2. Seismic/Geodetic Imaging Technologies
Velocity Macro Model Estimation From Nodes Data
Mathias Alerini
SINTEF Petroleum Research, Norway.
Abstract
Marine seismic nodes acquisition has been recently developed to improve the coupling of the receivers with sea floor, the vector fidelity of the seismic data and the positioning allowing an accurate repeatability. However, the distance between receivers is usually huge (typically 400m) and leads to problems for velocity model building for depth imaging. This is emphasized for slope tomography methods such a stereotomography. Indeed, such methods estimate the macro model from locally coherent events described by their two surface positions, their two-way traveltimes and their both slopes at source and receiver (i.e. the tangent to the local event in the common receiver gather and in the common source gather). A precise enough estimation of those tangents requires a dense enough acquisition on both sides, sources and receivers, condition which is not fulfilled with nodes data. Stereotomography has shown numerous advantages compare to other velocity model building methods such as an easier picking, speed and robustness. Using this approach with reflection seismic data recorded by nodes should lead to important results. I propose here a strategy to bypass the sparseness of the acquisition of nodes data so that it becomes possible to use stereotomography. This strategy uses the reciprocity of Green's functions after a redatuming of sources at receivers depth. The question which arises is whether such a sparse information is enough to estimate the velocity model. On a synthetic example, I can obtain relatively good results. An interesting conclusion appears: The stereotomography aims at flattening common image gathers in a similar way as differential semblance optimization, a migration based velocity estimation method. However, from nodes data, it is not possible to compute common image gathers as aliasing leads to strong artifacts. Therefore stereotomography flattens "non-existing" common image gathers and can be applied in situations where differential semblance optimization could not.
Last modified: Mon May 22 22:18:55 2006