A Modeling Study of Flow Diversion and Focusing in Unsaturated Fractured Rocks

Problems involving capillary barriers along sloping, layered unsaturated soils were first addressed by ZaslavThis study presents a systematic approach to analyze the flow sky and Sinai (1981a, 1981b). In the last few decades, diversion and flow focusing caused by the natural flow-barrier system several quantitative analyses of lateral water flow in in the unsaturated zone (UZ) of Yucca Mountain, Nevada, under amlayered porous media have been presented (Miyazaki, bient steady-state flow conditions. An existing analytical solution for analyzing capillary barrier in porous media has been extended to apply 1988; Ross, 1990, 1991; Steenhuis et al., 1991; Fayer et al., to the fractured porous rock. The new analytical solutions are used 1992; Oldenburg and Pruess, 1993; Yeh et al., 1994; to identify the critical layers and to provide the guidance for generation Stormont, 1995; Morel-Seytoux et al., 1996; Wilson, of a proper three-dimensional (3-D), site-scale numerical grid. A large1996; Warrick et al., 1997; Pan et al., 1997; Webb, 1997; scale 3-D numerical model (with more than a million grid blocks) has Ho and Webb, 1998; Morel-Seytoux and Nimmo, 1999). been developed with site-specific data to analyze the major flow Most investigations have focused on developing analytipatterns in the mountain. Our analyses show that large-scale lateral cal or numerical approaches for analyzing flow diversion flow could take place in the UZ under ambient conditions, as a result as the effect of either a single capillary barrier at the of capillary barriers formed at the contacts of heterogeneous rock layer contact interface or multiple but parallel substrata layers. This lateral flow runs generally toward the east (in the southern interfaces. What is lacking in the literature is a systempart) or southeast (in the northern part), which is consistent with the dip of the layer contacts. About 90% of the total lateral flow is found atic analysis of 3-D flow diversion and focusing in an to be conducted by only a few critical rock layers. Faults that penetrate unsaturated zone. In addition, understanding of the inthese rock layers act as vertical capillary barriers that stop the lateral teractions between barriers that are nonparallel or interflow. The combined effect of horizontal and vertical capillary barriers secting remains very limited. resulted in reduced percolation flow through repository horizon in During the early site characterization of the Yucca general but focused downward flow along those penetrating faults. Mountain UZ as a potential repository of high-level The model results were found to be consistent with the field water radioactive waste (Montazer and Wilson, 1984), the capsaturation. The findings of this study are consistent with a previously illary barrier and lateral flow concept was proposed. published two-dimensional (2-D) analysis and recent published modHowever, in situ direct measurement of the lateral flow eling results using field-observed Cl data. is technically difficult, if not impossible, mainly because of the dry conditions of Yucca Mountain. On the other hand, modeling studies, with proper conceptualization D and focusing of water flow in the unsatuand model parameters incorporated, can (i) provide a rated zone may occur in many natural subsurface complete picture of the complicated flow system for insystems. The main factor controlling such diversion and tegrating field measurements, (ii) indicate the measurfocusing is the spatial distribution of the flow barrier able variables to be used as indirect evidence to support (e.g., the capillary barrier formed at the contact interor deny a particular mechanism, and (iii) guide field faces between the strata of different soils and rocks with experiments and future modeling efforts to improve contrasting hydrological properties). In general, both our understanding of the system. As a result, several vertical and lateral heterogeneities exist in any natural numerical models (Rulon et al., 1986; Wittwer et al., layered formation. For example, a fault system, if pres1995; Moyer et al., 1996; Wu et al., 1998, 1999, 2000a) ent, could alter the spatial distribution of flow barriers have been used to explore the capillary-barrier phenomby breaking the continuity of those barriers, as well ena, but they came to quite different conclusions in as creating new barriers, through both displacement of terms of the amount of lateral flow that would occur in strata and alteration of hydrological properties. In a fractured tuffs. With improved understanding of the fractured unsaturated zone, fractures of varying intensite, scientists recently undertook a modeling study (Wu sity and geometry in different strata could further comet al., 2002b) of Yucca Mountain using several 2-D nuplicate the formation and distribution of these barriers. merical models. Their study identified some key hydroAs a result, lateral water movement and focusing of geologic conditions required to form capillary barriers downward flow may develop in the unsaturated zone and indicated possible large-scale lateral flow in the and lead to a complicated flow pattern, very different Yucca Mountain UZ. from what would occur in an unsaturated, layered, poOur objectives are (i) to present an extension of the rous medium system. analytical solutions of Warrick et al. (1997) to fractured rock, using an effective-continuum model (ECM), (ii) L. Pan, Y.-S. Wu, and K. Zhang, Earth Sciences Division, Lawrence to present a systematic analysis of the complicated 3-D Berkeley National Laboratory, Berkeley, CA 94720. Received 5 flow diversion and focusing in the UZ at Yucca MounMarch 2003. Original Research Paper. *Corresponding author (lpan@ lbl.gov). Abbreviations: CHn, Calico Hills nonwelded; ECM, effective-continuum model; PTn, Paintbrush nonwelded; TCw, Tiva Canyon welded; Published in Vadose Zone Journal 3:233–246 (2004).  Soil Science Society of America TSw, Topopah Spring welded; UZ, unsaturated zone; 2-D, two-dimensional; 3-D, three-dimensional. 677 S. Segoe Rd., Madison, WI 53711 USA