THE EFFECTS OF ELIMINATING FRACTURES BELOW SPECIFIED THRESHOLDS WHEN MODELING DISSOLVED-PHASE DNAPL MIGRATION IN CRYSTALLINE ROCK

Dense non-aqueous phase liquids (DNAPL), such as chlorinated solvents, are pervasive ground-water contaminants detected in industrial areas throughout North America. Characterizing the fate of such contaminants in porous media has been a challenging task. In fractured rock aquifers, characterizing the fate of DNAPL is greatly exacerbated due to the extreme variability in hydraulic properties of fractures and their complex connectivity. In addition, individual fractures have varying aperture, which, along with discontinuous branches of the conductive fracture network, may entrap DNAPL in immobile pools. Understanding the role of interconnected fractures of varying lengths and transmissivity (i.e. the network structure) in constraining fluid movement and chemical migration is critical in understanding the fate of the dissolved-phase components of DNAPL and assessing remediation strategies. In simulations of fractured rock, it is computationally impossible to represent explicitly all fractures, even when considering a discrete stochastic interpretation of the formation. A threshold must be introduced where fractures of given characteristics are eliminated from model simulations. The effects of variable network structure on dissolved-phase DNAPL migration are examined by considering a large suite of flow and transport simulations, using a range of network structures often reported in field studies. We evaluate the effect of removing fractures below specified cutoffs in the context of characterizing dissolved-phase DNAPL migration, and discuss the feasibility of correcting for their influence on simulated results. Preliminary results suggest that in systems where percolation is maintained, omitting fractures of the smallest lengths and transmissivity imposes a minor effect on transport predictions, but when connectivity intensity is largely dependent on omitted features a more pronounced effect occurs, leading to a dramatic variation in transport behavior.