Geoscientist Online

Prof. Andy Saunders* (Dept. of Geology, University of Leicester, UK) mounts a defence of the plume hypothesis - 6 August 2003

Do mantle plumes exist? This time last year I would have said "yes, of course". Like the electron, which is never seen, but whose effects are predictable and observed, plumes are elusive, hidden, enigmatic, and important. Now I find my certainties tested by a flurry of articles and letters1 vehemently purporting to show that the plume theory is a falsehood, and that the features we see at the Earth's surface attributed previously to mantle plumes are, apparently, caused by very different phenomena. One can be forgiven for being confused.

The plume model is criticised as being too vague and all-embracing to be testable by scientific method2. Its supporters are said to shift ground when questioned on their subject. I recall similar debates about ophiolites, but no-one now seriously questions the ophiolite model. Yes, plumes may well have different forms, different histories, variable (and inherently complex) compositions, and different behaviours. They probably even move laterally through the mantle! We imperfectly understand their formation and characteristics. This complexity is to be expected in models of a dynamic, evolving planet.

But let us consider the evidence in support of the plume model, and take a sideways glance at some alternative hypotheses.

The formation of the Hawaiian Islands and the Emperor Seamounts

Figre 1: The Hawaiian Islands photographed from above by NASA's SeaWiFS satellite. Credit: NASA/GSFC and Orbimage

The facts. 1) An average basalt production rate of about 0.013 km3 a-1 for at least the last 80 Ma(ref. 3). 2) Enhanced topography of the Pacific lithosphere, at least partly uncompensated and dynamic, forming the Hawaiian swell4. 3) A deep-seated velocity anomaly in the mantle5 underlying the volcanically-active islands. 4) Magmatism that ceases once the islands are carried away from the Hawaiian swell. 5) A clear time progression along the islands and seamounts3. Plume theory can elegantly satisfy these first-order observations. Other models, for example those which invoke magmatism at the tip of propagating faults, do not6.

Criticisms of the plume model (such as the suggestion that the Hawaiian plume moves through the mantle, or that there is no evidence that the Hawaiian plume ever had a start-up head) do not seriously challenge the validity of the model. They modify our understanding of the Hawaiian plume and of the model itself. I would argue that the evidence from Hawaii proves the existence of a discrete, focused, persistent and deeply-sourced convective mantle upwelling - my working definition of a plume - beneath the Big Island.

The formation of Iceland

Few doubt that Iceland results from anomalous production rates of basaltic magma. The crust here is thicker (up to 35 km thick7) than normal ocean crust, and the lithosphere has dynamic support from the underlying mantle. Without these, Reykjavik would be below sea level, and the 2003 Penrose "Beyond the Plume Hypothesis" meeting would, ironically, have to be held elsewhere. We also know that the seismic velocity beneath Iceland is anomalously low8.

The anomaly may represent hotter mantle, but we are not sure how much hotter; we can, however, be confident that it is more buoyant than the surrounding mantle. The anomaly is approximately cylindrical in form, at least in its upper part, and extends to at least 440 km(ref 8), but it could go much deeper9. To my mind this is a mantle plume, but because the original 1971 description10 of a plume includes phrases such as "manifestations of convection in the lower mantle", we are advised not to call it such11!

What are the alternative models? Gill Foulger's proposal12 that the excess magmatism on Iceland results from melting of subducted Iapetus ocean crust is definitely worth exploring, but three questions in particular need addressing. 1) Why is Iceland so restricted in space, when the source of this subducted ocean crustal material must be inherently large? 2) How does one form high-Mg basalts (picrites) by melting of basaltic crust? 3) Bearing in mind that Iapetus was south of the equator when it was formed and closed, how on Earth do Icelandic volcanoes tap into this source?

The formation of large igneous provinces (LIPs: continental flood basalts and oceanic plateaus)

In these provinces, large volumes of basalt were erupted in a geologically short interval of time. Some provinces persisted for longer than others. Many flood basalt provinces have a succeeding aseismic ridge and currently active volcanic centre (the ‘smoking gun')10. This is entirely consistent with plume models that predict large magmatic flux rates that wax and wane abruptly13, but then persist at lower rates for many millions of years. However, we shouldn't be surprised if these systems do not behave as they are described in textbooks.

We are left with the inescapable fact that the formation of LIPs involves a prodigious amount of energy to satisfy the latent heat of melting of the mantle. A hydrous mantle source would help reduce the melting temperature, but there is scant evidence that the source of LIP basalts is hydrated14. So, if we do rule out any sort of plume involvement, what would be the alternatives? Anderson and co-workers have suggested for several years that continental flood basalts originate from 'edge effects' induced at the margins of thick cratons15. The key evidence is the spatial association between some continental flood basalts and cratonic edges.

Figure 2: A 'snapshot' view of a theoretical situation involving several plumes in varying stages of development distributed around the globe is shown in the figure (right), kindly provided by Paul Tackley of UCLA. This may represent a typical pattern of plume distribution at any period of geological time.

The model proposes that asthenosphere beneath a thick craton is insulated as if by a thermal blanket. It warms as heat convects from depth, and then flows laterally towards cooler regions until it reaches the edge of the craton, where it ascends beneath thinner lithosphere. Here it decompresses and melts. Melting is focussed along the margin, forming a LIP. This model entirely plausible but, unlike the plume model, it has five first order problems in accounting for LIPs.

1. It is difficult to see how the edge effect can explain the abrupt flare-up of LIPs; there is no obvious on-off switch.
2. Such models cannot work for ocean plateaus such as the Ontong Java.
3. Edge-effect models advocate a lithosphere as thin as 50 km in order to operate (any thicker, and melt is not produced in sufficient quantity), which is probably far too thin to be realistic.
4. We do not find LIPs along every rifted craton margin.
5. The high temperatures required by picrites are unlikely to be achieved by edge-effect models. Interestingly, the ?edge effect? espoused by King and Anderson can be accommodated nicely within the plume model, inasmuch as hot plume mantle rising beneath a craton may flow laterally into regions of thinner lithosphere, where it melts16.

There are many other provinces I could have mentioned in support of the plume model; and I deliberately have not marshalled the geochemical arguments. Contrary to what has been said in recent letters in Geoscientist, I do not believe that there is a mafia out to silence the anti-plume lobby. I, for one, welcome the opportunity for this debate.

Few supporters of mantle plume theory push for every volcanic field to be plume-related, just as few supporters of the ophiolite model advocated that every pillow basalt represents an ancient ocean. Yes, the plume model may allow some to hide behind a convenient paradigm, but what is new there?

I find it curious that a scientific theory or model can cause so much anguish among some of my colleagues in the Earth Science community1. Perhaps similar angst was caused in the early days of plate tectonic theory. Many aspects of both plate tectonic and plume models remain imperfectly understood, yet because something doesn't fit our prejudice, we don't reject the entire theory outright; the model is refined to accommodate new data. The terms 'ugly' and 'inelegant', as used by some critics of plume theory2, are highly subjective epithets, easily returned to source, but serve little purpose.

My challenge to the Beyond-the-Plume lobby is this: present us with a viable, workable, and testable alternative hypothesis (or hypotheses) that accommodates the observations and data, and that sits comfortably with our understanding of the Earth's thermal and chemical structure.

This article was improved by comments from Tiffany Barry, Mike Branney, Laurence Coogan, Richard England, Godfrey Fitton, Ian Hill, Pamela Kempton, Andrew Kerr, Mike Norry, Marc Reichow, Steve Temperley, Pat Thompson, and Rosalind White. Defending such an embracing theory in limited space is fraught with difficulties, not least the sins of omission and error. These are mine alone. Andy Saunders.

• Back to the debate

References

Geoscientist 2003, Volume 13, issues 5, 7 and 8.

D.L. Anderson, Geoscientist, 13-8, 16-17 (2003).

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

e.g., A.B. Watts & U.S. ten Brink , J. Geophys. Res., 94, 10,473-10,500, 1989; L. Cserepes et al., Earth Planet. Sci. Letts., 178, 29-38. 2000.

e.g., Y. Ji. & H.-C. Nataf, Earth Planet. Sci. Letts., 159, 99-115, 1998; X. Li et al., Nature, 405, 938-941. 2000; D. Zhao, Earth Planet. Sci. Letts., 192, 251-265, 2003.

J.T. Wilson, Can. J. Phys. 41, 863-870, 1963.

R.K. Staples et al., J. Geophys. Res. 102. 7849-7866. 1997.

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

H. Bijwaard, H. & W. Spakman, Earth Planet. Sci. Letts.166, 121-126, 1999.

W.J. Morgan, Nature, 230, 42-43, 1971.

G.R. Foulger et al., Geophys. J. Int., 146, 504-530, 2001

G. Foulger, Astron. Geophys., 43, 6.19-6.23, 2002.

R.S. White & D. P. McKenzie, J. Geophys. Res. 94: 7685-7729, 1989; I.H. Campbell and R. W. Griffiths, Earth planet. Sci. Lett. 99: 79-93, 1990.

P.J. Michael, G-3, 1(1999GC00025), 1999.

S.D. King and D.L. Anderson. Earth Planet. Sci. Letts., 160, 289-296, 1998.

R.N. Thompson & S.A. Gibson, J. Geol. Soc. Lond 148: 973-977, 1991.