• 责任者：Lu Huiyi (Institute of Geoglogy and Geophysics ; Chinese Academy of Science ; Beijing 100029 ; China) ; Liu Yike
• 刊名：Chinese Journal of Geophysics
• 出版年：2013
• 卷期：56(9)
• 关键词：走时残差 ; traveltime residuals ; 层析技术 ; tomography
• 页码：p.3100-3108
• 图表：illus., 1 table, 28 refs.
• 分类号：1600
• issnNo：ISSN0001-5733
• isbnNo：CN11-2074/P

     Times New Roman">Seismic traveltime tomography method is an important technique which is used to reconstruct complex near-surface velocity models.This is an iterative inversion method,whose process requires repeated calculations of seismic ray traveltime.Therefore,one of key practical technologies of tomography is to research a traveltime calculation method which is with high efficiency and precision and can adapt to the complex model.MSFM times new roman’; mso-hansi-font-family: ’times new roman’">（Multi-stencils Fast Marching Methodtimes new roman’; mso-hansi-font-family: ’times new roman’">） proposed in the field of medical imaging will be introduced to traveltime calculations of seismic tomography.The method reduces large errors existed in the diagonal direction of standard FMM by generating new calculation stencils,which is based on coordinate rotation and make the diagonally adjacent grid points involve in the calculation.The calculation efficiency and accuracy of these methods will be analyzed and compared.The implementation scheme of MSFM for irregularly topographic complex near-surface model is studied.The local grid-refinement near source along the topography is adopted,which will largely improve the calculation precision by little calculation cost.Numerical analysis and experiments demonstrate that MSFM improves the accuracy while retaining the high efficiency of standard FMM.Test of traveltimes calculation forward and tomographic inversion for irregular model verifies the powerful applicable potency of MSFM applied in complex model.The study concludes that MSFM has very valuable application as an accurate and efficient traveltime calculation method in seismic traveltime tomography.