2005 Salt Lake City Annual Meeting (October 16–19, 2005)

Paper No. 8
Presentation Time: 3:35 PM

A DISCRETE FRACTURE NETWORK APPROACH TO QUANTIFYING HETEROGENEITY IN VARIABLY FRACTURED ROCK


SURRETTE, Megan J.1, ALLEN, Diana M.1 and JOURNEAY, Murray M.2, (1)Earth Sciences, Simon Fraser University, 8888 University Drive, Burnaby, BC V5A 1S6, Canada, (2)Geological Survey of Canada, 101-605 Robson Street, Vancouver, BC V6B 5J3, Canada, megans@sfu.ca

The use of discrete fracture network modeling in hydrogeology has typically been limited by the large amount of detailed fracture measurements required to accurately describe a fractured rock aquifer. For regional scale modeling, collection and representation of fracture data is not feasible, therefore, a common approach has been to represent the fractured rock mass by some equivalent property, and to treat it as a porous medium. This study applies the concept of hydrostructural domains, whereby fracture domains are represented and modeled using a stochastic, discrete fracture network approach (FracMan XP for MODFLOW and GMS). Approximately 9000 fractures measured at 125 stations in the Gulf Islands, British Columbia, Canada comprise the dataset. Analysis of the fracture data led to the development of a conceptual model with four hydraulically distinct hydrostructural domains – “highly” fractured interbedded sandstone and mudstone (<1.0 cm spacing), “less” fractured sandstone and mudstone (>1.0 m spacing), fault and fracture zones, and discrete faults and fractures. Hydrostructural domains are defined on the basis of changes in fracture intensity, and thus, characterize the distribution of relative permeabilities in the aquifer system. Models that statistically honor field data were constructed for representative stations of each of the four hydrostructural domains. Stochastic model results indicate that fracture set orientation is typically best-fit with Fisher and Bingham probability distribution functions, while lognormal and exponential distributions best describe fracture trace length for this data set; Kolmogorov-Smirnov and Chi-Squared statistics were used to evaluate the goodness-of-fit, and 90% significance was deemed acceptable. Model domains were populated with fractures based on these distributions. Preliminary model results indicate increased permeability resulting from greater fracture connectivity in hydrostructural domains with greater fracture intensity, such as fault and fracture zones. Away from fault and fracture zones permeability is substantially less.