全球P波波速三維構造之多重尺度層析成像(2/2)
Date Issued
2004
Date
2004
Author(s)
DOI
922611M002021
Abstract
It has already been 19 years since the first global
seismic traveltime tomography using the short period
teleseismic P-arrivals. Significant contributions made
upon improving our knowledge of the deep interior of
the Earth have been made. In recent years, due to the
fast accumulation of traveltime data of improved
quality, the continuously growing computing-power
and the development of inversion techniques, more
and more global models of P velocity heterogeneity,
each with their own nominal resolution, have been
reported from different research groups. However,
there are still significant controversies among these
published models in terms of highlighting the
structure of the deep interior of the Earth. These
ambiguities might be attributed to either different
adopted data sets or different schemes of
parameterization, regularization and inversion
algorithms. We have established communication
with Dr. Engdahl of the United States Geological
Survey (USGS), and obtain and compiled the updated
travel time data since Engdahl et al. (1998). They
have relocated hundred-thousands of events that are
well-constrained teleseismically by arrival-time data
reported to the International Seismological Centre
(ISC) and to the U.S. Geological Survey’s National
Earthquake Information Center (NEIC). We believe
that by adopting the same data set, it is a good
opportunity to clearly appraise the implications of the
inverted P-velocity model obtained from different
inversion techniques. Up to now, we have succeeded,
within this research, developing the first fully 3-D,
linear spherical wavelet bases that significantly
improves upon the discontinuity problem embedded
within the original spherical Haar basis developed by
our previous work (along with the derivation of the
3D Gaussian sampling and weighting functions
needed for the high order approximation of the
numerical integration that builds the Gram matrix); as
well as building the Empirical TravelTimes (ETT)
directly from the observed traveltime data and
constitute a better set of summary rays that both
reduces the total amount of observations and improves
the sampling geometry without sacrificing useful
information embedded within the raw data.
1
Preliminary inversion performed upon the subset built
from the 1998-2001 data has revealed results similar
to tomographic models built by other leading groups.
That is, we have very good correlation between the
P-velocity structure and the dynamics implied from
the surface tectonics such as the distribution of ridges
and subducted slabs within the top of upper mantle
(above about 400 km depth); and the pronounced
large-scale 2 L , lowermost mantle structure. It is
unfortunate that our current setup cannot handle such
humongous size of the Engdahl data set (over millions
of rays for just the 1998-2001 subset) even after we
have cleverly go through the summary ray processing
by the previously mention, new EET scheme. To
overcome this difficulty is certainly the most
important focus of our future work such that the
multiscale tomography that has been developed by our
group that is the only inversion scheme naturally
equipped with a non-stationary regularization scheme
and may potentially pushes the resolution to the upper
limit of what the data constraints might impose. We
will continue our research in this respect.
Subjects
Mantle P-heterogeneity
Global teleseismic earthquake relocation
Empirical traveltimes
Multiscale seismic tomography
Publisher
臺北市：國立臺灣大學海洋研究所
Type
report
File(s)![Thumbnail Image]()
Loading...
Name
922611M002021.pdf
Size
221.88 KB
Format
Adobe PDF
Checksum
(MD5):c60e97a7a2edf5c4485668e763e503f1