3D Variational Inference-Based Double-Difference Seismic Tomography Method and Application to the SAFOD Site, California

Hao Yang,Xin Zhang,Haijiang Zhang

2024-07-31

Abstract:Seismic tomography is a crucial technique used to image subsurface structures at various scales, accomplished by solving a nonlinear and nonunique inverse problem. It is therefore important to quantify velocity model uncertainties for accurate earthquake locations and geological interpretations. Monte Carlo sampling techniques are usually used for this purpose, but those methods are computationally intensive, especially for large datasets or high-dimensional parameter spaces. In comparison, Bayesian variational inference provides a more efficient alternative by delivering probabilistic solutions through optimization. The method has been proven to be efficient in 2D tomographic problems. In this study, we apply variational inference to solve 3D double-difference (DD) seismic tomographic system using both absolute and differential travel time data. Synthetic tests demonstrate that the new method can produce more accurate velocity models than the original DD tomography method by avoiding regularization constraints, and at the same time provides more reliable uncertainty estimates. Compared to traditional checkerboard resolution tests, the resulting uncertainty estimates measure more accurately the reliability of the solution. We further apply the new method to data recorded by a local dense seismic array around the San Andreas Fault Observatory at Depth (SAFOD) site along the San Andreas Fault (SAF) at Parkfield. Similar to other researches, the obtained velocity models show significant velocity contrasts across the fault. More importantly, the new method produces velocity uncertainties of less than 0.34 km/s for Vp and 0.23 km/s for Vs. We therefore conclude that variational inference provides a powerful and efficient tool for solving 3D seismic tomographic problems and quantifying model uncertainties.

Geophysics

What problem does this paper attempt to address?

### Problems Addressed by the Paper

The paper primarily addresses several key issues in three-dimensional seismic tomography:

1. **Nonlinearity and Non-uniqueness**: Seismic tomography is inherently a nonlinear and non-unique inverse problem, meaning there are multiple possible solutions. Therefore, quantifying the uncertainty of the velocity model is crucial for accurately determining earthquake locations and geological interpretations.

2. **Computational Efficiency**: Traditional Monte Carlo sampling methods can quantify uncertainty but are computationally expensive for large datasets or high-dimensional parameter spaces. In contrast, Bayesian variational inference offers a more efficient probabilistic solution through an optimization process.

3. **Regularization Constraints**: Traditional linearized methods solve for the optimal solution by minimizing the difference between observed data and model-predicted data. However, since inverse problems are typically underdetermined, regularization constraints are often needed to reduce uncertainty. Yet, regularization can lower model resolution and introduce subjective constraints.

### Main Contributions

1. **Application of Variational Inference to 3D Double-Difference Tomography**: The paper proposes a variational inference-based method to solve the 3D double-difference (DD) seismic tomography problem. This method uses both absolute and differential travel time data, avoiding regularization constraints and providing more reliable uncertainty estimates.

2. **Synthetic Test Validation**: Synthetic tests demonstrate that the new method generates more accurate velocity models than traditional DD tomography methods while providing more reliable uncertainty estimates.

3. **Real-world Application Validation**: The new method is applied to data recorded by a local dense seismic array near the San Andreas Fault in California. Results show that the velocity uncertainty generated by the new method is less than 0.34 km/s (for Vp) and 0.23 km/s (for Vs). This demonstrates the powerful and efficient capability of variational inference in solving 3D seismic tomography problems and quantifying model uncertainty.