Impact of horizontal spatial clustering in two-dimensional fracture networks on solute transport

Abstract

Listen

Translate article

This study explores the prevalence of fracture clustering in natural networks and its influence on fluid flow and advective transport. Fracture networks from a compilation of previously published maps that encompass igneous, metamorphic and sedimentary rock types, various geological and tectonic settings, and a broad distribution of length scales are analyzed using a two-point correlation technique to characterize the spatial scaling between fracture barycenters. The correlation dimension (Dc) ranges from 1.47 to 1.89 with a median of 1.8. These results suggest that fracture networks often exhibit spatial clustering and are rarely Poissonian as commonly assumed. Synthetic Discrete Fracture Networks (DFNs) with constant and variable transmissivity applied to all fractures are generated for three clustering (Dc = 1.6, Dc = 1.8, and Poissonian) and power law length exponent (α = 1.4, α = 1.8, and α = 2.2) scenarios. Median particle breakthrough times are shown to increase by 10–70% as spatial clustering transitions from Poissonian to highly clustered (Dc = 1.6). Median particle breakthrough times also increase, by a factor of 1.2–3.4 (20–240%), when the power law length exponent increases from α = 1.4 to α = 2.2. As clustering increases (i.e., Dc reduces), breakthrough curves become highly variable and exhibit features indicative of enhanced solute retention. The increase in median breakthrough times with clustering persists with different network size and transmissivity distribution. These results indicate that spatial clustering and length exponent both significantly influence transport. The use of power-law length distributions in generating fracture networks is relatively common, whereas the spatial organization of fractures is rarely included in transport studies of fractured rock. Structural heterogeneity introduced by spatially clustered organization of fractures leads to substantial complexity in flow paths which is not captured by Poissonian DFNs.