Abstract
Receiver function analysis is routinely used to isolate P-to-S converted
waves from a complex of earthquake recordings so as to explore crustal
and upper mantle structures and to infer possible geodynamic processes
within the Earth. In the last several years the number of deployments
of portable seismic arrays has been greatly increased. The conventional
receiver function method, which stacks receiver functions at a single
station, is not suitable for such a large amount of data. In this
thesis modifications of the receiver function method have been made.
Techniques of reflection seismology have been introduced into the
receiver function analysis. Modified receiver function method has
been successfully applied to the seismological data acquired in Tibet
and the central Andes. In these two Earth's largest and highest plateaus,
data of many available seismic broadband and short-period experiments
have been collected. In Tibet, data of the INDEPTH II and GEDEPTH
I experiments in southern Tibet and the PASSCAL 91/92 experiment
across the central Tibetan Plateau have been combined. A total number
of about 50 stations were distributed roughly in a NNE directed profile.
More than 900 receiver functions have been obtained. In the Central
Andes, more than 200 stations have been deployed within the experiments
of PISCO, CINCA, ANCORP, PUNA and KDS of the project of the SFB 267,
and the BANJO and SEDA broadband arrays of the PASSCAL experiments.
More than 640 teleseismic receiver functions have been obtained.
Results are summarized in the following. (1) Crustal thicknesses
under the two plateaus are reliably determined by teleseismic receiver
functions. P-to-S converted waves at the Moho are clearly seen under
the Tibetan plateau and under the Central Andean plateau. In southern
Tibet the Moho is 75-80 km deep. In northern Tibet it becomes shallower
to a depth of 55-60 km. In the Central Andes, the continental Moho
is 65-70 km deep beneath the Andean Plateau (it appears to be 15
km shallower beneath Puna than beneath Altiplano). The Moho abruptly
reduces its depth beneath the eastern edge of the Eastern Cordillera
(65-64.5 deg W) and remains at 45-50 km depth in the Subandes. Further
east there is another abrupt reduction of Moho depth between the
Subandes and the Chaco Plain. The Moho is 30-35 km beneath the Chaco
Plain. (2) Evidence of crustal-scale underthrusting is found in Tibet
as well as in the Andes. The INDEPTH data clearly show an intra-crustal
phase at a depth of 50-60 km in southern Tibet. This conversion boundary
is probably the evidence of the underthrust Indian crust. In the
Andean data a more than 300 km west-dipping intra-crustal converter
evidently marks the boundary of the underthrust Brazilian shield
crust. This boundary exists across the entire Altiplano and Puna
plateau from 20 km depth below the Eastern Cordillera to 40 km depth
below the Western Cordillera and the Precordillera. In both plateaus,
most of the thickened crust, if not all, can be attributed to the
crustal-scale underthrusting. (3) Plate boundaries are found to a
depth of about 250 km between the Indian and the Asian lithospheric
mantle and to a depth of about 120 km between the Nazca plate and
the South American plate. However, the nature of these boundaries
is different. In the Central Andes, the plate boundary is interpreted
as the oceanic Moho of the Nazca plate, above which a 10 km layer
of oceanic crust with lower seismic velocity suggests that the gabbroic
rocks do not completely transform to eclogite until a depth of 120
km. Most of the intermediate depth seismicity stops at the same depth,
suggesting a relation with phase transformation. In Tibet the observed
plate boundary of the two lithospheric mantles probably reflects
the temperature difference between the two plates. The cold Indian
mantle is subducted under the warm Asian mantle. The temperature
difference can be as high as 500-700 C resulting in large seismic
velocity contrast. (4) Interesting variations have been found in
the upper mantle discontinuities which are related to the plate collision
and subduction processes. In Tibet, the 410 km discontinuity is clearly
seen in its globally average depth in the south, and it is disturbed
and becomes complicated in the north. The 660 km discontinuity is
continuously displayed throughout the Tibetan profile. Similarly,
in the Central Andes, the 410 km discontinuity is only seen in the
western part of the profile. Near the subducted Nazca plate the 410
km discontinuity is not imaged coherently, which is obviously attributed
to the subduction complexity of the phase transformation of the mantle
rocks. It is interesting to see that the 660 km discontinuity is
depressed by about 30-40 km in the region of the cold Nazca slab,
which corresponds to a temperature reduction of 300-600 C within
the slab.
Links and resources
Tags