Top-Down Tectonics?

See allHide authors and affiliations

Science  14 Sep 2001:
Vol. 293, Issue 5537, pp. 2016-2018
DOI: 10.1126/science.1065448

This article has a correction. Please see:

T here are two competing models for mantle convection [HN1]. In the first, the mantle is stratified into two or more separate convecting regions. In the second, the whole mantle convects as a single unit. Recent progress in plate tectonics [HN2], seismology, solid-state physics, and mantle convection is providing strong support for stratified convection. The results may also help explain how plate tectonics relate to mantle convection. Upper mantle convection may be driven by plate tectonics, whereas the deep mantle [HN3] may convect in a completely different style.

Evidence for whole mantle convection comes primarily from seismology (1) [HN4]. Images of bright blue bands represent high-velocity seismic anomalies that appear to be slabs traversing the mantle [HN5]. The visual evidence for occasional slab penetration below 650 km (2) is usually taken as sufficient evidence for whole mantle convection. Whole mantle convection is also the reigning paradigm among geodynamic modelers because of the seismic evidence cited above and the similarity between the geoid [HN6] (the surface of constant gravitational potential that would represent the sea surface if the oceans were not in motion) and deep mantle seismic tomography [HN7] (which works much like medical x-ray tomography except that seismic velocities are imaged). Whole mantle convection simulations are also easier to do.

Arguments for stratified convection are more complex and harder to understand (2, 3). Pressure suppresses the effect of temperature on density, making it more difficult for the deep mantle to convect. It also suppresses the effect of temperature on seismic velocities, which are used by seismologists to map temperature variations. Ab initio calculations of mantle minerals (4, 5) indicate that subtle differences in seismic gradients and velocities may be compositional; even small changes in chemistry can stratify mantle convection. Furthermore, computer simulations of three-dimensional (3D) mantle convection with self-consistent thermal properties and variable heating (6) show thermochemical convection involving deep dense layers, which help explain the spatial and spectral features of tomographic models derived from seismic data [HN8].

An important measure of the vigor of convection and the distance from static equilibrium is the Rayleigh number [HN9], R. The smaller Ris, the harder it is for convection to occur. In a spherical shell, convection occurs spontaneously when Ris about 104 (7). Whole mantle convection models usually assume R > 107, but Tackley (6) [HN10] derives a value of only about 4000 for the base of the mantle. If the lower 1000 km of the mantle is isolated, R drops to 500.

These results have far-reaching implications. Small values of R imply that instabilities forming at the base of the mantle must be sluggish, long-lived, and immense. This is consistent with lower mantle tomography, which has shown that the deep mantle is characterized by two immense regions of low seismic velocity (8, 9), and makes it more plausible than previously thought for the mantle to be chemically stratified. Deep, dense layers need only be a fraction of a percent denser than the overlying layers to be trapped because thermal expansion is low and it is difficult to create buoyancy with available temperature variations and heat sources. The gravitational differentiation of the deep mantle may be irreversible, although mixing, overturn, and penetration may be possible at lower pressure and at an earlier stage of Earth history (10).

Equation of state modeling (which captures the equilibrium conditions of a system in terms of pressure, volume, and temperature) has shown that physical differences in the deep mantle, and across chemical interfaces in the mantle, must be very small and almost independent of temperature (4, 5, 11). What tools can seismologists then use to determine whether the mantle is chemically stratified? Most promising are spectral (8, 12), matched filter (9, 13), scattering, and correlation (9) techniques, as well as regional studies (2) and anisotropy (14) [HN11]. All these techniques support a seismic compartmentalization of the mantle with boundaries near 650 km, ∼1000 km, and ∼2200 km depth (9, 12, 15). Visual inspection of tomographic images (11) has also been used but has led to opposite conclusions (2, 16).

The most prominent seismic discontinuity in the mantle is at 650 km, but the boundary of the lower mantle should probably be placed at 1000 km, as proposed by Bullen and Jeffreys (2) [HN12]. Between 650 and 1000 km, steep subduction turns to predominantly horizontal flow; slablike features below 1000 to 1200 km are not connected to surface plates or presently subducting slabs (2) and have little correlation with subduction history (9). Furthermore, it has been inferred from anisotropy measurements (14) that the mantle is divided into two convective systems at 900 to 1000 km.

These inferences are bound to be controversial, but the evidence for a significant geodynamic boundary near 1000 km is as strong, although of a different kind, as the early evidence for other seismic discontinuities in the mantle (15). Whether the different mantle regions define independent compositional or convection regimes remains to be seen, but their existence provides constraints that challenge convection models and geochemical assumptions.

How does mantle convection relate to plate tectonics [HN13]? In 1900, Henri Bénard heated whale oil in a pan and noted a system of hexagonal cells. Lord Rayleigh [HN14] analyzed this pattern in terms of the instability of a fluid heated from below. Rayleigh-Bénard convection [HN15] has since become the classic example of thermal convection. In 1958, Pearson (17) [HN16] showed that Bénard's patterns were driven from above by surface tension. Bénard's patterns have also been used as the prototype far-from-equilibrium self-organized dissipative system [HN17].

There are several lessons to be learned from these experiments. First, things are not always as they seem. It seemed obvious that the system was driven from below and that the fluid was self-organizing via thermal buoyancy and viscous dissipation of the fluid. Actually, the system was driven and organized from above. Plate tectonics and mantle convection may also be organized and controlled from the top, not by surface tension but by the gravity-controlled compression that defines the plates and plate boundaries. The plates may also control the thermal evolution of the mantle (18), with resisting forces in the plates dominating over mantle viscosity.

Three ways to flow.

(Top) A fluid layer cooled from above or from the side, or heated from within, develops narrow cold downwellings that cool the interior. The downwellings are terminated by phase changes, density increases due to composition, or high viscosity. There are no active or hot upwellings. This resembles the upper mantle. (Middle) A high viscosity or isolated chemical layer cooled from above develops large cool downwellings. This mimics the mid-mantle (1000 to 2000 km). (Bottom) A deep, dense, high-viscosity layer with low thermal expansion overlying a hot region develops large, sluggish upwellings. This mimics the deep mantle.

Second, a far-from-equilibrium dissipative system is sensitive to small internal fluctuations and prone to massive reorganization. Such self-organization requires an open system, a large steady outside source of matter or energy, nonlinear interconnectedness of system components, multiple possible states, and dissipation. Plate tectonics is driven by negative buoyancy of the outer shell and appears to be resisted primarily by dissipation forces in the lithosphere (see the second figure).

Top-down tectonics?

The tectonic plates can be viewed as an open, far-from-equilibrium, dissipative and self-organizing system that takes matter and energy from the mantle and converts it to mechanical forces (ridge push, slab pull), which drive the plates. Subducting slabs and cratonic roots cool the mantle and create pressure and temperature gradients, which drive mantle convection. The plate system thus acts as a template to organize mantle convection. In contrast, in the conventional view the lithosphere is simply the surface boundary layer of mantle convection and the mantle is the self-organizing dissipative system.

If most of the buoyancy and dissipation is provided by the plates while the mantle simply provides heat, gravity, matter, and an entropy dump, then plate tectonics is a candidate for a self-organized system, in contrast to being organized by mantle convection or heat from the core. Stress fluctuations in such a system cause global reorganizations without a causative convective event in the mantle. Changes in stress affect plate permeability and can initiate or turn off fractures, dikes, and volcanic chains. The mantle itself need play no active role in plate tectonic “catastrophes.”

The difficulty in accounting for plate tectonics with computer simulations may be explained if plates are a self-organized system that organizes mantle convection, rather than vice versa. Upper mantle convection patterns should then be regarded as the result, not the cause, of plate tectonics. Whether the first-order features of plate tectonics emerge from this approach remains to be seen (19).

The mantle is usually considered as a homogeneous convecting layer that ex-presses itself at the surface in plate tec-tonics. Progress in understanding the base of the mantle, the mid-mantle, and the surface boundary layer show that this is much too simple a view. Theory shows that chemical stratification is difficult to detect with standard techniques. But a stratified mantle, along with the self-regulation of the plates, would slow down the cooling of Earth and postpone the inevitable heat death.

Thermochemical 3D convection simulations in spherical shell geometry and with self-consistent pressure-dependent thermodynamic properties and the possibility of deep undulating chemical interfaces will be required to test these ideas. If plate tectonics is a self-organizing system that also organizes mantle convection, then convection simulations need to allow multiple degrees of freedom so that all possible states can be explored.

HyperNotes Related Resources on the World Wide Web

General Hypernotes

Dictionaries and Glossaries

The Academic Press Dictionary of Science and Technology is provided by the publisher Harcourt.

The xrefer Web site offers a collection of scientific dictionaries and other reference works.

The Department of Geological and Atmospheric Sciences, Iowa State University, makes available a glossary of geologic terms.

The Treasure Trove of Physics is a glossary and reference resource maintained by E. Weisstein.

Web Collections, References, and Resource Lists

The Google Web Directory offers links to Internet resources on geology and geophysics.

Sci-Info from the University of Arizona Library offers a guide to Web geoscience resources.

Geophysics on the Internet is maintained by J. Butler, Department of Geosciences, University of Houston.

GeologyLink, presented by the College Division of the Houghton Mifflin Company, presents links to geology news and geology resources on the Internet. A geology glossary is provided.

Online Texts and Lecture Notes

Planet Earth and the New Geosciences is an online textbook by V. Schmidt and W. Harbert, Department of Geology and Planetary Sciences, University of Pittsburgh. A unit on continental tectonics and Earth's interior is included.

Visualizing Earth, an educational project funded by the National Science Foundation, provides illustrated introductions to geological processes.

The Past and Future of Planet Earth is a presentation by L. Moresi of the CSIRO Solid Mechanics Research Group, Australia. A section on the structure and layering of Earth is included.

V. DiVenere, Department of Earth and Environmental Science, Long Island University, C. W. Post campus, offers lecture notes on Earth's interior for a course on continental drift and plate tectonics.

J. Revenaugh, Department of Earth Sciences, University of California, Santa Cruz, provides lecture notes on Earth's interior for a course on geologic principles.

J. Louie, Nevada Seismological Laboratory, University of Nevada, Reno, offers illustrated lecture notes on the Earth's interior for a course on earthquakes and Earth structure.

J. Smyth, Department of Geological Sciences, University of Colorado, provides lecture notes for a physical geology course. A glossary is provided.

J. Butler, Department of Geosciences, University of Houston, provides lecture notes and other resources for a physical geology course. Presentations on Earth's interior and plate tectonics are included.

C. Lithgow-Bertelloni, Department of Geological Sciences, University of Michigan, offers lecture notes for a course on Earth as a dynamic system. A presentation on mantle convection and Earth's interior is included.

V. Cormier, Department of Geology and Geophysics, University of Connecticut, provides lecture notes for a geology course on Earth structure.

R D. Müller, Division of Geology and Geophysics, University of Sydney, Australia, offers lecture notes for a course on deep Earth structure and global tectonics

K. Regenauer-Lieb, Institute of Geophysics, Federal Institute of Technology, Zurich, provides lecture notes on the physics of Earth's mantle and core.

General Reports and Articles

Geochemistry, Geophysics, Geosystems, published by the American Geophysical Union and the Geochemical Society, is an online journal publishing interdisciplinary research in geophysics and geochemistry. An article by A. McNamara and P. van Keken titled “Cooling of the Earth: A parameterized convection study of whole versus layered models” was published 15 November 2000.

The U.S. National Report to International Union of Geodesy and Geophysics 1991–1994 includes a section of reviews on solid Earth dynamics.

The March-April 1995 issue of American Scientist had an article by M. Wysession titled “The inner workings of the Earth.”

The 16 June 2000 issue of Science was a special issue on Earth's dynamics. The Web supplement provides a selection of Science articles on the mantle and plate tectonics.

Numbered Hypernotes

1. The mantle and mantle convection. The Oxford Paperback Encyclopedia, provided by the xrefer Web site, has an entry for the mantle. The About Geology Web page provides a series of articles on the mantle. P. Allen and A. Densmore, Department of Geology, University of Dublin Trinity College, provides lecture notes on the mantle for a geology course. The Iowa State University glossary of geologic terms defines convection and convection cell. The geology section of Visualizing Earth provides an introduction to mantle convection. J. Schieber, Department of Geology, University of Texas at Arlington, offers an introduction to mantle convection in lecture notes on differentiation and plate tectonics for a course on Earth systems. S. King, Department of Earth and Atmospheric Sciences, Purdue University, offers a Mantle Convection Homepage. The Department of Earth, Atmospheric, and Planetary Sciences, Massachusetts Institute of Technology, makes available lecture notes by B. Hager for a course on mantle convection; information about models and a presentation of movies are included.

2. Plate tectonics. This Dynamic Earth: The Story of Plate Tectonics by W. J. Kious and R. Tilling is made available on the Web by the U.S. Geological Survey. Lecture notes on plate tectonics by B. van der Pluijm are included in the University of Michigan global change course. S. Nelson, Department of Geology, Tulane University, provides lecture notes on global tectonics for a physical geology course. J. Revenaugh provides lecture notes on plate tectonics for a course on geologic principles. H. Maher, Department of Geography and Geology, University of Nebraska, provides lecture notes for a course on plate tectonics. C. Lithgow-Bertelloni offers lecture notes for a course on plate tectonics. J. Braun, Research School of Earth Sciences, Australian National University, Canberra, provides lecture notes on tectonophysics.

3. The deep mantle. The 31 July 1998 issue of Science had an Enhanced Perspective by J. Tromp and A. Dziewonski titled “Two views of the deep mantle.” The 19 March 1999 issue had a report by S. Kaneshima and G. Helffrich titled “Dipping low-velocity layer in the mid-lower mantle: Evidence for geochemical heterogeneity,” a report by L. Kellogg, B. Hager, and R. van der Hilst titled “Compositional stratification in the deep mantle,” and a report by R. van der Hilst and H. Kárason titled “Compositional heterogeneity in the bottom 1000 kilometers of Earth's mantle: Toward a hybrid convection model.” The issue also had a News of the Week article by R. Kerr titled “A lava lamp model for the deep Earth.” The 20 March 1999 issue of Science News had an a article by R. Monastersky about these reports titled “A stirring tale from inside Earth.” The 4 May 2001 issue of Science had a News Focus article by R. Kerr titled “A lively or stagnant lowermost mantle?”

4. Seismology and Earth's interior. R. Phinney, Department of Geosciences, Princeton University, provides lecture notes on seismic waves in the Earth for a course on earthquakes, volcanoes, and other hazards. C. Ammon, Department of Earth and Atmospheric Sciences, St. Louis University, provides lecture notes on seismic waves and Earth's interior for a course on earthquakes. D. Schmitt, Geophysics Division, Department of Physics, University of Alberta, Canada, provides lecture notes on seismology and global structure for a geophysics course. Surfing the Internet for Earthquake Data is a collection of Internet seismology links maintained by S. Malone, Geophysics Program, University of Washington.

5. Slabs in the mantle. U. Christensen, Institut für Geophysik, Göttingen, Germany, provides an introduction to slabs. The U.S. National Report to IUGG, 1991-1994 included a chapter by S. Kirby titled “Intraslab earthquakes and phase changes in subducting lithosphere.” The 29 June 2001 issue of Science had a Perspective by H. Green titled “A graveyard for buoyant slabs?” about a report in that issue by W.-P. Chen and M. Brudzinski titled “Evidence for a large-scale remnant of subducted lithosphere beneath Fiji.” L. Wen, Department of Geosciences, State University of New York, Stony Brook, makes available a 1995 article by Wen and D. Anderson titled “The fate of slabs inferred from seismic tomography and 130 million years of subduction” (13) and a 1997 article by Wen and Anderson titled “Slabs, hotspots, cratons and mantle convection revealed from residual seismic tomography in the upper mantle” (20).

6. Geoid is defined in the Academic Press Dictionary of Science and Technology. E. Weisstein's Treasure Trove of Physics has an entry for geoid. S. Panasyuk, Department of Earth and Planetary Sciences, Harvard University, offers a presentation on the geoid. The Geoid Page of the U.S. National Geodetic Survey provides a definition of geoid, a note about the geophysics of the geoid, online papers, and Internet links.

7. Seismic tomography. Seismic tomography is defined in xrefer's Dictionary of Geography. H. Bijwaard offers an introduction to global seismic tomography. R. Faletic, Department of Physics, Australian National University, offers a presentation on seismic tomography. The 7 November 2000 issue of the Proceedings of the National Academy of Sciences had an article by T. Tanimoto and T. Lay titled “Mantle dynamics and seismic tomography.” A. Dziewonski, Department of Earth and Planetary Science, Harvard University, provides a history and overview of his research in seismic tomography. A chapter titled “Global seismic tomography of the mantle” by Dziewonski is included in the U.S. National Report to IUGG, 1991–1994. J. Lorenzo, Department of Geology and Geophysics, Louisiana State University, makes available a student project by J. Curry and E. Ferry on seismic tomography prepared for a course on reflection seismology. J. Ahern, School of Geology and Geophysics, University of Oklahoma, makes available a student project by Y. Chen on seismic tomography prepared for a course on solid earth geophysics. J. Louie offers a presentation titled “Velocity structure of the Earth” that discusses the discontinuities in the interior layers of Earth as determined by seismic tomography.

8. Mantle convection modeling and visualizations. The Harvard Seismology Group presents a Web page on 3D Earth structure. The Computational Geodynamics Group at Caltech provides mantle convection animations and a collection of Internet links. S. Zhong, Department of Physics, University of Colorado, offers an animated presentation titled “Thermal structure from models of mantle convection with surface plates, temperature-dependent and radially stratified viscosity” and a presentation on 3D convection models with plates. H. Cizkova, Department of Geophysics, Faculty of Mathematics and Physics, Charles University, Prague, provides animations of mantle convection models for a research project on thermal convection in the mantle with an impermeable boundary at a depth of 1000 km. D. Thorne, Computer Science Department, University of Kentucky, offers a slide presentation on modeling mantle convection. The Advanced Computing Laboratory at the Los Alamos National Laboratory makes available a conference presentation on mantle convection visualization on the Cray T3D. M. Sambridge, Research School of Earth Sciences, Australian National University, offers a presentation on a regionalized upper mantle seismic model of Earth. J. H. Davies, Department of Earth Sciences, Cardiff University, and Department of Earth Sciences, University of Liverpool, UK, offers a presentation on understanding mantle structure and dynamics with a section on mantle convection modeling and a section on comparison of circulation models with seismic tomography. The 3 April 1998 issue of Science had a report by H.-P. Bunge et al. tilted “Time scales and heterogeneous structure in geodynamic Earth models.” The Geodynamics Group, Department of Earth and Planetary Sciences, Harvard University, makes available a presentation by T. Becker and L. Boschi titled “A comparison of tomographic and geodynamic mantle models.”

9. Rayleigh number is defined in the Academic Press Dictionary of Science and Technology and in xrefer's Dictionary of Earth Sciences. E. Weisstein's Treasure Trove of Physics has an entry for the Rayleigh number. The Process Associates of America provides calculation tools for the Rayleigh number. An introduction to the Rayleigh number is provided in a slide presentation by M. Murphy titled “Secondary instabilities developed in the large-scale flow interacting with buoyancy in high Rayleigh number convection.”

10. P. Tackley is in Department of Earth and Space Sciences, University of California, Los Angeles; his home page provides presentations about his research and a selection of convection movies. The U.S. National Report to IUGG, 1991-1994 included a chapter by Tackley titled “Mantle dynamics: Influence of the transition zone.” The April-June 1998 issue of enVision, published by the National Partnership for Advanced Computational Infrastructure, had an an article about Tackley's research titled “As the world churns: Modeling convection in the earth's mantle.”

11. Anisotropy and anisotropic are defined in xrefer's Dictionary of Earth Sciences. E. Matzel, Department of Geological Sciences, University of Texas, Austin, includes a chapter on anisotropy in his Master's thesis titled “Evidence for anisotropy in the D” layer beneath Alaska.” H-R. Wenk, Department of Geology and Geophysics, University of California, Berkeley CA provides an introduction to the subject of anisotropic convection. A research overview titled “Seismic anisotropy and mantle flow in subduction zones” is provided by C. Hall, Seismology and Geophysics Group, Brown University.

12. Bullen and Jeffreys. A Dictionary of Scientists, available from xrefer, includes entries for Keith Edward Bullen and Sir Harold Jeffreys. The Australian Academy of Science makes available a biographical memoir by A. Hales about Bullen. The MacTutor History of Mathematics archive includes a biography of Jeffreys.

13. The relation of mantle convection to plate tectonics. The 16 June 2000 issue of Science had a review article by P. Tackley “Mantle convection and plate tectonics: Toward an integrated physical and chemical theory.” For a course on plate tectonics, H. Maher provides lecture notes on the relationship between convection currents and plate motions. J. Louie offers lecture notes on driving mechanisms of plate tectonics for a course on structure, tectonics, and earth physics. Van Der Hilst's Research Group, Massachusetts Institute of Technology, makes available a lecture by S. Zhong on mantle convection and plate tectonics from a seminar on mantle convection.

14. A biography of Lord Rayleigh (John William Strutt) is provided in the MacTutor History of Mathematics archive. The 1904 Nobel Prize in Physics was award to Lord Rayleigh.

15. Rayleigh-Bénard convection. A Dictionary of Science, available from xrefer, defines Bénard cells. M. Cross, Condensed Matter Physics group, California Institute of Technology, defines Rayleigh Bénard convection. E. Weisstein's Treasure Trove of Physics includes an entry for thermal convection (Rayleigh-Bénard convection). The Surko Research Group, Institute for Nonlinear Science, University of California, San Diego, offers a presentation on Rayleigh-Bénard convection. Self-Organization in Biological Systems, an online book by S. Camazine et al., includes a presentation on Bénard convection. A. Narasimhan, Mechanical Engineering Department Southern Methodist University, offers a presentation on Rayleigh-Bénard convection. The August 2001 issue of Physics Today had an article by L. Kadanoff titled “Turbulent heat flow: Structures and scaling” that discusses Rayleigh-Bénard systems.

16. Information about a tribute to J. R. A. Pearson upon his retirement is provided by the Dynamics of Complex Fluids program Web page. In 1986 Pearson received the Gold Medal of the British Society of Rheology.

17. Self-organizing systems. Self-organization and dissipative structure are defined in xrefer's Dictionary of Science. A FAQ on Self-Organizing Systems is provided on the CALResCo Web site. E. Decker, Department of Biology, University of New Mexico, Albuquerque, provides a tutorial on self-organizing systems.

18. D. L. Anderson is at the Seismological Laboratory, Division of Geological and Planetary Sciences, California Institute of Technology.

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article