Geoscience Reference
In-Depth Information
form magmatic and hydrothermal arc. The man-
tle wedge passing through this region may carry
at least several 1000 ppm H
2
ObyNAMs(and
phase A depending on the geothermal gradient)
to the transition zone, where the hydrous plumes
can occur to cause backarc and intraplate mag-
matism. The HBL transport H
2
O continuously
from the shallow depth to the transition zone,
during which various ascent mechanisms as in
Figure 13.3 operate to redistribute H
2
O within
the upper and transition zone mantle. In the next
section we discuss the water transport when the
slab penetrates into the lower mantle.
Figure 13.6 and Figure 13.7 show snapshots for the
case when
C
max
H
2
O
of the lower mantle is 0.21 wt %
(Murakami
et al
., 2002) and 0.001 wt % (Bolfan-
Casanova, 2005), respectively. The water trans-
port is incorporated into the model developed by
Nakakuki
et al
. (2010). The details are presented
in the caption of Figure 13.6. This model simu-
lates dynamic processes from slab subduction to
penetration into the lower mantle including the
buoyancy and rheological effects of olivine-series
phase transitions at 410-km and 660-km depths.
The model naturally reproduces plate-like litho-
spheric motion, and stagnation, avalanche and
retreat of the slab, without imposing the veloc-
ity at the surface or within the mantle. With
these parameters, including those related to the
phase transitions, the slab stagnates beneath the
660-km phase transition in a short duration, so
that the slab stagnation does not significantly in-
fluence the water transport in the models to be
presented in this section. We can, therefore, dis-
cuss the effect of the slab penetration into the
lower mantle on the water transport. The effects
of the slab stagnation above the 660-km phase
boundary for longer duration (e.g., due to the
steeper Clapeyron slope) have been discussed in
the former section.
We here consider only one-way effects of the
solid mantle flow on water transport. Dynamic
feedback from the water to the solid mantle con-
vection such as reduction of the density and
the viscosity is neglected. In this model, aque-
ous fluid, which is generated when the water
content exceeds
C
max
13.3.2 Slab penetration and water transport
to the lower mantle
The
C
max
H
2
O
value in the lower mantle is subject
to extensive debate and its estimate ranges from
10 ppm to over 1000 ppm depending partly on
the analytical methods and interpretations (see
Bolfan-Casanova (2005) for the details). One of
the explanations for the variable estimates is that
the major lower mantle minerals (perovskite and
magnesiow ustite) are essentially dry, and a fluid
remained at the grain boundary or fluid inclu-
sions in the crystals may be practically important
involving as high as several 1000 ppm H
2
O. Even
when the lower estimate is the case (i.e.,
C
max
H
2
O
in the major anhydrous minerals of the lower
mantle is
10 ppm), hydrous phases such as
phase D within the core part of cold slabs, if
they exist, may transport some water into the
lower mantle. After the breakdown of phase D
at
∼
H
2
O
shown in Figure 13.1, is
assumed to vertically migrate upwards and hy-
drate the overlying portion instantaneously. This
is an approximation for much faster transport
due to permeable flow or crack generation than
the solid mantle motion. The models to be pre-
sented here, therefore, cannot simulate the effects
of slow water transport comparable to the man-
tle flow due to small permeability or upwelling
plume formation from the buoyant hydrated layer
above the slab (Gerya & Yuen, 2003; Richard &
Bercovici, 2009). As a boundary condition, the
water is introduced into 6-km thick basaltic layer
that is placed on the surface of the subducting
1200 to 1500 km depth, water or hydrogen
may be transported as far as the bottom of the
lower mantle by reacting with metallic iron in
the lower mantle to form hydrous phases or iron
hydride (Ohtani, 2005). In the following, we show
how these variable estimates affect the water
transport to the lower mantle based on numerical
simulation when the subducting slab penetrates
into the lower mantle.
We have constructed numerical models for
water transport from the trench to the lower
mantle, with a special attention to redistribution
of H
2
O when the HBL enters the lower mantle.
∼
