Geology Reference
In-Depth Information
It turns out that plumes are also a significant source of heterogeneities. The
mass flow of plumes is less than that of subducted lithosphere, but only by a
factor of about 3. The rate of subduction of sea floor is 3 km
2
/yr (Chapter 6).
The strongly differentiated part of the lithosphere is about 60 km deep, so the
volume of heterogeneous lithosphere added per year to the MORB source is around
180 km
3
/yr. On the other hand, the volumetric flow rate of the Hawaiian plume is
estimated to be about 7.5 km
3
/yr (Chapter 7) and the Hawaiian plume carries about
10% of the global plume flow, so the global volumetric flow rate of plumes into
the upper mantle is about 75 km
3
/yr, around 40% of the plate flow, or 30% of the
combined flows into the MORB source.
Most of the plume heterogeneity will not be removed by melting, because only
10-20% of the plume material actually melts (Chapter 7), and the rest will be
stirred into the mantle [1]. Thus we should expect around one-quarter of mantle
heterogeneity to derive from plumes. These heterogeneities may ultimately derive
from subducted lithosphere, but they may also be modified in important ways prior
to their incorporation into a plume, as we will see.
10.4.2 Survival of heterogeneities
The early intuition of many geochemists was that mantle convection would keep the
mantle homogenised, and this led many to assume that the observed heterogeneity
required distinct layers to preserve chemical differences for long periods. This
intuition was probably fed by experience with fluid mixing in daily life, where, for
example, milk added to coffee can be homogenised with a quick stir of a spoon.
However, this experience is quite misleading, because the liquid in a coffee cup is in
a quite different flow regime from the mantle. The key difference is that water has
a sufficiently low viscosity that the momentum of moving fluid sustains its motion
and, even more important, generates smaller-scale eddies spontaneously. There is
a cascade of flow into smaller and smaller eddies, and the eddies of different scales
combine to mix the fluids very efficiently. This regime of flow is called turbulence.
In contrast, the mantle has such a high viscosity that momentum is completely
negligible and there is no turbulence. This means that the smaller-scale eddies that
so efficiently homogenise coffee are absent in the mantle. Homogenisation can
occur only by the slow shearing and stretching caused by the large-scale flow, and
this takes many orders of magnitude longer. Try stirring milk into honey and see
how much longer it takes. This regime of flow is called laminar flow.
Quite apart from this fundamental distinction, there has been some debate among
geophysicists about whether mantle homogenisation might take hundreds of mil-
lions of years or (many) billions of years. The rate of stirring depends on details of
the flow, such as whether there are flows at scales smaller than the plates, whether
