Multiparameter inverse scattering: computational approaches

Glen R. Young, Kristopher A. Innanen, Laurence R. Lines

The use of inverse scattering methods in the inversion of seismic data has been on the rise in exploration geophysics. With specific computational approaches it is possible to ascertain the material properties of the subsurface using scattered acoustic waves. We seek to determine multiple rock parameters such as density and bulk modulus from reflected seismic signals. At this early stage of investigation a basic approach is used based on straightforward inverse scattering equations. In this case we will examine how multiparameter inverse scattering in a constant 2D background works and what are the results of inverting synthetically generated data. A simulation was developed for this project in two parts. The forward modeling stage and the inversion. The forward modeling code is based upon CREWES MATLAB finite difference routines, this stage takes simple user generated velocity models and shot patterns and creates a set of synthetic shot profiles which are convolved with a Ricker source wavelet. The second stage takes these shot profiles and attempts a deconvolution and inversion to generate an inversion image. The inversion image is then examined for accuracy and various models are used to determine of physically realistic effects are present as would be seen in real seismic data.