introduction

The discrimination between shales and sands in the subsurface is key to oil exploration. Reservoirs usually exist in sand layers, sandwiched between large columns of shale layers. In a previous paper (Alkhalifah and Rampton, 1997), we showed that the anisotropy parameter $\eta$, extracted from surface seismic P-wave data, can be used, in a low resolution sense, to discriminate between shale-dominated layers and sand-dominated ones. The underlying theory is that shales induce anisotropy (Banik, 1984), positive $\eta$ in particular, and sands do not ($\eta \approx 0$). In our earlier paper (Alkhalifah and Rampton, 1997), we implemented the nonhyperbolic moveout inversion of Alkhalifah (1997) on a line that included spread length with offsets up to 6 km. The resulting $\eta$ estimates showed reasonable correlations with the reflections on seismic sections. Also, the estimates correlate well with gamma-ray well-log measurements, used as a shale estimate, which proves the credibility of these results.

In this paper, we use modified inversion tools, compared to the one used in our previous paper (Alkhalifah and Rampton, 1997), which although are based on the same underlying theory of nonhyperbolic inversion, provide more stable estimates of $\eta$ over a larger scale. The inversion now includes a damping process that constrains the $\eta$ values at depth, where the inversion is less stable, and a picking feature that helps constrain the NMO velocity and thus avoid multiples. We apply the inversion to two portions of a line from Trinidad and obtain 2-D plots of $\eta$ distribution in the subsurface at these two regions.