Geoscience Reference
In-Depth Information
slows the heating of the slab and cools the overlying wedge. A total value of about
2.5
10
5
Jkg
−
1
for all the dehydration reactions is probably not unrealistic. None
of these reactions takes place instantaneously (this point is discussed further in
Section 10.3.310.3.5). At low temperatures and pressures, the reactions proceed
more slowly than they do at higher temperatures and pressure; furthermore, under
very dry conditions, at great depth, transformation is also slow.
The released water moves out of the oceanic crust into the overlying upper
mantle. Dehydration probably begins with the onset of subduction but, at shallow
depths, has little effect on the overlying wedge of the overriding plate except
to stream fluid through it and metamorphose it. Dehydration continues as the
plate is subducted: a plate being subducted at an angle of 20
◦
at 8 cm yr
−
1
for
1Ma descends 27 km, with attendant heating, compression and metamorphic
dehydration. Figure10.6(b) indicates that dehydration of serpentine starts when
the top of the slab reaches a depth of roughly 70 km. At such depths, but shallower
than 100 km, water released from the slab is probably fixed as amphibole in the
overlying mantle wedge, or streams upwards if the wedge is fully hydrated; but, at
a depth of 100 km, the melting of wet overlying mantle begins. In the descending
slab, the wet solidus for basalt probably can be reached only at a depth of about
100-150 km, so melting of the subducted oceanic crust normally cannot take
place much shallower than this (but see the end of this section regarding some
special circumstances). The precise depth depends on such factors as the age (and
hence temperature and degree of hydration) of the slab and the angle and rate of
subduction. In reality, probably only the melting of the overlying wedge occurs
at 100 km; melting of the slab may not take place until it is much deeper, because
by this stage the slab must be highly dehydrated.
Since the released water is unlikely to be able to leave the subducting plate
and enter the overlying mantle wedge by porous flow, there must be some other
mechanism. A high pore pressure in this non-percolating water would act as a
lubricant, reducing friction on the subduction zone. This would facilitate slip and
allow intermediate-depth earthquakes to occur. A large earthquake could connect
sufficient water along the fault plane to initiate a hydrofracture. This hydrofracture
would then transport the water into the overlying mantle wedge, where it would
initiate partial melting. If there is down-dip tension in the downgoing slab, the
hydrofractures will propagate upwards perpendicular to the slab (perpendicular
to the least compressive stress). Beneath Japan intermediate-depth earthquakes
occur right at the top of the subducting Pacific plate and are due to brittle ruptures.
Figure 9.48 shows clearly that the low-velocity zones within the mantle wedge
beneath Japan are inclined within the wedge and that the subducting plate and the
volcanic arc are not directly connected vertically by low-velocity wet/molten/hot
material. The velocity structures imply strongly that the paths taken by the rising
volatiles and melts are along the shear zones in the mantle wedge, i.e., along the
flow-lines, which, in the lower part of the wedge, are parallel to the top of the
subducting plate. Both volatiles and melt take the easiest route to the surface,
which is not necessarily the shortest route.
×
