Northeastern Section (45th Annual) and Southeastern Section (59th Annual) Joint Meeting (13-16 March 2010)

Paper No. 4
Presentation Time: 2:35 PM

USING GIS AND DISCRETE FRACTURE NETWORK MODELS TO COMPARE THREE FRACTURE SAMPLING TECHNIQUES


MANDA, Alex K., Department of Geological Sciences and Institute for Coastal Science and Policy, East Carolina University, 387 Flanagan Building, East 5th Street, Greenville, NC 27858 and MABEE, Stephen B., Massachusetts Geological Survey, Univ. Massachusetts, 611 North Pleasant Street, Amherst, MA 01003, mandaa@ecu.edu

Various fracture sampling techniques are used by hydrogeologists, structural geologists and other workers to sample fractures in the field. These techniques include the single scanline, selection, and multiple scanline methods. The most demanding of these methods is the multiple scanline method followed by the single scanline method. The selection method is the least intensive method because it does not rely on deploying scanlines to sample fractures. Although these methods are commonly used to sample fractures for various purposes, a quantitative comparison of the three methods has yet to be undertaken. In this paper we test the three fracture sampling techniques against each other to determine which of the three methods is the most effective at sampling fractures. Field data and Geographic Information System (GIS) techniques were used to create a fracture map observed in layered dolomitic rocks with fracture sets that comprise persistently parallel fractures. GIS techniques were then used to derive fracture attributes from the fracture map using two versions of each of the three fracture sampling methods. The statistics collected with each method were then used to build synthetic fracture networks. These stochastic models were compared to a model that had the exact locations, sizes and orientations of fractures as those observed in the natural fracture network. Fracture intensities and permeabilities derived from each model were then used to assess the effectiveness of each method in recreating the natural fracture network. Results reveal that the selection and multiple scanline methods were more consistent than the single scanline method in honoring field data. However, the results also reveal that a well placed scanline may outperform the selection and multiple scanline methods. Thus, although a well placed scanline requires more time and labor to employ than the selection method, it may be ideal at smaller outcrops. According to the results, there is little to be gained from using multiple scanline methods because they are very intensive and do not perform better than the other methods. Owing to the amount of work that is involved in carrying out surveys with scanlines, we therefore suggest using the selection method in sampling persistently parallel fractures at large outcrops or where time is of the essence.