Abstract: Stochastic Modeling of Sequence Boundary Geometry

Alister C. MacDonald, Lars Magnus Falt, Anne-Lise Hektoen

Stochastic heterogeneity models should incorporate methods which adequately reflect modern geological concepts. Most techniques however only describe the spatial distribution of facies bodies, and omit some important sequence stratigraphic considerations.

Recognizing the importance of capturing uncertainty in the geometry of zonal bounding surfaces, a prototype stochastic model has been developed using a combination of marked-point (fibre) processes and Gaussian fields. The model is optimised to describe incised valley geometries in fluvial reservoirs and simulates a single surface in three stages: (1) the orientation and location of the valleys is modelled as fibres; (2) an "expectation surface," described by four 1-D Gaussian functions (width, maximum depth, sinuosity, and asymmetry), is generated to define the general geometry of the valleys. The cross sectional form is described by a power function where high values lead to the simulation of steep valley margins, and low values result in more gently dipping valley sides; and (3) a -D Gaussian function, which has different correlation structures in the valleys and interfluves, is simulated on top of the expectation surface to describe the detailed geometry of the bounding surface. The reservoir model framework is generated by simulating several of these (mutually erosive) surfaces from the base upwards.

The routine has been tested on a number of reservoirs. In low net-to-gross fluvial reservoirs, the geometry of the sequence boundaries controls the distribution of reservoir sandstones, and correct modeling is significant for designing an optimal infill drilling strategy. In high net-to-gross reservoirs, the erosional geometry controls the lateral extent of shale barriers. In this case, correct modeling has an impact on the prediction of local vertical permeability and potential coning problems.

AAPG Search and Discovery Article #90986©1994 AAPG Annual Convention, Denver, Colorado, June 12-15, 1994