orthogonal jointing during coeval igneous degassing and ...orthogonal jointing during coeval igneous...

For permission to copy, contact [email protected] 2003 Geological Society of America1492

GSA Bulletin; December 2003; v. 115; no. 12; p. 1492–1509; 9 figures; 1 table; Data Repository item 2003168.

Orthogonal jointing during coeval igneous degassing and normalfaulting, Yucca Mountain, Nevada

William M. Dunne†Department of Geological Sciences, University of Tennessee, Knoxville, Tennessee 37996, USA

David A. FerrillCenter for Nuclear Waste Regulatory Analyses, Southwest Research Institute, San Antonio, Texas 78238, USA

Juliet G. CriderDepartment of Geology, Western Washington University, Bellingham, Washington 98225, USA

Brittain E. HillDeborah J. WaitingPeter C. La FeminaCenter for Nuclear Waste Regulatory Analyses, Southwest Research Institute, San Antonio, Texas 78238, USA

Alan P. MorrisDepartment of Earth and Environmental Sciences, University of Texas, San Antonio, Texas 78249, USA

Randall W. FedorsCenter for Nuclear Waste Regulatory Analyses, Southwest Research Institute, San Antonio, Texas 78238, USA

ABSTRACT

An orthogonal system of tube-bearingjoints constitutes the oldest fractures in theTiva Canyon Tuff at Yucca Mountain, Ne-vada. The joints formed within a month ofignimbrite deposition, prior to major de-gassing. The system consists of (1) narrow,persistent, northeast-striking joint swarmswith trace lengths typically greater than 5m and between-joint spacings of less than20 cm and (2) northwest-striking swarmsthat have a more en echelon geometry andgreater between-joint spacings compared tothe northeast-striking swarms. Between-swarm spacing for both trends is ;50 m.Questions concerning the joints include thefollowing: (1) What was the origin of driv-ing stress(es) for formation of the joints,particularly as their orientations were notconsistent with the regional stress geometryat the time of their formation? (2) Whatmechanism caused the horizontal principalstresses to be reoriented so as to yield anorthogonal geometry? (3) What insightscan be developed for predicting joint ge-ometry in unexposed rock volumes by un-derstanding joint origin? These questions

†E-mail: [email protected].

are important because the joints and otherfractures may affect the performance of aproposed high-level nuclear waste reposi-tory within Yucca Mountain.

To address these questions, we use newand existing field data about joint geometryand the relationships of joints to degassingstructures, numerical modeling of fault be-havior, and work by previous authors. Ourinterpretation begins with the initial ignim-brite eruption and deposition during cal-dera collapse. The ignimbrites were depos-ited over a preexisting topography thatpossibly included a shallow northwest-trending basin in the Yucca Mountain area.During initial ignimbrite cooling, jointswarms formed as elements of orthogonalfumarolic ridge systems where degassingwas associated with vertical dilation. Bothjoint sets have unusual tubes that are inter-preted to have formed during dilation andsegmentation of joint faces resulting fromlithophysae inflation in the cooling ignim-brite deposit. Modeling supports the inter-pretation that a combination of regionalstresses and stress related to slip on localnormal faults controlled the orientation ofthe joint swarms and favored the formationof the northeast-striking joints first. The

faults might have moved in a stress fieldalready perturbed by caldera collapse. For-mation of the northwest-striking joints oc-curred after a local 908 switch of horizontalprincipal stress directions due to the pres-ence of the northeast-striking swarms, pos-sibly aided by differential compactionacross the northwest-trending basin. Tube-bearing joints occur in all lithophysae-bearing lithostratigraphic units of the To-popah Spring Tuff, which is thestratigraphic interval for the proposed nu-clear waste repository within Yucca Moun-tain. We conclude that the tube-bearingjoints formed in the same manner andshare similar geometric characteristics,adding a persistent joint population to theoverall fracture system that influences hy-drological and mechanical properties.

Keywords: orthogonal joints, welded tuffs,cooling joints, degassing tubes, perturbedstress fields, normal faults.

INTRODUCTION

We consider the origin of tube-bearingjoints in the Miocene Tiva Canyon Tuff (12.7Ma) and extrapolate the study to other tube-bearing joints at Yucca Mountain, Nevada.

Geological Society of America Bulletin, December 2003 1493

ORTHOGONAL JOINTING DURING COEVAL IGNEOUS DEGASSING AND NORMAL FAULTING

The faces of tube-bearing joints are distinc-tively ornamented by a network of linear de-pressions with curved bottoms (Figs. 1A, 1B)that are symmetrical across the two faces of ajoint. The tube-bearing joint system is intrigu-ing because (1) two nearly orthogonal jointsets define the system, (2) both joint sets areclearly associated with ignimbrite degassing,and (3) neither joint set is perpendicular to theapproximately east-west direction of regionalextension at the time of both sets’ formation(Zoback et al., 1981; Morris et al., 1996).Thus, the joints formed during ignimbritecooling in response to additional stress com-ponents that do not reflect the regional stressfield. These stress components also triggereda 908 switch of horizontal stress directions thatproduced both sets of the near-orthogonalsystem.

The joint system is also of interest becauseit is located above the proposed site for theUnited States of America’s first high-level ra-dioactive-waste repository in Yucca Moun-tain, Nevada. The joint system is a prominentcomponent of the pathways for shallowgroundwater infiltration through the capping,moderately to densely welded tuff into the re-pository (Winograd, 1971; Flint and Flint,1995; Buesch et al., 1996; Flint et al., 1996).Joints with morphologies similar to those ex-posed at the surface of Yucca Mountain arealso found in subsurface exposures in the Ex-ploratory Studies Facility (ESF) and EnhancedCharacterization of the Repository Block(ECRB) tunnels (Mongano et al., 1999;CRWMS M and O [Civilian RadioactiveWaste Management System Management andOperating Contractor], 2000). Joints that weinterpret to have formed early in the weldedignimbrites tend to be relatively large and rep-resent potential weaknesses that could influ-ence the stability of proposed excavationssuch as underground waste-emplacementdrifts. A better understanding of the geometryand origin of these joints provides a basis forbetter estimating their importance to the de-sign of underground openings.

In this paper, we (1) provide a summaryof the joint formation history based on newanalyses and published information for thedegassing-related joints in a part of the TivaCanyon Tuff; (2) identify the sources of thedriving stresses that may have perturbed theregional stress field to form this orthogonaljoint system; and (3) consider causes for the908 switch of horizontal stress directions thatcaused the orthogonal geometry. Although theanalysis is conducted in volcanic rocks, theresults are applicable to faulted sedimentaryrocks with elevated fluid pressures. To analyzethe origin of this joint system, we have used

existing field data (e.g., Morgan, 1984; Bartonet al., 1984, 1993; Barton and Larsen, 1985;Sweetkind et al., 1995a, 1995b; Throckmortonand Verbeek, 1995), newly gathered field-based fracture data employing a differentialGlobal Positioning System (GPS) and geo-graphic information system (GIS) approach,new observations of field relationships be-tween joints and volcanogenic features, andnumerical modeling of stress-field perturba-tion during active faulting.

DEVELOPING ORTHOGONAL JOINTSYSTEMS

Orthogonal systems of two vertical jointsets are common (Ver Steeg, 1942; Eyal andReches, 1983; Hancock, 1985; Stauffer andGendzwill, 1987; Dunne and North, 1990;Rawnsley et al., 1992; Martel, 1994; Caputo,1995; Olson, 1996; Fabbri et al., 2001). Nev-ertheless, our understanding of their origin isproblematic because the orthogonal geometryof the two sets requires at least one 908 changein local principal stress directions during jointsystem formation, so that each set forms nor-mal to the minimum principal compressivestress. Specifically, the minimum horizontalstress direction (Sh) must change by 908 orswitch with the direction of maximum hori-zontal stress (SH). Three groups of mecha-nisms have been proposed (Lachenbruch,1962; Stauffer and Gendzwill, 1987; Dunneand North, 1990; Caputo, 1995; Olson, 1996):(1) reorienting due to small stress fluctuationsbetween two near-equal horizontal principalstresses (Sh ø SH), (2) local stress switchingat the scale of individual joints, and (3) re-gional stress switching by 908. Switching twonearly equal horizontal principal stresses is anintuitively simple scenario because littlechange is required in the stress conditions togenerate the new stress geometry. Yet, whenthese two stresses are nearly equal, joints mayform in more than one orientation becausehorizontal stress conditions are essentially iso-tropic, which leads to the formation of randomor columnar cooling joints (Lachenbruch,1962; Pollard and Aydin, 1988) rather thanorthogonal joints. Thus, this stress state is un-likely to yield the orthogonal geometry. Re-cent numerical modeling has shown that localstress switching is plausible and could occuras a consequence of close spacing in layer-bounded joints (Bai et al., 2002), although ini-tial analysis by Martel (1994) indicated onlylimited potential. Regional switching may oc-cur as a result of regional tectonic changes instress state over time or relaxation of stressesas a result of uplift and erosion (Engelder,1985; Hancock, 1985; Stauffer and Gendzwill,

1987; Dunne and North, 1990; Rives et al.,1994; Caputo, 1995). These switches do nothave to occur in a short timeframe, whichmeans that two joint sets with an orthogonalgeometry may each form at different times inresponse to unrelated stress regimes (Gross,1993; Rawnsley et al., 1998). Thus, both lo-cal-scale and regional switching mechanismsmay each account for the formation of someorthogonal joint systems.

At Yucca Mountain, identification of themost likely mechanism and scale for trigger-ing changes in stress orientation that formedthe orthogonal joint network is needed. A va-riety of volcanic and tectonic driving stressesis feasible: (1) the region was actively extend-ing; (2) a large-scale caldera eruption had justoccurred; (3) the ignimbrite containing thejoints was still cooling, probably with differ-ent temperature gradients through the deposit;(4) the ignimbrite may have compacted dif-ferentially during joint formation; and (5) nor-mal faults near Yucca Mountain could havemoved seismically at that time. Consideringthe relative roles of these factors during jointformation is necessary to understanding themechanisms affecting stress orientations.

REGIONAL GEOLOGY

Tectonic Setting

The orthogonal joint system of interest is inthe welded ignimbrites of the Miocene TivaCanyon Tuff, which was emplaced across thelocation of the present Yucca Mountain insouthwestern Nevada at 12.7 Ma (Fig. 2A).Yucca Mountain consists of a thick accumu-lation of gently east-dipping Miocene tuff de-posits cut by an array of major north-striking,west-dipping normal faults, such as the Soli-tario Canyon, Bow Ridge, and PaintbrushCanyon faults, and northwest-striking dextralstrike-slip faults (Scott and Bonk, 1984; Dayet al., 1998a, 1998b; Ferrill et al., 1996a,1999a; Ferrill and Morris, 2001). These faultsaccommodated active crustal extension withinthe Basin and Range province through the Ce-nozoic to the present day (Scott, 1990; Wer-nicke, 1992; Axen et al., 1993; Morris et al.,1996; Day et al., 1998a).

Although this period of tectonic activity in-cludes the time of Tiva Canyon Tuff volca-nism, data supporting fault displacements co-eval with volcanism and cooling are limited.The Solitario Canyon fault has several splayfaults that exhibit displacement that decreasesupward to fault tips in the Tiva Canyon Tuff(Day et al., 1998a), which are interpreted toresult from normal faulting after or during de-position of the Tiva Canyon Tuff (Day et al.,

1494 Geological Society of America Bulletin, December 2003

DUNNE et al.



Figure 1. Photographs of joints in the upperlithophysae-bearing unit of the Tiva Can-yon Tuff. (A) Smooth cooling joint with tu-bular structures defines one face of looseblock (right side of photo), and later joint(at ;908) with rougher surface morphologycuts lithophysae (left side of photo). (B) Tu-bular structures on cooling joint surfaceshown in A. (C) Cooling joint swarm onpavement P100 (view toward the south-west). (D) Intersection of two cooling jointson pavement P100. (E) Oblique view ofhand sample showing smooth cooling jointwith tubular structures that are bowed ad-jacent to ;5-cm-diameter lithophysa. (F)View parallel to main cooling joint surfacein E shows deflection or bowing (arrows) ofjoint surface from overall planar geometry(dashed line). Total deflection is ;2 mm.Note in E and F that deflection of the cool-ing joint surface locally occurs as discretesteps across tubular structures. (G1) Detailof smooth cooling joint with tubular struc-tures. Inset shows small-aperture fracturesand large-aperture tubes cutting one face ofa cooling joint. Tubes and fracture tipshave bleached walls. The expanded imageshows fractures with apertures rangingfrom ,1 mm at fracture tips to nearly 1 cmalong major tubes. Regardless of aperture,fracture walls show bleaching indicative ofvapor-phase mineralization and alterationof adjacent rock. Unbleached areas awayfrom tube-bearing joints are part of thevertical joint surface that would have beenforced tightly closed by lateral stress causedby gas expansion and formation of litho-physae. (G2) Restoration of tubes shown inG1 by matching sides to show volume in-crease using block translation without ro-tation. (G3) Restoration of tubes shown inG1 by matching sides to show volume in-crease using block translation with minorrotation (counterclockwise indicated by228) of some blocks.N Figure 1. (Continued.)

1998a). From unpublished subsurface thick-ness data, Fridrich (1999) stated that some ofthe major north-striking faults in the YuccaMountain region were active between 13.1and 12.7 Ma, including the time of eruptionfor the Tiva Canyon Tuff. Finally, significantnormal-fault displacements occurred duringthe million years after deposition of the TivaCanyon Tuff, on the basis of abrupt changesin thickness of the overlying 11.6 Ma RainierMesa Tuff where it crosses faults (e.g., Chris-tiansen et al., 1977; Sawyer et al., 1994; Mon-

sen et al., 1992; Day et al., 1998a; Fridrich,1999).

Volcanic Setting

The Tiva Canyon Tuff of the PaintbrushGroup is a sequence of pyroclastic flow andminor tephra-fall deposits that most likelyerupted from the Claim Canyon caldera,which is partly preserved on the southern edgeof the younger Timber Mountain caldera com-plex (Fig. 2B, Table 1) (Byers et al., 1976;Buesch et al., 1996; Potter et al., 2002). The

Tiva Canyon Tuff consists mostly of moder-ately to densely welded ignimbrites with de-vitrification, welding, and vapor-phase alter-ation features that resulted from the eruptionof at least 1000 km3 of compositionally zonedrhyolitic magma at 12.7 6 0.03 Ma (Byers etal., 1976; Sawyer et al., 1994). The ;100-m-thick Tiva Canyon Tuff has two lithostrati-graphic members: a thin, crystal-rich uppermember and a thick, crystal-poor lower mem-ber (Buesch et al., 1996). These membersform a single cooling unit (Rosenbaum,1986). Our study was conducted within the


DUNNE et al.

Figure 2. (A) Geologic map of Yucca Mountain, Nevada. Map is based on geologic mapping of Day et al. (1998b) and shows proposedrepository outline (bright blue), topographic contours, distribution of Tiva Canyon Tuff, Exploratory Studies Facility (ESF) tunnel(projected surface trace shown by dark blue line), Enhanced Characterization of the Repository Block tunnel (ECRB) (projected surfacetrace shown by dark blue line); locations of pavements P100, P200, and P300; and the joint study area (dark blue rectangle) on LiveYucca Ridge. Repository area shown by cross-hatching. (B) Regional map showing location of Yucca Mountain with respect to FranRidge and the caldera complexes to the north. Base image is Landsat 5–Thematic Mapper acquired from the U.S. Geological Survey.Caldera coverage from Frizzell and Shulters (1990). Maps in both A and B are in Universal Transverse Mercator (UTM), Zone 11,North American Datum 1927 (NAD27).



Figure 2. (Continued.)

TABLE 1. STRATIGRAPHIC COLUMN FOR YUCCAMOUNTAIN

Formation Age(Ma)

Timber Mountain GroupRainier Mesa Tuff 11.6Paintbrush Canyon GroupBedded TuffTiva Canyon Tuff 12.7Yucca Mountain TuffPah Canyon TuffTopopah Spring Tuff 12.8Calico Hills Formation 12.9Crater Flat GroupProw Pass TuffBullfrog Tuff 13.25Tram Tuff .13.5

Note: After Sawyer et al. (1994).

20-m-thick upper lithophysae-bearing zone ofthe crystal-poor member (Buesch et al., 1996).

Variations in preemplacement topographyand depositional thickness affect the coolingbehavior of ignimbrites and potentially affecttheir fracture development. The base of the

Tiva Canyon Tuff is poorly exposed, so pre-emplacement topography is difficult to iden-tify. Also, the top of the Tiva Canyon Tuff iseroded, obscuring its thickness variations.Still, the spatial distribution of stratigraphicthicknesses for underlying tuffs from thePaintbrush Group may indicate the location ofdepocenters for the younger Tiva Canyon(Buesch and Spengler, 1998; Fridrich, 1999).If it is assumed that the ignimbrite tops wereflat after deposition, a northwest-striking basinfor the underlying Topopah Spring Tuff liesdirectly north of the study area where the unitthickens 50 m into the depocenter. The small-volume Yucca Mountain and Pah CanyonTuffs, which are immediately underneath theTiva Canyon Tuff, also show the effects of anorthwest-trending paleobasin and deposition-al control from a gentle topographic high thatlimited ignimbrite deposition immediately tothe south (e.g., Day et al., 1998b; Fridrich,1999).

Processes of Ignimbrite Cooling

Ignimbrite deposits analogous to the TivaCanyon Tuff are commonly emplaced at tem-peratures within 100 8C of magmatic temper-atures, because little cooling occurs duringflow except near the basal and upper contactsof the unit (Banks and Hoblitt, 1981; Bursikand Woods, 1996). The Tiva Canyon Tufflikely had a preeruption temperature of ;7008C (Lipman and Friedman, 1975). Uniformpaleomagnetic directions in the Tiva CanyonTuff indicate emplacement temperatures abovea Curie temperature between 580 and 640 8Cand that any potential rheomorphic deforma-tion occurred above the Curie temperature(Rosenbaum, 1986, 1993).

Degassing occurs rapidly during and afterignimbrite emplacement at a rate largely con-trolled by variations in permeability (Miller,1990; Riehle et al., 1995). Gas pressures ini-tially exceed lithostatic in the upper third of atuff (Riehle et al., 1995). Trapped gas canform lithophysal cavities through matrix ex-pansion or form other gas-escape structuresalong pathways to the surface. Gas pressureswithin an ignimbrite deposit likely fall belowlithostatic within one month after emplace-ment (Riehle et al., 1995). Gases continue toevolve from the devitrification of volcanicglass through crystallization of anhydrousminerals (e.g., Smith, 1960), triggering vapor-phase crystallization in ignimbrites with ma-trix porosities greater than ;20% (Sheridan,1970; Ragan and Sheridan, 1972). Subsequentdevitrification and vapor-phase alteration con-tinue to temperatures as low as 240 8C. Cool-ing to this temperature in the upper third ofthe Tiva Canyon Tuff probably occurred sev-eral years after emplacement (i.e., Lofgren,1971; Riehle et al., 1995).

The timing of joint development can beevaluated by using features related to postde-positional cooling of the ignimbrite. Earliest-formed joints would be associated with gas-escape structures such as fumaroles, whichoccur with noticeable alteration mineralization(e.g., Sheridan, 1970; Hildreth, 1983; Keith,1991; Buesch et al., 1999). Inflationary struc-tures such as lithophysae also form in the ear-liest stage of degassing. Joints that form be-fore the lithophysae would seldom cut orterminate in lithophysae, may be bent by rheo-morphic deformation during lithophysae for-mation, and would show the effects of vapor-phase mineralization on the joint faces.Compaction and welding in ignimbrites anal-ogous to Tiva Canyon Tuff typically continuefor periods on the order of 1 to 10 yr (Fried-man et al., 1963; Riehle, 1973; Bierwirth,1982, cited in Cas and Wright, 1987; Riehle


DUNNE et al.

Figure 3. Fracture and joint trace maps(modified from Barton et al., 1993) forcleared pavements in upper lithophysae-bearing zone of the crystal-poor member ofthe Tiva Canyon Tuff (after Barton et al.,1993). (A) Pavement P100. (B) PavementP200. (C) Pavement P300. Pavement loca-tions in Figure 2A. Thick lines—tube-bear-ing joints; thin lines—other fractures;dashed line—perimeter of mapped pave-ment; medium gray regions—abundantsmall fracture traces; dot-patterned re-gions—fracture faces; arrowhead symbol—north direction.

N

et al., 1995). Devitrification and vapor-phasemineralization likely continue after compactionceases, but probably also stop about a centuryafter emplacement (e.g., Keith, 1991; Riehleet al., 1995). Joints that are post–lithophysaeformation but synmineralization will likely cutor terminate in lithophysae and also show ef-fects of devitrification and vapor-phase min-eralization in adjacent rock walls. Finally,joints may form through cooling-induced con-traction of the tuff to ambient temperatures (e.g.,DeGraff and Aydin, 1987), which can occurmore than a century after deposition without as-sociated mineralization or alteration.

JOINT GEOMETRY AND ORIGIN

Since being identified as a potential site forthe permanent disposal of high-level nuclearwaste, Yucca Mountain has been a focus ofseveral fracture analyses. Fractures have beeninvestigated at the surface (e.g., Morgan,1984; Barton et al., 1984, 1993; Barton andLarsen, 1985; Sweetkind et al., 1995a, 1995b;Throckmorton and Verbeek, 1995), in newtunnels (e.g., Albin et al., 1997; Eatman et al.,1997; Mongano et al., 1999), and in boreholes(e.g., Carr, 1992). The early surface analysesused the then-novel approach of clearingpavements of ;200 to 250 m2 (Fig. 3) to in-vestigate fractures as two-dimensional net-works (e.g., Barton et al., 1993) rather thanrely on scan-line or anecdotal station data(Hancock, 1985; La Pointe and Hudson,1985). The tunnel data set consists of a rig-orously collected scan line that sampled 15fracture attributes and a periphery map of thetunnel walls for fracture traces greater than 1m in length. It is the largest and most thor-ough data set concerning fractures at YuccaMountain (CRWMS M and O, 2000, and ref-erences therein). Tunnels sample almost theentire Paintbrush Group, whereas surface data



Figure 4. Sequence of formation of cooling joints with tubes. (A) Contractional joint sur-face propagates through rock prior to the formation of lithophysae, so no joints terminateor cut lithophysae. (B) Cooling joint causes formation of adjacent brittle zone. As the rockexpands vertically with the formation of lithophysae, subhorizontal, tube-like fracturesdevelop within the brittle zone. Also, large lithophysae deflect joints. (C) Continued de-gassing of the rock utilizes these tubes for gas flow (arrows show gas flow). (D) An ex-panded view of a cooling joint shows that the two sides are mirror images of each other.

are mostly from the Tiva Canyon Tuff, whichdominates the outcrop on the crest and easternflank of Yucca Mountain.

The consensus of the surface investigations,which was also applied to the subsurfacework, was that the natural fractures at YuccaMountain include joints that have three originswith distinct relative ages determined fromabutting relationships (Hancock, 1985; Sweet-kind and Williams-Stroud, 1996). Volcaniccooling joints are the oldest, largest, most pla-nar natural joints, and they commonly havetube structures along their walls (Barton et al.,1984; Sweetkind and Williams-Stroud, 1996;Buesch et al., 1999) (Figs. 1A, 1B, 1C, 4).Next, tectonic joints, which postdate coolingand have geometries resulting from stressfields controlled by regional stress conditions,abut the cooling joints, are planar to curvi-planar, are larger when fewer older coolingjoints are present, and are smooth to rough.Unloading joints, which are youngest andformed near the present ground surface in re-sponse to surficial unloading by erosion, havehackly, irregular surfaces and are typicallysubparallel to the ground surface.

A subset of the ‘‘cooling’’ joints as definedby previous workers are the focus of this study.We use the descriptive term ‘‘tube-bearing’’ torefer to these joints, because they have aunique geometry and because an origin ascontractional fractures formed during coolingis questioned here. To the best of our knowl-edge after an extensive literature search, tube-bearing joints have only been reported in ig-nimbrite deposits at Yucca Mountain, Nevada.If, as discussed subsequently, these jointsformed prior to lithophysae formation, thenthe apparent rarity of tube-bearing joints inwelded lithophysae-bearing ignimbrites mayreflect the difficulty of forming brittle frac-tures before lithophysae formation in the firstfew days after pyroclastic flow emplacement.

Across much of Yucca Mountain, tube-bearing joints occur in two subvertical orthog-onal sets that are normal to flow foliation andin a locally developed third set that is sub-horizontal and parallel to flow layering. Thisthird set is rare, and its joints tend to havecurviplanar rather than planar geometries,which could simply result from vertical dila-tion during cooling and degassing. Joints inthe third set are not the focus of this study(Throckmorton and Verbeek, 1995; Sweetkindand Williams-Stroud, 1996). In the centralblock of Yucca Mountain, the orthogonal setstypically strike northeast (Figs. 3A, 1C, 5) andnorthwest (Figs. 1D, 4C, 5). The relative ageof the two orthogonal sets is not well con-strained from field data, as only a few abuttingrelationships are found (Figs. 1D, 3). For ex-

ample, at pavements P100 and P300 (Fig. 3),more northwest-striking tube-bearing jointsterminate against northeast-striking joints thanvice versa, and almost all northeast-strikingjoints that terminate against northwest-strikingjoints are less than 2 m in trace length, whichis much shorter than the typical size of thesejoint traces. Although neither geometric aspectis conclusive, the joints support an interpre-tation in which the majority of the northeast-

striking, tube-bearing joints are older than thenorthwest-striking ones.

When considering the geometry of the or-thogonal system at the scale of Yucca Moun-tain, based on anecdotal field observations,some workers have suggested that the two setsform a rectilinear pattern of joint swarms orclusters with swarm widths of 3–5 m andspacings between swarms of 150–200 m (Bar-ton et al., 1993; Sweetkind and Williams-


DUNNE et al.

Figure 5. Map of Yucca Mountain to Fran Ridge (see Fig. 1B) region showing strikes oforthogonal tube-bearing joints (modified from Throckmorton and Verbeek, 1995) and thelocations of Figures 2A and 6. Dots on joint-trend tips are station locations.

Stroud, 1996). The present study attempts toresolve the importance of the orthogonal setsof tube-bearing joints in the overall fracturenetwork by considering their orientations overareas of thousands of square meters as op-posed to hundreds of square meters.

Timing of Tube-Bearing Joints andVolcanogenic Features

Tube-bearing joints at Yucca Mountainhave two characteristics that we interpret toindicate that joint formation preceded litho-physae formation. First, the fact that somejoint walls within ;1 cm of a lithophysa showoutward deformation due to inflation of thelithophysa indicates that the walls are older(Figs. 1E, 1F). Second, tube-bearing joints sel-dom intersect lithophysae, although lithophy-

sae are common and the tube-bearing jointsare large (Figs. 1A, 1C, 1D). If joint propa-gation occurred in the presence of lithophysae,intersections and terminations with lithophy-sae should occur (Barton et al., 1993; Bueschet al., 1996), partly because some jointsshould have propagated toward existing lith-ophysae in response to local stress perturba-tions around the expanding void (Kirsch,1898; Lachenbruch, 1962).

Tube-bearing joints also have two featuresthat we interpret to indicate that joint forma-tion preceded tube formation. First, tubes arefound only on joint faces and not in the rockmass, where lithophysae are present. Second,some tubes pass through joint intersections,which indicates that these intersections areolder.

The millimeter- to centimeter-diameter

tubes form anastomosing networks, which iso-late areas of joint planes (Figs. 1B, 1D, 1G).Area edges match across tube walls, indicatingthat the tubes dilated joint faces. Restorationto a predeformation configuration (Fig. 1G)shows that tubes produced at least 15% ver-tical dilation of the joint face, accompanied bya few degrees of local rotation for some areas.This magnitude of vertical dilation is consis-tent with vertical expansion in the ignimbriteunit of 16% to 22% due to lithophysae for-mation described by Barton et al. (1993) (Fig.4). Thus, the tubes are degassing structures(Barton et al., 1993) that formed coevally withthe lithophysae in response to vertical inflationof the ignimbrite. Because the tube-bearingjoints are older than both the lithophysae andthe tubes and because gas pressure must haveexceeded lithostatic pressure for their infla-tion, the time frame for their formation boundsthe timing of joint development in a lithophysae-free ignimbrite. Thus, joint formation likelyoccurred within a month of tuff emplacement(i.e., Riehle et al., 1995).

Abundant vapor-phase silicates line the tubeinteriors, and devitrification selvages com-monly extend for half a tube diameter into theignimbrite matrix. In contrast, joint planes be-tween tubes show only minor vapor-phasemineralization and very limited amounts ofdevitrification into the joint walls (Fig. 1G).Thus, hot gases flushed through the tubes rath-er than along the joint faces, leaving most ofthe vapor-phase precipitation products on thetube walls. Further cooling permitted escapeof the less saturated gas vertically along thejoint.

Structural Geometry of Tube-BearingJoints

A sample area of 20,787 m2 with ;15%exposure was examined on Live Yucca Ridgeat Yucca Mountain to characterize the contri-bution of the orthogonal joint system to thetotal fracture network (Figs. 5, 6). Subhori-zontal tube-bearing joints are not exposed inthis area, so their geometry is not considered.

Both fracture traces and limits of exposedrocks were mapped to consider the influenceof the distribution and quantity of rock ex-posure on trace distribution. The sample areais about two orders of magnitude larger thanthat used for the cleared pavements, such asP100 (214 m2; Fig. 3A), which is within thenew sample area (Fig. 6).

We used a real-time kinematic differentialGPS (DGPS) (Novatel) with resolution of 62cm horizontal and 610 cm vertical (2s error)to map the topography of Live Yucca Ridgeand to create a fracture trace map. Fracture



Figure 6. Map of Live Yucca Ridge, Yucca Mountain, showing joint traces, faults, and rock exposures (pavements). Topographic contourinterval is 10 m, and location grid is the UTM, Zone 11, NAD27 system. Letters in rectangles are locations referred to in the text.

positions were mapped by using the GPS datacollector’s ‘‘feature tagging’’ option, wherethe position is ‘‘tagged’’ with a user-definedalphanumeric identifier. Fractures were taggedat their visible ends and at several points inbetween if they exceeded a few meters inlength. Fractures were sequentially numberedand designated as either bearing or lackingtubes. In addition, fracture orientations weremeasured with a Brunton compass and inputinto the data collector. The tagged data pointswere then connected in a GIS (Arcview) tocreate a trace map of the fractures. Other fieldobservations were measured and recordedwith compasses, field notebooks, and cameras.

To facilitate useful, but efficient data col-lection over such a large sample area, the min-imum recorded trace length of the fractures(i.e., cutoff length) was 2 m, and the minimumexposed area mapped was approximatelyequal to a circle with a diameter of 2 m. Thetruncation limit is an order of magnitude larg-er than the 0.2 m limit used to survey fracturetraces on pavements such as P100 (Barton etal., 1993). We used a larger truncation limitbecause previous work and our reconnaissancemapping indicated that (1) tube-bearing jointsdominantly have trace lengths greater than 2m, (2) previous pavement work (Barton et al.,

1993) provided sample sets for the geometryof smaller fractures, and (3) previous investi-gations at Yucca Mountain documented thetremendous increase in time necessary to ac-curately document the patterns for tracelengths of ,1 m (Sweetkind and Williams-Stroud, 1996).

P100 and the related swarm of northeast-striking tube-bearing joints are in a relativelywell exposed part of the ridge. However, vi-sual inspection of the mapped locations oftube-bearing joints shows that their abundancedoes not correlate to areas of best exposure(Fig. 6).

The true maximum length of tube-bearingjoints is difficult to ascertain because theirtrace lengths are so long that they exceed thedimensions of even the largest exposures andhence are censored (joint tips are not exposed)(Epstein, 1954; Priest, 1993; Mauldon et al.,2001). In a few cases, joints were traced byusing small discontinuous exposures between2-m-diameter exposures to partly overcomethe effects of censoring and determine maxi-mum trace length. These traces are visible asblack lines that are mostly not contained inregions of exposure in Figure 6. Even withthese efforts, true fracture tips were seldomfound for tube-bearing joints, and we are re-

stricted to saying that many tube-bearing jointshave trace lengths exceeding 10 to 15 m.

The two sets of tube-bearing joints havedifferent trace distribution patterns. Northeast-trending traces are generally longer (apparentmean trace length of 4.1 m vs. 3.6 m) and areeither grouped in narrow swarms (locations Aand B, Fig. 6) or occur as a few joints in nearalignment (locations C and D, Fig. 6). In con-trast, northwest-striking tube-bearing jointsoccur in wider swarms and have traces thatare more typically en echelon rather than col-linear (locations E and F, Fig. 6). Also, north-west-striking joints are less numerous (67 vs.126) than northeast-striking joints (locationsA and B, Figs. 6 and 7A).

The two sets of joint swarms form an ap-proximately rectilinear network of traces (Fig.6), as described by Barton et al. (1993). Theorthogonal distance between the two well-developed northeast-striking joint swarms is;150 m (Barton et al., 1993). This value isprobably an overestimate of the swarm spac-ing, because between the two well-developednortheast-striking swarms are two poorly de-veloped swarms, particularly around locationC in Figure 6, yielding a minimum swarmspacing of ;50 m for the northeast-striking


DUNNE et al.

Figure 7. Orientation data for joints on Live Yucca Ridge. Left column: Lower-hemisphere, equal-area plots of poles to joints. Middlecolumn: Unweighted rose diagrams of joint traces. Plot radius 5 20%. Right column: Length-weighted rose diagrams of joint traces.Plot radius 5 20%. (A) Tube-bearing joints; number of joints 5 272; cumulative trace length 5 1069 m. (B) Joints without tubes;number of joints 5 214; cumulative trace length 5 737 m.

set. The northwest-striking swarms also havespacings of ;50 m.

Tube-bearing joints have a greater cumula-tive trace length with a narrower orientationrange than joints without tubes (Figs. 6, 7).Even with censoring, field observations dem-onstrate that the traces of tube-bearing joints,particularly northeast-striking ones, are longerthan the traces of joints without tubes and thattube-bearing joints are more numerous. Thus,the swarms of tube-bearing joints provide arectilinear framework to the fracture network,whereas joints without tubes enhance fracture-network connectivity.

Fumarolic Association for Tube-BearingJoints

Many ignimbrite deposits have early-formed joints (e.g., Sheridan, 1970; Hildreth,1983; Keith, 1991), and near-orthogonal jointsin ignimbrites have previously been recog-nized (Keith, 1991). Early-formed joints in theValley of Ten Thousand Smokes and BishopTuff (Hildreth, 1983; Keith, 1991; Sheridan,1970) are associated with fumarolic degassingand have obvious gas-escape features such as

mineralized ridges at the top of the depositand vapor-phase alteration of joint walls.These fumarolic joints have not been reportedwith tubes, but neither have they been inves-tigated in units with lithophysae (e.g., Sheri-dan, 1970; Hildreth, 1983; Keith, 1991). Fu-marolic joints in these ignimbrite depositsoccur along ridges that are underlain by singlelong joints or narrow swarms of joints (seeFig. 3A in Sheridan, 1970), which is consis-tent with the geometry of the tube-bearingjoints at Yucca Mountain. Fumarolic jointsprovide the vertical pathways for gas escapeduring ignimbrite compaction and welding,and their host ridges have lengths from as lit-tle as tens of meters to greater than 1 km. Asa result, the underlying swarms of fumarolicjoints would have the same lateral persistenceas the ridges. Fumarolic joints formed duringridge construction are driven both by the ther-moelastic contraction expected for simplecooling joints (Engelder and Fischer, 1996)and by elevated fluid pressure (Secor, 1968)due to degassing. Given that tube-bearingjoints at Yucca Mountain are geometricallysimilar to fumarolic joints, are the oldestjoints, and formed in a lithophysae-free ma-

trix, we interpret them to have been very earlygas-escape pathways that may have served asthe plumbing system for fumaroles at thesurface.

Possible Stress Sources for DevelopingSubvertical Tube-Bearing Joints

Because the Tiva Canyon Tuff was depos-ited in a region undergoing east-west exten-sion (Zoback et al., 1981; Sawyer et al., 1994;Morris et al., 1996), and under the assumptionthat the orthogonal joints are mode I extensionfractures with wall-normal displacements,they might be expected to strike north andeast. However, across much of the YuccaMountain region, the two tube-bearing setsstrike northeast and northwest (Figs. 3, 5, 6,7). Thus, at least one additional stress com-ponent other than the remote regional stressesmust have operated to perturb the stress fieldfrom the regional stress state. Such stress per-turbation might result from crustal subsidenceduring caldera formation or from active nor-mal faulting during or just after ignimbritedeposition.

During ignimbrite cooling, temperature gra-



dients are greater vertically than horizontally.Although the formation of the subverticaljoints could not be driven by tension relatedto the vertical temperature gradient, their for-mation could be partly explained by horizontaltemperature gradients (Lachenbruch, 1962),although one might not expect them to be uni-form in direction across a region. As an aside,the vertical temperature gradient, and conse-quent differential thermal contraction, mayhelp explain the stress conditions for the for-mation of the rare subhorizontal tube-bearingjoints, although gas pressure may have had arole as well.

Another cooling-related effect is differentialcompaction where thicker ignimbrite depositsremain hotter for longer periods and compactmore than thin deposits because of their high-er total heat capacity. This variation in thick-ness creates a lateral stress gradient. For ex-ample, one set of fumarolic ridges in a systemof orthogonal ridges in the Valley of TenThousand Smokes, Alaska, was interpreted toform in response to differential compactionduring cooling of the ignimbrite (Keith, 1991).As the ignimbrite deposit cooled, ignimbritein the valley center compacted more than thatat the margins, creating tensional fissures par-allel to these margins. In another example forthe general case of a simple basin, Sheridan(1970) proposed that vertical joint formationcould be driven by differential compaction be-tween the depocenter (i.e., thick deposits,more compaction) and the basin margins (i.e.,thin deposits, less compaction). For the TivaCanyon Tuff at Yucca Mountain, if the depo-center of a northwest-trending basin existedjust to the north of the study area (Fridrich,1999), the steepest compaction and tempera-ture gradient would be along the short axis ofthe basin toward the depocenter. Thus, thegreatest cooling-generated tension would beoriented northeast-southwest and would pos-sibly generate northwest-striking verticaljoints.

The eruption that formed the Tiva CanyonTuff involved ejection of ;1000 km3 of mag-ma and collapse of the Claim Canyon caldera;7 km north of the present study area (Fig.2B; Byers et al., 1976; Sawyer et al., 1994;Potter et al., 2002). Such a collapse wouldhave perturbed the regional stress field direct-ly and likely would have triggered movementon nearby normal faults (e.g., Nostro et al.,1998). For example, volcanic eruptions withless than 1% of this volume are interpreted totrigger earthquakes on normal faults in regionswith active extension (e.g., Abe, 1992; Nostroet al., 1998, and references cited within). Theinterval between the two events is typically 1

to 10 yr, but could be much shorter for a largeevent.

The major normal faults in the YuccaMountain region were active within the100,000 yr interval before and during the mil-lion years after deposition of the Tiva CanyonTuff. The fault activity was in response to re-gional extension in the central Basin andRange province and is documented by thick-ness changes in tuffs adjacent to major faults(Christiansen et al., 1977; Sawyer et al., 1994;Day et al., 1998a; Fridrich, 1999). Thus, re-gional tectonic stresses with or without per-turbation from caldera formation, plus theincreased lithostatic load from deposition ofthe Tiva Canyon Tuff, could have reasonablytriggered normal faulting during ignimbritecooling. If one or more of the normal faultsin the area slipped, the regional stress fieldwould be locally perturbed, which could ex-plain an orthogonal joint geometry that didnot match horizontal stress directions for theregional stress field alone (e.g., Crider andPollard, 1998; Kattenhorn et al., 2000).

Ideally, one would like to quantitatively testthese possibilities individually and together.However, lack of data, limited understandingof the material behavior of complexly coolingignimbrites, and a lack of robustness of cur-rent mechanical models to handle some ofthese uncertainties preclude rigorous testing.The role of differential compaction and ther-mal gradients cannot be tested, owing to thepaucity of data about depositional thickness,possible basin geometry, and lateral tempera-ture variations during emplacement and cool-ing. Analytical models have been developedfor modeling stress perturbation around small-volume magma systems, but not for stress-field perturbation effects from very large vol-ume (;1000 km3) eruptions. Nevertheless, theavailable data for the orientation and distri-bution of tube-bearing joints, for normal-faultgeometry, and for possible regional stress con-ditions during the Miocene at Yucca Mountaindo allow us to test whether interacting normalfaults could have locally perturbed the regionalstress field sufficiently to lead to the formationof the swarms of fumarolic tube-bearingjoints.

NORMAL FAULTING ANDPERTURBATION OF THE REGIONALSTRESS FIELD

We use a numerical model to test the fol-lowing proposition: active faulting could havelocally perturbed the stress field to produce ajoint system with strikes different from ex-pected in the Miocene regional stress field.The tool we use is Poly3D (Thomas, 1993), a

numerical code based on a three-dimensionalboundary-element method. Supplemental de-scription of the numerical modeling procedureand results are available.1 The code solves theelastostatic equations for the stress and defor-mation fields around composite surfaces ofdisplacement discontinuity (Crouch and Star-field, 1990; Comninou and Dundurs, 1975; Je-yakumaran et al., 1992). Poly3D has beenbenchmarked against two-dimensional ana-lytical solutions and shown to give resultsaccurate to within a few percent (Crider andPollard, 1998). Advantages to using theboundary-element method are that it is com-putationally efficient and that we can specifydiscontinuities, such as faults, of any shape.A disadvantage is that we are limited to test-ing the deformation of a homogeneous, isotro-pic, and isothermal medium.

Poly3D has been extensively tested andapplied in the study of the mechanics ofnormal faulting, including fault interaction re-lated to slip distribution (e.g., Willemse et al.,1996) and linkage (e.g., Crider, 2001); stress-triggering of earthquakes (e.g., Crider et al.,2001); and, most relevant to this work, faultcontrol of the orientation of secondary struc-tures, including joints (e.g., Kattenhorn et al.,2000) and faults (Maerten et al., 2002). Werefer the reader to these works and to Thomas(1993) for further details about the model.

We modeled the fault system in the vicinityof the field study area with a simplified ge-ometry and very simple material properties.Our goal was to evaluate the hypothesis thatfault slip during cooling of the Tiva CanyonTuff may have influenced the orientation ofthe tube-bearing joints, not to reproduce thecomplete natural system.

Model Configuration

We modeled the four major faults that occurnear the field study area: the Solitario Canyonfault, the Iron Ridge fault, the Paintbrush Can-yon fault, and the southern segments of theBow Ridge fault (Fig. 8). The fault traces havebeen mapped by numerous workers (e.g.,Scott and Bonk, 1984; Day et al., 1998a; Pot-ter et al., 2002). The fault geometry at depthis not as well constrained, but is reasonablyinterpreted to be steeply dipping to a depth of4 km before becoming listric (Young et al.,1992; Ferrill et al., 1996b). We modeled onlythose segments with significant present-daytopographic expression, and thus we do not

1GSA Data Repository item 2003168, supple-mental description of the numerical modeling pro-cedure and results, is available on the Web at http://www.geosociety.org/pubs/ft2003.htm. Requestsmay also be sent to [email protected].


DUNNE et al.

Figure 8. Perspective view of model fault geometry. SCF—Solitario Canyon fault (blue),IRF—Iron Ridge fault (green), BRF—Bow Ridge fault (brown), PCF—Paintbrush Canyonfault (purple). Arrows indicate the direction of maximum (SH) and minimum (Sh) hori-zontal stresses. Surface traces of the faults are coincident with the free surface of themodel elastic half-space. Half-space is semi-infinite; bounding box is shown to illustrateperspective only. Dimensions of the bounding box are 14 km 3 19 km 3 5 km (east,north, depth). Rectangle at surface shows approximate area of Figures 5 and 9A. Dotsshow calculation grid for principal stresses. Inset: Stress vs. depth plot illustrating model’sboundary conditions. The vertical principal stress (SV) is equal to the lithostatic load (L).Horizontal principal stresses deviate from lithostatic by amounts equal to the appliedremote stresses ( and ). For the results shown in Figure 9, 5 210 MPa and 5r r r rs s s s2 3 2 328 MPa.

include the east-dipping segment of the BowRidge fault where it crosses Yucca Wash. Themodeled fault geometry (Figs. 8 and 9A) usessimplified fault traces from within the regioncovered by Figure 5 and from Day et al.(1998a) for the area outside Figure 5. Themodeled faults dip west at 708 to a depthgreater than 4 km. Because our interest is invery near surface phenomena, the deep, listricparts of the faults were not included. Smallerintrablock faults were also not included, sothat we could focus on the generalized effectsof stress perturbation by the major structures.

For the purposes of this analysis, we as-sumed that the Tiva Canyon Tuff is mechan-ically coupled to the rocks below and that de-formation is influenced by the stress state ofthe entire rock volume containing the modeledfaults. The model faults are embedded in asemi-infinite elastic half-space with standardcrustal rheology (Poisson ratio 5 0.25, shearmodulus 5 30 GPa). Although we expect thecooling Tiva Canyon Tuff to have been sig-nificantly less stiff, it represents only a small

volume of the modeled crustal section. Thus,we choose rheology appropriate to the largervolume. Model trials with order-of-magnitudevariation in shear modulus (3 GPa to 100GPa) show no difference in the resulting ori-entations of local stress directions (see foot-note 1). The faults intersect the free surface ofthe half-space and are allowed to slip freely(no friction). Thus, modeled faults will slip inresponse to any magnitude of differentialstress, and the stress drop across slipped faultsis complete.

Although it is not possible to obtain directmeasurements of the stress state in the Mio-cene, we assumed that deformation occurredin a normal-faulting stress regime with maxi-mum principal compressive stress vertical(SV), intermediate principal stress horizontaland north-south (SH), and minimum principalcompressive stress horizontal and east-west(Sh). This assumption is consistent with geo-logic estimates of the Miocene stress state, asinterpreted by Zoback et al. (1981). Modernvalues for the principal stresses provide a

framework for values used in the model. Fer-rill et al. (1999b) determined fluid-pressure–corrected principal-stress values at 1 km depthon the basis of stress measurements by Stocket al. (1985). These are SV 5 21 MPa (ap-proximately lithostatic), SH 5 17 MPa (4 MPaless than lithostatic), and Sh 5 11 MPa (10MPa less than lithostatic). By using the Mio-cene stress directions and the modern valuesas a guide, we applied to the model a linearlyincreasing lithostatic stress (L, equivalent tothe weight of overlying rock) acting in all di-rections (SV 5 L). To drive deformation, wereduced the horizontal stress in the north-south and east-west directions (Fig. 8 inset),as is observed in the modern stress data. Be-cause we were testing the orientation and notthe magnitude of Miocene stresses, the magni-tude of the boundary stresses is not as importantas the ratios among them. Several variationswere tested. We report one here, and others areavailable to the reader (see footnote 1).

Model Results and Interpretation

With the boundary conditions just de-scribed, the modeled faults slip simultaneous-ly, with maximum dip slip at the free surfaceand approximately at the center of the faulttrace (see footnote 1). Slip distribution alongeach fault is approximately elliptical, modifiedby interaction among the faults. The faultsalso show small components of strike slipwhere the strike of segments deviates fromnorth. For Sh equivalent to modern values, themean dip slip on the faults is less than 1 m.The modeled situation corresponds to a clusterof moderate-sized (Mw 5 5 to 6) earthquakes,such as is common in western North America,e.g., the 1993 Klamath Falls, Oregon, earth-quakes (Braunmiller et al., 1995), the Ridge-crest, California, sequence (Hauksson et al.,1995), or the much larger Dixie Valley–Fairview Peak, Nevada, sequence (Doser,1986).

In Figure 9, we present results for remotedriving stresses Sh 5 L 2 10 MPa (east-west)and SH 5 L 2 8 MPa (north-south). The re-sults are calculated at a preerosion depth of100 m. Vertical tensile fractures are expectedto form perpendicular to the local least hori-zontal stress. Outside the faulted region, adominant north-south fracture pattern perpen-dicular to the regional Sh is predicted (seefootnote 1). Near the faults, local perturbationof the principal stress orientations and conse-quent rotation of predicted fracture orienta-tions is observed. Figure 9 shows local ori-entations for Sh and predicted fractureorientations in the vicinity of Live YuccaRidge. These compare favorably to the ob-



Figure 9. Modeled stress trajectoriesand predicted joint orientations com-pared to observed structures. (A) Mod-eled fault traces and stress trajectoriesoverlaid on joint orientations fromThrockmorton and Verbeek (1995).Compare to Figure 5. Arrows show thedirections of Sh, and short lines show thecorresponding predicted joint orienta-tions. Other symbols as in Figure 5. Out-side the faulted region, east-west andnorth-south joint orientations are pre-dicted. In the block between the Solitar-io and Bow Ridge faults, joint orienta-tions are deflected to the northeast, incorrespondence with field observations.Arc shows 2 km radius from Live YuccaRidge and encircles modeled joint ori-entations plotted in B. (B) Rose diagramof predicted first-formed joint orienta-tions within 2 km of Live Yucca Ridge.Compare to Figure 7. (C) Angular mis-match between strike of observed jointsand predicted joint orientations. In thevicinity of Live Yucca Ridge, averagemismatch between the modeled jointsand observed northeast-striking set is128. Average mismatch between themodeled and observed northwest-striking set is 758. East of PaintbrushCanyon fault, the average mismatch be-tween the model and observations is 198for east-striking joints and 698 fornorth-striking joints.


DUNNE et al.

served orientations of the northeast-strikingtube-bearing joints. Within 2 km of Live Yuc-ca Ridge, the model predicts joint strike pri-marily between 0208 and 0708, with a meanstrike at 0518 and a mode at ;0408 (Fig. 9B).These predicted values compare favorablywith observed orientations of tube-bearingjoints. In the crustal block between the Soli-tario Canyon and Bow Ridge faults, predictedjoints are on average within 128 of the ob-served northeast-striking set and 758 divergentfrom the northwest-striking set. At Live YuccaRidge, the mismatch is 18 and 888. In the foot-wall of the Paintbrush Canyon fault, north-and east-striking joints are observed (Throck-morton and Verbeek, 1995), and the modelpredicts similar joint orientations.

Our model results show that slip on faultscan generate local stress perturbations that areconsistent with the orientations of northeast-striking joints observed at Yucca Mountain. Weemphasize that the model was not designed tobe an inversion of the joint-orientation data,nor was it tuned for best-fit results. We usethe model to test the proposition that fault slip,driven by geologically reasonable boundaryconditions, could have influenced the localjoint orientations observed to be differentfrom those expected to be produced by theMiocene regional stress field. The close cor-respondence between this simple mechanicalmodel and the observed northeast-strikingjoints suggests that active faulting synchro-nous with the emplacement of the Tiva Can-yon Tuff is a plausible mechanism for con-trolling the orientation of tube-bearing jointswithin the ignimbrite. Further, the results sug-gest that the northeast-striking joints wereformed before the northwest-striking set, as themodeled stress trajectories produce northeast-striking joints without the influence of thenorthwest-striking set.

FORMATION OF THE NORTHWEST-STRIKING JOINT SWARMS

If it is assumed that the northeast-strikingset of tube-bearing joints in the Tiva CanyonTuff was formed during volcanic degassingand active normal faulting, what was themechanism that formed the second orthogonalset with the necessary 908 switch of the hor-izontal stress directions? The likely possibili-ties are a regional-scale switch of stress direc-tions, changes in displacement behavior of thenormal faults, and more local-scale switchingdue to formation of the first-formed northeast-striking joints. Any of these mechanismscould have been aided by differential compac-tion in the postulated northwest-trendingbasin.

If a regional-scale switch is considered, thecentral Basin and Range in which YuccaMountain lies, underwent east-west extensionprior to 10 Ma (Zoback et al., 1981). Previouswork (Zoback et al., 1981; Wernicke, 1992;Axen et al., 1993) does not indicate that achange occurred in regional principal straindirections and, hence, regional principal stressdirections during this east-west extension.Consequently, we discount the possibility of aregional-scale switch of tectonic stresses.

Considering displacement behavior on thenormal faults, horizontal stress directionsmight switch if fault-slip directions changedsignificantly or if different faults or parts offaults slipped sequentially. We cannot com-pletely discount this possibility, although di-agnostic evidence for these scenarios is lack-ing for the joint system under investigation.

The possibility of a locally controlled switchdue to an additional perturbation of the stressfield by the formation of the northeast-strikingjoints remains. The fact that the northeast-striking joint swarms are better developedthan the northwest-striking swarms in terms ofjoint size and abundance within swarms pro-vides a geometric support for this possibility.The better development of the older set is atypical feature of orthogonal joint systemsformed by local stress release (Hancock, 1985;Caputo, 1995; Bai et al., 2002).

The case for local control of the switch ofthe stress directions is analogous to the for-mation of strata-bound layer-normal jointsforming in response to remote tension (Hobbs,1967; Gross et al., 1995). In this case, jointformation perturbs the stress field by causinga stress drop or shadow. Recent work by Baiet al. (2001) shows that when joint spacing ofthe older joint set is ,1.7 times the thicknessof the layer containing the joints and whenSh/SH . 0.2, the horizontal stress directions inthe rock mass between the first-formed jointsswitch. Given an interpreted penetration depthfor the northeast-striking joints of 20 to 30 m,which is the thickness of the upper lithophysae-bearing zone, and a swarm spacing of ;50 m,the spacing to thickness ratio in the Tiva Can-yon Tuff is 2.5 to 1.6, which is about the crit-ical threshold of 1.7 for stress switching whenthe ratio of horizontal stresses is .0.2, whichseems likely in this case, on the basis of thelikely stress conditions for fault slip and stressformation. Thus, a second perturbation of thestress field in the vicinity of Live Yucca Ridgeby the formation of the swarms of northeast-striking tube-bearing joints is a possiblemechanism for rotating the horizontal stressdirections to allow the formation of northwest-striking joints.

Care should be taken in applying this anal-

ogy to orthogonal swarm formation at YuccaMountain. The analogous case is applied tofracture networks of individual joints in brittlerocks with spacings on the order of centime-ters to meters, whereas the Yucca Mountaincase considers swarms of joints on the scaleof tens of meters that developed in a cooling,but hot, ignimbrite with a more complex ma-terial behavior. The analogous case relies onuplift and erosion to achieve near-surface con-ditions for formation of the cross joints,whereas at Yucca Mountain the process oc-curred quite close to the ground surface afterminimal burial. Also, whereas joint formationat Yucca Mountain is partly driven externallyby thermoelastic contraction due to cooling,the joint formation is also internally driven byfluid pressure generated by volcanic gases,which differs from the analogy. Still, the anal-ogy does offer an explanation. Numericalmodeling has the potential to allow the explo-ration of the possibility of extrapolating thisexplanation to the larger-scale case of jointswarms partly driven by internal fluid pres-sure. In fact, the fluid-driven case may offeran explanation for some cases of joint swarmdevelopment, which is a widely recognizedphenomenon (e.g., Hancock, 1985; Pollardand Aydin, 1988; Odling et al., 1999).

DISCUSSION

Cooling Joints Without Tubes

Cooling joints are ubiquitous features inmoderately to densely welded ignimbrite de-posits (e.g., Smith, 1960; Ross and Smith,1961; Cas and Wright, 1987). They form inresponse to tensional stress induced by ther-mal contraction during cooling (Lachenbruch,1962; DeGraff and Aydin, 1987; Engelder andFischer, 1996) and are not associated withductile deformation features such as tubes orlithophysae (Enlows, 1955; Ross and Smith,1961; Sheridan, 1970; Cas and Wright, 1987).Cooling joints are pseudocolumnar, smooth,or curviplanar; lack bleached walls; occur ina variety of orientations; and vary in lengthfrom 1 to 15 m. They also occur in ignimbriteswith fumarolic joints and terminate againstthose early-formed joints (e.g., Sheridan,1970).

By analogy with other welded ignimbrites,abundant cooling joints due to thermoelasticcontraction should be expected in the rockvolumes between early-formed fumarolicjoints at Yucca Mountain (e.g., Sheridan,1970). Such cooling joints have been recog-nized by previous workers at Yucca Mountain(Morgan, 1984; Barton et al., 1984, 1993;Barton and Larsen, 1985; Carr, 1992; Sweet-



kind et al., 1995a, 1995b; Throckmorton andVerbeek, 1995; Albin et al., 1997; Eatman etal., 1997; Mongano et al., 1999). Thus, thelithophysae-bearing part of the ignimbritecontains two systems of joints with origins re-lated to lithification and cooling of the depos-it: the fumarolic tube-bearing joints and theyounger cooling joints without tubes formedby thermoelastic contraction.

Distinguishing cooling joints from youngertectonic joints is problematic when only geo-metric characteristics are used. Previous work-ers have identified a series of characteristicssuch as size, planarity, and the smoothness offracture surfaces to distinguish cooling fromtectonic joints. These characteristics do distin-guish the relative age of two joint sets withouttubes, but they do not conclusively demon-strate that either set is of cooling origin. For-tunately, for subsurface joints and surfacejoints with limited to absent weathering, cool-ing joints can be identified by using the dif-ferent deuteric alteration facies within the TivaCanyon Tuff (cf. Buesch et al., 1996). Coolingjoint attributes, such as fill compositions andbleached rims, should correlate with degree ofdegassing, welding, and devitrification pro-cesses in the rock mass, which are attributesnot associated with tectonic joint formation(Sweetkind et al., 2003). If the relative abun-dance of cooling and tectonic joints in theTiva Canyon Tuff is deemed important to un-derstanding performance of the proposed re-pository, the previous assignment of a coolingor tectonic origin to joints on the basis of geo-metric characteristics would benefit from acareful reexamination of the joints for attri-butes indicative of ignimbrite cooling.

Implications for Joint Geometry in OtherLithophysae-Bearing Units at YuccaMountain

The focus of this paper is on tube-bearingjoints in the upper lithophysae-bearing zone ofthe crystal-poor member of the Tiva CanyonTuff. The results, however, have direct rele-vance to joints in other lithophysae-bearingparts of the Tiva Canyon and Topopah SpringTuffs. The best data set for evaluating the dis-tribution and characteristics of tube-bearingjoints within Yucca Mountain is the scan-lineand full-periphery mapping data from the ESFand ECRB tunnels (Mongano et al., 1999;CRWMS M and O, 1998, 2000). Tube-bearingjoints occur in the Tiva Canyon and TopopahSpring Tuffs sampled in the ESF and ECRBtunnels at Yucca Mountain. These joints moretypically have a northwest strike rather than anortheast strike and, in several exposures, oc-cur as closely spaced swarms (CRWMS M

and O, 1998, 2000; Mongano et al., 1999).Therefore, these subsurface units contain pre-lithophysae joints in the vicinity of the pro-posed repository.

The common understanding of the joints thatform after lithophysae in lithophysae-bearingunits is that those joints are generally not per-sistent and have a curviplanar form (Sweetkindand Williams-Stroud, 1996; CRWMS M andO, 1998, 2000; Mongano et al., 1999; Sweet-kind et al., 2003). This understanding is quiteappropriate for the cooling, tectonic, and un-loading joints that postdate lithophysae for-mation because lithophysae arrest fracturepropagation, inhibiting the development oflarge, throughgoing joints. However, the tube-bearing joints are prelithophysae and have ex-tensive individual and swarm geometry in theupper lithophysae-bearing zone of the TivaCanyon Tuff (Fig. 6). Tube-bearing joints inthe subsurface would be expected to have thesame extensive individual and swarm geom-etries, which would influence the mechanicaland hydrologic properties of the lithophysae-bearing rock mass. Whereas much of thelithophysae-bearing rock volume will be pop-ulated with discontinuous curviplanar frac-tures as previously described, locally exten-sive joint swarms will transect the volume,possibly creating fluid pathways and the po-tential for significant large joint-boundedblocks. Therefore, when considering the roleof joints in a fracture characterization of alithophysae-bearing unit, a distinction needs tobe made between the case of a rock volumewith and one without tube-bearing joints.

CONCLUSIONS

1. The oldest joints in the welded Tiva Can-yon Tuff at Yucca Mountain are tube bearingand consist of two orthogonal sets. They werepreviously interpreted to be related to ignim-brite degassing and dilation during cooling.Given that such processes likely occurredwithin one month of ignimbrite deposition, wethink that the entire orthogonal joint systemformed within that time frame.

2. Our proposed interpretation for the for-mation history of tube-bearing joints is con-sistent with field data and model results, in-cluding (a) initial pyroclastic eruption of theTiva Canyon Tuff and collapse of the ClaimCanyon caldera; (b) deposition of the ignim-brite over topography that may have includeda shallow northwest-trending basin; (c) devel-opment of a perturbed stress field due to acombination of thermal gradient, differentialcompaction, caldera collapse, and faulting; (d)slip on the Solitario Canyon, Bow Ridge, andPaintbrush Canyon faults, perhaps triggered

by caldera collapse and increased verticalloading associated with the newly depositedtuff sheet, is a strong candidate for perturbingthe regional stress field; (e) formation of thenortheast-striking joints with orientation con-trolled by the perturbed stress field in responseto gas pressure during fumarolic activity; (f)formation of the northwest-striking joints con-trolled by a stress switch due to the presenceof the extensive northeast-striking jointswarms in response to gas pressure and pos-sibly aided by differential compaction acrossthe northwest-trending basin; and (g) tube for-mation on both joint sets during degassing andlithophysae formation in the upper Tiva Can-yon Tuff. Modeling of fault-controlled stress-field perturbation does not completely excludeother possible contributions to perturbation ofthe local stress field. For example, with ap-propriate data, the roles of differential com-paction and thermal gradients could be explic-itly assessed. Similarly, a better understandingof the role of caldera collapse in perturbingstress fields and triggering fault motions mightprovide a sufficient basis for a quantitativeanalysis of this possibility. Nonetheless, wethink that this contribution represents a usefulfirst attempt to relate the roles of caldera col-lapse, ignimbrite deposition and cooling, andfault behavior and joint development to ex-plain the origin of a joint network with a veryinteresting formation history. This interpreta-tion also illustrates that a series of stresssources acting at a variety of scales, timespans, and magnitudes can combine to pro-duce an apparently simple orthogonal jointgeometry.

3. This new interpretation subdivides pre-viously interpreted cooling joints into prelith-ophysae tube-bearing joints related to gas es-cape and postlithophysae joints without tubesrelated to thermoelastic contraction. Conclu-sive identification of contractional coolingjoints depends on relating fracture attributes toprocesses during ignimbrite lithification.

4. Our conclusions regarding very early de-velopment of tube-bearing joints applies tosuch joints throughout the Paintbrush Group.Tube-bearing joints change the characteris-tics of the fracture geometry in a volume oflithophysae-bearing rock by adding persis-tent, planar, closely spaced fractures with pre-ferred orientations to the system. Northwestand northeast strikes of joint sets in both theTiva Canyon and Topopah Spring Tuffs indi-cate that stress-field perturbation was a recur-rent process during the deposition and coolingof both welded ignimbrite units of the Paint-brush Group.


DUNNE et al.

ACKNOWLEDGMENTS

This paper was prepared to document work per-formed by the Center for Nuclear Waste RegulatoryAnalyses (CNWRA) for the U.S. Nuclear Regula-tory Commission (NRC) under Contracts Nos.NRC-02-97-009 and NRC-02-02-012. The activitiesreported here were performed on behalf of the NRCOffice of Nuclear Material Safety and Safeguards,Division of Waste Management. The report is anindependent product of the CNWRA and does notnecessarily reflect the views or regulatory positionof the NRC. We thank James P. Hogan for assis-tance in the field and Rebecca Emmot for assistancewith manuscript preparation. Acknowledgment ismade to the Rock Fracture Project, Stanford Uni-versity, for use of Poly3D and Poly3DGUI. Criticalreviews by Terry Engelder, James P. Evans, PhilipJustus, Wesley C. Patrick, Chris Potter, LarryMcKague, and Wanda Taylor greatly improved themanuscript.

REFERENCES CITED

Abe, K., 1992, Seismicity of the caldera-making eruptionof Mount Katmai, Alaska, in 1912: Bulletin of theSeismological Society of America, v. 82, p. 175–191.

Albin, A.L., Singleton, W.L., Moyer, T.C., Lee, A.C., Lung,R.C., Eatman, G.L.W., and Barr, D.L., 1997, Geologyof the Main Drift—Station 28 1 00 to 55 1 00, Ex-ploratory Studies Facility, Yucca Mountain Project,Yucca Mountain, Nevada: Summary Report by theU.S. Bureau of Reclamation to the U.S. Departmentof Energy, DTN GS970208314224. 005.

Axen, G.J., Taylor, W.J., and Bartlet, J.M., 1993, Space-time patterns and tectonic controls of Tertiary exten-sion and magmatism in the Great Basin of the westernUnited States: Geological Society of America Bulle-tin, v. 105, p. 56–76.

Bai, T., Maerten, L., Gross, M.R., and Aydin, A., 2002,Orthogonal cross joints: Do they imply a regionalstress rotation?: Journal of Structural Geology, v. 24,p. 77–88.

Banks, N.G., and Hoblitt, R.P., 1981, Summary of temper-ature studies of 1980 deposits, in Lipman, P.W., andMullineaux, D.R., ed., The 1980 eruptions of MountSt. Helens, Washington: U.S. Geological Survey Pro-fessional Paper 1250, p. 295–314.

Barton, C.C., and Larsen, E., 1985, Fractal geometry oftwo-dimension fracture networks at Yucca Mountain,in Stephanson, O., ed., Proceedings of the Internation-al Symposium on Fundamentals of Rock Joints, Swe-den: Lulea, Sweden, Centek Publishers, p. 77–84.

Barton, C.C., Howard, T.M., and Larsen, E., 1984, Tubularstructures on the faces of cooling joints—A new vol-canic feature: Eos (Transactions, American Geophys-ical Union), v. 65, p. 1148.

Barton, C.C., Larsen, E., Page, W.R., and Howard, T.M.,1993, Characterizing fractured rock for fluid-flow,geomechanical, and paleostress modeling: Methodsand preliminary results from Yucca Mountain, Neva-da: U.S. Geological Survey Open-File Report 93-269,31 p.

Bierwirth, P.N., 1982, Experimental welding of volcanicash [B.S. honors thesis]: Victoria, Australia, MonashUniversity, 150 p.

Braunmiller, J., Nblek, J., Leitner, B., and Qamar, A., 1995,The 1993 Klamath Falls, Oregon, earthquake se-quence: Source mechanisms from regional data: Geo-physical Research Letters, v. 22, p. 105–108.

Buesch, D.C., and Spengler, R.W., 1998, Character of themiddle nonlithophysal zone of the Topopah SpringTuff at Yucca Mountain: High level radioactive wastemanagement, Proceedings of the 1998 InternationalConference: La Grange Park, Illinois, American Nu-clear Society, p. 16–23.

Buesch, D.C., Spengler, R.W., Moyer, T.C., and Geslin,J.K., 1996, Proposed stratigraphic nomenclature andmacroscopic identification of lithostratigraphic unitsof the Paintbrush Group exposed at Yucca Mountain,

Nevada: U.S. Geological Survey Open-File Report94-469, 45 p.

Buesch, D.C., Beason, S.C., and Spengler, R.W., 1999, Re-lations among welding, vapor-phase activity, crystal-lization, and fractures in the Tiva Canyon and Topo-pah Spring Tuffs, at Yucca Mountain: GeologicalSociety of America Abstracts with Programs, v. 31,no. 7, p. A476–A477.

Bursik, M.I., and Woods, A.W., 1996, The dynamics andthermodynamics of large ash flows: Bulletin of Vol-canology, v. 58, p. 175–193.

Byers, F.M., Jr., Carr, W.J., Orkild, P.P., Quinlivan, W.D.,and Sargent, K.A., 1976, Volcanic suites and relatedcauldrons of the Timber Mountain–Oasis Valley cal-dera complex: U.S. Geological Survey ProfessionalPaper 919, 70 p.

Caputo, R., 1995, Evolution of orthogonal sets of extensionjoints: Terra Nova, v. 7, p. 479–490.

Carr, W.J., 1992, Structures in continuously cored, deepdrill holes at Yucca Mountain, Nevada, with notes oncalcite occurrence: Los Alamos, New Mexico, SandiaNational Laboratories, SAND91-7037, 39 p.

Cas, R.A.F., and Wright, J.V., 1987, Volcanic succes-sions: Winchester, Massachusetts, Unwin Hyman Inc,p. 256–258.

Christiansen, R.L., Lipman, P.W., Carr, W.J., Byers, F.M.,Jr., Orkild, P.P., and Sargent, K.A., 1977, TimberMountain–Oasis Valley caldera complex of southernNevada: Geological Society of America Bulletin,v. 88, p. 943–959.

Comninou, M., and Dundurs, J., 1975, The angular dislo-cation in a half space: Journal of Elasticity, v. 5,p. 205–216.

Crider, J.G., 2001, Oblique slip and the geometry of nor-mal-fault linkage: Mechanics and a case study fromthe Basin and Range in Oregon: Journal of StructuralGeology, v. 23, p. 1997–2009.

Crider, J.G., and Pollard, D.D., 1998, Fault linkage: Three-dimensional mechanical interaction between echelonnormal faults: Journal of Geophysical Research,v. 103, p. 24,373–24,391.

Crider, J.G., Schaff, D.P., Pollard, D.D., and Beroza, G.C.,2001, Considering the third dimension in stress-trig-gering of aftershocks: 1993 Klamath Falls (OR) earth-quake sequence: Geophysical Research Letters, v. 28,p. 2739–2742.

Crouch, S.L., and Starfield, A.M., 1990, Boundary elementmethods in solid mechanics: Boston, Unwin Hyman,237 p.

CRWMS M and O (Civilian Radioactive Waste Manage-ment System Management and Operating Contractor),1998, Geology of the Exploratory Studies Facility—Topopah Spring Loop: Las Vegas, Nevada,BAB000000-0117-0200-00002 REV 01, 44 p.

CRWMS M and O (Civilian Radioactive Waste Manage-ment System Management and Operating Contractor),2000, Fracture geometry analysis for the stratigraphicunits of the repository host horizon: Las Vegas, Ne-vada, ANL-EBS-GE-000006 REV 00, 62 p.

Day, W.C., Dickerson, R.P., Potter, C.J., Sweetkind, D.S.,San Juan, C.A., Drake, R.M., II, and Fridrich, C.J.,1998a, Geologic map of the Yucca Mountain area,Nye County, Nevada: U.S. Geological Survey Mis-cellaneous Geologic Investigations Series Map I-2627,scale 1:24,000, 1 sheet.

Day, W.C., Potter, C.J., and Sweetkind, D., 1998b, Bedrockgeologic map of the central block area, Yucca Moun-tain, Nye County, Nevada: U.S. Geological SurveyMiscellaneous Geologic Investigation Series Map I-2601, scale 1:6000, 2 sheets.

DeGraff, J.M., and Aydin, A., 1987, Surface morphologyof columnar joints and its significance to mechanicsand direction of joint growth: Geological Society ofAmerica Bulletin, v. 99, p. 605–617.

Doser, D.I., 1986, Earthquake processes in the RainbowMountain–Fairview Peak–Dixie Valley, Nevada, re-gion 1954–1959: Journal of Geophysical Research,v. 91, p. 12,572–12,586

Dunne, W.M., and North, C.P., 1990, Orthogonal fracturesystems at the limits of thrusting: An example fromsouthwestern Wales: Journal of Structural Geology,v. 12, p. 207–215.

Eatman, G.L.W., Singleton, W.L., Moyer, T.C., Barr, D.L.,

Albin, A.L., Lung, R.C., and Beason, S.C., 1997, Ge-ology of the South Ramp—Station 55 1 00 to 78 177, Exploratory Studies Facility, Yucca MountainProject, Yucca Mountain, Nevada. MilestoneSPG42CM3: Denver, Colorado, U.S. Geological Sur-vey, 458 p.

Engelder, T., 1985, Loading path to joint propagation duringcycle: An example from the Appalachian Plateau,USA: Journal of Structural Geology, v. 7, p. 459–476.

Engelder, T., and Fischer, M.P., 1996, Loading configura-tions and driving mechanisms for joints based on theGriffith energy-balance concept: Tectonophysics,v. 256, p. 253–277.

Enlows, H.E., 1955, Welded tuffs of the Chiricahua Na-tional Monument, Arizona: Geological Society ofAmerica Bulletin, v. 66, p. 1215–1246.

Epstein, B., 1954, Truncated life tests in the exponentialcase: Annals of Mathematics and Statistics, v. 25,p. 555.

Eyal, Y., and Reches, Z., 1983, Tectonic analysis of theDead Sea rift region since the Late Cretaceous basedon mesostructures: Tectonics, v. 2, p. 167–185.

Fabbri, O., Gaviglio, P., and Gamond, J.-F., 2001, Diach-ronous development of master joints of different ori-entations in different lithological units within the samefore-arc basin deposits, Kyushu, Japan: Journal ofStructural Geology, v. 23, p. 239–246.

Ferrill, D.A., and Morris, A.P., 2001, Displacement gradientand deformation in normal fault systems: Journal ofStructural Geology, v. 23, p. 619–638.

Ferrill, D.A., Stamatakos, J.A., Jones, S.M., Rahe, B.,McKague, H.L., Martin, R.H., and Morris, A.P.,1996a, Quaternary slip history of the Bare Mountainfault (Nevada) from the morphology and distributionof alluvial fan deposits: Geology, v. 24, p. 559–562.

Ferrill, D.A., Stirewalt, G.L., Henderson, D.B., Stamatakos,J.A., Morris, A.P., Wernicke, B.P., and Spivey, K.H.,1996b, Faulting in the Yucca Mountain region: Criti-cal review and analyses of tectonic data from the cen-tral Basin and Range: U.S. Nuclear Regulatory Com-mission, CNWRA 95-017, NUREG/CR-6401, 92 p.

Ferrill, D.A., Stamatakos, J.A., and Sims, D., 1999a, Nor-mal fault corrugation: Implications for growth andseismicity of active normal faults: Journal of Struc-tural Geology, v. 21, p. 1027–1038.

Ferrill, D.A., Winterle, J., Wittmeyer, G., Sims, D., Colton,S., Armstrong, A., and Morris, A.P., 1999b, Stressedrock strains groundwater at Yucca Mountain, Nevada:GSA Today, v. 9, no. 5, p. 1–8.

Flint, L.E., and Flint, A.L., 1995, Shallow infiltration pro-cesses at Yucca Mountain, Nevada—Neutron loggingdata 1984–93: U.S. Geological Survey Water-Re-sources Investigations Report 95-4035, 46 p.

Flint, A.L., Hevesi, J.A., and Flint, L.E., 1996, Conceptualand numerical model of infiltration for the YuccaMountain area, Nevada, Yucca Mountain Milestone3GU1623M, Denver, Colorado: U.S. Geological Sur-vey, 174 p.

Fridrich, C.J., 1999, Tectonic evolution of the Crater Flatbasin, Yucca Mountain region, Nevada, in Wright, L.,and Troxel, B., eds., Cenozoic basins of the DeathValley region: Geological Society of America SpecialPaper 333, p. 169–195.

Friedman, I., Long, W., and Smith, R.L., 1963, Viscosityand water content of rhyolite glass: Journal of Geo-physical Research, v. 68, p. 6522–6535.

Frizzell, V.A., and Shulters, J., 1990, Geologic map of theNevada Test Site, southern Nevada: U.S. GeologicalSurvey Miscellaneous Geologic Investigations SeriesMap I-2046, 1:100,000 scale.

Gross, M.R., 1993, The origin and spacing of cross joints:Examples from the Monterey Formation, Santa Bar-bara coastline, California: Journal of Structural Geol-ogy 15, p. 737–751.

Gross, M.R., Fischer, M.P., Engelder, T., and Greenfield,R.J., 1995, Factors controlling joint spacing in inter-bedded sedimentary rocks: Integrating numericalmodels with field observations from the MontereyFormation, U.S.A., in Ameen, M.S., ed., Fractogra-phy: Fracture topography as a tool in fracture me-chanics and stress analysis: Geological Society [Lon-don] Special Publication 92, p. 215–233.

Hancock, P.L., 1985, Brittle microtectonics: Principles and



practice: Journal of Structural Geology, v. 7,p. 437–457.

Hauksson, E., Hutton, K., Kanamori, H., Jones, L., Mori,J., Hough, S., and Roquemore, G., 1995, Preliminaryreport on the 1995 Ridgecrest earthquake sequence ineastern California: Seismological Research Letters,v. 66, p. 54–60.

Hildreth, W., 1983, The compositionally zoned eruption of1912 in the Valley of Ten Thousand Smokes, KatmaiNational Park, Alaska: Journal of Volcanology andGeothermal Research, v. 18, p. 1–56.

Hobbs, D.W., 1967, The formation of tension joints in sed-imentary rocks: An explanation: Geological Maga-zine, v. 104, p. 550–556.

Jeyakumaran, M., Rudnicki, J.W., and Keer, L.M., 1992,Modeling of slip zones with triangular dislocation el-ements: Bulletin of the Seismological Society ofAmerica, v. 82, p. 2153–2169.

Kattenhorn, S.A., Aydin, A., and Pollard, D.D., 2000,Joints at high angles to normal fault strike: An expla-nation using 3-D numerical models in fault-perturbedstress fields: Journal of Structural Geology, v. 22,p. 1–24.

Keith, T.E.C., 1991, Fossil and active fumaroles in the 1912eruptive deposits, Valley of Ten Thousand Smokes,Alaska: Journal of Volcanology and Geothermal Re-search, v. 45, p. 227–254.

Kirsch, G., 1898, Die theorie der elastizitat und die be-durfnisse der festigheitslehre: Zeitschrift des VereinesDeutscher Ingenieure, v. 42, p. 797.

La Pointe, P.R., and Hudson, J.A., 1985, Characterizationand interpretation of rock mass joint patterns: Geolog-ical Society of America Special Paper 199, 37 p.

Lachenbruch, A.H., 1962, Mechanics of thermal contrac-tion cracks and ice-wedge polygons in permafrost:Geological Society of America Special Paper 70,65 p.

Lipman, P.W., and Friedman, I., 1975, Interaction of me-teoric water with magma: An oxygen-isotope study ofash-flow sheets from southern Nevada: Geological So-ciety of America Bulletin, v. 86, p. 695–702.

Lofgren, G., 1971, Experimentally produced devitrificationtextures in natural rhyolitic glass: Geological Societyof America Bulletin, v. 82, p. 111–124.

Maerten, L., Gillespie, P., and Pollard, D.D., 2002, Effectof local stress perturbation on secondary fault devel-opment: Journal of Structural Geology, v. 24,p. 145–153.

Martel, S., 1994, On the paradox of systematic, contem-poraneous, orthogonal opening-mode fractures, inNelson, R., and Laubach, S., eds., Rock mechanics:Rotterdam, Netherlands, Balkema, p. 801–808.

Mauldon, M., Dunne, W.M., and Rohrbaugh, M.R., 2001,Circular scanlines and circular windows: New toolsfor characterizing the geometry of fracture traces:Journal of Structural Geology, v. 23, p. 247–258.

Miller, T.F., 1990, A numerical model of volatile behaviorin nonwelded cooling pyroclastic deposits: Journal ofGeophysical Research, v. 95, p. 19,349–19,364.

Mongano, G.S., Singleton, W.L., Moyer, T.C., Beason, S.C.,Eatman, G.W., Albin, A.L., and Lung, R.C., 1999, Ge-ology of the ECRB Cross Drift—Exploratory StudiesFacility, Yucca Mountain Project, Yucca Mountain,Nevada. Milestone SPG42GM3: Denver, Colorado,U.S. Geological Survey, 46 p.

Monsen, S.A., Carr, M.D., Reheis, M.C., and Orkild, P.P.,1992, Geologic map of Bare Mountain, Nye County,Nevada: U.S. Geological Survey Miscellaneous Geo-logic Investigations Series Map 2201, 1:24,000 scale.

Morgan, T.L., 1984, Fracture characterization in the TivaCanyon Member, Paintbrush Tuff, Yucca Mountain[M.S. thesis]: Fort Collins, Colorado, Colorado StateUniversity, 120 p.

Morris, A.P., Ferrill, D.A., and Henderson, D.B., 1996, Sliptendency analysis and fault reactivation: Geology,v. 24, p. 275–278.

Nostro, C., Stein, R.S., Cocco, M., Belardinelli, M.E., andMarzocchi, W., 1998, Two-way coupling between Ve-suvius eruptions and southern Apennine earthquakes(Italy) by elastic stress transfer: Journal of Geophys-ical Research, v. 103, p. 24,487–24,504.

Odling, N.E., Gillespie, P., Bourgine, B., Castaing, C., Chil-es, J.-P., Christensen, N.P., Fillion, E., Genter, A., Ol-sen, C., Thrane, L., Trice, R., Aarseth, E., Walsh, J.J.,and Watterson, J., 1999, Variations in fracture systemgeometry and their implications for fluid flow in frac-tured hydrocarbon reservoirs: Petroleum Geoscience,v. 5, p. 373–384.

Olson, J.E., 1996, Post-compressional relaxation and or-thogonal fracture pattern development: Geological So-ciety of America Abstracts with Programs, v. 28, no.1, p. 56.

Pollard, D.D., and Aydin, A., 1988, Progress in understand-ing jointing over the past century: Geological Societyof America Bulletin, v. 100, p. 1181–1204.

Potter, C.J., Dickerson, R.P., Sweetkind, D.S., Drake, R.M.,II, Taylor, E.M., Fridrich, C.J., San Juan, C.A., andDay, W.C., 2002, Geologic map of the Yucca Moun-tain region, Nye County, Nevada: U.S. GeologicalSurvey Miscellaneous Geologic Investigations SeriesMap I-2755, scale 1:50,000, 37 p. pamphlet.

Priest, S.D., 1993, Discontinuity analysis for rock engi-neering: New York, Chapman and Hall, 527 p.

Ragan, D.M., and Sheridan, M.F., 1972, Compaction of theBishop Tuff, California: Geological Society of Amer-ica Bulletin, v. 83, p. 95–106.

Rawnsley, K.D., Rives, T., Petit, J.-P., Hencher, S.R., andLumsden, A.C., 1992, Joint development in perturbedstress fields near faults: Journal of Structural Geology,v. 20, p. 1641–1661.

Rawnsley, K.D., Peacock, D.C.P., Rives, T., and Petit, J.-P.,1998, Joints in the Mesozoic sediments around theBristol Channel Basin: Journal of Structural Geology,v. 20, p. 1641–1661.

Riehle, J.R., 1973, Calculated compaction profiles of rhy-olitic ash-flow tuffs: Geological Society of AmericaBulletin, v. 84, p. 2193–2216.

Riehle, J.R., Miller, T.F., and Baley, R.A., 1995, Cooling,degassing and compaction of rhyolitic ash flow tuffs:A computational model: Bulletin of Volcanology,v. 57, p. 319–336.

Rives, T., Rawnsley, K.D., and Petit, J.-P., 1994, Analoguesimulation of natural orthogonal joint set formation inbrittle varnish: Journal of Structural Geology, v. 16,p. 419–429.

Rosenbaum, J.G., 1986, Paleomagnetic directional disper-sion produced by plastic deformation in a thick Mio-cene

orthogonal jointing during coeval igneous degassing and ...orthogonal jointing during coeval igneous...

Documents