Product has been added to the basket

The Great Plumes Debate - Part 2

Correspondence in Plumes debate, received after Jim Natland's contribution (If not plumes - what else? 19 September 2003)

Time to test plume alternatives

From Rex Pilger (received 19.9.03)

Sir, The arguments against deep mantle plumes continue to accumulate. Natland's1 summary of the arguments: geothermal, geochemical, petrologic, and tomographic must be challenging to plume proponents. But, does the alternative hypothesis, intraplate stress fields, buy us anything more?

In papers2-5 published some twenty years ago, I argued for stress fields along the same lines, following the suggestions of Solomon and Sleep.6 What was most bothersome about the stress argument at that time still applies today: We have examples of hotspot traces that clearly formed in semi-symmetrical fashion on opposite sides of a spreading centre: I showed via plate reconstructions that the Tuamotu and Nazca ridges are "mirror-images", just as Morgan7 had originally proposed (however, Morgan's other suggestion, that the Line Islands represent continuation of the Tuamotu trace, appears to be incorrect).

If we have mirror image traces that formed from a common source centered on a spreading centre (we might add Iceland, Walvis-RioGrande, and Kerguelen-Ninetyeast to the list), how can a shear stress field propagate across a spreading centre, producing anomalous volcanism on either flank? I suggested that once melting results due to plate fracturing and thinning, the melting anomaly might become self-sustaining, fed from below; thus when it encounters the spreading centre, it produces enough additional volcanism to generate the twin traces. However, I was never completely persuaded of the viability of this proposal. After all, the Tuamotu trace initiates abruptly across the Marquesas fracture zone, and the nature of Andean subduction beneath Peru indicates that the mirror image of the Tuamotu trace exists on the subducting Nazca plate. I don't think plate fracturing is enough.

In my previous communication,8 I reviewed some newly recognised evidence9 that demonstrates persuasively (to me at least), that virtually all of the Cenozoic hotspots/traces of the Pacific plate owe their origin and persistence to lithospheric structure. That is, variations in plate thickness (controlled by age) provide the context for the hotspots to emerge, perhaps with the additional control of variable fertility of the underlying asthenosphere and mesosphere. In other words, all of these hotspots (except Hawaii and Louisville) are demonstrably shallow in origin. And, the only reason we can't include the two bigger, long-lived hotspots is because their oldest extents are hidden in subduction zones, along with the plate structure.

I want to encourage the geochemical and petrological modellers to get to work testing what I have empirically shown. Allow the effect of sudden depressurisation of asthenosphere/mesosphere due to replacement of thicker, overlying lithosphere by thinner lithosphere (as a result of movement over the mesosphere of a fracture zone separating older from younger plate). Use the thermal models of plate thickening versus age, and try different possible mesosphere petrology recipes. See if you can produce the volumes of magma that we observe. And, see if you can produce the cross-grain gravity lineations10 while you're at it, from the same plate-thickness control.11 Develop the plate thickness-partial melting idea and, as an alternative or as an addition, the intraplate stress models advocated by others.

It's time to move on: let's begin testing alternatives to plumes. The energy devoted to plume plundering is better devoted to developing and testing viable alternatives, at least until we have indisputable tomographic evidence of the reality of deep mantle plumes. Whether we accept Kuhn's scientific revolution paradigm, our human experience tells us an existing theory isn't buried until a better model is established.

Rex H. Pilger, Ph.D., Senior R&D Program Manager , Landmark Graphics Corporation, 1805 Shea Center Drive, Suite 400, Highlands Ranch, Colorado 80129, (303) 675-2446, wirefree: (303) 589-3802; Email


Natland, J., 2003, If not plumes... what else? , Letters, GSL, September 19, 2003,
Pilger, R. H., and Handschumacher, D. W., GSA Bull., 92, 437-446, 1981.
Pilger, R. H., GSA Bull., 92, 448-456, 1981.
Pilger, R. H., JGR, 87, 1825-1834, 1982.
Pilger, R. H., 1984, JGSL, 141, 793-802, 1984.
Solomon, S. C., and Sleep, N. H, JGR, 79, 2557-2567., 1975.
W. J. Morgan, GSA Mem. 132, 7-22,1972.
Pilger, R. H., Plumes and Hotspots, Letters, Discussion, GSL, August 11, 2003,
R. H. Pilger, Geokinematics, Prelude to Geodynamics, Springer-Verlag, Berlin, 338 p. (2003).
W. F. Haxby, and J. F. Weissel, J. Geophys. Res. 91, 3507-3520 (1986).
R. H. Pilger, unpublished ms (2003).

How can a crack propagate across a ridge axis?

From Norm Sleep* (received 19.9.03)

I spent a lot of time trying to find alternatives to plumes from the 1970s until about 1986. For example, it is easy to show that slab pull should create intraplate tension more or less perpendicular to the Hawaiian chain. Thus there may be a propagating crack in the plate.

This leads to testable predictions. The crack should tap ordinary asthenosphere not much hotter material. The swell will form by delamination of the base of the lithosphere.

I was going to show that the plume was not kinematically compatible with the nose of the swell being upstream from the volcanoes and end the issue once and for all. I used a simple kinematic theory for the interaction of the plume material with plate drag. When I plotted my first guess of the shape of the nose of the swell it was right on. My disproof had instead given an explanation for the (map) shape of the swell. There are of course better dynamic models now, but it does show that antiplume people can change their minds if they form testable hypothesis.

In retrospect, the oblique lineaments of the volcanoes probably indicate that the stress is not exactly perpendicular to the chain. This is bad news for the chain being a propagating crack. A propagating crack should not cross ridge axes like hotspots seem to do.


Approximations to reality

From Norm Sleep* (received Sunday 21 September 03)

Sir, Plumes, like plate tectonics, are an approximation that helps one visualize the complicated flow pattern in the Earth. In analogy, one can say that a tornado is in the atmosphere. The tornado exists and is a useful concept (especially if one lives in a caravan park) but one cannot define exactly when it commences and when it ceases or where its edges are. The flow pattern in and around the tornado is much complicated than a simple vertical updraft.

Rigid plates are a useful approximation, but one that is clearly wrong everywhere if one looks in enough detail. One does not abandon plate tectonics because there are intraplate earthquakes, like in Yorkshire, or because the active zone of the ridge axis is wider than a hammer. Rather, one examines the geology and physics of plate interiors and ridge axes.

Plumes are part of the flow pattern in the mantle that impinges on near surface regions where we see hotspots. The stationary cylindrical vertical plume that is independent of the rest of the mantle is sometimes a useful approximate starting concept, but sometimes a concept that clouds thinking. One can understand a lot about the Earth from kinematic plate tectonics. One can understand other things from plume theory where the top of plumes are sources of hot buoyant material. One must examine the whole Earth to understand other issues like the relative motion of hotspots and the interaction of plates (dead slabs) and plumes at the base of the mantle.

Which hotspots are associated with plumes and where the plumes are at depth are valid unsolved questions. So is the fate of plume material once it reaches sublithospheric depths.

It is also productive to formulate testable quantitative alternatives to plumes that explain hotspots. It is not productive to press a cartoon concept of plumes to the point it cannot work and then say that plumes do not exist.


Reply to Prof. Sleep from Gill Foulger

Sir, Prof. Sleep's latest contribution illustrates well the fundamental problem that restricts the diversity of thought regarding the origins of melt anomalies. It is written as though we may assume that plumes exist without the slightest shadow of doubt. As long as this mind-think is pervasive, and transferred to students, progress will be limited to second-order advances of little consequence.

It does not help to compare plumes to things for which there is abundant evidence, e.g. tornadoes, implying that there is similarly compelling evidence for plumes. It does not help either to plead that it is OK if plumes are only vaguely defined. If the existence of a plume somewhere is to be accepted, convincing evidence for a) high temperatures and b) deep upwelling must be found. Until this is achieved, all the possibilities that are consistent with the observations should be entertained, including the possibility that the melt anomaly may arise from locally fusible mantle at normal mantle temperatures. (Gillian R. Foulger)

Popper, Plumes and Palaeomagnetism

From Ted Irving HonFGS (received 23.9.03)

Sir, Regarding Gillian Foulger's introduction of Popper into the plume debate, and Andy Saunder?s question about whether there was similar 'angst' in the 'early days' of plate tectonics, the answer is Yes. I* introduced Popper into the mid-century mobilism debate, as a consequence of attending, in 1963, a seminar series given by him on the logic of scientific discovery. I found, much to my astonishment, that he was (with his gift for aphorisms) describing rather accurately and eloquently what we was doing. We were then engaged at the Australian end of the palaeomagnetic pole test for continental drift. That the Earth's magnetic field was on average a geocentric axial dipole (GAD) was our basic assumption.

By the late 50s we knew that if the GAD was correct then continents had drifted in much and the same fashion as Wegener suggested. We knew from the work of Hospers and others that the GAD was, within our errors, correct for the past 10Ma, but prior to that was less certain, although we had written many papers arguing that it was applicable for the later half of the Phanerozoic. Fixism was then the ruling theory, and our critics, perhaps sensing defeat, fastened on to what they believed was our Achilles? Heel, the magic word 'non-dipole', and used it at every opportunity. However, according to Popper, their non-dipole hypothesis was unfalsifiable and therefore unscientific, because any possible observation could be accomodated by introducing ad hoc further harmonics in the time vaying geomagnetic field.

Happily the 'angst' soon evaporated, when that other aspect of paleomagnetism, reversals recorded in the sea floor, confirmed mobilism, and hence the legitimacy of my Popperian argument.

Finally I should note that the 'angst' to which Andy Saunders referred is very evident in the current debate about the Late Palaeozoic GAD, Permian palaeogeography, and the relative merits of Pangaea A and B.

*Irving,E., 1964, Paleomagnetism and its Application to Geological and Geophysical Problems, Wiley, p.133.

A curate's egg - Marc Davies, a PhD student at the Open university, UK, reflets on the Penrose Meeting. (Received 2 October 2003)

Equal time plea

From Alexei Ivanov*

Sir, The classical lower mantle plumes of Jason Morgan [1] were used to explain the mechanism of uprising convection in the Earth?s mantle and for explaining the unidirectional propagation of volcanism within lithospheric plates. The suggested features of such plumes (e.g. head-and-tail structure, stability in the mantle, excess of temperature etc.), however, have not been confirmed so far. This lack of corroboration has been taken by Gillian Foulger to show that existence of mantle plumes is just a hypothesis - and an unsuccessful one at that. One of the major counterarguments of the plume supporters has been that there is no alternative explanation that is any better. Let me claim the opposite.

Geologists have great expectations of deep seismic tomography of the mantle. Large progress in that field has been achieved in recent years, but plume head and tail structures have not been revealed [2]. Plume supporters generally attribute this to the low resolution of deep seismic tomography. But the question is, why should we waste our time talking about what we cannot see (plumes) and overlook things we can see in these seismic tomography images?

In 2001 Sergei Balyshev and I noted morphological similarity between the low velocity anomaly beneath Africa and high velocity anomalies elsewhere (undoubtedly connected to subducting slabs [3] ). We used a simple model to calculate that subducted slabs can be radioactively heated on a time-scale of 1-2 billions years and hence due to heating can be converted from high-velocity to low-velocity structures. In 2003 Jeroen Ritsema and Richard Allen have written in their outstanding paper [2]: ?? [low-velocity anomalies] appear to connect to broad lower-mantle low-velocity anomalies beneath the Pacific and Africa via convoluted paths, reminiscent of high-velocity anomalies in the lower mantle ? that have been interpreted as remnants of the Farallon and Tethys slabs?. They did not go one step further, however, and make the genetic link between those anomalies that are so similar in shape and so different in velocity. Is it right time to make this step?

The idea that low-velocity anomalies are re-heated slabs is not so strange as it sounds. Yet in 1988, Paul Silver and his colleagues [4] suggested a model of penetrative convection in which modern subducting slabs are downwelling convective flow and ancient warmed up slabs are upwelling convective flow. This model with support of radioactive heating of slabs can account for almost all observations thought to be attribution of the plume model [5].

I would like to finish this letter by saying that the plume model is certainly not the only one. The propagating crack model for temporal migration of intraplate volcanism and penetrative convection model for mode of whole-mantle convection (and maybe some other models) should be paid as much attention as plumes have had in the last thirty years. Only comparing of observations with various models can give an answer on how the Earth actually works.


Morgan W.J. Convection plumes in the lower mantle. Nature, 1971, v.230, pp. 42-43.
Ritsema J., Allen R.M. The elusive mantle plume. Earth Planet. Sci.Let., 2003, v. 207, pp. 1-12.
Balyshev S.O., Ivanov A.V. Low-density anomalies in the mantle: ascending plumes and/or heated fossil lithospheric plates? Doklady Earth Sciences, 2001, v. 380, pp. 858-862.
Silver P.G., Carlson R.W., Olson P. Deep slabs, geochemical heterogeneity, and the large-scale structure of mantle convection:Investigation of an enduring paradox. Ann. Rev. Earth. Sci., 1988, v.16, pp. 477-541.
Ivanov A.V. Plumes or reheated slabs? October 3, 2003

*Alexei V. Ivanov, Senior Researcher at the Institute of the Earth?s Crust, Siberian Branch of the Russian Academy of Sciences, Irkutsk, Russia