Product has been added to the basket

If not plumes, what?

Prof. James Natland (Rosenstiel School of Marine and Atmospheric Science, University of Miami) answers the charge that antiplumers have no coherent counter-theory.

The point/counterpoint between Andy Saunders and others (1, 3) and Gill Foulger (2) suggests that a summary statement of difficulties with the plume hypothesis and indications of a likely alternative should be attempted. Saunders and others (3) deplore what they perceive as the "aggressive nihilistic attacking" of the plume hypothesis, which they assert is supported by "few positive alternatives." 

The situation is not quite this grim; there is definitely a crescendo of doubt and confusion about plumes that even Saunders and colleagues are finding difficult to ignore. The question now is whether a general framework for an alternative, which must always be preceded by expressions of doubt and confusion, can be perceived. I believe it can. 

Bear in mind that prior to plate tectonics there was no general theory to account for all manifestations of volcanism on the Earth's surface. Plate tectonics offered a clear explanation for the distribution of volcanism at spreading ridges and island arcs. What we are considering now is whether mid-plate volcanism, which more or less is "everything else", can be subsumed under a single unifying theory of planetary volcanism, or whether two independent mechanisms are required. 

Plume theorists hold that the most important cause of mid-plate volcanism is unrelated to modern plate motions, coming instead from deeper convective disturbances within the Earth that are fueled by heat from the core. The lead alternative is whether much of mid-plate volcanism, but including all linear-chain volcanism, is instead, like that of ridges and arcs, directly related to the plate-tectonic cycle in the shallow mantle. The latter is neither an ill-formulated, philosophically unjustified, obscure, nor even an altogether recent idea. Aspects of fracture-controlled explanations for Pacific volcanism were advocated by James Dwight Dana as early as 1849 (4). However, the matter is very complex, and a full explanation embraces many disciplines. 

Much of the following discussion is based on collections of articles, tutorials, references and abstracts posted on Opening of the web site, which is maintained by Gill Foulger, followed an extensive e-mail correspondence among Don Anderson, Warren Hamilton, Gillian Foulger, Dean Presnall, Jerry Winterer, Mike O'Hara, Seth Stein, Carol Stein, Anders Meibom, me and others. To save space and give a sense of the flavour and immediacy of that correspondence, I use names as pointers rather than complete references and the reader can consult the website and references therein for fuller treatments. Many of these ideas were also presented and discussed at a recent Penrose Conference, Plume IV, Beyond the Plume Hypothesis. Most pertinent to the issues raised by Saunders are contributions by those named above, Hetu Sheth, Bruce Julian, Peter Vogt, Henry Dick, David Green, Marjorie Wilson, Adam Dziewonski, Enrico Bonatti, Alan Smith, Scott King and Anders Meibom. Conference abstracts are also posted on the web site. 

Thus hot spots are not hot by measurement of heat flow (Stein, following 5), or by dint of petrology. High-temperature estimates based on MgO geothermometry using glass and olivine compositions for Proterozoic and younger komatiites as well as modern picrites are model-dependent (e.g., they assume a single homogeneous mantle source, or a single liquid line of descent for all olivine in a rock). They often do not take into account profound alteration that can increase the MgO content of many of those rocks (Natland). Some komatiites also have arc-like (boninitiic) geochemical affinities. The most primitive picritic Hawaiian tholeiites erupt at temperatures only 100-150o higher than primitive abyssal tholeiites (Presnall, 6). Because of the presence of water and CO2, the mantle solidus is much lower than previously thought using dry-mantle experimental analogs (Presnall, 6). Therefore, incipient melting is not restricted to the hearts of long narrow conduits beneath major volcanoes, but occurs widely, perhaps even globally, within the low velocity region beneath the lithosphere. 

After 30 years of increasingly sophisticated development, seismic tomography has not yet seen evidence linking anything in the lower mantle to an active volcanic region on the surface of the Earth, even beneath huge magmatic provinces like Iceland or Yellowstone (Foulger). Slow regions deep in the mantle are typically coloured red, suggesting warm rock, but could as easily represent concentrations of water, CO2 or other lithologic variability (Anderson, Foulger, Presnall). The statistical significance of anomalies observed in some large portions of the mantle and their continuity to the surface are uncertain (Julian). Some fraction of the low-velocity zone in the lower mantle will correlate with hotspots even if they are unrelated (Anderson). The most resolvable features of plumes (lithospheric and transition-zone thinning, plume heads, spreading out of hot material below the plate and at 650 km) are absent in seismological models (Julian, Foulger). Carefully selected and cropped cross-sections, vertical exaggeration and saturated colour scales contribute to the confusion (Hamilton, Anderson). The vertical travel time of seismic waves beneath Hawaii is the same as the global average (Julian). 

The so-called ocean-island basalt (OIB) geochemical signature is no proof of any particular deep geometry to the mantle source (Natland); instead the general petrogenesis of the lavas indicates shallow sources that may alternatively be arranged in laterally and vertically heterogeneous layers beneath the lithosphere (Natland) rather than long vertical pipes. OIB geochemistry may not even represent mantle heterogeneity, coming about instead by processes of wall-rock reaction, assimilation and mixing in magma chambers and conduits (7, 8, O'Hara). Alternatively, it represents different ways the same mantle is sampled by ridges as opposed to mid-plate volcanoes (Anderson, Meibom). 

OIB itself is a misnomer, since tens of thousands of seamounts that are neither islands nor parts of linear volcanic chains are made of the same sorts of basalts (Natland). E-(enriched) MORB is alkali olivine basalt with OIB geochemistry on spreading ridges. Bulk lithologic heterogeneity of the mantle source (varieties of peridotite, garnet pyroxenite, eclogite) is important in producing the so-called OIB signature as well as variations in depleted MORB; e.g., Na8 etc. (Natland, 9). Such heterogeneity also reduces the global range of potential temperatures required for mantle melting. Large-scale lithologic heterogeneity involving ocean crust and adjacent melt-depleted abyssal peridotite is re-introduced to the shallow mantle at subduction zones as fast as it is created at spreading ridges. This heterogeneity is too large and profound for it ever to re-integrate with adjacent mantle and produce a homogeneous source material for later volcanism. A shallow layer beneath the lithosphere of somewhat molten, enriched, volatile-rich mantle material is a better explanation than plumes for the very widespread distribution of seamounts, small and large, old and young, and very long volcanic ridges that occur on major portions of the Pacific and other plates (Natland, Winterer). 

Although high 3He/4He indicates something about ages of mantle components, it tells us nothing about their original depth or distribution (Meibom). A high ratio merely means isolation of helium, captured with other volatiles in fluid inclusions in minerals, from U and Th, which are retained in melts that are generally excluded from such inclusions (10). Thus 4He is low and remains invariant in old peridotitic or cumulate gabbroic/ultramafic protoliths (Anderson). We can expect such material to be present in the upper mantle (Meibom) and the mantle beneath ancient continental crust that, in the lee of rifting, often winds up in the middle of ocean basins (Jan Mayen, Seychelles, Elan Bank at Kerguelen, possibly Iceland) and becomes involved in partial melting beneath ridges and at the active ends of linear volcanic chains. Ancient peridotitic slabs may also be present in the shallow mantle. 

Parallel traces to island chains are a consequence of stress fields acting across plates (Anderson, Hamilton, Winterer, Natland). There are clear examples of non-hot spot ridges (Winterer) and island chains directly influenced by strong stress fields (Natland); these also have "OIB" geochemistry. In the long history of the Pacific plate, stress differences across the plate have become stronger and more consistently aligned as a result of growth of the plate and the increasing length and geometry of subduction zones at its northern and western extremities. This forced a shift from widely scattered Mesozoic seamounts that have no age progressions in the far western and central Pacific to parallel linear chains that are strongly age-progressive since about 50 Ma (Natland, Winterer). Subduction is the driving engine not just of plate motions, but it largely controls the intra-plate stress field. 

Subducting plates don't descend (7) or rarely descend (Anderson) below the 660km discontinuity. Nothing can rise out of the lower mantle (Bullen's region D) because of its strong density stratification (Anderson), thus geochemical recycling of EMI, EMII, HiMu, etc., is restricted to the upper mantle. Although estimates of the meaning of that alphabet-soup consistently invoke involvement of material from the Earth's surface (mature sediment, pelagic clay, altered ocean crust), many geochemists take for granted the necessity of driving these materials all the way down to near the core-mantle boundary (despite buoyancy problems and low solidus temperatures), picking up some other signatures, and then resurfacing, where the densities and ease of melting - ignored to this point - now assist in the model. 

Alternatively, none of this is necessary and instead arc-continent and continent-continent collisions trap blocks of oceanic lithosphere in sutures produced in the past Philippine Seas and continental collisions of this world that later are exposed when continents rift and ocean basins open. The trapped slabs contribute to heterogeneities and large magma volumes seen in places like Iceland (Foulger, Anderson, Natland). Which choice, a full-circuit down to the core with a slab and up with a deep mantle plume, passing twice through the full thickness of the mantle, or tapping of old shallow sutures, is the more fantastic? 

Eclogite should not be considered a monolithic lithology during partial melting (Natland). Incipient melting of subducted ocean crust in the form of eclogite produces rhyodacite (11), which frequently erupts on Iceland, although there it may also have other origins. More extended partial melting of eclogite (12) produces ferrobasalt, which is quite abundant on Iceland, then olivine tholeiites, but not picritic glass. The bulk composition of abyssal gabbro obtained in long drilled sections closely approximates that of primitive Icelandic tholeiite, which is non-picritic, and which also has the trace-element characteristics of abyssal olivine gabbro cumulates, not MORB liquid (13,14). Ocean crust itself is not merely simple basalt, but includes an extremely diverse gabbroic layer that already contains a small percentage of tonalite/trondhjemite veins and a much greater percentage of oxide gabbro containing abundant magmatic ilmenite and magnetite plus a small percentage of titanian magmatic amphibole (15, 16). The ilmenite transforms largely to rutile in the eclogite facies, a mineral that is abundant in many Alpine eclogites (17). No melting experiments have been carried out using rutile eclogite as a starting material. 

Nevertheless, temperatures of crystallization of abyssal gabbro span from about 1200-700oC (Natland), suggesting that partial melting of the same assemblage transformed to eclogite at any given pressure will begin at quite low temperature, and may be nearly complete at temperatures required for formation of basaltic melt from peridotite in the same melting domain in the mantle. The source of heat for large-scale eclogite melting will be the huge volume of warm mantle enveloping a subducting slab (Anderson). Subducting slabs in narrow closing ocean basins and backarc basins are much warmer than those at the subduction margins of old, huge plates, and do not require much reheating to become neutrally buoyant and even partially molten in the shallow mantle. Most of them will not sink into the lower mantle, and their readily fused basaltic crust will add to the fertility of the upper mantle. If eclogitic portions of slabs in a melting domain are part of a steeply dipping suture, this geometry will add to melt productivity and crustal thickness above them (18). 

The plume model has become so flexible that it cannot be tested. Thus plumes may either have a plume head or not; the plume head can be hot - and produce hot (?) komatiite - or not; plume tails can blow in the mantle wind or not; they can pipe material laterally for extremely long distances, or not; they can come from deep mantle, or not; the plume tails themselves can be hot, or not; some plumes provide enriched mantle to melt sources, others depleted material, still others both enriched and depleted material. Some scientists will go to extraordinary lengths to rationalise plumes. Thus we read a recent proposal that seamounts near the East Pacific Rise are actually sensing enriched material injected in the shallow mantle very remotely at Hawaii (19), five thousand kilometres away. Thus a pipe-like mass, suggestively coloured red to indicate warmth (even though it could also represent chemical heterogeneity), observed in upper-mantle tomography between Samoa and Tahiti presumably supplies both (by splitting of the top of the plume?). This may look reasonable on a global-scale projection reduced to column width in a journal article, but the projected intersection of the mass on the Earth's surface is actually thousands of kilometres from either island group. 

This much, so greatly telescoped, should give readers an idea of the extent of our concerns and conjectures, and where the terms of future discussion can best be framed. The growing alternative to plumes is that shallow stresses are transmitted across plates from subduction boundaries. These and stresses induced by regional and local loading serve to produce patterns of fracture within both continental and ocean crust. Within the continents, major physical breaks and sutures in deep lithospheric structure also influence the distribution of volcanism (e.g., Yellowstone, 20). 

The transition between the lithosphere and asthenosphere corresponds to the distribution of a widespread small melt fraction in the shallow asthenosphere that contains, distilled within it, the essence of isotopic heterogeneity observed at point volcanic sources, line volcanic sources, and widely distributed volcanic sources, all of which must be explained in any general theory. The lithosphere-asthenosphere transition dips generally toward trenches, but is itself uneven, as a result of transform offsets, asthenospheric heterogeneity (different fertilities of material in the asthenosphere), shallow asthenospheric convection, and melt ponding. Mid-plate magmatism results from stress-induced lithospheric fracturing from above and magma overpressure and buoyancy from below. Parallel volcanic alignments result from consistently oriented stress fields acting across plates and local concentrations of the magnitude of stresses. Age progressions result from propagation of those fractures in directions of least principal stress at rates at, or faster than, the rate of progression of plates over the asthenosphere. Concatenations of stress pattern, underlying asthenospheric fertility (related to ancient subduction cycles and continental suturing), shallow asthenospheric convection, and melt ponding dictate locations of both transient LIPs and large, long-lived linear chains like Hawaii. Changes in patterns of fracture orientation such as the Hawaiian-Emperor bend result from changes in stress patterns acting across plates that occur especially when subduction geometries change as major spreading ridges are consumed at trenches along migrating triple junctions (e.g., Kula Ridge at about 50 Ma in the north Pacific), or continents collide. 

Saunders et al. (3) dismiss the adequacy of fracture and plate mechanisms in controlling linear-chain volcanism at places like Hawaii, but apart from expressing astonishment that this could possibly be considered instead of plumes, they offer no argument against it. Geophysicists have tended to develop plume models to the near exclusion of models of how far-field stresses might induce parallel lithospheric fracturing over distances of thousands of kilometers on individual plates. The absence of such theory, however, places fracture theory at about the level of plume theory when it was first proposed by Jason Morgan (21, 22), and before it was augmented by injection experiments (23). However, as Anderson (24, this issue) reasons, there is serious reason to question the assumptions that underlie such models and experiments. A partially molten and variably fertile asthenosphere removes most of the basis for inferring a deep heat source. 


1) A.D. Saunders, 'Mantle plues: an alternative to the alternative', Geoscientist, 2003. h
2) G. Foulger, 'Making the evidence fit the plume, Geoscientist, 2003.
3) A.D. Saunders, G. Fitton, and L. Coogan, "The Emperor strikes back', Geoscientist, 2003.
4) J.D. Dana, 'Geology', In C. Wilkes, United States Exploring Expedition, 10: Philadelphia (C. Sherman), with Atlas: New York (Putnam), 1849.
5) R.P. Von Herzen., M.J. Cordery, R.S. Detrick, and C. Fang, 'Heat-flow and the thermal origin of hot spot swells - the Hawaiian swell revisited', J. Geophys. Res., 94, 13,783-13,799, 1989.
6) D.H., Green, T. J. Falloon, S.M. Eggins and G.M. Yaxley, 'Primary magmas and mantle temperatures', European J. Min. 13: 437-451, 1999.
7) W.G. Hamilton, 'The closed upper-mantle circulation of plate tectonics', In S. Stein, and J.T. Freymueller (Eds.), Plate Boundary Zones, AGU Geodyn Ser. 30: Washington (Amer. Geophys. Union), 359-410, 2002.
8) M.J. O'Hara, and R.E.Mathews, 'Geochemical evolution in an advancing, periodically replenished, periodically tapped, continuously fractionated magma chamber', J. Geol. Soc., 138: 237-277, 1981.
9) J.H. Natland, 'Partial melting of a lithologically heterogeneous mantle: Inferences from crystallization histories of magnesian abyssal tholeiites from the Siqueiros Fracture Zone', In A.D. Saunders and M. Norry (Editors), Geol. Soc. London, Spec. Publ. Geological Society of London, pp. 41-77, 1989.
10) J.H. Natland, 'Capture of mantle helium by growing olivine phenocrysts in picritic basalts from the Juan Fernandez Islands, SE Pacific', J. Petrol. 44, 421-456, 2003.
11) A. Yasuda, T. Fujii, and K. Murita, 'Melting phase relations of an anhydrous mid-ocean ridge basalt from 3 to 20 GPa: Implications for the behavior of subducted oceanic crust in the mantle', J. Geophys. Res. 99:9401-9414, 1994.
12) K. Ito, and G.C. Kennedy, 'The composition of liquids formed by partial melting of eclogites at high temperatures and pressures', J. Geol., 82: 383-392, 1974.
13) K. Breddam, 'Kistufell: Primitive melt from the Iceland mantle plume', J. Pet., 43: 345-373, 2002.
14) C. Chauvel, and C. Hemond, 'Melting of a complete section of recycled oceanic crust: Trace element and Pb isotopic evidence from Iceland', Geochem. Geophys. Geosys., 1: 1999GC000002, 2000.
15) H.J.B. Dick, J.H.Natland, J.C. Alt, W. Bach, D. Bideau,, J.S. Gee, S. Haggas, JG.H.. Hertogen, G. Hirth, P.M. Holm, B. Ildefonse, G.J. Iturrino, B.E. John, D.S. Kelley, E. Kikawa, A. Kingdon, P.J. LeRoux, J. Maeda, P.S. Meyer, D.J.Miller, H.R. Naslund, Y.-L. Niu, P.T. Robinson, J. Snow, R.A. Stephen, P.W. Trimby, H.-U. Worm, and A. Yoshinobu, 'A long in situ section of the lower ocean crust: Results of ODP Leg 176 drilling at the Southwest Indian Ridge', Earth Planet. Sci. Lett. 179, 31-51, 2000.
16) J.H. Natland, and H.J.B. Dick, Formation of the lower ocean crust and the crystallization of gabbroic cumulates at a very slowly spreading ridge. Jour. Volcanol. Geotherm. Res. 110: 191-233, 2001.
17) W.G. Ernst, 'Mineral chemistry of eclogites and related rocks from the Voltri Group, western Liburia, Italy', Schweiz. Mineral. Petrogr. Mitt. 56: 293-343, 1976.
18) G.R. Foulger, D.L. Anderson and J.H. Natland, 'Iceland is fertile: The melt anomaly results from extensive melting of subducted crust in the Caledonian suture, not from a plume' submitted to J. Volcan. Geotherm. Res., 2003.
19) Y.-L. Niu, M.Regelous, I. Wendt, R. Batiza, and M.J. O'Hara, 'Geochemistry of near-EPR seamounts: importance of source vs. process and the origin of enriched mantle component', Earth Planet. Sci. Lett., 199: 327-345, 2002.
20) R.L. Christiansen, G.R. Foulger, and J.R. Evans, 'Upper mantle origin of the Yellowstone hotspot', Bull. Geol. Soc. Am., 114, 1245-1256, 2002.
21) W.J. Morgan, 'Convection plumes in the lower mantle' Nature, 230: 42-43, 1971.
22) W.J. Morgan, ' Plate motions and deep mantle convection', In R. Shagam, R.B. Hargraves, W.J. Morgan, F.B. Van Houten, C.A. Burk, H.D. Holland, and L.C. Hollister (eds.), Studies in Earth and Space Sciences, Geol. Soc. Am. Mem., 132: 7-22, 1972.
23) R.W. Griffiths, and I.H. Campbell, ' Interaction of mantle plume heads with the Earth's surface, and onset of small-scale convection. Journal of Geophysical Research 96, 18295-18310, 1991.
24) D.L. Anderson, 'The smoking gun' Geoscientist, 2003.
Prof. James H Natland (Rosenstiel School of Marine and Atmospheric Science, University of Miami, Miami, FL 33149 USA email: