Dynamic Source Model Research Articles

Radiated seismic energy based on dynamic rupture models of faulting

By modeling spontaneous ruptures, we study the mechanism dependence of radiated seismic energy from three hypothetical crustal events, 30° dipping reverse fault, 60° dipping normal fault, and a vertical strike‐slip fault, and the 1994 blind‐thrust Northridge earthquake. Embedded in a homogeneous half‐space, all three hypothetical faults have the same area and are subjected to the same shear and normal stress conditions and frictional parameters. Dynamic simulations produce apparent stress of 0.53 MPa, 0.23 MPa, and 0.34 MPa for the reverse, normal, and strike‐slip faults, respectively. The energy distribution on a distant surface shows that a large fraction of energy is concentrated in the forward direction of rupture propagation. We use spontaneous rupture models to compute the radiated energy from the 1994 Northridge earthquake. The initial stress drop distribution is based on a kinematic slip distribution. Using a linear slip‐weakening friction law, we modify both the initial stress and yield stress until the dynamic rupture produces near‐source synthetics that are consistent with the data. The total radiated seismic energy from our model is 6.0 × 1014 J; seismic moment 1.47 × 1019 Nm; apparent stress 1.5 MPa; fracture energy 3.2 × 1014 J; and slip‐weakening distance 0.25 m. The energy flux distribution is heterogeneous with strong directivity effects. These results suggest that correcting for directivity could be difficult, but necessary, for teleseismic and regional estimates of radiated energy. Dynamic source models constrained by ground motions can provide a stable and accurate energy estimate for large earthquakes.

Strong Ground-Motion Prediction from Stochastic-Dynamic Source Models

In the absence of sufficient data in the very near source, predictions of the intensity and variability of ground motions from future large earthquakes depend strongly on our ability to develop realistic models of the earthquake source. In this article we simulate near-fault strong ground motion using dynamic source models. We use a boundary integral method to simulate dynamic rupture of earthquakes by specifying dynamic source parameters (fracture energy and stress drop) as spatial random fields. We choose these quantities such that they are consistent with the statistical properties of slip heterogeneity found in finite-source models of past earth- quakes. From these rupture models we compute theoretical strong-motion seismo- grams up to a frequency of 2 Hz for several realizations of a scenario strike-slip Mw 7.0 earthquake and compare empirical response spectra, spectra obtained from our dynamic models, and spectra determined from corresponding kinematic simulations. We find that spatial and temporal variations in slip, slip rise time, and rupture prop- agation consistent with dynamic rupture models exert a strong influence on near- source ground motion. Our results lead to a feasible approach to specify the vari- ability in the rupture time distribution in kinematic models through a generalization of Andrews' (1976) result relating rupture speed to apparent fracture energy, stress drop, and crack length to 3D dynamic models. This suggests that a simplified rep- resentation of dynamic rupture may be obtained to approximate the effects of dy- namic rupture without having to do full dynamic simulations.

Effects of resonant pumping on the temporal and spatial evolution of a laser produced lithium plasma

The spatial and temporal distribution of neutral and singly ionized lithium atoms has been investigated when slab targets are irradiated by ∼1 μs pulses of both resonant and nonresonant radiation from a flashlamp pumped dye laser. The ion population is enhanced when the laser is tuned to resonance with the 2s–2p line of neutral lithium and there is evidence for efficient ion production at distances up to 7.5 mm from the target surface. In particular, the rate of ionization is increased and the maximum ion population is attained between 120 and 200 ns earlier in the resonant case depending on distance above the target surface. However, the on-resonance ion population close to the surface is found to decrease significantly some 200 ns into the laser pulse. Ion and neutral drift velocities were measured by direct observation of the surface of the expanding plasma plume at a laser flux close to 8×108 W cm−2 and were consistent with the value predicted by a dynamic source model for accelerated plasma expansion, as was the value for the electron temperature derived from line intensity ratios. Power density thresholds for ion production have been studied and in particular, the threshold for Li2+ production is established.

Resonance of a fluid‐driven crack: Radiation properties and implications for the source of long‐period events and harmonic tremor

A dynamic source model is presented, in which a three‐dimensional crack containing a viscous compressible fluid is excited into resonance by an impulsive pressure transient applied over a small area ΔS of the crack surface. The crack excitation depends critically on two dimensionless parameters called the crack stiffness, C = (b/μ)(L/d), and viscous damping loss, F = (12ηL)/(ρƒd2α), where b is the bulk modulus, η is the viscosity, ρƒ is the density of the fluid, μ is the rigidity, α is the compressional velocity of the solid, L is the crack length, and d is the crack thickness. The first parameter characterizes the ability of the crack to vibrate and shapes the spectral signature of the source, and the second quantifies the effect of fluid viscosity on the duration of resonance. Resonance is sustained by a very slow wave trapped in the fluid‐filled crack. This guided wave, called the crack wave, is similar to the tube wave propagating in a fluid‐filled borehole; it is inversely dispersive, showing a phase velocity that decreases with increasing wavelength, and its wave speed is always lower than the acoustic velocity of the fluid, decreasing rapidly as the crack stiffness increases. The source spectrum shows many sharp peaks characterizing the individual modes of vibration of the crack; the variation of spectral shape, both in the number and width of peaks, is surprisingly complex, reflecting the interference between the lateral and longitudinal modes of resonance, as well as nodes for these modes. The far‐field spectrum is marked by narrow‐band dominant and subdominant peaks that reflect the interaction of the various source modes. The frequency of the dominant spectral peak radiated by the source is independent of the radiation direction. The frequency, bandwidth, and spacing of the resonant peaks are strongly dependent on the crack stiffness, larger values of the stiffness factor shifting these peaks to lower frequencies and decreasing their bandwidth. The excitation of a particular mode depends on the position of the trigger and on the extent of the crack surface affected by the pressure transient. Fluid viscosity decreases the amplitudes of the main spectral peaks, smears out the finer structure of the spectrum, and greatly reduces the duration of the radiated signal. The energy loss by radiation is stronger for high frequencies, producing a seismic signature that is marked by a high‐frequency content near the onset of the signal and dominated by a longer‐period component of much longer duration in the signal coda. Such signature is in harmony with those displayed by long‐period events observed on active volcanoes and in hydrofracture experiments. The very low velocity which is possible in a crack with high stiffness (C ≥ 100) also provides an attractive explanation for very long period tremor, such as type 2 tremor at Aso volcano, Japan, without the requirement of an unrealistically large magma container. The standing wave pattern set up on the crack surface by the sustained resonance in the fluid is observable in the near field of the crack, suggesting that the location and extent of the source may be estimated from the mapping of the pattern of nodes and antinodes seen in its vicinity. According to the model, the long‐period event and harmonic tremor share the same source but differ in the boundary conditions for fluid flow and in the triggering mechanism setting up the resonance of the source, the former being viewed as the impulse response of the tremor generating system and the latter representing the excitation due to more complex forcing functions.

被害調査から推定される地震動の卓越方向と地震断層モデル

Modes and degrees of the damages caused by the Onikobe earthquake (M=4.9) of July 5, 1976 were investigated in special reference to the predominant direction of the ground motion. The directions of falling and slip of gravestones near the focal region were restricted to a narrow range of directions and these data were found useful for estimating the predominant direction of the ground motion. On the basis of the seismological data such as the hypocenter location, the focal mechanism and aftershock distribution, various dynamicsource models were tested if they could explain the predominant direction near the focal region. As a result, precise determination of the hypocenter location became possible, and the finally selected model explained the systematic pattern of the predominant directions observed near and around the focal region.

A dynamic source model for the San Fernando earthquake

abstractWe model the San Fernando earthquake as a propagating rupture in a half-space, using for the slip-time-history on the fault plane analytical expressions which approximate the slip functions of dynamic crack models obtained by Das and Aki (1977a, b). We synthesize the strong ground motions and accelerations at the Pacoima Dam site and compute the teleseismic signals for different models of cracks. Three major featuras of the data–the strong pulse associated with the beginning of the rupture, the high acceleration phase on the Pacoima Dam records, and the presence of ripples on the teleseismic seismograms–which are not compatible with a smooth rupture process, are well explained by a crack with barriers model where the rupture encounters, along the fault plane, barriers or obstacles of high strength materials which may remain unbroken after the passage of the rupture front. A high-stress drop (400 to 500 bars) is required in the hypocentral area to explain the high-amplitude short-duration first pulse of the teleseismic records. This indicates a high level of tectonic stress in the area. A study of the earthquake series following the main shock shows that the aftershocks which took place in the region where major slipping occurred during the earthquake may represent the release of some of the barriers.