S43A-01
Statistical seismology helps to constrain fault plane characteristics: New aspects of the Mw=6.9 Cariaco earthquake
The spatial distribution of the seismological parameter b obtained from the Gutenberg-Richter relation can map out changes in the state of stress and/or material properties. This dual capability of the seismic b-value has been utilized in laboratory experiments and seismotectonic studies to infere fault structures which can influence earthquake rupture initiation and propagation. The local b-value study we performed using the aftershock sequence of the 1997, Mw=6.9 earthquake of Cariaco- Casanay in North-eastern Venezuela determines the heterogeneity of the rupture plane and indicates areas of high co-seismic moment release and stress drop. Our new contribution is a slip distribution from a 3D finite element model of the co-seismic rupture. A best fit model is obtained by taking as independent constraints the co-seismic slip observed by 3 GPS stations. Correlating b-value map and slip distribution shows a clear coincidence of high post-seismic b-value patches with areas of large co-seismic slip. We further examined radiated seismic energy and focal mechanisms of aftershocks to look for evidences on fault plane segmentation and took previous local tomography studies as a reference for changes in material properties. In our presentation we will give an interpretation of the observed parameter distributions and correlations along the rupture plane in terms of source characteristics and rupture propagation. First results seem to favour a bi- lateral co-seismic rupture propagation.
S43A-02
Slip-Weakening Distance Estimated from Near-Fault Stations
We estimate the slip-weakening distance directly from the seismograms recorded at near-fault stations, taking into account its estimation error. To do this, probable spatio-temporal smoothing effects during continuous rupture propagation on the fault are evaluated through numerical modeling. For this purpose, we use a 2-D in- plane shear crack propagating with a constant rupture velocity and with a Yoffe-type source time function, which is a kinematically good approximation of dynamic slip weakening behavior. From this modeling, the displacement and velocity waveforms expected at short distances away from the fault are calculated. As a proxy of a real slip- weakening distance Dc, Dcf, which is defined as a slip at the time of peak slip velocity, can be calculated as a function of distance from the fault. This technique is applied to the observed records from the 2000 western Tottori, Japan, earthquake (Mw6.6) and also from the 2002 Denali, Alaska, earthquake (Mw7.9), both of which are mainly strike-slip events. We estimated Dcf at GSH station (~100 m away from the fault) as about 0.3 m for the Tottori earthquake, and that at PS10 station (~3 km from the fault) as about 2.5 m for the Denali earthquake, within a possible error of about 30 %. We confirmed that these estimates are not much affected by the spatio-temporal smoothing effects.
S43A-03
Nonextensivity in Geological Faults?
Geological fault systems, as the San Andreas fault (SAF) in USA, constitute typical examples of self-organizing systems in nature. In this work, we have considered some geophysical properties of the SAF system to test the viability of the nonextensive model for earthquakes developed in [R. Silva, G.S. França, C. Vilar, J.S. Alcaniz, Phys. Rev. E 73 (2006) 026102]. To this end, we have selected earthquake events, ranging in the magnitude interval 2 < m < 8, from the National Earthquake International Center catalogs (NEIC, 2004 - 2006) and the Bulletin of the International Seismological Centre (ISC, 1964 - 2003). For values of the nonextensive parameter q ~ 1.68, it is shown that the energy distribution function deduced in above reference provides an excellent fit to the NEIC and ISC SAF data. A brief discussion on a possible influence of the fault mechanisms (reverse, normal or strike-slip) on the predictions of this nonextensive model is also presented.
S43A-04
Seismic source parameters and geodynamical features in the Vrancea region, Romania
The earthquakes clustered in a confined volume at intermediate depths beneath the Vrancea (Romania) seismic region are severely affecting an extended area of Romania and its neighboring countries. During the last century, five events with magnitudes larger than 6.5 occurred within this narrowly confined focal volume. A better understanding of the source process of Vrancea earthquakes is of crucial importance in order to properly assess the ground motion level that has to be expected from future large events and to model the geodynamics of this complex intra-continental convergence area. A substantial new amount of high-quality earthquake data have been recently gained through the progress of seismic networks on the Romanian territory within the cooperation programme with the University of Karlsruhe (Germany): Collaborative Research Centre 461 programme (Bonjer et al., 2000) and the tomography experiment CALIXTO'99 (Wenzel et al., 1999). On the basis of the new digital accelerometer and velocity data (Kinemetrics K2 and broadband instruments) we retrieve the source parameters of small-to-moderate earthquakes (duration magnitude below 5.8) generated in the Vrancea subducting slab. Empirical Green's function deconvolution and spectral ratio methods are applied for a set of 130 earthquakes. Pairs of collocated events with similar focal mechanism are selected in order to retrieve source parameters and to inspect the source scaling properties over the entire seismic active depth domain (between 60 and 180 km). The tests of stability of the source time function parameters are carried out using all the elements (9 short-period and 1 broad band borehole instruments over a 5km x 5km area) of the BURAR array, installed in the northern part of Romania (Bucovina area) within a cooperation with Air Force Technical Applications Center - Florida (USA). Source parametrization and scaling are correlated with clustering properties of earthquake time, space and size distributions as well as with differences in specific source mechanisms or/and structural inhomogeneity properties along the subducting lithosphere. Our results confirm previous research focused on seismicity, tomography and seismic source scaling in the Vrancea region revealing significant subducting slab irregularity which was assumed to be in connection with differences in the physical, geochemical and tectonic processes at different scale lengths.
S43A-05
Stress regime from active earthquake faulting in the Basin of Mexico
The Trans-Mexican Volcanic Belt is characterized by E-W fault systems forming structural features such as graben and semi-grabens. Some of these structures are associated with seismic events, indicating active faulting. In this study earthquake fault parameters have been determined using broad-band (BB) data and first motion polarities in the Basin of Mexico. Simultaneous inversion of P-wave polarities and near-field BB waveform modeling using a grid-search approximation was performed and source parameters of 29 low-to-moderate-size earthquakes located within the Basin of Mexico were estimated. The stress field in the Basin of Mexico is basically defined by the principal stress sigma-3 oriented N346° and the principal stress sigma-1 oriented N140° plunging at 38° and 49° respectively. The stress field along the Central TMVB is continuous with small gradual rotations in the sigma-1. Our results from focal mechanisms associate with E-W fault system confirm that the Basin of Mexico is a transtensional stress regime probably related to slip partition from the Mid American Trench.
S43A-06
A study of source parameters and attenuation of earthquakes in a regime of subduction in the east of Venezuela using Genetic Algorithms
A sample of 50 earthquakes associated with the regime of subduction in the east of Venezuela has been analyzed in order to determine the following source parameters: seismic moment, stress drop and attenuation Q. The spectral content of the seismograms of this sample was studied with Brune´s model as a basic reference, while the numerical treatment- by means of Genetic Algorithms. The studied seismograms were obtained through the network of the Venezuelan Seismological Foundation, with FUNVISIS broadband sensors detecting noise signals for calculating Q. The spectral analysis of seismograms of this sample reveals preliminary values for M0 between 6.E+9 to 8.5E+12 Nm. When considering attenuation Q for frequencies of 2, 5, 4, 8 and 10 Hertz, a variable Q was obtained from 125 to 590, indicating a different tectonic behavior compared to a sample of superficial seismicity of greater attenuation for the same geographic region. Additionally, the correlation between Q(P) and Q(S) for the earthquakes of the relevant region was investigated.