Premium
Effects of simplifying fracture network representation on inert chemical migration in fracture‐controlled aquifers
Author(s) -
Wellman Tristan P.,
Shapiro Allen M.,
Hill Mary C.
Publication year - 2009
Publication title -
water resources research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.863
H-Index - 217
eISSN - 1944-7973
pISSN - 0043-1397
DOI - 10.1029/2008wr007025
Subject(s) - aquifer , fracture (geology) , plume , advection , geology , dispersion (optics) , network model , residence time (fluid dynamics) , representation (politics) , flow (mathematics) , soil science , mechanics , econometrics , geotechnical engineering , groundwater , mathematics , computer science , data mining , meteorology , geometry , thermodynamics , geography , physics , optics , politics , law , political science
While it is widely recognized that highly permeable ‘large‐scale' fractures dominate chemical migration in many fractured aquifers, recent studies suggest that the pervasive ‘small‐scale’ fracturing once considered of less significance can be equally important for characterizing the spatial extent and residence time associated with transport processes. A detailed examination of chemical migration through fracture‐controlled aquifers is used to advance this conceptual understanding. The influence of fracture structure is evaluated by quantifying the effects to transport caused by a systematic removal of fractures from three‐dimensional discrete fracture models whose attributes are derived from geologic and hydrologic conditions at multiple field sites. Results indicate that the effects to transport caused by network simplification are sensitive to the fracture network characteristics, degree of network simplification, and plume travel distance, but primarily in an indirect sense since correlation to individual attributes is limited. Transport processes can be ‘enhanced’ or ‘restricted’ from network simplification meaning that the elimination of fractures may increase or decrease mass migration, mean travel time, dispersion, and tailing of the concentration plume. The results demonstrate why, for instance, chemical migration may not follow the classic advection‐dispersion equation where dispersion approximates the effect of the ignored geologic structure as a strictly additive process to the mean flow. The analyses further reveal that the prediction error caused by fracture network simplification is reduced by at least 50% using the median estimate from an ensemble of simplified fracture network models, and that the error from network simplification is at least 70% less than the stochastic variability from multiple realizations.