Geoscience Reference
In-Depth Information
expected in cold regions and a thinner transition
zone in hot regions. It would also be expected
that the topographies of the 410 and 660 km
discontinuities are anti-correlated. It is difficult
to estimate the temperature of the transition
zone, as seismic tomography only provides us
with information on the local seismic velocity
perturbation. Many studies use seismic velocity
perturbations as a proxy for temperature pertur-
bations, assuming that high velocity corresponds
to low temperature and low velocity with high
temperature. This idea ignores compositional
heterogeneity in the transition zone, which may
also play a role. But as the amount of compo-
sitional heterogeneity in the transition zone is
currently unknown, using velocity perturbations
is a good starting point.
Lebedev
et al
. (2002) showed that a correla-
tion between the topography of the '410' and
the '660' and the predictions for the Clapeyron
slopes of olivine phase transitions indeed exists
for a region in East Asia and Australia, using Pds
receiver functions and local shear wave tomogra-
phy. This result is not surprising, as the SS pre-
cursor and receiver function maps in Figure 10.7
also show a thickened transition zone in these
regions. Such a correlation has also be seen on
a global scale (Lawrence & Shearer, 2006b) using
Pds receiver functions, and some correlation has
been seen in global SS precursor studies (Houser
et al
., 2008). However, the additional phase tran-
sitions in garnet around 660 km depth will change
the temperature dependence of the discontinuity
depth, especially at high temperatures. SS precur-
sors show a slight correlation between the 410 and
660 km discontinuity topography which might be
explained by the garnet transition being domi-
nant at higher temperature (Houser & Williams,
2010). Ritsema
et al
. (2009) used transition thick-
ness estimates from SS precursors to estimate
transition zone temperature and found variations
of 200K at over 1000 km wavelengths regionally.
They also found that the average mantle poten-
tial temperature is better fit by a mechanically
mixed mantle of basalt and harzburgite than by a
homogeneous mantle.
10.4 Upper Mantle
Inspecting Figure 10.4 again, we find that we
can see much more than just the transition zone
discontinuities. Especially the Lehmann discon-
tinuity at 220 km depth can be seen in the SS
precursors and Pds receiver functions, but with
a much smaller number of observations than the
transition zone discontinuities suggesting that it
is not a global reflector. There is also a suggestion
for additional reflectors from around 300-350 km
depth, seen in all three data types. Here, we will
discuss the seismic observations of these addi-
tional upper mantle reflectors.
10.4.1 Lithosphere-asthenosphere boundary
The lithosphere-asthenosphere boundary (also
called LAB) is not defined as either a phase tran-
sition or a compositional boundary in terms of
major elements. In fact, it is a change in thermal
and mechanical properties from the lithosphere
which only transports heat by conduction, to
the asthenosphere which takes part in mantle
convection. A range of data types have been used
to observe the LAB. Li
et al
. (2007) used Sdp
receiver functions to study the LAB under the
United States and found an average depth (also
called lithosphere thickness) of 70 km and that
the discontinuity was relatively sharp, occurring
over less than 20 km. Rychert and Shearer (2009)
used Pds receiver functions for a global study of
the LAB, finding a depth of 80 km in tectonically
active regions, 95 km in shields and 70 km in
ocean islands. They realize that their value for
shields of 95 km is quite small and suggest that
they might be imaging a different boundary than
the LAB in these regions.
In a different approach, Priestley and McKenzie
(2006) used surface waves, in particular funda-
mental and higher mode Rayleigh waves, which
are sensitive to S wave velocity. They assume
that S wave velocity is mainly controlled by tem-
perature, and use this to find the lithospheric
thickness. With thismethod, themaximum litho-
sphere depth is most often much larger that the
