Roadmap | The review process | Home
   More on Hell

What the Hell is that?

Don L. Anderson

Seismological Laboratory, California Institute of Technology, Pasadena, CA 91125

Will Sager presents a heart warming history of his epiphany [Ed: See What the Hell is Hawaii? page]. My petrology friends often use the parable of a visitor from another planet and her inability, according to them, to understand rocks unless she is taken aside by an experienced petrologist who explains the mysteries to her. She then understands that the mantle is composed of homogeneous isothermal refractory pyrolite and that all melting anomalies are due to the importation of high absolute temperature material from deep inside the Earth, heated by the core. The canons of Bowen are revealed to her. The trick of "adding back" olivine to infer the temperatures of melt sources, assuming the composition of a primitive parent, are explained to her delight.

Her less indoctrinated planetary colleagues would be unable to understand Sager’s laments. Why should age progressions represent convective velocities? What is the matter with volcanism and fissures on a stationary plate? What about the thousands of small volcanoes not in a line? How can anything besides the inertial frame be "fixed" in a spinning convecting deforming planet? When James Dwight Dana visited the Hawaiian islands, he was like a first-time extra-terrestrial visitor. Dana (1849), developed the giant fissure hypothesis to explain the age progressions and parallelism of the island chains in the Pacific. This possibility was critiqued, along with others, by Clague & Dalrymple (1989) just 140 years later. The absolute fixity of hotspots and their very high temperatures were regarded as absolute constraints that ruled out all options except the fixed hot plume.

Giant fissures were found wanting because there was no obvious way to have melt, unless it was already there–the partially molten and inhomogeneous asthenosphere hypothesis. The assumption of most marine scientists was, and is, that the upper mantle is subsolidus and uniform and that melting requires importation of heat from the deep mantle or frictional heating of rapidly moving plates. The perceived requirements of fixity and high temperature, and an homogeneous shallow mantle, favored the plume idea and other ideas were shoved onto the back burner. Subsequently, many volcanoes were dated and few showed the expected age progressions or lack of motion with respect to others. Since it was now known that volcanic islands were fixed, and were spots, the observations were interpreted in this light. It was too late to say “never mind” and start afresh. New hotspots were placed at places where the age progressions along a chain were irregular; this explains the proliferation of new plumes, all in a row, in the south Pacific. Other irregularities were explained by numerous changes in the directions and velocities of the plates or the poles of rotation. Hotlines were proposed. The radii of plume-spots were increased to retain the plausibility of fixity. Plume juice was allowed to flow for long distances to fix up other problems. All islands became plumes.

Regions where the fissure hypothesis was most plausible, based on the tectonics (e.g., Samoa, Juan Fernandez, San Felix, Afar, Kelvin fracture zone) were also treated as plume tracks. Most hotspots are currently on, or started on, ridges or fractures zones. In fact, many hotspots are named after the fracture zones they are on (e.g., Canaries, Azores, Jan Mayen). Some are clearly reactivated abandoned ridges. But these observations or concerns were also ignored.

As problems mounted, the concepts of mantle roll, mantle wind, group motions, westward plate drift, Euler pole jerks, long-distance lateral transport, deformable plates, hidden plate boundaries, and true polar wander (TPW) were introduced. There is nothing wrong with TPW but it has often been used with hotspots as a known fixed reference frame, just to fix up discrepancies with the plume hypothesis. The other complications removed the elegant simplicity of the fixed plume hypothesis, in which knowledge of the plate motions and single hotspots could predict all the island chains in the world. Modeling now requires a convection and a viscosity model, and a starting depth, radius and time for each plume, plus adjustable plate motions and deformations. Hotspot motions can be satisfied but at enormous sacrifice in simplicity and plausibility.

Some volcanoes on the Earth’s surface are now attributed to features on the core some 7000 km away and in a different ocean. Most hotspots are not hot and do not have high heat flow, but these problems have been rationalized away. The most common method to rationalize problems is the Texas Sharpshooter Method, well known to statisticians. Re-label the hotspots that fit as Real Plumes or Primary Plumes, and relegate the others to secondary or tertiary status; they are not really plumes. This is also known as the Pick-and-Choose Method.

Meanwhile the melting problem has been solved. The upper mantle is heterogeneous and contains substantial low melting blobs. The solidi of parts of the mantle are well below the minimum melting temperature of refractory peridotite. Small volcanoes are being found that have sources that underlie old lithosphere and in what was thought to be particularly cold mantle. Flexure of the plate allows magma to access the surface. The problem is not "why are there volcanoes in some places" but "why are there not volcanoes everywhere?" Relative fixity can be understood with asthenospheric counterflow models, and the stability of the global stress pattern. We know that some, perhaps many, oceanic islands and plateaus are underlain by continental crust and we know that the lower continental crust can delaminate and enter the mantle as mafic blobs. We know that the deep mantle geotherm is unlikely to be an adiabat. Because of radioactive heating and subduction cooling it must have a maximum at depth.

Although the negatives of the plume hypothesis outweigh the positives, it has been difficult for many in the Earth science community to abandon the idea. It keeps getting "fixed up". We now know that the hallmarks of the plume hypothesis–fixity, parallelism, and high heat flow–are not valid, but this has not been enough to abandon what has become an albatross. Even the basic assumptions are now known to be wrong–an homogeneous, isothermal upper mantle. When these assumptions are dropped, there may be no need for the plume hypothesis, but these amendments have also have been incorporated into the hypothesis.

The visitor from another planet would look at the evidence and would come up with a sensible hypothesis. If a marine geophysicist were to take her aside and present her with the plume theory, the visitor would certainly say “What the Hell is that?” (e.g., Anderson, 2007). Such a strange idea would not be publishable on the home planet.


  • Dana, J.D., 1849, Geology, in Wilkes, C., ed., United States Exploring Expedition, Volume 10,
    with atlas: Philadelphia, C. Sherman, New York, Putnam, p. 756 pp.
  • Clague, D. A., and Dalrymple, G. B., 1989, Tectonics, geochronology, and origin of the Hawaiian-Emperor volcanic chain, in Winterer, E. L., et al., eds., The Geology of North America, The eastern Pacific and Hawaii, Decade of North American Geology, vol. N, Geol. Soc. Amer., Boulder, CO, p. 188-217.
last updated 26th July, 2007