Effects of fracture geometry on fluid flow through the Monterey Formation, California : an application of a 3D discrete fracture simulator

Outcrops of the fractured and folded sedimentary Monterey Formation located along or near the coast of central California were sampled to characterize the geometry of the fracture system and determine the impact of the fracture geometry on rock mass permeability. A scanline mapping technique was used to obtain: orientation, trace length, spacing, mineral infill, termination, censoring, and shape data for over 1700 individual discontinuities. Analyzing the general fracture characteristics, using cluster analysis of the poles to fracture planes, and taking into account the different structural settings, it was determined that there were two to three fracture sets at each of six different locations. Four of these locations, were part of a "small-scale" fracture survey where short scanlines (≤ 30 m) sampled fractures 0.25 m or greater in trace length from individual outcrops. The other two locations, were a composite of sites sampled for "large-scale" fractures (extensive breccias and faults) with trace lengths greater than 3 m, using long (100 m to 1400 m) scanlines. The strike of the dominant fracture set at most small-scale sites was sub-parallel to bedding dip direction. Generally the fractures in the subordinate set had shorter trace lengths that terminated against and were approximately perpendicular to the fractures making up the dominant set. The two large-scale sites had three sets, one sub-horizontal, one sub-parallel to bedding strike, and the other sub-parallel to bedding dip direction, all with similar mean lengths. -- Statistical methods were used to correct for some of the biases, associated with scanline mapping, in the trace length, spacing and orientation data. Straight lines, drawn through the data points on probability plots show that die fracture trace length and spacing data can be approximated by a lognormal distribution. The statistics of trace length, density, and orientation obtained from the field data, along with assumed aperture distributions were used in a 3-D discrete fracture flow simulator to evaluate the impact of the variability in fracture orientation, fracture interconnectivity, and aperture on rock mass permeability. Simulations of rock mass cubes with side length of 8 to 18 meters for the small-scale data and 100 to 150 meters for the large-scale data, found that the fracture geometry was well connected and imparted a strong anisotropy to flow in the horizontal section and less so in the vertical section, when the same aperture was assigned to each fracture. When the apertures for individual fractures were generated using a lognormal distribution the computed directional permeabilities became even more anisotropic to flow