 |
(tel) 617.253.8027 (fax) 617.253.6385 |
|
(tel) 617.253.8027 (fax) 617.253.6385 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
Determination of Earthquake Source Parameters From Inversion of Body Waves
John L. Nabelek
Submitted
to the Department of Earth, Atmospheric, and Planetary Sciences on January
24, 1984 in partial fulfillment of the requirements for the degree of Doctor
of Philosophy
Abstract
Earthquakes
provide us with important information about ongoing tectonic processes. In
the past, only the arrival time data and rough measurements of the amplitudes
of the recorded seismograms were used in the analysis of large numbers of
earthquakes using formal inversion techniques. Detailed studies utilizing
the information contained in the full seismograms were limited in number because
of the time consuming trial-and-error techniques involved; only recently have
comparable automatic techniques begun to emerge. An increase in the global
coverage by digital networks has certainly provided an impetus for such endeavors.
This thesis develops a technique for extracting average (point source) earthquake
properties from the teleseismic body wave data. The technique enables us to
obtain precise estimates of the source strength and mechanism and its depth
and time history for earthquakes of magnitude 4.5 and higher. For larger events,
it also provides an estimate of the location of the measure of the fault dimensions
and the direct rupture. Parameterization in terms of line source is also presented.
For complex events, multiple sets of parameters can be estimated.
Numerical tests indicate that for simple events and typical station coverage
the accuracy of the depth determination using long—period data is approximately
1 KM. The source orientation can be determined to within 1-3° and the
seismic moment to within a few percent. For small events, the accuracy in
depth determination can be further improved by using short—period data.
Numerical tests also show that inversion using source parameterizations which
ignore the effects of source duration give biased depth estimates.
The technique was applied to four earthquakes of different size and complexity:
the 1975, Lice, Turkey earthquake; the 1982, New Brunswick, Canada earthquake
and its main aftershock; and the 1981 El Asnam, Algeria earthquake. Two of
these events, Lice and El Asnam, broke the surface of the earth and were studied
extensively in the field. Parameters such as fault length, displacement and
slip direction (at least on the surface), are therefore well established.
The results of the inversions are consistent with these observations.
The New Brunswick earthquake was studied in detail using both long—
and short—period data. It was found to be associated with a stress drop
of approximately 960 bars. In this study attenuation of P waves at ~l Hz for
paths from New Brunswick was also investigated. The strongest attenuation
(t*+0.6—0.8s) was observed for South American stations.
The inversion techniques were also applied to two underground nuclear explosions:
the August 18, 1979 Kazakhstan, USSR event and the March 26, 1970 (Handley),
N.T.S., event. It was shown that the teleseismic body waves can resolve the
isotropic component of the moment tensor. Deconvolution of the instrument
response from the short—period records also helped greatly in stabilizing
the inversion. In general, the P—wave travel paths from Kazakhstan are
characterized by very low attenuation with values of t* as low as 0.3s for
paths to stations CHTO (Thailand), BCAO (Central Africa), and ANTO (Turkey).
The largest t* value (1 s) was observed for station ANMO (New Mexico).
