RAYMER, D.G. * and J-M.KENDALL, University of Leeds, G.J.BEAUDOIN, M.C.MUELLER and R.R.KENDALL,Amoco Exploration and Production

Abstract: Measuring the anisotropy of salt in the Mahogany oil field, Gulf of Mexico

Salt flows and deforms during emplacement of evaporitic structures. Since it is known from experiment and theory that deformation in evaporites can lead to the preferential alignment of the constituent crystals (which are themselves elastically anisotropic) such a preferred alignment produces an anisotropic aggregate. In this work we investigate evidence of salt anisotropy in data and use numerical simulations of crystal alignment to constrain our interpretations. The aim is to measure in-situ anisotropy in a large scale salt structure over the Mahogany oil field in the Gulf of Mexico. The data we consider comes from 3 sources: surface seismics, cross-dipole sonic logs and VSPs.

Elastic constants derived from numerical simulations of flow-induced petrofabrics (Raymer et al., 1998) are used to build simple models of an anisotropic salt sill. Two types of deformation regimes are considered; that due to simple shear and that due to planar extension. In addition to these two models we also consider a hybrid model of the two deformation regimes and a reference isotropic salt model. The anisotropy is calculated in terms of Thomsen parameters (Thomsen, 1986) as these are closely related to properties which can be measured with seismic surveys. Thomsen parameters for the simple shear case show very little azimuthal variation, but the planar extension case shows considerable azimuthal variation (Fig.1). Thomsen's delta parameter (a measure of near-vertical anisotropy) for the simple shear case has a value of delta = -0.08 for all azimuths. In the case of planar extension, delta is negative along the line of flow and positive orthogonal to this. These results suggest that it may be possible to identify styles of deformation based on measurements of anisotropy parameters.

Synthetic data is used to test a method for measuring delta which involves measuring non-hyperbolic move-out in anisotropic media (Tsvankin and Thomsen, 1994). Seismograms for surface recorded reflections from the top and base of salt are generated using ATRAK, a ray tracer for 3D, multilayered, inhomogeneous, anisotropic media (Guest and Kendall, 1993). The move-out velocities are calculated using standard isotropic methods and the interval velocity for the salt is calculated using Dix's equation. Due to non-hyperbolic move-out, this interval velocity is not the true velocity of the medium.This velocity and the vertical velocity, which can be calculated from VSP data, are then used to obtain delta.We investigate the sensitivity of delta estimates with selected maximum offset. The optimum offset for achieving a match to the actual delta values (calculated from elastic constants) occurs for offsets between a third and a half of the distance to the base salt.

The Mahogany structure is approximately 10km long and 1km thick and results from a horizontal emplacement. This was the first oil field at which sub-salt production occurred. Surface seismic data comes from two 4-component OBS lines close to the well. Semblance velocities obtained from CDP supergathers in the region of interest are used to calculate the interval velocities.The chosen CDP's are those which give clear and consistent top and base salt reflections for a range of neighbouring CDPs.The vertical velocities in the salt are obtained from the VSP at the discovery well (Fig.2).

Delta estimates for the line closest to the discovery well give an average value of delta =-0.13 for maximum offsets of 1.5km (Fig.3). This is in good agreement with the predictions from the numerical simulation of halite deformation. The method used has a few key assumptions: flat layering, no lateral heterogeneity, velocities at the well are the same as those at the seismic line and that the deviation from hyperbolic move-out is entirely due to anisotropy. Further modelling has shown that the results are quite insensitive to low levels of dip (<10 degrees) in the reflecting layers.

Our hypothesis of anisotropic salt structures is further tested using crossed-dipole sonic data which calculates Thomsen's gamma parameter (a measure of azimuthal shear-wave anisotropy). Initial results suggest low gamma implying limited but non-zero cross-line energy. Furthermore, current analyses of shear-wave splitting in VSP data shows significant splitting in the salt.

The results of this work suggest that salt at Mahogany is anisotropic and the effects on wave propagation could be quite significant. Ignoring this anisotropy in sub-salt imaging could lead to costly mislocation of targets.

AAPG Search and Discovery Article #90923@1999 International Conference and Exhibition, Birmingham, England