For most of the year, Nassau groupers are solitary. The individual fish, banded in mottled beige and brown and capable of growing to over 50 pounds and about four feet long, skulk in reefs and seagrass beds throughout the Caribbean, ambushing prey and swallowing it whole.
But each winter, the groupers group, sometimes swimming more than a hundred miles to return to familiar spawning sites. Thousands of groupers may congregate in a given location. And while there are many factors, including water temperature, that influence the fishes’ amorousness, it generally waxes and wanes with the lunar cycle, peaking for up to five orgiastic nights following a full moon.
Fifty years ago, the largest spawning aggregations reportedly numbered in the tens of thousands. Nassau grouper was among the most common commercially fished species in the Caribbean. This is no longer the case. Their seasonal gatherings and inquisitive disposition made them relatively easy pickings for fishermen and a target for over-exploitation. Nassau grouper are now considered endangered by the IUCN and threatened by NOAA under the Endangered Species Act.
How many fish are there, and where are they? How big? How many males; how many females? These seemingly basic questions, on which sustainable commercial fishing and aquatic conservation both depend, require information that can be difficult to gather and imprecise when it is. We’re up here and the fish are down there; the waters are vast and the depths inscrutable.
Traditionally, marine scientists might scoop up samples of fish with a large net, assess the physical or genetic attributes of what they find, and then extrapolate to entire populations in larger areas. Under the right circumstances, though there is another approach that can help answer some of these difficult questions, a method that has until recently been largely ignored as a source of information about the health of fish populations. And the Nassau grouper is a good candidate for study, because the Nassau grouper makes sound.
The Nassau grouper has an alarm call, a series of low, booming grunts. It also makes a sound associated with courtship, a bit like a rumbling wubwub dubstep bass drop. Once a sound has been identified with a species, its presence can be mapped in time and space—to monitor, for example, the advance of a species considered invasive. A link between a specific sound and a specific behavior, like the courtship rumble, provides even more information.
“We’ve been developing our technology to detect when and where they are spawning,” said marine biologist Michelle T. Schärer, of HJR Reefscaping in Puerto Rico, who found the association between sound and courtship behavior in Nassau grouper. This information can help set regulations to protect the grouper when and where they are most vulnerable, while limiting the burden on fishermen.
Nassau groupers are not the only fish that make sound, and there is a growing field of research focused on listening to fish with specialized underwater microphones, called hydrophones. This aural approach offers a minimally invasive way to survey an area for a species of interest. Scientists have only recently overcome many of the technical limits to recording fish, allowing researchers to explore a relatively untapped source of data, and hopefully of insight.
The field is mostly still in early stages. And one grand goal of conservation and fisheries, the ability to measure fish populations by sound, or even roughly estimate them, is still a long leap beyond the current understanding of the best-studied fish. Still, in cases like the Nassau grouper, listening could reveal critical behavioral patterns. “That’s very powerful, and very difficult information to get by conventional means,” said Rodney Rountree, an inveterate fish listener. “It’s not just about counting fish.”
“People have known for probably hundreds of years that some fish make noise,” said Joe Warren, an underwater acoustician at Stony Brook University. He notes that the drums, the grunts, and the croakers, sometimes called frogfish, are all named for the sounds they make.
But human ears have historically only picked up on the loudest and most obvious noisemakers. Exactly how many others of the maybe 33,000 species of fish make detectable and identifiable sounds is unknown. Acousticians have studied relatively few, but it’s clear that there’s great diversity in fish physiology, and accordingly, the ways that they sense and produce sound. “Fish can produce drumming sounds, popping sounds, clicking sounds — all sorts of onomatopoeia,” said Rosalyn Putland, an underwater acoustician at the University of Minnesota Duluth. “Sometimes those words don’t exist yet, of what these sounds should be called.”
There’s stridulation, scraping of fins, bones, or teeth; drumming, expansion and contraction of an air-filled swim bladder; or the expulsion of air from that swim bladder. This last group of sounds—amounting to belches, coughs, and farts—come from some common, recognizable types of fish. They include eels, herring, catfish, and salmon. While many of these noises are probably incidental and non-communicative, they may still be used to identify species, Rountree said. This was recently the case with several freshwater fish species he was studying, including three kinds of trout.
Sound-making fish may have been overlooked in modern times because the tools of marine science— boat engines, diving gear — produce noise of their own. But a hydrophone left to silently listen can capture a clearer picture of underwater sound. Stationary hydrophones can be anchored to the seafloor or floated up into the water column and left for months. Carrie Wall, who leads an archive of underwater acoustic data at the National Centers for Environmental Information, said she has even mounted hydrophones to underwater gliders, instrument-laden drones that can be programmed to transect a body of water for weeks at a time.
“There are a lot of sounds out there made by fish, and we have no idea what they are,” Wall said.
Sound is a useful source of information to fish whenever visibility is low: at night, when the waters are deep and dark, or when the surroundings are a bit murky. These same conditions make it difficult to verify that a particular sound was produced by a particular fish, because connecting audio with a particular species of fish typically requires visual evidence from underwater cameras. And that proof must be collected in the the wild if the fish doesn’t make a full range of sounds in a laboratory setting.
Marine acoustic tools have roots in submarine warfare. Scientists can interpret the echoing pings of sonar as masses of fish and other sea life. And during the Cold War, under the direction of the U.S. Navy, a woman named Marie Fish began identifying Atlantic fish by sound as a way of distinguishing them from watercraft. The two sound-based approaches—active and passive, echolocating and listening—are now regarded as complements to one another. Sonar, for example, is less effective with bottom-dwelling fish and samples a smaller area, but it provides biomass estimates and is able to locate quiet fish.
In the early days of underwater recording, there were many technical challenges. Salt water damaged electrical equipment. Short-lived batteries and scant storage capacity limited the total time that could be recorded. But field researchers and engineers have addressed many of the practical problems, developing tricks for waterproofing their equipment, all while smaller, longer-lasting batteries and data solutions became commercially available. Underwater acousticians now report two major challenges to passive acoustic monitoring of fish: confirming the origins of unknown sounds; and processing the enormous volume of data that improved recording devices can now collect.
There’s also enough variety in the behavior of sound-making fish that recording techniques must be tailored to the species being studied. Putland and colleagues have triangulated the location of toadfish living in turbid water by recording with an array of hydrophones. Minute delays in the arrival of male toadfishes’ honking “boat whistles” were enough to calculate differences in distance from each recording device. While this fine-grained locating would not be possible with many other detection instruments, it’s only feasible in this case because male toadfish build nests from which they call out to potential mates. They are largely sedentary.
Tasks like this require significant computational resources and mathematical know-how, expertise that has an already interdisciplinary field collaborating with computer and data scientists. “It’s no longer the ability to collect data, it’s the ability to analyze that’s the limiting factor,” Wall said.
For example, Schärer and colleagues have identified sounds from seven different Caribbean grouper species and counting. But months worth of recordings can also take months for researchers to label, listening and poring over visualizations called spectrograms. So they’ve begun experimenting with automation to distinguish grouper calls by species, even applying the type of artificial intelligence called deep learning to a library of grouper sound.
“The problem with this type of sound is it’s not like the human voice,” said collaboratorAli K. Ibrahim, a Ph.D. student in electrical engineering at Florida Atlantic University. That is, compared to human voices, grouper calls are often low-pitched, and fall within the same range as many types of environmental noise. Lower frequency sounds also travel farther than high frequency sounds, so hydrophones pick up rumbling from miles away. The group’s current approach requires pre-processing to help pick out the fishy signal from the ambient ocean noise.
Fish sounds are part of a larger soundscape which includes all manner of biological noises, including less-studied underwater life, such as shrimps and lobsters. There are physical sounds, like wind, waves, and weather; even earthquakes and shifting ice. And increasing attention is being turned, experts say, to the effects of anthropogenic noise on fish and other animals, many of which rely on sound to navigate through the world, whether they produce it themselves or not.
Fish sounds are also just one of many potential sources of information about fish populations, our effects on them, and their reactions to changing climate and environments. Rountree likes to say, “It’s another tool in your toolbox.” But it’s one that remains mostly unhoned and unexplored.
There’s an ocean of sound. What can we do with it?