Why Whales Sing Underwater

A fin whale's call lasts less than a second and sits near 20 hertz, a pitch most people feel more than hear. Yet hydrophones (underwater microphones) have confirmed those pulses at distances of about 350 kilometers. Whales manage this by combining two physical facts: low-pitched sound loses very little energy in seawater, and a layer of ocean called the SOFAR channel traps sound and carries it sideways. This article covers how that channel works, how whales make such sounds, what humpback song and sperm whale clicks actually show, and what rising ocean noise does to them.

What the SOFAR channel is and why it exists

SOFAR channel, short for Sound Fixing and Ranging channel and also called the deep sound channel, is a horizontal layer of ocean where the speed of sound reaches a minimum. Sound waves that wander above or below that layer bend back toward it, so a signal launched near the layer stays confined instead of spreading in every direction. At mid-latitudes the layer sits at roughly 1,000 meters (about 3,300 feet) deep, according to the Discovery of Sound in the Sea project, and its exact depth shifts with latitude and season.

The minimum exists because temperature and pressure push sound speed in opposite directions. Near the surface, sun-warmed water carries sound quickly, but the water cools with depth and sound slows down. Below about 1,000 meters the temperature levels off near 2 degrees Celsius, so pressure takes over and sound speeds up again with depth. The minimum sits where one effect hands off to the other.

Think of the layer as a valley in a landscape of sound speed. A marble rolled along the valley floor stays in the valley, because every sideways nudge is met by a slope that rolls it back. Sound behaves the same way: rays that tilt upward enter faster water and curve back down, and rays that tilt downward enter faster water and curve back up.

How oceanographers found it

The discovery came from physics, not from biology. In spring 1944, Maurice Ewing and J. Lamar Worzel sailed from Woods Hole, Massachusetts, aboard the research vessel Saluda, which lowered a deep hydrophone. A second ship set off 4-pound explosive charges at depth, at distances of up to about 900 miles, and the Saluda recorded a distinctive signal that built to an abrupt climax (DOSITS history of the SOFAR channel). In 1946, Leonid Brekhovskikh and colleagues in the Soviet Union independently found the same effect while experimenting in the Sea of Japan.

The name came from a rescue idea. The U.S. Navy imagined that survivors of a downed aircraft or sunken ship could drop a small charge set to explode in the channel, letting listening stations triangulate their position from thousands of kilometers away. That is where "sound fixing and ranging" comes from.

In the 1950s, the Navy's Project Jezebel built bottom-mounted hydrophone arrays that became the Sound Surveillance System (SOSUS), designed to track submarines. Early operators logged strange low-frequency sounds nicknamed the "Jezebel Monster." They were later identified as the calls of blue and fin whales (DOSITS).

Do whales really use the channel on purpose?

Researchers have long suspected that baleen whales (the filter-feeding group that includes blue, fin, and humpback whales) benefit from the channel, because their calls are loud, low, and travel far. Evidence for deliberate positioning is thinner than popular accounts suggest. The channel is a property of the ocean, not an adaptation, and whales that call near the surface or in shallow water still reach neighbors, just over shorter distances. The safest statement is that low-frequency calls are well suited to long-range propagation and that long-range detections of fin and blue whale calls are well documented.

Detection ranges also vary. A NOAA technical memorandum on SOSUS reported a confirmed detection distance of about 350 kilometers for fin whale pulses, and noted that detection distances vary with seasonal ocean conditions (NOAA Technical Memorandum AFSC-107). That is far enough to matter for communication, and it is also a reminder that "across an ocean" is a ceiling under ideal conditions rather than a typical day.

Why low frequencies travel so much farther underwater

Frequency, not just volume, decides how far a sound survives in seawater. As a sound travels, the water converts some of its energy into heat, a process called absorption, and that loss climbs steeply with frequency. At the same starting loudness, a high-pitched sound fades over a much shorter distance than a low-pitched one.

A foghorn makes the point. Ships use deep, low blasts because a low note carries across miles of open water, while a piccolo would be lost within a few hundred meters. The same principle explains why a neighbor's subwoofer is easy to hear through a wall while the vocals are not: low frequencies lose less energy to whatever medium they cross, whether that medium is drywall or seawater. For a refresher on how pitch, wavelength, and energy relate, see our explainer on how sound works.

Fin and blue whale calls sit very low. The Discovery of Sound in the Sea project describes blue whale calls as typically 15 to 40 hertz, often below the range of human hearing, and fin whale 20-hertz pulses as downward sweeps from roughly 40 hertz to 13 hertz. Healthy young adults hear down to about 20 hertz, so some of this singing is at or beyond the floor of our hearing. If you want the human side of that limit, our piece on how hearing works covers it.

Humpback whales take a different approach. Their songs use a much broader band of pitch, with harmonics reaching above 20 kilohertz in some measurements (Au et al., Journal of the Acoustical Society of America, 2006). That breadth costs range, since the higher components fade faster, but it buys a richer palette of sounds. Every communication system trades reach against information density, and these species sit at different points on that trade.

Species	Typical call frequency	Reported source level	Likely function
Blue whale	15-40 Hz (song and D calls)	About 189 dB (Antarctic blue whale)	Song; D calls are short downsweeps given by both sexes and thought to be contact calls
Fin whale	20-Hz pulses sweeping about 40 to 13 Hz	Up to about 190 dB	Male song bouts; a separate 40-Hz call is linked to feeding areas
Humpback whale	Broad band, harmonics above 20 kHz	About 151 to 173 dB maximum within a song	Male breeding-season song; possibly male-male ordering
Sperm whale	Below 100 Hz to 30 kHz, most energy 5-25 kHz	Peak-to-peak levels up to 236 dB reported	Echolocation clicks; social codas

Source levels are in decibels referenced to 1 micropascal at 1 meter. Sources: DOSITS species pages; Au et al. (2006); Møhl et al. (2003). Peak-to-peak and average levels are not interchangeable, so compare across rows with care.

Direct comparison to airborne decibels is misleading, because underwater sound uses a different reference pressure than sound in air. A 189 dB whale call is not "louder than a jet engine" in any meaningful way, even though headlines sometimes say so. Our article on the loudest sounds ever recorded explains why decibel scales need a stated reference before they can be compared.

Why do blue whale calls travel farther than humpback songs?

Blue whale calls pack most of their energy into a narrow, very low band around 15 to 40 hertz, which loses little energy to absorption. Humpback songs spread energy across a wide band that includes much higher components, and those fade faster, which shortens their practical range. Exact ranges depend on depth, season, and background noise, so published figures vary.

How whales actually make the sound

Whales do not have vocal cords in the way people do, and for decades how baleen whales produced such low sounds was unclear. In 2024, a team led by Coen Elemans and W. Tecumseh Fitch published a study in Nature (volume 627, pages 123 to 129) titled "Evolutionary novelties underlie sound production in baleen whales." They blew air through larynxes taken from three stranded whales: a minke, a humpback, and a sei whale. They found that the arytenoid cartilages, small structures that in most mammals help open and close the airway, had been reshaped into a large U-shaped rigid structure that presses against a fatty cushion inside the larynx. A computational model of that arrangement predicted the natural frequency range of the whales' calls.

The finding comes with a trade-off the authors emphasized: the same anatomy that produces deep, loud calls also limits how flexibly the larynx can change pitch. That matters for noise, because a whale that cannot easily shift its voice upward has fewer options when ships raise the background level. The researchers also noted that the larynx of an air-breathing mammal that feeds underwater has another job, which is keeping water out of the airway, so the structure serves several demands at once.

Toothed whales, the group that includes sperm whales, use a completely different system. According to DOSITS, sperm whale sounds originate in the nasal passages and sacs, and the spermaceti organ is thought to help focus them. The result is a stream of clicks rather than a sustained tone, built for aiming sound at prey and listening to the echo. Our article on how bats hunt with echolocation describes the same basic idea in the air: send out a sharp pulse, then read the returning echo to build a picture of the surroundings.

Payne and McVay, and what humpback song structure actually shows

The scientific study of whale song as a structured phenomenon traces to a 1971 paper. Biologists Roger Payne and Scott McVay published "Songs of Humpback Whales" in Science (DOI 10.1126/science.173.3997.585), describing a hierarchy. The shortest continuous sound a human ear can pick out is a unit. Units combine into phrases, repeated phrases form themes, and a string of themes makes up a song that is then repeated, sometimes for as long as about half an hour in Payne's accounts. That vocabulary is still used by researchers today.

Two features make humpback song scientifically interesting. First, the song changes. All singing males in a population sing a version of the same song at a given time, and the song evolves from season to season, with phrases added and others dropped. Second, the changes spread. In a study published in Current Biology in 2011, Ellen Garland and colleagues analyzed recordings from populations across the western and central South Pacific over 11 years and found that new song types moved repeatedly in one direction, eastward, like cultural ripples (abstract at the University of St Andrews). Press coverage of the work described the songs starting in the large eastern Australian population and reaching French Polynesia, with the authors attributing the direction to the size of the source population.

The closest human parallel is a pop song. A tune starts in one scene, passes to neighboring groups, and displaces the previous hit within a year or two, long before anyone can say exactly who passed it on. The whales are not carrying the songs in their bodies from one ocean to another; the evidence points to males hearing and learning songs from neighbors.

Who sings, and when?

According to DOSITS, male humpbacks sing on the breeding grounds, and recent work has found them singing on feeding grounds as well. Proposed functions include a male breeding display, male-male social ordering, and spacing of reproductively active males. Those are hypotheses; the song's exact role has not been pinned down, and several of the proposals may be true at once.

Does humpback song carry meaning?

Humpback song has hierarchical structure, repeats with precision, and changes culturally, which is strong evidence that it functions as communication, probably tied to mating. What the evidence does not show is that individual elements refer to specific things the way words do. Structure alone does not establish language: birdsong is highly structured too, and no ornithologist claims sparrows are discussing the weather. Whether humpback song carries referential content remains an open question.

Blue and fin whales: songs built for distance

Blue whale song, as DOSITS summarizes it, consists of regularly repeated units that last about 15 to 20 seconds each and can combine into phrases and bouts. Blue whales also produce D calls, short downward-sweeping calls given irregularly by both sexes and thought to serve as contact calls among groups of two or more whales. Song types vary geographically, which lets scientists tell populations apart from sound alone.

Fin whales are more regular still. Their familiar 20-hertz pulse is the most commonly reported fin whale sound and has been recorded worldwide. DOSITS reports that pulses occur singly or in doublets and triplets, with 6 to 37 seconds between notes, and that males produce song bouts that can last up to 32.5 hours, broken by silent gaps of 120 to 150 seconds roughly every 10 to 15 minutes. Some pulses carry an extra component near 85 to 140 hertz, and the frequency of that component differs by region, near 85 hertz off the Western Antarctic Peninsula and near 99 hertz off eastern Antarctica. Regional dialects are not only a humpback phenomenon.

Why sing for a day and a half at a stretch? The usual hypothesis is advertisement: a male signaling his presence and condition over long distances to females and rival males. Fin whales are hard to observe directly, and the reasoning rests on circumstantial patterns such as the song's seasonality, so treat it as a leading hypothesis rather than a settled function.

The 52-hertz whale

One of the odder cases in bioacoustics came from the Navy's hydrophones. In 1989, a Woods Hole Oceanographic Institution team picked up an unusual signal in the North Pacific. In a 2004 paper in Deep Sea Research Part I, "Twelve years of tracking 52-Hz whale calls from a unique source in the North Pacific," William Watkins and colleagues concluded that it came from a single animal whose movements seemed unrelated to the presence of other whale species. The call sat near 52 hertz, higher than the typical 10 to 40 hertz calls of blue whales and the roughly 20-hertz pulses of fin whales.

The species of the caller is still unknown. Proposed explanations include an undiscovered species, a hybrid of a blue and a fin whale, and an anatomical abnormality. Popular retellings often add that the whale is "the loneliest in the world," which is a story about human feelings. The data show only that a single animal produced a distinctive call that tracked its own path through the ocean.

Sperm whales speak in a different acoustic language entirely

Sperm whales produce clicks across a very wide range, from below 100 hertz up to 30 kilohertz, with most of the energy between 5 and 25 kilohertz (DOSITS). Several click patterns serve different jobs: directional clicks for echolocation, regularly spaced clicks interrupted by "creaks" while foraging, very rapid clicking during the final approach to prey, and codas, which are stereotyped patterns of clicks used in social situations. DOSITS notes that dives reach about 1,000 meters and that breath-holds can last up to 90 minutes, with squid as the main prey. In the dark at that depth, sound is the main sense available.

The loudness of the clicks is remarkable. A 2003 study by Bertel Møhl and colleagues in the Journal of the Acoustical Society of America reported peak-to-peak source levels as high as 236 dB re 1 micropascal at 1 meter. That figure is a peak value, not an average, which is one reason it is not comparable to the roughly 189 dB average-style figure for blue whales. For more on sound extremes, see the loudest sounds ever recorded, which places these animals in context.

What the 2024 coda study found

Codas are the socially used click patterns, and they have been studied for decades by the Dominica Sperm Whale Project. In May 2024, Pratyusha Sharma and colleagues, including Project CETI researchers, published "Contextual and combinatorial structure in sperm whale vocalisations" in Nature Communications. They analyzed 8,719 codas from about 60 whales recorded between 2005 and 2018. The paper identifies four features, which it calls rhythm, tempo, rubato, and ornamentation. Rhythm and tempo behave as context-independent building blocks. Rubato (slight stretching and compressing of the pattern) and ornamentation (an extra click added to some codas) vary with context, and the authors report that the repertoire contains nearly an order of magnitude more distinguishable codas than previously recognized.

That is evidence of a combinatorial system, meaning a small set of elements reused and recombined to produce many distinct signals, the way a limited set of sounds in human speech builds a large vocabulary. The author list includes computer scientists, and Project CETI is built around applying machine learning to whale recordings, because thousands of recordings are too many for a human listener to sort by ear.

The study does not claim that codas carry word-like meaning. The authors describe a coding system and leave semantics open. That caution is deliberate: a combinatorial structure shows that the signals are systematic and context-sensitive, which is a real and significant finding, but it is a much weaker claim than saying sperm whales have grammar or reference.

How scientists listen

Most of what we know about whale sound comes from passive acoustic monitoring, which means leaving hydrophones in the water and recording whatever passes. The method has several advantages. It works at night, in bad weather, and in places where whales rarely surface within sight of a ship. It can run for months, and it hears whales that are submerged and silent to the eye. SOSUS, the Cold War array that started the field by accident, still shows what this approach can do over basin-scale distances.

Active sound works differently. Sonar sends out a pulse and measures the echo, as explained in our article on how sonar maps the seafloor. Sperm whale echolocation is a natural version of the same technique.

Even with decades of recordings, the ocean is poorly sampled. Our overview of why the ocean is still mostly unexplored shows how much of the volume has never been instrumented, and any claim about "all" whale calls should be read as a claim about the places and years where someone left a hydrophone.

Why the ocean has gotten louder, and what that costs whales

Low-frequency background noise in the open ocean has risen. Frisk's 2012 analysis in Scientific Reports, titled "Noiseonomics," found an increase of about 3.3 dB per decade between 1950 and 2007 and tied it mainly to commercial shipping and global economic growth. The author notes an important limit: the low-frequency data come largely from the northeast Pacific, so the trend is better documented there than everywhere else.

Shipping, seismic surveys for oil and gas, and naval sonar all produce substantial energy at low frequencies, and that is where fin and blue whales call. This is the heart of the problem. The frequencies that allow a container ship's engine noise to travel so far through the ocean are the same ones that allow a whale's call to do the same.

The clearest physiological evidence comes from a natural experiment. Rosalind Rolland and colleagues had been collecting fecal samples from North Atlantic right whales in the Bay of Fundy, Canada, to measure stress-related hormone metabolites. After the September 11, 2001 attacks, ship traffic in the bay fell, noise dropped by about 6 decibels, and the largest reduction was below 150 hertz. Baseline stress-hormone metabolites were lower in samples from the quieter period and rose again in later years as ship traffic and noise returned (summary in ScienceDaily). The paper, published in Proceedings of the Royal Society B in 2012, described this as evidence that low-frequency ship noise may be associated with chronic stress in whales. It is an association drawn from a single bay, one species, and a natural experiment, so it is suggestive rather than a controlled demonstration.

Behavior changes too. In a 2010 study in Biology Letters, Susan Parks and colleagues used tags on 14 right whales and found that as background noise rose, the whales increased their call amplitude, a pattern comparable to a person raising their voice in a loud room (paper on PubMed Central). Changes in call frequency and duration were not significant over the short periods studied. The authors suggested the change might help preserve communication range, and noted that this still needs testing.

Louder calling is not free. It takes more energy, and it can make animals easier for passive acoustic monitors to detect, which is useful for conservation but shows how tightly call behavior is tied to noise. The Elemans larynx study adds a further consideration: if the anatomy limits how far a baleen whale can shift its voice, the room to adapt may be narrower than for animals with more flexible vocal organs. That is an inference from the anatomy, not a measured result.

Source of low-frequency noise	Where it overlaps with whales	What the evidence shows
Commercial shipping	Broad low-frequency energy, strongest below a few hundred hertz	Background rise of about 3.3 dB per decade, 1950 to 2007, in northeast Pacific data
Quieter shipping periods	Same band as blue and fin whale calls	6 dB drop after 9/11 coincided with lower right whale stress hormones in one bay
Natural noise (wind, waves, rain)	Mostly higher frequencies	Whales evolved in this baseline; it varies with weather and season

What people get wrong about whale song

The most common misconception is treating "structured" as synonymous with "language." Humpback song has had its hierarchy documented since 1971, and sperm whale codas showed combinatorial structure in 2024. Neither finding establishes syntax, grammar, or reference in the linguistic sense. Popular coverage has sometimes blurred that line more than the papers do, and the researchers behind the coda study describe a coding system, not a language.

A second error is assuming all whales communicate alike. Blue and fin whale pulses, humpback song, and sperm whale clicks are three acoustically and functionally different systems, shaped by different anatomy, social lives, and habitats. Blue and fin whales appear built for range. Humpbacks appear built for structured display. Sperm whales split their toolkit between an echolocation system for hunting in darkness and a social coding system for group life. Lumping them into "whale song" hides the more interesting story, which is how differently evolution solved different problems.

A third error is thinking whales built the SOFAR channel, or even that they invented the use of it. The channel is a pre-existing feature of the ocean, found by oceanographers with explosives and hydrophones in 1944 and again in the Sea of Japan in 1946. Whales did not create it. What the evidence supports is a good fit between low-frequency calls and long-range propagation, and good documentation of long-range detections, not proof that every call is aimed at the channel.

A fourth error is the idea that sound travels better in water simply because water is quieter. Sound travels faster in water than in air, but the long distances come from the combination of low absorption at low frequencies and the way the ocean's layered structure guides sound. Deeper, colder, higher-pressure water behaves differently from the sunlit surface, as readers curious about the deep ocean can see in our piece on the deepest places in the ocean.

What remains unsettled

Several basic questions are still open, and an honest summary has to say so.

• Function of humpback song. Breeding display, male-male ordering, and spacing are all proposed, and the relative weight of each is not settled.
• Meaning of sperm whale codas. The 2024 study shows structure and context sensitivity but no demonstrated reference.
• The identity of the 52-hertz caller. The species is unknown and the hybrid hypothesis is one of several.
• How much noise matters. Stress hormones and louder calling are documented in specific settings. Population-level consequences are harder to measure and are still being studied.
• How far calls really carry. Confirmed detection distances such as about 350 kilometers for fin whale pulses are real, but they depend on season, depth, and background noise.

The takeaway

Whales sing underwater by exploiting two documented physical facts. Low-frequency sound loses little energy to absorption, and the ocean's sound-speed minimum guides sound sideways for long distances. Fin and blue whales produce deep, loud calls whose range has been confirmed at hundreds of kilometers in some conditions, while humpbacks build elaborate songs that spread culturally across ocean basins, and sperm whales run a separate click system that includes socially used codas with combinatorial structure.

None of that structure has been shown to be language, and readers should treat any claim that scientists have "decoded" whale speech with caution until the underlying papers say so. What is settled is that low-frequency noise from ships has risen, and that in at least one well-studied bay, quieter water coincided with lower stress hormones in right whales.

---

Read next:

• Why Deep Sea Fish Look So Strange
• How Salmon Find Their Birth River
• Why Flamingos Stand On One Leg

FAQ

Q: How far can a whale's call travel underwater? A: It depends on the species, the frequency, the season, and background noise. A NOAA technical memorandum on SOSUS reported a confirmed detection distance of about 350 kilometers for fin whale pulses, and noted that detection distances vary with seasonal ocean conditions. Higher-pitched sounds, such as parts of humpback song, travel shorter distances than 20-hertz fin whale pulses.

Q: What is the SOFAR channel and who discovered it? A: The SOFAR (Sound Fixing and Ranging) channel is a horizontal layer of ocean, at roughly 1,000 meters deep at mid-latitudes, where the speed of sound is at its minimum and sound waves bend back toward the layer. Maurice Ewing and J. Lamar Worzel tested it in spring 1944 with explosive charges, and Leonid Brekhovskikh independently found it in the Sea of Japan in 1946.

Q: Why do blue and fin whale calls use such low frequencies? A: Seawater absorbs low-pitched sound far less than high-pitched sound, so low calls keep their energy over longer distances. Blue whale calls are typically 15 to 40 hertz, and fin whale pulses sweep from about 40 hertz down to 13 hertz, according to the Discovery of Sound in the Sea project. Their larynx anatomy, described in a 2024 Nature study, also suits deep sounds.

Q: Do humpback whale songs actually mean something, or is it just structure? A: Payne and McVay's 1971 paper showed a hierarchy of units, phrases, and themes, and later work showed that new songs spread across populations over years. That strongly suggests a communicative role tied to breeding. Whether individual elements carry specific referential meaning, like words, has not been shown.

Q: Do sperm whales have a language? A: Not as far as anyone has demonstrated. A 2024 Nature Communications study of 8,719 codas from about 60 whales found a combinatorial coding system built from rhythm, tempo, rubato, and ornamentation, with the last two varying by context. The authors describe structure and do not claim word-like meaning or grammar.

Q: Is human-caused ocean noise actually affecting whale communication? A: There is evidence that it does in specific cases. Low-frequency background noise rose by about 3.3 dB per decade from 1950 to 2007 in the data analyzed by Frisk. A 2012 study found lower stress-hormone metabolites in right whales during a quieter period after September 11, 2001, and a 2010 study found right whales calling louder as noise increased. Population-level effects are harder to measure.

Sources

• Discovery of Sound in the Sea, University of Rhode Island: History of the SOFAR channel, the sound speed minimum, blue whale, fin whale, humpback whale, and sperm whale sound pages.
• Parks, S. E., Johnson, M., Nowacek, D., and Tyack, P. L. (2011). Individual right whales call louder in increased environmental noise. Biology Letters 7(1), 33-35. PubMed Central.
• Garland, E. C., et al. (2011). Dynamic horizontal cultural transmission of humpback whale song at the ocean basin scale. Current Biology. University of St Andrews record.
• Rolland, R. M., et al. (2012). Evidence that ship noise increases stress in right whales. Proceedings of the Royal Society B 279, 2363-2368. ScienceDaily summary.
• Payne, R. S., and McVay, S. (1971). Songs of humpback whales. Science 173, 585-597.
• Elemans, C. P. H., et al. (2024). Evolutionary novelties underlie sound production in baleen whales. Nature 627, 123-129.
• Sharma, P., et al. (2024). Contextual and combinatorial structure in sperm whale vocalisations. Nature Communications 15.
• Møhl, B., et al. (2003). The monopulsed nature of sperm whale clicks. Journal of the Acoustical Society of America 114, 1143-1154.
• Frisk, G. V. (2012). Noiseonomics: the relationship between ambient noise levels in the sea and global economic trends. Scientific Reports 2, 437.
• Au, W. W. L., et al. (2006). Acoustic properties of humpback whale songs. Journal of the Acoustical Society of America.