Abstract

SUMMARY The accuracy and applicability of the Gaussian beam method for the computation of synthetic seismograms in absorbing 2-D laterally inhomogeneous media is discussed in this paper. A computer program was developed for the computation of P-SV- and SH-waves along horizontal and vertical seismic profiles in absorbing 2-D inhomogeneous media with first-order discontinuities in density and velocities. Subdividing the model into triangles with linear density and velocity laws not only allows flexible modelling of complicated structures but also results in fast kinematic and dynamic ray tracing. With this program it is possible to compute phases which are specified by single or multiple reflections, refractions and/or conversions at the different discontinuities in the medium. The width and phase-front curvature of each beam is controlled by its complex beam parameter e. Along with several e-options proposed by other authors, a newly developed option to select the beam parameter was tested for a large set of models. The new option correlates the width of each beam to the size of the triangles the beam has passed through and thus to the structure of the medium. Slowly varying media represented by large triangles will give broad Gaussian beams whereas complicated and rapidly changing media which have to be described by many small triangles will produce narrow Gaussian beams. For the computation of reference seismograms the reflectivity method, the frequency-domain finite-difference method and the finite-difference method were used. The conclusion from these tests is, that the newly developed e-option is the optimum choice, whereas the other options sometimes fail and do not give accurate seismograms for critical incidence and at caustics.

Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call