Abstract: Stochastic Simulation of Deep Water Oligocene Turbidites (Campos Basin) Guided by a Stratigraphic Model Based Seismic Inversion

Santos, Rogerio A.; Olinto G. Souza Jr.; Mauro R. Becker and Ricardo R. P. Alves - Petrobras/E&P/Cenpes

A stochastic model comprising a truncated Gaussian field has been used to generate an improved 3-D earth model, which attempts to represent a turbidite succession, in terms of lithotypes, which is the reservoir of a giant oil field in deep-water (600 - 1200 m) at SE Brazil. The goal of this 3-D reservoir characterization is to derive more realistic images of the inter-well heterogeneity using integrated geophysical and geological data.

The reservoir is a dip-elongated Oligocene turbidite fan, NW/SE oriented, which were strongly cut by moderate to low sinuosity deep-sea channels. Reservoir fault systems have their main direction along and perpendicular to the fan orientation and are closely related to salt movement. This reservoir sandstone covers an area of 130 Km², with thickness reaching 25 m. It shows average porosity of 30% and high permeability ( above 1000 mD).

The seismic volume was acquired in 1986 and reprocessed in 1996. This reservoir oriented reprocessing was done to improve the amplitude recovering, the well- seismic phase deconvolution and multiple attenuation, and also to obtain a more detailed velocity field for stacking and migration. The result of this reprocessing, which enhanced the vertical seismic resolution, allowed a better understanding of either the stratigraphic and structural model for the field (Fig. 1). Although there is an excellent correlation between amplitudes and sand thickness and a good correspondence between time and depth, due to a detailed velocity analysis, the individualization of each high resolution stratigraphic sequence can not be done from the 3D seismic amplitudes. But using acoustic impedances derived from constrained seismic amplitudes inversions, the correlation among the stratigraphic sequences and the seismic data is improved. The seismic resolution after inversion is controlled not only by the time sampling of logs, but by a kind of "adjustable" seismic resolution for each stratigraphic sequence.

The stratigraphic model and the depositional elements of the turbidite system are the framework used to constrain the 3-D gridded geological model, which represents the reservoir architecture in terms of spatial distribution of the lithotypes. The stochastic model attempts to constrain 3-D earth models with 2-D seismic information related to the proportions of lithotypes. A 3-D proportion matrix is therefore required. This matrix is interpolated from well data constrained by seismic data (Fig. 2). The main philosophy behind this method is to establish a relationship between the spatial distribution model of the lithotypes proportions and the realization of a Gaussian random function in the analyzed stratigraphic interval.

AAPG Search and Discovery Article #90933©1998 ABGP/AAPG International Conference and Exhibition, Rio de Janeiro, Brazil