Xq. Li et J. Nabelek, DETECTING SLOW, LONG-DURATION SLIP OF LARGE EARTHQUAKES USING VERY LONG-PERIOD ORBITAL SURFACE-WAVES, Geophysical journal international, 124(2), 1996, pp. 483-501
In this paper, we present a simple yet effective technique for measuri
ng the source time-delay spectrum from very long-period world-circling
surface waves without relying on assumed earth models or calculation
of synthetic seismograms. By using auto-correlation and self-convoluti
on functions derived from whole seismograms of large earthquakes recor
ded at a seismic station it is possible to isolate the source effects
and precisely (+/-5 s) determine the source delay spectrum at periods
as long as 600 s. Measurements at such long periods are necessary for
detecting slow precursory slip and slow creep after the main 'fast' se
ismic phase of large earthquakes. We apply this technique to five rece
nt large earthquakes: the 1989 Macquarie Ridge (M(W), 8.2), the 1992 N
icaragua (M(W) 7.6), the 1993 Mariana Islands (M(W) 7.8), and the 1994
deep events in Fiji (M(W) 7.6) and Bolivia (M(W) 8.3). Our results ar
e largely in agreement with the results obtained by other investigator
s. The Mariana Islands, Fiji, and Bolivia earthquakes had a normal rup
ture duration. The Nicaragua earthquake exhibits a source time-delay s
pectrum with the delay increasing from 31 s at the 100 s period to 50
s at the 600 s period, indicating an unusually slow rupture process. T
he Macquarie Ridge earthquake is anomalous by the discrepancy between
the Love-wave and Rayleigh-wave estimates. Love waves, in the period b
and of 150 to 600 s, show a normal, flat source-delay spectrum with a
delay of about 28 s. The Rayleigh-wave spectrum agrees with the Love-w
ave spectrum for periods smaller than 200 s and greater than 400 s; ho
wever, it shows a significantly shorter source delay (about 15 s) in t
he 250 to 350 s period range. The unusual nature of the Macquarie Ridg
e earthquake's surface wave spectrum was pointed out by Ihmle, Harabag
lia & Jordan (1993) who also provided an explanation in terms of a slo
w, precursory moment release. This explanation is not consistent with
the flat Love-wave spectrum determined here.