Broadband analysis of the extended foreshock sequence of the Miyagi-Oki earthquake of 12 June 1978

1982 ◽  
Vol 72 (6A) ◽  
pp. 2017-2036
Author(s):  
George L. Choy ◽  
John Boatwright
Keyword(s):  
Stress Drop ◽  
Dynamic Stress ◽  
Main Shock ◽  
Body Waves ◽  
The Third ◽  
Radiated Energy ◽  
Dynamic Stress Drop ◽  
Static Stress Drop ◽  
Stress Drops ◽  
Seismograph Network

abstract The Miyagi-Oki earthquake of 12 June 1978, a large (Ms 7.8) interplate thrust event, occurred in a region which had not experienced earthquakes of magnitude greater than 7 since 1938. A sequence of four moderate-sized (5.4 < mb < 6.1) earthquakes encircled the rupture zone of the Miyagi-Oki earthquake over a period of 2 yr before the main shock. Broadband displacement and velocity records of body waves recorded digitally by stations of the Global Digital Seismograph Network are analyzed to determine the static and dynamic characteristics of the sequence. These characteristics include moment, radiated energy, dynamic and static stress drop, and apparent stress. Inversions of duration measurements made on the velocity waveforms permit quantifying the complexity of an event as well as constraining its rupture geometry. Intervals of 7 to 8 months separated the first three events; the main shock occurred 4 months after the third event. The rupture process of the third event was relatively complex; the event also had a substantially higher dynamic stress drop (175 bars) than did the stress drops of the first two events (9 and 10 bars, respectively). These observations suggest that the third event was an interme-diate-term precursor to the main shock. The fourth event, a short-term precursor to the main shock, occurred about 8 min before the main shock. Its dynamic stress drop (20 bars) was lower than that of the third event but higher than that of the first two events.

A dynamic model for far-field acceleration

1982 ◽  
Vol 72 (4) ◽  
pp. 1049-1068
Author(s):  
John Boatwright
Keyword(s):  
Stress Drop ◽  
High Frequency ◽  
Dynamic Stress ◽  
The Body ◽  
Body Waves ◽  
Far Field ◽  
Rupture Velocity ◽  
Frequency Level ◽  
Average Change ◽  
Dynamic Stress Drop

abstract A model for the far-field acceleration radiated by an incoherent rupture is constructed by combining Madariaga's (1977) theory for the high-frequency radiation from crack models of faulting with a simple statistical source model. By extending Madariaga's results to acceleration pulses with finite durations, the peak acceleration of a pulse radiated by a single stop or start of a crack tip is shown to depend on the dynamic stress drop of the subevent, the total change in rupture velocity, and the ratio of the subevent radius to the acceleration pulse width. An incoherent rupture is approximated by a sample from a self-similar distribution of coherent subevents. Assuming the subevents fit together without overlapping, the high-frequency level of the acceleration spectra depends linearly on the rms dynamic stress drop, the average change in rupture velocity, and the square root of the overall rupture area. The high-frequency level is independent, to first order, of the rupture complexity. Following Hanks (1979), simple approximations are derived for the relation between the rms dynamic stress drop and the rms acceleration, averaged over the pulse duration. This relation necessarily depends on the shape of the body-wave spectra. The body waves radiated by 10 small earthquakes near Monticello Dam, South Carolina, are analyzed to test these results. The average change of rupture velocity of Δv = 0.8β associated with the radiation of the acceleration pulses is estimated by comparing the rms acceleration contained in the P waves to that in the S waves. The rms dynamic stress drops of the 10 events, estimated from the rms accelerations, range from 0.4 to 1.9 bars and are strongly correlated with estimates of the apparent stress.

A Source Physics Interpretation of Nonself-Similar Double-Corner-Frequency Source Spectral Model JA19_2S

2022 ◽  
Author(s):  
Chen Ji ◽  
Ralph J. Archuleta
Keyword(s):  
Stress Drop ◽  
Wave Speed ◽  
Dynamic Stress ◽  
Corner Frequency ◽  
Rupture Velocity ◽  
Scaling Relation ◽  
Shear Wave Speed ◽  
Dynamic Stress Drop ◽  
Static Stress Drop ◽  
Frequency Source

Abstract We investigate the relation between the kinematic double-corner-frequency source spectral model JA19_2S (Ji and Archuleta, 2020) and static fault geometry scaling relations proposed by Leonard (2010). We find that the nonself-similar low-corner-frequency scaling relation of JA19_2S model can be explained using the fault length scaling relation of Leonard’s model combined with an average rupture velocity ∼70% of shear-wave speed for earthquakes 5.3 < M< 6.9. Earthquakes consistent with both models have magnitude-independent average static stress drop and average dynamic stress drop around 3 MPa. Their scaled energy e˜ is not a constant. The decrease of e˜ with magnitude can be fully explained by the magnitude dependence of the fault aspect ratio. The high-frequency source radiation is generally controlled by seismic moment, static stress drop, and dynamic stress drop but is further modulated by the fault aspect ratio and the relative location of the hypocenter. Based on these two models, the commonly quoted average rupture velocity of 70%–80% of shear-wave speed implies predominantly unilateral rupture.

Source physics interpretation of non-self-similar double-corner frequency source spectral model JA19_2S

2021 ◽  
Author(s):  
Chen Ji ◽  
Ralph Archuleta
Keyword(s):  
Aspect Ratio ◽  
Stress Drop ◽  
Dynamic Stress ◽  
Static Stress ◽  
Corner Frequency ◽  
Rupture Velocity ◽  
Dynamic Stress Drop ◽  
Self Similar ◽  
Static Stress Drop ◽  
Frequency Source

<p>Source spectral models developed for strong ground motion simulations are phenomenological models that represent the average effect that the source processes have on near fault ground motion. Their parameters are directly regressed from the observations and often do not have clear meaning for the physics of the source process. We investigate the relation between the kinematic double-corner frequency (DCF) source spectral model JA19_2S (Ji and Archuleta, BSSA, 2020) and static fault geometry scaling relations proposed by Leonard (2010). We derive scaling relations for the low and high corner frequency in terms of static stress drop, dynamic stress drop, fault rupture velocity, fault aspect ratio, and relative hypocenter location. We find that the non-self-similar low corner frequency  scaling relation of JA19_2S model for 5.3<<strong>M</strong><6.9 earthquakes is well explained using the fault length scaling relation of Leonard’s model combined with a constant rupture velocity. Earthquakes following both models have constant average static stress drop and constant average dynamic stress drop. The high frequency source radiation is controlled by seismic moment, static stress drop and dynamic stress drop but strongly modulated by the fault aspect ratio and the hypocenter’s relative location. The mean, scaled energy  (or apparent stress) decreases with magnitude due to the magnitude dependence of the fault aspect ratio. Based on these two models, the commonly quoted average rupture velocity of 70-80% of shear wave speed implies predominantly unilateral rupture.</p>

The foreshock sequence of the 1986 Chalfant, California, earthquake

1988 ◽  
Vol 78 (1) ◽  
pp. 172-187
Author(s):  
Kenneth D. Smith ◽  
Keith F. Priestley
Keyword(s):  
Main Shock ◽  
Slip Surface ◽  
Body Waves ◽  
Aftershock Distribution ◽  
Time Period ◽  
Main Shocks ◽  
Short Period ◽  
Local Shear ◽  
Stress Drops ◽  
Foreshock Activity

Abstract The ML 6.4 Chalfant, California, earthquake of 21 July 1986 was preceded by an extensive foreshock sequence. Foreshock activity is characterized by shallow clustering activity, including 7 events greater than ML 3, beginning 18 days before the earthquake, an ML 5.7 foreshock 24 hr before the main shock that ruptured only in the upper 10 km of the crust, and an “off-fault” cluster of activity perpendicular to the slip surface of the ML 5.7 foreshock associated with the hypocenter of the main shock. The Chalfant sequence occurred within the local short-period network, and the spatial and temporal development of the foreshock sequence can be observed in detail. Seismicity of the July 1986 time period is largely confined to two nearly conjugate planes; one striking N30°E and dipping 60° to the northwest associated with the ML 5.7 foreshock and the other striking N25°W and dipping 70° to the southwest associated with the main shock. Focal mechanisms for the foreshock period fall into two classes in agreement with these two planes. Shallow clustering of earthquakes in July and the ML 5.7 principal foreshock occur at the intersection of the two planes at a depth of approximately 7 km. The seismic moments determined from inversion of the teleseismic body waves are 4.2 × 1025 and 2.5 × 1025 dyne-cm for the principal foreshock and the main shock, respectively. Slip areas for these two events can be estimated from the aftershock distribution and result in stress drops of 63 bars for the principal foreshock and 16 bars for the main shock. The main shock occurred within an “off-fault” cluster of earthquakes associated with the principal foreshock. This cluster of activity occurs at a predicted local shear stress high in relation to the slip surface of the 20 July earthquake, and this appears to be the triggering mechanism of the main shock. The shallow rupture depth of the principal foreshock indicates that this event was anomalous with respect to the character of main shocks in the region.

Source dimensions and stress drops of small earthquakes near Parkfield, California

1984 ◽  
Vol 74 (1) ◽  
pp. 27-40
Author(s):  
M. E. O'Neill
Keyword(s):  
Stress Drop ◽  
Small Region ◽  
Static Stress ◽  
Depth Range ◽  
Zero Crossing ◽  
Source Dimensions ◽  
Duration Magnitude ◽  
Short Period ◽  
Static Stress Drop ◽  
Stress Drops

Abstract Source dimensions and stress drops of 30 small Parkfield, California, earthquakes with coda duration magnitudes between 1.2 and 3.9 have been estimated from measurements on short-period velocity-transducer seismograms. Times from the initial onset to the first zero crossing, corrected for attenuation and instrument response, have been interpreted in terms of a circular source model in which rupture expands radially outward from a point until it stops abruptly at radius a. For each earthquake, duration magnitude MD gave an estimate of seismic moment MO and MO and a together gave an estimate of static stress drop. All 30 earthquakes are located on a 6-km-long segment of the San Andreas fault at a depth range of about 8 to 13 km. Source radius systemically increases with magnitude from about 70 m for events near MD 1.4 to about 600 m for an event of MD 3.9. Static stress drop ranges from about 2 to 30 bars and is not strongly correlated with magnitude. Static stress drop does appear to be spatially dependent; the earthquakes with stress drops greater than 20 bars are concentrated in a small region close to the hypocenter of the magnitude 512 1966 Parkfield earthquake.

Source parameters for aftershocks of the Oroville, California, earthquake

1984 ◽  
Vol 74 (4) ◽  
pp. 1101-1123
Author(s):  
Jon Fletcher ◽  
John Boatwright ◽  
Linda Haar ◽  
Thomas Hanks ◽  
Art McGarr
Keyword(s):  
Stress Drop ◽  
Dynamic Stress ◽  
Depth Interval ◽  
Causal Mechanism ◽  
Overburden Pressure ◽  
Data Set ◽  
State Of Stress ◽  
Aftershock Sequences ◽  
Least Squares Fit ◽  
Stress Drops

Abstract A suite of 111 strong-motion accelerograms for 14 aftershocks of the Oroville, California, earthquake (ML = 5.7, 1 August 1975) that range in local magnitude (ML) from 2.8 to 5.2 has been analyzed to obtain estimates of seismic moment (Mo), source radius (ro), and stress drop (Δσ) in addition to the focal parameters of location, depth, and fault-plane solution. This data set, which is unusually complete for near-source (Δ ≲ 20 km) on-scale readings, allows for greater precision in the calculation of various measures of stress difference as represented by the Brune stress drop, the apparent stress, the arms stress drop, and the dynamic stress drop. In addition, the seismicity following each aftershock and state-of-stress seem to correlate with particular estimates of stress drop. Seismic moments were calculated from the asymptotic long-period spectral levels which were corrected for the radiation pattern of a double-couple point source. They range from 1.4 × 1021 dyne-cm for a ML = 2.8 shock to 3.3 × 1023 dyne-cm for a ML = 5.1 event. A least-squares fit between ML and the logarithm Mo yields log M 0 = ( 1.36 ± 0.22 ) M L + ( 16.8 ± 1.1 ) for M L ≧ 4.3 and log ⁡ M 0 = ( 1.1 ± 0.14 ) M L + ( 18 ± 0.51 ) for M L ≦ 4.1. These relationships are qualitatively in agreement with the response of the Wood-Anderson instrument to a Brune pulse. Stress drops from the Brune formulation range about 14 to 170 bars. Stress drop is correlated with depth in that the deepest events have the largest stress drops and no large stress drops occur at the shallow depths. Apparent stresses are smaller than the Brune stress drops and show a weaker depth dependence over the depth interval for which they are available. The stress drop calculated from the rms of acceleration (arms) was approximately constant at about 90 bars for 5 of the 7 larger events analyzed; the two high values of 160 and 190 bars were obtained only for the two events which had marked aftershock sequences of their own. These results may be interpreted in terms of the state-of-stress, simple fracture criteria, and mechanisms for the generation of aftershocks. The increase with depth of the envelope of the Brune stress drops may be caused by an increase in shear stress from overburden pressure. Smaller stress drop events can occur at any depth interval. The causal mechanism of aftershocks is not known, but probably includes a change in the frictional properties of the fault, suggesting that the arms stress drop is a measure of the frictional or dynamic stress release.

On the determination of radiated seismic energy and related source parameters

1984 ◽  
Vol 74 (2) ◽  
pp. 395-415
Author(s):  
D. J. Doornbos
Keyword(s):  
Stress Drop ◽  
Source Parameters ◽  
Seismic Energy ◽  
Source Size ◽  
Static Stress ◽  
Central Moments ◽  
Radiated Energy ◽  
Radiated Seismic Energy ◽  
Static Stress Drop

Abstract The determination of radiated seismic energy on the one hand, and of source size and static stress drop on the other, depends in principle on a representation of different parts of the source spectrum. In practice with band-limited data from a sparse network, the required source parameterization is often the same. Spectral models parameterized by the source's central moments of degree zero and two are introduced as an approximation to the general representation of the amplitude spectrum in terms of the central moments of even degree. Phase spectra are not used, apart from polarity. These models are shown to simulate well the principal features of common circular and Haskell type of models, including the corner frequency shift of P waves with respect to S waves, and the relation between rupture velocity and maximum seismic efficiency. Spectral bandwidths and the determination of radiated energy and apparent stress are contrasted to time domain pulse widths and the determination of source size and static stress drop in these models. The consequences of a reduced number of source parameters are examined, in particular for circular models and point source approximations; in these cases, results for radiated energy can be obtained in closed form. The scaling of radiated energy with moment is assumed to be linear for simple sources, but in stochastic models of complex sources the scaling may be between linear and quadratic. A relatively large increase of radiated energy with moment would be accompanied by an underestimate of source size and an overestimate of stress drop. However, the determination of radiated energy may still be correct.

Sign in / Sign up

Export Citation Format

Share Document