Open AccessNumerical simulation of seismic wave propagation in realistic 3‐D geo‐models with a Fourier pseudo‐spectral method
Author(s) -
Klin Peter,
Priolo Enrico,
Seriani Geza
Publication year - 2010
Publication title -
geophysical journal international
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.302
H-Index - 168
eISSN - 1365-246X
pISSN - 0956-540X
DOI - 10.1111/j.1365-246x.2010.04763.x
Subject(s) - discretization , attenuation , fourier transform , grid , convolution (computer science) , aliasing , seismic wave , wave propagation , geology , computer science , algorithm , geometry , mathematical analysis , geophysics , mathematics , optics , physics , filter (signal processing) , machine learning , artificial neural network , computer vision
SUMMARY We present a procedure for the numerical computation of the seismic waves in realistic 3‐D geo‐models. The procedure is based on a parallel implementation of the Fourier pseudo‐spectral method (FPSM) on staggered grids. FPSM solves spatial derivatives with optimal accuracy in domains which are discretized in easy‐to‐handle, structured meshes. The application of the presented procedure is therefore advantageous, in terms of both accuracy and simplicity of use, whenever the geo‐model can be adequately sampled by a structured grid. To be applicable to a wide range of geophysical problems, the procedure implements the most important features needed for modelling the seismic waves propagation, that is intrinsic attenuation, seismic velocity anisotropy and irregular topography. The intrinsic attenuation is implemented by means of the generalized zener body GZB mechanical model. Orthorhombic symmetry is considered for the description of the seismic anisotropy. Irregular topography is approximated by a staircase of cubic cells and treated using an original approach based on discrete Fourier transforms of arbitrary length. Sharp impedance contrasts in the geo‐model are handled by suitable averaging of the material properties to reduce phase misalignments and staircase effects due to the structured grid. Absorbing boundaries based on the convolutional perfectly matching layer technique are adopted. Finally, the efficiency of the proposed procedure is illustrated by a number of tests and examples.