Improved Modeling of Deepwater Reservoir Architecture and Shapes with Multipoint Geostatistics Applied to a Deepwater Outcrop Analog, Tanqua Karoo Basin, South Africa
David S. McCormick1, Daniel Tetzlaff1, Roy Davies2, Tuanfeng Zhang3,
Nneka Williams1, Claude Signer1, and Dave M. Hodgson2
1 Schlumberger-Doll Research, Cambridge, MA
2 Liverpool University, Liverpool, United Kingdom
3 Stanford University, Stanford, CA
We have implemented and tested 2D and 3D multipoint geostatistics algorithms (Snesim and FilterSim) for geological modeling using a 3D digital dataset from the EU-funded industry–academic NOMAD project (Novel Modelled Analogue Data for more efficient exploitation of deepwater hydrocarbon reservoirs), of deepwater outcrops in the Tanqua Karoo Basin. Data included sedimentary facies logs and paleocurrent information (including borehole image logs).
Our multipoint geostatistical implementations demonstrate the ability of this family of algorithms to efficiently and faithfully reproduce the shapes and texture of geological facies while honoring a large number of hard data, rotation and scaling fields, and soft probability fields, all in a reservoir model comprising ten's of millions of cells. We have used stationary “training images” with 3-5 facies to model the proximal to distal facies relationships seen in outcrop. The results show the ability to honor hard data, soft constraints, and complex geological relationships, results that cannot be achieved with conventional two-point geostatistics or object modeling.
One of the virtues of these algorithms is their ability to honor multiple soft probability fields when probability distributions overlap; there is no requirement for exclusive categorical membership with sharp cut-offs. This promises to allow better use of seismically derived attribute data for modeling of complex reservoirs.
The examples presented here use data from one of these fan models to demonstrate the power of the multipoint geostatistical algorithms for producing highly detailed reservoir models that mimic complex geological shapes and textures that will honor the flow behavior of the reservoir.