Page 416 - Global Tectonics

398 CHAPTER 12

dimensional spherical convection models of the mantle 200
in which the viscosity of the lower mantle was 30 times Lower mantle
that of the upper mantle. They found that not only was
the wavelength of the resulting convection greater but Discontinuity?
that long linear downwellings formed from the upper km 100
boundary layer; both effects making the pattern of con- D
vection very comparable to that deduced for the mantle.
ULVZ
ULVZ
The convective pattern also had greater temporal
stability. 0
Researchers also have investigated the effect on Fuzzy CMB Core rigidity zone (CRZ)
Outer core
mantle convection of the endothermic phase change at
a depth of 660 km, the base of the transition zone. For
plausible physical characteristics of this phase change Figure 12.13 Cartoon of the D″ layer where it is
the results suggest that it might inhibit but not prevent hotter than its average temperature. These regions
the passage of upwellings and downwellings through it. include an ultra-low-velocity zone (ULVZ), thought
This is consistent with the results of seismic tomogra- to be characterized by partial melt and chemical
phy that indicate that the transition zone has some heterogeneity, chemical and melt scatterers throughout
and, possibly, the points of origin of plumes (redrawn,
effect but that it is not sufficient to impede whole mantle

with permission, from Garnero, 2000. Annual Review of
convection (Montelli et al., 2004b).
Earth and Planetary Sciences, 28. Copyright © 2000,
The chemical heterogeneity of layer D″ (Section
Annual Reviews).
12.8.4) means that it acts as a thermochemical, rather
than a thermal boundary layer. Indeed where it is
hottest it is essentially a thermal boundary layer over a
chemical boundary layer, the ultra-low velocity zone geologic time. Within the supercontinent cycle (Section
(ULVZ). Upwellings of the low viscosity, low density 11.5) there are times when subduction zones are initi-
thermal boundary layer at these points entrain the low ated, as a result of continental break-up, and terminated
viscosity but higher density chemical boundary layer to by continent–continent collision. Such events could
a height of 50–100 km depending on the strength of the initiate changes in the gross pattern of convection in the
upwelling (Fig. 12.13). Analogue experiments (Davaille, mantle and even change the distribution of mass within
1999) indicate that the nature of the upwelling depends the Earth causing a change in the location of the rota-
on the ratio of the stabilizing chemical density anomaly tional axis, that is, the axis about which the moment of
to the destabilizing thermal density anomaly. If this inertia is a maximum. This would be particularly true
is greater than 1, a plume-like upwelling forms; if it if the initial development of subduction zones includes
is approximately 0.5, thermals (broad upwellings or a build-up of subducted material in the transition zone
domes) are produced. In either case the entrainment of that ultimately avalanches down into the lower mantle.
the dense chemical boundary layer is thought to stabi- Such True Polar Wander (Section 5.6) is thought to have
lize the location of the plume or thermal upwelling occurred between 130 and 50 Ma ago (Besse & Cour-
(Jellinek & Manga, 2004). However, as a result of tillot, 2002), a time period bracketed by the break-up of
the greater stability ratio, plumes will tend to be very Pangea, with the initiation of subduction zones, and the
long-lived. collision of India with Eurasia and a major change in
If this general picture of convection in the mantle is the rate of subduction in this zone. The change in direc-
correct the roles of subduction zones and a chemical tion of the Hawaiian–Emperor seamount chain, and the
boundary layer at the base of the mantle are crucial in change in the relative motion between the Pacifi c and
determining the pattern and nature of the convection. Indo-Atlantic hotspot reference frames (Section 5.5)
Indeed it could be argued that the location of subduc- also occurs at the time of the Indian collision, 40–50 Ma
tion zones is most fundamental in that they not only ago. Thus, these too might reflect the consequent

determine downwellings occur but also where the changes in the thermal regime and pattern of convec-
boundary layer at the core–mantle boundary is hottest, tion in the mantle, and, hence, the relative positions of
and hence where upwellings occur. However, subduc- the two major convection cells within the African and
tion zones are transient features within the context of Pacifi c hemispheres.