Adaptive 3D Joint Inversion of Direct, Reflected and Refracted Arrivals

Abstract

The paper shows how it is possible to use the available data from conventional seismic surveys, usually finalised to oil exploration or crustal studies, for shallow water environment studies. In particular, we show here that the joint inversion of direct, reflected and head waves provides a better reconstruction of the sea water, the shallowest Earth layers and their inhomogeneities in depth. It doubles in fact the available ray paths and traveltimes, so increasing the statistical and mathematical robustness of the obtained estimates. Furthermore, it allows primary and multiple reflections to be better discriminated, and expands the area illuminated by the elastic waves. The reconstruction of complex interfaces and of the velocity field in 3D can be done automatically, by minimising the spreading of the reference points estimated by reflected and refracted arrivals. Moreover, adaptive irregular grids provide a reliable estimate of the velocity field, by fitting the local resolution to the available ray paths and to the detected velocity anomalies. We defined an automatic procedure based on the Voronoi polygons, able to adapt the grid shape to the boundaries of the velocity anomalies. These algorithms work well in 3D for detecting velocity anomalies that are both very small (e.g. due to sea currents) and large (e.g. due to gas lenses below the seafloor). An application to a seismic profile from the Antarctica’s offshore shows that the tomographic velocity field enhances the resolution of the pre-stack depth imaging. There are several applications of the proposed techniques, beyond seismic exploration, as shallow water acoustics and offshore engineering, e.g. in the site studies for platforms and pipelines. A promising application of the proposed technique is however the study of the marine environment. The sound velocity at sea is directly related to the water temperature and salinity. Since the volume of the wave propagation in a seismic survey is relevant, i.e.. of the order of several kilometres, the estimated velocities are statistically robust, and so it is in particular its average regional value. This data can change during the different seasons and, of course, during the years. It is more and more common to have available time-lapse 3D surveys (the so-called 4D ones), acquired by the oil companies to monitor the hydrocarbon reservoir depletion. In these and similar cases, we can exploit the tomographic data to detect large scale, long term climatic changes.