Roadmap | The review process | Home

Discussion of

"On alternative models for the origin of time-progressive volcanic chains"


V. Puchkov

23rd May, 2008, Carlo Doglioni
I think this is a helpful piece of work. However, the model by myself and Cuffaro (Doglioni et al. (2005)) has been misinterpreted. None of the world's plate boundaries are fixed relative to the mantle, and this can be easily demonstrated kinematically. The so-called plume trails argued to be moving against the mantle flow (implicit in the net rotation of the lithosphere required by any reference frame and supported by the asymmetry of orogens and rift zones) are volcanic trails that are stationary and close to a ridge or transtensional area. These have been interpreted as shallow, unfixed wet spots. Those geometries are in fact on both sides of the ridge (therefore they do not represent any preferred kinematics), and they neither represent nor contradict the evidence for mantle flow. Otherwise they are along the 90°E ridge and the Laccadive transtensional boundary.

A model for this type of volcanism is proposed by Doglioni et al. (2005), Figure 9. In this model the volcanism does not represent the mantle motion relative to the deep mantle. In the paper by Gripp & Gordon (2002) there is a description of the net rotation in HS3 where, for example, Africa is moving "W-ward", contrary to the volcanic trails in the eastern Atlantic.

25th May, 2008, Victor Puchkov
I did not try to interpret the model of Cuffaro & Doglioni (2007) and therefore I could not misinterpet it. My intention was to see what are the consequences of this model because they are the best test for any theory. The direction of the mantle flow in this model changes between NE and SE, and directions of relative lithosperic plate motions in most of cases are opposite to the mantle flow, changing between NW and SW. The melt anomalies are entrained by the mantle flow and move in the same direction as the mantle flow, but more slowly. If this is correct, the time progression of the volcanic trails must always be directed in the NW-SW direction. Unfortunately such a conclusion conflicts with my map. Thus, something is wrong either with the model of Cuffaro & Doglioni (2007) or with the map I compiled from many independent sources.

The way in which of Cuffaro & Doglioni (2007) explain the cases which contradict there model is merely inventing new, additional models. For example, the model suggested ad hoc for the Southern Atlantic (Doglioni et al., 2005; Figure 9) implies that the hot (or wet) spots sit under the MAR and produce symmetric volcanic ridges. But again, this is not the case according to my information and map: neither Tristan, nor Helena sit on MAR and they did not produce symmetrical patterns.

As far as I understand, yet another ad hoc model is suggested by Cuffaro & Doglioni (2007) for the Reunion and Kerguelen hotspots, implying a transtensional nature for the 90°E and Laccadive tectonic lines. But there are still the Easter, Galapagos, Madeira and Canaries to explain, with E- and NE-ward time-progressions

On the other hand, I did not need any additional models, easily correlating my map with the ITRF-2005 scheme, and exactly following the classical Wilson and Morgan explanation.

26th May, 2008, Carlo Doglioni
I have been aware of these points since I first saw a global asymmetry along an undulated flow pattern of plate motion. I was puzzled as to why there was so much evidence for global tuning, and about the contrasting kinematics suggested by some volcanic trails. The asymmetries across rifts, the opposing characters of orogens, such as the dip of the foreland monocline, the subsidence in the foredeep, the elevation of orogens regardless of the nature and age of the subducting lithosphere, their metamorphic evolution, the types of rocks involved, the depth of the basal decollement, the gravimetric signature, the heat flow signature, the kinematics of the subduction hinge, and a number of other different markers are all verifiable data in favor of a polarized flow (Doglioni et al., 1999; 2007). That is why I studied the origin of plumes and I concluded that volcanic tracks may have different origins (wet spots, shear heating, rifts, transtensional shears, etc.) but have in common a shallow source (e.g., Smith & Lewis, 1999). So our model is not ad hoc, but based on verifiable data. This evidence has to be disproved before our model of global tuning can be rejected.

The existence of a flow of plates was proposed by myself in 1990 based on the first-order tectonic features, but it has been independently confirmed by space geodesy (e.g., Crespi et al., 2007). The westward drift is added to this flow, but it is an independent observation and also not an artifact. The net rotation of the lithosphere is based on a number of independent types of evidence. We can question what causes it, and its real velocity, but not its existence. In Scoppola et al. (2006) you may find a review and a new model on the subject. The westward drift has been envisaged and/or measured, and discussed by many researchers, including Wegener, Rittmann, Le Pichon, Bostrom, Jordan, Knopoff, Moore, Wang, Gordon, Ricard, Sabadini and myself. The ITRF is artificially anchored to the Earth's center of mass, and it is fixed a priori without any net rotation (the sum of all plate motion vectors is constrained to be zero), disregarding what the lithosphere is doing relative to the mantle (see papers by Argus on that subject). Therefore it is not representative of any "absolute" plate motion.

Regarding the concern about volcanoes which Prof. Puchkov infers to be fed by the deep mantle, the paper by Bonatti (1990) on wet spots is particularly relevant. These do not need to be exactly on the ridge. Rifting – "fissures" – have also been invoked, e.g., by Natland & Winterer, 2005).

Figure 1 shows water variations along the mid-Atlantic ridge. There is a good correlation with the so-called hot spots. As you know, the asthenosphere is richer in water, which lowers the melting temperature. Therefore the so-called hot spots, when they are in truth wet spots, are actually cool spots! I am not saying that all plume trails are wet spots. I do not think Hawaii is a wet spot. In my opinion, the Hawaiian track is really an intraplate plume. But nevertheless thermo-barometry and geochemistry support an asthenospheric source depth.

I reiterate that:

  1. the lithosphere is decoupled from the asthenosphere and the deep mantle (see shear wave splitting for example), and
  2. if the source of plume trails comes from within or even above the decoupling surface, the trails of age-progressive volcanism do not represent the plate motion relative to the mantle, especially when they are on or close to plate boundaries (including transform zones) which all move relative to the mantle.

Again, note the motion of Africa in HS3 with respect to the eastern Atlantic trails calculated by Prof. Puchkov.

The arguments proposed to disprove the net rotation of the lithosphere with "westerly" moving plume tracks are, from my point of view, indirect but strong evidence for the shallow nature of plumes.

Figure 1. Variations in water and sodium along the mid-Atlantic ridge, from Ligi et al. (2005).

27th May, 2008, Don Anderson
I agree with the comments of Carlo Doglioni except for a few semantic issues. It is not helpful to refer to all melting anomalies as "plumes" since this term has a fluid dynamic and conventional usage meaning. The alternatives to mantle plumes include shallow features such as wetspots, coldspots or fertile-spots. Mantle metasomatism even makes CO2- or carbonatitic-spots. "Mantle plumes" must be hot and anchored in a deep thermal boundary layer.

Fertile blobs on the other hand can melt at ambient mantle temperature and can originate as delaminated arc crust, from above. Even Hawaii need not be hotter than ambient mantle under a large long-lived plate. Fertile blobs can also originate as subducted seamount chains or temporarily subducted young buoyant oceanic crust; there is a lot of such crust and many seamount chains are currently entering subduction zones. The mass flux of these exceeds the hotspot flux. The central Pacific, under Hawaii, is more likely to be fueled by these kinds of fertile blobs than by delaminated supercontinent crust that seems to fuel the Indian and south Atlantic ocean fertility spots.

In the asthenospheric counterflow models, the hotspot motions on a given plate will be approximately parallel. There is no agreement on what to call these various alternatives to mantle plumes but it is confusing to call them "plumes" since they have none of the attributes of the Wilson-Morgan-Campbell plumes. They may, in fact reflect an entirely different form of convection, more akin to salt fingers, dikes, Galileo thermometers, and diapirs than to the boundary layer convection, heated from below, that motivated the plume hypothesis. The actual form of convection involved in melting anomalies is driven by internal heating and phase changes, not heating from below combined with thermal expansion. "Hotspots" are actually due to high homologous temperature rather than to high absolute temperature. The results are essentially the same for seismology and petrology except that predicted temperatures and heatflows are much lower.


last updated 26th May, 2008