The Guns of Barisal and Anomalous Sound Propagation


  1. Introduction
  2. Anomalous Sound Propagation
  3. References


Barisal is a small port on the Tetulia mouth of the River Ganges, on the northern shore of the Bay of Bengal, about 130 mi. east of Calcutta, and about the same distance south of Dacca, in Bangladesh. From the 1870's there were reports of hearing sounds resembling cannon fire, coming from the south or southeast. They occurred mainly from February to October, very seldom from November to January, and the source was unknown. The sounds were often multiple, in groups of two or three.

Similar sounds, called in the Netherlands and Belgium mistpoeffers, or in Italy brontidi, or marinas, or in the Phillipines retumbos, or elsewhere fog guns, were heard from time to time. They were reported from Passamaquoddy Bay in southwestern New Brunswick, in Belgium and Scotland, at Cedar Keys, Florida, Lough Neagh in Ireland, and in Western Australia, and in Victoria State. They were reported on an Adriatic island in 1824, at Franklinville, NY in 1896, and in northern Georgia. Lewis and Clark heard them on 4 July 1805. Though often heard at coastal locations and beside lakes, they were also heard away from bodies of water, and were described as booms, like thunder, or the discharge of cannon.

You may read in encyclopedias that the Guns of Barisal are supposed to be caused by earth movments too feeble to be felt. Earthquakes can make noises, but not when no movements are felt, and the noises are not like guns. The Guns, indeed, show no correlation with local earthquake activity. It is quite improbable that earth movements are the cause. When a search was made for possible sources, nothing was ever found. Sometimes it was thought that they might come from blasting a few miles away, but when this was investigated, nothing was found. The Barisal guns themselves came from the vastness of the Bay of Bengal. Thunderstorms were not reported at the time of the observations, but thunderstorms cannot usually be heard more than a few kilometers away. That the Guns were often heard near water means only that such locations were quiet enough to make unusual sounds noticeable. Not all the Guns were heard near water, anyway.

The Wikipedia article on Barisal has only one sentence dealing with the Guns, where it is suggested that they are connected with the tidal bore in the Meghna estuary. Tidal bores are regular phenomena, and if they somehow caused the Guns here, the Guns would be heard regularly, which they are not.

Seneca Lake in New York was the site of noises that were thought to come from the release of bubbles of natural gas from deep in the lake, observed as early as 1903. The bubbles were not ignited, of course, and if they were they would only have burned quietly. They were thought to come from the Oriskany formation, a sandstone that had been drilled for natural gas in the vicinity. These "guns" were still around in 1934, when the gas wells had been exhausted. There were also the "Moodus noises" of the lower Connecticut valley, known to the Indians and heard from 1709 to 1729, 1852 and 1885, and were heard again in 1897. They were described as a thunder and a roar. Unexplained explosions were heard near Deerfield, NH in 1846. All these noises are mysterious, but somewhat unlike the Barisal guns and similar booms.

An apparently different phenomenon was first noticed in 1666, during an engagement of the English and Dutch fleets in the Channel on 1 June. The sounds of the guns were heard in London, but not on the South Downs, Deal or Dover, all points between the battle and London. This was recorded in the diaries of John Evelyn and Samuel Pepys as a remarkable occurrence, that the winds brought to them the noise of the guns, but not to the people in between. In 1904 a large explosion at Förde in western Germany was accompanied by similar results. There was a small area in which the sound arrived directly, then a zone of silence, and finally the sound was heard again, at a distance of 100 to 200 km from its source. G. von der Borne explained this as the return of an acoustic wave reflected from a hydrogen or helium zone in the upper atmosphere, and he was supported by van Everdingen, who observed large explosions in the early stages of the World War. It was remarked that the noise of guns firing in Flanders could often be heard in the east of England, especially in the summer, but rarely in winter. When explosions were later made for the explicit purpose of investigating this effect, it was found that the apparent time of propagation showed that the waves reaching the zone of abnormal audibility had travelled on a path considerably longer than the direct one. It became clear that the sound was being reflected in the upper atmosphere.

A munitions factory at Silvertown, an industrial area in the East London Docks on the north bank of the Thames, exploded in January 1917. It could be heard in an elliptical area with a major axis of about 150 km in the NW-SE direction, and a minor axis of about 50 km, centred on the site of the explosion. This is the zone of normal audibility. Beyond, in the zone of silence, the explosion was not heard. About 120 km to the northeast, in Lincolnshire and Norfolk, the explosion was again heard, in a similar but somewhat larger elliptical area, the zone of anomalous audibility. In this region, the report was multiple, consisting of two or more bangs at brief intervals, a consequence of multiple paths.

Reports of strange sounds seem to be very much rarer now, and it is possible that they are being ignored in these noisy times. While rocking gently at anchor on a dark, still night on the Bay of Bengal, such booms will attract attention, but in a noisy city they would never be regarded. Since there are so many sources of bangs and booms in modern times, an odd bang from a great distance would never be recognized. Even in the past reports, the unusual sounds were normally thought to arise in the vicinity. I think it possible that the Barisal guns, and similar phenomena, are the sounds of thunder carried by anomalous propagation over several hundred kilometers to the points where they are heard. In many cases, the sound is described as much like thunder, and to come from the direction of the horizon, with no apparent source, and often in multiple, all of which is characteristic of noises heard by anomalous propagation.

Anomalous Sound Propagation

Very little was known about the upper atmosphere in the years immediately following the World War. The lapse rate in the troposphere was well known from balloon and aircraft ascents. The stratosphere had been discovered by Teisserenc de Bort in 1899, and the tropopause located at an altitude of 10 or 12 km. It was thought that each atmospheric constituent was then distributed independently according to its molecular weight, so that eventually, above about 100 km, the lighter gases, hydrogen and helium, would predominate. This was the basis for von der Borne's explanation of the return of sound to the ground. This explanation was attacked on the ground that the very low atmospheric pressure at high levels would remove most of the energy from the sound wave. Erwin Schrödinger showed that this was not the case, since the amplitude of motion would increase to maintain the energy of the wave, although there would be absorption, as predicted by Rayleigh. However, the altitude required was much too high to transmit sound without excessive attenuation, and so there was still a difficulty.

In 1922, Lindemann and Dobson published results from meteor observations that suggested a temperature of 300K at a height of 60 km. If the temperature was this high at such a reasonable altitude for sound propagation, then a good explanation of anomalous propagation was at hand, avoiding the difficulties with hydrogen and helium. It was not yet known that the photochemical production of ozone created a temperature maximum in the upper atmosphere that was comparable to ground-level temperatures. If θ is the angle that a sound ray makes with the horizontal, and c(z) is the speed of sound at altitude z, then the quantity cos θ/c(z) = constant, from the usual law of refraction. Then, if θ is the inclination of the ray at z = 0, the ray will be reflected at an altitude z such that cos θ/c(0) = 1/c(z), or c(z) = c(0) cos θ. For small angles θ, c(z) does not have to be much larger than c(0).

An additional effect is also important. As far as returning a wave to the ground, the velocity of the wind in the direction of propagation can be added to the velocity relative to the air. Since the temperature maximum is indeed about equal to the ground temperature, a wind in the direction of propagation will encourage return to the ground, while a contrary wind will oppose it. It was, indeed, found that in the winter, when the stratospheric winds were westerly in Europe, there was a zone of anomalous audibility to the east, but none to the west. In the summer, when the winds were reversed, the zone of audibility was moved to the west. On occasion, a second zone of silence and a second zone of anomalous audibility were observed, the sound making a double skip. Surface winds will have no strong effect, except in the launching of the wave.

At the time, there was no way of investigating the upper atmosphere directly. Balloons could only ascend to the stratosphere. Therefore, anomalous propagation of sound became of great interest as a means of probing the upper atmosphere. In 1923, a series of large explosions was set off from La Courtine in central France. These were the first experiments in which accurate time measurements were possible. By analyzing the results, Whipple found that the top of the stratosphere was at 32 km, and that the temperature increased to the value at the ground, 290K, at 46 km. He assumed a stratospheric temperature of 210K.

The present-day standard atmosphere has a ground temperature of 288K, stratospheric temperature 217K at 12 km, with an increase beginning at 25 km, and a maximum temperature of 283K at 50 km. It is clear that Whipple's results are very close to the truth. They were the very first good measurements of temperatures in the upper atmosphere. Whipple carried out further experiments with gunfire. A gun at Shoeburyness could be heard at Grantham, 185 km away to the north. Guns were fired from three points just east of London, and sounds were received at Birmingham, Bristol, Cardiff, Nottingham, Exeter and North Walsham, using hot-wire microphones. An array of three microphones, with accurate time measurements, allowed the determination of the angle of incidence. Experiments were carried out in December, 1932 with four explosions at Oldebroek in Holland, showing the expected zone of audibility to the east, and none to the west.

At 50 km, a typical temperature is 283K, pressure 0.66 mmHg or 0.88 mb, density 1.1 g/m3, molecular weight 28.966 (same as at ground level), number density 2.25 x 1022 m-3, collision frequency 6.06 x 106 s-1, mean free path 75 μm, and speed of sound 337 m/s. Ionization is negligible, since 50 km is well below the D level of the ionosphere. Sound propagates quite successfully under these conditions. Above 50 km, the temperature again decreases, goes through a minimum, and then increases steadily. Ground level temperature is attained again at about 110 km altitude, but here the mean free path is only about 20 cm, so severe absorption except at very low frequencies is to be expected. The molecular weight is about 28.82, already beginning to decrease due to photochemical dissociation. There is also ionization at these levels, which could affect sound propagation. It is conceivable that low-frequency sound could return from such heights, but no observation has been reported.

The absorption of sound is very small, especially in dry air. The Rayleigh-Kirchhoff "classical" absorption due to viscosity is proportional to the square of the frequency, and depends on the ratio of the wavelength to the mean free path of the air molecules. Water vapor introduces a considerable added absorption. However, the upper atmosphere is remarkably dry, so this should have an effect only near the ground. The scattering of sound by turbulence also causes a decrease in intensity, as does the usual spreading. The principal effect of what absorption there is will be to attenuate the higher frequencies, converting a sharp crack into a boom. The initial sound wave from an explosion is usually a supersonic shock wave for a distance, and will certainly be characteristically modified during long-distance propagation.

It was discovered after World War II that at times there was an acoustic velocity minimum in the neighborhood of the tropopause. This meant that sound created at this level would be reflected both above and below, and would be channeled or ducted near this minimum, as sound is in the sea. Not only is the sound absorption very low in the atmosphere, but the spreading in a duct is only 1/r, instead of the usual 1/r2

The hot-wire microphone was often used as an acoustic detector, especially for infrasound. It was invented around the time of the First World War for use in acoustic artillery ranging, and an improved model was developed by Tucker and Paris [Phil. Trans. Roy. Soc. A221, 390 (1921)] and is described in Wood, pp. 303-305. It consists of a grid of fine platinum wire suspended in a hole in a mica disc closing the mouth of a Helmholtz resonator. A current is passed through the wire to heat it, while its resistance is measured with a Wheatstone bridge. The periodic air currents established when a signal is received cool the wire (at twice the frequency of vibration) and change its resistance. It is very useful for infrasonics, where impedance matching is more important than speed. The hot-wire microphone has no diaphragm to add inertia. It is useful for frequencies as high as 512 Hz, and is quite sensitive. Hot-wire microphones were often used in long-distance sound investigations.


W. R. Corliss, Handbook of Unusual Natural Phenomena (Glen Arm, MD: The Sourcebook Project, 1977). pp. 368-385.

F. J. W. Whipple, Propagation of Sound in the Atmosphere, Quart. Jour. Roy. Met. Soc., 61, 285 (1935). Also Nature, 118, 309 (1926).

E. G. Richardson, ed., Technical Aspects of Sound, vol. II (Amsterdam: Elsevier, 1957). pp. 9-14.

E. Schrödinger, Zur Akustik der Atmosphäre, Physikalische Zeits. 18, 445-453 (1917).

A. Wood, Acoustics (New York: Dover, 1960). pp. 169-177.

G. von der Borne, Physikalische Zeits. 11, 483 (1910). Explosion at Förde, Westfalen in 1904.

E. van Everdingen, Proc. Roy. Acad. Sci. Amsterdam, 18, 923 (1915).

Return to Waves Index

Composed by J. B. Calvert
Created 25 February 2003
Last revised 10 December 2008