Optimal staggered-grid finite-difference schemes by combining Taylor-series expansion and sampling approximation for wave equation modeling

Yan, Hongyong; Yang, Lei; Li, Xiang-Yang. 2016 Optimal staggered-grid finite-difference schemes by combining Taylor-series expansion and sampling approximation for wave equation modeling. Journal of Computational Physics, 326. 913-930. https://doi.org/10.1016/j.jcp.2016.09.019

Full text not available from this repository. (Request a copy)

Official URL: http://dx.doi.org/10.1016/j.jcp.2016.09.019

Abstract/Summary

High-order staggered-grid finite-difference (SFD) schemes have been universally used to improve the accuracy of wave equation modeling. However, the high-order SFD coefficients on spatial derivatives are usually determined by the Taylor-series expansion (TE) method, which just leads to great accuracy at small wavenumbers for wave equation modeling. Some conventional optimization methods can achieve high accuracy at large wavenumbers, but they hardly guarantee the small numerical dispersion error at small wavenumbers. In this paper, we develop new optimal explicit SFD (ESFD) and implicit SFD (ISFD) schemes for wave equation modeling. We first derive the optimal ESFD and ISFD coefficients for the first-order spatial derivatives by applying the combination of the TE and the sampling approximation to the dispersion relation, and then analyze their numerical accuracy. Finally, we perform elastic wave modeling with the ESFD and ISFD schemes based on the TE method and the optimal method, respectively. When the appropriate number and interval for the sampling points are chosen, these optimal schemes have extremely high accuracy at small wavenumbers, and can also guarantee small numerical dispersion error at large wavenumbers. Numerical accuracy analyses and modeling results demonstrate the optimal ESFD and ISFD schemes can efficiently suppress the numerical dispersion and significantly improve the modeling accuracy compared to the TE-based ESFD and ISFD schemes

Item Type:	Publication - Article
Digital Object Identifier (DOI):	https://doi.org/10.1016/j.jcp.2016.09.019
ISSN:	00219991
Date made live:	10 Jan 2017 09:58 +0 (UTC)
URI:	https://nora.nerc.ac.uk/id/eprint/515731

Item Type:

Publication - Article

Digital Object Identifier (DOI):

https://doi.org/10.1016/j.jcp.2016.09.019

ISSN:

00219991

Date made live:

10 Jan 2017 09:58 +0 (UTC)

URI:

https://nora.nerc.ac.uk/id/eprint/515731