Product has been added to the basket

The emperor strikes back

Feathers fly in the plume debate as the allegedly naked theory takes to the catwalk - 22 August 2003

Dr. Gill Foulger1 has mounted an eloquent and vigorous riposte against Andy Saunders's article in support of plumes in Geoscientist2. Indeed, Dr. Foulger's article could be read as

a source of innocent merriment.3
We had wished to avoid a tit-for-tat or Oh-yes-it-is, Oh-no-it-isn't type of dialogue, but it is necessary to address in more detail some of the individual points raised by Dr Foulger. Unfortunately, the Devil is the detail, and we'd like to thank Dr Foulger for prompting us to expand on some of the points that have been raised.

Theories, or models, are often multi-facetted, and to dismiss one part does not indicate a fatal flaw in the whole enterprise. Examples of the ophiolite and plate tectonic models were cited2 simply in order to illustrate this, and not to use them as direct evidence that the plume model is (or is not) valid. We don't fully understand plate tectonics, but most people accept the model. This doesn't stop us refining and improving the model, and seeking the scientific truth - which is out there, somewhere.

Let us revisit to those fair places, Hawaii and Iceland, to enjoy fire and ice.


None of the observations listed in (2) were 'modified' to fit the plume model. The average output rate of magma along the Hawaiian-Emperor chain is indeed stated by Clague4 to be 0.013 km3a-1. The analogy with magma production rates at a mid-ocean ridge is misleading, however; the Hawaiian and Emperor magmas have to be produced beneath 100 km thick lithosphere, which inhibits decompression melting. Therefore, to produce a 'mere' 0.013 (average rate) or 0.13 km3a-1 (current rates) of Hawaiian magma is difficult, and requires high mantle temperatures5. Without these high source temperatures, it is difficult to see how the basalt could be produced.

Comparison with the South Pacific Superswell is very misleading, as the wavelength is much longer than the Hawaiian swell6. The Hawaiian Swell is both focused and transient, as the Pacific plate crosses the upwelling (plume) in the underlying mantle. The South Pacific Superswell is a much broader structure, and its origins are unclear.

There are many examples of time-progressive volcanism where there is no suggestion of a plume; the ocean ridges, for one. Absence of a process in one setting does not preclude it from another, though! And it is difficult to see how the prolonged time progression along the Emperor-Hawaii chain (>5000 km and >90 Ma, and with a bend!) can be produced purely by, for example, fracture propagation7 or magmatism in transform faults8. Lithospheric structure can, of course, play a role in controlling the location of magma chambers, conduits, etc.

Absence of a thermal anomaly around Hawaii. The issue here is that there is no, or only a minor, thermal anomaly on the Hawaiian swell that is attributable to lithosphere heating by the proposed plume9. (Clearly, any thermal model must accommodate the fact that there is a lot of magma being erupted onto and into the Pacific plate.) Given that the lithosphere beneath Hawaii is about 100 km thick, and that the thermal time constant for such lithosphere is about 30 Ma (assuming a thermal diffusivity of 10-6 m2s-1), this means that any conductive thermal pulse coming solely from the plume will not reach the surface for several millions of years, by which time the plate will have moved 100s of kilometres downstream. This may be part of the reason why no anomaly is seen on the Hawaiian Swell, even far to the WNW of the active islands9. An additional explanation is that the swell has a strong dynamic component to its uplift, and it is not produced solely by basal heating (and expansion) of the lithosphere. This, in turn, reduces the requirement for the amount of heat conductively transferred into the lithosphere9.

The objection that the Hawaiian system doesn't exhibit a plume head is an interesting one. Most likely, it was expressed as an ocean plateau and subducted, but this may not be provable, unless pieces were also obducted and we can study them, or seismic observation can 'see' the plateau on its route through the mantle. (It is probably not true that all ocean plateaus cannot subduct10. It depends on their buoyancy, and unless this is known, their fate is difficult to predict. But an old, cold plateau could subduct.) Unfortunately, we don't know the age of the Hawaiian plume, as the oldest surface products are likely to have been subducted at the Kurile Trench. Alternatively, perhaps not all plumes begin life with large heads.
Hawaii To dismiss deep-mantle seismic velocity anomalies as 'discredited' requires a statement to say where they are discredited, and by whom. Otherwise it's hearsay. The big challenge for (or to?) the plume model (in Hawaii and elsewhere) will be when there is sufficient resolution in seismic tomography techniques (or similar) to image narrow vertical anomalies in the lower mantle11. Either they are there and we've yet to find them, or they simply aren't there. If this is not a testable hypothesis, then we don't know what is!

Geochemical arguments were not discussed in (2) because of lack of space; to develop these arguments fully in 1200 words is difficult. Geochemistry alone does not 'prove' the existence of plumes, but it can help link mantle sources to Earth models. Geochemistry also allows us to predict depths and extents of melting, and source temperatures, and these help to corroborate the plume model. As pointed out by Stuart Weinstein12, it is very difficult to see how a Hawaiian basalt could be generated (e.g., via crack propagation models) from the uppermost asthenosphere, which has very different isotopic compositions (i.e., MORB mantle source). Rare gas13 and radiogenic isotope14 geochemistry of Hawaiian (and Icelandic) basalts tells us that their sources are NOT the same as for MORB, and by implication that these melts are unlikely to derived from a shallow sub-lithospheric mantle reservoir. Elevated 186Os/188Os ratios in Hawaiian picrites suggests that their source interacted with the Earth's core at some time15.

What other models can so explain Hawaii and the seamounts, and explain the Big Bend, the Hawaiian swell, and account for the compositions of Hawaiian basalts? The composition of Hawaiian and Emperor lavas is not MORB-like, so it is difficult to see how they come from the shallow mantle alone. The challenge is to produce a viable alternative model to account for Hawaii. Clearly, if people can disprove that a plume exists beneath Hawaii, then the plume model is weakened, but at the present time the evidence is best supported by the plume model. If there is an alternative, please let us hear it; we are open to alternative ideas. Many people await a clear exposition of the crack propagation model (especially one that can accommodate the Big Bend), and that they can then test.


It is very difficult to produce the thick Icelandic crust without recourse to higher temperatures in the source region16. We agree that the seismic velocity anomaly beneath Iceland17 does not require higher temperatures, but given the volume of basalt, it seems tortuous logic to assume that the temperature is the same as in the surrounding mantle.

Volatiles may play a role during melting, but the amount of volatiles in the Icelandic mantle is very unlikely to be sufficient18 to enable us to reduce the temperature requirement of the source to ambient. In any case, the volatile species are likely to be lost to the first melt fractions and removed from the system, which is then left anhydrous.

'More fusible material' in the mantle source regions might be an explanation for some of the excess magmatism in Iceland (and LIPs in general)19, but cannot account for the whole of the 30+ km of the Icelandic crust. It has an inherent problem in explaining the compositions of Icelandic basalts. If we make the reasonable assumption that the mantle source is a mixture of peridotite and more fusible eclogite, then the incompatible element content of this mixture is likely to be greater than the peridotite alone. Melting of this hybrid source rock will initially generate primary melts with high contents of incompatible elements - too high to form most Icelandic tholeiites - and consequently it is necessary to melt substantial amounts of peridotite to dilute the trace element budget. This requires additional heat energy. Note that we do not rule out eclogite being present in the Icelandic mantle source, as it can help explain the isotopic variability of Icelandic basalts, but it alone (i.e., without additional thermal energy) does not provide a ready explanation for the large volumes of magma.

Picrites? Glasses (=liquids) with nearly 12% MgO have been recovered from Iceland, but they are rare20. According to the recent IUGS classification21, these are Mg-rich basalts, but not quite picritic melts. Andy Saunders2 did not state that such liquids are proof of high temperatures in the source regions (although this would help!), because similar MgO levels have been found in some MORB glasses22. The point here is that it is very difficult to produce high-MgO melts from a purely basaltic source, such as remelting of subducted Iapetus crust. Complete melting of basalt produces basalt; partial melting of such basalt produces a more silicic melt, depending on the amount and conditions of melting. Readers who are still with us may be interested to note that substantial volumes of picritic melts were produced during the initial phases of the North Atlantic Igneous Province, in Baffin Island, West Greenland, SE Greenland, and Scotland (e.g., Rum)23, and these do testify to anomalously high source temperatures. Was the source hotter then, or is it a sampling issue now? Dense picritic magmas could have difficulty erupting through magma chambers24 in the Icelandic rift zones, and fail to reach the surface. Good scientific logic, but we risk being falsely accused of deus ex machina.

Large Igneous Provinces

Despite what Gill Foulger1 (and Hetu Sheth25) state, most LIPs do indeed have a time-progressive volcanic trace leading from the LIP26:
  • Deccan-Chagos/Laccadive,Reunion Island hotspot
  • Kerguelen/Rajmahal/Bunbury, Ninetyeast Ridge, Isles Kerguelen/Heard Island hotspot
  • Madagascar, Madgascar Rise, Marion Island hotspot
  • Paraná/Etendeka, Rio Grande Rise/Walvis Ridge, Tristan da Cuhna hotspot
  • Columbia River Basalts, Snake River Plain, Yellowstone hotspot
  • North Atlantic Igneous Province, Greenland-Faroe Ridge, Iceland hotspot
  • Ethiopian Flood Basalts, Afar and/or East African hotspot
Older provinces are more difficult to relate to extant hotspots, because of plate (and probably plume) movements. The plume may also have dissipated in the intervening years.
  • Karoo/Ferrar, uncertain
  • Siberian Traps, uncertain
  • Ontong Java, Louisville hotspot?
The Siberian Traps are an unfortunate example, because we know so little about the high-Arctic region, but there is a lot of Mesozoic basalt on the Arctic Shelf which may well represent the successor to the Siberian plume. The uplift associated with the Siberian plume was probably mostly in the adjacent West Siberian Basin, not on the Siberian craton.

The plume model does account for the short burst of magmatism found in most LIPs. It may be a function of sudden increased mantle flux rates (a plume start-up 'head', perhaps27?), or decompression melting of hot mantle due to rapid extension above a developed plume26, and with or without an eclogite component19. The plume model may not be adequately understood, but it is the one that is most at ease with the observations.

'LIPs can be explained equally well by other models'. Sure, we don't dispute this. But plumes do tend to, let's say, have the edge on other models. Again, let's hear these other models, explained on these pages. We know of no other model that can explain the formation of LIPs so elegantly as the plume model.

The Big Game being played out here is really between the Top-Downers who advocate that the lithospheric plates can explain most of the features we see at the Earth's surface, versus those who believe that the deep Earth has an important role to play. We find it difficult to envisage that convective systems are not strongly influenced by the strong temperature contrasts at the major thermal boundary layers deep in the Earth, especially at the core-mantle boundary (where the increase in temperature may be in excess of 1000C) and at the 670 km discontinuity (depending on whether mantle convection is layered (up to 1000C) or not (no temperature difference))28.

We remain curious about the angst that plume models have generated in the community. To say that the articles and letters that have been generated in Geoscientist are vehement is an understatement. We hope that in this reply (and in 2) we?ve managed not only to defend the plume model, but also to play the ball into the anti-plume court. We have no vested interest in defending the plume model to the bitter end. Who wants to support a failed model? But so far the anti-plume groups have not convinced us that their models are valid. We hear plenty of aggressive nihilistic attacking of the plume model, but few positive alternatives. The plume model has stood the test of time for almost as long as the plate tectonic model, and frankly we don't see it collapsing now.

The original challenge2 to the anti-plume group remains, but unanswered.


1. G. Foulger, 'Making the evidence fit the plume', Geoscientist, 2003.

2. A.D. Saunders, 'Mantle plumes: an alternative to the alternative', Geoscientist, 2003.

3. Gilbert, W.S. The Mikado.

4. D.A. Clague & G.B. Dalrymple, U.S.G.S, Prof. Paper 1350, Table 1.6, 1987.

5. E.g., D. McKenzie & M.J. Bickle, J. Petrol. 29: 625-679, 1988; S. Watson & D. McKenzie, J. Petrol. 32: 501-537, 1991.

6. M.K. McNutt & K. M. Fischer, In: B. H. Keating, P. Fryer, R. Batiza & G. W. Boehlert (eds) Seamounts, Islands, and Atolls, American Geophysical Union Geophysical Monograph, 43: 25-34, 1987.

7. E.g., D.H. Green, Phil. Trans. Roy. Soc. Lond., 268: 707-725, 1971.

8. B.H. Erikson et al., Nature, 225, 53-54, 1970; D. Handschumacher, Nature, 244, 150-152, 1973.

9. R.P. Von Herzen et al., J. Geophys. Res., 94: 13,783-13,799, 1989.

10. M. Cloos, Geol. Soc. Amer. Bull. 105: 715-737, 1993; A.D. Saunders et al., Lithos 37: 81-95, 1996;

11. J. Ritsema & R.M. Allen. Earth Planet. Sci. Letts., 207, 1-12, 2003.

12. S. Weinstein, 'Stop that Whining', Letter to Geoscientist, 2003.

13. E.g., M.D. Kurz et al., Nature 297: 43-46, 1982; D.W. Graham et al. Earth Planet. Sci. Letts., 160: 241-255, 1998; K. Breddam et al., Earth Planet. Sci. Letts., 176: 45-55, 2000.

14. E.g., W.M. White & A. W. Hofman, Nature 296: 821-825, 1982; A. Zindler &S. R. Hart, Ann. Rev. Earth Planet. Sci. 14: 493-571, 1986; J. Blichert-Toft et al., Science 285: 879-882, 1999.

15. A.D. Brandon, et al. Earth Planet. Sci. Letts., 174, 25-42, 1999.

16.There is considerable debate about the temperature of the present-day Iceland plume; see, for example, J.-G. Schilling, Nature, 352: 397-403, 1991; N.M. Ribe et al., Earth Planet. Sci. Letts., 134: 155-168, 1995; R.S. White et al., J. Geol. Soc. Lond. 152: 1039-1045,1995.

17. C. Wolfe et al., Nature 385: 285-247, 1997; G.R. Foulger et al., Geophys. J. Int., 142, F1-F5, 2000.

18. A.R.L. Nichols et al. Earth Planet. Sci. Letts., 202: 77-87, 2002.

19. M.J. Cordery et al., J. Geophys. Res. 102: 20179-20197, 1997.

20.T.H. Hansteen, Contribs. Mineral. Petrol. 109: 225-239, 1991; S. Révillon et al., Lithos 49: 1-21, 1999.

21. M.J. Le Bas, J. Petrol., 41: 1467-1470, 2000.

22. E.g., E.M. Klein & C. H. Langmuir, J. Geophys. Res. 92: 8089-8115, 1987, and references therein.

23. E.g., D.B. Clarke & A.K. Pedersen, In: A. Escher and W.S. Watts, eds. Geology of Greenland., Copenhagen, Geological Survey of Greenland: 365-385, 1976; R.W. Kent, J. Geol. Soc. Lond. 152: 979-983, 1995; R.C.O. Gill et al. (1992). In: B.C. Storey, T. Alabaster & R.J. Pankhurst (eds.), Magmatism and the Causes of Continental Breakup, Geol. Soc. Lond. Sp. Pub., 68: 335-348, 1992; S. Révillon, N.T. Arndt, et al., Lithos 49: 1-21, 1999; J.G. Fitton et al., In: A.D. Saunders, H.C. Larsen & S.W. Wise, Jr. (eds), Sci. Res., Ocean Drilling Program, College Station, Texas, 152: 331-350, 1998.

24. R.S.J. Sparks et al., Earth Planet. Sci. Lett. 46: 419-430, 1980; E. Stolper & D. Walker, Contribs. Mineral. Petrol. 74: 7-12,1980.

25. Sheth, H., 'Deccan Claptrap', Letter to Geoscientist, 2003,
26. E.g., W.J. Morgan, Nature 230: 42-43, 1971; R.S. White & D. P. McKenzie, J. Geophys. Res., 94: 7685-7729, 1989.

27. M.A. Richards et al., Science 246: 103-107, 1989; I.H. Campbell & R. W. Griffiths, Earth Planet. Sci. Lett. 99: 79-93, 1990.

28. D.L. Turcotte & G. Schubert, Geodynamics, Cambridge University Press, 2002.