Complex Features in Sedimentology and Truncated Pluri-gaussian Simulations

Helene Beucher, Ecole des Mines Centre de geostatistique, 35 rue St Honore, Fontainebleau, 77305, France, phone: 33 01 64 69 47 71, fax: 33 01 64 69 47 05, beucher@cg.ensmp.fr, Brigitte Doligez, Institut Français du Pétrole, Dept. of Geology & Geochemistry, 1 et 4 avenue de Bois Preau, 92506 Rueil-Malmaison Cedex, France, and François Geffroy, Ecole des Mines Centre de Geostatistique, Fontainebleau, France.

In the domain of reservoir characterization, the facies encountered exhibit a great variety of features. These anisotropies usually result from successive and independent sedimentation process, reef deposit in marine environment for instance or from post sedimentation process as diagenesis.

Among the different models used for simulating categorical variables, the pluri-gaussian model is an efficient tool that allows us to reproduce complex arrangements of several lithofacies.

This model is an extension of the well-known truncated gaussian method. In addition to the usual parameters (facies order, proportion curves and structure of the underlying gaussian random function) it requires the specification of a second gaussian random function. The two random functions can be correlated and possibly shifted. Finally, the threshold scheme rules the arrangement of the facies. These parameters must also account for non stationarity. Such a model can be used to simulate heterogeneities as various as argillaceous plugs in a fluvial environment or algal mounds on carbonate platforms. It also makes it possible to combine different physical process such as diagenesis or fracturation combined to sedimentation.

The most recent developments involve the inference of the numerous parameters, the integration of complementary information (global sand-shale ratio, seismic attributes, trends from analog outcrop studies...).

This paper reviews the characteristics of the method. It focuses on practical examples to address different geological problems characterized by complex relationships and contacts between the facies. Finally, a number of images using particular couples of gaussian function to simulate realistic fluvial deposit are presented, providing an outlook on very promising future applications of this method.

SEPM: Quantitiative Stratigraphy and Geostatistics: New Approaches and Applications
AAPG Annual Meeting 2003: Energy - Our Monumental Task Technical Program