How Elephants Communicate Through the Ground: The Strange Acoustics of Seismic Signaling

The schoolroom version of animal communication says that mammals signal through visible displays and audible vocalizations, with the audible range running from about 20 hertz to 20 kilohertz. Elephants do something stranger. They produce low-frequency rumbles at 14-24 hertz, mostly below human hearing, that travel through air for several kilometers and through the ground as seismic surface waves for much longer distances. They also detect these signals through specialized mechanoreceptors in their feet. The result is a communication system that operates across landscapes large enough that the participating elephants cannot see, smell, or directly hear each other in any conventional sense.

The discovery rewrote the model of how large mammals coordinate across the savanna. It is also one of the more dramatic cases of a perceptual modality that humans cannot intuitively imagine.

The puzzle that started the investigation

Katy Payne, an acoustic biologist at Cornell, noticed in the mid-1980s that captive elephants at the Washington Park Zoo in Portland seemed to coordinate behavior across enclosures without obvious cues. She felt vibrations in the air around the elephants that she could not hear. Returning with low-frequency-capable recording equipment, she found that the elephants were producing intense calls in the 14-35 Hz range — below the lower limit of human hearing for most adults but well within the range of large-membrane microphones.

The wild observation that followed was that the same calls had been recorded for years on elephant videotapes as a kind of low-frequency rumble that researchers had assumed was equipment noise or wind. Once Payne and her collaborators went looking for the signals deliberately, they were everywhere. Caitlin O'Connell-Rodwell extended the work in the 1990s and 2000s, documenting that elephants also produced ground-coupled seismic waves and were capable of detecting them.

The physics of low-frequency propagation

The choice of frequency band is not arbitrary. Sound attenuation in air is roughly proportional to frequency squared — a 14 Hz signal travels much farther than a 1000 Hz signal at the same source power. Elephant calls at 14-24 Hz can be detected at 4-10 kilometers in air under good conditions, and the source level of an elephant rumble (around 117 dB at 1 meter) puts the detection threshold for other elephants well above ambient noise at those distances.

The seismic channel is different physics. A loud air-coupled rumble couples into the ground through the elephant's feet and produces a Rayleigh surface wave that propagates through the top few meters of soil. Soil is a much better long-distance medium for low frequencies than air. The seismic signal can be detected at 16-32 kilometers depending on substrate, against a quieter background noise floor than air provides. The same call exists in both channels simultaneously.

The detection mechanism

How elephants detect seismic signals is one of the more interesting parts of the story. The candidate mechanisms are Pacinian corpuscles in the foot pads — the same vibration-sensitive mechanoreceptors mammals use for fine touch perception — and bone conduction through the foot bones and skeleton to the inner ear. Both probably contribute, with different frequency sensitivities.

The behavioral demonstrations are striking. O'Connell-Rodwell's experiments at Mushara waterhole in Namibia in the early 2000s showed elephants orienting toward distant seismic sources, pressing their feet to the ground in postures consistent with seismic listening, and changing behavior in response to seismically replayed predator alarms produced kilometers away. The "freeze and listen" posture — feet flat to the ground, ears forward, body still — appears to be specifically a seismic-attention mode.

The 2017 work by Caitlin O'Connell-Rodwell with Tim Rodwell using ground-coupled vibrations at controlled amplitudes demonstrated that elephants could distinguish between seismic recordings of "anwr" calls (a specific call associated with cohesion) from familiar versus unfamiliar elephants. The detection was not just of the seismic signal but of identity information encoded within it.

The vocabulary

Elephant vocalizations are not random rumbles. Joyce Poole's catalog and subsequent work has identified dozens of distinct call types with specific contexts. The "contact rumble" coordinates group movement across distance. The "musth rumble" advertises reproductive status. The "alarm rumble" warns of specific threats. The "greeting rumble" accompanies physical reunion. Some calls have known semantic content that other elephants respond to in predictable ways.

The 2014 work by Joseph Soltis at Disney's Animal Kingdom showed that the calls have signature acoustic features that identify individual callers — the elephant equivalent of voice recognition. A receiving elephant can tell who is calling, not just that an elephant is calling.

The coordination problem

The deeper biological observation is that elephant social structure requires cross-landscape coordination at scales that visual and acoustic-air-only communication cannot support. Matriarchal family groups maintain stable membership over decades. Bond groups (clusters of families) coordinate movement across home ranges of hundreds of square kilometers. Clan-level coordination during migrations involves dozens of family groups moving in approximate synchrony toward distant resources.

The infrasonic and seismic channel makes this coordination feasible. A matriarch's contact rumble can be heard by relatives 10 kilometers away. A musth advertisement can travel 4-10 kilometers in air and farther in the ground. An alarm call propagating through both channels can warn elephants who were not visually present at the inciting event.

The conservation implication

The seismic channel introduces an unexpected conservation concern. Industrial activity — mining, oil exploration, large-vehicle traffic — produces low-frequency vibrations in the same band that elephants use for communication. The 2019 work by Beth Mortimer at Oxford using fiber-optic sensors in the ground showed that the elephant communication band overlaps substantially with anthropogenic seismic noise in regions where mining or development is active. The functional impact on elephant social coordination has not been fully characterized, but the obvious hypothesis is that it degrades the channel they have evolved to use.

This is the seismic analog of the more familiar acoustic-noise-pollution problem for marine mammals. The remediation options are not obvious — quieting heavy industrial activity in elephant range is not a small ask — but the conservation question is one that did not exist as a concept before the communication system was characterized.

The convergent cases

Elephants are not unique in seismic communication. The 2007 work by Peggy Hill at Tulsa documented seismic signaling in golden moles, kangaroo rats, and naked mole rats. The 2010 work on elephant seals demonstrated infrasonic vocalizations. The classic 1976 work by Lewis and Narins on white-lipped frogs showed seismic component to mating calls. The pattern is that any animal large enough to produce strong low-frequency vibrations and sensitive enough to detect them at a distance can use the seismic channel as a communication medium, and many do.

The unifying observation is that the ground is a low-frequency communication medium with very different properties from air, and biology has explored it in lineages from amphibians through rodents through elephants over hundreds of millions of years. Humans, with our pedestrian dynamic-range hearing and our long-distance reliance on the higher frequencies that vision-dominated primates use, simply never noticed.

The deeper observation

The story of elephant seismic communication is a small example of how parochial human perceptual experience is. The signals were always there. They were recorded for decades on equipment that captured them. The behaviors associated with them were observed for centuries by people who lived alongside elephants. What was missing was the conceptual framework that suggested that a "rumble" might be a structured signal at a frequency below human hearing, that the ground might be a communication medium, and that elephants might detect signals through their feet. The framework arrived in the 1980s, the measurements caught up in the 1990s and 2000s, and the picture that emerged is of a communication system as rich and structured as any vertebrate language, operating in a perceptual space that no human had thought to look at. The universe of possible signaling systems is far larger than the small slice that human-style sensory equipment makes obvious.