On anisotropy and inhomogeneity parameter estimation using traveltimes

TL;DR

This paper develops a method to estimate anisotropy and inhomogeneity parameters from seismic traveltimes in layered models, validating it with synthetic and real data, and analyzing the impact of noise on the estimates.

Contribution

It introduces a least-squares inversion approach for anisotropic inhomogeneous models using traveltime data, including noise robustness analysis.

Findings

Method successfully estimates parameters in synthetic models.

Noise severely limits parameter estimation accuracy.

Real-data application demonstrates practical utility.

Abstract

We consider an anisotropic inhomogeneous model to simulate measured vertical-seismic-profile traveltimes. In this model, we assume that velocity increases linearly with depth and anisotropy is the result of elliptical velocity dependence. Using a series of sources in a line to a single receiver, we minimize the least-squares residual between measured and modelled traveltimes to estimate the anisotropy and inhomogeneity parameters of the subsurface. To verify the approach, we construct synthetic traveltime data for a one- and two-layer model and perform the traveltime inversion. We justify the convergence-ensuring specifications and model-parameter restrictions that are intrinsic to the parameter estimation. To determine the approach's practicality, we assess the reliability of results under the influence of noise. From this assessment, we discern the noise threshold for both models and determine that the noise restriction is severely restrictive for increasing model complexity. We estimate parameters, using the same methodology, for a real-data case and conclude with a discussion of results.