How Volcanoes Work


  1. Introduction
  2. Magma and Eruptions
  3. The Earth's Crust and Causes of Vulcanism
  4. Examples
  5. References


This article attempts to present a readable account of the "theory of volcanoes" for nonspecialists who would like to understand this impressive class of natural phenomenona better. I say "class" because volcanoes and similar happenings are quite various, and not at all the same things. The volcanoes themselves will not be described in any detail; for this large body of information, the reader is referred to the References, where there are not only text references, but also web links to current events. Volcanic activity is a reality, and the reader will find it very interesting to study images of active and extinct volcanoes to see explicitly what we are talking about. The image of El volcán Fuego de Colima in the title is borrowed from the website in the References.

Volcanic phenomena, for the purposes of this article, include the following: central volcanoes, the familiar type; calderas, large explosive features; fissure eruptions, creating sheets of basalt; and diatremes, like the kimberlite pipes from which diamonds are obtained. Of these, central volcanoes and calderas are occurring at present or in the recent past, while fissure eruptions and diatremes are not. Volcanism specifically refers to phenomena with surface expressions. Related and similar activity occurs beneath the surface, such as the emplacement of dikes and sills, tabular features predominantly vertical and horizontal, respectively, called intrusives or hypabyssal rocks. Even the large features called batholiths have a similar source, although they are not intrusive but created in place. Hot springs and other hydrothermal or geothermal features also spring from the same source, but involve meteoric (ground) water as well.

The characteristic action of a volcano, in the popular view, is the emission of lava, molten rock. Until recently, even the scientific view was that the earth was formed as a molten ball, of which the surface has now cooled and solidified into the solid crust on which we live. Cracks in this crust would permit some of the hot lava below to break through on the surface, creating a volcano. Such a simple picture is not consonant with the structure of the earth, or of the distribution and properties of volcanoes. When it became clear from seismology that the earth was too solid for this, and that the probable temperature distribution was not suitable, geologists conceived that somehow the rock just beneath the crust, at a depth of 50 km or so, was melted, at least in places, either because of the release of pressure, or the "attraction" of heat to the region. The latter concept, so at variance with the second law of thermodyamics, was vaguely thought due to the decrease of thermal conductivity in the region, also a very hypothetical supposition. At any rate, it saved the pool of lava concept for the theory of volcanoes.

Magma and Eruptions

Lava before it is poured out at the surface seems to have properties beyond that of the simple molten rock, so geologists have called it magma. To avoid laborious circumlocution, I'll often use the terms lava and magma, and even rock, interchangeably where the distinction is obvious. However, I do not think of magma as necessarily fluid, as will become clear. The dictionary says that "lava" comes from Latin "lavare," to wash, but I cannot see the connection; there is no Latin word for lava, apparently. Magma possibly comes from a Greek word meaning to "knead," as "material kneaded," which is easier to see, but still not very satisfactory.

The types of rock discharged by volcanoes, not by any means exclusively as molten lava, are classified as basic, acidic and carbonitic. Carbonitic lavas are very special cases, consisting of things like sodium carbonate. Typically, such a lava may be 32% CO2, 30% Na2O, 13% CaO, 7% K2O, 8% H2O, and the rest various constituents. One of the rare volcanoes of this type is Oldoinyo Lengai in the Rift Valley of Africa. These are genuine igneous rocks, very different from the usual lavas. Of the ordinary lavas, basic lavas are the most common type by far. The major ingredient is calcic feldspar (plagioclase, mainly anorthite), perhaps 46%, then pyroxene, commonly as augite, 37%, and smaller amounts of olivine (iron and magnesium silicate) and iron ores. Basic lavas may differ greatly in composition, some containing major amounts of olivine (olivine-basalt), and different kinds of feldspar. With only small amounts of feldspar, they become pyroxenite or peridotite. Basic lavas contain very little free silica, SiO2. Olivine basalt contains no silica. Basic lavas melt to a mobile fluid at a little above 1000°C, a red-orange heat. If they solidify quickly, a black glass called obsidian is formed; if cooling is very slow, the coarsely crystalline rock gabbro results. Intermediate rates of cooling, seen in intrusives, give dolerite or diabase. Widespread sheets of basalt are also called "trap" rock from the Swedish for "steps" on account of its appearance in eroded hills. Basalt is famous for columnar structure, resulting from internal tensions upon cooling. The columns are vertical in a sill, horizontal in a dike. Like all lavas, eruptive basalt is accompanied by dissolved gases, which gives a cinder-like vesicular appearance to surface layers, from which the gas escaped. All basic lava is dark-colored.

Acid lavas contain free silica, perhaps 30%, but are mainly alkali feldspars (orthoclase), 52%, and 12% micas, usually biotite, which has soaked up most of the magnesium and iron. They melt at a considerably higher temperature than basic lavas, 1500°C and upwards, and are very sticky and viscous, flowing with great difficulty and never far. They cool rapidly to a glass, also called obsidian, but more commonly to the foam called pumice. Cooling very slowly, the result is granite if the predominant feldspar is orthoclase, and granodiorite if plagioclase predominates. Volcanic rocks are rhyolite, rhyodacite or dacite with increasing plagioclase. It is much more common for acid lavas to be erupted as a suspension of fine ash in hot gas than as lavas, and when this aerosol solidifies, it is difficult to distinguish from rhyolite, which was presumed to solidify from lava. In fact, most rhyolite probably was never molten, and is a kind of welded tuff. Tuffs are made from coarser pyroclastics, and are very typical of acidic vulcanism. Acid lavas can be light in color.

An interesting intermediate lava is andesite, named for its occurrence in the Andes cordillera. It is mainly plagioclase feldspar with less than 50% anorthite (more would put it closer to basalt), and biotite. There is little free silica, so it melts at a lower temperature and flows more readily. It is usually porphyritic, containing phenocrysts of minerals that have formed before the fine-grained groundmass solidified. It seems to be basic magma that has risen through and reacted with continental rocks. Similar rock in which orthoclase is predominant is called trachyte. The coarse-grained (plutonic) equivalents are diorite and syenite, respectively. Andesitic lavas are very common, often mixing basic and acidic characteristics. Around the Pacific "Ring of Fire" the separation of the andesitic lavas on the continental side from basaltic lavas on the oceanic side was so distinct that it was called the andesite line long before the appreciation of plate tectonics. This line coincides with the ring of oceanic trenches at subduction boundaries.

The style of eruption of a volcano depends very much on the type of lava and the content of active magma fluids. At one end of the scale, there are the quiet eruptions of Hawaiian type, of fluid basaltic lava with a modest amount of gases released on cooling. At the other end are the catastrophic caldera-forming explosive eruptions of acid lavas with high gas content, which have occurred at Krakatoa, Lake Taupo, Crater Lake, Yellowstone, Jemez, Santorini and many other places. The high pressure of the depths, tens of thousands of atmospheres, is communicated to the surface through a vent of low specific gravity (because of all the gas), and the eruption disgorges a huge cloud of white-hot pyroclastics and gas. Between these two extremes is every gradation. An andesitic pile like Mt. St. Helens or Popocatépetl sometimes exudes lava quietly, sometimes spits showers of pyroclastics, and sometimes explodes violently. Different types of magma may erupt at different times in the same volcano, but usually the type is fairly consistent. The explosiveness of acidic lavas is due to the tendency to form plugs near the surface that allow pressure to build up dangerously. Mont Pelée on Martinique showed the eruption of clouds of hot aerosols, the nuées ardentes, the plug forming the spine, and an explosion. The Leeward and Windward Islands of the Caribbean form a volcanic island arc at a subduction boundary, with a trench in front of them, just like the islands of the Ring of Fire. The volcano Kick-'em-Jenny at the southern end of the Windward Islands is a submarine volcano, only projecting 160 m above the ocean.

The largest volcanic eruptions are fissure eruptions of huge amounts of hot, mobile basaltic lavas that spread into large sheets burying the topography. The Deccan traps of India, which erupted at the end of the Cretaceous, and the Columbia and Snake river lavas of Washington, Oregon and Idaho, erupting in the Tertiary, are examples. The Deccan was poured out above a mantle hot spot that is now under the island of Réunion. It is easier to keep open a conduit of smaller cross-sectional area, so roughly circular conduits giving rise to central volcanoes with their typical conical shape are more common. In Iceland, fissure eruptions ending in a line of central volcanoes are evidence of this. The external forms of central volcanoes are many, but three ideal types can be identified. Hot, fluid lavas usually make shield volcanoes, like those of Hawaii. The less fluid the lava, the greater the slope, which seldom exceeds 10°. At the other end of the scale, explosive eruptions of tephra, a general term for ash, cinders and pyroclastics, give rise to steep-sided cinder cones, with slopes of up to 40°, with a bowl-shaped crater at the top ("crater" is Greek for a wine bowl of similar shape). These are generally rather small, since large ones would have been blown away in catastrophic caldera eruptions. Between these two is the familiar stratovolcano, composed of alternate layers of lava and tephra. Stratovolcanoes can be lofty and symmetrical, with upper slopes of about 35°, like Fuji in Japan and Klyuchevskoy in Kamchatka. The large volcanoes of the Cascade Mountains are stratovolcanoes. The lava is typically andesite, which varies in nature, so alternating quiet and noisy eruptions.

There is no precise distinction between an "active" volcano and an "extinct" volcano. A volcano that is emitting steam now and then, or has erupted in the past 50 or 100 years, or near which small earthquakes are observed now and then, is usually called "active." A volcano that is completely cold, with no earthquakes in the vicinity, and which has not erupted for a few thousand years is usually safely termed "extinct." When the active fluids have been exhausted, that is the end of that particular episode. Vesuvius had done nothing for hundreds of years before it blew up in 79, but has erupted occasionally ever since. Mt. St. Helens was a complete surprise, but all these volcanoes of the Pacific Northwest should really be considered active. Yellowstone shows hydrothermal activity, which means that vulcanism is lurking beneath the surface, and could be renewed at any time. The North Island of New Zealand is in a similar state. In both cases, the renewal could be a caldera explosion, preceded by a symphony of earthquakes building to a crescendo. Hawaiian eruptions are continuous and mild, Etna's are continuous and rowdy, while caldera explosions are very seldom and very violent. Eruptions are usually signaled by earthquakes and earth movements, but the magnitude of an eruption is unpredictable. Volcanic activity that occurrs once, then ceases and becomes extinct, is called monogenetic.

The Earth's Crust and Causes of Vulcanism

The earth's crust consists of a layer of basaltic rock about 12 km thick covering its surface. Here and there are mounds of acidic rock thick enough to press the base of the basalt to about 35 km. The base of the basaltic layer is called the Mohorovicic discontinuity or Moho, discovered by seismology in 1909, at which the velocity of seismic waves increases abruptly. This is the transition to the mantle, whose rocks are denser and more rigid than those of the crust. At least this part of the mantle is probably composed mainly of olivine, (Fe,Mg)2SiO4, an ultrabasic rock, so called for its lack of aluminum and feldspars. Another component is eclogite, composed of half garnet and half pyroxene, with some quartz. Both have been found in the samples of mantle rock that we find in diatremes. When these rocks are brought to the surface (usually underwater) they react with water to form an ophiolite suite of rocks that always marks a present or past spreading centre. The ophiolite suite consists of serpentine (after which it is named; ophios = snake), chlorite, epidote and albite. Albite is sodic plagioclase, epidote is a calcium aluminium iron silicate, chlorite is a green flaky mineral, and serpentine is a green, banded metamorphic rock made of antigorite (dark-green scaly), chrysotile (greenish fibres) and lizardite (scaly white). These are very unusual minerals, and their occurrence was long a mystery until their association with plate tectonics was discovered. The mantle is very deep and remarkably homogeneous, extending down to the earth's core at about 3000 km depth (out of 6371 km to the centre). The core occupies roughly 1/8 of the earth's volume, the mantle 7/8, and the crust is negligible. Some people believe that the crust has a role in vulcanism, but this is an outside chance.

Another, older, division of the earth considers the lithosphere to extend down to about 50 km, with the asthenosphere down to about 1200 km. The lithosphere was supposed to be rigid, and the asthenosphere, named by J. Barrell in 1914, apparently not. At least, it was not so rigid that isostatic equilibrium was hampered. Seismology disturbed this picture by showing that the asthenosphere was more rigid than steel, and that rock became more rigid with increased pressure, not less. The pressure at 50 km was calculated to be sufficient to overcome the shear strength of rocks, so that no open voids could exist at greater depths. There is no evidence of any discontinuity at this level, so it probably has little significance. The lithosphere-asthenosphere division still seems to be current in American geology, with plates consisting of the lithosphere floating on the asthenosphere. Much of this confusion over rock flowing has been eliminated through realizing that time is an important parameter in strength and flow, and that materials that seem to be rigid over short time intervals will flow like liquids over long ones. Therefore, any distinction between "liquid" and "solid" is not absolute, and they are not mutually exclusive categories. Many common examples of this, like wax, pitch, salt and ice were available, but were not noted except by Arthur Holmes and a few others.

Another extremely important factor, perhaps the dominating factor in vulcanism, is the efficiency of light, hot fluids, mainly water but also gases such as CO2, in the transmission of heat and the production of chemical change. Such fluids almost certainly make granite by reaction with gneiss in deeply-buried rocks. I shall assume here, as is likely, that the occurrence of such fluids is an essential and causative element in all volcanic processes, not an incidental one. It is the concentration of bubbles of such fluids at the mantle-core boundary that produces hot spots, such as those under Hawaii and Yellowstone, and many other places. These bubbles can be very persistent, and cause only intermittent and gradual eruptions. At active subduction zones, ocean water is carried down with the subducting plate, as well as water and sialic components in the sediments that stick to the plate instead of being scraped off, so the fluids are constantly being recruited. This has produced the "Ring of Fire" around the borders of the Pacific, as well as on segments of the Mediterranean-Indonesian volcanic trend. Together with vulcanism on spreading ridges (much of which takes place unseen beneath the ocean), where olivine basalt is directly brought to the surface, this largely explains the geographical distribution of volcanoes. Volcanoes are not scattered uniformly or randomly over the earth's surface, and are not a consequence of the average structure of the earth. That is, they have specific causes, and the major cause is the occurrence of active fluids rising from the mantle.


There are examples of all four kinds of vulcanism in the western United States, most of it Tertiary or Holocene. There is a chain of active andesite volcanoes in the Cascades, such as Ranier, Hood, Whitney and, of course St. Helens, that erupted in the early 1980's. Calderas are notable at Crater Lake, Yellowstone and the Valles Caldera in northern New Mexico. Tertiary fissure eruptions covered Oregon, Washington and Idaho with plateau basalts. There are clusters of diatremes between Laramie, Wyoming and Ft. Collins, Colorado, and in the Laramie Range of Wyoming. Beneath the Coast Ranges are granite batholiths. There has been a subduction zone along the west coast at least since the end of the Palaeozoic. Nevada and California (as well as points north and west) rode in on the subducting plate to be plastered against the craton, while a large piece rode under the continent, as India has ridden under Tibet, supplying lots of active fluids and acidic rocks. First the western part was raised to a mountainous level by the end of the Mesozoic, then the mass moved on eastwards, letting the Basin and Range drop and be dragged open, while the Rocky Mountain area rose to its present elevation, upper layers faulting in huge blocks. This was accompanied by an immense amount of volcanic material, both basic sheet basalts and acidic pyroclastics, that blanketed the area in the Tertiary.

Yellowstone is the end of the path of a particularly active source of fluids, that first filled the Snake River plain with basalt, and poured out the Absaroka rocks in the vicinity of Yellowstone. Then a large acidic caldera was formed by a more recent violent eruption, about 180,000 or 70,000 ybp, that blew out the Plateau Rhyolite and sank into the space vacated. The location of the caldera is shown in the map at the right. Fluids still rise in sufficient amounts to provide hot springs and 182 geysers. Although geysers involve mostly meteoritic water, the heat is provided by fluids from below. If it were just "hot rocks" hot springs and geysers would be much more common in all areas, and the rocks would have rapidly cooled anyway. On New Mexico-Colorado boundary there is the Raton plateau basalt, eroded to a prominent tableland, a field of basaltic shield volcanoes in the Raton-Clayton field along US 64 of somewhat later date, such as Sierra Grande and Mt. Dora, and finally a collection of cinder cones, of which Capulin Mountain is an excellent example, which erupted only a few thousand years ago with lava flows and pyroclastics. It is shown in the map at the left (adapted from USGS). The Raton-Clayton volcanic field is not far from the majestic Spanish Peaks near Walsenburg in southern Colorado, eroded Tertiary stocks in a swarm of basaltic dikes. They may never actually have been volcanoes. The small, eroded cinder cone or stock called the Huerfano is just east of I-25 north of Walsenburg.

The Jemez (or Valles) caldera, 14 miles in diameter, is west of Los Alamos on New Mexico Route 4. The name Valles is given to the later caldera, Jemez sometimes restricted to the ash flow of 2 mybp in the same area. The pink Bandelier Tuff is the acidic material that was expelled over a wide area around the site of the eruption, as was the similar rhyolite and obsidian of Yellowstone. The eruption was preceded by extensive basaltic lava flows, which characterize this whole region, and can be seen best in the walls of the canyon that the Rio Grande cut through them west of Taos, after they filled its valley in the Rio Grande graben. The floor of the caldera is still evident, with domes of volcanic rock. The Valle Grande once held a lake like that at Crater Lake, but it is now drained. The Valles eruption took place in the Pleistocene, and there was considerable fumarole activity afterwards.

A cinder cone 1731 m high visible to the northwest of Carrizozo, New Mexico is at the head of a dark lava flow that extends to the southwest a considerable distance, to opposite Three Rivers. US 380 crosses it west of Carrizozo, and another smaller flow further on, with two vents to the north. This is very recent (holocene) activity. NM 3 passes just west of Taos Volcano a few miles north of Taos. This activity was late Tertiary, as is the basalt flow that filled the Rio Grande rift valley to the west as far as Tres Piedras, with widely scattered cinder cones. West of Albuquerque on I-40, near Grants (MP 102), is Mount Taylor, one of the very few stratovolcanoes in the Rocky Mountain area, and important in Indian legend. Unfortunately, its top has been blown off by an eruption at the end of the Tertiary, but what is left is still 11,300 feet high. There is a large plateau of late-Tertiary basalt flows extending to the northeast, with numerous cinder cones.

Colorado and Wyoming are rather poor in volcanic sights, except for Yellowstone, of course. The Spanish Peaks are merely deeply eroded volcanic stocks, not stratovolcanoes, but it is worth looking at the dikes in this area. At Dotsero, on I-80 east of Glenwood Canyon, there is a small volcano northeast of the town, now being quarried for volcanic cinder to make concrete blocks. Some sources claim this is a maar, a ring of tuff blown out in an eruption, with a lake in the collapsed centre. However, topographic maps do not show such a round lake. There are various areas of Tertiary basalt, extensively in the San Juan, lesser amounts west of Gunnison, on Grand Mesa, on the White River Uplift and the Rabbit Ears Range, but no volcanoes to be seen. Wyoming has the Leucite Hills, northeast of Rock Springs and at the western end of a notable sand area. This is an area of Pleistocene volcanism, and pumice is quarried in the area. Leucite is a white, hard (6) and light (2.5) pseudocubic mineral with the composition KAlSi2O6, a feldspathoid in which each KAlO2 unit is accompanied by 2SiO2, while in orthoclase 3SiO2 is required. Therefore, there was no free silica in the lava, since there was not even enough to make all feldspar. Leucite weathers more readily than orthoclase. Outside of this, volcanic rocks are found in Wyoming only in the northwest as an extension of the Snake River vulcanism, and a little in the northeast as an extension of the Black Hills vulcanism (Devil's Tower). It is clear that not much went on, volcanically speaking, this far into the interior of the continent, until subduction pushed some water and rock under the area in the Tertiary and created some hot spots.


Volcano Live, by Australian volcanologist John Seach, has information on the earth's volcanoes and links to live webcams, as well as volcano news. Use this good site to access the volcanocams at Klyuchevskoy and Shiveluch in Kamchatka. There are AVI files of eruptions here, that are well worth viewing, so that you will know what an eruption of an andesitic volcano is like from personal experience.

Cascades Volcanic Observatory is an excellent website loaded with information, and a portal to other USGS sites. The material is public domain, but should be attributed out of courtesy. The main interest is, of course, United States volcanoes, but general information is included, with excellent maps and graphics.

The website Colima of the University of Colima is an excellent source of volcano information, with many views of this active andesitic stratovolcano. There are both English and Spanish versions.

Arthur Holmes, Principles of Physical Geology, 2nd ed. (New York: Ronald, 1965). Chapter XII. For Oldoinyo Lengai, see p. 1067.

Halka Chronic, Roadside Geology of New Mexico (Missoula, MT: Mountain Press, 1987). New Mexico has a wealth of volcanic sites, chiefly in the north and west, and along the Rio Grande rift. The volcanic area between Raton and Clayton is easy to see, as is the Jemez caldera, and are worth the visit.

Halka Chronic, Roadside Geology of Colorado (Missoula, MT: Mountain Press, 1987). Colorado has the Spanish Peaks, and Huerfano butte, accessible from Interstate 25. The San Juan mountains are volcanic, but there is little to see but mountains. Many volcanic caps are to be seen, as at Castle Rock and the Table Mountains near Golden. The Southern Rockies from New Mexico to Montana are not accompanied by much vulcanism.

D. R. Lageson and D. R. Spearing, Roadside Geology of Wyoming (Missouls, MT: Mountain Press, 1988). Wyoming has the outstanding Yellowstone area in the northwest. Yellowstone is Quaternary, but the great Absaroka volcanics in which it sits are Eocene. Oligocene rocks (White River fm) are largely ash blown in from the west, and loess, often weathering to bentonite. Devil's Tower in the far northeast is a remarkable basalt feature with columnar jointing.

A good tour of volcanic and geologic features of the New Mexico-Wyoming area would extend from Yellowstone on the north to Jemez caldera on the south, via Raton and the Spanish Peaks, then westward to Monument Valley in northern Arizona, with a side trip to the San Rafael Swell via Moab and Green River, then up over Utah 150 across the Uinta Range, then north via US 191 and 189 through the Overthrust Belt to Jackson, and finally to Yellowstone. This could be accomplished in about 10 days or a fortnight, either by RV or motel. The best time is mid-September to November, when Utah 150 is open, or else early summer, but you will have more company then. North to south is better in the autumn, south to north in the spring. A number of National Parks line the route, but they are usually packed with people, regulations and armed guards, and best avoided. Yellowstone, of course, cannot be avoided. More recommendations for such a tour may be developed later. There are, unfortunately, no active volcanoes in this region, but we can hope.

Return to Geology Index

Composed by J. B. Calvert
Created 19 February 2003
Last revised 22 February 2003