Quantitative Geologic Description of Natural Fracture Systems (FOOTNOTE ^dagger)

J. H. Howard, R. C. Nolen-Hoeksema

A quantitative method for characterizing natural fracture systems differs from previous work by emphasizing small volumes of rock that may contain fractures and that collectively comprise a rock-fracture system. This

FOOTNOTE ^dagger. This paper does not necessarily represent the views of Standard Oil Production Company, for which J. H. Howard is currently employed.

description method is more similar to descriptive methods used in sedimentology than to traditional fracture description methods used in structural geology.

A "fracture system," as discussed here, consists of "domains." Domains are smaller volumes of the rock-fracture system that have been selected for measurement and quantitative summaries. "Samples" are parts of domains that are large enough to capture part of the fracture network yet small enough to be worked with routinely. The absolute size of samples is a fixed part of the description. Individual samples do not have to include fractures. However, for samples that do contain fractures, the usual measurements (such as width or orientation) are taken.

Data on length, width, and number of fractures per sample (including samples with no fracture) are summarized by histograms and distribution functions. Data on fracture orientations are summarized by contoured stereonets. These figures describe the fracture system within a domain, assuming that the sample data sets are adequately large to describe the local fracture population.

Monte Carlo combination of length, width, and number of fractures per sample for a domain leads to suites of estimated fracture porosities and permeabilities for samples within a domain. Spatial variation of the entire fracture system can be described by maps showing key aspects of each domain. These may include distribution functions and stereonets (as appropriate) of the basic parameters, as well as estimated porosities and permeabilities and/or selected representative values for these functions (e.g., the mean).

AAPG Search and Discovery Article #91037©1987 AAPG Southwest Section, Dallas, Texas, March 22-24, 1987.