3D structure from 2D seismic reflection surveys

Marine and land seismic re ection data collected for academic purposes have been largely acquired by 2D pro ling. Because of streamer feathering during marine studies and land acquisition on local roads that typically wind through the countryside, seismic pro ling is characterized by midpoint scattering and the planned 2D surveys result in swath 3D data sets. Although this midpoint scattering is known to be a serious hindrance to good imaging, causing several standard 2D imaging procedures to severely underperform, it is generally ignored during data processing. I will show that the midpoint scattering problem in 2D studies can commonly be turned to an advantage by processing that takes into account the 3D character of the data and of the geology. I developed and tested on synthetic and eld data both a prestack and poststack approach to extracting 3D structural information from swath 3D data. In the poststack case, better signal alignment before stacking can be achieved by revising the traveltime equation and including the crossdip terms in the moveout calculations. In the prestack case, I investigated the potential of 3D prestack Kirchho migration to directly construct a 3D image volume of all re ectors viewed by the survey. For both developed approaches, the nal products are signi cantly more accurate and more informative images of the subsurface structures than that obtained by the conventional 2D approach, with important implications for subsequent interpretation.