The Singing Sands: Dunes That Boom a Musical Note for Kilometres
Slide down the steep face of Sand Mountain in Nevada on a dry summer afternoon and the whole hillside answers back. For a few seconds the slope gives out a deep, organ-like note that seems to rise from the ground and hang in the air, low enough to buzz in the ribs. Press a palm flat against the sand and the sound is felt as much as heard, a slow throb travelling up through the grains into the hand. Hikers reach for the same comparisons every time: a drone, a moan, a distant cello, a low-flying propeller aircraft passing somewhere below the horizon. The dune is singing, and it has been unsettling travellers for at least seven centuries.
Somewhere around 1275, crossing the Badain Jaran dunes on the edge of the Gobi, Marco Polo described a desert that seemed inhabited. The air, he wrote, filled with the sound of instruments, drums and the clash of unseen armies. Travellers of his day blamed spirits of the sand, said to call men away from the caravan by name. Polo was not the only chronicler to set the sound down. Medieval Arab writers, Chinese chroniclers and desert nomads across three continents left the same testimony in their own words.
Six centuries later Charles Darwin, ashore in Chile during the voyage of the Beagle, recorded the same astonishment. His informants described a bellowing hill they called El Bramador, the roarer, which was said to sound whenever the sand was disturbed. Like Polo, Darwin was reporting something the local people treated as an ordinary feature of their landscape rather than a marvel.
The effect is documented today at roughly 35 desert sites. Among them are the Kelso Dunes and the Eureka Dunes of California, Sand Mountain in Nevada, the Badain Jaran megadunes of China and the coastal dunes of Oman. Every one of these places shares the same basic recipe: a large dune with a steep lee face, dry weather, and loose surface sand.
Modern recordings have turned the traveller's tale into a set of numbers. A booming dune produces a sustained drone that can hold for minutes and carry for kilometres. It is built around one dominant note, most often between 70 and 105 hertz, with fainter harmonics stacked above it. Close to the slope the sound can pass 100 decibels, loud enough to drown a conversation, which is why early witnesses reached for armies and instruments to describe it. The instruments confirm the impression: this is a genuine musical tone, produced by loose grains and nothing else. There is no throat here, no string, no reed and no pipe. Beach sand is a separate and lesser phenomenon, the brief high squeak underfoot that some coasts produce. The deep boom of a desert dune is rarer, louder and far stranger.
The sound begins when sand avalanches down the steep lee face of a large dune, loose grains shearing over a firmer bed beneath. For a long time the hardest question was not the volume but the coherence. Millions of grains tumble independently, yet together they produce a single clear note rather than a formless hiss; a landslide of gravel roars, it does not hum. Measurements point to synchronisation. The grains fall into step and vibrate in unison, so that a whole patch of the surface pulses together like the diaphragm of an enormous loudspeaker, pushing the air in rhythm. On that much most researchers now agree.
The laboratory work that followed is where the record becomes precise. The French physicists Stephane Douady and Simon Dagois-Bohy recorded two very different dunes: one near Tarfaya in Morocco that hummed a clean tone near 105 hertz, and another near Al-Askharah in Oman that only growled across a smear of frequencies. They carried the Omani sand back to their laboratory in Paris and ran it down a controlled chute. When they sieved out a narrow band of grains between roughly 200 and 250 microns, the noisy sand sang a clear tone. Uniform grains gave a pure pitch; mixed sizes gave a muddle. They reported that the frequency tracks the shear rate of the flowing layer and scales with grain diameter, so that coarser sand sings lower.
At Caltech, Nathalie Vriend and Melany Hunt spent seasons measuring the booming dunes of California with geophones buried in the slope. Their field data showed that the booming frequency did not track particle diameter. They proposed instead that a few tens of centimetres of dry, loose surface sand, lying over firmer damp sand below, act as a natural waveguide, a resonating channel whose depth and internal wave speeds fix the note much as the length of an organ pipe fixes its pitch. Below a cutoff frequency, they said, no sound can propagate at all, which would explain why a dune has to reach a certain size before it booms. The model also accounted for a seasonal drift, since the moist layer rises and falls through the year and the channel changes depth with it. Their 2007 paper was titled 'Solving the mystery of booming sand dunes'.
Within months a third physicist, Bruno Andreotti of Paris, published a formal Comment in the same journal disputing almost every step. His own field recordings, made by walking avalanches down dunes in Morocco, led him to argue that the booming frequency is set by the shear rate at the flowing surface itself. Collisions between grains, he wrote, excite elastic waves that ripple along the top of the dune and then feed back to synchronise the very collisions that made them, a self-reinforcing loop he called wave-particle mode locking.
The record gained one more hard fact in 2012. Dagois-Bohy, Douady and their colleagues poured singing sand down a chute onto a hard laboratory plate and made it boom with no dune, no buried moist layer and no waveguide beneath it at all. Their paper was titled 'Singing-sand avalanches without dunes'.
Alongside the physics runs a set of well-attested conditions. The song demands well-sorted, rounded, silica-polished grains, and some researchers point to a thin varnish of iron and silica gel on the surface of each grain. Above all it demands dryness. A little moisture makes the grains cling and the note dies. The same dune that booms in dry heat falls silent through a damp winter and finds its voice again when the sand dries out. Wet a singing sample in the laboratory and it goes mute; dry it and the tone returns. Samples carried home from a booming dune have been known to lose their voice for months and then recover it.
Conclusions and Open Questions
Three explanations are on the table, and each fits part of the evidence while stumbling over the rest.
The first, proposed by Douady and Dagois-Bohy, holds that the note is written into the grains themselves, set by grain size and the shear rate of the flowing layer. Its strength is that it is portable: pour the right sand anywhere and it should sing, dune or no dune, which is exactly what the Paris chute experiments showed. Its weakness is that out on real dunes the careful Caltech measurements refused to line up with grain diameter at all.
The second, proposed by Vriend and Hunt, holds that the pitch is fixed by a resonating near-surface channel of dry sand over damp. It neatly explains the seasonal drift of the note and why a dune must reach a certain size before it can boom. Its weakness is the 2012 benchtop result: an avalanche on a hard plate boomed with no dune and no damp layer under it. Field reports that the note changes little no matter how the avalanche is triggered also sit awkwardly with a pure resonance account.
The third, proposed by Andreotti, holds that the singing surface tunes itself from the dynamics of the flow rather than from any structure underground. It splits the difference and matches much of the field data, but it is the hardest of the three to test cleanly, and it has not won the other camps over.
Some argue that all three are holding parts of the same animal: the synchronised avalanche makes the sound, and near-surface structure, where it exists, colours it. That would be a tidy settlement, but nobody has cleanly stitched the benchtop experiment to the desert, and that seam is still open.
Beneath the argument sits an older question none of the camps has closed. Why does one dune sing while its neighbour, raised by the same wind out of the same quarry of grains, stays mute? Dryness looks decisive, yet no published study has mapped the microclimates of the silent dunes against the singing ones. One theory holds that the real rarity lies less in the sand than in the narrow band of weather above it, in which case a mute hill might be carrying perfectly capable sand and sit one long drought away from booming. That idea is testable and, so far as the published record shows, untested.
The open questions are easy to list and hard to answer. Are a coastal dune in Oman and a desert megadune in China even the same instrument, or do they sing for genuinely different reasons that researchers keep forcing under one law? Why is the note so insensitive to how an avalanche is started? What part, if any, does the iron and silica gel coating actually play? And why can a sample lose its voice for months in a jar and then recover it?
Legend and laboratory have met in the middle without quite shaking hands. The desert is not haunted, and the marvel that unnerved Marco Polo comes down to grains, shear and resonance. But it is a peculiar kind of solved mystery, one in which researchers can summon the sound on demand from a bucket of sifted sand and still cannot agree on what, exactly, they are hearing.