Sensitivity kernels for wave-equation migration velocity analysis

摘要: The success of migration velocity analysis methods is strongly dependent on the characteristics linearized tomographic operator that inverted to estimate updates. To study properties wave-equation analysis, we analyze its sensitivity kernels. Sensitivity kernels describe dependence data space elements small changes model elements. We show MVA depend frequency content recorded and background model. computed assuming presence a salt body in these are drastically different from idealized ifat raysi. Consequently cannot be approximated by articial fattening geometrical rays. Furthermore, our examples illustrate potential nite-frequenc y as well frequency-dependent nature illumination for subsalt regions.