Page 415 - Global Tectonics

THE MECHANISM OF PLATE TECTONICS 397

the core–mantle boundary. As Davies (1993) has aptly being entirely passive. Meguin & Romanowicz (2000)
put it, the plate mode is crucial in cooling the mantle, and Montelli et al. (2004b) note that there is evidence
by the creation of oceanic lithosphere, and the plume in their mantle tomographic models for lateral fl ow in
mode releases heat from the core. The heat released by the upper mantle from the African upwelling to the
the plate mode is thought to be much greater than that Atlantic and Indian Ocean ridges, and from the Pacifi c

released from the core as the mantle is heated internally upwelling to the East Pacific Rise. If so this would com-

by radioactivity. One might expect therefore that plete the elusive route of the return flow from sub-
the plate mode is dominant. These two very different duction zones to mid-ocean ridges, or at least provide
modes of convection need not necessarily be strongly one such route.
coupled. However it is noteworthy that the two major The scale, or wavelength, of this gross pattern of
upwellings at the present day, beneath southern Africa convection in the mantle is greater than that predicted

and the south central Pacific, are at the centers of the by analogue experiments and early numerical models
6
expanding ring of subduction zones around what was assuming a Rayleigh number greater than 10 . It tran-
Gondwana and the contracting ring of subduction spires that this is because these models assumed uniform
zones around the Pacific respectively, and hence distant viscosity throughout the convecting layer. In the Earth’s

from the cooling effect of the subducting slabs that mantle the viscosity varies with both temperature
appear to extend to the core–mantle boundary ((Plate and pressure. For the relevant temperature gradient in
12.2 (between pp. 244 and 245), Fig. 12.12). It is also the mantle the effect of increasing pressure with depth
striking that these two active upwellings do not corre- almost certainly means that the viscosity of the

spond directly to mid-ocean ridges. This is consistent lower mantle is significantly greater than that of the
with the interpretation of the upwelling beneath ridges upper mantle. Bunge et al. (1997) investigated three-

E Pacific

MOR
MOR

African
superswell
Pacific
superswell Core

MOR

660 km
discontinuity
W Pacific

Figure 12.12 Cartoon showing an approximately equatorial section through the Earth and illustrating the
possible relationship of subduction zones, superswells, plumes, and mid-ocean ridges (MOR) to the gross pattern of
circulation in the mantle. Note that deep-seated or primary plumes, such as Afar, Reunion, Tristan, Hawaii, Easter, and

Louisville, are peripheral to the superswells, and that secondary plumes are common above the Pacific superswell. The
mid-ocean ridges are a passive response to the plate separation and not systematically related to the main convective
pattern.