Module 4: Whale Song & Communication

Male humpback whales (Megaptera novaeangliae) on their tropical breeding grounds produce hierarchically structured songs. Payne & McVay (1971) identified four levels of organization:

A singer repeats its song with the themes in the same order, for hours on end. All males in a given population sing the same song at a given time. But the song changes progressively: themes are added, dropped, and modified over the course of a breeding season. Most remarkably, these changes sweep across populations. Noad et al. (2000) documented how a novel song from the Indian Ocean (west Australia) replaced the ancestral east Australian song over two years as it traveled eastward — a clear case of cultural transmission and population-scale fashion cycles in non-human animals.

1.1 Functional Hypotheses

Despite fifty years of study, the function of humpback song remains debated. Leading hypotheses:

• Sexual selection (male display): analog to bird song; a complex learned performance signaling fitness or condition to females.
• Male-male signaling: coordination or competition between males; song stops or changes when two singers meet.
• Sonar/biosonar: acoustic probing of the environment to locate other whales or features (Mercado 2018).

Most likely the song serves multiple functions simultaneously. The cultural revolution phenomenon (sudden adoption of a new song across a population) suggests an element of social learning beyond pure genetic or developmental programming.

The SOFAR channel (Sound Fixing and Ranging), also called the deep sound channel or the Munk channel, is a horizontal waveguide at around 1000 m depth in the temperate ocean. It arises because the sound speed in seawater has a minimum at depth, creating an acoustic “minimum of c” that traps horizontally-propagating rays.

2.1 Sound Speed in Seawater

Sound speed in seawater depends on temperature, salinity, and pressure:

Near the surface in temperate oceans, T is high (~20°C) and z is shallow — both push c higher. At depth, T drops but z rises; T wins first, making c fall to a minimum. Below ~1 km, pressure dominates and c rises again. The sound speed minimum (“SOFAR axis”) sits at ~1000 m in mid-latitudes, shallower in the tropics, essentially at the surface in polar seas where there is no thermocline.

2.2 Acoustic Waveguide

A ray launched near the SOFAR axis and with small grazing angle never reaches the surface or bottom — it is continuously refracted back toward the axis (Snell's law applied in a continuously varying medium). The mode is analogous to light in a graded-index optical fiber. Geometric spreading for a trapped cylindrical mode is only \(r^{-1/2}\) (vs \(r^{-1}\) for spherical spreading), so transmission loss scales as:

At 20 Hz (blue whale call frequency), absorption is tiny:\(\alpha \approx 0.001\,\text{dB/km}\). Combined with cylindrical spreading, a blue whale vocalization at source level 188 dB can reach thousands of kilometers before attenuating below background ocean noise. Before the advent of extensive ship traffic, blue whales may have been able to communicate across entire ocean basins.

2.3 Blue Whale Calls

Blue whales (Balaenoptera musculus) produce some of the most intense biological sounds on Earth: ~188 dB re 1 μPa at 10–40 Hz. Nine regional populations have distinct call types (the “songs” of various oceans), and puzzlingly the peak frequencies of these calls have been shifting downward over decades — roughly 30% drop from 1960 to 2010. Proposed explanations include population recovery (if only the largest males sing, their fundamental frequencies are lower), sexual selection for lower pitch, or cultural drift.

Beyond sonar clicks for prey detection, sperm whales produce social codas— short patterned click sequences (typically 3–20 clicks with distinctive inter-click intervals). Codas are used for social communication at the surface, not for foraging. Rendell & Whitehead (2003) demonstrated that sperm whales in the Pacific partition into acoustic clans: each clan uses a distinct repertoire of codas, and clan boundaries transcend geographic ocean regions. A sperm whale raised in a particular clan learns its coda repertoire and retains it for life. These are among the clearest documented examples of animal culture outside humans.

3.1 Coda Syntax

A coda is notated by its inter-click intervals, e.g. the common Caribbean “1+1+3” coda has one click followed (after a pause) by three quick clicks. Different clans use distinct “identity codas” that are particularly diagnostic of clan membership (Gero et al. 2016). Acoustic analysis suggests that coda vocal learning in sperm whales is comparable to dialect learning in humans or song-type culture in birds.

3.2 Dolphin Signature Whistles Revisited

Introduced in Module 3, the dolphin signature whistle functions as an individual acoustic label. Janik (2013) compiled the experimental evidence:

• Each bottlenose dolphin develops a unique signature whistle by 1–2 months of age
• Signature whistles are used upon separation (mother-calf, pod members)
• Dolphins copy each other's signature whistles to address specific individuals (King & Janik 2013)
• Signature whistles remain stable for decades (demonstrated in wild populations for >10 years)

This is arguably the closest analog to human names in any non-human animal.

Industrial shipping has transformed the acoustic environment of the oceans. Low-frequency (10–500 Hz) ambient noise has risen by ~3 dB per decade since 1950 (Andrew et al. 2002); in shipping lanes the increase exceeds 15 dB over the same period. This is precisely the frequency band used by blue, fin, and bowhead whales.

5.1 Masking

The effective communication range for a whale call declines with ambient noise. If the detection threshold is\(\text{NL} + \text{DT}\), the maximum range is:

A 15 dB increase in ambient noise reduces communication range by up to 10^15/20 = 5.6×. For a blue whale with pre-industrial range ~5000 km, this contraction puts the animals within hailing distance of only ~900 km — still impressive, but a dramatic reduction in the acoustic “world” available to a social animal.

5.2 The Lombard Effect in Whales

North Atlantic right whales and other species exhibit a Lombard effect — raising their call amplitude in proportion to ambient noise, just as humans speak louder in a loud room. Parks et al. (2011) measured right whale up-call source levels increasing by ~0.5 dB for every 1 dB rise in background noise. This physiological compensation is incomplete (Lombard ratio ~0.5), so effective communication range still contracts with noise. It also has metabolic costs: louder calls require more energy.

5.3 The 2020 Pandemic as Natural Experiment

The COVID-19 pandemic reduced marine traffic dramatically in spring 2020. Measurements in the Strait of Georgia (Thomson & Barclay 2020) showed a 1.5 dB drop in underwater noise. Concurrent studies of whale vocalization behavior suggested that whales responded immediately — calling less often at higher amplitude, or calling more often at lower amplitude. The pandemic quieting constitutes perhaps the clearest evidence that chronic noise pollution was constraining cetacean acoustic behavior.

5.4 Seismic Airgun Impacts

Oil and gas exploration uses arrays of compressed-air guns that fire every 10–20 seconds, producing broadband impulsive sounds at source levels of ~260 dB re 1 μPa. Airgun surveys continue for months and propagate detectably across ocean basins. Nowacek et al. (2015) documented cessation of sperm whale foraging, fin whale abandonment of breeding grounds, and blue whale avoidance behavior in response to airgun surveys. Cumulative exposure during seismic surveys is among the highest anthropogenic acoustic impacts a cetacean can experience short of direct military sonar.

Vocal learning — modifying vocalizations based on auditory experience — is rare in mammals. Among mammals it has been documented in humans, cetaceans, pinnipeds, elephants, and bats. Cetaceans are the most accomplished mammalian vocal learners outside our own species: both odontocetes (dolphins, orcas, belugas) and mysticetes (humpbacks) exhibit production learning, in which individuals acquire new sound categories by listening.

6.1 Signature Whistle Development

Bottlenose calves develop a signature whistle over the first 6–12 months of life. They seem to listen to a range of whistles in their social environment and then select a pattern that is distinct from both their mother's and from other close associates (Fripp et al. 2005). This deliberate differentiation ensures the signature is uniquely identifying. Some calves copy elements of their mother's signature; others show more independent innovation.

6.2 Orca Dialect Learning

Orca pods have pod-specific dialects consisting of several stereotyped call types. Juvenile orcas learn these calls from their mothers and close relatives during the first year of life (Ford 1991). Transient and resident orca ecotypes have completely different repertoires despite overlap in range — a pure cultural distinction. Captive orcas have been observed to imitate the calls of companions from different wild populations, confirming vocal production learning directly.

6.3 Interspecific Imitation

Captive belugas (Ridgway et al. 2012) and killer whales (Musser et al. 2014) have been documented imitating human speech sounds — the only non-primate mammals reliably documented doing so. The beluga NOC at the National Marine Mammal Foundation produced recognizable “out!” calls; an orca named Wikie learned to mimic short English words. These abilities suggest a vocal learning capacity comparable to some bird species and well beyond anything documented in non-human primates.

6.4 Neural Substrates

Vocal learning requires a neural circuit linking the forebrain motor cortex to the brainstem motor neurons that drive the sound-production organs. In song birds this circuit is well characterized (HVC → RA). In cetaceans the analogous pathway includes strong direct projections from motor cortex to nucleus ambiguus (which innervates the phonic lips and laryngeal muscles) — a feature shared only with humans among mammals. The neuroanatomical substrate for cetacean vocal learning thus appears homologous to the substrate of human speech.