Applying Bat Metaheuristic Algorithm for Building Shear Wave Velocity Models from Surface Wave Dispersion Curves

2017 ◽  
Author(s):  
R. Poormirzaee
Keyword(s):  
Surface Wave ◽  
Shear Wave ◽  
Wave Velocity ◽  
Shear Wave Velocity ◽  
Wave Dispersion ◽  
Dispersion Curves ◽  
Metaheuristic Algorithm ◽  
Surface Wave Dispersion ◽  
Velocity Models

Using Deep Learning to Derive Shear-Wave Velocity Models from Surface-Wave Dispersion Data

2020 ◽  
Vol 91 (3) ◽  
pp. 1738-1751
Author(s):  
Jing Hu ◽  
Hongrui Qiu ◽  
Haijiang Zhang ◽  
Yehuda Ben-Zion
Keyword(s):  
Surface Wave ◽  
Shear Wave ◽  
Wave Velocity ◽  
Shear Wave Velocity ◽  
Southern California ◽  
Wave Dispersion ◽  
Surface Wave Dispersion ◽  
Velocity Models ◽  
Dispersion Data ◽  
Phase Group

Abstract We present a new algorithm for derivations of 1D shear-wave velocity models from surface-wave dispersion data using convolutional neural networks (CNNs). The technique is applied for continental China and the plate boundary region in southern California. Different CNNs are designed for these two regions and are trained using theoretical Rayleigh-wave phase and group velocity images computed from reference 1D VS models. The methodology is tested with 3260 phase–group images for continental China and 4160 phase–group images for southern California. The conversions of these images to velocity profiles take ∼23  s for continental China and ∼30  s for southern California on a personal laptop with the NVIDIA GeForce GTX 1060 core and a memory of 6 GB. The results obtained by the CNNs show high correlation with previous studies using conventional methods. The effectiveness of the CNN technique makes this fast method an important alternative for deriving shear-wave velocity models from large datasets of surface-wave dispersion data.

P- and S-wave velocity models from surface wave dispersion curves data transform

2017 ◽  
Author(s):  
Valentina Socco ◽  
Farbod Khosro Anjom ◽  
Cesare Comina ◽  
Daniela Teodor
Keyword(s):  
Surface Wave ◽  
Wave Velocity ◽  
Wave Dispersion ◽  
Dispersion Curves ◽  
S Wave ◽  
Surface Wave Dispersion ◽  
Velocity Models

scholarly journals Probabilistic neural network tomography across Grane field (North Sea) from surface wave dispersion data

2020 ◽  
Vol 223 (3) ◽  
pp. 1741-1757
Author(s):  
S Earp ◽  
A Curtis ◽  
X Zhang ◽  
F Hansteen
Keyword(s):  
Neural Network ◽  
Monte Carlo ◽  
Surface Wave ◽  
Shear Wave ◽  
Wave Velocity ◽  
North Sea ◽  
Shear Wave Velocity ◽  
Wave Dispersion ◽  
Surface Wave Dispersion ◽  
Dispersion Data

SUMMARY Surface wave tomography uses measured dispersion properties of surface waves to infer the spatial distribution of subsurface properties such as shear wave velocities. These properties can be estimated vertically below any geographical location at which surface wave dispersion data are available. As the inversion is significantly non-linear, Monte Carlo methods are often used to invert dispersion curves for shear wave velocity profiles with depth to give a probabilistic solution. Such methods provide uncertainty information but are computationally expensive. Neural network (NN) based inversion provides a more efficient way to obtain probabilistic solutions when those solutions are required beneath many geographical locations. Unlike Monte Carlo methods, once a network has been trained it can be applied rapidly to perform any number of inversions. We train a class of NNs called mixture density networks (MDNs), to invert dispersion curves for shear wave velocity models and their non-linearized uncertainty. MDNs are able to produce fully probabilistic solutions in the form of weighted sums of multivariate analytic kernels such as Gaussians, and we show that including data uncertainties as additional inputs to the MDN gives substantially more reliable velocity estimates when data contains significant noise. The networks were applied to data from the Grane field in the Norwegian North sea to produce shear wave velocity maps at several depth levels. Post-training we obtained probabilistic velocity profiles with depth beneath 26 772 locations to produce a 3-D velocity model in 21 s on a standard desktop computer. This method is therefore ideally suited for rapid, repeated 3-D subsurface imaging and monitoring.

scholarly journals Shear-wave Velocity Structure of Greenland from Rayleigh-wave Analysis

2016 ◽  
Vol 20 (1) ◽  
pp. 1-11 ◽  
Author(s):  
V. Corchete
Keyword(s):  
Rayleigh Wave ◽  
Shear Wave ◽  
Wave Velocity ◽  
Shear Wave Velocity ◽  
Velocity Structure ◽  
Wave Dispersion ◽  
Dispersion Curves ◽  
Depth Range ◽  
Velocity Models ◽  
Rayleigh Wave Dispersion

<p>The elastic structure beneath Greenland is shown by means of S-velocity maps for depths ranging from zero to 350 km, determined by the regionalization and inversion of Rayleigh-wave dispersion. The traces of 50 earthquakes, occurring from 1990 to 2011, have been used to obtain Rayleigh-wave dispersion data. These earthquakes were registered by 21 seismic station located in Greenland and the surrounding area. The dispersion curves were obtained for periods between 5 and 200 s, by digital filtering with a combination of MFT (Multiple Filter Technique) and TVF (Time Variable Filtering). Later, all seismic events (and some stations) were grouped to obtain a dispersion curve for each source-station path. These dispersion curves were regionalized and inverted according to the generalized inversion theory, to obtain shear-wave velocity models for a rectangular grid of 16x20 points. The shear-velocity structure obtained through this procedure is shown in the S-velocity maps plotted for several depths. These results agree well with the geology and other geophysical results previously obtained. The obtained S-velocity models suggest the existence of lateral and vertical heterogeneity. The zones with consolidated and old structures present greater S-velocity values than the other zones, although this difference can be very little or negligible in some case. Nevertheless, in the depth range of 15 to 45 km, the different Moho depths present in the study area generate the principal variation of S-velocity. A similar behaviour is found for the depth range from 80 to 230 km, in which the lithosphere-asthenosphere boundary (LAB) generates the principal variations of S-velocity. Finally, the new and interesting feature obtained in this study: the definition of the base of the asthenosphere (for the whole study area and for depths ranging from 130 to 280 km, respectively) should be highlighted.</p><p> </p><p><strong>Estructura de velocidad de cizalla de Groenlandia obtenida de análisis de onda Rayleigh</strong></p><p><strong><br /></strong></p><p><strong>Resumen</strong></p><p>La estructura elástica bajo Groenlandia es mostrada por medio de mapas de velocidad de onda para profundidades variando desde cero a 350 km, determinada por la regionalización e inversión de la dispersión de onda Rayleigh. Las trazas de 50 terremotos, ocurridos desde 1990 hasta 2011, han sido usados para obtener datos de dispersión de onda Rayleigh. Estos terremotos fueron registrados por 21 estaciones sísmicas localizadas en Groenlandia y el área circundante. Las curvas de dispersión fueron obtenidas para periodos entre 5 y 200 s, por filtrado digital con una combinación de MFT (Técnica de Filtrado Múltiple) y TVF (Filtrado en Tiempo Variable). Después, todos los eventos sísmicos (y algunas estaciones) fueron agrupados para obtener una curva de dispersión para cada trayecto fuente-estación. Estas curvas de dispersión fueron regionalizadas e invertidas de acuerdo con la teoría de la inversión generalizada, para obtener modelos de velocidad de cizalla para una rejilla rectangular de 16x20 puntos. La estructura de velocidad de cizalla obtenida a través de este procedimiento es mostrada in los mapas de velocidad de onda S representados para varias profundidades. Estos resultados muestran buen acuerdo con la geología y con otros resultados geofísicos obtenidos previamente. Los modelos de velocidad de onda S obtenidos sugieren la existencia de heterogeneidad lateral y vertical. Las zonas con estructuras antiguas y consolidadas presentan mayores valores de velocidad de onda S que las otras zonas, aunque esta diferencia puede ser muy pequeña o despreciable en algún caso. No obstante, en el rango de profundidad de 15 a 45 km, las diferentes profundidades del Moho presentes en el área de estudio generan la principal variación de velocidad de onda S. Un comportamiento similar es encontrado para el rango de profundidad desde 80 a 230 km, en el cual la frontera litosfera-astenosfera (LAB) genera las principales variaciones de velocidad de onda S. Finalmente, debería ser destacada la nueva e interesante característica obtenida en este estudio: la definición de la base de la astenosfera (para el área de estudio completa y para profundidades variando desde 130 a 280 km, respectivamente).</p>

Layered anisotropy beneath the Japan Sea and NE China from inversion of surface wave dispersion using rj-MCMC method

2020 ◽  
Author(s):  
Yanzhe Zhao ◽  
Zhen Guo ◽  
Xingli Fan ◽  
Yanbin Wang
Keyword(s):  
Surface Wave ◽  
Shear Wave ◽  
Wave Velocity ◽  
Shear Wave Velocity ◽  
Japan Sea ◽  
Wave Dispersion ◽  
Azimuthal Anisotropy ◽  
Ne China ◽  
Surface Wave Dispersion ◽  
The Japan Sea

&lt;p&gt;The surface wave dispersion data with azimuthal anisotropy can be used to invert for the wavespeed azimuthal anisotropy, which provides essential dynamic information about depth-varying deformation of the Earth&amp;#8217;s interior. In this study, we adopt an rj-MCMC (reversible jump Markov Chain Monte Carlo) technique to invert for crustal and upper mantle shear velocity and azimuthal anisotropy beneath the Japan Sea using Rayleigh wave azimuthally anisotropic phase velocity measurements from Fan et al. (2019). The rj-MCMC implements trans-dimensional sampling in the whole model space and derives the distribution for each model parameter (shear wave velocity and anisotropy parameters) directly from data. Without the prejudiced parameterization for model, the result can be more credible, from which some more reliable estimates for shear wave velocity and azimuthal anisotropy as well as their uncertainties can be acquired. Our preliminary results, together with shear wave splitting observations, show a layered anisotropy beneath the Japan Sea and NE China, suggesting the complicated mantle flow that is controlled by the subduction of the Pacific plate and the large-scale upwelling beneath the Changbaishan volcano.&lt;/p&gt;

Sign in / Sign up

Export Citation Format

Share Document