere explains that the sound varies with the time of day and the temperature of the water, but a shrimp colony is never silent.
“One of the biggest challenges we face is dealing with the huge amount of noise created by the pistol shrimp and the reflections of all those sounds off the surrounding area,” says Laferriere.
Interpreting these reflections is quite a challenge as, unlike traditional sonar, the location of the sound source is unknown. Again, the solution is to use modern software.
Laferriere’s team developed clever algorithms to analyze the sound and select a single click, first calculating the location of the shrimp, then the path taken by the reflected sound, and finally deducing where it was reflected.
To understand the return sound, Laferriere’s team needed to create computer models to determine which echoes came from background objects and could be ignored. Eliminating these echoes highlights objects moving through the environment — which could be fish, submarines or unmanned underwater vehicles.
Once again, the final solution will be a set of smart hydrophones with on-board computing, capable of processing shrimp sounds and determining the location of possible targets of interest in the region.
Other teams on the Pals project have been carrying out similar studies. Northrop Grumman researchers are working on another sonar system using shrimp, and a US Navy team is researching the general sounds of reefs and how intruders affect them.
All promise an immense sensor network that will cover wide regions for extended periods, with most of the hardware conveniently provided by nature. Only the hydrophones will need repairs or replacements.
It will work?
“The DARPA study will be a really important innovation if it succeeds,” says Sidharth Kaushal, an expert on naval military equipment at the British think tank RUSI. “An ecosystem of scattered living sensors in permanent fluctuation is, in principle, tempting.”
In principle, but not necessarily in practice. Kaushal has his doubts because previous projects using marine life to detect submarines have not been successful.
German submarines were even identified because of the effect on bioluminescent plankton, which emit a bright light when disturbed — and one of these submarines was reportedly sunk thanks to this effect in World War I. But later attempts to use this phenomenon more broadly, with special sensors looking for light sources over a larger area, have met with little success.
“Soviet and American efforts during the Cold War to use them systematically came to nothing,” says Kaushal. “Partly because they had no way of differentiating false positives, like the reaction of a passing whale, from the real object.”
It is not yet known what the quality of the distinction that the Pals will be able to make between a submarine and a shark, for example. Lori Adornato believes that the combination of marine organisms and modern intelligent algorithms will provide a reliable “path warning” to guide more traditional submarine hunters to check for a possible intruder.
Adornato says that the technologies developed for the Pals project could also be used for scientific research, monitoring reefs and other underwater environments with a set of sensors.
“These low-impact observation systems can be developed in many different environments without harming the ecosystem established by nature,” she says.
Tuning in to the sounds produced by normal marine life and learning how they change would offer researchers a cheap, eco-friendly way to track the impact of human activities underwater. This would be useful for projects like wind generators, oil extraction and deep sea underwater mining, as all we would need to do is listen to nature.
The project focuses on species common in US territorial waters, so their replication in other regions would not necessarily be easy. But technology, in general, can be applied more broadly.
The Pals project has completed its first phase, which was a feasibility study for two different approaches, monitoring the reactions to intruders of reef-dwelling species and the sonar of the pistol shrimp. The developers are now working on a second step, to demonstrate how their solutions work in controlled testing, in the Northern Hemisphere summer of 2022.
Have you watched our new videos on YouTube? Subscribe to our channel!