Mantle Flow Beneath Pacific Lithosphere Subducted at Kamchatka
If Pacific oceanic lithosphere
slides over an asthenosphere that largely decouples it from underlying
mantle circulation, then what happens to the asthenospheric decoupling
layer at subduction zones? Co-Author Jonathan
Lees and I
have used earthquakes that occurred within the subducted Pacific plate
to characterize the upper mantle flow field beneath the slab. S waves from these earthquakes leave
the slab and become split into orthogonally-polarized fast and slow
waves in the seismically anisotropic upper mantle region below the
slab. They then cross the isotropic region of the lower mantle
and become split again by anisotropy beneath recording stations in
Eurasia and North America. We correct for the known shear wave
splitting at these stations to isolate the splitting from the sub-slab
Figure 1, above, shows
the topography and bathymetry of the Kamchatka
and environs. Pacific oceanic lithosphere subducts beneath the
continental landmass along the Kurile-Kamchatka Trench. Note,
Trench ends abruptly at its northeastern intersection with the Aleutian
Trench. The latter is a strike-slip fault which cuts through the
lithosphere, so the subducted slab actually has a lateral termination.
Figure 2. Seismicity of the
Kamchatka subduction zone, 1963-1995 NEIC,
atop topography and free-air marine gravity anomalies [Sandwell and
Smith, 1997]. Earthquake depths as per key, lower
right. Isodepth contours
seismicity constructed by us also shown, depths as labeled. Heavy
Kamchatka and Aleutians morphological trenches.
Figure 3. Results of shear-wave splitting measurements at
the SEKS network
stations [Peyton et al., 2001]. Squares
are station locations. Bars parallel to
polarization directions of SK(K)S and PKS phases,
length of bars scaled
splitting delay times, see key uppe left. Note 90° change in
trends from trench-parallel along most of Kamchatka Peninsula
between 55-56°N, interpreted by Peyton
et al.  as
around the edge of the Pacific lithosphere subducting beneath
Schematic of source-receiver ray paths used in this study.
the Kamchatka slab are recorded at distant stations where
upper mantle shear wave
splitting parameters are known. Corrections for
shear-wave splitting on the receiver
side of the ray paths allow isolation of splitting
engendered in the region beneath the
Kamchatka slab. Note that the upper mantle wedge
above the slab is not sampled at
all by these waves.
Figure 5a. Seismogram of S wave from
event 97136 recorded at station KEV,
Finland. S wave used in
our measurement (gray boxed areas) is clearly isolated from
expected arrival times of other shear phases, marked by vertical dashed
two traces rotated into calculated fast and slow reference frame.
Bottom two traces
show effect of linearization via energy minimization of component
minimum eigenvalue of the polarization matrix.
Figure 5b. Waveform coherence for split
\fIS\fR of event 97136 shown
in Fig. 5a. Good correspondence of fast and slow waves shown in
fast-slow frame (top left) and with delay corrected (top right).
particle motion diagnostic of shear-wave splitting seen in bottom left
plot, corresponding to waves shown in top left. Linear particle
after correction for splitting shown in bottom right.
Figure 5c. Result
of shear-wave splitting parameter grid search over
180° azimuth and 0-4 s delay time. Contours of energy on the
seismogram corresponding to the minimum eigenvalue component of
the polarization matrix shown. Star indicates best splitting
fast polarization azimuth and delay time.
Figure 6. Shear-wave splitting
results (red bars) for our study. Results of
Peyton et al.  also shown
(blue bars) for comparison. Note the delay time
the two data sets are different: for source-side S wave splitting
bottom right; for SK(K)S
splitting key is at top left.
Figure 7. Polar histogram of source-side shear-wave
splitting fast trends.
Note predominant trends are parallel to local Kamchatka trench strike
Figure 8. Null splitting
results from our study. These represent fast anisotropy
in one of either two
orthogonal directions, as shown, or possibly no anisotropy
Figure 9. Schematic block
diagram of upper mantle flow beneath Pacific
lithosphere subducting at Kamchatka.
Pacific plate lithosphere shown in
light blue. Orange arrow shows upper mantle flow around the slab
View is from the NE looking SW.
Figure 10. Vertical section across Kamchatka slab edge
168.0°E to 57.5°N, 159.0°E. Seismicity within 100 km
of the section
line plotted as small black squares. Large star is location of
98027, and large triangle marks location of SEKS network station on
SK(K)S and S
waves ray paths shown schematically
as dashed lines from station and 98027 hypocenter. Splitting along
these paths is not trench parallel, as it is beneath Kamchatka proper,
hence inferred anisotropy differs between the locales: double
arrows parallel to Pacific plate motion indicate anisotropy within the
unsubducted Pacific asthenospheric channel; vector arrow heads
(bulls-eyes) indicate trench-parallel anisotropy and flow beneath Kam-
chatka. The width of the trench-parallel flow field is limited by
Island and 98027 measurements. Assuming the asthenospheric channel
and its underlying mesosphere maintain their layered form into the sub-
duction zone, the maximum width of the trench-parallel flow field is
This Project has been supported by the Geophysics Program of
National Science Foundation We
also gratefully acknowledge data and support from the IRIS Consortium.