Data Resolution and Scale-dependent Fracture Clustering: Implications for Deformation Mechanisms
Thursday, 17 December 2015
Poster Hall (Moscone South)
Fracture spacing data collected from scanlines and wells at various resolutions are commonly analyzed for the purposes of aquifer and reservoir characterization. It has been previously found for a certain set of nested fracture networks with similar fractal dimensions that differences in scale of observation hence, resolution lead to differences in clustering. It remains however, to be established whether differences in resolution of fracture spacing data can lead to significant differences in clustering behavior for a wider range of datasets. Most studies on fracture analysis either consider the cumulative frequency of spacing data without regard to the actual sequence of the spacing values or compute an average spacing that may not delineate clustered fractures. The coefficient of variation parameter, Cv is often used to differentiate between clustered, random, and unclustered fractures in a scanline but does not address the issue of scale-dependent variations. Lacunarity is a parameter that has been previously used for quantifying the scale-dependent clustering of spatial patterns and recently, this technique has also been used for identifying scale-dependent pattern changes from scanline data. The current research illustrates the application of this technique for delineating differences between scale-dependent clustering attributes of data collected at various resolutions along the same scanline. Specifically, data were collected at different resolutions from two outcrop exposures, a cliff-section and pavement, of the Cretaceous turbititic sandstones of the Chatsworth Formation widely exposed in southern California (USA). For each scanline, low resolution aerial photographs and high resolution ground measurements are analyzed for scale-dependent clustering attributes. In terms of scale-dependent lacunarity, higher resolution data show larger values than their respective low-resolution counterparts. It is postulated that lower resolution data captures fracture zones that have relatively uniform spacing while higher resolution data capture both thin and short splay joints and shear joints that form fracture clusters. Therefore, it may be concluded that data resolution is critical for identifying deformation mechanisms and their products.