Source mechanism and mb?Msanalysis of aftershocks of the San Fernando earthquake

Aftershocks of the 1971 February 9 San Fernando, California, earthquake were studied using: (1) local body-wave spectra (0.2–75 Hz) from two high-dynamic range, wide-band digital instruments (free period = 1 s) and one analogue instrument (free period = 3.8s) located at hypocentral distances of 10–30 km; (2) seismograms from six short-period instruments in the California Institute of Technology (CIT), Southern California network, at hypocentral distances of 35–300 km; (3) for the larger events, surface-wave magnitudes (Ms) and spectra (≈ 0.03–0.1 Hz) from long-period instruments in the WWSSN at hypocentral distances of 2–20 degrees; (4) body-wave magnitudes (mb) from short-period instruments in the Canadian network, at hypocentral distances of 16–42 degrees; and (5) magnitude determinations, hypocentral locations and fault plane solutions of the US Geological Survey and CIT. The spectra of all of the small (M ≈ 1–3.5) and about half of the larger (M ≈ 3.5–4) events have a single ‘corner’ frequency (between 3 and 30 Hz), below which they are approximtely constant and above which they are proportional to about ω−2 or ω−3. The spectra of the remaining larger events have two corner frequencies: one between 0.1 and 1 Hz, below which they are constant, and another between 3 and 10 Hz, above which they are proportional to about ω−2 or ω−3. For the events with two corner frequencies, the spectral amplitude at the first corner is two to eight times larger than at the second. One interpretation of these observations is that the small events and about half of the larger events are produced by faulting limited in time (≈ 0.1 s) and space (≈500m), whereas for the remaining larger events, the initial rupture ‘grows’ in time (for ≈ 2 s) and/or in space (to ≈ 6 km), thus increasing the total moment and causing a second corner frequency. Alternatively, the lower corner frequency might be attributed either to fore-slip and/or after-slip with a relatively slow dislocation velocity, This difference in spectrum shape may explain why many previous studies have observed a wide range of surface-wave excitation for events with similar local magnitude. Interpretation of the spectra using the Brune (1970, 1971) model results in a wide range of stress drops, from about 300 bar to less than a bar. There is an apparent upper limit on stress drop of about 300 bar and this may be related to the total tectonic stress operating. The events with two corner frequencies can be interpreted as partial stress drop events (stress drop ~ l/5 effective stress). The mb–Ms data for 3 4. Therefore, unless the mb-Ms data for underground nuclear explosions with 3 < mb < 4 deviate from the line for mb > 4, explosions and earthquakes will discriminate for 3 < mb < 4. For some of the largest events, source finiteness may have somewhat enhanced discrimination.