An efficient approach of the pseudospectral method for modelling of geometrically symmetric seismic wavefield

Hiroshi Takenaka, Yanbin Wang, Takashi Furumura

Research output: Contribution to journalArticle

9 Citations (Scopus)

Abstract

The pseudospectral method is a high-accuracy numerical modelling technique that requires less computer memory and computation time than the traditional techniques such as the finite-difference method. These advantages of the pseudospectral method have enabled us to practically apply this method to modelling realistic problems that have complex structure and source models. However, a major drawback of such numerical schemes for discrete grid models is that even for rather a simple structural model they require as much computational requirements (e.g. computation time and memory) as for an entirely complex structural model with the same size of the simple one. We actually need to employ idealised simple models, such as a model with geometrical symmetry, to investigate basic phenomena of seismic waves, to develop new techniques, or to choose optimal values of some computational parameters for more complex modelling. In this paper we propose an efficient approach of an economical pseudospectral method for calculation of wavefields in models symmetric with respect to a vertical plane or two orthogonal vertical planes. Using this approach, the wavefields only need to be computed in a half or quarter domain of the models, so that the computer memory and computation time can be reduced ideally by half or quarter, respectively, as compared with the calculation of the entire models.

Original languageEnglish
Pages (from-to)73-79
Number of pages7
JournalEarth, Planets and Space
Volume51
Issue number2
DOIs
Publication statusPublished - Jan 1 1999
Externally publishedYes

ASJC Scopus subject areas

  • Geology
  • Space and Planetary Science

Fingerprint Dive into the research topics of 'An efficient approach of the pseudospectral method for modelling of geometrically symmetric seismic wavefield'. Together they form a unique fingerprint.

  • Cite this