DRUZHININ, ALEXANDER, and COLIN MACBETH, British Geological Survey, West Mains Road, Edinburgh EH9 3LA, UK

Abstract: High-Resolution Prestack Sub-Basaltic Imaging

A problem for exploration on the Atlantic Margin is the wide spread occurence of Early Tertiary basalt lavas. Standard CDP processing does not produce a satisfactory image. Moveout velocities show very little because of strong multiples. Scattering phenomena and P-wave energy localization in the basalt structure give rise to a very low signal to noise ratio.We have attempted to develop a special prestack depth migration processing scheme for high-resolution subbasaltic imaging. Such scheme does not require any a priori information such as source signature or interval velocity model. The dataset represents a single 6 km line from a known basalt covered area with severe water bottom topography. In the first stage, wavelet transform filtering together with standard pre-processing tools such as bandpass, f-k filtering and f-x deconvolution remove most of multiples and show distinct diffractions associated with the fractures and faults in the target zone. An anti-aliased numerical version of the Berryhill's datuming formula has been applied to original offset data to move the acquisition level down to the sea floor. Despite the obvious aperture limitation, such a continuation makes it possible to suppress waterlayer multiples and preserves amplitudes. Next, we apply the prestack depth migration itself. Instead of operating directly on the observed data, we consider the above extrapolated field. This allows more flexibility in the type of migration that can be performed without the knowledge of velocity model. We begin with the modified common-shot Kirchhoff-type migration formula which combines the signals according to a new moveout correction. Such correction can be obtained using a local spherical approximation of the upgoing wave-field in accordance to the Gelchinsky's idea. Among the estimated parameters are the velocity just below the sea floor and the emergence angle of the zerooffset ray. Conventionally, such information is extracted with stack power maximization scanning the data. The depth section is designed in the depth range 0.6-7 km with the water layer depth varying from 0.6 km to 1.3 km. It is compared to standard NMO/DMO/stack processing. Results show that the method is capable to produce a clear image of the sub-basalt structure. The best image is obtained for a model with 10% effective anisotropy.The final depth image provides a consistent geological interpretation.