The latest explanation of the mysterious 'song of the dunes' — an eerie booming drone emitted by some sand dunes — is stoking the controversy fuelled by rival theories.

Sliding scale: the sound of dune avalanches could depend on their depth. Credit: M. FALZONE/LIFESIZE/GETTY

Research on this striking natural phenomenon became something of a battleground after two groups, previously collaborators, put forward two opposing theories. Now, a team at the California Institute of Technology (Caltech) in Pasadena, led by mechanical engineer Melany Hunt, says that they are both wrong1.

“There are strong feelings in this field,” says physicist Michael Bretz at the University of Michigan in Ann Arbor. “It'll take a while longer to get it sorted out. But the explanations keep getting better.”

The noise of the 'singing' dunes can be very loud, audible for up to ten kilometres. “It's really magnificent,” says physicist Stéphane Douady at the Ecole Normale Supérieure in Paris, who proposed one of the competing theories. Marco Polo described it on his journeys through the Gobi desert in the thirteenth century, attributing the sound to evil desert spirits. There is a sand-covered hill in northwestern China called Mingsha Shan, which means singing sand mountain.

The effect is clearly related to avalanches of sand, and can be triggered by people sliding down the slopes. One of the first attempts at a scientific explanation came from Ralph Bagnold, an army engineer who fell in love with the North African deserts during the Second World War. He suggested that the noise was caused by sand grains colliding, and that the frequency of the sound was determined by the average time between collisions. This implies that the frequency depends on the size of the individual grains, increasing as the grains get smaller.

Douady and his students Bruno Andreotti and Pascal Hersen focused on these collisions during a research trip in Morocco in 2001. Douady decided that for the moving grains to generate a single sound frequency, their motions must become synchronized. This synchronization, he argued, comes from standing waves set up in the sliding layer. The noise is loud because the surface of the dune acts like a giant loudspeaker membrane.

But Andreotti came up with a slightly different explanation. The synchronization of grain motions, he said, comes from waves in the sand below the sliding layer, which then act back on the moving grains, 'locking' their movements together.

It might seem like a small distinction, but Douady and Andreotti could not resolve their differences, and in the end they published separate papers offering their explanations2,3. Andreotti now works at a separate lab in Paris.

But both explanations have serious problems, according to Hunt. For one thing, the measurements made by her team on booming dunes in Nevada and California seem to show that the booming frequency doesn't depend on the grain size at all. What's more, the previous theories imply that all dunes should be able to 'sing' when an avalanche takes place. But in fact not all dunes sing — that's why Mingsha Shan got its name, for example.

You can take a cupful of this sand and excite it with your finger.

Andreotti proposed that 'silent' dunes aren't dry enough, or have grains of the wrong shape. But Hunt and her colleagues think that the answer lies deeper than this. They say that dunes have to be covered in distinct layers of sand to create a boom. Their measurements of vibrations in the sand — made with an array of 'geophones' like those used to monitor seismic waves in earthquake studies — showed that the speed of these seismic waves increases in abrupt steps as the sand gets deeper. At 1.5 metres below the surface of one dune, for instance, the speed of the seismic waves increased from 180 metres per second to 310 metres per second.

The Caltech researchers think that this layered structure enables the surface to act as a kind of waveguide for acoustic energy, rather like the way an optical fibre channels light. So although they agree that the boom is transmitted to the air by a loudspeaker effect of the dune surface, they think that the frequency is set by the width of the waveguide layer of sand. Dunes that do not have this layered structure, such as smaller dunes, do not sing at all.

This is unlikely to be the last word on the matter, however. For one thing, the strange properties of the sand in 'booming dunes' are not just found in large structures. “You can take a cupful of this sand and excite it with your finger,” says Peter Haff, a geologist at Duke University in Durham, North Carolina. “You can feel it vibrating, like running your finger over a washboard. But you can take sand from other parts of the dune, and there's nothing you can do to make it vibrate.” Haff says that, although these theories may offer part of the answer, “there must be something else going on at a small scale”.

Douady agrees. “The problem for the Caltech theory is that we can recreate these sounds in the lab,” he says. He thinks that the sand layering might play a role in modifying the sound, but that it is “just a decoration” to the basic mechanism of booming.