Comparing seismograms

Since 1966, seismological instrumentation has changed. Early earthquakes were recorded on paper or perhaps on film. Now, we use digital recordings from broadband seismometers. These modern systems have many advantages: we know the transfer function of the instrument and can recover the true ground motion; the dynamic range is large, so we can analyse both small and large signals; and the frequency band of observations is wide, so we can observe both high and low frequency signals. And finally, on September 28, 2004, another magnitude 6 earthquake occurred at Parkfield. Unlike the events in 1922, 1934 and 1966, this earthquake started well to the SE of the town of Parkfield, and ruptured NW. The rupture stopped about where the 1966 event had initiated (Langbein et al, 2005).

Fortunately, modern seismometers have been installed at the stations which Bakun and McEvilly (1979, 1984) used. We simulated the waveforms which would have been recorded for the 2004 Parkfield earthquake if old seismographs had still been running at the regional stations and at DBN.

At the regional stations, there are clear differences in the waveforms. These are probably related to the differences in the hypocenter and rupture direction of the 2004 event as compared to the earlier events. Figure 13.7 shows a comparison of P-wave records from the NS component of Wood-Anderson (WA) seismographs. The record for the 1934 is dashed and for 1966 solid (Bakun and McEvilly, 1979). These recordings are shown for the stations at MHC and SBA. Below each are WA traces for the 2004 event simulated from Berkeley Seismological Laboratory's station MHC and station SBA, which is a station of the Southern California Seismic Network operated by Caltech and the USGS. The records are on the same scales and have been aligned on the P-wave. The WA records for the 2004 quake are clearly different from those of the earlier events. The Wood-Anderson amplitudes are proportionally larger at MHC than at SBC, and some phases arrive relatively earlier at MHC and later at SBC. These observations are consistent with NW rupture for the 2004 event, as compared to the SE rupture in 1966 and 1934.

Figure 13.8: Comparison of the 2004 earthquake with the events in 1922, 1934 and 1966, as recorded at the teleseismic station DBN by a Galitzin seismograph. The amplitudes are given in mm recorded on the photographic paper. The records have been low pass filtered at 0.1 Hz. Below the seismograms for each event the recordings for the four events have been overplotted to emphasize their similarities.
\begin{figure}\begin{center}
\epsfig{file=hell05_3_2.ps, width=8cm}\end{center}\end{figure}

Nearly 10,000 km from their epicenter, the events in 1922, 1934 and 1966 were recorded on light-sensitive paper by a set of horizontal Galitzin seismographs at De Bilt (DBN), the Netherlands. These seismographs were operated until 1994, when they were replaced by a 3-component, long period, digital seismometer, located at the same site. The modern instrument recorded the 2004 Parkfield earthquake.

In the records from DBN, there is some variation in the amplitudes of surface wave packets of the four events (Figure 13.8), although their phases seem to match quite well. For the comparison with the 2004 event, we have redigitized the photographic recordings of the horizontal components of the Galitzin seismographs for the 1922, 1934 and 1966 earthquakes. The waveform of the 2004 event closely resembles those of the 1922 and 1934 events. The 1966 event deviates from the other three, especially for high frequencies. Amplitudes are comparable within the uncertainty in the gain, supporting the idea of a characteristic Parkfield earthquake (Bakun and McEvilly, 1984).

Berkeley Seismological Laboratory
215 McCone Hall, UC Berkeley, Berkeley, CA 94720-4760
Questions or comments? Send e-mail: www@seismo.berkeley.edu
© 2005, The Regents of the University of California