Introduction to this special section: Imaging migration

M imaging methods are slowly but surely trending toward the solution of an inverse problem, not only for the velocity-estimation phase but also for the migration step. The general direction is toward the conventional imaging twostep process of velocity estimation followed by migration being superseded by the single full-waveform inversion (FWI) process. However, today, advanced imaging workflows are hybrid, and FWI is complemented by more conventional imaging methods. The gradual transformation of imaging into inversion is made possible not only by significant increases in computing power, facilitating methods that iterate with migrationlike operators, but also by theoretical advances that better exploit information in the acquired data. Moreover, modern acquisition provides data with increasingly long offsets and wide-azimuth coverage, higher signal-to-noise ratio (S/N), and broader bandwidth. Acquisition increasingly makes use of multicomponent sensors that allow sampling of elastic wavefields, which contain additional information that can be used to better characterize the physical properties of the subsurface. This special section aims to explore some of the above-mentioned trends and to highlight methods that build on existing imaging technology while reaching beyond the current state of the art. Three articles describe the application of least-squares migration, which is an example of linearized waveform inversion. Huang et al. and Zeng et al. demonstrate the improvement in bandwidth, resolution, and S/N achieved by posing the migration problem as an inversion. As computational efficiency keeps improving, the use of inversion in the migration phase of the imaging process is likely to increase, particularly for data sets that suffer from severely irregular data sampling and from poor illumination caused by complex overburden. Wong et al. demonstrate that linearized inversion can go beyond the imaging of primary events, which is the main goal of conventional imaging, to image multiple reflections generated by known multiple generators, such as the seafloor and sea surface. Primary and multiple reflections can be inverted simultaneously to form an image that is superior to those obtained by separately inverting primaries and multiples. Multiples provide additional coverage of the subsurface, thus increasing its illumination and better constraining the geologic structure in complex geologic environments. Vigh et al. and Ratcliffe et al. discuss other ways in which waveform inversion and migration are increasingly connected and complementary. Both articles show how FWI can be used as a preprocessing step before high-frequency reverse time migration (RTM). Because of computational cost, current applications of FWI do not yet exploit the full bandwidth of acquired seismic data. Thus, images produced by FWI are between long-wavelength models produced with conventional velocity-analysis techniques and high-resolution images obtained by migration. Nevertheless, the improved resolution of velocity models estimated by FWI has a significant positive impact on the final RTM results. BIONDO BIONDI and DAVE NICHOLS, Stanford, California PAUL SAVA, Golden, Colorado