Many reports about earthquakes have suggested that the escape of gases was a major effect, both before, during and after the quakes.

The modern theory has it that some subterranean forces, of unknown origin, gradually build up strains in the crustal rocks, up to the breaking point. The earthquake is then supposed to denote the moment of fracture of that rock.

Many features of earthquakes seem to have no explanation in this theory.

Why would there be many occasions of multiple large quakes over a period of a few days to months? Would the rock not break in all the locations in which it is already stressed to near breaking point, at the time it is violently shaken? Why would the ground shake sometimes for periods longer than a minute? Why would quakes cause tsunamis, the massive ocean waves? A brief tremor, however fierce, would not have such an effect. Perhaps the modern earthquake research had omitted the consideration of effects due to the sudden movements and the rapid large changes of volume that gases may cause. We shall therefore discuss the huge eruptions that have brought up diamonds, and we might well ask whether there may not be smaller ones much more frequently. Are they the initiating events for earthquakes as well as for volcanic eruptions?

Eye-witness accounts strongly suggest that gas eruptions are the initiating events, but in modern times not much attention is paid to such information, because it is considered too uncertain. Instead the effects that can be measured with accuracy, such as a gradual increase of the strains in rocks and the relation of this to earthquakes, has now become the main subject of research in this field in the US. The overriding importance of this research would lie in finding a method for the prediction of earthquakes, but no such method has so far been found.

One city has been successfully evacuated two hours before a massive earthquake, and thereby probably many thousands of lives were saved. This was the city of Haicheng in China, in February of 1975. That prediction was based almost entirely on gas-related phenomena. (See the description in "Eye-witness Accounts of Several Major Earthquakes," this Web site.)

In the modern geological literature the movement of gases in the crust is rarely considered. Perhaps this is still due to the widespread misconception that no pore-spaces could exist deeper than about 10 km. Even when a most violent volcanic gas explosion has occurred, this is often discussed as due to gases that have come out of solution as the lava is exposed to lower pressure. Why would one not consider that a massive bubble of gas had entered the lava channel and raced to the top?

Diamonds, a very pure form of carbon, tell us quite a detailed story about the physical and chemical conditions in the earth, below about 150 km, and this information bears on the origin of petroleum. Chemical theory and the experience in making artificial diamonds, show that high pressures of the order of 45 kilobars are needed to produce this dense crystal. Such pressures are found in the Earth only at a depth of 150 kilometers or more, and it is somewhere at such depths that natural diamonds must have been formed. The temperature there exceeds 1000°C.

The geologic situation in which diamonds occur shows that unusual gigantic eruptions were involved. Although many diamonds are found dispersed in river gravels, the only concentrated deposits are in the rare "pseudovolcanic" structures called "kimberlite pipes." These are deep, vertical shafts, usually filled with a mixture of rock types, including the diamond-bearing rock called kimberlite (Figure 1). Most of the known kimberlite pipes are in South Africa and Siberia, but there are some in North America, in Australia, and probably also in Brazil, where they may be well hidden under younger sediments.

Kimberlite Pipe
© K.G. Cox, 1978Model of a Kimberlite pipe.
The model is based on several South African pipes which have been exposed at various levels by erosion, not only on the one near the town of Kimberley, which gave this formation its name. The model was devised by J.B. Hawthorne of DeBeers Consolidated Mines, Ltd.

"The diamond pipes serve as a window that gives us a look into the Earth. There is probably no other group of rocks that originated from even remotely as great a depth as have these" (Kennedy and Nordly, 1968).

The pressure and temperature at a depth of 150 or 200 kilometers are in the right range for carbon to crystallize as diamond. But how did the carbon become concentrated? We cannot reasonably suppose that concentrations of large pieces of pure carbon were formed in the outer mantle by the diffusion through the rock of the dispersed individual carbon atoms.

Minerals of high purity are usually formed in the Earth by a process that involves the flow of some liquid through cracks and pore spaces in the rock. In some particular circumstance of pressure, temperature or chemical surroundings, a component of such a fluid precipitates, and thus builds up a concentrated deposit in the pore spaces. We then have to suppose that a fluid containing carbon, percolated through rock spaces and precipitated concentrated carbon. Veins of diamond would then be built up in these pore spaces, and a later eruption might bring fragments to the surface. What are the fluids that could be responsible for precipitating and concentrating the carbon?

The mere existence of the diamonds at these depths proves that unoxidized carbon exists there. The two types of fluids that one may consider for the concentration process would be carbon dioxide and methane, the latter possibly associated with heavier hydrocarbon molecules also. Tiny pore spaces in diamonds have been analyzed and found to contain small amounts of highly compressed gases, among which the carbon-containing ones were both carbon dioxide and methane (Melton and Giardini, 1974). It is clear, therefore, that not only unoxidized carbon, namely the diamond itself, but also methane, can exist down there. Thermodynamic calculations have shown that both these gases are stable in the upper mantle at diamond forming depth, and either could be responsible. The following indications would seem to favor methane. Methane is generally much more abundant in the crust than CO2, and appears to be streaming up from deeper levels. Secondly, of the gases contained in diamonds, nitrogen is by far the most abundant. One has to judge that nitrogen had something to do with the deposition of the diamond.

It follows that pore spaces in which fluids can flow exist at these depths, and that mineralization processes, leading to great concentrations of certain substances, can be active there, just as they are at shallower levels. Fluid pressures equaling the rock pressures seem to be widespread, at least in the crust and outer mantle, and this is a matter of great significance, both for the chemical processes and for the methods of ascent of fluids to the surface.

The existence of the kimberlite pipes shows that high concentrations of gas can build up, and have been building up, and these concentrations can explode a hole through 150 kilometers of overlying dense rock. Quite large bubbles of high-pressure gas must have been assembled to do this, and only an inhomogeneous mantle containing volatile-rich materials could be responsible.

A gas eruption, rather than a volcanic transport to the surface, is required to maintain the diamonds. The stable form of carbon at low pressures is graphite, but if diamonds are cooled sufficiently rapidly as they are brought to lower pressures, they are maintained as unstable but super-cooled crystals. At surface temperatures, they are then effectively stable. We see that the evidence from the diamonds is very simple and clear. Unoxidized carbon can and does exist in the outer mantle. It can be brought up without becoming oxidized; it is associated with a variety of hydrocarbon molecules, both within inclusions in diamond and also in other materials brought up in the eruptions. Volatile-rich regions exist in the mantle, so that high pressure gas bubbles become assembled there that can force their way violently through all the overlying rocks. This clearly shows that the Earth has an unmixed, inhomogeneous mantle, and that there is a high concentration of carbonaceous material in many areas of the globe.

I am presenting here a selection of eye-witness accounts of major quakes, showing that gases, and in particular combustible gases are frequently in evidence.

The great earthquake series in New Madrid (Mid Mississippi) in 1811 - 1812

The report by contains the following items:
On the 16th day of December, 1811, at two o'clock in the morning, the inhabitants of New Madrid were aroused from their slumbers by a deep rumbling noise like many thunders in the distance, accompanied with a violent vibratory or oscillating movement of the earth from the southwest to the northeast, so violent at times that men, women, and children caught hold of the nearest objects to prevent falling to the ground.

It was dangerous to stay in their dwellings, for fear they might fall and bury them in their ruins; it was dangerous to be out in the open air, for large trees would be breaking off their tops by the violence of the shocks, and continually falling to the earth, or the earth itself opening in dark, yawning chasms, or fissures, and belching forth muddy water, large lumps of blue clay, coal, and sand, and when the violence of the shocks were over, moaned and slept, again gathering power for a more violent commotion.

On this day twenty-eight distinct shocks were counted, all coming from the southwest and passing to the northeast, while the fissures would run in an opposite direction, or from the northwest to the southeast.

On a small river called the Pemiseo at that time stood a mill owned by a Mr. Riddle. This river ran a southeast course, and probably was either a tributary of the St. Francis or lost itself in those swamps. This river blew up for a distance of nearly fifty miles, the bed entirely destroyed, the mill swallowed up in the ruins, and an orchard of ten acres of bearing apple trees, also belonging to Mr. Riddle, nearly ruined. The earth, in these explosions, would open in fissures from forty to eighty rods in length and from three to five feet in width; their depth none knew, as no one had strength of nerve sufficient to fathom them, and the sand and earth would slide in or water run in, and soon partially fill them up.

Large forest trees which stood in the track of these chasms would be split from root to branch, the courses of streams changed, the bottoms of lakes be pushed up from beneath and form dry land, dry land blow up, settle down, and form lakes of dark, muddy water.

One family, in their efforts to reach the highlands by a road they all were well acquainted with, unexpectedly came to the borders of an extensive lake; the land had sunk, and water had flowed over it or gushed up out of the earth and formed a new lake. The opposite shore they felt confident could not be far distant, and they traveled on in tepid water, from twelve to forty inches in depth, of a temperature of 100 degrees, or over blood heat, at times of a warmth to be uncomfortable, for the distance of four or five miles, and reached the highlands in safety.

On the 8th of February, 1812, the day on which the severest shocks took place, the shocks seemed to go in waves, like the waves of the sea, throwing down brick chimnies level with the ground and two brick dwellings in New Madrid, and yet, with all its desolating effects, but one person was thought to have been lost in these commotions.

The morning after the first shock, as some men were crossing the Mississippi, they saw a black substance floating on the river, in strips four or five rods in breadth by twelve or fourteen rods in length, resembling soot from some immense chimney, or the cinders from some gigantic stove-pipe. It was so thick that the water could not be seen under it. On the Kentucky side of the river there empties into the Mississippi river two small streams, one called the Obine, the other the Forked Deer. Lieutenant Robinson, a recruiting officer in the United States army, visited that part of Kentucky lying between those two rivers in 1812, and states that he found numberless little mounds thrown up in the earth, and where a stick or a broken limb of a tree lay across these mounds they were all burnt in two pieces, which went to prove to the people that these commotions were caused by some internal action of fire.

About four miles above Paducah, on the Ohio River, on the Illinois side, on a post-oak flat, a large circular basin was formed, more than one hundred feet in diameter, by the sinking of the earth, how deep no one can tell, as the tall stately post-oaks sank below the tops of the tallest trees. The sink filled with water, and continues so to this time. The general appearance of the country where the most violent shocks took place was fearfully changed, and many farms were ruined.
After reading this and several reports about other earthquakes that are quite similar, I find it very hard to understand how there can be any opposition to the notion that the eruption of gases is connected with earthquakes, and possibly a major cause of them. I know of no way in which an area of land could suddenly sink by tens of feet, except by the release of large amounts of gases whose pressure had previously held open a large total volume of pore-spaces in the underlying rocks.

The same consideration applies to the creation of the earthquake-related ocean waves called tsunamis. A rapid and very large change in some volume is necessary to set up these waves, and that volumetric change has to be of a magnitude similar to the volume of ocean water that has been displaced to make either the negative or the positive phase of the great wave. Again sinking of an area of ocean floor due to the sudden escape of gases would be a possibility as would the rapid expansion of gases that make their way from the ocean floor to the surface. There are various reports of violent bubbling of areas of the ocean, and even of flames emerging out of the water.

Another feature of earthquakes that seems incompatible with the theory of shear strain in the rocks reaching breaking point are the deep source earthquakes. Earthquakes are known at depths down to 700 kilometers, and the pressure there is so great that sudden fracture cannot occur. The friction between two masses that slide against each other would be so great that this would far exceed any mechanical breaking strength of any rock. Any movement at such depths would occur only as a gradual adjustment proceeding in step with the driving force that causes the movement. This implied that another process must be going on down there and finding the answer to that may also then explain the features of shallower earthquakes that have so far remained unexplained, but that appear in seismic investigation quite similar to the deep ones.

We have two recent examples: on June 8, 1994, a very large earthquake registering 8.2 emanated from 600 kilometers below Bolivia. Not far away in time and space, in 1970, there was a powerful deep earthquake in Colombia.

The southern island arc of Indonesia and its continuation into Burma and the mountains of southern China is a very long belt that shows many features that show themselves along the whole length. Earthquakes make clear that it is related to an underlying structure of very large dimensions. The two other features that follow this same arc all the way are active volcanoes and the commercial production of oil and gas. While the belt was defined by the frequent occurrence of small quakes, it is also the region of the highest frequency of large quakes. In the 75 years between 1897 and 1972, there were ten earthquakes of magnitude eight or larger along this belt. There are no signs of a progressive shift of some land masses against others, and the rock stress situation is surely totally different in the folded mountains of Burma as in the volcanic island arc of Indonesia.

There are other features of earthquakes that have also to be considered. There are places that are distinctive "earthquake spots." There is a spot in northern Norway where for a long time one could almost be guaranteed to feel an earthquake in any 24 hour period. These were weak earthquakes, not much above the level at which one could feel them, but there was no faultline that was slipping, no accumulation of any deformation of the surface, it just kept shaking in an area that was about 12 kilometers across. A very similar story comes from two places in the United States, one is on the western tip of Flathead Lake in Montana, the other is in Arkansas, near the small town of Enola. Both of those have been active in recent times, and the one in Arkansas is known to have been active some 80 years ago.

Another earthquake spot is on the north shore of the St. Lawrence River, most interestingly just in a large meteorite impact structure ("astrobleme") called Charlevoix. The large meteorite struck there some 350 million years ago, and detailed evidence of this impact has been obtained. Despite the length of time that has elapsed since then, it seems that even now the area has not settled down and some activity is still clearly centered there. Some earthquakes that can be felt occur there every few days, and microquakes are registered extremely frequently. In this case, the proximity to the major faultline of the St. Lawrence River complicates the discussion somewhat but, nevertheless, the concentration of the seismic activity to the 30 mile diameter impact area is quite evident.

Such spots clearly need a different explanation from that of plates shearing against each other. Possibly the explanation has to do with gases forcing their way up and causing fractures in the rock to open and shut repeatedly.

We have investigated in some detail the Arkansas and the Charlevoix spots, and in the course of this discovered that they both contain a most intriguing feature which has shed further light on this type of occurrence. This is the presence of clusters of earth mounds that stand abruptly out of the alluvial plain. From a few feet to 40 feet in height and up to 200 feet or so in the horizontal dimensions, they are composed internally just of the clay and sand of the local alluvium, and no good reason has been offered to account for their origin.

The association in both areas of these strange mounds with locally concentrated seismic activity cannot reasonably be ascribed to chance. While such mounds do occur elsewhere, dense clusters of them are extremely rare, and an explanation for them is required. One cannot argue that the shaking of the ground of the earthquakes would itself cause what appears to be a substantial extrusion from below.

A class of a much larger type of feature is known and referred to as "mud volcano." It is also strongly related to earthquake activity. Mud volcanoes are mountains that are in the general shape of a volcano, sometimes but not always with an open hole on top and with steep sides sloping down to the plain below. The sides are made of rock debris, which presumably was ejected at the top as a mixture of such debris with water. Huge fields of mud volcanoes exist in several areas of the globe. The best known ones and the largest are in Azerbaijan on the north slopes of the Caucasus. Large eruptions of individual mud volcanoes are common there and the gases that propel the eruption are usually flammable and become ignited at the time, presumably by electrostatic sparks resulting from the friction of fast moving rock grains. Flames to a height of two kilometers have been photographed from one mud volcano whose orifice measures 120 meters across.

The gases coming out of mud volcanoes have often quite unusual composition and contain elements that are known to be at a high concentration in the mantle of the earth and at a much lower concentration in the sediments and in the outer crust. They clearly represent a very different chemical environment from that of the sedimentary cover.

The mounds on the earthquake spots in Enola, Arkansas and in Charlevoix on the St. Lawrence River, can be attributed to the same class of phenomenon as mud volcanoes, only on a much smaller scale.

Gases that stream up out of cracks during earthquakes are also frequently flammable. In the collection of eyewitness reports, flames are frequently a feature. Also in recent times, the great earthquake in San Francisco in 1906 was accompanied by large fires, and it was said at the time that this was due to the fracture of gas pipes in the ground. That may well have been the case; however flames were also seen on hills nearby that had no gas pipes and also on roads and fields in nearby San Jose. The Armenian earthquake of 1990 showed a line of burnt bushes along a visible faultline.

Large vertical displacements of areas of land can be understood if a mass of gas had previously held open pore spaces in the rocks below, and thereby raised the ground, and if these pore spaces had suddenly made connections to the surface and rapidly exhausted the gas. Such volumetric changes occurring in a matter of seconds can then account for the large tsunamis and for the flames often seen in earthquakes. As methane appears to be the most common gas in the rocks, it would seem reasonable to expect that methane would be the principal gas responsible, just as it is known in the case of mud volcanoes. The mud volcanoes merely show the locations in which earthquakes and gas eruptions are particularly frequent, and locations in which large amounts of underground mud have been generated by the frequent agitation of ground water in some fine-grained alluvial sediments.

Can the emission of gases be used for precursory information?

There were two observations before the earthquake at Loma Prieta on October 17, 1989 that seemed to be gas related and are clearly just prior to the earthquake (Reimer, 1990 and Fraser-Smith, 1989). But these observations were made for different purposes, unrelated to earthquake research, and yet they constitute the best earthquake-predictive observations. One was the observation of the amount of helium in a shallow well, which showed a sharp increase a day before the quake (Figure 2). I suppose that this represented an increased flow of gases upwards through the rocks, that had gathered up the helium that had accumulated in the pores.

Helium Graph
© Reimer, 1990
The other observation was that of a low radio frequency noise that is not normally present, also seen just before the quake (Figure 3); I attribute this to the interruption and reconnection of earth currents normally flowing in the groundwater, as these current paths are interrupted and re-connected by the bubbles of insulating gases that stream through the pores of the rock. Would these and other gas-related precursory effects not form the best line of earthquake investigation, to devise the most important of all, a predictive capability?

Magnetic Field Data
© Fraser-Smith 1989
The eye-witness stories of the past are all ignored or not even known to the present investigators; they are certainly not mentioned much in the modern earthquake literature. See the related documentation describing historical accounts of many large quakes (in "Eye-witness Accounts of Several Major Earthquakes," this Web site).


Cox, K. G. (1978). Kimberlite pipes. Scientific American 238 (4).

Fraser-Smith, A.C. et al. (1989). STAR Laboratory, Stanford University, Stanford, CA 94305.

Kennedy, G.C. and Nordly, B.E. (1968). The genesis of diamond deposits Econ. Geol. 63, 495-503.

Melton, C.E. and Giardini, A.A. (1974). The composition and significance of gas released from natural diamonds from Africa and Brazil. Amer. Mineralogist 59, 775-782.

Reimer, G.M. (1990). Helium increase. Nature, 347, 342.