skip to main content

Title: Optimal Transport Based Seismic Inversion:Beyond Cycle Skipping
Full-waveform inversion (FWI) is today a standard process for the inverse problem of seismic imaging. PDE-constrained optimization is used to determine unknown parameters in a wave equation that represent geophysical properties. The objective function measures the misfit between the observed data and the calculated synthetic data, and it has traditionally been the least-squares norm. In a sequence of papers, we introduced the Wasserstein metric from optimal transport as an alternative misfit function for mitigating the so-called cycle skipping, which is the trapping of the optimization process in local minima. In this paper, we first give a sharper theorem regarding the convexity of the Wasserstein metric as the objective function. We then focus on two new issues. One is the necessary normalization of turning seismic signals into probability measures such that the theory of optimal transport applies. The other, which is beyond cycle skipping, is the inversion for parameters below reflecting interfaces. For the first, we propose a class of normalizations and prove several favorable properties for this class. For the latter, we demonstrate that FWI using optimal transport can recover geophysical properties from domains where no seismic waves travel through. We finally illustrate these properties by the realistic application of more » imaging salt inclusions, which has been a significant challenge in exploration geophysics. « less
Varadhan, S.R.S.
Award ID(s):
1913129 1913209
Publication Date:
Journal Name:
Communications on Pure and Applied Mathematics
Sponsoring Org:
National Science Foundation
More Like this
  1. Seismic full-waveform inversion aims to reconstruct subsurface medium parameters from recorded seismic data. It is solved as a constrained optimization problem in the deterministic approach. Many different objective functions have been proposed to tackle the nonconvexity that originated from the cycle-skipping issues. The analogy between objective functions in the deterministic inversion and likelihood functions in Bayesian inversion motivates us to analyze the noise model each objective function accounts for under the Bayesian inference setting. We also show the existence and wellposedness of their corresponding posterior measures. In particular, the theorem shows that theWasserstein-type likelihood offers better stability with respect tomore »the noise in the recorded data. Together with an application of the level-set prior, we demonstrate by numerical examples the successful reconstruction from Bayesian full-waveform inversion under the proper choices of the likelihood function and the prior distribution.« less
  2. Summary The contiguous United States has been well instrumented with broadband seismic stations due to the development of the EarthScope Transportable Array. Previous studies have provided various 3D seismic wave speed models for the crust and upper mantle with improved resolution. However, discrepancies exist among these models due to differences in both data sets and tomographic methods, which introduce uncertainties on the imaged lithospheic structure beneath North America. A further model refinement using the best data coverage and advanced tomographic methods such as full-waveform inversion (FWI) is expected to provide better seismological constraints. Initial models have significant impacts on themore »convergence of FWIs. However, how to select an optimal initial model is not well investigated. Here, we present a data-driven initial model selection procedure for the contiguous US and surrounding regions by assessing waveform fitting and misfit functions between the observations and synthetics from candidate models. We use a data set of waveforms from 30 earthquakes recorded by 5,820 stations across North America. The results suggest that the tested 3D models capture well long-period waveforms while showing discrepancies in short-periods especially on tangential components. This observation indicates that the smaller-scale heterogeneities and radial anisotropy in the crust and upper mantle are not well constrained. Based on our test results, a hybrid initial model combining S40RTS or S362ANI in the mantle and US.2016 for Vsv and CRUST1.0 for Vsh in the crust is compatible for future FWIs to refine the lithospheric structure of North America.« less
  3. State-of-the-art seismic imaging techniques treat inversion tasks such as full-waveform inversion (FWI) and least-squares reverse time migration (LSRTM) as partial differential equation-constrained optimization problems. Due to the large-scale nature, gradient-based optimization algorithms are preferred in practice to update the model iteratively. Higher-order methods converge in fewer iterations but often require higher computational costs, more line-search steps, and bigger memory storage. A balance among these aspects has to be considered. We have conducted an evaluation using Anderson acceleration (AA), a popular strategy to speed up the convergence of fixed-point iterations, to accelerate the steepest-descent algorithm, which we innovatively treat as amore »fixed-point iteration. Independent of the unknown parameter dimensionality, the computational cost of implementing the method can be reduced to an extremely low dimensional least-squares problem. The cost can be further reduced by a low-rank update. We determine the theoretical connections and the differences between AA and other well-known optimization methods such as L-BFGS and the restarted generalized minimal residual method and compare their computational cost and memory requirements. Numerical examples of FWI and LSRTM applied to the Marmousi benchmark demonstrate the acceleration effects of AA. Compared with the steepest-descent method, AA can achieve faster convergence and can provide competitive results with some quasi-Newton methods, making it an attractive optimization strategy for seismic inversion.« less
  4. We have developed a memory and operation-count efficient 2.5D inversion algorithm of electrical resistivity (ER) data that can handle fine discretization domains imposed by other geophysical (e.g, ground penetrating radar or seismic) data. Due to numerical stability criteria and available computational memory, joint inversion of different types of geophysical data can impose different grid discretization constraints on the model parameters. Our algorithm enables the ER data sensitivities to be directly joined with other geophysical data without the need of interpolating or coarsening the discretization. We have used the adjoint method directly in the discretized Maxwell’s steady state equation to computemore »the data sensitivity to the conductivity. In doing so, we make no finite-difference approximation on the Jacobian of the data and avoid the need to store large and dense matrices. Rather, we exploit matrix-vector multiplication of sparse matrices and find successful convergence using gradient descent for our inversion routine without having to resort to the Hessian of the objective function. By assuming a 2.5D subsurface, we are able to linearly reduce memory requirements when compared to a 3D gradient descent inversion, and by a power of two when compared to storing a 2D Hessian. Moreover, our method linearly outperforms operation counts when compared with 3D Gauss-Newton conjugate-gradient schemes, which scales cubically in our favor with respect to the thickness of the 3D domain. We physically appraise the domain of the recovered conductivity using a cutoff of the electric current density present in our survey. We evaluate two case studies to assess the validity of our algorithm. First, on a 2.5D synthetic example, and then on field data acquired in a controlled alluvial aquifer, where we were able to match the recovered conductivity to borehole observations.« less
  5. Full waveform inversion (FWI) and least-squares reverse time migration (LSRTM) are popular imaging techniques that can be solved as PDE-constrained optimization problems. Due to the large-scale nature, gradient- and Hessian-based optimization algorithms are preferred in practice to find the optimizer iteratively. However, a balance between the evaluation cost and the rate of convergence needs to be considered. We propose the use of Anderson acceleration (AA), a popular strategy to speed up the convergence of fixed-point iterations, to accelerate a gradient descent method. We show that AA can achieve fast convergence that provides competitive results with some quasi-Newton methods. Independent ofmore »the dimensionality of the unknown parameters, the computational cost of implementing the method can be reduced to an extremely lowdimensional least-squares problem, which makes AA an attractive method for seismic inversion.« less