In the conventional finite-frequency traveltime tomography, the traveltime Fréchet derivative is usually derived using the Born- or the Rytov approximation, which requires that the magnitude of velocity perturbations should be small enough to satisfy the weak-scattering assumption. For large-scale and strong velocity perturbations, the generalized Rytov approximation (GRA) has been proposed, which can predict the accumulated phase-change of the forward scattered wave, making it more suited for traveltime inverse problems. To mitigate the source uncertainties, we introduce the doubledifference (DD) approach to calculate the traveltime shift. By combining GRA with the DD approach, we propose a new finite-frequency traveltime sensitivity kernel, i.e. the GRA-based double-difference traveltime sensitivity kernel, which inherits the advantage of both the DD measurement and the GRA method. To solve the associated traveltime inverse problem, we use a matrix-free Gauss-Newton algorithm to accelerate the convergence rate, where the Hessian-vector product is computed approximately using the ray approximation, making it more computationally efficient. Therefore, the proposed method is quite promising for 3-D problems.
Presentation Date: Monday, October 12, 2020
Session Start Time: 1:50 PM
Presentation Time: 2:15 PM
Location: Poster Station 8
Presentation Type: Poster
