Revealing small-scale diffracting discontinuities by an optimization inversion algorithm

Small-scale diffracting geologic discontinuities play a significant role in studying carbonate reservoirs. The seismic responses of them are coded in diffracted/ scattered waves. However, compared with reflections, the energy of these valuable diffractions is generally one or even two orders of magnitude weaker. This means that the information of diffractions is strongly masked by reflections in the seismic images. Detecting the small-scale cavities and tiny faults from the deep carbonate reservoirs, mainly over 6 km, poses an even bigger challenge to seismic diffractions, as the signals of seismic surveyed data are weak and have a low signalto-noise ratio (SNR). After analyzing the mechanism of the Kirchhoff migration method, the residual of prestack diffractions located in the neighborhood of the first Fresnel aperture is found to remain in the image space. Therefore, a strategy for extracting diffractions in the image space is proposed and a regularized L-2-norm model with a smooth constraint to the local slopes is suggested for predicting reflections. According to the focusing conditions of residual diffractions in the image space, two approaches are provided for extracting diffractions. Diffraction extraction can be directly accomplished by subtracting the predicted reflections from seismic imaging data if the residual diffractions are focused. Otherwise, a diffraction velocity analysis will be performed for refocusing residual diffractions. Two synthetic examples and one field application demonstrate the feasibility and efficiency of the two proposed methods in detecting the small-scale geologic scatterers, tiny faults and cavities.