ALBERT

Description: This paper investigates which mechanism of Lg generation dominates in low-velocity source media, which is important because of the central role of Lg in discrimination and yield estimation of nuclear explosions. The mechanisms investigated are surface P -to- S conversion ( pS ), generation directly by the nonspherical component of the explosion source volume, and Rg -to- S scattering. We identify and quantify observations that distinguish between mechanisms. We also specifically test the assumptions of previous work that concluded that Rg scattering is the dominant mechanism. To do so, we analyze and simulate records of adjacent, normally buried and overburied Nevada test site (NTS) explosions, and analyze deep seismic sounding (DSS) explosion Quartz 3 data. The data analyses and simulations consistently indicate that pS is the dominant source of explosion Lg in low-velocity source media, that nonspherical source components could also contribute significantly to Lg , and that scattered Rg contributes less, except possibly at very low frequencies. For NTS overburied versus normally buried explosions, we compare Lg -to- Pg spectral ratios, corner frequencies, and tangential versus vertical and radial Lg spectral nulls. We perform simulations for the NTS to compare the contributions to Lg of pS , direct S from a CLVD, and scattered Rg . Quartz 3 data show that Rg spectral nulls vary with azimuth and differ from corresponding Sg and Lg spectral nulls, counter to assumptions required by the Rg scattering hypothesis.

Description: We evaluate the mechanisms responsible for generation of shear waves by explosions in high-velocity source media by identifying, quantifying, and modeling observations that can distinguish between commonly suggested mechanisms. We review the literature to identify regional observations that have been or can be used to distinguish between two or more mechanisms. We supplement these historical observations with new measurements of the Semipalatinsk test site (STS) event Lg and P amplitudes at Borovoye and model the observations with nonlinear source models, Rg -to- Lg upper bound calculations, and wavenumber integration synthetic seismograms for point explosions and CLVDs. Direct generation of shear waves by the nonspherical component of the source volume is consistent with the regional Lg amplitude versus yield relationship, while S * and Rg -to- Lg scattering are not. We also analyze and model a large set of Degelen explosion records from approximately 10 to 90 km. The local Sg spectral corner frequency is lower than the Pg corner frequency by approximately the source P -to- S velocity ratio, which is consistent with shear waves directly generated by the source, and inconsistent with Sg being the result of pS , S * , or Rg -to- Lg scattering. The local Sg and Rg spectra are distinctly different. Taken together, results from previous work and new observations presented here support the conclusion that explosions in high-velocity source media dominantly generate shear waves directly, through the nonspherical part of the nonlinearly deforming source volume.

Description: Pulse-like ground motions, which are mainly concentrated in near-fault regions, can induce severe damage on structures. This article presents a simple and effective method used for quantitatively identifying pulse-like ground motions. First, zero velocity point method is proposed based on the properties of trigonometric functions. This method can reveal the time-domain characteristics of ground-motion velocity time history and can be used to detect the potential velocity pulse which is included in corresponding original ground motion. Second, pulse-like ground motions are classified into two classes: single-pulse-like and multi-pulse-like ground motions. Further, single-pulse-like is divided into significant-single-pulse-like and non-significant-single-pulse-like, while multi-pulse-like is divided into double-pulse-like, three-pulse-like, and four-pulse-like. Third, energy method is adopted to formulate the identification criterion for each type of pulse-like ground motion. Finally, the energy parameters of the detected potential velocity pulse for a large number of ground motions are calculated, and the proposed identification criteria are applied to identify whether these ground motions are pulse-like or not. A comparison of the methodology proposed in this study with previous identification methods is presented. Analysis results indicate that the methodology proposed in this study can accurately distinguish pulse-like and non-pulse-like ground motions.

Description: A large earthquake (M (sub w) 7.6) occurred near Chi-Chi, Taiwan, on 20 September 1999 (UTC) and was followed by many moderate-size aftershocks in the following days. The two largest aftershocks with magnitudes M (sub w) 6.1 and 6.2, respectively, were used as empirical Green's functions (EGFs) to retrieve the source time functions (STFs) and further image the temporal-spatial rupture process of the mainshock. For each station, two types of STFs were retrieved, one from P phases and another from S phases. A total of 178 STF individuals were retrieved for source process analysis of the event. From the STFs retrieved, firstly, similarities appeared on the STFs from most stations except that several STFs in special azimuths looked different or odd because of the focal mechanism difference between the mainshock and the EGF aftershocks; and secondly, systematic shape-variation with azimuth appeared. The analysis of the STFs indicated that this event consisted of two subevents; on the average, the second event was about 7 sec later than the first one, and the first event was about 15% larger than the second one in the moment-release rate. The total duration time of earthquake rupture process was about 26 sec. From the image of the static slip distribution, there were two slip-concentrated areas on the fault plane, and their centers were about 45 km away from each other. The maximum slip of about 6.5 m appeared on the northern one. The rupture area, with the slip greater than 0.5 m, was about 80 km long and 60 km wide. The maximum stress drop was about 25 Mpa (250 bar), and the average stress drop on the entire fault was 9.2 MPa (92 bar). From the snapshots of the temporal-spatial variation of slip and slip rate, the rupture initiated at the southern end and stopped at the northern end of the fault, which suggested an overall unilateral rupture at a rupture velocity of about 2.5 km/sec. The depth of the initiation locus was about 20 km. The rupture duration times on the subfaults were estimated. A good correlation was noticed between the rupture duration time and slip amplitude of the subfaults: the larger the slip amplitude, the longer the rupture duration time of the subfault. The maximum duration time for the subfaults was 16 sec, which was about two-thirds the total duration time of the earthquake rupture process. In the beginning and near the end of the source process, the rupture propagated like a self-healing pulse of slip.