Seismic imaging in thrust‐belts with rugged topography — A 3D modeling approach

Abstract

Summary Seismic image quality in foothills and thrust-belts with rugged topography is often very poor with conventional low-fold sparse 3D orthogonal surveys. The most often cited reason is the high-level of shot-generated noise backscattered from near-surface heterogeneities and irregular topography. Consequently, many depth imaging projects in thrusted terrains have been disappointing. The principal motivation for this study is to assess in a systematic and quantitative way the challenges raised by the conventional acquisition-to-imaging workflow in foothills exploration context. Specifically, here we investigate the impact of the sparse seismic acquisition geometry in mountainous terrains with difficult access and rough topography. Using 3-D Acoustic Finite-Difference (AFD) modeling we analyze both homogeneous and heterogeneous near-surface velocity models with the same background reflectivity. We then evaluate the impact of sparse 3D orthogonal surveys on image quality when using the exact velocity model and accurate imaging algorithms. We show that in absence of near-surface heterogeneities but rugged topography, the sparse 3D orthogonal geometries may image correctly only the deeply buried targets. In the real world, however, we often have a very heterogeneous near-mid subsurface with short-wavelength (relative to receiver spread-length) velocity perturbations and thickness variation. This generates significant amount of reverberated body-wave refracted energy, which is subsequently scattered by the rugged topography and occurs as fuzzy “noise” throughout the entire shot record. With low-fold sparse 3D acquisitions, this scattered primary energy is spatially aliased in all domains, acting as very energetic signal-generated “noise” difficult to handle with conventional pre-processing. As a result, even the most advanced imaging algorithms such as Reverse Time Migration (RTM) and Beam, fail to yield interpretable images of the subsurface even when the nearsurface/background velocity model is known.