Showing papers in "Bulletin of the Seismological Society of America in 1981"
TL;DR: In this paper, a magnitude-independent shape based on geometrical spreading and anelastic attenuation was proposed for the attenuation curve, which decouples the determination of the distance dependence of the data from the magnitude dependence.
Abstract: We have taken advantage of the recent increase in strong-motion data at close distances to derive new attenuation relations for peak horizontal acceleration and velocity. This new analysis uses a magnitude-independent shape, based on geometrical spreading and anelastic attenuation, for the attenuation curve. An innovation in technique is introduced that decouples the determination of the distance dependence of the data from the magnitude dependence. The resulting equations are log A = − 1.02 + 0.249 M − log r − 0.00255 r + 0.26 P r = ( d 2 + 7.3 2 ) 1 / 2 5.0 ≦ M ≦ 7.7 log V = − 0.67 + 0.489 M − log r − 0.00256 r + 0.17 S + 0.22 P r = ( d 2 + 4.0 2 ) 1 / 2 5.3 ≦ M ≦ 7.4
where A is peak horizontal acceleration in g , V is peak horizontal velocity in cm/ sec, M is moment magnitude, d is the closest distance to the surface projection of the fault rupture in km, S takes on the value of zero at rock sites and one at soil sites, and P is zero for 50 percentile values and one for 84 percentile values.
We considered a magnitude-dependent shape, but we find no basis for it in the data; we have adopted the magnitude-independent shape because it requires fewer parameters.
1,010 citations
TL;DR: In this paper, it was shown that the Green9s functions for an elastic layered medium can be expressed as a double integral over frequency and horizontal wavenumber, which can be exactly represented by a discrete summation.
Abstract: Green9s functions for an elastic layered medium can be expressed as a double integral over frequency and horizontal wavenumber. We show that, for any time window, the wavenumber integral can be exactly represented by a discrete summation. This discretization is achieved by adding to the particular point source an infinite set of specified circular sources centered around the point source and distributed at equal radial interval. Choice of this interval is dependent on the length of time desired for the point source response and determines the discretized set of horizontal wavenumbers which contribute to the solution. Comparisons of the results obtained with those derived using the two-dimensional discretization method (Bouchon, 1979) are presented. They show the great accuracy of the two methods.
928 citations
TL;DR: In this article, the authors examined the root-mean-square acceleration (a rms ) characteristics of these records for 0 ≦ t − R/β ≦ T d in terms of the relation a rms = 0.85 ( 2 π ) 106 2 Δ σ ϕ R f max f o where Δ π is the earthquake stress drop, and showed that all 16 earthquakes have stress drops, as determined by record values of rms, very nearly equal to 100 bars (within a factor of 2).
Abstract: Analysis of more than 300 horizontal components of ground acceleration written by the San Fernando earthquake, eight other moderate-to-large California earthquakes, and seven Oroville aftershocks reveal that these acceleration time histories are, to a very good approximation, band-limited white Gaussian noise within the S -wave arrival window; the band limitation is defined by the spectral corner frequency f 0 and f max , the highest frequency passed by the accelerograph or the Earth9s attenuation, and the S -wave arrival window is (0 ≦ t − R/β ≦ T d ), where R is distance, β is shear-wave velocity, and T d is the faulting duration. An examination of the root-mean-square acceleration ( a rms ) characteristics of these records for 0 ≦ t − R/β ≦ T d in terms of the relation a rms = 0.85 ( 2 π ) 106 2 Δ σ ϕ R f max f o where Δ σ is the earthquake stress drop, yields the surprising result that all 16 earthquakes have stress drops, as determined by record values of a rms , very nearly equal to 100 bars (within a factor of 2). The source dependence of a rms thus depends solely on the parameter 1 / f o , which increases only as the one-sixth power of seismic moment for constant stress drop earthquakes. Put another way, model and record a rms are in agreement within a factor of 2 approximately 85 per cent of the time for Δσ = 100 bars and knowledge of 1 / f o . On the basis that acceleration time histories are finite-duration, band-limited, white Gaussian noise, for any of which a rms is fixed by Δσ = 100 bars and 1 / f o , we can estimate the peak accelerations ( a max ) for all of these records with considerable accuracy (50 per cent or less). The relation is a max = a rms 2 In ( 2 f max f o ) , where a rms is defined above. With less accuracy, this relation fits the peak acceleration set of Hanks and Johnson (1976) as well, again with Δσ = 100 bars. At a fixed, close distance, we determine the magnitude dependence of a max to be log a max ∝ 0.30 M for 4 ≲ M = M L ≲ 6 1 2 , close to that recently determined empirically by Joyner and Boore (1981) for 5.0 ≦ M ≦ 7.7, their coefficient on M (moment magnitude) being 0.25 ± 0.04. In the model presented here, the magnitude dependence of peak acceleration is a function of faulting duration alone; larger earthquakes have larger peak accelerations because they last longer, not because they are intrinsically more powerful at the high frequencies controlling peak acceleration. These well-behaved characteristics of high-frequency strong ground motion also suggest that the stress differences which develop in the course of crustal faulting are comparably well behaved, both in the average stress release across the characteristic source dimension and in the spectral composition and distribution of stress differences that develop across smaller dimensions.
743 citations
TL;DR: In this article, the Haskell-Thompson transfer matrix method is used to derive layer stiffness matrices which may be interpreted and applied in the same way as stiffness matrix in conventional structural analysis, and the exact expressions are given for the matrices, as well as approximations for thin layers.
Abstract: The Haskell-Thompson transfer matrix method is used to derive layer stiffness matrices which may be interpreted and applied in the same way as stiffness matrices in conventional structural analysis These layer stiffness matrices have several advantages over the more usual transfer matrices: (1) they are symmetric; (2) fewer operations are required for analysis; (3) there is an easier treatment of multiple loadings; (4) substructuring techniques are readily applicable; and (5) asymptotic expressions follow naturally from the expressions (very thick layers; high frequencies, etc) While the technique presented is not more powerful than the original Haskell-Thompson scheme, it is nevertheless an elegant complement to it The exact expressions are given for the matrices, as well as approximations for thin layers Also, simple examples of application are presented to illustrate the use of the method
712 citations
TL;DR: In this paper, the authors used strong-motion data recorded within 50 km of the rupture zone to study near-source attenuation characteristics of horizontal peak ground acceleration for worldwide earthquakes of magnitudes 5.0 to 7.7.
Abstract: Strong-motion data recorded within 50 km of the rupture zone were used to study near-source attenuation characteristics of horizontal peak ground acceleration for worldwide earthquakes of magnitudes 5.0 to 7.7. The data base consisted of 229 horizontal components of peak acceleration recorded from 27 earthquakes, including the 15 October 1979, Imperial Valley earthquake. These data were found to be adequately represented by the functional relationship PGA = a exp ( b M ) [ R + C ( M ) ] − d
where PGA represents the mean of the peak values scaled from the two horizontal components of each recording, M is Richter magnitude, and R is distance from the fault rupture zone. Peak acceleration was found to be lognormally distributed with a standard error representing a 45 per cent increase in the median estimate. The regression analysis statistically confirmed the results of earthquake simulation studies that have predicted peak acceleration to become independent of magnitude and distance in the near field. An extensive sensitivity study showed that predictions based on our attenuation relationships are stable with respect to reasonable model and parameter variations. An analysis of residuals was used to investigate the behavior of peak acceleration with respect to various earthquake, site, and recording parameters, the more significant findings being: (1) a similarity in the level of acceleration recorded on soil or rock; (2) larger than average accelerations recorded at sites located on shallow soils or in areas of steep topography; (3) larger than average accelerations associated with earthquakes having reverse fault mechanisms; and (4) lower than average accelerations recorded in large embedded structures.
450 citations
TL;DR: In this article, Bayesian statistical methods are used to incorporate a priori information about arrival-time variance into the derivation of hypocenter confidence ellipsoids, permitting a more realistic calculation of critical parameters in the case where the number of stations is small.
Abstract: Improved methods for single-and multiple-event hypocenter determinations are developed and applied to the problem of locating earthquake clusters in the South-Central Pacific Ocean. Bayesian statistical methods are used to incorporate a priori information about arrival-time variance into the derivation of hypocenter confidence ellipsoids, permitting a more realistic calculation of critical parameters in the case where the number of stations is small. The diagonal elements of certain projection operators, called “data importances” by Minster et al. (1974), are used to evaluate network balance. The hypocentroid of an event cluster is defined to be the average location of events within the cluster, and the deviations of individual hypocenters from the hypocentroid are called cluster vectors . The problem of estimating the cluster vectors can be decoupled from the problem of estimating the hypocentroid by a simple but fundamental mathematical result, here termed the hypocentroidal decomposition theorem. The algorithm based on this analysis appears to have many advantages over other published methods for multiple-event location, both in its efficient use of available information and its computational speed. The application of this method to three clusters of shallow intraplate seismicity in the South-Central Pacific, designated Regions A, B, and C, demonstrates that the seismicity within each cluster is very localized; the rms lengths of the cluster vectors for each group of epicenters are estimated to be only 9, 6, and 12 km, respectively. Estimates of the epicentroids are
348 citations
TL;DR: The effect of alluvium on strong ground motion can be seen by comparing two strong-motion records of the Coyote Lake, California, earthquake of 6 August 1979 (M L = 5.9) as discussed by the authors.
Abstract: The effect of alluvium on strong ground motion can be seen by comparing two strong-motion records of the Coyote Lake, California, earthquake of 6 August 1979 ( M L = 5.9). One record at a site on Franciscan bedrock had a peak horizontal acceleration of 0.13 g and a peak horizontal velocity of 10 cm/sec. The other, at a site 2 km distant on 180 meters of Quaternary alluvium overlying Franciscan, had values of 0.26 g and 32 cm/sec, amplifications by factors of 2 and 3. Horizontal motions computed at the alluvial site for a linear plane-layered model based on measured P and S velocities show reasonably good agreement in shape with the observed motions, but the observed peak amplitudes are greater by a factor of about 1.25 in acceleration and 1.8 in velocity. About 15 per cent of the discrepancy in acceleration and 20 per cent in velocity can be attributed to the difference in source distance; the remainder may represent focusing by refraction at a bedrock surface concave upward. There is no clear evidence of nonlinear soil response. Fourier spectral ratios between motions observed on bedrock and alluvium show good agreement with ratios predicted from the linear model. In particular, the observed frequency of the fundamental peak in the amplification spectrum agrees with the computed value, indicating that no significant nonlinearity occurs in the secant shear modulus. Computations show that nonlinear models are compatible with the data if values of the coefficient of dynamic shear strength in terms of vertical effective stress are in the range of 0.5 to 1.0 or greater. The data illustrate that site amplification may be less a matter of resonance involving reinforcing multiple reflections, and more the simple effect of the low near-surface velocity. Application of traditional seismological theory leads to the conclusion that the site amplification for peak horizontal velocity is approximately proportional to the reciprocal of the square root of the product of density and shear-wave velocity.
255 citations
TL;DR: In this article, the authors proposed a split node technique for finite element numerical computations, where the value of the displacement at a single node point shared between two elements depends upon which element it is referred to, thus introducing a displacement discontinuity between the two elements.
Abstract: This paper outlines a new method, the “split node technique” for introducing fault displacements into finite element numerical computations. The value of the displacement at a single node point shared between two elements depends upon which element it is referred to, thus introducing a displacement discontinuity between the two elements. We show that the modification induced by this splitting can be contained in the load vector, so that the stiffness matrix is not altered. The number of degrees of freedom is not increased by splitting. This method can be implemented entirely on the local element level, and we show rigorously that no net forces or moments are induced on the finite element grid when isoparametric elements are used. This method is thus of great utility in many geological and engineering applications.
246 citations
TL;DR: In this paper, the authors examined the seismic gaps and recurrence periods of large, shallow interplate earthquakes along the Mexican subduction zone, combining information from a catalog of nineteenth century 9.9s earthquakes, some relocated epicenters of the early part of this century, source parameters of recent large earthquakes, and redetermined magnitudes of great, shallow earthquakes.
Abstract: Seismic gaps and recurrence periods of large, shallow interplate earthquakes along the Mexican subduction zone are reexamined after combining information from a catalog of nineteenth century9s earthquakes, some relocated epicenters of the early part of this century, source parameters of recent large earthquakes, and redetermined magnitudes of great, shallow earthquakes of this century. Tehuantepec and Michoacan gaps have not experienced a large shock in this century and perhaps none in the past century; they are either aseismic or have anomalously large repeat times. Guerrero, Jalisco, and Ometepec regions presently appear to have a high seismic potential. Observed average repeat times of large earthquakes ( M s ≳ 7.4) in six regions (east, central, and west Oaxaca, San Marcos, Petatlan, and Colima) are between 32 to 56 yr. Data of this century indicate that the strain is released mostly in large events ( M s ≳ 7.4). A simple dislocation model with parameters obtained from the studies of recent earthquakes explains the observed recurrence periods quite well. The b value for this zone is not meaningful, an observation which is of significance for seismic risk estimation. Most of seismic moment (or, equivalently, seismic energy) release since 1800 appears to occur for 15 yr followed by relative quiescence in the next 15 yr.
224 citations
209 citations
TL;DR: In this article, a moment tensor and double couple inversion technique is applied to near-regional waveform data from a small earthquake associated with the Koyna Reservoir which occurred 13 September 1967, as the second in a group of three events with very similar waveshapes, but differing amplitudes.
Abstract: The treatment of the seismic source inverse problem, when diverse forms of waveform data are available, is simple and elegant using a moment tensor formalism. If earth structure is known and its effects predictable in terms of vertically inhomogeneous elastic-layered models, then all types of wave phenomena (e.g., surface waves, body waves, leaky modes, etc.) for a purely deviatoric moment tensor point source may be represented by, at most, a sum of three Green9s functions. For an arbitrary symmetric moment tensor point source, one additional Green9s function is needed for the P - SV system. However elegant this formalism may be for posing the linear inverse problem, the major difficulties lie in earth structure unknowns and resultant nonlinearities in the Green9s functions which can cause significant trade-offs with source parameters. A hybrid inversion procedure is set up to gain insight into the probable unknowns in particular problems by incorporating both a linearized least-squares gradient method for the moment tensor or double couple, and smoothed time function parameters, and a nonlinear systematic trial-and-error search for moment tensor or double couple parameters for several assumptions of Green9s function. The inversion technique is applied to near-regional waveform data from a small earthquake associated with the Koyna Reservoir which occurred 13 September 1967, as the second in a group of three events with very similar waveshapes, but differing amplitudes. The magnitudes of the second and third events are smaller by −0.2 and −0.8 units, respectively, compared to the first. The absolute magnitude for the first event is poorly constrained but is estimated to be 4.0 to 4.5 rather than the previously published value of 5.5 to 6.0. From the similarity of waveshapes, all three events are inferred to have the same mechanism and occurred within about 2 km of the same hypocenter. The results from moment tensor and double couple inversions for event 2 data indicate that source depth was 5 km and that left-lateral faulting occurred on a plane with a strike of N20°E ± 5°, dip of 90° ± 15°, and a rake of 0° ± 35°. The inferred far-field time function is approximately 3 sec in duration, unusually long for the seismic moment of 9 × 10 22 dyne-cm, yielding a possible stress drop of about 0.05 bars. A fault map was constructed from LANDSAT image interpretation and shows a predominance of NNW to NNE striking faults in the Koyna area which is used to infer the appropriate nodal plane in the inversion results. These faults tend to define a broad en-echelon zone which parallels the Western Ghats in this area.
TL;DR: A reanalysis of the available data for the 1966 Parkfleld, California, earthquake (Mr -- 5½) suggests that although the ground breakage and aftershocks extended about 40 km along the San Andreas Fault, the initial dynamic rupture was only 20 to 25 km in length as mentioned in this paper.
Abstract: A reanalysis of the available data for the 1966 Parkfleld, California, earthquake (Mr -- 5½) suggests that although the ground breakage and aftershocks extended about 40 km along the San Andreas Fault, the initial dynamic rupture was only 20 to 25 km in length. The foreshocks and the point of initiation of the main event locate at a small bend in the mapped trace of the fault. Detailed analysis of the P-wave first motions from these events at the Gold Hill station, 20 km southeast, indicates that the bend in the fault extends to depth and apparently represents a physical discontinuity on the fault plane. Other evidence suggests that this discontinuity plays an important part in the recurrence of similar magnitude 5 to 6 earthquakes at Parkfield. Analysis of the strong-motion records suggests that the rupture stopped at another discontinuity in the fault plane, an en-echeion offset near Gold Hill that lies at the boundary on the San Andreas Fault between the zone of aseismic slip and the locked zone on which the great 1857 earthquake occurred. Foreshocks to the 1857 earthquake occurred in this area (Sieh, 1978), and the epicenter of the main shock may have coincided with the offset zone. If it did, a detailed study of the geological and geophysical character of the region might be rewarding in terms of understanding how and why great earthquakes initiate where they do.
TL;DR: In this paper, near-source ground motion at four azimuths but constant epicentral range is computed from a buried circular strike-slip fault in a half-space.
Abstract: Near-source ground motion at four azimuths but constant epicentral range is computed from a buried circular strike-slip fault in a half-space. Particle acceleration, velocity, and displacement at each station on the free surface is computed in the frequency band 0.0 to 5.0 Hz. The assumed dislocation is derived from the Kostrov (1964) displacement function for a continuously propagating stress relaxation. The azimuthal variations in the amplitudes and waveforms directly result from spatially varying slip on the fault, spatially varying radiation pattern over the fault, and the magnitude and direction of the rupture velocity. The near-source ground motions are dominated by the rupture in the direction of the receiver.
Using a 100-bar effective stress (initial stress minus sliding friction) in a Poisson solid with β = 3.0 km/sec the shear wave speed, and shear modulus μ = 3.0 × 10^(11) dyne/cm^2, the simulated earthquake has a moment M_o = 4.5 × 10^(25) dyne-cm. Using a rupture velocity of 0.9β, the peak acceleration is 1195 cm/sec^2 and velocity 10^4 cm/sec for the receiver directly on strike. For a receiver 30° off strike, the maximum acceleration 236 cm/sec^2 occurs on the vertical component.
TL;DR: A global tectonic regionalization, GTR1, has been constructed for use in seismological studies of aspherical heterogeneity as mentioned in this paper, where three oceanic regions are defined by equal increments in the square root of crustal age (25- and lO0-m.y.m. isochrons).
Abstract: A global tectonic regionalization, GTR1, has been constructed for use in seismological studies of aspherical heterogeneity. Three oceanic regions are defined by equal increments in the square root of crustal age (25- and lO0-m.y. isochrons), and three continental regions are distinguished according to their generalized tectonic behavior during the Phanerozoic (Precambrian shields and platforms, Phanerozoic platforms, Phanerozoic orogenic zones and magmatic belts). The regionalization is presented both as a tectonic map and by its discretization on a 5 ° by 5 ° geographic grid. GTR1 is reasonably successful in accounting for the large-scale geographic variations in body-wave, surfacewave, and free-oscillaUon data, and it should prove useful in the seismological testing of the competing theories of tectospherm development.
TL;DR: More than 1800 earthquakes have occurred during the first 9 yr of filling of the 300m deep Nurek Reservoir in Tadjikistan as mentioned in this paper, which is more than four times the average rate of activity in the region prior to the start of filling.
Abstract: More than 1800 earthquakes (1.4 > M > 4.6) have occurred during the first 9 yr of filling of the 300-m deep Nurek Reservoir in Tadjikistan. This is more than four times the average rate of activity in the region prior to the start of filling. The increased seismicity has occurred in a series of bursts, the two most intense of which were related to rapid increases in water level during the first two stages of filling—to 105 m in 1972 and to 205 m in 1976. All periods of high seismicity take place when the water level is higher than it has been previously or within 10 m of its previous maximum. If the water level drops more than 10 m below its previous maximum, the level of seismicity decreases. All of the largest earthquakes and most of the bursts of activity are triggered by decreases in the rate of filling of the reservoir. Once the water level is more than 10 m above the previous maximum, the potential for increased seismicity is high. Extremely small changes in filling rate can then trigger the onset of activity. For example, the largest earthquakes all followed decreases in filling rate of approximately 0.5 m/day; and in a number of cases, increased seismicity began soon after the reservoir started to empty by rates as small as 0.2 m/day 2 . The response in seismicity to decreases in the filling rate is rapid. Increased activity follows abrupt decreases in filling rate with delays as short as 1 to 4 days. As the reservoir has approached its maximum size, extending 40 km upstream from the dam, the area of induced seismicity has increased as well. The first induced earthquakes in 1971 were located 10 to 15 km southwest of the reservoir. From 1972 to 1978, activity migrated into the immediate reservoir area and followed the growth of the reservoir upstream. The first stage of activity in 1971 to 1972 was characterized by low b values and included the largest earthquakes of M = 4.6 and M = 4.3 in November 1972, when the water level first exceeded 100 m. From 1973 to 1979, activity was confined to the immediate reservoir area and b values were higher. Although the water level has risen to over 250 m, no induced earthquakes larger than M = 4.1 have occurred since November 1972.
TL;DR: In this article, ground acceleration was recorded at a depth of about 3 km in the East Rand Proprietary Mines, South Africa, for tremors with −1 ≦ M L ≦ 2.38 · M L.
Abstract: Ground acceleration was recorded at a depth of about 3 km in the East Rand Proprietary Mines, South Africa, for tremors with −1 ≦ M L ≦ 2.6 in the hypocentral distance range 50 m R ≦ 1.6 km. The accelerograms typically had predominant frequencies of several hundred Hertz and peak accelerations, a , as high as 12 g . The peak accelerations show a dependence on magnitude, especially when expressed as dynamic shear-stress differences, defined as σ˜ = ρRa , where ρ is density. For the mine tremors, σ˜ varies from 2 to 500 bars and depends on magnitude according to log σ˜ = 1.40 + 0.38 · M L . Accelerograms for 12 events were digitized and then processed to determine velocity and, for seven events with especially good S/N , displacement and seismic source parameters. Peak ground velocities v ranged up to 6 cm/sec and show a well-defined dependence one earthquake size as measured by M L or by seismic moment, M o . On the basis of regression fits to the mine data, with −0.76 ≦ M L ≦ 1.45, log Rv = 3.95 + 0.57 M L , where Rv is in cm 2 /sec, and log Rv = −4.68 + 0.49 log M o . These regression lines agree excellently with the corresponding data for earthquakes of M L up to 6.4 or M o to 1.4 × 10 26 dyne-cm. At a given value of M L or M o , a , at fixed R , shows considerably greater variation than v and appears to depend on the bandwidth of the recording system. The peak acceleration at small hypocentral distances is broadly consistent with ρRa = 1.14 Δτ r o f s / β , where Δτ is stress drop, r o is the source radius, β is shear velocity, and f s is the bandwidth of the recording system. The peak velocity data agree well with Rv = 0.57 β Δτ r o /μ, where μ is the modulus of rigidity; both expressions follow from Brune9s model of the seismic source and were compared with data for events in the size range 5 × 10 16 ≦ M o ≦ 1.4 × 10 26 dyne-cm. Measurements of the source parameters indicated that, as for earthquakes, the stress drops for the tremors range from 1 to 100 bars and show no consistent dependence on M o down to M o = 5 × 10 16 dyne-cm.
TL;DR: In this paper, the authors measured deformation measured within networks that span the San Juan Bautista-Cholame segment of the San Andreas fault in central California yields slip rates similar to those measured across a 100-to 300-m-wide zone by repeated alinement array surveys.
Abstract: Periodic measurements of fault-crossing networks with a side length of 1 to 3 km are being made to monitor deformation across fault zones in California. The distance measurements are made with a Hewlett-Packard 3800 or 3808 electronic distance meter and have a maximum standard deviation of 5 mm. Deformation measured within networks that span the San Juan Bautista-Cholame segment of the San Andreas fault in central California yields slip rates similar to those measured across a 100- to 300-m-wide zone by repeated alinement array surveys. Fault slip rates increase from near 0 to 32 mm/yr between San Juan Bautista and Bitterwater Valley in step-like increments. From Bitterwater to Slack Canyon slip rates vary between 26 and 32 mm/yr. Slip rates decrease southwestward of Slack Canyon to 3 mm/yr at Cholame. In contrast, Geodolite measurements of deformation across a 20-km-wide zone are consistent from San Juan Bautista to Slack Canyon and imply a 32 ± 2 mm/yr slip rate. Deformation across the Calaveras fault accounts for the difference between Geodolite and near-fault slip rates between San Juan Bautista and Bear Valley, although the zone of deformation is wider than 2.5 km just south of Hollister. At Bear Valley, measurements of a short-range network crossing the Paicines fault imply a slip rate of 10 ± 3 mm/yr during the period 1976 to 1979. From Slack Canyon to Cholame, Geodolite measurements show a constant decrease in the rate of shallow slip.
TL;DR: In this paper, a detailed study of the earthquakes associated with the Rocky Mountain Arsenal disposal well is presented, where long-period surface-wave studies are used together with P-wave first motions to show that the 10 April, 9 August, and 27 November 1967 earthquakes occurred at depths of 3 to 5 km and were characterized by normal faulting along a northwest striking fault plane.
Abstract: A detailed study of the earthquakes associated with the Rocky Mountain Arsenal disposal well is presented. Long-period surface-wave studies are used together with P -wave first motions to show that the 10 April, 9 August, and 27 November 1967 earthquakes occurred at depths of 3 to 5 km and were characterized by normal faulting along a northwest striking fault plane. A joint hypocenter relocation of 103 microearthquakes of a data set of 279 recorded between 1967 and 1968 shows a hypocenter pattern striking N50°W, with most of the events located about 5 km northwest of the disposal well at depths between 3 to 8 km. A fault plane dipping southwest is tenuously suggested by those earthquakes with depths less than 5 km. Modeling of near-field seismoscope observations lend support to the focal mechanisms derived.
TL;DR: A number of seismic source models for underground nuclear explosions have been developed over the past two decades as discussed by the authors, including the spherically symmetric compressional source model, the wave conversion source model and the tectonic strain release source model.
Abstract: A number of seismic source models for underground nuclear explosions have been developed over the past 2 decades. These models include the spherically symmetric compressional source model, the wave conversion source model, the tectonic strain release source model, the spall slapdown source model, and the near-regional source model. These model are reviewed in this study and are shown to be inconsistent with various geophysical data associated with underground nuclear explosions. In particular, the Rayleigh and Love wave signals generated by underground nuclear explosions have not been explained satisfactorily by any of these source models. To explain the observed explosion data, it may be necessary to model the explosion seismic source as a sequence of mechanisms producing seismic signals. These mechanisms all act within the first few seconds following the explosion detonation. One of the most important of these mechanisms is probably explosion-induced thrust faulting.
TL;DR: In this article, the authors determined approximate Lg attenuation coefficient values for paths in eastern North America at periods between 2 and 4 seconds where no data were previously available, and these new data, together with reported values at a period of 1 sec, are consistent with values predicted by a frequency-dependentent Qβ of the crust in which Qβ varies as ω 0.2.
Abstract: Approximate Lg attenuation coefficient values are determined for paths in eastern North America at periods between 2 and 4 sec where no data were previously available. These new data, together with reported values at a period of 1 sec, are consistent with values predicted by a frequency-dependentent Qβ of the crust in which Qβ varies as ω 0.2. The new data are inconsistent with values which would be predicted by models having prominent maxima or minima in Qβ −1 within the period range 1 to 5 sec.
Lg attenuation coefficients were computed for the frequency-independent crustal Qβ models of Cheng and Mitchell (1981) for the Basin and Range Province and Colorado Plateau of the Western United States. The predicted values at 1 sec for the Colorado Plateau and Basin and Range Province are about twice as large, and three times as large, respectively, as those predicted for the Eastern United States. Recently reported values of Q for 1-sec Lg in those regions are consistent with the attenuation coefficients and Q values of Lg predicted by the Colorado Plateau and Basin and Range Province models. A model with a small degree of frequency dependence of Qβ would be consistent, but is not required by the data.
TL;DR: In this paper, an efficient and accurate method of estimating the sensitivity of seismic hazard calculations to statistical uncertainties in models and parameters is demonstrated. But this method requires knowledge of the earthquake magnitude and distance that contribute most to the probability of exceedence of a chosen acceleration level; the methods estimate sensitivities using point-source seismichazard approximations for which closed-form solutions are available.
Abstract: Efficient and accurate methods of estimating the sensitivity of seismic hazard calculations to statistical uncertainties in models and parameters are demonstrated. These models require knowledge of the earthquake magnitude and distance that contribute most to the probability of exceedence of a chosen acceleration level; the methods estimate sensitivities using point-source seismic-hazard approximations for which closed-form solutions are available. An additional result is that the use of Bayesian estimates for seismicity and ground motion parameters in the hazard analysis produces unbiased Bayesian estimates of the seismic ground motion hazard, due to the almost linear relationship between ground motion amplitudes at a given probability level, and parameter uncertainties. Application of these methods to the San Francisco, California, Bay area indicates a coefficient of variation (cov) of the 500-yr acceleration of about 0.4 at sites close to major faults, and a cov of about 0.2 at sites 50 km to the east of the major east bay faults. These cov9s result from statistical uncertainty in the depth of energy release, the activity rate and Richter b value for each fault, and the mean acceleration-attenuation relationship. A similar analysis in the central Mississippi Valley area indicates a cov in 500-yr acceleration of 0.4 near the major faults, with a value of about 0.3 at distances greater than 50 km. The sources of statistical uncertainty in this region are the depth of energy release as well as its location, the activity rate and Richter b value for each fault, and the mean acceleration-attenuation function.
TL;DR: In this paper, a method for calculating the focal mechanism of an earthquake from the distribution of the ratio of the amplitudes of SV to P waves has been developed and tested, based on information taken from the hypocenter solution for the earthquake, a factor for each station that corrects for the effect of the free surface on the observed wave amplitudes.
Abstract: A procedure for calculating the focal mechanism of an earthquake from the distribution of the ratio of the amplitudes of SV to P waves has been developed and tested. The input consists of information taken from the hypocenter solution for the earthquake, a factor for each station that corrects for the effect of the free surface on the observed wave amplitudes and the observed values of the ratio of the vertical component of P to the vertical component of SV . The procedure starts with the selection of a slip direction. The program then finds the best fault strike and dip corresponding to that slip direction, with “best” measured by the smallest scatter of the dip for any strike. All three fault parameters are then adjusted by an iterative least-squares adjustment. ( SV/P ) z is a strongly nonlinear function of the fault parameters, so the solution found by the procedure is inherently nonunique. The acceptable solutions can be quite well constrained if the slip direction can be estimated initially on the basis of independent information. A variety of tests of the procedure have been carried out on real and synthetic data. Given a set of amplitude data for an earthquake in Bear Valley, California, and told only that the mode of slip was predominantly strike-slip, the program “found” the San Andreas fault, i.e., converged to a fault with strike and dip close to the known values. Further work on a German earthquake served to bring out some of the ambiguities of the solutions. Analysis of data from a brief swarm of earthquakes in southwest Germany showed that the method yields useful mechanisms for small events and revealed a rotation of the focal mechanisms during the swarm. Processing of amplitudes from synthetic seismograms gave a fault plane close to the known input, but marked disagreement between the SV amplitudes on the synthetic seismograms and the amplitudes predicted by simple dislocation theory is unresolved.
TL;DR: In this paper, a new crustal model with linear velocity gradients within layers does as good a job of locating earthquakes on south Hawaii as any model yet published, using travel-time residuals from local earthquakes and consistency of focal mechanism solutions as tests.
Abstract: A new crustal model with linear velocity gradients within layers does as good a job of locating earthquakes on south Hawaii as any model yet published. Incorporating linear gradients means the model can be simpler and free of artificial velocity discontinuities. Using travel-time residuals from local earthquakes and consistency of focal mechanism solutions as tests, it is seen that a low-velocity zone at the base of the crust is not required.
TL;DR: Several transmitting boundaries have been proposed for the numerical analysis of wave propagation in continua of infinite extent as discussed by the authors, which have been grouped into three broad classes: (1) elementary boundaries (Dirichlet or Neumann boundary conditions); (2) highly absorbing local boundaries; and (3) consistent transmitting boundaries.
Abstract: Several transmitting boundaries which have been proposed for the numerical analysis of problems of wave propagation in continua of infinite extent are reviewed. They are grouped into three broad classes: (1) elementary boundaries (Dirichlet or Neumann boundary conditions); (2) highly absorbing local boundaries; and (3) consistent transmitting boundaries. A closed-form comparison for the problem of time-harmonic antiplane line load on a stratum shows that elementary boundaries produce strong reflections, local boundaries have good absorption characteristics, and consistent transmitting boundaries are perfect absorbers.
TL;DR: In this paper, it was shown that the corner frequency shift is a very common condition of the far-field body waves of earthquakes, with no discernible dependence on earthquake source strength, hypocentral distance or depth, or recording device.
Abstract: The very common seismological observation that the P waveform is enriched in high-frequency motion relative to the S waveforms of the same earthquake manifests itself, in spectral studies of the earthquake mechanism, as the “corner frequency shift,” the general although not ubiquitous tendency for the P -wave corner frequency f o( P ) to be greater than the S -wave corner frequency f o( S ). In point source, time-domain modeling studies of the earthquake mechanism which follow the recipe of D. V. Helmberger and C. A. Langston that explicitly suppresses the corner frequency shift, it is an equally common result that the synthetic S waves are systematically enriched in high-frequency motion relative to the observations. Almost three dozen spectral analysis and time-domain modeling studies are recapitulated in this one to conclude: (1) the corner frequency shift is a very common condition of the far-field body waves of earthquakes, with no discernible dependence on earthquake source strength, hypocentral distance or depth, or recording device; and (2) the corner frequency shift is the manifestation of an intrinsic property of earthquakes, source finiteness. Anelastic attenuation estimates for shear waves determined on the basis of a source model derived from P waves are likely to be strongly biased to high values of Q -1, if the shear excitation is estimated directly from the equivalent far-field compressional excitation, i.e., with no allowance for source finiteness. The corner frequency shift, moreover, places strong constraints on admissible earthquake source models; point-source and Haskell-type dislocation models will not be among them.
TL;DR: In this article, the authors analyzed P_n propagation as a function of azimuth across a 28-station, 150-km aperture subarray of the SCARLET network centered near the central Transverse Ranges, California.
Abstract: We analyze P_n propagation as a function of azimuth across a 28-station, 150-km aperture subarray of the SCARLET network centered near the central Transverse Ranges, California. We selected signals from 81 earthquakes and explosions with epicentral distances ranging from 150 to 400 km, covering all azimuths except a 40° gap from the southwest and a lesser gap from the northeast direction. For each source the apparent velocity of P_n was determined using a one-norm measure of misfit. The apparent P_n velocity does not show any systematic variation with epicentral distance but exhibits a strong azimuthal dependence. Our preferred interpretation calls for a slightly dipping (2° to N40W) planar moho, with 3 to 4 per cent anisotropy of subcrustal material. Transverse isotropy with a nearly horizontal symmetry axis is sufficient to explain the data; the direction of sagittal symmetry is N50W. The isotropic velocity of P_n is 7.8 km/sec. In contrast, a higher (8.1 km/sec) P_n velocity is found in the Mojave block, with no indication of anisotropy. These observations are consistent with a subcrustal model of the Pacific-North America plate boundary where ductile flow is characterized by simple shear in a vertical plane with strike parallel to the direction of relative plate motion.
TL;DR: In this paper, the authors investigated the body force equivalents for a seismic dislocation occurring in an anisotropic source medium and study radiation patterns of seismic body waves resulting from them.
Abstract: We investigate body force equivalents for a seismic dislocation occurring in an anisotropic source medium and study radiation patterns of seismic body waves resulting from them. The point source representation of the equivalent body forces is obtained following a result of Kosevich (1962, 1965). Green9s tensor for an anisotropic medium is calculated using a far-field approximate method by Kosevich and Natsik (1964). Radiation patterns of seismic body waves are obtained by a straightforward convolution operation on the equivalent forces with the approximate Green9s tensor. The seismic dislocation occurring in an anisotropic source medium is equivalent in general to the sum of three orthogonal dipole forces with different magnitudes, for which the seismic moment tensor has a nonzero trace. Because of the third dipole force which never appears for an isotropic medium, a significant distortion of the radiation patterns occurs in a direction near the null vector. Nodal lines of P -wave radiation patterns are separated into isolated loops and/or secondary nodal lines appear. In directions where group velocity differs from the corresponding phase velocity, the effect of the medium transfer response on the polarities of body waves seems to be larger than that in other directions. The combination of the effects of source forces and medium transfer response distorts the radiation pattern.
TL;DR: In this article, Finite-difference calculations have been performed for the simple case of vertically incident SV and P waves impinging on a step change in surface elevation, and they predict scattered Rayleigh waves with amplitudes as large as 0.4 times the amplitude of the surface motion of the incident waves in the absence of any topographic relief.
Abstract: Scattering of body waves to surface waves is a means for converting energy with essentially infinite horizontal wavelength to motion with short wavelengths. This process, of interest to engineers designing structures with large horizontal dimensions such as pipelines, tunnels, and bridges, has received little attention from seismologists. Finite-difference calculations have been performed for the simple case of vertically incident SV and P waves impinging on a step change in surface elevation. These calculations predict scattered Rayleigh waves with amplitudes as large as 0.4 times the amplitude of the surface motion of the incident waves in the absence of any topographic relief, even for incident wavelengths several times the height of the step. Both vertical and inclined steps are examined.
TL;DR: In this paper, a vertical seismic array at the Richmond Field Station on the margin of San Francisco Bay, California, has provided some excellent experimental data showing the effects of unconsolidated sediments on ground motion during local earthquakes.
Abstract: A vertical seismic array at the Richmond Field Station on the margin of San Francisco Bay, California, has provided some excellent experimental data showing the effects of unconsolidated sediments on ground motion during local earthquakes. The experiments are well controlled in that good estimates of the material properties of the sediments are available, with the largest uncertainty being in the attenuation. Three components of acceleration have been recorded in bedrock below the sediments, within the sediments, and at the surface of the sediments from two local earthquakes. The maximum accelerations at the surface are between 1.5 and 4.3 times those in the bedrock. Analytical calculations which assume vertically propagating waves and which are suitable for highly attenuating materials explain the major differences between the bedrock records and the surface records for both vertical and horizontal components. Maximum accelerations, duration, frequency content, general character of the records, and response spectra are all in approximate agreement. Some of the energy arriving late in the records can best be explained by surface waves propagating horizontally in the sediments.
TL;DR: In this paper, a detailed body-wave analysis of the rupture process of two deep earthquakes was carried out by deconvolving the long and short-period channels to produce broadband records of body wave ground displacement and velocity.
Abstract: The digital data base provided by the SRO and ASRO stations is exploited in a detailed body-wave analysis of the rupture process of two deep earthquakes. The data are processed by simultaneously deconvolving the long- and short-period channels to produce broadband records of body-wave ground displacement and velocity. These pulse shapes show significant complexity and directivity with frequency content up to 5 Hz. To model these waveforms theoretical seismograms are generated by using the full wave theory to simulate the effects of propagation in the Earth, and by using dynamically realistic source models, where the source complexity is modeled as a sequence of simple subevents. In order to model the high-frequency content of the body waves, the AFL Q model is used, as Q models derived from long-period data are inadequate. By combining the relative locations of the subevents with the inferred rupture areas, we determine complete rupture histories for the events, including estimates of the dynamic and static stress drops (between 40 and 60 bars), the rupture velocity, and the rupture geometry. The results of this modeling are interpreted in terms of the barrier model proposed by Das and Aki in which the rupture complexity is controlled by the distribution of fracture strength.