Misfit function for full waveform inversion based on Earth Mover's Distance with dynamic formulation

TL;DR

This paper introduces a novel Earth Mover's Distance-based misfit function for full waveform inversion, improving robustness against local minima and time shifts, with an efficient GPU-accelerated optimization method.

Contribution

It develops a dynamic formulation of EMD for FWI, addressing computational challenges and demonstrating effectiveness on synthetic and real seismic data.

Findings

EMD-based FWI reduces cycle-skipping issues.

The proposed method increases computation time by only 11%.

Application to SEG 2014 data shows reliable velocity estimation.

Abstract

Conventional full waveform inversion (FWI) using least square distance (LSD) between the observed and predicted seismograms suffers from local minima. Recently, earth mover's distance (EMD) has been introduced to FWI to compute the misfit between two seismograms. Instead of comparisons bin by bin, EMD allows to compare signal intensities across different coordinates. This measure has great potential to account for time and space shifts of events within seismograms. However, there are two main challenges in application of EMD to FWI. The first one is that the compared signals need to satisfy nonnegativity and mass conservation assumptions. The second one is that the computation of EMD between two seismograms is a computationally expensive problem. In this paper, a strategy is used to satisfy the two assumptions via decomposition and recombination of original seismic data. In addition, the computation of EMD based on dynamic formulation is formulated as a convex optimization problem. A primal-dual hybrid gradient method with linesearch has been developed to solve this large-scale optimization problem on GPU device. The advantages of the new method are that it is easy to implement and has high computational efficiency. Compared to LSD based FWI, the computation time of the proposed method will approximately increase by 11% in our case studies. A 1D time-shift signals case study has indicated that EMD is more effective in capturing time shift and makes the misfit function more convex. Two applications to synthetic data using transmissive and reflective recording geometries have demonstrated the effectiveness of EMD in mitigating cycle-skipping issues. We have also applied the proposed method to SEG 2014 benchmark data, which has further demonstrated that EMD can mitigate local minima and provide reliable velocity estimations without using low frequency information in the recorded data.