Applications of the Generalized Finite Differences Method in Viscoelastic Media

Abstract:

The meshless generalized finite difference method is applied to solve problems of seismic wave propagation in viscoelastic media, with irregular distribution of nodes. Schemes in generalized finite differences are obtained for the decoupled wave systems P-SV and SH in homogeneous media, and the stability of the schemes is analyzed. These schemes are used to solve certain particular case and the accuracy and easyness of its application is shown.

Media has not been submited for this Presentation

2-D Modeling of the Crustal Structure of Merida Andes, Venezuela, from Wide-Angle Seismic and Gravity Studies

Abstract:

The Merida Andes (MA) is an orogeny delimiting the Maracaibo block to the west, in response to the subduction of the Caribbean Plate beneath South American continent. 2-D seismic and gravity modeling of the crustal structure of MA was carried out along the profile Central Andes with a length of 380 km from the coast of Falcon, crossing perpendicular to the mountain range of the MA until its southeastern end in the Barinas-Apure basin. In the seismic data obtained from 11 shot points with 0.2 to 1 tons of explosive charges and recorded by 480 Texan recorders, we observed critical PmP reflections at distances of about 60 km for the northern part of the profile (Falcón Basin), and 120 to 90 km for the central and southern part of the profile, corresponding to the axis of the orogen and the Barinas Basin, respectively. Derived values for the depth of the Moho discontinuity range from 29 km depth for the Falcon basin, to 40-53 km close to the core of the chain, and 35 km at the southern end of the profile in Barinas basin. The crustal root is defined with a maximum depth of 53 km, which is displaced with respect to the highest part of the chain at this segment, approximately 10 km towards the northwest. 325 gravity stations were acquired along the profile and modeled together with satellite gravity data in a high resolution 2D gravity forward model, which confirms the strong variations of the thickness of the crustal root and its asymmetry. Cenozoic sediments with a P-wave velocity (Vp) ranging from 2 to 4 km/s are underlain by Cretaceous sediments, (Vp = 4 to 5 km/s) with a maximum depth of 10 and 6 km in the Maracaibo and Barinas-Apure basins, respectively. The crystalline basement shows Vp between 5.5 and 6.3 km/s and a density of 2.78 g/cm3 down to about 15 km beneath the basins and 25 km beneath the orogen. The lower crust is modeled with Vp of 6.5 to 7 km/s and a density of 2.84 g/cm3, underlain by the upper lithospheric mantle, with Vp greater than 7.7 km/sand a density of 3.22 g/cm3. The density model suggests the existence of an incipient A-subduction of continental South America towards the Maracaibo block.

Media has not been submited for this Presentation

Anatomy of Old Faithful from Subsurface Seismic Imaging of Yellowstone National Park, Upper Geyser Basin

Abstract:

The Upper Geyser Basin (UGB) in Yellowstone contains one of the highest concentrations of hydrothermal features on Earth including the iconic Old Faithful geyser. Although this area has been the focus of many geological, geochemical, and geophysical studies, the shallow (<200 m) subsurface structure remains poorly characterized due to limited instrument implementations in this delicate and sensitive environment. The recent availability of seismic dense arrays (large-N) permits an environmental-friendly approach to investigate the detailed crustal structure from the low-cost and easy-deployed geophones. To probe the detailed structure in relation to the hydrothermal plumbing of the UGB, we deployed large-N arrays of 3C 5-Hz geophones in both November of 2015 and 2016, composed of 133 stations with ~50 m spacing, and 519 station locations, with an ~20 m spacing, respectively. We constructed cross-correlation functions (CCFs) and extracted Rayleigh-wave signals between 1-10 Hz via seismic signals excited by nearby hydrothermal features. We observe a clear lateral velocity boundary at 3.3 Hz frequency that delineates a higher phase velocity of ~1.6 km/s in the NE and a lower phase velocity of ~1.0 km/s in the SW corresponding to the local geologic formation of rhyolitic and glacial deposits, respectively. We also image a relatively shallow (10-60 m deep) large reservoir with an estimated porosity 30% located ~100 m southwest of Old Faithful from the significant spatial-dependent waveform distortions and delays between 5-10 Hz. This reservoir is likely controlled by the local geology with a rhyolitic deposit in the NE acting as a relatively impermeable barrier to vertical fluid ascent. In addition to the static structure, we observe temporal variations in both phase and amplitude from the minutely CCFs with regard to the potential influences from instrument resonance, seismic source, and structure. The preliminary results of variations will be demonstrated and discussed.

Media has not been submited for this Presentation

Jizhaigou ms7.0 Earthquake High-Frequency Ground Motion Simulation and Zhangzha Town Intensity Estimation

Abstract:

Jiuzhaigou occurred Ms7.0 earthquakes which caused serious human and economic losses. We simulated the strong ground motion characteristic using the Empirical Green Function Method. We also estimated the ground motion intensity of Zhangzha town which lost earthquake record data. All the stations are within the scope of around 100 km and all the stations’ PGA are larger than 10gal. Because of lacking appropriate aftershock record data, we try to use the aftershocks of Wenchuan earthquake and Dingxi earthquake as green function to simulate Jiuzhaigou earthquake in the first time. The simulated results as a whole can indicate the characteristics of the ground motion intensity especially the high frequency component. PGA, Duration and Response Spectrum are also fitted well between the simulated values and observed values. This attempt also indicated that using other earthquake’ s aftershock to simulate this event is feasible when lacking aftershock records. The PGA of Zhangzha town is estimated about 180gal-300gal. We also discussed the criterion about how to choose aftershock as green function.

Media has not been submited for this Presentation

Towards Quasi-Automated Estimates of Directivity and Related Source Properties of Small to Moderate Southern California Earthquakes with Second Seismic Moments Using Weighted Stacked EGFs

Abstract:

We develop a method for quasi-automated estimation of directivity, rupture area, duration, and centroid velocity of earthquakes with second seismic moments. The method is applied to small to moderate earthquakes in southern California. P and S phase picks are given by a 1-D ray tracing algorithm and cataloged event locations. These are refined for deconvolution by using a grid search on zero-crossings within a short time window around the automated P/S picks. Source Time Functions (STFs) of target events are derived using deconvolution with a stacked empirical Green’s function (seGf). The use of seGf suppresses non-generic source effects such as directivity in individual eGf’s. The seGf for each target event is based on stacking individual eGfs (normalized by seismic potencies) selected by spatial and magnitude criteria as well as performances in the projected Landweber deconvolution. A weighted stack of eGfs, with weight coefficients grid searched and determined by waveform fits, helps further to correct inaccuracies of focal mechanisms. Compared with a single eGf, analysis with a weighted stack can significantly improve waveform fit and typically allows getting STFs at 5-10 more stations. The method is suitable for analysis of large seismic datasets and it works for target events in southern California with magnitudes as small as 3.5. Most events analyzed so far have significant directivities.

Media has not been submited for this Presentation

Evaluating inversion techniques for soft sites in Windsor, Ontario, Canada

Abstract:

This paper will illustrate a joint inversion analysis of the ambient noise data collected at six soft sites in Windsor, Ontario, Canada, using two different joint inversion methods. We also have geotechnical (borehole and CPT) data near the sites for verifying the applicability of these methods. For both the methods we will take fundamental mode Rayleigh wave dispersion curve and H/V curve as an input. For the first method, most of energy propagation is assumed to be in the vertical direction. VS profile estimation will be done by matching the input curves with theoretical fundamental mode Rayleigh wave dispersion curve and the theoretical Rayleigh wave ellipticity curve, arriving at a best fit model iteratively. This inversion technique is based on the Neighborhood Algorithm (NA), which is a stochastic direct search method to find a global minimum. The previous search results guide the newer searches to find local minima and then these local minima are analysed to arrive at a global minimum. For the second method, the theoretical fundamental mode Rayleigh wave dispersion curve will be matched iteratively with the input dispersion curve, and the theoretical H/V curve (comprising of Rayleigh, Love, and/or body wave contributions) will be matched iteratively with the input H/V curves to minimise the misfit and arrive at a best fit model to get an estimate of VS profile. This method is based on the theory of diffuse field assumption (DFA), which states that a diffusion-like field is generated when random seismic sources generate ambient noise. In DFA, the Green’s function is proportional to the autocorrelation of seismic wavefields in frequency domain. VS depth profiles estimated by these two methods will be then validated using the geotechnical data. This study will help us in understanding the contribution of different wave types in the H/V curves calculated using the ambient noise data and in determining the most efficient and applicable method for site assessment.

Media has not been submited for this Presentation