Page 417 - Global Tectonics

THE MECHANISM OF PLATE TECTONICS 399

12.10 PLUMES point or arête (Jellinek & Manga, 2004). As a result
the upwelling of the thermal boundary layer that pro-
duces these peaks is focused into a narrow, cylindrical
conduit. The temperature difference between this
Certain volcanic hotspots at the Earth’s surface appear plume and the surrounding mantle is probably 200–

to be essentially fixed with respect to the Earth’s deep 300°C, implying more than two orders of magnitude
interior and to provide an absolute reference frame reduction in the viscosity across the boundary layer
for plate motions for the past 40 Ma (Section 5.5). The between them. If partial melt is present in the upwell-
fi xed nature of hotspots such as Hawaii, fi rst suggested ing thermal boundary layer this would also lower the
by Wilson (1963), led Morgan (1971) to propose that viscosity. Partial melt entrained from the ULVZ may
they are located over plumes of mantle material upwell- be required to explain the osmium isotopic ratios in
ing from the lower mantle or even the core–mantle certain hotspot lavas that are thought to indicate that
boundary. The plume hypothesis has been, and con- the source of the osmium is the outer core (Brandon
tinues to be, controversial because it has proven dif- et al., 1998).

ficult to provide unequivocal evidence of such plumes Numerical and analogue models of these hot, low
(Foulger & Natland, 2003). There is now, however, a viscosity plumes, originating in the deep mantle,
growing body of evidence, from both modeling and suggest that plume shape and mobility are controlled
observational data, that some hotspots at the Earth’s by the magnitude of the viscosity contrast with the
surface may be fed by narrow plumes of high tem- surrounding mantle (Kellogg & King, 1997; Lowman
perature, low viscosity material rising from essentially et al., 2004; Lin & van Keken, 2006). As the contrast
fixed points on the core–mantle boundary. There is increases, the plume conduit becomes narrower and

also theoretical and empirical evidence that the source its head becomes broad and mushroom-shaped as hot
material for other hotspots may be derived from much material is able to move upward more effi ciently (Fig.
shallower depths, within the mantle transition zone 12.14). This model, with a mushroom-shaped plume
or the uppermost part of the lower mantle, or even head and a long, thin tail extending to the depth of
from immediately beneath the lithosphere; the latter origin, has achieved widespread application. Neverthe-
being a passive response to various forms of litho- less, numerical models also predict a great variety of
spheric break-up (Anderson, 2000). The suggestion that plume shapes and sizes in cases where density contrasts
there are three types of hotspot, in terms of their due to chemical variations in the lowermost mantle
depth of origin, has been deduced by Courtillot et al. are incorporated into models of plume formation
(2003), mostly by a consideration of the roles of the (Section 12.9) (Farnetani & Samuel, 2005; Lin & van
three potential thermal boundary layers in the mantle. Keken, 2006).
However, there is substantial support for this from the In general, the model of a narrow, mushroom-

results of seismic tomography (Montelli et al., 2004a, shaped plume fits well with the initial expression of
2004b). Hotspots that are underlain by low seismic some hotspots in terms of continental fl ood basalts

velocities in the upper mantle only appear to be limited or oceanic plateaux, reflecting the arrival of the plume
in number. Examples in Fig. 5.7 are Bowie, Cobb, head beneath a thinned lithosphere, and the subsequent
Galapagos, East Australia and, surprisingly perhaps, trace of the hotspot, in the form of a volcanic ridge or
Iceland, although it is underlain by a very large upper line of volcanoes, produced by the tail. Courtillot et al.
mantle anomaly (Montelli et al., 2004b). Iceland is (2003) suggest that these hotspots be termed primary
anomalous also in terms of geochemical indicators hotspots (Section 5.5). They also suggested that the
3
4
of deep mantle origin, notably the ratios of He/ He lifespan of primary hotspots might be approximately
187
186
and Os/ Os in the lavas (Foulger & Pearson, 2001; 140 Ma. Those initiated within the past 100 Ma, such as
Brandon, 2002). Montelli et al. (2004b) describe the Afar and Reunion, are still active; those that are 100 to
tomographic anomaly beneath Yellowstone as being 140 Ma old, that is Louisville and Tristan, might be
virtually nonexistent. failing, and those that formed more than 140 Ma ago,
Laboratory experiments indicate that the peaks that such as Karoo and Siberia, have no active trace. Theo-
develop on the ULVZ in the D″ layer, where it is suf- retical arguments predict that such large plume heads
ficiently dense, form where ridges between embay- and long-lived tails must originate in a thermal bound-

ments in the surface of the ULVZ meet at an elevated ary layer at great depth, presumably layer D″ at the