There is a word in English for a sound that should not exist. Brontide — from the Greek brontē, meaning thunder — is a low, muffled rumble like distant thunder, heard in mountainous regions, often with no storm anywhere in sight. It is not the thunder of a cloud. It is, by the accounts of those who have studied it most closely, the thunder of the ground. The earth, under strain, makes a sound, and in certain places, at certain times, people can hear it.
The word is rare, and the phenomenon is rarer still in the scientific literature, but it is old. Brontides have been reported for centuries, in Italy above all, where they carry a dozen local names — brontidi, marina, balza, lagoni, bomba, rombo, boato, bonnito, mugghio, baturlio, tromba, rufa — each one a different community’s attempt to describe the same impossible noise. One of the earliest documented reports came from an island in the Adriatic Sea in 1824. They have been heard since on calm summer days in the Bay of Fundy in Canada, on Lough Neagh in Northern Ireland, in Scotland, in Passamaquoddy Bay in New Brunswick, at Cedar Keys in Florida, at Franklinville in New York in 1896, and in northern Georgia. Wherever they are heard, the description is the same: distant but inordinately loud thunder, while no clouds are in the sky large enough to generate lightning.
A sound with two voices
What makes a brontide different from the other members of the skyquake family — the Barisal guns of Bengal, the Seneca guns of New York, the mistpoeffers of the Low Countries — is that it has been studied where it lives, in the seismic country of north-western Italy, by people who listened to it for years and learned to tell its voices apart.
At a small observation centre for seismic precursors near Pinerolo and Barge, in Piedmont at the foot of the Alps near the French border, researchers documented the phenomenon in detail. Their findings, published in the Open Journal of Earthquake Research in 2018, describe a sound with two distinct characters, depending on where it is produced. When a brontide occurs in the air, it is very short, with no echo, and can be confused with a gunshot; analysed closely, its timbre resembles the snap of a percussive wooden instrument. When it is produced in the subsoil, it takes on a different character altogether: sometimes it sounds like the closing of large underground doors, very brief; sometimes it resembles a drum roll, with a local tremor, lasting longer; sometimes it comes from a subterranean landslide, a muffled noise with a tremor that can be felt underfoot, like the sensation of a thunderbolt striking less than a kilometre away.
The researchers also recorded details that no one who had only read about brontides would have expected. When the natural explosions take place in the air, sulphureous smells are sometimes perceived. Dogs, they found, give the alarm first — usually before the human observers notice anything, and in one case about forty minutes before a quake. And the sounds are easily confused with something far more ordinary: quarry explosions in the Luserna stone quarries at Rorà, less than ten kilometres from the study centre, which produce a similar bang and must be checked against the record before any brontide is counted.
A warning before the quake
The most provocative claim in the Piedmont research is that brontides are seismic precursors — that they announce earthquakes, sometimes hours before the ground moves. The evidence for this is specific and dated, and it is worth setting out in full, because it is the strongest case ever made for the idea.
On 2 December 2013, more than fifteen brontides were perceived at the observation centre in a single day, all of them announced, the researchers wrote, by the behaviour of the dogs. On the same day, instruments recorded a high emission of radon gas — radon-222, a radioactive gas that seeps from the ground when rock is under strain — and anomalies in the intensity of the geomagnetic field. The next day, 3 December 2013, a local earthquake of magnitude 2.6 struck at a depth of ten kilometres, with its epicentre less than ten kilometres away, on the Italian-French border. On that occasion, the researchers concluded, both the high radon emission and the brontides were seismic precursors, with a temporal and spatial forecast: the sounds and the gas had pointed, a day ahead, to a quake in a particular place.
A second case followed the same pattern. In the monitoring record, a rising peak of radon emission preceded the earthquake of Finale Emilia — the destructive 2012 quake in northern Italy — and during that anomalous increase there were numerous brontides. The researchers’ conclusion was careful but clear: brontides are often precursors of earthquakes of high magnitude, though not of all seismic activity, and they are heard especially in seismic regions along sea coasts, over lakes, or near quicksand — places where the ground is soft enough to move and loud enough to be heard doing it.
It is important to say what this does and does not prove. A magnitude 2.6 earthquake is a small event, and a single day of brontides before it is a correlation, not a demonstration that brontides can predict earthquakes in general. The radon connection is suggestive — radon is a recognised, if unreliable, precursor signal — but radon spikes happen without earthquakes too, and earthquakes happen without radon spikes. What the Piedmont record does establish is something narrower and more interesting: that in one well-instrumented location, over years of observation, a particular kind of natural sound repeatedly occurred in the company of measurable strain in the ground, and that the people listening learned to read it.
Why the earth should make a sound at all
The physics behind a brontide, if the seismic explanation is right, is straightforward in outline and difficult in detail. Rock under stress does not fail all at once. It cracks, grain by grain, and each tiny failure releases a pulse of energy — some of it as a tremor in the ground, some of it as a pressure wave in the air. A single crack is inaudible. Thousands of them, in a volume of rock that is being squeezed toward failure, can add up to a sound: a rumble, a drum roll, a boom. Gas helps. When rock fractures, it can release gas trapped in its pores, and a sudden release of gas — from a pocket of methane, or from water flashing to steam — can produce a sharp report, like a gunshot, or a longer roar.
There are competing explanations, and they have to be taken seriously. One is that the noise is very distant thunder, focused anomalously as it travels through the upper atmosphere, so that a storm hundreds of kilometres away sounds as if it were overhead. Another is that the sounds are microseismic — tiny tremors too small to register on a seismograph but large enough to be heard when they happen close to the surface. A third is that they are gas escaping from underground pockets, with no earthquake involved at all. And a fourth, the most mundane, is that they are meteors entering the atmosphere, whose sonic booms can sound exactly like distant thunder on a clear day.
None of these explanations excludes the others. The word brontide, like the word skyquake, may describe a family of sounds with several different causes, and the Italian brontidi — tied as they are to a seismic landscape, to radon, and to the behaviour of dogs — may be only one member of that family. What is striking is how consistent the descriptions are across centuries and continents: a low thunder, a clear sky, a sound that seems to come from everywhere and nowhere.
Listening to the ground
The brontide is the least famous member of the skyquake family, and in some ways the most important. The Barisal guns were a colonial curiosity. The Seneca guns became a piece of folklore. The brontide became, in the hands of a small group of Italian researchers, a scientific instrument — a sound to be recorded, measured, and correlated with radon, magnetism and the behaviour of animals, in the hope that it might one day give warning of a quake.
That hope has not yet been fulfilled. No seismic network in the world uses brontides as a forecasting tool, and no one has shown that the sounds can predict an earthquake with the reliability that a warning system would need. But the Piedmont record stands as a careful, multi-year observation of a real phenomenon, made by people who knew the difference between a quarry blast and a sound from the earth, and who wrote down what they heard, day after day, for years. In a field full of anecdotes, that is rare. In a world where earthquakes still arrive without warning, it is worth remembering that the ground, sometimes, speaks first — and that in the mountains of Italy, there are people who have learned to listen.
What is documented, and what is not
Documented facts: Brontide, from the Greek brontē meaning thunder, is a low, muffled sound like distant thunder heard in mountainous regions, often attributed to seismic activity. In Italy the phenomenon carries many local names, including brontidi, marina, balza, lagoni, bomba, rombo, boato, bonnito, mugghio, baturlio, tromba and rufa. One of the earliest reports came from an Adriatic island in 1824; brontides have also been reported in the Bay of Fundy, Lough Neagh, Scotland, Passamaquoddy Bay, Cedar Keys in Florida, Franklinville in New York in 1896, and northern Georgia. A study published in the Open Journal of Earthquake Research in 2018, based on years of observation at a seismic precursor centre near Pinerolo and Barge in Piedmont, Italy, describes brontides in the air as short, gunshot-like sounds and brontides in the subsoil as drum rolls or underground door-closings with local tremors, sometimes accompanied by sulphureous smells. The study records more than fifteen brontides on 2 December 2013, together with high radon-222 emissions and geomagnetic anomalies, one day before a magnitude 2.6 earthquake on the Italian-French border, and notes numerous brontides during a radon increase before the 2012 Finale Emilia earthquake.
Unverified or disputed claims: That brontides can reliably predict earthquakes — the Piedmont cases are correlations in a small number of events, and no seismic network uses brontides as a forecasting tool. That all brontides have a seismic cause — distant thunder focused through the upper atmosphere, microseismic activity, gas release and meteors are all viable explanations for some reports. Claims that dogs can forecast earthquakes from brontides are anecdotal, based on observations at one centre, and have not been tested independently.