The effect of the near surface on seismic land data can be so severe that static corrections are insufficient. Full-waveform inversion followed by redatuming may be an alternative, but inversion will work only if the starting model is sufficiently close to the true model. As a first step toward determining a viscoelastic near-surface model, we assume that existing methods can provide a horizontally layered velocity and density model. Because near-surface attenuation is strongest, we propose a method to estimate the P-wave attenuation based on viscoacoustic finite-difference modeling. We compare energy decay along traveltime curves of reflection and refraction events in the modeled and observed seismic data for a range of attenuation parameters. The best match provides an estimate of the attenuation. First, we estimate only the attenuation of the top layer and study the sensitivity to depth and velocity perturbations. Then, we consider multiple layers. We apply the method to synthetic and real data and investigate the effect of source wavelet and topography. The method is robust against depth and velocity perturbations smaller than 10%. The results are sensitive to the source wavelet. Incorporating the surface topography in the computed traveltimes reduces the uncertainty of the attenuation estimates, especially for deeper layers.
Acoustic Gaussian beam migration is an attractive imaging method because it is flexible with input geometry, efficient, and accurate in imaging multipath arrivals. However, one of the hurdles that this method must overcome in production processing is its extension to use multimeasurement data, as recently allowed by novel acquisition technologies. This is inevitable when the compensation of the ghost effect is best corrected within a true-amplitude imaging process, a necessity for amplitude-variation-with-offset work. For this purpose, I introduced a novel formalism for vector-acoustic imaging, based on Green’s function theory, which can remove the ghost effect and produce amplitudes on reflectors that are proportional to the reflection coefficients. I established a theoretical framework with Gaussian beam representations of Green’s functions, including the weighted beam-stacking approach that reduced the cost of computation. I extended my formulas to use the steep-descent (i.e., stationary phase) approximation. Then, I explained the impact of this approximation on the illumination and the event continuity and sharpness. I also studied the special case of acoustic imaging corresponding to using single-measurement (i.e., pressure) data. I applied the derived formulations to realistic synthetic multisensor data (North Sea) using a research code of Gaussian beam migration. The numerical examples demonstrated that I can improve the illumination of the final images and obtain wide-bandwidth reflectivity maps.
Most common deghosting techniques operating on multimeasurement marine seismic data rely on a ghost model to combine pressure and vertical velocity. The ghost model provides information needed for the optimization of the signal-to-noise ratio in the broadband results. These techniques are, in general, sensitive to the accuracy of such model and can suffer from perturbations, especially at high frequencies: for instance, the coarse sampling in the crossline direction often forces these techniques to rely on a 2D ghost model, and this can induce significant inaccuracies in complex scenarios. Other modelindependent techniques combine the pressure and vertical velocity without the use of a ghost model. Although these suffer less from 3D propagation effects, they often produce suboptimal results for what concerns the S/N. We introduce a novel model-independent deghosting technique called the "Bayesian PZSUM." This technique utilizes the multimeasurement second-order statistics to estimate the upgoing wavefield optimally in the minimum mean square error (m.m.s.e) sense. We demonstrate the effectiveness of the technique on both synthetic and real data.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
customersupport@researchsolutions.com
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.