VZJ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (5)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Evans, J. P.
Right arrow Articles by Bradbury, K. K.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Evans, J. P.
Right arrow Articles by Bradbury, K. K.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Evans, J. P.
Right arrow Articles by Bradbury, K. K.
Related Collections
Right arrow Vadose Zone Processes and Chemical Transport
Right arrow Fractured Rock
Published in Vadose Zone Journal 3:602-623 (2004)
© 2004 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA

ORIGINAL RESEARCH

Faulting and Fracturing of Nonwelded Bishop Tuff, Eastern California

Deformation Mechanisms in Very Porous Materials in the Vadose Zone

James P. Evans* and Kelly Keighley Bradbury

Department of Geology, Utah State University, Logan, UT, 84322-4505
* Corresponding author (jpevans{at}cc.usu.edu).

Received 11 April 2003.



   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 Geologic Setting of the...
 METHODS
 RESULTS
 CONCLUSIONS
 APPENDIX
 REFERENCES
 
Field and microstructural studies were conducted on the basal, nonwelded and partially welded portions of the rhyolite Bishop Tuff in eastern California to examine the nature and processes of brittle deformation in these units. The nonwelded tuff consists of a variable sequence of finely laminated to massive pumice-rich deposits, fine-grained ash, and pyroclastic glass erupted from the Long Valley Caldera. The deposits are experiencing east–west extension in the hanging wall of the White Mountain fault, and small-displacement faults and fractures cut the tuff. Deformation in the Bishop Tuff occurred by fracturing associated with faults, and by slip along narrow faults with smooth, often mineralized surfaces. Localization of fracturing appears to be a function of welding. Units with a greater degree of welding have a greater abundance of fractures associated with faults, whereas nonwelded portions typically have a narrow deformation band–type faults with little or no associated damage. Microstructural observations show that transgranular fractures lie along grain boundaries of pumice and feldspar phenocrysts, and these fractures are often filled with calcite. These deposits appear to have behaved as an open-cell foam with a low strength, but with a cohesion that allowed the support of a differential stress to failure that resulted in subvertical open fractures and faults. These results demonstrate how brittle deformation may be manifested in nonwelded deposits in the vadose zone, and impart an anisotropy in which flow would be enhanced vertically and impeded horizontally. The Bishop Tuff is analogous to other nonwelded tuffs in the western USA. Thus, these results have implications for understanding deformation and flow in a variety of arid regions.

Abbreviations: PTn, Paintbrush Tuff • SEM, scanning electron microscopy • XRD, X-ray diffraction


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
Geologic Setting of the...
 METHODS
 RESULTS
 CONCLUSIONS
 APPENDIX
 REFERENCES
 
DEFORMATION OF UNCONSOLIDATED to poorly consolidated deposits in the unsaturated zone has received relatively little attention from the structural geology community. Faults and fractures in such deposits may significantly impact fluid flow in the vadose zone (Winograd, 1971; Sigda and Wilson, 2003). Studies in unconsolidated sediments (Rawling et al., 2001; Cashman and Cashman, 2000) and nonwelded tuffs (Wilson et al., 2003) document the presence of brittle structures in the unsaturated zone. Relatively few rock mechanics–based studies have examined these types of deposits or studied their impact on flow. Thus, relatively little is known about the mechanical processes that are responsible for deformation in unconsolidated deposits seen in the vadose zone.

In addition to unconsolidated sediments, volcanic deposits are common in basins of the arid west and record Quaternary deformation. The flow of fluids in fractured and faulted, nonwelded tuffs has emerged as a key technical issue in the characterization and design of the proposed geologic repository for high-level nuclear waste at Yucca Mountain, Nevada (Flint et al., 2001a, 2001b, 2001c, 2003), and in some basins the presence of faults may affect groundwater flow (Sigda and Wilson, 2003). Owing to the relatively poor exposure, difficulty in accessing nonwelded tuffs, and difficulties in sampling nonwelded tuffs, there are few direct observations of naturally deformed tuffs. Winograd (1971) speculated that flow in the unsaturated zone would be greatly affected by fractures in nonwelded tuffs, and that fracture density would be a function of welding. Wilson et al. (2003) is among the first detailed study of deformation in nonwelded tuff, and they show that deformation-band faults are a common structure, with the degree of fracturing a function of porosity of the tuff.

We examine nonwelded and poorly welded tuffs of the Bishop Tuff, eastern California, to determine the likely structure and composition of faulted and fractured nonwelded tuffs. The faulted and fractured nonwelded and poorly welded ignimbrite and ash-fall deposits of the Bishop Tuff are exposed in a plateau north of Bishop, CA, and extension across the Owens Valley has resulted in a series of normal faults that cut the tuff (Bateman, 1965; Pinter, 1995; Ferrill et al., 1999a). Our work is restricted to a series of highly porous tuffs with little or no devitrification, so we concentrate on the nature of deformation in a more restricted lithology than Wilson et al. (2003), but for a wide range of porosity.

In this paper, we present results of a study of the normal faults as viewed in vertical sections and couple these data with previous work on the rock mechanics and thermal structure of nonwelded tuff to

This paper provides the geologic descriptions of the faults at the outcrop scale, and detailed microstructural observations of the deformed rocks. These field-based studies also provide the geologic framework for in situ hydrologic tests of these deposits (Fedors et al., 2001; Dinwiddie et al., 2002a, 2002b), which examine in some detail the flow properties of undeformed and deformed nonwelded tuff. In addition to its application to the hydrogeology at Yucca Mountain, this work provides insight into fractures and faults in other unconsolidated to poorly consolidated rocks in general (Rawling et al., 2001), processes by which unconsolidated deposits deform in shear (Cashman and Cashman, 2000), and the behavior of such deposits in the unsaturated zone.


   Geologic Setting of the Bishop Tuff in Volcanic Tableland
TOP
 ABSTRACT
 INTRODUCTION
 Geologic Setting of the...
METHODS
 RESULTS
 CONCLUSIONS
 APPENDIX
 REFERENCES
 
The Volcanic Tableland (Fig. 1) is a geomorphically distinct region underlain by the Tableland lobe ignimbrite deposit of the Bishop Tuff (Hildreth, 1979). The Bishop Tuff (Gilbert, 1938; Bateman, 1965; Wilson and Hildreth, 1997) consists of a series of airfall and ashflow tuffs erupted 758.9 ± 1.8 ka (Sarna-Wojcicki et al., 2000) from the Long Valley Caldera, 40 km to the northwest of the present study area (Bailey et al., 1976). The basal sequence of the Bishop Tuff is a non- to poorly welded tuff, which represents basal fall and surge deposits that resulted from the eruption of the caldera (Wilson and Hildreth, 1997, 1998). Above this ignimbrite is a moderately to densely welded ash-flow sequence with an uppermost, densely welded, resistant tuff that forms the distinctive surface of the tablelands. These ignimbrites overlie a sequence of fine-grained lacustrine deposits (Hollett et al., 1991).



View larger version (72K):
[in this window]
[in a new window]
  		Fig. 1. Geology of the study area. (A) Location and regional geology of the Bishop study area, Bishop, CA. Location of the Volcanic Tablelands between the Sierra Nevada to the west and the White Mountains to the east. Normal faulting changes polarity from the east-dipping Owens Valley fault to the west-dipping White Mountain fault. Qbt, Quaternary Bishop Tuff; Kg, Cretaceous granitoids of the Sierra Nevada; Pzu, undifferentiated Paleozoic sedimentary rocks of the White Mountains; WMF, White Mountain Fault. The Long Valley caldera is the source of the Bishop Tuff. (B) Traces of the major normal faults in the study area, along with the locations of the study sites, as mapped from air photos and the work of Ferrill et al. (1999).

 
The Volcanic Tableland exhibits a classic sequence of north-striking normal faults (Bateman, 1965) for which map-view and displacement-length relationships have been described in some detail (Dawers and Anders, 1995; Pinter, 1995; Ferrill et al., 1999a; Ferrill and Morris, 2001). The normal faults are the result of Quaternary extension across the northern end of the Owens River Valley, and the Tableland faults may represent 2 to 5% extension at the level of the Bishop Tuff. Cross-sectional views through some of these faults and their associated fractures are afforded at the southern end of the Tableland in an area called the Chalk Bluffs and are the focus of the work reported here (Fig. 1). The normal faults of the Tableland are due to active normal faulting in the Owens Valley. Studies of the 1986 Chalfant earthquakes (Cockerham and Corbett, 1987; Lienkaemper et al., 1987; Smith and Priestley, 2000) and subsurface geologic models (D. Ferrill and A. Morris, personal communication, 2002) show that the structures of the region are complex. The Volcanic Tableland faults accommodate strain in the hanging wall of the White Mountain fault, which bounds the Owens Valley on the eastern side. The Tableland lies in a stepover zone in an oblique slip setting where, south of the Tableland, the east-dipping Owens Valley fault has both right-lateral and normal slip (Martel, 1989). East and north of the Tableland, the White Mountain fault zone dips west and exhibits normal and right-lateral slip (dePolo et al., 1993; Smith and Priestley, 2000; Kirby et al., 2002). Smith and Priestley (2002) suggested that the Volcanic Tableland is a region of right-later slip, transferred from the White Mountain fault to the southeast.


   METHODS
TOP
 ABSTRACT
 INTRODUCTION
 Geologic Setting of the...
 METHODS
RESULTS
 CONCLUSIONS
 APPENDIX
 REFERENCES
 
We use a variety of geologic approaches to describe and interpret the faulted and fractured units of the Bishop Tuff and the structures within the rocks. To establish the map-scale structure of the fault zones at the southern end of the Tableland, we used air photo analysis of 1:20000 scale black and white air photos and geologic mapping traverses at a scale of 1:24000.

Detailed outcrop-scale investigations were performed at four sites where faults and/or fracture systems are well exposed (Fig. 1). At these sites, stratigraphic sections were measured and described, fault zone geometries were mapped at scales ranging from 1.5 m:1 cm to 1 m:5 cm, and fracture characteristics were inventoried along linear scanlines. Detailed samples for thin section and any laboratory analyses were collected at most of these sites. We used standard optical petrography techniques and scanning electron microscopy (SEM) to determine deformation mechanisms and examine the nature and history of fluid flow of the fault related rocks and fractured rocks. X-ray diffraction (XRD) studies were conducted on samples from one of the sites. The SEM work was performed at the University of Oregon Micro Analytical Facility on a JEOL JSM-6300VX scanning electron microscope. The microscope has a Link eXL energy dispersive X-ray detector for elemental analyses, and a back-scattered electron detector. Most of our observations were made in back-scattered mode at 10 to 20 kV and a working current of 20 µA. The XRD analyses were performed on standard dry glass-mounted powdered samples at 2° step intervals from 2 to 59°. We used a modified Jolly balance method to determine porosity on our samples (and calculated from other reported values). Aspects of sampling and porosity determinations unique to the collection of these nonwelded units are outlined in the Appendix.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 Geologic Setting of the...
 METHODS
 RESULTS
CONCLUSIONS
 APPENDIX
 REFERENCES
 
Stratigraphy
We report on the stratigraphy of the sites we examined by presenting our own stratigraphic section (Fig. 2) in conjunction with the work of Wilson and Hildreth (1997). Remarkably, despite its importance in understanding Plinian eruptive processes, the analysis by Wilson and Hildreth (1997) appears to be among the first detailed stratigraphic analyses of the tuffs in the Tableland region (Ewert and Harpel, 2000). The stratigraphy and lithologies described here are based on our observations and the work of Bateman (1965) and Wilson and Hildreth (1997). The total thickness of the Bishop Tuff in the western edge of the study area is 142 m and at the eastern edge of the study area the Bishop Tuff is 70 m. Our stratigraphic sections emphasize field observations of the hydrologic and mechanical properties of the deposits, and the textures observed in thin section.



View larger version (31K):
[in this window]
[in a new window]
  		Fig. 2. Stratigraphic section of the Bishop Tuff in the study area.

 
We subdivide the Bishop Tuff into four major lithologic, and presumably, mechanical and hydrostratigraphic units on the basis of composition, bedding characteristics, and degree of welding (Bateman, 1965; Hildreth, 1979; Wilson and Hildreth, 1997, 1998, 2003) (Units A–D, Fig. 2 and 3a) . These are the lowermost fall deposits and are primarily pumice-rich fallout tephra (Gilbert 1938; Bateman, 1965; Wilson and Hildreth, 1997); an intermediate, slightly to partially welded ignimbrite deposit, equivalent to unit Ig2Ea of Wilson and Hildreth (1997); an ignimbrite roughly equivalent to Ig2Eb (Wilson and Hildreth, 1998), which consists of a lower non- to poorly welded ashflow with an upper zone of slightly welded ash that in places is sintered; and an uppermost densely welded ash-flow tuff that forms the caprock of the Tableland. The Bishop Tuff had been interpreted as a classic deposit representing basal air-fall deposits, overlain by a surge deposit, and capped by an ash-flow tuff (Sheridan, 1970), but Wilson and Hildreth (1997)( 2003) indicated that complex interlayering of the fall and surge deposits exist across the deposit. Approximately 15 km north of the study area fossil fumarole vents lie in the Bishop Tuff (Gilbert, 1938; Bateman, 1965; Sheridan, 1970) and represent hydrothermal alteration due to heated meteoric water rising through the tuff (Holt and Taylor, 1998). These cylindrical structures and associated vesiculated textures are not observed in our study area, and as we will show later, little hydrothermal water–rock interaction appears to have affected the protolith.



View larger version (81K):
[in this window]
[in a new window]
  		Fig. 3. Outcrop photographs of undeformed Bishop Tuff. (A) Entire sequence from near Chalk Cove site. Lower white sequence is the nonwelded airfall deposits, overlain by Units B–D. (B) Close-up photographs of the basal thin-bedded airfall deposits. (C) Massive nonwelded sequence. (D) Upper portion of Ig2Ea and Ig2eb of Wilson and Hildreth (1997). Note variation in fracture density upsection, which corresponds to variations in welding.

 
Air-Fall Deposits (Unit A)
Across the entire region, a series of ash and pumice-rich fall deposits are exposed in cliffs, road cuts, and riverbank exposures. These comprise the basal fall deposits of the Bishop Tuff that are 5 to 20 m thick across our study area. These deposits have been subdivided into nine subunits by Wilson and Hildreth (1997); we generalize these subunits into three sequences.

The basal part of the sequence consists of alternating thin- to medium-bedded greenish-gray angular to subangular, pumice-rich layers that are moderately sorted and normally graded near the upper contact (Fig. 3b). Thin sections show that these deposits consist of 200 to 2000 µm vesiculated glass fragments and pumice embedded in a <100-µm grain size matrix of glass and ash fragments. This unit was emplaced as a normal-fall deposit (Wilson and Hildreth, 1997, 1998).

The middle portion of the basal, nonwelded, internal Bishop Tuff (Fig. 3c) consists of beds ranging from 15 to 45 cm thick with an overall sequence thickness of about 1.6 m. Bedding is thin to massive, with white, tan, and gray sequences of pumice-rich deposits, where pumice fragments may reach 30 cm in length. Pumices are angular to subrounded, moderately sorted, and for the most part exhibit normal fining upwards character. These deposits are high porosity, clast-supported and pumice rich.

The upper portion of Unit A consists of 40- to 70-cm-thick beds of planar and minor cross-bedded lapilli to block-sized pumice fragments, with some lithic fragments at the topmost part. In some localities these deposits contain 10 to 25° cross laminae that are 0.5 to 15 cm thick, defined by pumice-rich horizons. In some cases these units have climbing ripples, scours, and inverse grading (Fig. 4b) . These deposits are interpreted as being the result of a hybrid emplacement mechanism in which surge and flow mechanisms occurred together during deposition (Wilson and Hildreth, 1998).



View larger version (35K):
[in this window]
[in a new window]
  		Fig. 4. (A) Sketch map of the Borrow Pit site and fracture characterization zones. (B) Fracture pavement map illustrates spacing and trace style of fractures within the poorly to slightly welded Bishop Tuff. Persistent fracture network mapped in pavement area have a horsetail-splay geometry to the south-southwest and begin to die out west of Pit 1.

 
Porosities of the basal nonwelded sequence are 38 to 60% with densities of 1.1 to 1.5 g cm–3, with most of the samples between 45 and 57% (our data; Sheridan, 1968; Ragan and Sheridan, 1972).

Middle Ignimbrite Deposit (Unit B)
This sequence was mapped as Ig2Ea by Wilson and Hildreth (1997) and termed the Tableland sequence by Hildreth (1979). The ignimbrite deposits are relatively massive and contain crystal-rich pumices with two pyroxenes (Wilson and Hildreth, 1997). The lower portion, Ig2Ea, is approximately 30 to 40 m thick at the western end of the exposures of the Tableland and <10 m thick at the eastern end. This unit varies from slightly welded at its base to moderately welded, glassy, and nearly sintered in some places. It consists of a pumice, ash, and lithic-rich sequence. Intense vapor-phase alteration is present in places (Wilson and Hildreth, 1997). A distinct ledge is locally present at the top of Ig2Ea (Fig. 3a) and forms a distinctive marker used to determine fault offsets at the Chalk Bluffs site. Porosity of this sequence decreases from 45% at the base to about 20% near the upper part of the unit (Bateman, 1965; Sheridan, 1968; Ragan and Sheridan, 1972), and densities range from 1.3 to 1.55 g cm–3 (Wilson and Hildreth, 2003).

Upper Ignimbrite Deposit (Unit C)
This deposit is equivalent to the upper quarter of Ig2Eb of Wilson and Hildreth (1997) and consists of a 40- to 80-m-thick sequence that contains a large component of rhyolite fragments in the lithic portion of the deposit. The unit is nonwelded to poorly welded and becomes densely welded over a vertical section of 20 to 30 m from its base. Porosity decreases from about 20% to below 10% (Gilbert, 1938; Bateman, 1965; Sheridan, 1968), and densities are 1.2 to 1.5 g cm–3 (Wilson and Hildreth, 2003).

Upper Cooling Unit (Unit D)
This is the moderately to densely welded cap to the Tableland (Fig. 3) and is equivalent to the upper cooling unit of Sheridan (1968). The sequence is a fine-grained vitric tuff with lapilli to block-sized pumice fragments. Euhedral crystals of quartz, sanidine, Na-plagioclase ± biotite ± lithic fragments lie in the glassy matrix. The porosity of this unit is 10 to 15% (Bateman, 1965; Sheridan, 1968), with rock densities of 1.2 to 1.7 g cm–3 (Wilson and Hildreth, 2003).

Outcrop-Scale Analysis of Faults
We describe fault and fracture zones at four sites that range in throw from 0 to 18 m. The faults cut all four units described above, but we focus on structures where faults cut Units A through C, the massive to well bedded nonwelded to moderately welded tuff. We describe the structures from the least to greatest amount of fault throw based on outcrop-scale observations. We then discuss the representative microstructures, from the various sites, which document the mechanisms of brittle fracture and faulting in ignimbrites and how these results relate to other types of unconsolidated deposits in the unsaturated zone.

Borrow Pit: 0 m Slip
The Borrow Pit site (Fig. 1 and 4) is an excavation into the middle part of Unit B. Here, the rocks are a white to light pink, non- to slightly welded massive ash and pumice-bearing tephra. The excavation has a floor that slopes 4 to 7° west, and covers an area of about 3000 m2. The exposure of the fractures is along the northern and southern 3-m-high vertical walls, and on the floor in the northeast part of the exposure (Fig. 4). The locality is also the site of in situ dye tracer and falling-head permeametry hydrologic tests (Fedors et al., 2001). The excavation lies in the footwall of two normal faults. The trace of a steeply east-dipping normal fault lies approximately 10 m east of the excavation, and these fractures may be the result of bending of the footwall block. The deposits are cut by a small-displacement vertical fault on the western edge of the exposure, and a larger west-dipping normal fault lies about 40 m west of the pit. Thus, this area may represent a damage zone related to these two faults.

Deformation here is marked by closely to widely spaced fractures with centimeter-scale faulting at the west edge of the site, near Pit 3 (Fig. 4). On the eastern edge of the exposure, fracture densities are 10 to 15 fractures per meter, with fracture trace lengths of 10 cm to >2 m long in vertical profile (Fig. 5b) , and 1 to 2 m long in plan view (Fig. 4). The fractures are planar to subplanar, and typically exhibit a smooth millimeter-thick calcite coating. In map view, fractures form Y-type junctions, in which fractures join adjacent fractures at low angles. The eastern part of the southern vertical face (Fig. 5) lies along strike of the pavement map (Fig. 5c). Thus these closely spaced fractures impart a high degree of vertical fracture anisotropy to the rock near the fault zone. This region might be regarded as a broad damage zone (Caine et al., 1996), which can have important implications for mesoscopic hydrologic properties of the tuff (Fedors et al., 2001). Fracture spacing decreases westward, away from the proximal normal fault, to a value of one to three fractures per meter, as measured on scan lines (Fig. 6b) .



View larger version (80K):
[in this window]
[in a new window]
  		Fig. 5. Details of the Borrow Pit. (A) Vertical outcrop face of the southern Borrow Pit site and (B) corresponding fracture trace maps underlying test pit sites of dye and infiltration tests at the Borrow Pit site (Fedors et al., 2001). Persistent fracture network lies within the immediate hanging wall of two macroscopic faults. Two major high-angle sets are characterized by thin discrete surfaces filled with calcite or wider distributed zones of hairline fractures stained with iron oxides. Slickenlines were observed on several fracture surfaces. Hydrologic tracer tests (Pits 1 and 2) by Fedors et al. (2001) suggest fractures enhance matrix flow vertically by constraining lateral flow. (C) North-northeast view of the floor of the Borrow Pit (shown in Fig. 4b) and its vertical faces. (D) Vertical fault at the northwest corner of the Borrow Pit site. Smooth mineralized fault surface consists of calcite and has nearly pure dip-slip slickenlines.

 


View larger version (53K):
[in this window]
[in a new window]
  		Fig. 6. Photomosaic and projected vertical fault map for the Crucifix site. (A) Exposure of west-dipping normal fault that offsets thinly bedded, highly porous and nonwelded, pumice-rich hybrid to streaky fallout tephra deposits (Wilson and Hildreth, 1998) interbedded with minor thin ignimbrite sequences pf the Crucifix site. Conjugate normal faults and numerous centimeter-scale displacement faults occur within the footwall damage zone. Ferrill et al. (2000) suggested these faults formed by sequential rather than simultaneous slip. (B) We estimated a minimum offset of 8 m from measurements along the Crucifix fault. The hanging wall damage zone extends about 5 m to the west, whereas the footwall damage zone extends about 10 m to the east. Fault core is heterogeneous and is filled with calcite and/or coated with clay, or consists of a comminuted gouge material with abundant pumice clasts (Fig. 8). Vertical fractures are closely spaced in the hanging wall and typically get deflected as they intersect less-welded and/or coarser pumice-rich layers. Subsidiary fractures are open or have millimeter to 1-cm-thick calcite and silica coatings. Numbers indicate sample locations for X-ray diffraction and thin-section analyses.

 


View larger version (30K):
[in this window]
[in a new window]
  		Fig. 8. X-ray diffraction patterns for undeformed and deformed samples from the Crucifix site. (A, B) samples from the thick pumice-rich unit (see Fig. 7b for location) show marked contrast in the amount of noncrystalline (glass) material in the primary deposit vs. a quartz–calcite rich fault zone. (C, D) Analysis of a thinner ash-rich sequence shows the presence of feldspar phenocrysts and glass, which in the fault zone becomes a more ordered quartz–feldspar cataclasite.

 
Along the western edge of the northern vertical exposure, a small fault with approximately 40 cm of normal slip is exposed (Fig. 6d). The slickenline on the fault surface is developed on the calcite coating, which is smooth and 2 to 6 mm thick. This suggests the mineralized surface formed before faulting.

Crucifix Site: 7 m Throw
A series of faults with 0.5 to 7 m of throw are exposed along a length of approximately 70 m along the Chalk Bluffs road at the Crucifix site (Fig. 1). The site described here is about 0.3 km east of our stratigraphic section, and the faults cut nonwelded, finely laminated fall deposits interbedded with slightly welded ignimbrite beds 20 to 60 cm thick and a massive bed at the top of the basal Unit A. Wilson and Hildreth (1998) identified these deposits as unique hybrid to normal fallout tephras related to local wind currents during eruption events. The deposits share characteristics similar to surge and flow deposits, including textures typically associated with reworked basal surge deposits (Bateman, 1965; Wilson and Hildreth, 1997).

The main Crucifix fault is a steeply dipping planar normal fault with a 7-m offset and a very narrow fault core. The fault zone contains a moderately deformed hanging wall and an intensely deformed footwall damage zone as inferred from fracture persistence (Fig. 6). The immediate hanging-wall damage zone extends for about 5.5 m. The footwall damage zone is about 6 m thick measured horizontally from the fault surface and contains numerous centimeter-scale antithetic, synthetic, and en echelon faults. Crossing conjugate normal faults with 50 to 80 cm of offset connect the main fault traces (Fig. 6). Displacement on these faults increases upwards, and outcrop relationships suggest sequential rather than simultaneous slip along the faults (Ferrill et al., 2000). The fault is characterized by a discrete fracture surface with a multilayered central core composed of fine-grained gouge and/or zones of distributed faulting, as indicated by numerous centimeter-scale faults (Fig. 7) . The central fault core ranges in thickness from 1.5 to 30 cm and consists of layers of fine ash gouge, tectonized clay, and calcite mineralization (Fig. 7b). Fracture networks are typically open and vertical to steeply dipping. Deformation in thicker, more resistant beds that have a higher ash content is characterized by closely spaced vertical and subvertical fractures (up to 20 per meter; Fig. 7c) and fracture density that increase locally near small fault surfaces. Mineralized fracture surfaces up to several millimeters thick are in some cases filled with calcite and/or silica and are connected to the main fault trace and to smaller faults within the damage zone.



View larger version (89K):
[in this window]
[in a new window]
  		Fig. 7. Photographs of the Crucifix fault zone and sampling localities denoted by numbers. Photos highlight fault zone architecture. (A) Note drag geometry and deformation of fault gouge appears similar to that of a clay smear process (Yielding et al., 1997) parallel to the main slip surfaces. (B) Close-up view of fault core. In hand sample we observed a decrease in grain size in the central fault core gouge and numerous small-scale en echelon slip surfaces. X-ray diffraction of Sample BT-43-01 yields calcite mineralization on the outer slip surface and quartz and feldspars comprise the central gouge zone. (C) Closely spaced fractures between two faults that show the relationship between grain size and degree welding and fracture density. The white ash-rich beds (A) have numerous fractures that do not persist into the nonwelded pumice-rich layers (P).

 
X-ray diffraction analyses of host and fault rock reveal that the primary constituents of the host rock include quartz, feldspar, silica glass ± calcite. Compositional variations associated with the faults include the presence of calcite (≤1 cm thick) along the main fault slip surfaces whereas quartz and feldspars are the primary minerals within the central fine-grained gouge material (Table 1, Fig. 8) . The XRD results also suggest that faulting may induce devitrification. Analysis of samples from about 10 to 20 cm from the fault show a broad and scattered XRD signature suggestive of a significant amount of nonordered silica glass, whereas samples from the fault do not exhibit this signature (Fig. 8).


View this table:
[in this window]
[in a new window]
  		Table 1. Summarized results of X-ray diffraction analyses at the crucifix site.{dagger}

 
The fault zone deformation elements described at the Crucifix site may contribute to porosity reduction and the formation of localized barriers to fluid flow along the fault that potentially may disrupt lateral flow and permeability within the surrounding host rock. Meanwhile, an increased intensity of open fractures in the fault damage zone may locally increase permeability and enhance vertical flow in this region. The presence of crossing conjugate normal faults in the footwall and connected to the main Crucifix fault may also create additional permeability anisotropies within the damage zone (Ferrill et al., 2000).

Chalk Cove: 6–8 m of Throw
The Chalk Cove site is located along Chalk Bluffs Road, west of the Crucifix site (Fig. 1) and is approximately 2 km east of the stratigraphic Section 57 of Wilson and Hildreth (1997). The Chalk Cove fault is a north-striking, nearly vertical normal fault with a minimum displacement of 9 m (Fig. 9) . It juxtaposes the upper portion of the lower, poorly to slightly welded, massive ignimbrite of the Bishop Tuff (Unit B, Fig. 3) and the basal part of Unit B in the west block against a massive lower air-fall deposit on the eastern block (Fig. 9a). Up section, structures in Unit C can be observed. Where the fault cuts the densely welded Unit D, the fault trace is a zone of large tilted blocks and rubble, and no fault contact is discernable.



View larger version (73K):
[in this window]
[in a new window]
  		Fig. 9. Chalk Cove study site and fault zone. (A) View of fault showing fault zone architecture of a single fault that flares upward into a wide zone of fractures in the densely welded tuff. (B) Fault core composed of 1-mm to 5-cm-thick clay smear or gouge; heterogeneous mixed zones with varying degrees of deformation. Immediate damage zones on either side of fault have fracture intensity values of 11 fractures per meter. Within first 3 m of hanging wall, fracture intensity values increase to 33 fractures per meter. (C) Slip surface is very narrow in the lower part of the exposure.

 
At the lowermost exposure, the main fault has an extremely narrow fault core with a relatively broad hanging-wall damage zone and a narrow footwall damage zone (Fig. 9). The fault is characterized by a discrete polished surface with a 1 to 3 mm thick fault core (Fig. 9a–9c). This surface becomes progressively curvilinear and splays up-section into several distributed slip surfaces forming a 1-m-wide fault zone in the upper part of Unit B and into Unit C (Fig. 9b). This results in an overall fanning upward geometry as the fault zone changes from a discrete slip surface to a distributed zone of large trace length fractures parallel and subparallel to the main fault trace (Fig. 9 and 10) . The transition variation in fracture density across the fault zone occurs where the tuff has a transition in porosity to <15 to 20%. Thus, the degree of welding, and/or devitrification, which dictates porosity, appears to control the macroscopic deformation mechanism at this site.



View larger version (29K):
[in this window]
[in a new window]
  		Fig. 10. Schematic sketch of the Chalk Cove fault illustrating interpretation of a "funnel-shaped" geometry. Open fractures in welded units narrow to a discrete slip surface. Predominate strike orientations of fault and hanging wall fracture surfaces are north–south to northeast–southwest. Numerous high-angle slip vectors were measured along the trace of the main fault.

 
The fault zone has identifiable architectural and hydrological elements (Fig. 10), including a hanging-wall damage zone, hanging-wall and footwall mixed zones, and a central core zone (Chester and Logan, 1986; Caine et al., 1996; Heynekamp et al., 1999). The damage zone on the west side is approximately 5 m thick and contains abundant, open, iron oxide-stained fractures. The footwall damage zone is 0.05 to 1 m thick (Bradbury et al., 2002). The central fault core varies from a discrete 2-mm slickenlined slip surface to two major slip surfaces with a 5-cm-thick foliated ash central fault core gouge. Initial grain size analyses of samples within the fault core (Fig. 9b) suggest a decrease in grain size within the central fault gouge core. Note, however, that results vary per sample locality, suggesting a heterogeneous internal fault zone core.

Horton Creek Site: 18 m of Throw
The Horton Creek site (Fig. 1 and 11) exposes the largest displacement fault studied in the area. The fault orientation is 010°/070° east, and the fault cuts the pumice-rich ignimbrite (Unit A), which here is a massive- to medium-bedded sequence that in places is composed entirely of poorly to nonwelded and noncemented pumice fragments in the hanging wall. Thin-bedded pumice-bearing beds and the fine-grained pre-Bishop Tuff lacustrine silts lie in the footwall (Fig. 11a–11d).



View larger version (111K):
[in this window]
[in a new window]
  		Fig. 11. Horton Creek outcrop. (A) The Horton Creek fault and fault traces interpreted on the photo. (B) Exposed fault contact between older lacustrine deposits and younger air-fall sequences of the Bishop Tuff. (C) Nonwelded ignimbrite and air fall deposits about 30 m north of the fault trace. Pumice fragments and glass have an average grain size of 0.9 to 1.9 m or coarse to very coarse sand particle size, show inverse grading, and are overall very poorly sorted, angular grains.

 
The fault here consists of a narrow planar contact (Fig. 11b) composed of fine cataclastically deformed tuff. The fault zones is approximately 50 to 60 cm wide, and little deformation is observed next to the fault, except for several subparallel fractures in the hanging wall. Displacement along the fault is concentrated on two surfaces 1 to 2 cm thick, comprised of very fine-grained, pink to red ultracataclasite. Three to 10 m east in the hanging wall, vertical discontinuities cut the tuff (Fig. 11c). Along much of the trace of these features, there are no surfaces that display any shear displacement nor any evidence of the origin of these features. They may be fractures in the hanging wall of the fault and are eroded out due to preferential weathering. Near the upper portion of the outcrop, some of the planar structures are filled by a thin layer of calcite.

In summary, these field observations of the fractured and faulted Bishop Tuff show two general types of faults. Very narrow, planar faults with little associated deformation form in the nonwelded to poorly welded tuff, whereas wide faults in zones of fault-parallel fracturing form in rocks that have been moderately to densely welded.

Microstructures and Deformation Mechanisms
We examined the textures and microstructures of the faulted and fractured nonwelded units to determine how the faults and fractures formed, and how they might impact fluid flow. Most of the fractures in the study area are coated with calcite or calcite–silica coatings, suggesting either a penetrative flow in the system or ubiquitous near-surface coatings that flowed down from the present day ground surface. We examined the microstructures with optical and scanning electron microscopy, using sample preparation techniques outlined in the appendix. We use terminology of Fisher and Schmincke (1984) for the textures of these rocks.

Texturally, the undeformed basal fall deposits of Unit A consist of glass-dominated matrices of Y-shaped glass shards, lapilli of pumice, and small quartz and feldspar grains (Fig. 12 and 13a) . Delicate filaments and vesicle shapes within pumice lapilli suggest little post emplacement compaction and grain–grain or grain–matrix support of the undeformed deposit. Large subangular to subrounded feldspar phenocrysts lie in a glass-shard and ash-rich matrix (Fig. 12). Little devitrificaion is observed in the rocks except near fractures and faults. Little birefringence is seen in thin section, and XRD analyses do not show the presence of amorphous silica or neocrystalized feldspar.



View larger version (78K):
[in this window]
[in a new window]
  		Fig. 12. Mosaic of backscattered scanning electron micrograph of nondeformed nonwelded tuff from the Horton Creek site. Large feldspar phenocrysts (F), vesiculated pumice lapilli (P), and quartz grains (Q) lie in a matrix of irregularly shaped glass shards and ash. Intragranular fractures in the quartz may be the result of rapid cooling from the caldera. Images acquired with 10 KV accelerating voltage.

 


View larger version (115K):
[in this window]
[in a new window]
  		Fig. 13. Optical photomicrographs of textures of nonwelded tuffs. (A) Plane-polarized light image of undeformed tuff from the Horton Creek site. Pumice (P) at right contains elongate vesicles and numerous smaller pumice lapilli in a matrix of glass shards and ash. (B) Plane-polarized light image of open fractures (F) in a feldspar phenocryst that appear to be truncated at the pumice grain boundary at center of image. (C, D) Fragmentation of tuff samples from the Crucifix site. Fractured quartz grains (Q) lie next to relatively undeformed pumice (P) and a matrix of cataclastically deformed grains (C). (E) Cross-polarized light image of cataclastically deformed sample at the Crucifix site (BT-50). Birefringent material is calcite filling the fault. (F) Calcite-filled cataclasite in a fault zone from the Crucifix site. CC, calcite.

 
Microstructures of fractured nonwelded tuffs reveal the processes by which brittle deformation occurs in these units. Samples were collected from the tips of fractures and were carefully collected and prepared so as to preserve the natural fractures. The progressive brittle deformation can be seen in the microstructures at the thin-section scale (Fig. 13). Undeformed, nonwelded tuff (Fig. 12 and 13a) evolves to zones of fractured phenocrysts (Fig. 13b and 13c), and notably the intragranular fractures appear more intense in the quartz and feldspar grains rather than in the pumice. Hence, intragranular fractures may be the result of tectonic deformation, or may be the result of cracking during cooling during and immediately after eruption. Grain-scale fracture gives way to narrow zones of cataclasis (Fig. 13e and 13f) that have a calcite-rich matrix.

Optical and scanning electron micrographs reveal that fractures that cut the pumice-rich sequences of the fall deposits follow grain boundaries, and that little deformation is recorded adjacent to the fracture. Deformation at the fracture tip appears to die into small grain boundary fractures, and the walls of these open fractures are relatively smooth (Fig. 14b) . Connectivity between the fractures and the matrix porosity is variable in the single fractures. In some cases the fracture walls appear to be unconnected to the matrix (Fig. 14) whereas in other cases the fractures have re-entrants that create an irregular fracture wall (Fig. 14b). Note that in this case the pores that lie in contact with the fracture are typically much smaller than the width of the fracture, and are not well connected to the matrix.



View larger version (83K):
[in this window]
[in a new window]
  		Fig. 14. Single fractures in nonwelded tuff. (A) Mosaic of plane-polarized light photomicrographs of a sample parallel to and 1 cm from the Chalk Cove fault. Fracture (f) that traverses the sample cuts dense ash and glass-shard matrix with quartz (Q) phenocryrsts, lithic fragments (L), and an ash matrix (M). Fabric density is due to compaction and deformation associated with the fault. (B) Mosaic of backscattered scanning electron micrograph of natural fracture from the Crucifix site. Fracture (f) lies along the phenocryst and lapilli boundaries. Irregular fracture boundary is produced by the interaction of the fracture and the primary pores. Note that the matrix pores are in general much smaller than the fracture aperture, and the pore spaces (black regions) are not well connected. Q, quartz; F, feldspar; P, pumice.

 
Zones of cataclasis and deformation band faults (Wilson et al., 2003) represent the next stage in the development of the thin faults. The narrow faults may consist of a zone of angular to subrounded feldspars and pumice grains in a fine-grained matrix of comminuted glass and calcite (Fig. 15a) . Large pumice grains lie adjacent to these faults and the grain-scale strength of the pumices are shown by the intact nature of the grain (Fig. 15a), where the thin bubble walls are not broken. Note also that open fractures may lie along the edge of these zones of cataclasis (Fig. 15a). The other expression of slip is narrow slip surfaces (Fig. 15b) that contain thin calcite surfaces. The presence of calcite in these zones and along the narrow slickensided surfaces observed at the outcrop (Fig. 15d) indicates that fluid flow has occurred along the surface.



View larger version (155K):
[in this window]
[in a new window]
  		Fig. 15. Cataclastic deformation and development of deformation band fault in the nonwelded tuff. (A) Mosaic of backscattered scanning electron image of a fault zone comprised of feldspar and pumice fragments in a calcite-ash-glass shard matrix. Fault zone lies below a large pumice grain (P) at the top of image. Note the lack of internal deformation of the pumice grain, despite the large volume of pore space (black) and the thin walls between the pores. Fault zone has fractures (f) along either side of fault; Q is quartz, (B) Thin zone of cataclasite with calcite (C) along the surface. Matrix consists of nonwelded tuff.

 
The impact of the fractures associated with the deformation on fluid flow is shown in Fig. 16 . Fractures that were probably open are filled with calcite with acicular crystal habit (Fig. 16a). Filling of fractures was in some cases associated with subsequent deformation (Fig. 16b) that has both calcite and iron oxide coatings. Fine-grained cataclasite with entrained calcite (Fig. 15a and 16c) attests to early calcite deposition and subsequent shear deformation. Swarms of fractures also lie in fine-grained fault gouge (Fig. 16d).



View larger version (115K):
[in this window]
[in a new window]
  		Fig. 16. Evidence for fluid flow associated with faults. (A) Open fracture filled with calcite. Cross-polarized light image with quarter-wave plate in. (B) Calcite-filled vein (C), with iron oxide fill (dark material) subparallel to the fine-grained fault gouge of the Horton Creek fault. P is pumice. Cross-polarized light image. (C) Open fractures (f) cutting pumice (P) and feldspar (F). Red iron-oxide staining lies in a fault zone. Cross-polarized light image with reflected light. (D) Fine-grained fault gouge that cuts calcite (C) at the Crucifix site. Cross-polarized light image.

 
Interpretations
The observations presented here document the brittle deformation of the highly porous nonwelded tuffs. We address two questions:

Brittle deformation of unconsolidated deposits is an increasingly well-documented phenomenon. Small faults in unconsolidated sediments have been documented in late Tertiary deposits in New Mexico (Heynekamp et al., 1999; Rawling et al., 2001), in sands of northern California (Cashman and Cashman, 2000), in experiments in deformation of sands (Finno et al., 1997; Saada et al., 1999) and in nonwelded tuff (Wilson et al., 2003). These data all show that localized brittle failure can occur in unconsolidated deposits.

A significant difference between the observations cited above (with the exception of Wilson et al., 2003) and our data is that we see similar zones of cataclasis, and Mode I fractures, in materials with ≥50% porosity (Sheridan, 1968). One might think that such materials would deform by bulk grain flow. However, the dominant deformation mechanism appears to be brittle fracture followed by cataclasis to form deformation-band faults (Wilson et al., 2003). We suggest that several factors come into play to create the structures we see here. There are few geologic analogs to the deformation of these pumice-rich nonwelded units, but there are similar materials in the field of engineered materials—those of cellular solids or brittle foams (Gibson and Ashby, 1999). Four factors play into creating brittle failure in these materials:

  1. While the bulk material porosity may be >50% for these deposits, a significant fraction of this porosity is in the intragranular porosity of pumice lapilli (Ragan and Sheridan, 1972), and thus, the effective porosity in the most porous portions of the deposit may be closer to 35 to 40% (Day, 1993), and where ash content is high, even lower.
  2. These deposits are overlain by as much as 150 m of variably welded tuff. Thus, at the time these faults and fractures developed, there were subjected to as much as 2.0 MPa of lithostatic load, so deformation would have occurred when the grains were constrained.
  3. The fall and surge deposits were probably emplaced at temperatures of at least 200 to 250°C and were buried by a thick sequence of hot ash, perhaps as hot as 650 to 730°C (Wilson and Hildreth, 1998: Wallace et al., 2003), which cools slowly over years (Fisher and Schmincke, 1984; Riehle et al., 1995; Wallace et al., 2003). This would have resulted in sintering and welding at grain contacts (Bierwirth, 1982, as cited in Cas and Wright, 1986, p. 256–258; Ragan and Sheridan, 1972).
  4. The deformation observed here is the result of rapid loading from below by slip during earthquakes and related shaking, and the rate of loading is approximately 10–1 s–1.

We suggest that these deposits are brittle, open-cell ceramic or metal foam. The relatively fine-grained materials and/or the contact strength at bubble walls provide a cohesive strength to the "nonwelded" unit. While in outcrop the nonwelded tuff typically crumbles, the units often support near-vertical rock faces, indicating that it has some cohesive strength. In a constrained subsurface environment loaded at seismic slip rates, these materials had a small, but measurable cohesive strength. Such materials, when subjected to a variety of loads, will fail (Gibson, 1989; Gibson and Ashby, 1999) by brittle crushing. Failure in engineered materials occurs where the strength of the thin cell walls is exceeded in the purely brittle case (Gibson, 1989). Many of the fractures examined here appear to lie along grain boundaries, suggesting the contact strengths at these grain contacts are low. Several of the fractures imaged with SEM suggest a process zone formed in front of the cracks, where microcracking represents a cohesive zone ahead of the crack tip.

Mechanical properties for nonwelded tuffs are scarce, and most rock mechanics literature on the topic examines rocks that are slightly denser and less porous than the lower part of the Bishop Tuff (Price and Bauer, 1985; Day, 1993; Moon, 1993a, 1993b; Schultz and Li, 1995). Day (1993) is among the few studies of truly nonwelded tuff. He shows that nonwelded portions of the Bandilier Tuff have cohesion of 5.7 kPa and a uniaxial compressive strength of 5.7 kPa. In a confined ring uniaxial compression test, the void ratio only decreased from 1.3 to 1.12, showing that much of the porosity is mechanically isolated. This suggests that the pumice fragment walls have strength and that intergranular porosity is lower than the total porosity. Thus, when the pumice is compressed, deformation would first collapse intergranular pores. The data for very high porosity samples studied by Price and Bauer (1985), Moon (1993b), and Schultz and Li (1995) in general agree with these results. The available porosity data from these experiments suggest that 15 to 25% of the total porosity is isolated, probably within pumice fragments. Thus, the effective porosity would be about 35%, making these units hydrologically equivalent to a porous granular material.

This brittle deformation, driven by the east–west extension in the region, creates a strong anisotropy imparted by the subvertical fractures and faults at the outcrop to map scales. The coatings along the fractures and faults, and some of the in situ permeability testing (Dinwiddie et al., 2002a, 2002b), shows that these features can impart a strong control on vertical fluid flow in these unconsolidated deposits. The faults create a significant reduction in permeability and porosity, and thus may create a barrier to lateral flow (see also Heynekamp et al., 1999; Sigda et al., 1999). Fracture walls are also either relatively smooth, or connect to a matrix porosity that has much smaller pore throat sizes than do the fractures. Therefore, these fractures would likely promote thin-film flow (Doe, 2001; Or and Tueller, 2000). Thus, in regions where poorly consolidated and highly porous tuffs are in the unsaturated zone, and where the hydraulic gradient is vertical, the fine-grained faults would likely impede horizontal dispersion, and the fine-grained surfaces might provide relatively smooth paths for fluids vertically downward (see Sigda and Wilson, 2003).

Fractures in these deposits may likewise increase the vertical velocity of flow. While the nonwelded tuffs of the Bishop tuffs have extremely high porosities, the microstructures examined here suggest that the pore spaces along the fracture walls are very small, and there is relatively little interconnected porosity in the nonwelded tuff adjacent to the fractures. The fractures provide much larger, and reasonably well-connected, zones through which water may flow.

The structure of the faults as they cut more densely welded tuffs upsection also has hydrologic implications. Closer to the surface, the faults consist of funnel shaped fracture zones around the main fault, which tapers to the thin faults seen in the nonwelded tuff (Fig. 11). This structure would collect water from a relatively large area along the fault and focus flow to the narrow slip surface. This broad zone of deformation is manifested at the surface as large blocks at the base of the fault scarps. DePolo (quoted in Smith, 1987) reported that during an Mw 6. 4 earthquake under the Tableland, the jointed rock mass " jostled in an independent manner." Thus, the upper portions of the faults have both slip and a broad zone of fracturing associated with them.

The impact of faults and fractures in poorly welded tuff in the unsaturated zone is an important issue at Yucca Mountain (Flint et al., 2001a, 2001b, 2001c). The distributions of faults and fractures have implications for Yucca Mountain and the 10000 yr performance of the repository site (Ferrill and Morris (2001), including controlling permeability architecture and groundwater flow rates and paths in fractured tuff. Fractures and faults in the Yucca Mountain site are more complex than at Bishop, with a variety of orientations and ages (Sweetkind et al., 1997a, 1997b). At Yucca Mountain sheet-like fractures and pipe-like features cut the nonwelded Paintbrush Tuff, which is very similar to the Bishop Tuff. These surfaces and linear structures may be the fast pathways inferred to exist from the chlorine isotopic data (Levy et al., 1999; Wolfsberg et al., 2000). The data we present here suggest that fractured and faulted Paintbrush Tuff would conduct fluids vertically.

Flint et al. (2001c) reviewed the evolution of thought regarding the conceptual model for unsaturated flow in the nonwelded part Paintbrush Tuff (Ptn). The Ptn is similar in many ways to the Bishop Tuff, and the current hydrologic model suggests that "Most of the infiltrating water passes through the fractures of the Tcw [authors addition: the welded tuff above the Paintbrush Tuff] to be slowed during transition to matrix flow in the PTn except where faults or broken zones disrupt the PTn, providing fast pathways for a small component of flow" (emphasis added). The results of numerical models for flow in the PTn (Flint et al., 2003) also indicate that vertical flow through the PTn via faults and fractures would dominant over lateral diversion of flow. This is in contrast to the model of Wang and Narasimhan (1985), which suggests that significant amounts of matrix imbibition occur before fractures in these rocks is important. Recent analyses of 36Cl data (Campbell et al., 2003) show that faults that cut the PTn provide pathways for fast flow in the system, and in situ experiments with faulted PTn indicate that as wetting increases in the matrix adjacent to faults, fluid flow travels farther along the fault surface (Salve et al., 2003).

Our observations of fractures and faults in the Bishop Tuff support the conceptual model of Flint et al. (2001a)(2001b, 2001c) and the results of Flint et al. (2003), Campbell et al. (2003), and Salve et al. (2003). Hydrologic tests of the nonwelded Bishop Tuff examined how much of the total infiltrations may occur along the fracture and faults (Fedors et al., 2001; Bradbury et al., 2002; Dinwiddie et al., 2002a, 2002b). We suggest that the smooth surfaces created along faults may promote fast water flow as rivulets or even sheet flow along the fault surface. Microscopic observations of natural fractures in the Bishop Tuff show that the aperture of the fracture is much greater than the aperture of pores adjacent to the crack. This suggests that the water would more likely flow along a vertical fracture rather than overcome the capillary pressure needed to enter the matrix via the small pores. Finally, in situ dye tracer tests (Fedors et al., 2001) showed that vertical matrix flow in the nonwelded tuff between filled fractures penetrates more deeply than in nonfractured nonwelded tuff, perhaps because of the constraints placed on the lateral dispersion of water (Fedors et al., 2001). The most important unresolved question in these rocks is not whether fractures and faults enhance the vertical flow of water in unsaturated nonwelded tuffs, but rather what the relative proportion of fracture flow is relative to matrix infiltration away from faults and fracture zones.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 Geologic Setting of the...
 METHODS
 RESULTS
 CONCLUSIONS
APPENDIX
 REFERENCES
 
The data presented here, along with the relatively few data on the fracture characteristics of other nonwelded tuffs and rock mechanics data, indicate that nonwelded tuffs have small, but finite cohesive strengths, and that when loaded in a triaxial state of stress, will deform either by Mode I fractures or by shearing and faulting. The mode of deformation is a function of welding, with the highly porous nonwelded, pumice-rich tuffs deforming by the formation of narrow, smooth deformation band faults. Fracture density near faults increases in more welded tuff. Furthermore, we show that these fractures and faults can be surfaces along which fluids may travel and create a hydraulic anisotropy. In the case of such deformed deposits in the unsaturated zone, and in regions where the tuffs experience extension, the orientation of the fractures and the hydraulic gradient create much faster vertical flow paths than horizontal. This behavior is probably enhanced in these pyroclastic deposits, where the nonwelded tuffs are overlain by very low-porosity welded ash-flow tuffs, where the fractures provide the only connected flow paths. If the fracture densities are greater in these welded units, then recharge at the surface would be focused into the faults at depth.


   APPENDIX
TOP
 ABSTRACT
 INTRODUCTION
 Geologic Setting of the...
 METHODS
 RESULTS
 CONCLUSIONS
 APPENDIX
REFERENCES
 
Sample Collection and Preservation Techniques
Collecting and preparing samples in deformed, non- to poorly welded, pumice-rich tuffs in a manner that preserves the fabrics and does not introduce new fractures within the samples is very difficult. We adapted field techniques used in the soil sciences (e.g., Puentes et al., 1992; Vepraskas et al., 1991) and for preparation of poorly consolidated materials (Fox et al., 1993; Middleton and Kraus, 1980). Samples of undeformed tuff were collected in one of three ways:

  1. We formed a semicircular depression 3 to 6 cm deep at the top of a scraped surface and poured a slow-acting epoxy resin into the depression. The resin was allowed to seep into the tuff, and the hardened sample was collected the next day. We also applied a negative pressure to scraped surfaces next to the depression, but are not certain that this helped suck resin into the rocks.
  2. Some samples were collected by chipping and prying from the outcrop; we immediately immersed the samples in a fast-acting epoxy resin.
  3. Some samples were collected by cutting blocks up to 40 cm on a side with a portable saw, and lifting the sample out of the outcrop and placing it on a wood or cardboard platform. The sample was then immediately stabilized by pressing wood or cardboard coated with fast-acting foam insulation against all sides of the sample and allowing the foam to compress the sample slightly.

Deformed samples were collected in one of two ways. Some samples could be carefully excavated from the outcrop by carving with a knife or chisel, and stabilized with foam insulation, or by pouring epoxy resin into fractures and along fault surfaces. These samples or peels were then cut out of the rock face, and stabilized by placing the sample into a small plastic resealable bag. This bag was then set into spray-foam insulation in an outer resealable bag. This provided lateral support for the sample during transport but keeps the insulation material from penetrating the sample.

In the laboratory, the foam-stabilized samples were carefully opened such that a small free face was exposed. Extremely friable samples were baked before epoxy for several hours at <100°C. Slow-acting thermally activated petrographic epoxy (Petropoxy 154) was poured on the free face, and a high vacuum was applied to the sample. Successive exposure and vacuum cycles ensured that the entire sample was stabilized with an inert epoxy. The sample was then heated to 125°C for 8 h to set the epoxy. Thin section chips were then cut from the samples, and these chips were epoxied again as needed.

Porosity values for the nonwelded portion of the Bishop Tuff are rare. We determined porosity in the lab, and supplemented these data with values from the literature. Most published data give density values, and we used the same calculation as in our laboratory determination to find porosity. We isolated samples on the nonwelded tuff in a vacuum-sealed bag to prevent water from seeping into the pores, and measured its density with respect to water by determining the displaced water volume. Porosity is given by:

where {phi} is porosity, {gamma}sample is specific gravity of the sample, and {gamma}constituents is the specific gravity of the solid constituents, which we take to be 2.5 g cm–3, based on mineral and glass of 30% glass (2.4 g cm–3) and 70% feldspar (2.55 g cm–3) and quartz (2.67 g cm–3) presented in Hildreth (1979) and Wilson and Hildreth (1998). This technique allowed us to measure porosities of samples with large pumice clasts, which are difficult to sample with image analysis techniques without damaging the sample.


   ACKNOWLEDGMENTS
 
Numerous discussions with Randy Fedors, David Ferrill, Cynthia Dinwiddie, Craig Forster, and Dani Or provided much insight into the nature of flow in the unsaturated zone, and the relative importance of fracturing in unconsolidated materials. Don Fiesinger helped us interpret some of the textures of these rocks. Field assistance by Jason Heath, Betty Paepke, and Richard Heermance, and laboratory assistance by Stephanie Carney and Stacy Peterson are gratefully acknowledged. Thanks to John Donavan at the University of Oregon Electron Microbeam facility for assistance with SEM work, and to Peter Kolesar at Utah State for assistance with the XRD analyses. The views expressed here are solely those of the authors, who take full responsibility for all statements herein. This paper represents the results of work conducted through 9 May 2002 by both authors.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 Geologic Setting of the...
 METHODS
 RESULTS
 CONCLUSIONS
 APPENDIX
 REFERENCES