• Readers
  • Authors
  • Partners
  • Students
  • Libraries
  • Advertising
  • Contacts
  • Language: Русская версия
Section: Earth Sciences
Reflections of Eternal Flames

Reflections of Eternal Flames

Only few people have heard of such a geological phenomenon as mud volcanism. This is a process of natural degassing of buried sedimentary rocks rich in organic substances. The majority of mud volcanoes are concentrated in areas that are rich in deposits of oil and hydrocarbon gases. The gases (mostly methane) usually escape quietly, but sometimes this process ends with real eruptions. Moreover, these eruptions are often accompanied by "fiery events" (ignition of gas jets), the duration of which can vary from minutes to several years.

Siberian researchers, together with their Israeli colleagues, have found a province of unknown ancient mud volcanoes in the Hatrurim formation near the Dead Sea, which they called the Levantine province. Scientists suggest that fiery eruptions of the mud volcanoes of the Levantine province could have caused the death of the Biblical cities Sodom and Gomorrah


October 2005. The Hatrurim Basin in the Negev Desert, Israel. The Dead Sea lying deep below and the Jordanian border of the Dead Sea rift rising in the front, about one kilometer above. Regular piles of sediments deposited by a warm Cretaceous sea around the basin and hundreds of straw-yellow cones of rock medley…

Dr. Evgeny Vapnik, our colleague from Beer Sheva University, invited me (E. S.) to come to Israel as an expert in natural coal fires and in the so-called combustion metamorphism, which had taken twenty years of my professional life by the time.

However, the event that brought me to Hatrurim and then to its “eternal flames” actually happened long before. It was in November 1986, in the Ilmen reserve (Southern Urals), where I began my carrier after the university. I was going to leave for Novosibirsk Akademgorodok, when Boris V. Chesnokov, a brilliant Uralian mineralogist, handed me over a roll of microfilm and said, instead of parting wishes:

– You’ll find something incredible therein. Nobody has understood it yet. No idea should be swept aside, anything may be in that place. Come on, read and think it over.

On arrival in Novosibirsk, I had the film printed out and got a half-blind copy of a book titled The Mineralogy of the Hatrurim Formation, Israel. That was a plain description of 122 minerals by Shulamit Gross, a lady with a biblical name.

Most of minerals she described were rarities. Some were simple anhydrous oxides or silicates that usually make part of grogs and cement, and can form nowhere but on the Earth’s surface, under very high temperatures (up to 1000 º C). Others were high-water crystal hydrates similar to products of cement hydration. Both types surprisingly coexisted in a piece of rock. Of course, for me it was puzzling why and how those queer mineral assemblages could appear among carbonates around the Dead Sea. Neither could I get the answer from Shulamit Gross: she avoided any genetic interpretations.

It was not before nineteen years later that I got a chance to see Hatrurim myself. Every route made us ever more sure the Hatrurim geology would be hardly consistent with solid fuel combustion (Bentor et al., 1963), an idea popular with Israel geologists. Natural coal fires, which created particular combustion (fire born) landscapes in the American Great Plains or in the Kuznetsk Coal Basin in Siberia, always occurred in dry areas when the groundwater table was deep below the surface. In those places, they were the fired rocks that defined the terrain being much more resistant to erosion than the ambient coal-bearing formations. In the Hatrurim Basin, instead, hard fused rocks were set among strange sediments in tens of dispersed foci, mostly one or two meters thick. The material below bore signature of alteration by aggressive saline water while the sediments right under the conical hills were all cut with hydrothermal veins. Obviously, water and fire were inseparable companions in Hatrurim.

Especially striking were concentric ridges on cone tops consisting of high-temperature rocks welded in monolith hornfels. There had been fifteen such fired rings discovered by 2005. Judging by the geological setting, extremely high temperatures acted around those holes (10—15 m in diameter) with fine-grained mishmash inside. In vain did we try to overlook the scattered combustion foci from Zohar mountain rising high above the basin hoping to spot any systematic pattern: it was absolutely chaotic.

The high-temperature foci in the Hatrurim Basin were obviously local and lying shallow under the ground. The key discoveries of the first field trip were almost vertical narrow conduits filled with brecciated rocks lying exactly beneath the fired rings. It prompted our first hypothesis: the Hatrurim rocks may be associated with explosion pipes and the thermal alteration with venting inflammable gas. I told the Hatrurim story to Dr. I. Novikov, my colleague in Novosibirsk, and a month later he uttered a few words which sounded striking then but have become common today: “They are mud volcanoes.”

Once the key words had been spelled, there remained quite a routine thing of getting material evidence (it was indeed material, mostly mineral one). Svetlana Kokh became the first mineralogist to succeed Shulamit Gross, thirty years after, in studying the Hatrurim minerals. The site she chose was Nabi Musa Hill (Arabian for Prophet Moses Hill), a crater complex of a fossil mud volcano in the western side of the Jordan River, a perfect sample of the Hatrurim Formation.

E. V. Sokol

Mud volcanism results from natural degassing of buried sediments impregnated with organic matter whereby light gaseous hydrocarbons vent into air together with erupting water, mud, rock clasts, and occasionally oil. A universal definition of mud volcanism in terms of the buoyancy force was suggested by Kopf (2002), who interpreted it as buoyant rise of fluidized disintegrated sediments. Provinces of mud volcanism abound in flows of gas (mostly methane) with different flow rates. Gas seeps out slowly, but sometimes water- and gas-rich mud bursts to the surface to erupt in different ways, either as smooth mud flows or as explosions that erase the volcanic cones. 

Mud volcanoes are localized mainly in areas of past or present oil and gas generation (Kholodov, 2002; Kopf, 2002), such as, for instance, the Caspian petroleum province. Geological surveys in large mud-volcanic fields worldwide performed in the first half of the 20th century led to oil discoveries. Since the 1970s, when shelf oil development became technologically feasible, the studies of mud volcanism have shifted more to submarine provinces. Then mud volcanoes were found in the Black, Mediterranean, North, Barents, and Caribbean seas, in the gulfs of Mexico and Bengal, in the western Atlantic ocean, etc. 

Today more than 1800 onshore and submarine mud volcanoes are known, which are grouped in several large provinces in mobile belts, on continental rises and shelves, especially within the Alpine-Himalayan and Pacific belts. A few mud-volcanic fields exist among salt diapirs or in areas of extremely rapid sedimentation (such as sea-shelf levee deltas of large rivers).


Mud volcanoes were everyday reality for people who lived nearby, which remained imprinted in geographic names. That was the case of the Kerch-Taman’ and Caspian provinces (Kovalevsky, 1940; Shnyukov et al., 2005). The first genetic classification of mud volcanoes belongs to Cossacks who distinguished two types: a “peklo” (burner) and a “blevak” (spitter) corresponding to the fire and mud eruptions, respectively: Azov Peklo, Chernomorskoe Peklo, Gorelaya (burned) Mountain as opposed to Gnilaya (rotten) Mountain, etc. The same opposite polarities of something burning and swampy appear in Turkic and Persian local names in the Kerch Peninsula, in Azerbaijan and Turkmenia: Jau-Tepe (Enemy’s Mountain), Kainar-Ja (Boiling Place, Boiling Hill); Lok-Batan, At-Batan (Batan means quag; Lok-Batan means Camel Sunk in Quag), Otman-boz-dag (Shooting Gray Mountain). The western side of the Caspian Sea is an area of fire-related names. Toponyms alluding to eternal flame occur most often near Baku and Derbent. The gas flares of the Dagogni deposit were described as early as in the 5th century AD as “fire that comes out from a rock lying at the near-sea road to the Caucasus,” referring to the Derbent Pass (Kovalevsky, 1940). The gas flow rate was so high that people used that gas for burning lime and bricks till 1916. However, the most current is the Turkic name Otazh-Kaya or Atashkya (Fire Cliff), which originally had a religious meaning of sacrifice fire. This is the name used now for burning gas flares

Fire effects: eyewitness evidence

Eruptions of mud volcanoes in Iran, Azerbaijan, Kurdistan, in the Taman’ Peninsula, and in Trinidad Island are often accompanied by fire events, from a few minutes to years long. Fire effects were observed in twelve out of seventy seven eruptions in the Taman’ province from 1818 through 2005 (Shnyukov et al., 2005). “Flame eruptions” in Azerbaijan were as many as 40 % of 196 events between 1810 and 1974, of which about half were with gas flares that rose from 40 to 200 m high, one reaching 1000 m high (Yakubov et al., 1974).

A large flame eruption of Lok-Batan mud volcano, about 10 miles far from Baku, on 15 January 1887, became known due to a report by H. Siögren (1887), one of first explorers of mud volcanism. That is how he described the event (quoted from (Kovalevsky, 1940) and (Yakubov, 1941)).

Fifteen minutes before the eruption, gas began to escape into air with noise resembling effervescence from a steam boiler. “A while after the eruption began, the gas-air mixture exploded as the heating reached the hydrocarbon ignition point.” According to eyewitness evidence, the flare looked like a 400—500 m gas fountain. It was as if daylight had driven the night away for a moment. The glow was so bright that one could read a newspaper in Ajikabul Village 70 km south of Lok-Batan. At more than a mile far from the volcano, people felt unbearable heat, and the noise drowned the whistles of steam trains. The wind made the flame swaying and took mud and stones off to the south to throw them down about two miles far away.

Having climbed up the volcano forty days after the eruption, Siögren discovered gas to keep venting, with roar, from the crater fissures, and to go on burning. The combusted breccia around became brick-red and locally vitrified. As Siögren estimated, Lok-Batan had spitted out about 250,000 ton of mud for a few hours.

The flares are still more impressive when come from submarine eruptions. E. F. Shnyukov saw a flame arc over the Black Sea which was clearly visible from the shore during the Crimea earthquake of 1927. The arc traced the deep fault that generated the seismic event.

There were many reports of fire effects in the Caspian Sea as well. The captain of vessel Salyants who witnessed the mud volcano eruption of 1927 in the Kuman’ Bank mentioned that “the fire first appeared at the gas column top over the sea and only then fell down.” The same thing happened during the eruption of 1923 in Los’ Island: flame spread down from the gas column top to burn dry grass and birds’ backs. A similar disaster recurred in 1940 when a descending flame that accompanied a mud volcano eruption scorched the backs of animals and birds in Bulla Island.

Note that fire in gas flares always goes from top to bottom, either swiftly or slowly, according to decreasing flow rate, and the combustion front moves to the zone of fuel excess. This effect of flame influx is well known in technology (Knorre, 1955). When descending to the level of fractured hilly fields, fire can produce stationary foci of rock combustion. Gas flares in the hilly plains of the Apsheron Peninsula could keep burning for two years between mud volcano eruptions (Kovalevsky, 1940). The related high-temperature zone located from 2 to 12 m below fractured rocks was easily spotted from glow (red at 850—900 ° C, yellow at 1000—1050 ° C, and white at 1200—1300 ° C).

Back to Hatrurim

After a year of hard work we, to our surprise, had come to recognize an unknown province of fossil mud volcanism in an exhaustively documented region. Later we called it the Levantine province.

Although the earliest known mud-volcanic sites are as old as Cambrian-Ordovician (542 to 444 Ma), reliable finds of even small remnants of pre-Quaternary volcanoes are very few. Relicts of mud volcanoes are hard to identify because the cones composed of friable brecciated material degrade rapidly without leaving traces while no distinct diagnostic criteria are available. These criteria are, however, of utmost importance as provinces of mud volcanism, both past and present, are associated with petroleum reservoirs.

The Levantine province turned out to be a unique place where tens of perfectly preserved main and parasitic craters, conduits, and gryphons were found in the Negev and Judean deserts, and then in the Transjordanian plateau. How did they get so well preserved? They survived due to combustion of rocks in the craters maintained by permanently burning methane and, on the other hand, due to carbonatization and crystallization of calcium hydrosilicates, which transformed the originally friable sediments into something like “aged” concrete. The transformation was the job of ultra-alkaline water attendant with mud volcano eruptions in the area. Some artesian alkaline springs in today’s Maqarin complex in Jordan have water with the hydrogen ion exponent (pH) up to 12.9. Thus, they were jointly fire, water, and the extremely dry climate of the Middle East deserts that have conserved the Hatrurim fossil mud volcanoes.

As the features of the geological bodies and the mineralogy of fired rocks have prompted, there were several methane combustion regimes in the time of mud volcanic activity in the Levantine province. Gas was burning mostly under a layer of viscous sediment brought from greater depths, as it often happens in the present Apsheron and Taman’ provinces. The network of communicating fractures, a customary companion to eruptions, provided both the oxidizer supply into the burn zone and the evacuation of hot reaction products. With the time on, the low-grade (weakly altered) rocks of the upper layer were weathered away while the underlying fused rocks endured erosion and held on as armoring surfaces.

Foci of combustion metamorphism can reappear many times in a mud volcano section and thus imprint the spells of flame eruptions alternating with smooth flow of fluidized mud. Furthermore, the rhythmic pattern records periodical increases in gas pressure at the volcano root, i.e., methane influx from the main gas reservoir. The combustion temperature inferred from marker minerals turned out to be quite ordinary for this kind of rocks, in the range 750 to 1000 º C. The fired rings (see above) most likely were produced by gas flares after the extrusion of gas and mud had created a crater ridge. The fine grain sizes of hornfels and the scarcity of melting foci evidence that gas was burning for a short time: paralava (melt rock) exists in sporadic lenses formed at temperatures no hotter than 1220ºC.

The very causes of gas ignition during mud-volcanic eruptions remain unclear. According to what is known about active mud volcanism in the Kerch-Taman’ and Caspian provinces, the sequence of events is most often as follows: underground boom and ground motion – gas venting through fractures – explosion – a large gas outburst – appearance of a gas column – local gas ignition in the air (or an explosion in the air accompanied by ignition) – descent of flame down the gas column.

This scenario has a simple explanation in terms of physics of combustion and explosion. Gas in near-surface reservoirs always grows overpressurized before an eruption. The excess pressure releases in an explosion and in the ensuing eruption which begins with a violent gas outburst. The resulting shock wave propagates through the rising gas column whereby gas experiences heating-related adiabatic compression at the wave front while a zone of decompression and cooling forms at the back. The shock wave forming in a combustible mixture causes its spontaneous ignition, even in the absence of any other fire source (Knorre, 1955). In fewer eruption scenarios, it is the burning flare that detonates as the flame front surface increases and spreads over the whole mixture.

Fire, water, and old myths

The flame eruptions have always terrified the eyewitnesses. The reflections of those “eternal flames” must remain stored in ancient texts but are not so easy to pick. As S. A. Kovalevsky and E. F. Shnyukov, an expert in the Black Sea mud volcanism, agree, they were the Taman’ mud volcanoes that Homer must have referred to in his Odyssey. This was likely the place where he put the entrance to Pluto’s underworld in a “bare land of sorrow.” The hypothesis stems from paleogeographic reconstructions implying the path of Argo to go along the paleo-Kuban’ River, across the gulf of Taman’ and the Akhtaniz liman. It was exactly the way the ship should have taken to pass through a chain of “fire-spitting mountains,” the mud volcanoes of Gorelaya, Karabetova, Tsimbaly, Boris-and-Gleb, and Akhtaniz.

Few available isotope ages of the Hatrurim Formation, along with our paleogeographic reconstructions (Sokol et al., 2010), indicate that flame eruptions repeated many times between 3 Myr and 100 Kyr. As for the more recent history of Hatrurim, it appears in the local place names recognized for the first time by Avnimelech (1964), a geologist from Israel. One field of Hatrurim rocks, Kefar Uriah, is located near a village called Amwas (misspelled Greek Emmaus meaning a hot spring), which later transformed in Hammat in the Hebrew language (Avnimelech, 1964). The name is mentioned in the Talmud, though hot springs are no longer active in the area. Possibly, people in ancient Greek colonies witnessed the final active phase of the Levantine mud-volcanic province. Therefore, in the biblical time, the province may have been at the same stage the Kerch province is living through now, with few large eruptions on the background of dying activity.

At this very point, there arises a guess, moot and so far impossible to prove. It concerns a striking detail in the biblical myth of Sodom and Gomorrah: fire coming down from the skies. The fire was thus attributed to God’s punishment rather than to a foray of enemy neighbors. The two cities would have perished under a falling celestial flame. This is a thing one hardly can invent but can see: Remember flame eruptions, like the Caspian one, which can burn bushes, animals, and birds as the pressure of gas flow decreases and the flame goes down to the ground. Although archaeologists have never found the scorched cities, the story still sounds alive in the local names; the Sodom pillars of salt (diapir) is not far from the Hatrurim basin. Of course, the Old Testament story has many other interpretations, and anyone may choose what he or she likes.

In our trips through Israel and Jordan, we often came across Late Paleolithic non-finished or finished stone tools and flakes on the eroded tops of fossil mud volcanoes. It seemed strange why should people have taken pains of climbing up the hill instead of picking chert from primary deposits widespread all around at their feet. However, our ancestors proved to be wise enough. We understood it when Evgeny Vapnik and Igor Novikov discovered a primitive stone-tool workshop on the summit of an old mud volcano in 2007. The ancient population of the Hatrurim Basin obviously preferred to use natural chert clasts ejected from mud volcanoes rather than work hard to crush the monolith rock. The round pebbles high up the crater are the strongest pieces that have gone through mechanic sorting in the “mixer” of the volcanic conduit during an eruption.

We wondered whether it was a local feature or the Paleolithic community all over Eurasia profited from mud volcanic provinces elsewhere? There are reports available on the subject, the earliest being by Kovalevsky (1940), who found a Paleolithic quartzite scraper among the extrusion products of active Akhtarma-Pashalinskaya mud volcano in Apsheron. Early Paleolithic encampments in the Taman’ peninsula are located on the slopes of active mud volcanoes (Bogatyri, Tizdar, and Tsimbaly) or nearby. The tools found there are made of silicified dolomite present in ejecta (Vasiliev et al., 2008). Tel Ubeidiya, the oldest Paleolithic site (1.4 Myr) of Eurasia, is located in the Jordan River valley only 25 km east of the Maqarin complex that belongs to the Hatrurim Formation.

What other draws the landscapes of the mud volcanic province could have had for the ape man? They were likely the hot springs or medicinal mud, but especially the abundant and diverse salts, which were also inviting for large herbivorous animals.

Eduard Shteber (Russified from Stöber), a Russian oil chemist, suggested a surprising definition of man in his booklet What a Chemist Thinks about the Origin of Man published in 1928: “The man is an animal living in symbiosis with fire.” The idea of the book is that before arriving at symbiosis with fire, a primate needed some evolution-scale significant time for getting used to stay nearby, to overcome the innate animal fear of flame. Getting the habit required a special kind of fire: the one not being a disaster. According to E. Shteber, it can have been nothing but the burning small gas flows, i.e., stationary fire sources, such as the eternal flame of mud volcanoes.

Finally, the last link in the chain of hypotheses: The Caspian and some Mediterranean provinces of mud volcanism could offer to primitive humans the supplies of fire, water (saline), salt, and stone material for tools. Readers fascinated by this idea can try to compare the presumable paths of Early Paleolithic migrations with the map of mud volcanic belts.

April 2007. We’re driving across another hill along the new roadway Jerusalem – Jericho, when a stunning picture captures us. A huge crater, filled with extruded chert and dolomite clasts, opening in a 30 m fan among dazzling white and yellowish sedimentary rocks. We must be the first geoscientists to see, with our own eyes, a full section of a fossil mud volcano. Its body contains numerous veins of melt rocks, at depths from 7 to 25 m below the surface, which follow narrow tortuous fractures, the former channels of hot gas flows.

As we estimated later from indicator minerals and melt inclusions (Sokol et al., 2010), melting began at 1500 º C, the temperatures that only gas flares can maintain in natural near-surface conditions. The extruded material included pieces of Nubian-type red sandstone, a large subeconomic petroleum reservoir rock in Arabia. Therefore, the Levantine gas pools must have fed multiple mud volcanoes for quite a long time and may be still rich. Hopefully, a revision of the petroleum potential of this wonderful volcanic province is not too far off…

Bentor, Y. K., Gross, S. & Heller, L. Some unusual minerals from the «Mottled Zone» complex, Israel // American Mineralogist, 1963. V. 48. P. 924—930.
Gross S. The mineralogy of the Hatrurim formation, Israel // Geological Survey of Israel, Bulletin, 1977. No. 70. P. 80.
Kholodov V. N. On the origin of mud volcanoes // Priroda, 2002. No. 11. P. 47—58 (in Russ.).
Knorre G. F. What is Combustion? M.-L.: Gosenergoizdat, 1955.223 pp. (in Russ.).
Kopf A. J. Significance of mud volcanism // Reviews of Geophysics, 2002. V. 40 (2), P. 1005—1012.
Kovalevsky S. A. Mud Volcanoes in Southern Caspian Area (Azerbaijan and Turkmenia). Baku: Azgostoptekhizdat, 1940. 200 pp. (in Russ.).
Shnyukov E. F., Sheremetiev V. M., Maslakov N. A. et al. Mud Volcanoes in the Kerch-Taman’ Region. Krasnodar :GlavMedia, (2005). 176 pp. (in Russ.).
Shteber E. What a Chemist Thinks about the Origin of Man. Vladikavkaz: Ingushetia Printing House, 1928. 27 pp. (in Russ.).
Sokol E., Novikov I., Zateeva S., Vapnik Ye., Shagam R., Kozmenko O. Combustion metamorphishm in Nabi Musa dome: new implications for a mud volcanic origin of the Mottled Zone, Dead Sea area. Basin Research, 2010. V. 22, P. 414—438.
Svensen H., Planke S., Polozov A. G., Schmidbauer N., Corfu F., Podladchikov Y. Y., Jamtveit Bj. Siberian gas venting and the end-Permian environmental crisis // Earth and Planetary Science Letters, 2009. V. 277, P. 490–500.
Vasiliev S. A., Derevianko A. P., Matishov G. G., Amirkhanov Kh. A., Shchelinsky V. E., Velichko A. A., Medvedev G. I., Vishnyatsky L. B., Kulakov S. A. , Titov. V. V. (Eds.). Early Paleolitic of Eurasia: New Discoveries. Abstracts, Intern. Conf., Krasnodar – Temryuk, September 1—6. Rostov-Don, 2008. 208pp.
Yakubov A. A. Mud volcanoes in the western Apsheron Peninsula and their relation to petroleum potential. Baku: Izd. AzFAN, 1941. 102 pp. (in Russ.).
Yakubov A. A., Ali-Zade A. A., Rahmanov R. R., Mamedov Yu. G. Catalog of Recorded Mud Volcano Eruptions in Azerbaijan (from 1810 through 1974). Baku, 1974. 33 pp. (in Russ.).

We thank E. Vapnik, I. Novikov, L. Gilat, A. Sokol, and V. Kalugin, our friends and companions in travels and discussions, who shared with us our field life as well as our disappointments and discoveries

The study was supported by grant 09-05-00285 from the Russian Foundation for Basic Research and grant MK-6750.2010.5 from the President of the Russian Federation

Like the article? Share it with your friends

Subscribe to our weekly newsletter