Product has been added to the basket

The Great Plumes Debate (2003)

Gill Foulger gets serious in the search for tangible evidence of plumes

To refer to this Debate, please cite as follows:

Foulger, G. et al., 2003; The Geological Society of London Great Plumes Online Debate.

How this section is structured

This debate consists of letters and longer essays accessible via hotlinks to the left. After Gill Foulger's "Plates Plumes and Popper", these letters and essays are ordered with the oldest on the bottom and the youngest on the top. Letters received after Jim Natland's contribution (19.09.03) are accessible on Latest Correspondence via the link below.

Plumes, plates and Popper

Gillian Foulger writes: In 1963, Wilson1 suggested that there is a hotspot in the mantle, which is forming the time-progressive Hawaiian island chain as the Pacific sea floor moves over it. Within a few years, this hotspot had migrated to the core-mantle boundary, replicated itself into 20 plumes, each 3000 km tall and with a giant plume head, and these were driving the newly discovered tectonic plates2. It was pointed out that, given the rate at which plumes were multiplying, someone urgently needed to prove that they don’t exist before it was too late. Despite this timely warning, by 1999 the number of deep mantle plumes in the Earth had reached a critical 5,2003. Fortunately, the crisis was averted in 2003 when their number crashed to only 94. These extraordinary fluctuations in the plume population probably tell us more about Earth science than the Earth.

Unlike the plate-tectonic theory, whose predictions have been repeatedly borne out by such observations as earthquake mechanisms and plate motion rates measured by GPS, the fundamental predictions of the plume hypothesis have not been confirmed, despite three decades of intensive work. These predictions are a) narrow, vertical seismic structures traversing the whole mantle, b) high temperatures, and c) relative fixity of "hotspots".

Where mantle tomography has imaged proposed plume locations, anomalies are found to be confined to upper mantle, or even lithospheric depths, e.g., Yellowstone, Tristan da Cuhna and the Azores5,6. Heat-flow measurements and petrology provide little evidence of the high magma temperatures postulated for deep plumes, e.g., at Hawaii, Louisville and Iceland7,8. Hotspots move relative to one-another at rates of the order of centimeters per year, and many island chains originally assumed to be time-progressive are not9. Physical models suggest that classical plumes may be impossible because the enormous pressure in the deep mantle suppresses the buoyancy of hot material10.

Plume enthusiasts have responded to these challenges with creativity. Carefully truncated cross sections, with colour scales cranked up, give noisy images the illusion of strong anomalies traversing the mantle11. Proposed depths to plume heads have been increased in order to preclude measurable heat to flow at the surface12. “Mantle wind” is invoked to explain why hotspots move relative to one another.
Iceland At Iceland, hypothetical plume motion has been calculated assuming fixity relative to other hotspots13, and the theoretical “track” used as an a-priori assumption in subsequent work.

Plumes have been proposed to come from almost any depth, to rise vertically or tilt, to flow for long distances laterally, to have narrow or broad conduits, to have no plume head, one head, or multiple heads, to produce steady or variable flow, to be long or short-lived, to speed up or slow down, to have a source that is either depleted, enriched, or both, and to have either high or low 3He/4He. Such model elaboration is a sign of a hypothesis in trouble.

The single, unifying theme in plume science is the unquestioning belief of most Earth scientists in their existence, indicated by the use of the term “the” plume when discussing one's area of interest. The possible non-existence of a plume or plumes may have never have been suggested to many students of Earth science. Alternative mechanisms are rarely mentioned and papers suggesting them are even more rarely accepted by mainstream journals.

A few Davids are, nevertheless, currently challenging the plume Goliath. They are meeting a foe armoured heavily with assumptions, belief and faith, with only the scientific method to hand. Correct scientific method requires us to proliferate and falsify theories, not to attempt to prove them correct, adapt them to fit unpredicted data, or to consider only one14. We might do better by our students if we spent a little less time on plumes and a little more on Popper.

A common argument is that we have to accept the plume hypothesis because it is the only game in town. If Lavoisier had taken that attitude we would all still be designing ad hoc adaptions to the phlogiston theory to account for new, paradoxical observations. Furthermore, an exciting new generation of alternative models for volcanic regions is emerging, which explains them as the “by products” of plate tectonic processes.
Hotspots Figure: Hotspots proposed to originate from the core-mantle boundary (red), the base of the upper mantle (yellow) and in the lithosphere (green). (Courtillot et al., 2003. Figure by G. Sella).

Volcanism requires both a source of melt (apparently without exceptionally high temperatures) and extension of the Earth's surface, to allow it to escape. The reintroduction of basalt into the shallow mantle at subduction zones causes inhomogeneity and locally enhanced fertility in the form of eclogite, from which exceptionally large volumes of melt may be extracted at relatively low temperatures15. Intraplate deformation causes extension remote from plate boundaries.

This often occurs along pre-existing lines of weakness such as transform zones and old sutures. The latter probably are also the sites of old eclogite-bearing slabs that were trapped in the lithospheric suture when continents collided. Anomalous volcanism traditionally attributed to plumes commonly occurs at such locations e.g. Tristan da Cunha, the Deccan Traps, Yellowstone, Iceland and many of the Pacific volcanic chains5,16,17.

This new movement is particularly exciting in that it could radically simplify our view of convection in the Earth. The plume hypothesis as it is applied today requires that Earth dynamics is driven by two independent modes of convection - plate tectonics and plumes. The former is driven by forces at plate boundaries - ridge push and slab pull, and the other is driven by heat from the Earth's core. The new theory suggests that only one of these is needed - plate tectonics. Volcanism that appears to be anomalous either in its location, distribution or volume rate then results from the inhomogeneity imparted to the mantle by plate tectonics and intraplate deformations that occur preferentially along pre-existing lines of weakness. The possibility of such a radical simplification alone is a strong hint that may something important may be going on here.

To be fair to plumes, however, they have served geologists well for over 30 years. In 2002 several hundred papers about them are listed by the Science Citation Index, so the subject is paying a lot of mortgages. The assume-a-plume approach has also relieved researchers of the hard work of thinking up new theories - a welcome relief in these days when we are all expected to publish six papers a year or else. However, the time may have come to retire plumes on a full pension and make way for younger theories - or are they really older ones recycled?

References cited:

J. T. Wilson, Can. J. Phys. 41, 863-870 (1963).
W. J. Morgan, Nature 230, 42-43 (1971).
B. D. Malamud, D. L. Turcotte, Earth planet. Sci. Lett. 174, 113-124 (1999).
V. Courtillot, A. Davaillie, J. Besse, J. Stock, Earth planet. Sci. Lett. 205, 295-308 (2003).
R. L. Christiansen, G. R. Foulger, J. R. Evans, Bull. Geol. Soc. Am. 114, 1245-1256 (2002).
J. P. Montagner, J. Ritsema, Science 294, 1472-1473 (2001).
C. Stein, S. Stein, Astron. Geophys. 44, 1.8-1.10 (2003).
K. Breddam, J. Pet. 43, 345-373 (2002).
A. A. P. Koppers, J. P. Morgan, J. W. Morgan, H. Staudigel, Earth planet. Sci. Lett. 185, 237-252 (2001).
D. L. Anderson, Science 293, 2016-2018 (2001).
H. Bijwaard, W. Spakman, Earth planet. Sci. Lett. 166, 121-126 (1999).
M. Liu, C. G. Chase, J. geophys. Res. 94, 5571-5584 (1989).
L. A. Lawver, R. D. Muller, Geology 22, 311-314 (1994).
K. Popper, The Logic of Scientific Discovery (Hutchinson, London, 1959).
M. J. Cordery, G. F. Davies, I. H. Campbell, J. geophys. Res. 102, 20,179-20,197 (1997).
A. D. Smith, Terra Nova 5, 452-460 (1993).
G. R. Foulger, Astron. Geophys. 43, 6.19-6.23 (2002).

*Dr Gillian R Foulger (University of Durham, UK) is currently Visiting Scientist with the Volcano Hazards Team, US Geological Survey, 345 Middlefield Road, MS 910, Menlo Park, CA 94025-3591, USA. email:


For correspondence subsequent to Natland (below), please view Part 2 of the debate.
  • If not plumes - what else? Prof. Jim Natland (University of Miami) answers the charge that anti-plumers have no coherent model to replace plume theory

Letter from Gillian Foulger to the Site Editor

Dear Ted,

Many thanks for bringing to everyone's attention the very welcome coverage of this subject (Fuming over Plumes, by Anjana Ahuja, Times T2, 18 September, 2003 p14).

It is a pity, nonetheless, that the impression is given of tremendous emotion. I suppose this was done in order to "sex it up" for the readership, but it is distracting and irrelevant. The Geol Soc webpage presents a frank, measured debate on the subject, which usually isn't possible because scientific exchanges tend to be talks where there is essentially no time for "debate". It is a pity, therefore, that words such as "bitter", "virulence", "acrimonious" and "impassioned" are used. Isn't it possible to have a good, professional discussion about a scientific theory without having it written off as mere emotion?

I would add that we can "see" both electrons and the Earth's core. Electrons are "seen" for example from the tracks of ionised particles they leave in bubble chambers. The Earth's core is "seen" with seismic waves - there is no shadow of doubt of its existence. The human eye is not the only detector that can "see" things. However, neither seismic waves nor anything else can "see" plumes. The most recent authoritative summary of this issue was presented by Adam Dziewonski at the Penrose conference (see his abstract: In particular, beneath Hawaii, seismic velocities are fast compared with the rest of the southwest Pacific, as was clearly shown years ago by multiple ScS bounces from the 1973 M 6.2 Hawaii earthquake ( The Times article says seismic velocities under Hawaii are slow, which is the opposite of the truth. It is regrettable that wrong information is constantly repeated and continues to hamper progress. (Gillian Foulger)
Ted Nield comments: Strong emotion may indeed hamper progress in the scientific arena, but it is the making of public profile. Oddly, scientists getting excited about things is very good PR for them (and the Society). And you're right, Gill, it's the emotion that sells the story. What has become interesting about the plumes debate for journalists is that it changes the story from a very unsaleable "here's something you've never heard about that might not be true" to "here's some scientists getting hot and bothered about something you have never heard of" which gives the story human interest (which sells it) curiosity value (which makes folk read it) and counteracts the image that scientists often have (and which many, unfortunately, like to encourage) of being bloodless crankers of the handle on the scientific method...

Plumes and hotspots

From Rex Pilger*

Sir, Since W. Jason Morgan (WJM) introduced the mantle plume hypothesis1,2, subsequent research has tended to focus on the two distinct facets of the proposal: (1) Hotspots form an absolute (I prefer to use proper') reference frame. (2) Hotspots represent cylindrical plumes flowing from the base of the mantle. Yet, these two facets are coupled intimately if we attempt to rethink the kinematic logic of the hypothesis. Evidence assembled by WJM and subsequently supported by Minster et al.3 suggested that contemporary hotspot traces represent a consistent reference frame. If indeed hotspots represent a global system, then they must originate from the lower mantle, deeper than the deepest subduction zones; subduction zones are in motion relative to one another, requiring relative motion of the shallow mesosphere on either side of the zones. Further a deep source would require very rapid transport to the base of the lithosphere - thus cylindrical plumes of rapidly rising material.

There was a logical flow to WJM's deep mantle plume hypothesis, dependent upon evidence for hotspots forming a globally consistent proper reference frame. However, as we fast-forward some thirty years, we now realize that the hotspots of the Atlantic and Indian Ocean (except for the northernmost Atlantic) form a proper reference frame, as do the hotspots of the Pacific Ocean, but the two hotspots systems are essentially independent. The recent global reference frame is a contemporary accident. For years the inconsistency between the two sets of hotspots was explained in terms of a possible hidden plate boundary in Antarctica - a boundary that can now be shown to have inadequate displacement to explain the discrepancy.4,5 In other words, the Pacific Ocean and Atlantic-Indian Ocean hotspot systems are separate and distinct, for at least the last 80 m.y. Therefore, following the kinematic logic above, the hotspot reference frame(s) no longer need be within deeper mantle.

There is evidence that the hotspot reference frames are in fact shallow: (1) Intracontinental contemporary stress6 and paleostress7 fields are consistent in orientation with motions in the hotspot reference frame. (2) The oldest portions of minor hotspot traces of the eastern Pacific Ocean are genetically related to the age-thickness structure of the overlying Pacific plate.7 (3) Cross-grain gravity lineations of the eastern Pacific, which parallel plate motions in the hotspot frame8 are restricted in their occurrence to thinner, younger plate, similar to the minor hotspot traces.9 (4) Motion of the Pacific (or 'Hawaiian') hotspot set relative to the Atlantic-Indian Ocean ('Tristan') set between 80 and 25 Ma is subparallel with motion of the North and South American plates relative to the Tristan set - in other words, the Cordilleran and Andean subduction zones appear to have displaced the hotspot set beneath the Pacific Ocean plates as they advanced to the west.7

In sum, hotspot frames are shallow and minor hotspots originate at shallow depths. The big hotspots may originate from deeper mantle via WJM's plumes, but their shallow-most portions are fixed within the local proper reference frame set - within what are termed 'mesoplates', of which there are three: Hawaiian, Tristan, and 'Icelandic'.7 Mesoplate boundaries are largely determined by deep subduction zones and relative motions of the mesoplates.

I leave seismic, geochemical and petrologic arguments for and against plumes to others. However, I recall a conversation with a prominent basalt petrologist: I expressed the hope that petrologists such as he would provide the constraints we geophysicists need to understand the mantle. With a distressed look, he said that, on the contrary, it was his hope that geophysicists would provide the constraints that petrologists need.

*Landmark Graphics Corporation, 1805 Shea Center Drive, Suite 400, Highlands Ranch, Colorado 80129 United States of America (303) 675-2446

References cited:

W. J. Morgan, Nature 230, 42-43 (1971).
W. J. Morgan, Geol. Soc. Am. Mem. 132, 7-22 (1972).
J. B. Minster, T. H. Jordan, P. Molnar, E. Haines, Geophys. J. Royal Astron. Soc. 36, 541-576 (1974).
S. C. Cande, C. A. Raymond, J. Stock, W. Haxby, Science 270, 947-953 (1995).
C. A. Raymond, J. M. Stock, S. C. Cande, Geophys. Mon. 121, Am. Geophys. Union, 359-375 (2000).
M. L. Zoback, et al., Nature 341, 291-298 (1989).
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).

Plumes, cells and the mantle wind of change

From Don Findlay

Sir, Professor Saunders's promising argument (see below) that "the evidence from Hawaii proves the existence of a discrete, focused, persistent and deeply sourced convective mantle upwelling..." for the Pacific founders on the word 'convective', thus hiding the titillating plumage of the model beneath the drab garb of plate tectonics. That is, if he intends to support his argument by the use of his posted images. Diapir rise is not convection, and neither does it imply it, yet his images clearly show that a collapsing diapir at the terminus of its natural reach is the concept under discussion. Rightly, I think.

Convection? Inscripted in the ocean floors from the ridges to the subduction zones of plate tectonics, there is no geological evidence for any more than a single half-cycle of 'convective overturn' (which is, in effect, diapir rise), and the modelled images cited above reinforce the point. Moreover spatially, oceanic ridges once supposedly defining upwelling convection do not sit easily with the tomography-defined 'plumes' of today. So, what about convection? How do the ridges that once evidenced 'convection' match with the plumes of diapir rise today? Where, for example, is the 'convective return' for the "convective mantle upwelling" of the Hawaiian plume mentioned by Professor Saunders, presuming it is different from the subduction zone partnered with the East Pacific Rise?

It seems in this discussion we are finally (and not before time) witnessing the troubled transmutation from the CONVECTIVE cell of yesterday to the DIAPIRIC plume of today. This lexical pas-de-deux is the middle step in the conceptual dance that plate tectonics must execute if it is to get out of the blind alley it has landed itself in through overlooking some obvious and fundamental aspects of ocean floor structure.

Once the words/steps are in place, the ground will be set for the meme-driven machinery of consensus to change, from convection-driven plate tectonics to plume/diapir-driven (driven? ...or following?) Earth Expansion. The change in thinking from convection cell to mantle plume is substantial, and caution should be observed in the transfer of plate-baggage. Moreover, the question whether plumes are potentially 'drivers' or 'followers' in the larger global dynamics should be addressed.
  • An alternative to the alternative - Prof. Andy Saunders (Leicester University, UK) defends the plume hypothesis and challenges its detractors to come up with an alternative
  • "Ptolemy, Piltdown, phlogiston, polywater and... plumes?" by Don L Anderson (CalTech, Pasadena, Ca) - from the August 2003 issue of Geoscientist

Expansive view

From Don Findlay*

Sir, Have plumes had their day? Certainly. Within a plate tectonic paradigm at any rate, which requires recycling of mantle diapirs as convection cells. The main 'plume' among them however, the Pacific, will almost certainly gain respectable rebirth as the driver of Earth expansion, to which plate tectonics is inevitably and inexorably headed. Not only 'plumes' (in the sense of plate tectonics), but plate tectonics itself, has had its day.

Tough, but it's the way it is.


Stop that whining

from Dr. Stuart A. Weinstein, Geophysicist.

Sir, The fundamental problems that never seem to be addressed by those who suggest hotspot volcanism sources are shallow are mutli-fold. One of the fundamental problems with placing hotspot sources in the shallow mantle is that it provides no explantion for hotspot tracks which have gaps. When a ridge migrates over a hotspot, spreading produces a break in the track, but hotspot activity as far as we can tell is not interupted by the process. If the source was shallow, one would expect it to be exhumed. If it is a crack, one wonders why the crack doesn't eventually spread into the rifted region. Second, there is the age progression. The age progression switches polarity. That is, suppose plate motion is to the north, the hotspot track gets older northward. After a ridge migrates over it, and relative motion switches direction, the hotspot track gets older southward. An example of a scenario like this is exhibited by the Kerguelen hotspot and resulting tracks.

Next, we have the longevity of hotspots. Hotspots can remain active for well over 100 myr. It remains to be demonstrated whether or not "crack" volcanism can remain active that long. If crack volcanism exists, than why are the basalts produced ocean island basalts and not mid-ocean ridge basalts?

I have by no means exhausted my criticisms. I also appreciate Foulger's wanting to make the task of mantle convectioneers easier. However, simplicity at the cost of violating other observations is never worth it.

I am not claiming that all hotspots that have been so designated over the past decades are indeed formed as a result of plume activity, but to make the claim that plumes don't exist or that the crack hypothesis is better than the plume hypothesis, may get rapid attention from the press, but less rapid consideration from the geodynamical community. In which case I suggest Foulger spend more time crafting her hypothesis, addressing objections, and spend less time whining that reviewers are holding back progress. Don Anderson has been discounting the plume hypothesis for decades, and I haven't heard him whine about reviewers holding back progress.

Where Popper fails us

From John Hernlund*

Sir, I would like to speak up against some of the non-plume ideas expressed in this particular thread of letters. I enjoy Gillian Foulger's work, and wish her great success in improving our understanding of volcanism on the Earth (and perhaps other planets). I hope this continued discussion will lead us all to a better watering hole. Here, I will only address the shortcomings of this particular approach, based on an appeal to Popper's philosophy of science (Ref. #1).

The rhetorical devices used in Dr. Foulger's piece to argue against the plume hypothesis are clever, but flawed at a fundamental level. These devices are also over-used in Professor Foulger's own recent publications, some of which I have had the opportunity to read, and enjoy. The problem here is the following: Dr. Foulger's piece presents the Modus Tollens argument:
  • If deep-seated mantle plumes, then A,B,C,D,...
  • Not A,B,C,D,...
  • Therefore, not plumes
Recalling our college logic classes, we know that such an argument is valid, so long as the premises are true. Here is where we encounter the essential problem in Professor Foulger's piece: none of the proposed premises bear much resemblance to the truth. Thus she proceeds by handicapping the hypothesis by her own false premises, which she then proceeds to attack. When persons point out that Dr. Foulger's premises are incorrect, she then accuses them of making 'ad hoc' adjustments to the hypotheses. Perhaps the essential problem is that nobody is properly taught the plume hypothesis as a student, thus perpetuating these types of misunderstandings.

If we are going to try falsifying a hypothesis, then we ought to be working on the right premises, otherwise we cannot make any firm progress.

But this presents other problems as well. Perhaps a brief (and extreme) example helps. If I want to test the hypothesis that it is not raining outside today, then I must construct a test, preferably using a solid logical format. If I present this argument:
  • If it is raining outside today, then George Bush has oranges for breakfast
  • George Bush didn't have oranges for breakfast.
  • Therefore it is not raining outside.
This is another modus tollens argument, where I have made the absurdity of the premise fairly clear. Obviously, somebody eating oranges 3000 miles away doesn't have anything to do with whether it is raining outside. This is where Popper failed us. It is the construction of the premises to be tested that confounds scientific progress, and it is where the true business of science is forged (e.g. Ref. #2). However, you can also see why it is absurd to suggest that a better premise is a sign of a 'troubled hypothesis' in this case. We cannot, on the basis of the above argument, reject the idea of rain fall. Just make a better test, e.g. going outside and see if you get wet! Is that ad hoc? Or is that just common sense? Cheers!


Popper, Karl, The Logic of Scientific Discovery, Hutchinson, London, 1959.
Laudan, Larry, Progress and its Problems: Toward a Theory of Scientific Growth. Univ. Calif. Press, Berkeley, 1977.

Gillian Foulger replies:

Prof. Hemlund has misunderstood my article.

I do not assert that because the predictions of the plume hypothesis are not borne out, therefore plumes do not exist.

I assert that because the predictions of the plume hypothesis are not borne out, it is unsafe to assume, a priori, that this model is correct.

I point out that the plume hypothesis has become so flexible that it seems nothing could be observed that could rule it out. Thus it appears that the plume hypothesis, as it is applied today, is unfalsifyable. If so, then it is unscientific, and is hindering our progress in understanding the Earth because it removes the perception of need to develop and examine alternative theories.

I hope this clarification is useful.

Plumes in a balloon

From Seth Stein*

Sir, From my perspective as a former editor (JGR), reviewer, and author on some hotspot-related papers, I wouldn't blame anyone for the state of thinking and publications about hotspots and mantle plumes except ourselves, as in Pogo's dictum "we have met the enemy and they are us." In the absence of any other clear model, we've accepted very vague ideas about plumes and allowed them to be the null hypothesis for excess ridge or intraplate volcanism. The hypothesis evolved from fairly rigorous criteria (deep low velocity anomaly at present site of volcanism with age progression and near-fixity) to a point where plumes don't have to meet any particular test. Hence the hypothesis now always works with appropriate site-specific modifications, but increasingly doesn't tell us anything or predict anything, especially about structures formed in the past. It does, however, make it harder to offer non-plume explanations.

This is reminiscent of the joke where lost balloonists call to the ground 'where are we'? and are told 'you're in a balloon'. The answer is absolutely true and totally useless.

Fortunately, we're starting to ask the hard questions needed to decide what it means to say that something is a plume, and to decide if, when, and where the concept is useful.

*Department of Geological Sciences Northwestern University, Evanston, IL 60208

Hotspots fight back!

From Dr Ray Kent (University of Leicester)

Sir, I was intrigued to see alternatives to the mantle hotspot hypothesis being presented as 'Radically different, alternative models [that] ... if substantiated ... may lead to a first-order paradigm shift.' This suggests that such alternative hypotheses offer fresh, exciting insights; they are poised to flood the literature with new data and interpretations in order to brush away the established order of things. There is even some suggestion of long-lived conspiracies between referees and journal editors to stifle emerging non-plume hypotheses in their infancy, and thereby to suppress 'the truth'.

The reality is somewhat different. Some folk have never accepted the plume hypothesis, and their views, far from being suppressed, are widely published in the geological literature. For example, the 'big daddy' of non-plume hypotheses, Don Anderson, is credited with having published more than 30 'non-plume' papers in the past 8 years, most of which appear in respected journals. Similarly, Alan Smith, Gill Foulger and others allied to the 'Andersonian' cause have produced a rash of papers in the past 15 years, all championing non-plume models. Hardly a case of being stifled at birth by evil editors keen to hang on to the old ways!

One might also object to the suggestion that non-plume models are somehow 'new', and hence it is only a matter of time before such hypotheses are proven. There were non-plume models long before the first plume model, and the two have existed (more or less peaceably) side-by-side for forty years. Thus far, neither side has been able to prove their case conclusively. This suggests to me that 'the truth' is probably somewhere between the two opposite ends of the spectrum, i.e. some but not all of the largest volcanic outpourings were fed by hotspots in the mantle, whereas other, smaller, outpourings have their origin in plate tectonics.

To conclude, the best way forward is to continue talking to one another rather than throwing stones at one another's glass houses. This means genuine debate must occur, preferably in a nice (plume-top) setting like Hawaii ... !

Deccan claptrap

From Dr Hetu Sheth (Department of Earth Sciences, Indian Institute of Technology (IIT) Bombay,)

Sir: Proposing mantle plumes as explanations for anything and everything has been fashionable in recent years for Earth scientists. The original plume hypothesis posulated plumes that were deep-sourced, narrow, fixed, 'enriched', abnormally hot, etc. Competing non-plume models were rejected without sufficient consideration. Today's plumes can be hot or cool. They can be hydrous while at the same time hot. They can be fixed or swaying in the mantle, narrow or very broad ('superplumes'), with cylindrical, tabular or any other imaginable shape (as required), and deriving from any depth in the mantle. They can be 'enriched' or 'depleted', with just the chemical flavour required to explain the data. Large volumes of homogeneous magma in a volcanic province become 'compelling evidence for a plume origin', but if significant chemical heterogeneity is found, it becomes 'evidence that the plume is heterogeneous'. Ad hoc escape routes abound. Entire plume heads can be missing (e.g., Hawaii), or plume tails (e.g., Siberia, Ontong Java).

I started my student career as a plume enthusiast but soon felt the idea was nonsense. Four years ago I published in mainstream journals against the plume head-tail model for the Deccan Traps of India (despite hostile reviewers, some of them understandably anonymous). I hoped for debate and reconsideration. What I have seen since is total silence and indifference, however. Papers citing the Deccan as an excellent example of a plume-generated province are pouring in every month even now. In fact, I do not know of a single flood basalt-hotspot track pair where all the predictions of the plume head-tail model are fulfilled without requiring special pleading.

Gill Foulger's article in your magazine is excellent, as are the letters by readers. Thank you for your coverage of our views.

Plumes - the truth will out

Reply by Gillian Foulger to John Wright (see below)

Dear John, Thank you for your supportive words in response to my recent article Plumes, Plates and Popper in Geoscientist. You are expressing sentiments shared by many geologists who think the Earth is made of rocks rather than numbers on a spreadsheet or resolution kernels in a computer's memory.

The plume hypothesis has disenfranchised geologists in the same way that Kelvin disenfranchised them in the early 20th Century by his authoritarian insistence that, despite their observations, the Earth had to be very young because it was still warm. Geologists have been persuaded that their observations are inconsequential to identifying the source of volcanic regions by the insistence that it is so deep in the Earth than only geophysics can reach it. Then, when geophysics cannot find it, we are told that it must be because it is too narrow to be detected. However, if regions of anomalous volcanism are indeed sourced at shallow depth, it may be that geology and geochemistry are capable of probing essentially the entire depth interval of any consequence to what is at the surface. I encourage geologists to step outside of the plume framework and interpret their observations in the way the data suggest, rather than forcing them into an ill-fitting, a priori model.

Gillian R. Foulger 4 June, 2003

The plume's a balloon

Open Letter to Dr Gillian Foulger, from Prof. John Wright (Open University)

Dear Gillian, Oh, I did so greatly enjoy your blast (Geoscientist May 2003, p. 16-17) at the plume hypothesis - which even now is difficult to dignify with the status of theory, given the inherent inconsistencies you describe. Some thirty years ago (golly, judging from the photograph, that's before you were born!), I railed against some of the worst excesses promulgated by advocates of the hypothesis (refs. 1-3), who seemed to see plumes where none existed. My particular beef was the proposition that a plume gave rise to the (granitic) White Mountain magmas in New Hampshire, and that the same plume is now under the (basaltic) Azores volcanoes. Equally hare-brained (to me) was the suggestion that another plume was responsible for the Nigerian Younger Granites, and that this one ended up under Ascension Island. Mind you, I cannot deny that the hypothesis has a superficial appeal, even for me, especially as regards the Hawaiian island chain.

However, I readily concede that it's the only example that seems to work at all, and I understand there are several other island chains in the Pacific that are without regular age progression and may lack convincing evidence of underlying plumes.

But where do you go from here? You have the enormous advantage of youth on your side, for I have to confess that, with advancing years I have become less and less familiar with the relevant literature; in addition to which, you and your colleagues must know and understand a hell of a lot more than we did back in the 1970s. All the same, you have a hard row to hoe, a long steep hill to climb, if you hope to convince the Earth science fraternity that plumes are a figment of their collective imagination. Can you really bring them round to your view that the mode of convection that drives plate tectonics is the "only game in town"? Plans to send a probe to the centre of the Earth may appear to have overtones of lunacy (ref. 4), but such a project might perhaps contribute to resolving such questions. Anyway, all power to your elbow, and best wishes for success.


Nature, vol. 244, p. 565-7 (1973)
Ibid. vol. 247, p. 25-6 (1974),
Ibid. vol. 248, p. 365-6 (1974).
Ibid. vol. 423, p. 239 (2003).

Non-plumian professing

From Tony Doré (Statoil)

Sir, I thoroughly enjoyed Gill Foulger's recent article on mantle plumes ("Plumes, plates and Popper", Geoscientist v. 13 no. 5).

I have been puzzled about Iceland for a long time. Specifically, what bothers me is the assumption of an Iceland hotspot track based on working out where the plume centre "must have been" assuming a fixed hotspot reference frame. So by that reckoning, the plume started in the Labrador Sea at about 60 Ma, and moved southeastward across Greenland until by some amazing coincidence it resides under the spreading ridge/plate boundary at present day. And yet, if you look at the hotspot "track" in the present ocean, it is distributed symmetrically either side of Iceland (the Greenland Faroes Ridge), not just on one side as it would be if it had migrated in from the northwest. Furthermore, if you look at the Thulean volcanism, representing earliest known "plume" activity in the Paleocene, it extends approximately the same distance into the continents either side of the N Atlantic (NW to the Baffin Island coast, SE to Lundy Island) - i.e. it is symmetrically distributed about Iceland again. All of this empirical evidence cries out that the hotspot must have always resided under the plate boundary. Conversely, there is hardly a scrap of corroborative evidence for a hotspot trail moving across Greenland. But, if you accept these observations, a deep-seated plume fixed relative to the moving plates cannot be the explanation for Iceland. I have put this evidence to experts on the Iceland hotspot (I am not one) and, although they don't deny it, they seem unwilling to take this issue on.

I still think the Hawaiian seamount chain is best explained by the plate moving over a fixed heat source. However, other hotspots such as Iceland don't fit the plume dogma - and especially the fixed hotspot reference frame - at all well. It's great to see somebody tackling this issue head on.

From Alan Smith, CIE-UNAM, Temixco, Morelos, Mexico

Sir, I have long considered the plume model to be a bandwagon characterised by a monkeys&typewriters-style approach of looking for every possible variation without ever pausing to consider the possibility of fundamental flaws in the whole concept. The monopoly it has enjoyed is a result of the peer review system whereby any criticism or presentation of an alternative to a ruling theory needs a majority vote from proponents of the ruling theory, and editors allow reviewers to reject manuscripts for the most trivial of reasons. To say the plume concept has served the geological community well is overgenerous, more like it has set back a true understanding of the geodynamics of the Earth's interior by three decades.