principles and applications of microearthquake networks

72 3. Dnta Processing Procedures
correct, then the up direction on the seismogram of a vertical-component
station corresponds to a compression in ground motion, whereas the down
direction corresponds to a dilatation. Since station polarity may be accidentally
reversed, it should be checked periodically by examining the
directions of first P-motions of large teleseismic events. For example,
Houck et ul. (1976) carried out this check for the USGS Central California
Microearthquake Network. About 15% of the stations were found to be
reversed.
Magnitudes of microearthquakes are frequently estimated from signal
durations because of the difficulties in calibrating large numbers of stations
in a microearthquake network, and in measuring maximum amplitudes
and periods from recordings with limited dynamic range. Signal
duration of microearthquakes has been defined in several ways as is discussed
in Section 6.4. Recently, the Commission on Practice of the International
Association of Seismology and Physics of Earth's Interior has
recommended that the signal duration be defined as "the time in seconds
between the first onset and the time the trace never again exceeds twice
the noise level which existed immediately prior to the first onset."
3.5.2. Interactive Method
The batch method just described has a serious drawback in that it is
time consuming to go back to the seismic records for correcting errors and
to enter the corrections into a computer. If one or a small group of earthquakes
is processed at a time, then errors can be corrected more efficiently.
Furthermore, if the data processing procedure is under computer
control, then an interactive dialogue between the analyst and the computer
can speed the entire process. Basically, a computer can perform
bookkeeping chores well, but it is more difficult to program the computer to
make intelligent decisions. There are two approaches in the interactive
method. In the first approach, the seismic traces are digitized and stored
in a computer for further processing. In the second approach, the seismic
traces are available only as visual records.
For the first approach, several interactive systems for processing microearthquake
data have been developed. For example, Johnson (1979)
described the CEDAR (Caltech Earthquake Detection and Recording)
system for processing seismic data from the Southern California Seismic
Network. This system consists of four stages of data processing. In the
first stage, an on-line computer detects events and records them on digital
magnetic tapes. The remaining stages of data processing are performed on
an off-line computer using the digital tapes. In the second stage, the detected
events are plotted, and the analyst separates the seismic events