HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Dinwiddie, C. L. Articles by Ferrill, D. A. Search for Related Content PubMed Articles by Dinwiddie, C. L. Articles by Ferrill, D. A. GeoRef GeoRef Citation Agricola Articles by Dinwiddie, C. L. Articles by Ferrill, D. A. Related Collections Capillary Barriers Field-Scale Studies Spatial Variability Fractured Rock

ORIGINAL RESEARCH

Fault Zone Deformation Overprints Permeability of Nonwelded Ignimbrite: Chalk Cove Fault, Bishop Tuff, Bishop, California

^a Center for Nuclear Waste Regulatory Analyses, Geosciences and Engineering Division of Southwest Research Institute, 6220 Culebra Road, San Antonio, TX 78238
^b Utah Faults, Fractures, and Fluids (UF³), Innovation Campus, Utah State Univ., 1770 Research Park Way, Ste. 183, North Logan, UT 84341
^c now at Dept. of Earth, Material, and Planetary Sciences, Geosciences and Engineering Division of Southwest Research Institute
^d now at U.S. Nuclear Regulatory Commission
^* Corresponding author (cdinwiddie{at}swri.org)

Received 2 May 2005.

	ABSTRACT

TOP EXECUTIVE SUMMARY ABSTRACT INTRODUCTION Background Objectives LOCATION AND DESCRIPTION OF... STRUCTURAL AND HYDROGEOLOGICAL... RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Deformation-induced secondary heterogeneities associated with a steeply-dipping normal fault in the nonwelded Bishop Tuff were examined in a multidisciplinary study using in situ gas permeability tests, structural mapping, and laboratory analyses as analogs for fault deformation features within the poorly exposed Paintbrush nonwelded hydrogeologic unit at Yucca Mountain, Nevada. Three features of fault-zone deformation were identified that may act to constrain lateral flow in unsaturated nonwelded ignimbrites. First, development of a fault gouge, with a decrease in grain size caused by grain comminution, decreases the permeability of this element with respect to the host rock. Second, long open fractures paralleling the fault have an increased permeability with respect to the host rock. Third, small-scale fractures and grain rotation in adjacent matrix blocks induce an interconnected porosity not seen in host rock. Intrinsic permeability variation beyond what is observed in the host rock reflects the effect of small-scale deformation in the matrix adjacent the fault. These features can reduce the continuity of potential capillary or permeability barriers, and thus limit any redistribution of percolation that would result from lateral flow diversion.

Abbreviations: PTn, Paintbrush nonwelded [unit] • TCw, Tiva Canyon welded [unit] • TSw, Topopah Spring welded [unit] • XRD, X-ray diffraction

	INTRODUCTION

TOP EXECUTIVE SUMMARY ABSTRACT INTRODUCTION Background Objectives LOCATION AND DESCRIPTION OF... STRUCTURAL AND HYDROGEOLOGICAL... RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
INTRINSIC PERMEABILITY is one of a few primary variables that influence fluid flow in porous media, and small-scale (centimeter-scale) permeability heterogeneity plays a substantial role in the migration of fluids in many unsaturated zone flow problems. Primary geological heterogeneities on the micrometer to meter scale are caused by variations in physical deposition and compaction processes throughout space and time, and by postdeposition but precompaction biological processes. The sum of these processes may lead to the occurrence of distinct beds or laminae, internal physical and biological sedimentary structures, and variations in pore morphology. Secondary geological heterogeneities may be an effect of either postcompaction geochemical processes, which cause solution, precipitation, cementation, recrystallization, and authigenesis, or postcompaction physical processes, including tectonism. Given the large variability of possible combinations of processes that can affect the character of the geosphere, geological heterogeneity is usually evident when permeability measurements are made at the centimeter scale. Of the many types of secondary geological heterogeneities, this study focused on those caused by deformation in the form of faulting and fault-related fracturing of host rock.

Within pyroclastic flow deposits, permeability heterogeneities on the centimeter to meter scale may be associated with variations in petrophysical controls (e.g., porosity, degree of welding and recrystallization, grain mineralogy), variations in devitrification, occurrence of interbedded ash-fall or surge deposits, and the presence of deformation structures such as cooling joints, open fractures, mineralized fractures, and fault cores composed of comminuted gouge and clays. Nonwelded pyroclastic flow deposits, or ignimbrites, are often weakly consolidated, but have sufficient strength to undergo localized brittle failure (Wong et al., 2001; Wilson et al., 2003; Evans and Bradbury, 2004). Studies by Sigda and Wilson (2003) and Winograd (1971) found that fault and fracture deformation of weakly consolidated materials may significantly affect the preferential flow direction of fluids in the unsaturated zone.

Within unsaturated nonwelded ignimbrites, a subvertical fault often functions as a permeability barrier to cross-fault flow. In a typical fault system, the fault core is subject to the greatest amount of deformation. The fault core is more comminuted, or finer grained, than the surrounding rock, and thus tends to be a relatively low-permeability element that acts as a permeability barrier to cross-fault flow for all flow conditions, and as a conduit for vertical flow during low-flow unsaturated conditions.

Fractures in damage zones surrounding a subvertical fault may act as conduits to vertical gravity flow (Evans and Bradbury, 2004) and as barriers against large-scale lateral flow. Damage zones may contain both open and mineralized fractures. Open fractures are relatively high-permeability elements, and, if subvertically oriented, they preferentially promote vertical drainage of unsaturated flow. Filled fractures may or may not have higher permeability than adjacent rock material. Nevertheless, Fedors et al. (2001) demonstrated that the presence of subvertical, caliche-filled fractures within a nonwelded ignimbrite can enhance vertical matrix flow by constraining water to flow within the matrix adjacent to fractures.

Theoretically, the extent of lateral flow associated with bedded, dipping, nonwelded ignimbrites depends, in part, on the degree of lateral heterogeneity found in the units. To help evaluate potential length scales for lateral flow in the poorly exposed nonwelded ignimbrites at Yucca Mountain, field studies linking hydrology and structural geology were undertaken within the well-exposed nonwelded units of the Bishop Tuff near Bishop, CA. Yucca Mountain is the site of a potential high-level radioactive waste and spent nuclear fuel repository in southern Nevada.

	Background

TOP EXECUTIVE SUMMARY ABSTRACT INTRODUCTION Background Objectives LOCATION AND DESCRIPTION OF... STRUCTURAL AND HYDROGEOLOGICAL... RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Detailed studies pairing hydrogeological and structural characterization of small faults and fractures within nonwelded ignimbrite on the scale of less than a millimeter to a few meters are limited. Part of the problem resides in a misconception that because matrix flow is dominant in nonwelded ignimbrites, fractures and faults are sparse. Fedors and Ferrill (2002) estimated from data in the Exploratory Studies Facility at Yucca Mountain an average fault spacing of 3.65 m (12 ft) in nonwelded units. The average dip of the faults is estimated to be 70° and at a high angle to stratigraphic layering. Hinds et al. (2003) noted that centimeter- to meter-scale fractures comprise the majority of fractures in welded and nonwelded tuffs at Yucca Mountain. Models representing the effects of these features are absent.

Several workers have examined the hydrogeologic significance of faults and fractures in unconsolidated or poorly consolidated sandstone and ignimbrite deposits with porosity values similar to the Bishop Tuff (e.g., Antonellini and Aydin, 1994; Manzocchi et al., 1998; Heynekamp et al., 1999; Sigda et al., 1999; Sigda and Wilson, 2003; Wilson et al., 2003).

Wilson et al. (2003) observed both shear fractures and deformation bands in nonwelded, crystallized ignimbrites (Bandelier Tuff, New Mexico). Their work suggests that the primary controls on mode of failure within an ignimbrite are grain-contact area and strength. An increase in grain-contact area through postdepositional crystallization provides a rigid framework and a subsequent increase in the mechanical strength of the rock through which fractures may propagate (Moon, 1993; Wilson et al., 2003). Deformation band fault zones were preferentially altered and cemented relative to the host rock in the Wilson et al. (2003) study, implying localized fluid flow within the unsaturated fault zone. This observation is similar to that of sandstones in which fault-zone mineralization occurs between bounding slip surfaces and suggests a preferential flow path parallel to the direction of the deformation bands (Antonellini and Aydin, 1994).

Sigda et al. (1999) and Sigda and Wilson (2003) showed that small faults within poorly lithified sediments may act as vertical conduits to flow under high matric potentials in the unsaturated zone, whereas small faults may be barriers to fluid flow in the saturated zone. For any fault zone, the architecture of intrinsic permeability ultimately will depend on the grain-size distribution of the host rock, the type and magnitude of deformation, and the degree of cementation, mineralization, and welding (Heynekamp et al., 1999).

Lithologic heterogeneities, structural discontinuities, and hydraulic properties of ignimbrites vary spatially at the meter scale in tectonically active settings (Flint et al., 2003). High-porosity materials, such as the nonwelded deposits within the Bishop Tuff and Paintbrush Group at Yucca Mountain, favor strain-hardening deformation processes and the development of highly connected, small-scale fracture systems (Manzocchi et al., 1998). The distribution, characteristics, and spatial variability of these small-scale features and their hydraulic properties are necessary to accurately evaluate the effects of lateral flow within faulted nonwelded ignimbrites in the unsaturated zone.

The work we describe was motivated by models of flow in the Paintbrush nonwelded hydrogeologic unit at Yucca Mountain. This unit is located stratigraphically above the site of a potential geologic repository for high-level radioactive waste and spent nuclear fuel. Of particular interest is the potential for conditions favorable to the formation of flow barriers of various types and orientations that may exist within the Paintbrush nonwelded unit (Wu et al., 2000; Fedors et al., 2002; Flint et al., 2003). The overall interaction, predominant behavior, and influence of such flow barriers on repository performance are not completely understood. For example, based on numerical modeling of subhorizontal capillary and permeability barriers within the Paintbrush nonwelded unit by Wu et al. (2000, 2003) and Bechtel SAIC Company, LLC (2001), flow barriers at lithologic contacts between fine- and coarse-grained subunits may cause large-scale lateral diversion of water to faults located down dip and outside the repository footprint. However, heterogeneity within hydrogeologic subunits and gradational contacts between subunits would disrupt the continuity of flow barriers and limit the extent of lateral flow diversion (Ho and Webb, 1998; Waiting et al., 2001; Fedors et al., 2002; Flint et al., 2003). Furthermore, Fedors and Ferrill (2002) and Bechtel SAIC Company, LLC (2001) suggested that the presence of steeply dipping fractures and faults and the occasional absence of the Paintbrush nonwelded unit would limit the lateral extent of flow diversion.

Previous field and laboratory investigations within the Bishop Tuff have shown that deformation caused by small faults may induce local changes in matrix properties that lead to vertical flow (Fedors et al., 2001, 2002; Fedors and Ferrill, 2002). At a minimum, small faults and fractures may cause flow pathways to become more tortuous, while at a maximum, they may compartmentalize the host rock by physically disconnecting the matrix blocks on either side of structural features (Fedors et al., 2002). Results from dye tracer tests indicate that faults and fractures can constrain lateral flow in non- and poorly welded ignimbrites in the unsaturated zone (Fedors et al., 2001, 2002). The presence of faults and fractures may also explain the periodicity of vertical flow breakthrough within the Paintbrush nonwelded hydrogeologic unit, particularly when fault surfaces are oriented at a high angle to bedding, which is typical of this unit. The heterogeneity of matrix properties within nonwelded ignimbrites, including primary depositional textures, devitrification or vapor phase crystallization, and secondary structural features, affects the continuity and effectiveness of capillary and permeability barriers (Ferrill et al., 2000; Fedors et al., 2001, 2002; Flint et al., 2003).

A critical assumption in conceptual model development for numerical models of unsaturated zone flow in gently dipping, layered, nonwelded ignimbrites is whether or not there is significant variation in intralayer hydrologic properties. Numerical experiments by Ho and Webb (1998) showed that models assuming heterogeneity of hydrologic properties result in constraints on the scale of lateral flow. An improved understanding of the effect of faults on heterogeneity and the potential for lateral flow in layered nonwelded ignimbrites is needed. The intent of this field study was to evaluate the potential effects of small faults and fault-zone deformation on the heterogeneity of weakly consolidated, nonwelded ignimbrites.

	Objectives

TOP EXECUTIVE SUMMARY ABSTRACT INTRODUCTION Background Objectives LOCATION AND DESCRIPTION OF... STRUCTURAL AND HYDROGEOLOGICAL... RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Large-scale unsaturated zone flow models with homogeneous layers may predict lateral diversion of percolating water atop layer interfaces that act as either permeability or capillary barriers to downward flow. Knowledge of small-scale heterogeneity induced by high-angle faults and fractures gives rise to a question of the degree to which geologic units are laterally continuous, and thus effective as flow barriers. To examine this question, field and laboratory results from a fault in the Bishop Tuff, a natural analog site for the Paintbrush nonwelded hydrogeologic unit at Yucca Mountain, are presented, including:

• analysis of lithology and deformational features, which provide the geological framework for evaluating the architecture of intrinsic permeability in nonwelded and deformed ignimbrite
  
• characterization of the nature of intrinsic permeability variation in terms of the effects of secondary discontinuities (i.e., faults and fractures)
  
• demonstration that faults and fractures produce increased hydrologic heterogeneity with respect to the host rock
  

	LOCATION AND DESCRIPTION OF RESEARCH AREA

TOP EXECUTIVE SUMMARY ABSTRACT INTRODUCTION Background Objectives LOCATION AND DESCRIPTION OF... STRUCTURAL AND HYDROGEOLOGICAL... RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Similarities between the nonwelded units of the Bishop Tuff (Bishop, CA) and the Paintbrush Group (Yucca Mountain, NV) enabled this analog study (Fig. 1 ). While subunits of the Bishop Tuff are the focus of the study, the Paintbrush nonwelded hydrogeologic unit is discussed in enough detail (e.g., Moyer et al., 1996) to demonstrate the analogy.

View larger version (46K): [in this window] [in a new window]   Fig. 1. Stratigraphic comparison between the Paintbrush nonwelded hydrogeologic unit and the Bishop Tuff.

View larger version (141K): [in this window] [in a new window]   Fig. 2. Topography and geographic location of the research area near Bishop, California, in relation to that of Yucca Mountain, Nevada. Painted, shaded-relief map generated from USGS 7.5' quad DEMs, UTM coordinates, Zone 11. Resolution is 30 m, NAD 27 datum.

View larger version (118K): [in this window] [in a new window]   Fig. 3. Aerial photograph of the Chalk Cove fault exposure at the erosional southern margin of the Volcanic Tableland; view to the north.

View larger version (81K): [in this window] [in a new window]   Fig. 4. Chalk Cove fault: (a) photograph annotated to show the fault trace, the locations of Transects CCHW1 and CC2/CC3, and lithologic subunits; view to the north-northeast; (b) photograph illustrating wedge-shaped system of fractures in the hanging wall; (c) schematic illustration of the wedge-shaped architecture of the fault deformation zone.

Because the Paintbrush nonwelded hydrogeologic unit is generally inaccessible and some subunits are poorly exposed in outcrop, exposures of the Bishop Tuff were identified as easily accessible natural analog sites for field investigation of unsaturated zone properties and flow processes (Fedors et al., 2001, 2002). The Bishop Tuff represents a single eruptive event, in contrast to the multiple eruptive events observed within the approximately 12-Ma Paintbrush nonwelded hydrogeologic unit at Yucca Mountain (Fig. 1). Based on field observations of bedding, composition, grain size, nature of particle cohesion, and degree of welding, the Bishop Tuff has been informally divided into four mappable hydrostratigraphic units (Fig. 1; Fedors et al., 2001; Evans and Bradbury, 2004). The Paintbrush nonwelded hydrogeologic unit is similar to the Bishop Tuff in the following ways:

• Both formations are composed of massive, uniform rhyolitic ignimbrites and bedded tephra-fall deposits, although the Paintbrush nonwelded hydrogeologic unit has a number of interlayered flow-fall sequences whereas the Bishop Tuff has one cycle of tephra fall deposits underlying a nonwelded ignimbrite that grades to a welded ignimbrite. The difference in the number of sequences, and thus layering, is not important for assessment of lateral intralayer heterogeneity.
  
• Both formations have tectonic deformation histories characterized by extension, which were accommodated by normal faulting (Martel, 1989; Spengler et al., 1993; Dawers and Anders, 1995; Pinter, 1995; Ferrill et al., 1999, 2000; Ferrill and Morris, 2001; McGinnis et al., 2005)
  
• Comparable subunits of the nonwelded Bishop and Paintbrush tuffs have comparable centimeter-scale permeabilities (Table 1). Permeability values for the Paintbrush nonwelded hydrogeologic unit were developed from well cores and are likely biased toward low values due to limited core recovery of poorly consolidated subunits. Thus, the Yucca Mountain Tuff permeability value in Table 1 likely reflects the more welded portion of the unit.
  
• Nonwelded ignimbrite units of the Bishop and Paintbrush tuffs have comparable porosities and bulk densities (Table 1).
  

View this table: [in this window] [in a new window]   Table 1. Comparison of Paintbrush nonwelded hydrogeologic unit and Bishop Tuff nonwelded properties: intrinsic permeability, k; porosity, ; and bulk density, .

	STRUCTURAL AND HYDROGEOLOGICAL CHARACTERIZATION OF FAULTED NONWELDED TO SINTERED BISHOP TUFF

TOP EXECUTIVE SUMMARY ABSTRACT INTRODUCTION Background Objectives LOCATION AND DESCRIPTION OF... STRUCTURAL AND HYDROGEOLOGICAL... RESULTS DISCUSSION CONCLUSIONS REFERENCES

Detailed Fault Zone Mapping
Mapping focused on characterization of the geometry of the Chalk Cove fault zone and documentation of related deformational features that may influence the permeability architecture of the fault zone. Subunits and the visible surface trace of the fault were delineated (Fig. 4a). Mapping was completed both parallel and perpendicular to the fault core and associated damage zones. Detailed tape and compass mapping of fractures was performed along two approximately 6-m transects, CCHW1 and CC2/CC3 (Fig. 5 ), oriented perpendicular to the main Chalk Cove fault surface (Fig. 6 and 7) . The distance between the two mapped transects is 17 m along the fault trace. The nature of deformational features was investigated, and their distribution was mapped relative to the location of gas permeability measurements. Between Transects CCHW1 and CC2/CC3, the exposed length of the Chalk Cove fault zone and surrounding fracture system was mapped at a 1:40 scale. Orientations of fault and fracture surfaces follow the right-hand rule, where the dip direction is clockwise from the reported strike azimuth.

View larger version (22K): [in this window] [in a new window]   Fig. 5. Geologic map of Chalk Cove fault study site in map view, illustrating main fault trace and hanging-wall fracture systems. Detailed fracture mapping areas of Fig. 6 and 7 are shaded in gray. Transect CCHW1 is downsection and Transect CC2/CC3 is upsection (see Fig. 4). Red shading and traces represent fault core gouge, small slip surfaces, or iron-oxide stained zones. Sample locations are identified.

View larger version (41K): [in this window] [in a new window]   Fig. 6. Fracture pavement map of Transect CCHW1, located at x = 0 m in Fig. 5. Sample locations for gas permeability studies, liquid permeability studies, and laboratory samples are identified.

View larger version (38K): [in this window] [in a new window]   Fig. 7. Fracture pavement map of Transect CC2/CC3, located at x = 17 m in Fig. 5; mapping extends to x = 22 m. Sample locations for gas permeability studies, liquid permeability studies, and laboratory samples are identified.

Fracture Intensity Surveys
Fracture intensity data were obtained directly from detailed pavement maps of both the lower (Transect CCHW1) and upper (Transect CC2/CC3) transects (Fig. 4, 6, and 7) for comparison with permeability measurements along both transects. The number of fracture traces intersecting the transect lines was divided by the total distance of the transect line to yield an average fracture intensity value for each pavement map. To better understand the fracture distribution associated with the Chalk Cove fault, fracture intensity data also were collected at two vertical exposures at x = 5 m and x = 22 m along the fault trace (Fig. 4 and 5).

Grain Size, Microstructural, and Compositional Analyses
The grain size of nonwelded, matrix-supported ignimbrite in the vicinity of the Chalk Cove fault is highly heterogeneous at the centimeter scale and is poorly to very poorly sorted. Because of large spatial variability in median grain size, which reflects the presence of pumice and lithic fragments, grain-size reduction caused by fault zone deformation is difficult to assess except in the fault gouge where fines dominate. To better understand enrichment of fines at the Chalk Cove site, the individual weight percentage of samples from the central fault core, mixed zones, and host rock was calculated.

Samples for microstructural analyses were extracted from fresh surfaces excavated at the outcrop. These unconsolidated samples required impregnation with epoxy before thin-section preparation. To estimate macro- and microporosity, blue dye was added to the epoxy. Lithologic composition variations and microstructures were examined using a transmitted light petrographic microscope. Descriptions from Schulz and Evans (1998), Snoke et al. (1998), Blenkinsop (2000), and Mitra and Ismat (2001) were used to identify microstructures and deformation mechanisms.

X-ray diffraction (XRD) analyses using standard methods were performed to assess spatial variations in composition within the fault damage zone.

Small-Drillhole Minipermeameter Survey
Small-scale permeability measurements have traditionally been made by inducing one-dimensional gas flow through a cylindrical core plug (Hassler, 1944). More recently, such laboratory measurements have been made by inducing multidimensional gas flow through a rock sample with various configurations of the conventional surface-sealing gas minipermeameter probe (American Petroleum Institute, 1998). The lack of cohesion of friable nonwelded ignimbrite leads to difficulties during sample collection, making standard sampling methods impractical. The small-drillhole minipermeameter probe (Molz et al., 2002; Dinwiddie et al., 2003; Dinwiddie, 2005) was developed for in situ application inside a small hole drilled into an outcrop to circumvent such sampling difficulties.

The small-drillhole minipermeameter probe (Fig. 8a ) is designed for use in holes with a maximum depth of 10 cm. The probe is inserted into a previously drilled and vacuumed hole (Fig. 8b) until the faceplate contacts the conical end of the drill hole. An annular rubber tip seal undergoes axial compression, causing the seal to radially expand like a packer against the sides of the drill hole. The packer is used to seal the probe to the distal end of the drill hole, while isolating the injection zone through which pressurized nitrogen gas is introduced to the porous medium. Pressure within the sealed-off region is maintained above atmospheric, so that nitrogen gas enters the porous medium, flows around the tip seal, and exits to the rock surface at ambient pressure. After steady-state conditions are achieved, several pressure and flow rate pairs are recorded. Given an assumption of homogeneity within the averaging volume, the measured injection pressure, flow rate, and a numerically determined geometrical factor describing the flow system (Dinwiddie et al., 2003), the effective gas permeability of the porous medium surrounding the drill hole (Molz et al., 2003) is calculated with the standard semianalytical inverse solution (Dinwiddie et al., 2003). Pressure and flow rate data pairs from each location are analyzed for the presence of high-velocity flow effects, and corrections are made if warranted. In the arid, windy outcrop environment of the Chalk Cove fault exposure, water saturation is naturally low; thus, the use of effective gas permeability data as a surrogate for intrinsic permeability is thought to be appropriate.

View larger version (127K): [in this window] [in a new window]   Fig. 8. (a) Small-drillhole minipermeameter probe. (b) Drilled test holes for gas permeability measurement at Transect CC2/CC3. (c) Constant-head permeameter near Transect CC2/CC3; view to the south-southwest.

A portable, homogeneous ceramic check source was used to calibrate the small-drillhole minipermeameter system, ensuring data quality by enabling measurement of the apparent permeability of the ceramic check source in the laboratory before travel, in the field following travel but before data collection, and again daily in the field. Additionally, the pressure transducer and three flow meters were each calibrated to NIST-traceable standards at the beginning and end of each field expedition. When necessary, calibration curves were used to correct pressure or flow rate data.

Sampling holes at the Chalk Cove site were drilled inclined 45° downward to the west. As with detailed fault zone mapping, gas permeability data were collected along Transects CCHW1 and CC2/CC3. Gas permeability data along Transect CCHW1 were generally collected on 10- to 20-cm centers (Fig. 6). Gas permeability data associated with the name CC2 were collected during early fieldwork and consist of low spatial resolution data (50-cm centers) (Fig. 7). Gas permeability data associated with the name CC3 were collected during subsequent fieldwork and consist of higher spatial resolution data (10-cm centers). Fault gouge permeability was not measured because the instrument averaging volume (Molz et al., 2003) would have consisted of both fault gouge and adjacent ignimbrite. The permeability of large open fractures was not measured because of instrument range limits.

Constant-Head Permeameter Measurements
A constant-head liquid permeameter (Reynolds and Elrick, 1986) was used to determine field-saturated hydraulic conductivity of the nonwelded ignimbrite adjacent the Chalk Cove fault (Fig. 8c). Shallow boreholes with known dimensions were drilled at varying distances from the fault. The permeameter was used to establish and maintain a constant head of water inside the borehole and to measure the steady-state flow rate of water into the porous medium. At equilibrium in a homogeneous porous media, a bulbous saturated zone develops around each shallow borehole. The shape of this averaging volume is captured by a numerically determined geometrical factor, as documented by Reynolds and Elrick (1986). This factor is used with the diameter and hydraulic head of the borehole and the steady-state outflow rate to calculate field-saturated hydraulic conductivity. Fifteen field-saturated hydraulic conductivity measurements were collected with the liquid permeameter to augment the more spatially detailed gas permeability data set and to provide additional confidence in the gas minipermeametry technique. The resulting field-saturated hydraulic conductivity data (cm s^–1) were converted to effective aqueous permeability (millidarcies) for comparison with gas permeability data. The zone of influence is larger for the constant-head liquid permeameter than for the gas minipermeameter, and thus more likely to be affected by heterogeneities such as large open fractures in close proximity to the borehole. Away from larger fractures, effective liquid and gas permeability data should be comparable because their corresponding relative permeabilities are approximately unity, and thus, they are both thought to be appropriate surrogates for the intrinsic permeability of the rock mass.

	RESULTS

TOP EXECUTIVE SUMMARY ABSTRACT INTRODUCTION Background Objectives LOCATION AND DESCRIPTION OF... STRUCTURAL AND HYDROGEOLOGICAL... RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
The Chalk Cove fault is nearly vertical with a visible surface trace length of approximately 22 m. The Chalk Cove fault offsets lower Bishop Tuff basal pumice-fall deposits in the footwall and a massive deposit of pink to moderate orange, highly porous, glassy, nonwelded to sintered ignimbrite of the Bishop Tuff in both the hanging wall and footwall (Fig. 4a; Wilson and Hildreth, 2003; Evans and Bradbury, 2004). The terminology of Wilson and Hildreth (2003) is followed for welding variations: nonwelded implies a noncohesive deposit that can be disaggregated easily by hand, while sintered implies a cohesive deposit that requires a hammer to fracture, but lacks a eutaxitic texture. Compared with bulk density and porosity of the sintered zone, the bulk density is greater and the porosity is lower within the nonwelded unit (Wilson and Hildreth, 2003; Evans and Bradbury, 2004). Because of fault exposure and accessibility, structural and hydrologic data were collected primarily within the hanging-wall block of the Chalk Cove fault where the material correlates with mechanical and hydrostratigraphic Unit B (Fig. 1 and 4) described by Evans and Bradbury (2004) and with Subunit Ig2Ea of Wilson and Hildreth (2003).

Detailed Fault Zone Mapping
Mapping indicates the Chalk Cove fault has a minimum observed displacement of 8.6 m (Evans and Bradbury, 2004). The Chalk Cove fault is characterized by a funnel-shaped geometry in profile (Fig. 4; Evans and Bradbury, 2004), which appears as a horse-tail splay geometry in map view (Fig. 5). The fault damage zone is strongly asymmetric with more deformation in the hanging wall than in the footwall (Fig. 5). The fault contact downsection consists of a 2-mm-thick discrete mineralized slip surface with slickenlines and a relatively narrow damage zone. The fault damage zone widens continuously upsection and is mappable into a distinctive ledge-forming horizon (equivalent to the cryptic-ledge of Wilson and Hildreth, 2003) that consists of a thin zone of partially sintered and devitrified ignimbrite. This ledge forms a transitional unit between the top of Unit B or Subunit IG2Ea and the base of Unit C or Subunit IG2Eb (Wilson and Hildreth, 1997, 2003; Fig. 1 and 4). Further upsection, near the surface of the Volcanic Tableland (Fig. 3 and 4a), the fault trace is obscured by large tilted blocks and talus from the overlying welded ignimbrite (Unit D or IG2Eb welded, Fig. 1) cover the area. The location of detailed fracture mapping areas associated with Transects CCHW1 and CC2/CC3 are shown in Fig. 5, and fracture maps for each transect are presented in Fig. 6 and 7. Some fractures and anastamosing slip surfaces exhibit iron-oxide staining at the CCHW1 stratigraphic level (similar to features described by Evans and Bradbury, 2004). In an exposed outcrop directly above the detailed mapping area of Transect CCHW1 (Fig. 4b), slip along the fault is transferred into a series of smaller opening-mode splay fractures and cataclastic slip or deformation.

The Chalk Cove fault may be divided into mappable architectural and hydrogeological elements including a hanging-wall damage zone, hanging-wall and footwall mixed zones, and a central core zone (Fig. 9a through 9c; Caine et al., 1996; Heynekamp et al., 1999; Rawling et al., 2001; Evans and Bradbury, 2004). Upsection, the main fault strand connects with smaller fault splays and stepover faults (Fig. 5). In the upper exposure near Transect CC2/CC3 (Fig. 7) the fault zone widens to a 1-m-thick zone (Fig. 9c) including two major slip surfaces bounding a 5-cm-thick foliated silty to clayey ash gouge (Fig. 9d). Alternating pods of comminuted grains and less deformed material occur between both the hanging-wall and footwall bounding slip surfaces, forming mixed zones similar to those described by Heynekamp et al. (1999). Within the mapped area for the permeability surveys, the hanging-wall damage zone west of the fault contact ranges in thickness between 3.5 and 5 m. The footwall damage zone east of the fault contact is much thinner, ranging in thickness between 0.05 to 1 m. Little deformation is observed outside the boundaries of these damage zones.

View larger version (144K): [in this window] [in a new window]   Fig. 9. Outcrop photos from the Chalk Cove fault site: (a) 2-mm-thick fault contact near station 0 along Transect CCHW1 (Fig. 6); (b) sample locations and fault zone architecture at x = 9 m (Fig. 5); (c) fault zone architecture at x = 22 m (Fig. 5 and 7); view to south; (d) 5-cm-thick central fault core gouge in CC2/CC3 map area (Fig. 5 and 7). F = central fault core or contact; HW = hanging-wall block; FW = footwall block.

A series of moderately to steeply dipping hairline fractures (<1 m long) are oriented parallel and oblique to the main system. These smaller fractures are heterogeneously clustered and often form zones of interconnected anastomosing strands near fracture terminations or between major fractures. Abundant iron-oxide staining is associated with this system.

Fracture Intensity Surveys
Fracture mapping along Transect CCHW1 yields minimum fracture intensities equal to 7 fractures m^–1 for long trace-length fractures (>20 m) in the hanging-wall damage zone (Fig. 5 and 6). A single fracture is mapped along the second half of Transect CCHW1 at y = 4.35 m (see Fig. 6 inset), and host rock is inferred to be present at greater distances from the fault. Relatively little deformation is observed in the footwall compared with the hanging wall near Transect CCHW1 (Fig. 6). For example, fracture intensities measured within the footwall at x = 5 m along the fault trace (Fig. 5) are <1 fracture m^–1. Along Transect CC2/CC3 (Fig. 7), however, minimum fracture intensities increase to 20 to 40 fractures m^–1 in the hanging-wall damage zone and less than 20 fractures m^–1 in the footwall damage zone. The upsection increase in fracture intensity is associated with the presence of numerous short trace-length fractures intersecting the transect line both parallel and oblique to the main fracture system. These fracture intensities represent a minimum value because although numerous short trace-length (centimeter scale) hairline fractures exist, they either do not intersect the transect line or are below the scale of map resolution. Several of these fractures occur within zones of intense iron-oxide staining and may reach fracture intensity values on the order of 50 fractures m^–1. Such zones are present between (i) the main fault and 4 m into the hanging wall at 10 x 12 m along the fault trace (Fig. 5); (ii) 0.8 to 1.1 m, and 2.0 to 2.7 m within the CCHW1 map area (Fig. 6); and (iii) 1.6 to 2.0 m, and 3.0 to 4.0 m within the CC2/CC3 map area (Fig. 7).

Grain Size, Microstructural, and Compositional Analyses
The individual weight percentage of fines that were smaller than 0.125 mm (4) are provided in Table 2 for a suite of samples collected near x = 9 m along the fault trace (Fig. 5). These data show an increase in fines within both the fault zone core and the mixed zones relative to within the surrounding host rock.

Nonwelded ignimbrite host rock near the lower fault exposure consists of large cuspate to platy glass shards, phenocrysts and pumice clasts surrounded by a fine-grained ash matrix with little interconnected pore space. Within the fault damage zone, however, features associated with HW Fracture 3 (Sample BT-99, Fig. 6 and 10a ) include flaking and rotation of quartz clasts, intragranular microfaults (not shown), fractures that follow cleavage planes in feldspar phenocrysts, and open fractures with parallel zones of fine clay-size particles and darker alteration. Microporosity associated with a fine-grained matrix appears interconnected with macroporosity of adjacent open fractures, given the observed disribution of blue epoxy.

View larger version (127K): [in this window] [in a new window]   Fig. 10. Photomicrographs showing: (a) flaking and rotation of quartz clasts, and open fractures that follow grain boundary contacts near HW Fracture 3 (Fig. 6, Sample BT-99); (b) a zone of cataclasis and an opening-mode microfracture in the footwall mixed zone and the fault core contact (Fig. 5, Sample BT-74), viewed in plane polarized light; (c) indistinct micro-scale layering within the internal fault core where deformation is localized in narrow cataclastic bands of aligned angular shards and phenocrysts, with an opening mode microfracture that parallels cataclastic layering (Fig. 5, Sample BT-82); (d) sintering in the hanging-wall zone at approximately 0.5 m west of the Chalk Cove fault (Fig. 5, Sample BT-82); (e) locally decreased grain size (zone of comminution) within the central fault zone core (Fig. 5, Sample BT-85). Deformation within the fault-zone core includes dilational microfractures that parallel bands of comminuted host rock. Note how the extension fracture locally follows the margin of a pumice clast.

Microscopic layering of the internal fault core (BT-82, Fig. 5 and 10c) includes deformation localized within narrow cataclastic bands of aligned angular shards and phenocrysts, and microfractures that parallel the cataclastic bands. Grain size in the cataclastic bands decreases toward the fault core contact. Fine material represents a clay-size pseudomatrix (Rawling and Goodwin, 2003) created by cataclastic deformation processes within the fault zone. Boundaries of fine-grained bands range from gradational and wavy to sharp and planar.

The hanging-wall zone near the upper fault exposure [0.5 m west of the Chalk Cove fault (Sample BT-85, Fig. 5 and 10d)] includes sintered zones with indistinct boundaries between interlocking grain contacts and glassy shards that form a weak zone of annealing immediately around the grain contacts (Mitra and Ismat, 2001; Moon, 1993). Sintering is associated with locally increased brittle deformation.

Thin section and grain size analyses reveal that the 5-cm-thick central fault core (Sample BT-103, Fig. 7 and 10e) exhibits comminuted material and decreased grain size relative to the host rock. Deformation within the fault core includes dilatant microfractures parallel to bands of comminuted host rock. Blue epoxy permeates an open fracture and the surrounding comminuted material. The fracture follows the edge of a pumice clast boundary and forms splays at the tips. Zones of rotated quartz phenocrysts parallel an open fracture, suggesting shear displacement similar to the formation of deformation bands. Deformation intensity and grain sizes are highly variable within the mixed zones surrounding the fault core.

Results from compositional analyses using X-ray diffraction indicate both the fault zone and the surrounding host rock largely consist of the same minerals and glass, including quartz, sanidine, and other feldspars. In the lower fault exposure (Fig. 5 and 6) where the central fault core is only 2 to 10 mm thick, the fault core is composed predominately of quartz and calcite. In the upper exposure (Fig. 5 and 7) the fault core is composed of approximately 5 cm of comminuted material, consisting of quartz and minor glass. There is a slight devitrification in the central fault core relative to the glassy host rock.

Permeability Data
Gas permeability data indicate that the intrinsic permeability of nondeformed, nonwelded host rock at the Chalk Cove site is very homogeneous relative to the intrinsic permeability of matrix blocks in the fault-deformation zone (Fig. 11a , Table 3). Permeability within the damage zone increases by more than one order of magnitude above that of the nondeformed host rock (Fig. 11a and 11b).

View larger version (29K): [in this window] [in a new window]   Fig. 11. Gas and liquid permeability data measured along (a) Transect CCHW1; (b) Transect CC2/CC3. The fault is at y = 2.0 m (6.6 ft) along Transect CC2/CC3.

In situ gas permeability measurement of structurally deformed unconsolidated rocks is not accomplished without difficulties. Gas permeability data are lacking from Transect CCHW1 (Fig. 11a) in poorly cohesive zones where many iron-oxide stained fractures below the scale of map resolution run parallel to the fault (Fig. 6). The pressure of the minipermeameter probe seal against the walls of the drill holes in these locations caused the grains to compress, widening the drill holes and allowing gas leakage to occur. One might infer large intrinsic permeability in these zones. The inability to supply nitrogen at sufficiently high rates also precludes measurements of gas permeability when large-scale open fractures intersect the measurement volume; therefore, gas permeability data reflecting the effect of large open fractures are not presented.

Transect CC2/CC3
The entire CC2/CC3 map area (Fig. 7) is within the fault-damage zones because of the wedge-shaped architecture of the fault (Fig. 4 and 5) and the nature of the available exposure surface. Gas permeability data from the matrix blocks along Transect CC2/CC3 are elevated relative to gas permeability data measured within the damage zone along Transect CCHW1 (Fig. 11a and 11b). Gas permeability data are lacking from the CC2 data set (Fig. 11b) in a poorly cohesive zone consisting of many iron-oxide stained fractures below the scale of map resolution (Fig. 7). As a result, corresponding CC3 data were collected about 1 m (0.3 ft) to the north (Fig. 7) to achieve leak-proof seals. The detailed heterogeneity evident in the CC3 data set, but lacking from the CC2 data set, confirms the need for high-resolution sampling when investigating fault-related secondary heterogeneities. Liquid permeability and gas permeability data along this transect are in good agreement (Fig. 11b).

The solid circle plotted at y = 2.5 m along the CC2 transect represents some uncertainty in the measurements taken at this location as a result of an inoperable A-to-D converter for the 0 to 2000 cm³ min^–1 flow meter during the field campaign (Fig. 11b). The two flow rates at which permeability was computed for this location were near the maximum measurable flow rate for the intermediate flow meter. Additional data from the flow meter with the highest range would have been helpful but were not available.

Statistical Analysis
Gas permeability data collected along Transect CCHW1 are analyzed as a single distribution, and as two separate distributions with the location of the farthest fracture from the fault at y = 4.35 m serving as the boundary between the fault damage zone and the nondeformed host rock of the hanging wall (Table 3, Fig. 6 inset). Data collected along Transect CC2/CC3 are analyzed as a single distribution, and as two separate distributions with the location of the fault core at y = 2.0 m serving as the boundary between the footwall and hanging-wall data (Table 3, Fig. 7). The statistics in Table 3 represent a minimum level of permeability heterogeneity because of sampling bias related to lack of permeability data for the fault gouge (a low permeability element) and large-scale open fractures (high permeability elements).

Roughly the same number of observations were made in the hanging wall along Transect CCHW1 (n = 40) as were made within the hanging wall of Transect CC2/CC3 (n = 37). Comparing these data sets, permeability upsection in the hanging wall is, on the average, more than a factor of two greater than permeability downsection. When taking into account the best measure of central tendency for each distribution, permeability within the hanging wall is, on the average, greater than permeability within the host rock and within the footwall damage zone. The coefficient of variation of host rock permeability is significantly smaller than that for both the hanging wall and footwall damage zones.

Intrinsic permeability of the ignimbrite matrix is likely influenced by lithology, degree of welding, and small-scale fracture intensity. Although one might hypothesize that small-scale fracturing may be correlated with large-scale mappable fracturing, spatial profiles of gas permeability and calculated fracture density (Fig. 12a and 12b) and a scatterplot of bin-averaged gas permeability versus calculated fracture density (Fig. 12c) illustrate poor correlation. The poor correlation observed in Fig. 12 may arise because of heterogeneities within the matrix and irregular distribution of the small- and large-scale fractures. Small-scale fracture systems are significant within these deposits because they may locally increase permeability, and depending on their orientation with respect to the main fracture system, may increase connectivity within the entire fracture network.

View larger version (25K): [in this window] [in a new window]   Fig. 12. Spatial profiles of fracture density and gas permeability data using the same vertical scales for (a) Transect CCHW1, and (b) Transect CC2/CC3. (c) Scatterplot to assess how well the gas permeability data may be correlated with available fracture intensity data.

	DISCUSSION

TOP EXECUTIVE SUMMARY ABSTRACT INTRODUCTION Background Objectives LOCATION AND DESCRIPTION OF... STRUCTURAL AND HYDROGEOLOGICAL... RESULTS DISCUSSION CONCLUSIONS REFERENCES

Grain-size data suggest an increase in the percentage of fine particles in the fault core relative to that observed in nondeformed ignimbrite. Thin-section observations suggest foliation and slip surfaces may impart anisotropy to the fault gouge permeability; that is, permeability perpendicular to the gouge along the fault surface may be less than permeability in the gouge laterally and vertically along the fault surface. The presence of a fine-grained fault gouge would act as a permeability barrier to lateral flow for all flow conditions and as a conduit for vertical flow during low-flow conditions because a fine-grained fault gouge will tend to wick water by capillary attraction.

Fracture mapping delineates large-scale extension features that could affect lateral flow. Hanging Wall Fractures 1 through 5 are inferred to be relatively large aperture, open fractures. Fracture intensities range up to 50 fractures m^–1, although generally the intensities are highly variable and an average for all damage zones, if calculated, would be much lower. Subvertical fractures would constrain lateral flow by providing a vertical conduit for water in a variably saturated system containing interlayer capillary or permeability barriers. Capillary theory also suggests that large-aperture subvertical fractures would hinder lateral flow because capillary attraction is greater within the rock matrix than within open fractures. Filled subvertical fractures may function as either capillary or permeability barriers against lateral flow depending on the hydrologic properties of the fracture filling relative to the properties of the adjacent rock. Previous work using ponded dye tracer tests in the Bishop Tuff (Fedors et al., 2001, 2002) demonstrated the capability of filled fractures to both constrain lateral flow and promote vertical flow of water through adjacent matrix material.

Deformation features likely cause an increase in heterogeneity of hydrological properties that, for subvertical faults, would be manifest in the lateral direction, or along a stratigraphic unit. Intra-unit heterogeneity is known to reduce or eliminate large-scale lateral flow (e.g., Ho and Webb, 1998) associated with bedded units. In situ gas permeability data from a fault zone was used in this study to characterize heterogeneity in matrix blocks separated by mesoscale tectonic features. Matrix permeability in deformed blocks ranged from high background values to more than an order of magnitude increase, and calculated statistics represent a minimum level of heterogeneity for the site because fault gouge and open fracture permeability was not measured. The observed variation in permeability values is believed to be caused by spatially variable matrix block deformation. Thin-section analysis of microscale features within the matrix blocks identified grain rotation and alignment, pods of comminuted grains and less deformed material, and the development of connected porosity. Thus, the increased magnitude and variability of gas permeability in the deformation zone surrounding the fault are believed to reflect the mixed deformation observed in the field and in thin section.

In addition to fault deformation features observed across (or transverse to) the fault, variation in deformation was also noted along the fault trace. Lithologic variation along the fault trace from nonwelded (Transect CCHW1) to grain-boundary sintering (Transect CC2/CC3) coincides with the broadening of the fault zone. Thin-section and field observations of the Bishop Tuff support Wilson et al. (2003), who suggest that low-porosity welded units deform by transgranular fracture and that high-porosity nonwelded glassy to devitrified units form shear deformation bands or fractures depending on local variations in the degree and nature of crystallization. High-porosity materials favor strain-hardening deformation and the development of connected small-scale fracture systems (Manzocchi et al., 1998).

Sintering, which creates interlocking grain contacts, may locally increase brittle deformation features. The slight sintering present in the study area near Transect CC2/CC3 is thus consistent with development of connected porosity related to small-scale fracturing and faulting. Upsection increases in welding, comparable to that seen in the Yucca Mountain Tuff, corresponds with an increase in the width of the damage zone and the number of laterally continuous large-scale fractures adjacent the fault (Fig. 4b).

In summary, fault-related deformation in adjacent matrix blocks resulted in increased magnitude and variability of matrix permeability, thus demonstrating that deformation features are not only present in nonwelded ignimbrites, but also noticeably increase heterogeneity to levels beyond that of the host rock. Field and laboratory analyses suggest deformation of nonwelded ignimbrite occurred by both extension and shear fracturing.

Unsaturated zone flow models may predict capillary and permeability barriers at gently sloping layer interfaces between nonwelded ignimbrite units at scales of meters to a kilometer, depending on lateral heterogeneity and the nature of model layer contacts. Assessing field evidence in combination with conceptual and numerical models, Flint et al. (2003), Fedors et al. (2002), and Fedors and Ferrill (2002) suggested that lateral flow in the Paintbrush nonwelded hydrogeologic unit of Yucca Mountain physically should be limited to a scale of tens of meters. Large- and small-scale tectonic faulting and fracturing and related affects on heterogeneity were suggested by these authors as mechanisms that may constrain the areal extent of lateral flow at Yucca Mountain.

In stratigraphic section, the Paintbrush nonwelded hydrogeologic unit is composed of alternating massive ignimbrite and tephra-fall interbeds. Some subunits are not laterally continuous, and contacts between subunits range from gradational to sharp unconformities. Deformation features, such as small faults and fractures are present in multiple orientations, and fracture intensity generally decreases from massive ignimbrite to tephra-fall interbeds. Small faults are known to cross all subunits of the Paintbrush nonwelded hydrogeologic unit, as observed from surface exposures on the west flank of the volcanic ridge, tunnel exposures within the Exploratory Studies Facility, and vertical boreholes. From a detailed line survey in the Exploratory Studies Facility (Eatman et al., 1997; Barr et al., 1996), Fedors and Ferrill (2002) estimated an average fault spacing of 3.65 m, and an average fault dip of 70°. These faults are oriented at a high angle to stratigraphic layering. The type of deformation in and adjacent to small faults in the Paintbrush nonwelded hydrogeologic unit has not been investigated in detail that is comparable to that presented here, but descriptions of deformation from small faults (Sweetkind et al., 1996) suggest similarity with that observed in the nonwelded Bishop Tuff.

Although it is likely that localized lateral flow along capillary or permeability barriers is associated with some subunits of