Prestack waveform inversion of three-component, two-azimuth surface seismic data for azimuthally dependent anisotropic parameters using a parallelized nondominated sorting genetic algorithm

Abstract

Summary Consideration of azimuthal anisotropy, at least to an orthorhombic symmetry is important in exploring the naturally fractured and unconventional hydrocarbon reservoirs. Full waveform inversion of multicomponent seismic data can, in principle, provide more robust estimates of subsurface elastic parameters and density than inversion of single component seismic data; also azimuthally dependent anisotropy can only be resolved by carefully studying of multicomponent data acquired and processed along different azimuths. Multicomponent seismic inversions however, require the optimization of multiple objectives, one for each data component. Such multiobjective optimization problems are in general nonlinear with nonunique solutions, known as the Paretooptimal solutions. Therefore it is appropriate to treat the objectives as a vector and simultaneously optimize each of its components so that the entire Pareto-optimal set of solutions could be estimated. The fast nondominated sorting genetic algorithm (NSGA II) is a robust stochastic global optimization algorithm, capable of handling such multiobjective problems, however, its computational intensity increases exponentially with increasing number of objectives and the model parameters to be estimated. In addition, an accurate extraction of subsurface azimuthal anisotropy requires multicomponent seismic data acquired at a fine spatial resolution along many source-to-receiver azimuthal orientations. Because routine acquisition of such data is prohibitively expensive, they are typically available along two or at most three azimuthal orientations at a spatial resolution where such an inversion could be applied. Here, we propose a novel two-step multi-objective inversion methodology using a parallelized version of NSGA II for waveform inversion of three components (P-P, P-SV, P-SH) seismic data along two different azimuths. In the first step, we use the near-offset prestack data along a single azimuth and invert them using an isotropic assumption to obtain an estimate of the vertical P- and Swave velocities. In step 2, we then estimate the azimuthally dependent anisotropic properties and density by inverting the entire long-offset, two-azimuth data. We demonstrate the applicability of the method using synthetic data generated from a multilayer model based on a real well log. We document that the proposed method can reliably extract subsurface elastic parameters and density from multicomponent seismic data.