Full-waveform inversion of intensity-focused seismic data

Abstract

Full-waveform inversion (FWI) is a promising tool for the comprehensive analysis of seismic data because it involves modeling the complete elastodynamic phenomena in 3D for the estimation of medium parameters. However, the application for reservoir characterization is still limited in achieving the inverse problem solution using reasonable computational resources and the required model resolution. Following a data-driven approach, we have implemented the isotropic elastic FWI with sensitivity emphasis on a prescribed target zone of the propagation volume, showing that the method is useful in improving the estimation over the target zone compared with nonlocalized methods. Our method uses the superposition principle to construct multipoint sources that focus the primary wave intensity (seismic illumination) over the specific target area. Thus, the secondary wavefield scattered from the target area to the receivers increases, providing improved information on the model parameters that control the property heterogeneity within the target zone. We design two types of focused macrosource (FMS) configurations that accomplish this objective: the convergent FMS and the combined beam FMS. Although multipoint seismic sources are not implemented in the field for practical reasons, they can be synthetized computationally using the superposition principle, interpolation, and the current seismic survey data (point-source) to compose the FMS observed gathers, which are then used as input data for the FWI. We perform elastic FWI synthetic tests in a 2D modified Marmousi model to compare the inversion performance with focused and nonfocused data, under the same computational conditions, with and without the presence of noise. Our tests show faster convergence and improved estimation in the target zone with the use of the intensity-focused seismic data for FWI; improved focusing effects are expected in 3D. Being target-oriented, the method is suitable for reducing computational requirements in elastic FWI application for reservoir description.