Hot Spots and What Lies Beneath Them

An Old Idea Sparks New Debate
Wrangling Over Waves
Elemental Discord
Large Igneous Problems
On the Horizon

An Old Idea Sparks New Debate

A hypothetical model of a mantle plume, in a time sequence (left to right) showing temperature and several lines of tracer points. By Geoff Davies. Used by permission. URL:

The theory of plate tectonics neatly explained many of the mysteries of earth science: volcanoes, earthquakes, and continental drift. However it posed a mystery: why volcanism refuses to be confined to the margins of tectonic plates, but often appears as "hot spots" scattered across the planet. Hot spots are characterized as relative topographic highs that are capped by active or recently active volcanoes. They persist for about 100 million years. To solve the problem, geologists turned to the Mantle Plume Model, which posits the existence of uprisings from the core/mantle boundary.

J. Tuzo Wilson was the first to propose a mechanism for the creation of volcanic island chains. Inspired by Hawaii, Wilson envisioned a column of magma erupting to the surface. As tectonic plates move over this column, a chain of islands, growing progressively older with distance, develops. W. Jason Morgan of Princeton expounded upon this by proposing a mode of convection independent of plate tectonics. Mantle plumes are created by the temperature differential at the core/mantle boundary, he suggested, when a hotter region within the mantle rises as a result of decreased density. The plumes' narrow tails (~100 km) are rooted in the lower mantle, below the level of vigorous convection associated with plate tectonics. The plume head gradually enlarges itself and becomes cooler, while the tail remains hot (Morgan, 1971). In November 2003 he won the prestigious National Medal of Science for his work.

The idea is not difficult to grasp, but its simplicity belies the current discord that surrounds the theory. Plumes have never been an area of complete harmony. Geologists have argued over the years about their number and depth. In addition, there are degrees of belief in plumes; some believe they are responsible for all anomalous volcanic activity, while others think they only cause a small fraction of it. The arguments on both sides have been heating up and G.R. Foulger, a seismologist at the University of Durham in England, is often at the forefront of the fray, publishing numerous papers criticizing the Mantle Plume Model and maintaining a Web site at (

From this photo of Biscuit Basin, it is not difficult to see that something unusual is going on under Yellowstone. But exactly what is contentious. From URL:

The alternative to the Mantle Plume Model lies not in refuting the existence of hot spots (which are pretty plain to see) but rather in the completeness of plate tectonic theory - that plumes are not necessary to explain the various processes in question. The advocates of this model point out that the Mantle Plume theory was originally formulated to explain the existence of stationary plumes, yet it currently encompasses the Icelandic hot spot, which is clearly on the move (Foulger 2002).

Return to Top
Some suggest that the debate stems from a lack of a formal definition for the word "plume" (Wikipedia 2004). Others contend that the word is ill-defined because of the frequent ad hoc changes made to plume theory in order to reconcile conflicting data. "If plume hypothesis cannot be adapted to fit the observations, then the observations are commonly adapted to fit the hypothesis," declares Foulger (Geology News 2003a). Yet others take issue with the current state of the investigations undertaken to verify plume theory, insisting that a statistical approach is necessary for a truly comprehensive theory (Anderson, 2003).

The question is more than merely academic. Volcanism releases carbon dioxide, which can lead to global warming; if the carbon dioxide is released into the ocean, ocean currents can be affected. Plumes could be the cause of massive eruptions that have altered the environment on earth drastically, perhaps to the point of mass extinction. They may also have a role in the formation of mountains and the reversing of the earth's magnetic poles (Larson 1995). There is almost no end to the variety of methods used to detect plumes, and their implications to earth science. This Web site will explore the most frequently cited geochemical and geophysical evidence for mantle plumes, and as well as their more profound implications.

Wrangling Over Waves

Seismic tomography is the state of the art for mantle plume detection. It involves the use of seismic waves, which can be categorized as S or P waves. S waves are transverse waves which displace the ground perpendicularly to the direction of wave propagation. P waves are compressional, and alternately compress and dilate the earth in the direction of wave propagation. These waves generally travel twice as fast as S waves and can travel through any type of material. Both P and S waves are created by earthquakes, although seismologists can also generate them artificially. The time that the waves arrive at seismic stations can then be used to calculate the waves' speed through the Earth. By combining the data from many earthquakes across the globe a three dimensional map of wave speed through the earth can be generated. Surface waves, which cause displacement along the ground, can also be used to detect subsurface features; the process is then called surface-wave tomography.

Beginning in the early 1980's seismologists began to accumulate seismological data on the earth's interior. Seismic wave speeds differ with pressure, temperature, and rock composition. Typical speeds for P waves are 330m/s in air, 1450m/s in water and about 5000m/s in granite. Thus P wave velocity can be used to determine how far a plume extends into the mantle. Seismologists have recently discovered a 5- to 40-km-thick region at the base of the mantle where P wave velocities are depressed by as much as ten percent from the overlying mantle. This area is termed the ultra-low-velocity zone (ULVZ) and is conjectured to coincide with partial melting of the mantle. After an extensive survey, some geologists concluded that there was less than one percent chance that the correlation between the locations of hot spots and those of ULVZ's occurred strictly by chance (Williams 1998).

For the purpose of analysis, seismic waves are seen as a collection of lines. As the waves pass through molten rock, they slow. However, with a narrow body such as a plume, the waves can later regain speed as they are influenced by adjacent parts of the wave. Guust Nolet and F.A. Dahlen of Princeton propose that viewing a plume as a hollow banana is a more useful model. Only the peel of the banana is detectable to the curving ray path (Nolet 2003). Graduate student Raffaella Montelli of Princeton used this new technique, analyzing 87,806 seismic recordings. She concluded that deep mantle plumes are indeed present beneath Hawaii, Tahiti, and Easter Island. Some hotspot plumes (Reunion in the Indian Ocean and the Azores in the Atlantic) branch off of a superplume that rises beneath the South Pacific and Africa (Kerr, 2003).

Return to Top

Were the Hawaiian islands really made by a mantle plume? Inquiring geophysicists want to know.
Taken from U.S. Fish and Wildlife Service (National Image Database).

Seismology is unfortunately not an exact science. Seismic wave velocities are not highly sensitive to temperature, partial melt, or the chemical compositions that could be indicative of plumes, so resolution is limited. Conflicting conclusions regarding the presence of plumes at a given location result. For example in a single year (1998) three differing conclusions on the Hawaiian islands were presented. P-wave travel time shows low velocities in the upper mantle, according to Bijwaard et al (Bijwaard 1998), but Katzman et al reported high upper mantle velocities beneath in the Hawaiian islands (Katzman 1998). In addition, Ji and Natof claimed to have detected a low velocity anomaly in the lower mantle 200 km northewest of Hawaii using scattered P waves (Ji 1998).

Some of this conflict may be due to the different ways of interpeting the data. "It seems to me that the critics don't understand some aspects of the basic problem," says Donald J. DePaolo, a geophysicist at UC Berkeley. "For example, they claim that the mantle beneath Iceland is not hot, and the mantle beneath Yellowstone is not flowing upward. My view of the data in the papers they cite is that it leads straight to the conclusion that the rock is hot and flowing upward, just as we expect for mantle plumes" (AIG 2003).

Because of these difficulties in detection and interpretation of the results, the number of accepted hot spot areas on the earth varies from year to year. In 1999, the number of hot spots peaked at 5200; now less than ten remain to be explained by this process. In 2004, Montelli et al reported that P-wave velocities clearly show a plume model at work in six locations: Ascension, Azores, Canary, Easter, Samoa, and Tahiti (Montelli, 2004).

Elemental Discord

Another line of evidence popular with plume enthusiasts is the ratio of helium-3 to helium-4. A higher ratio is characteristic of deep mantle origin, they argue. Similar information can be gleaned from isotopes of the elements neodymium, strontium, lead, and hafnium. The ratio of 3He/4He increases over time as 4He is produced by the decay of uranium and thorium. The present day atmospheric 3He/4He ratio is 1.39x10-6, and is referred to as RA. Geochemists infer deep origins whenever 3He/4He are in excess of 9 to 10 RA. These ratios have been found at hot spot locations such as Hawaii, and are consistently different than the basalts of the mid-ocean ridges.

Foulger, naturally, takes issue with this line of reasoning. The concentration of 3He predicted by this model would be as high as that found in gas-rich chrondriic meteorites, she insists (Foulger, 2002). This conflicts with cosmological data. A better interpretation would be that the high 3He/4He ratio arises from a deficiency in 4He in the upper mantle caused by low U+Th areas, and thus low rate of addition of radiogenic 4He.

Many hot spot volcanic systems display a temporal variation in their 3He/4He ratios. At Mauna Loa, Hawaii, 3He/4He ratios decrease from 18-20 RA about 250,000 years ago to 8 to 9 RA today. "Plume contamination" is the standard explanation for these findings. On its way to the crust, the magma could be affected by material from the upper mantle or seawater.

Large Igneous Problems

From time to time the earth spews forth vast quantities of lava, which blanket the earth in basalt many km thick. These areas, which exist both on land and sea, are called large igneous provinces (LIGs). To account for them, geophysicists have developed a two-fold model of mantle convection. Ninety percent of the mantle heat is dissipated by orderly convection cells. However, ten percent remains to erupt in massive plumes (Coffin, 1993). The events that result in LIGs are commonly called "superplumes."

A vast LIG exists beneath the Western Pacific, where ocean floor there was covered in layers of basalt dating from the mid-Cretaceous, much younger than tectonic theory stipulates. This LIG is larger than any others on earth, and it led geophysicist Roger Larson to conclude that superplumes have built vast suboceanic plateaus with "pulses" of lava. Formation of ocean crust doubled at onset of pulse and tapered off over the next 709 to 80 million years (Larson 1995).

Others have gone on to say that superplumes are not just responsible for LIGs, but are also driving forces of plate tectonics. In 1999 Richard Kerr suggested that two vast superplumes cool the earth from opposite sides (the Pacific and Africa). These are critical to the stability of the core, which would otherwise collapse under the weight of subducting plates. Surface waves, earthquake waves, and S waves were used to locate a mass of magma beneath Southern Africa, which bends to the northeast and rise beneath the Afar Triangle in the form of a plume (Kerr, 1999).

Mantle convection could be related to the geomagnetic field. Heat dissipation from the core by plume convection may upset the convective cycle of the geodynamo, causing a decline in geomagnetic reversals (Larson 1991). The Cretaceous Long Normal Superchron (CLNS), an unusually long period of consistent polarity, supports this idea. Others suggest that increased plume flux and heat removal would instead increase the rate of geomagnetic reversal (Gubbins 1994). Return to Top

On the Horizon

Mantle plumes are an interesting topic and new studies are emerging almost every day. From the recent controversy, it seems that geologists have come to one conclusion: there are several types of plumes, not just a single category. Otherwise, the topic remains as divisive as ever. Because of of the vast literature accumulated on the subject, anyone wishing to characterize a particular area as the result of a mantle plume will be faced with an array of conflicting studies. Some suggest that scientists' imperative to publish prodigiously has created the need for a theory that does not really explain anything. Foulger notes (tongue-in-cheek) that "the assume-a-plume approach has also relieved researchers of the hard work of thinking up new theories, a welcome relief in these days when we are all expected to publish six papers a year or else" (Geology News 2003b). A better understanding of plumes should be a top priority. They could elucidate how the cooling of the earth drives mantle convection; where the mantle s compartments lie; and the formation of large igneous provinces. They can also elucidate such far-reaching topics as the formation of the lithosphere and chemical constitution of the mantle (Kerr, 2003). Luckily, new technology offers hope of this. A deployment of large networks of broadband ocean-bottom seismic instruments could fill the gap in the data (Solomon 2000). In addition, recently developed electromagnetic techniques, which are more sensitive to rock composition and temperature, could be used in concert with seismic wave profiling for a more exact image of the earth's interior (Tajima 2000).

Return to Top


Anderson, Don L. 2003. "A Short History of the Plume Hypothesis: The Inside Story." URL: Retrieved 4/18/2004.

AIG 2003. "Volcanoes Inspire Fiery Debate." URL: Retrieved 4/16/2004.

Bijwaard, H., Spakman, W. and Engdahl, E. R. 1998. Closing the gap between regional and global travel time tomography. J. Geophys. Res. 103:30,055-30,078. Coffin, Millard F. and Eldholm, Olav 1993. Large igneous provinces. Scientific American. 270:10,42-49.

Foulger, G.R. 2002. Plumes, or Plate Tectonic Processes? Astronomy & Geophysics 43:6.19-6.23.

Geology News 2003a. Making the evidence fit the plume. Geological Society Homepage Retrieved 4/13/04.

Geology News 2003b. Plumes, plates and Popper. Geological Society Homepage URL: Retreived 4/13/04.

Gubbins, D. 1994. Geomagnetic polarity reversals - a connection with secular variation and core-mantle interaction. Rev. Geophys. 32 (1):61 - 85.

Ji, Y., and Nataf, H.C. 1998. Earth Planet. Sci. Lett. 159:99-115. Katzman, R., Zhao, L. and T. H. Jordan 1998. J. Geophys. Res. 103:17,933-17,971. Kerr, Richard A. 1999. Great African plume emerges as a tectonic player. Science. 285:5425,187-188.

Kerr, Richard A. 2003. Plumes from the Core Lost and Found. Science. 299:5603,35-36.

Larson, Roger L. 1995. The mid-Creatceous superplume epidsode. Scientific American. Vol. 272, Issue 2, p. 82-86.

Larson, R. L. and Olsen, P. 1991. Mantle plumes control magnetic reversal frequency, Earth Planet. Sci. Lett. 109:437-447.

Montelli, Raffaella; Nolet, Guust; Dahlen, F.A.; Masters, Guy; Engdaul, E. Robert; and Hung, Shu-Huei 2004. Finite-frequency tomography reveals a variety of plumes in the mantle. Science. 303(5656):338-343.

Morgan, W.J. 1971. Convection plumes in the lower mantle. Nature. 230:42-43.

Nolet, G. and F.A. Dahlen 2000. Wave front healing and the evolution of seismic delay times. J. Geophys. Res. 105:19043-19054.

Solomon, Sean C. 2000. Seismic Imaging of Mantle Plumes: Progress and Prospects. Plume 3 Conference, Kohaloa, Hawaii. URL: Retrieved 4/20/2004.

Tajima, Fumiko 2000. Modeling of mantle electrical conductivity anomalies associated with an upwelling hot plume. Berkeley Seismological Laboratory. URL:http:// Retrieved 4/21/2004.

UC Berkeley Campus News 2002. Press Release. URL: Retrieved 14/14/2004.

19.Wikipedia 2004. Mantle Plumes. URL: Retrieved 14/18/2004.

20. Williams, Q.; Revenaugh, J.; Garnero, E. 1998. A correlation between ultra-low basal velocities in the mantle and hot spots. Science. 281:5376,546-549.

Report for ES 767 Global Tectonics
Cheryl Sedlacek, April 2004