1 Petro-Canada Oil and Gas
2 Husky Oil Operations Ltd.
Conventional core description methods and core description tools generally prohibit or limit the potential of directly undertaking geostatistical analysis of sedimentary sequences. As a result, valuable information is missed which can have a significant impact on sedimentological interpretations, and on reservoir characterization and simulation.
The reservoir units in the Terra Nova field were deposited in a sandy braid-plain/braid-delta setting which prograded predominantly from the head of the Jeanne d'Arc Basin in several units. Each tectonically-triggered depositional cycle generally began with the incursion of conglomeratic alluvial fans, fan deltas or conglomerate-filled incised valleys which extended from the south-easterly, faulted margin of the basin.
To further enhance our understanding of the reservoir for reservoir characterization and modeling, probe permeameter measurements were taken at 2.5 cm intervals from each of the cored intervals through the Jeanne d'Arc Formation in five wells. At the same time, each core was re-logged using both conventional means and a spreadsheet. To create a numerical data base, the key sedimentological attributes were assigned numerical codes which were matched against depth, log-derived porosity and the probe permeability data in a spreadsheet for each well, using the same sample spacing as the probe permeameter data. With this data base, it was then possible to undertake statistical, geostatistical and spatial statistical analysis of the core description data, including variogram analysis and the development of vertical proportion curves for each progradational unit.
The results of this analysis directly impacted our understanding of the following:
1. The vertical permeability heterogeneity and variability within the reservoir, and its direct link to a sedimentological hierarchy of vertical length scales.
2. The interpretation of vertical and lateral variability of sedimentary facies, and the re-interpretation of some parts of the depositional setting/model.
3. The identification and selection of appropriate modern and ancient reservoir analogs.
4. The derivation of `geologically-based' reservoir rock units as the basis for our geostatistical modelling of the reservoir, as opposed to `petrophysically-based rock units', and a basis for differentiating those units in which capillary trapping may have a greater or lesser potential impact on reservoir performance.
5. The intrinsic length scales within the sedimentary succession of each progradational unit with respect to the determination of appropriate grid cell size and up-scaling factors.
6. The need to use a non-stationary model to properly portray the spatial variability of both geological and reservoir properties.
Through the quantification of core descriptions, the sedimentological and reservoir engineering aspects of a reservoir come together directly. Different questions are asked, usually resulting in new ideas, and a much better understanding of the reservoir architecture and what is important.
AAPG Search and Discovery Article #90928©1999 AAPG Annual Convention, San Antonio, Texas