IMAGING OF OFFSET VSP DATA WITH AN ELASTIC ITERATIVE MIGRATION SCHEME

VSP data are usually acquired in order to obtain high-resolution image s of complex structures in reservoirs and near boreholes. We present a n elastic iterative migration scheme which has few limitations regardi ng the complexity of the geology, and where the macromodel for both P- and S-wave velocities is automatically improved and updated at each i teration. We avoid wavefield separation (up/down and P/S) and the simp lifying assumptions of small dips underlying most such methods. The mi gration scheme is based on elastic inversion theory. The wavefield ext rapolation is based on a high-order, coarse-grid, finite-difference so lution to the elastic two-way wave equation. At each iteration, the ma cromodel is updated using a gradient method, in which the gradient is computed by correlation of forward-modelled fields with back-propagate d residual fields. The first iteration of the migration scheme is equi valent to elastic reverse-time migration with an imaging condition sim ilar to Claerbout's principle. Both P- and S-wave reflections contribu te to the images. In numerical examples with both synthetic and real o ffset VSP data, we find that increasing the number of iterations impro ves the image quality. Images based on both P- and S-wave energy give more near-well information and higher spatial resolution than images o n only acoustic energy. In the real data example we show that the iter ative migration scheme can image a relatively complex geological struc ture. A fault and a small graben intersecting the well can be identifi ed.