AAPG Annual Convention and Exhibition

Datapages, Inc.Print this page

Full Use of Dipmeter Data for Geocellular Property Modeling in the McMurray Formation, Alberta

Abstract

Dipmeter data has traditionally been used to assist interpretation of geologic structure, to infer depositional environments, and determine paleocurrent direction. However, dipmeter data is seldom used as direct conditioning for population of geocellular model properties. We discuss the preparation and use of dipmeter data for interpolation of Facies and Gamma Ray properties in a fluvial point bar deposit. Our example is from the McMurray Formation in Alberta, Canada. Dipmeter data in fluvial point bar deposits are noisy, containing signals from trough cross bedding, subaqueous dune foresets, slumping, mud clast conglomerates, and inclined heterolithic sediments (IHS). Trough cross bedding and dune foresets indicate paleocurrent direction, while the orientations of IHS beds contain information about the architecture of bar-scale lateral accretion packages. When populating geocellular model properties, it is these roughly co-planar, dipping IHS orientations that tell us how to correlate the beds between wells. We have developed a processing method that measures the variation of dipmeter orientations within a moving window, which allows isolation of the coherent bed orientation signal of the co-planar IHS beds. Once the dipmeter data has been high-graded, it can be upscaled into a model framework, interpolated throughout the model volume, and then used as a Locally Varying Azimuth to orient variograms/training images in the property modeling algorithms. The workflow is as follows: (1) Create bed co-planarity “coherency” scores using moving depth windows in dipmeter data (2) Apply coherency score filtering and other criteria (i.e. exclude certain facies) to high-grade dipmeter readings for use in model conditioning (3) Supplemental bed dip data from geophysical survey interpretations can be added to the dataset (4) Decompose the bed dip azimuths and dip angles into unit circle vectors dX, dY, and dZ (strategy to avoid azimuth aliasing) (5) Upscale the dX, dY, and dZ coherent bed orientation data into a model framework (6) Interpolate dX, dY, and dZ through the geomodel framework using inverse distance squared weighting (7) Recombine unit circle vector properties into Dip Azimuth and Dip Angle (8) Replicate upscaled core and wireline log data around the wells according to interpolated Dip Azimuths and Dip Angles to enhance local honoring of coherent bed orientation data (9) Perform property modeling using the bed orientation properties as Locally Varying Azimuth for steering variograms