E&P NOTE

Joint sets that enhance

cause the contemporary tectonic stress field favors the propa-gation of hydraulic fracture completions to the east-northeast,

AUTHORS

Terry Engelder Department of Geo-sciences, Pennsylvania State University, Univer-sity Park, Pennsylvania 16802; jte2@psu.edu

Terry Engelder, a leading authority on the recentMarcellus gas shale play, received his B.S. degreefrom Pennsylvania State University (1968), hisM.S. degree from Yale University (1972), andhis Ph.D. from Texas A&M University (1973). Heis currently a professor of geosciences at PennState and has previously served on the staff ofthe U.S. Geological Survey, Texaco, and Lamont-Doherty Earth Observatory. He has written 150research papers, many focused on fracture inDevonian rocks of the Appalachian Basin, anda book, Stress Regimes in the Lithosphere.

Gary G. Lash Department of Geosciences,State University of New York, College at Fre-donia, Fredonia, New York 14063

Gary Lash, an authority on various aspects ofthe Middle and Upper Devonian shale succes-sion of western New York, received his B.S.degree from Kutztown State University (1976)and his M.S. degree and Ph.D. from Lehigh Uni-versity (1978 and 1980, respectively). Beforeworking in western New York, Lash was involvedin stratigraphic and structural investigations ofthrusted CambrianOrdovician deposits of thecentral Appalachians.

Redescal S. Uzctegui Departamento deCiencias de la Tierra, Simn Bolvar, Caracas,Venezuela; present address: Intevep, S.A., Apar-tado Postal 76343, Caracas 1070A, Venezuela

Redescal Uzctegui is a professor of structuralgeology at the Universidad Simn Bolvar andan I&D (Instruction & Development) associatedprofessional in tectonics and structural geologyat Petroleos de Venezuela, S.A. (PDVSA)-Intevep.He received his B.A. degree in geology at theUniversidad Central de Venezuela and his Ph.D.in geosciences from the Pennsylvania StateUniversity. His current focus of research is thegeometry and evolution of structures and frac-tures in the Perij and Andes de Mrida foothills,and the Maracaibo Basin in Venezuela.

ACKNOWLEDGEMENTS

We recognize numerous former students whoseB.S., M.S., and Ph.D. theses over the past 30 yearsfracture stimulation from vertical wells intersects and drainsJ2 joints. Horizontal drilling and subsequent stimulation ben-efit from both joint sets. By drilling in the north-northwestsouth-southeast directions, horizontal wells cross and drain J1joints, whenever present. Then, staged hydraulic fracturestimulations, if necessary, run east-northeast (i.e., parallel tothe J1 set) under the influence of the contemporary tectonicstress field thereby crosscutting and draining J2 joints.

Copyright 2009. The American Association of Petroleum Geologists. All rights reserved.

Manuscript received February 29, 2008; provisional acceptance May 20, 2008; revised manuscriptreceived March 2, 2009; final acceptance March 23, 2009.DOI:10.1306/03230908032production from Middle andUpper Devonian gas shalesof the Appalachian BasinTerry Engelder, Gary G. Lash,and Redescal S. Uzctegui

ABSTRACT

The marine Middle and Upper Devonian section of the Ap-palachian Basin includes several black shale units that carrytwo regional joint sets (J1 and J2 sets) as observed in outcrop,core, and borehole images. These joints formed close to or atpeak burial depth as natural hydraulic fractures induced byabnormal fluid pressures generated during thermal matura-tion of organic matter.When present together, earlier J1 jointsare crosscut by later J2 joints. In outcrops of black shale on theforeland (northwest) side of the Appalachian Basin, the east-northeasttrending J1 set is more closely spaced than thenorthwest-striking J2 set. However, J2 joints are far more per-vasive throughout the exposed Devonian marine clastic sec-tion on both sides of the basin. By geological coincidence,the J1 set is nearly parallel the maximum compressive normalstress of the contemporary tectonic stress field (SHmax). Be-AAPG Bulletin, v. 93, no. 7 (July 2009), pp. 857889 857

contributed in meaningful ways to the finalconclusions of this article. These former studentsinclude LaMichelle Arnold, Paul Bembia, RandyBlood, Pat Case, Pat Dwyer, Mark Fischer, AmyFreeman, Ben Haith, Michael Gross, John Kroon,Alfred Lacazette, John Leftwich, Staci Loewy,Steve Marshak, David McConaughy, RichardPlumb, Laura Savalli, Michael Scanlin, AnthonySoricelli, Amy Whitaker, and Amgad Younes.We also thank colleagues who, through collab-oration, contributed to the conclusions herein,including Marc Sbar, Peter Geiser, and KeithEvans. Early versions of this article were read byKent Bowker, Ron Nelson, and Barry Tew. Thiswork was supported by New York State EnergyResearch and Development Authority (NYSERDA),U.S. Department of Energy (DOE), and NationalResearch Council as far back as the 1970s. Theseeds for recognizing the important function ofnatural hydraulic fracturing were planted in agrant from the National Science Foundation(EAR-83-04146). The Gas Research Institute (GRI)and Penn State's Seal Evaluation Consortium(SEC) contributed significant funding as well.Data collection and writing of this article weresupported by NSF EAR-04-40233, Departmentof Energy (DOE)/National Energy TechnologyLaboratory (NETL) (WVU subcontract-08-686F-PSU TO 41817M2173), and Appalachian Frac-ture Systems, Inc.The AAPG Editor thanks the following reviewersfor their work on this paper: Kent A. Bowker,Ronald A. Nelson, and Berry H. Tew, Jr.

858 E&P NoteINTRODUCTION

DevonianMississippian gas shale in the Appalachian Basin isparticularly susceptible to joint growth, an observation datingfrom the early 19th century geological survey of New Yorkstate (Hall, 1843). By the early 20th century, geologists recog-nized that fracture by joint growth in black shale differs in ori-entation and density when comparedwith joint growthwithingray shale and siltstones of the Appalachian Basin (Sheldon,1912). Mapping of joints throughout the northern Appala-chian Basin revealed that more than one black shale forma-tion, including those of theMarcellus Formation (Ver Straetenand Brett, 2006), hosts the same east-northeaststriking jointset (Parker, 1942). Moreover, mapping of east-northeastjoints in black shale of the Illinois Basin confirmed that the af-finity between joint growth andDevonian black shale extendsbeyond the confines of the Appalachian Basin (Campbell,1946). Viewed in hindsight, data collected through the mid-20th century point to a commonpost-Devonian growthmech-anism that originated within black shale formations accumu-lating over more than 30 m.y. of the Middle to Late Devonian(Figure 1).

The timeline for industrial development of fractured De-vonian shale in the Appalachian Basin starts with productionfrom the Dunkirk Formation at Fredonia, New York, in 1821and from the Marcellus Formation within the Naples field,New York, in 1880 (Van Tyne, 1983). Production began in theHuron (Dunkirk equivalent) black shale of the Big Sandy field,Kentucky, in 1914 (Hunter and Young, 1953). From 1821, ittook industry more than a century to recognize the criticalfunction that natural fractures have in economic gas produc-tion (Browning, 1935). Through 1953, natural-fracture-aidedgas production fromDevonian black shale inKentucky,mainlyfrom the Huron (Dunkirk equivalent) black shale of the BigSandy field, had approached nearly 1 tcf (Hunter and Young,1953). Indeed, some of the early wells in the fractured Huronand Rhinestreet formations produced for 50 yr (Vanorsdale,1987). By the 1970s, however, matrix porosity was consid-ered to exert the principal control on long-term productionfrom black shale, thereby subordinating natural fractures tothe function of high-permeability pathways (Smith et al.,1979). Initially, interconnectivity with natural fractures wasachieved by shooting nitroglycerin in gas wells, but by the1980s, this practice was supplanted by nitrogen, carbon diox-ide, and foam-fracturing agents (Sweeney et al., 1986).

Horizontal drilling arrived in theAppalachianBasin in 1944as a means of producing heavy crude oil from the Venango

Figure 1. Stratig-raphy of NorthAmerican blackshales with a focuson the northerncentral Appala-chian Basin. Onlythe names of blackshale units are given.The vertical scalefor units outsidethe inset box istime and not thick-ness. The Devonianblack shales aredated using cono-donts and tied toabsolute time withash beds (Over,2002, 2007; Stasiukand Fowler, 2004;Kaufmann, 2006;Over et al., 2009).Many erosionalsurfaces punctuatethe deposition ofDevonian blackshale so sectionsare mostly discon-tinuous. A strati-graphic section ofthe Catskill deltasouthwest of Leroy,New York, is scaledfor thickness (insetbox). The blackshales of westernNew York includeDunkirk (DK), PipeCreek (PC), Rhine-street (RS), Middle-sex (MS), Geneseo(GS) (found in cen-tral New York), andMarcellus (MAS).Engelder et al. 859

sandstone following the depletion of its solution gas(Overby et al., 1988).Devonianblack shale ofWestVirginia was tested with high-angle or slant drillingstarting in 1972, and a horizontal test was com-pleted in 1987 (Yost II et al., 1987a). The 1987 lat-eral was aimed S37E because several subsurfacedata sets mostly from the U.S. Department of En-ergy Eastern Gas Shales Project (EGSP) showedthat this orientation crossed the highest density ofnatural fractures, an east-northeast set (Cliffs Min-erals, 1982;Yost II et al., 1987b).At about the sametime, the Devonian Antrim gas shale was tested inthe Michigan Basin with slant wells designed tocrosscut natural fractures (Hopkins et al., 1998).Full-scale development of the Huron (Dunkirkequivalent) of the Big Sandy field, Kentucky, com-menced in the 2000s where one in five horizontalwells (of 80 drilled to date) flows with enough vol-ume to forgo stimulation, a clear indication ofwell-bores crossingeast-northeast joints (Gerber, 2008).When stimulation is required, nitrogen and foamare commonly used as the fracturing medium inthe underpressured Huron (Dunkirk equivalent)black shale.

Unlike the relatively shallow Huron (Dunkirkequivalent) black shale, every economic horizontalwell in deeper DevonianMississippian gas shalerequires stimulation with technological break-throughs, including slickwater fractures and multi-stage fracturing (Fontaine et al., 2008). These stimu-lation techniqueswere perfected in theCretaceousCotton Valley of east Texas and Mississippian Bar-nett black shale of the Fort Worth Basin, Texas(Walker et al., 1998), and soon applied toother blackshale plays, including the Woodford (Oklahoma),Fayetteville (Alaska), Marcellus (Maryland, NewYork, Pennsylvania, Ohio, West Virginia), and Hay-nesville (Louisiana) (Figure 2). The first horizontalwells in the Marcellus Shale of Pennsylvania datefrom 2006, and 47 horizontal wells have spud datesthru 2008 according to Pennsylvania Departmentof Environmental Protection records. As of early2009, the most voluminous 24-hr flow test showed24.5mmcf/day inWashingtonCounty, Pennsylvania.

The Marcellus black shale of the AppalachianBasin constitutes one class of unconventional res-ervoir from which production is optimized if hor-860 E&P Noteto the presence of unhealed fractures at depth. Theeconomic importance of unhealed fractures in gasshales lends exigency to an extended discussionof the present understanding of the origin, orienta-tion, and distribution of joints in theMarcellus andother black shales of the Appalachian Basin.

Marcellus and Its Tectonic Heritage

The Appalachian Basin contains eight major blackshale units, most placed at or near the bottom ofstratigraphic groups such as the Middle DevonianHamilton (Ettensohn, 1985) (Figure 1). The lowerpart of the Hamilton Group comprises more thanone black shale, including the Marcellus Forma-tion whose basal unit, the Union Springs Member,contains greater than 10% total organic carbon(TOC) (Werne et al., 2002; Desantis et al., 2007).The prospective thickness of the Hamilton Groupreaches 400500 ft (122152 m) where the over-lying Oatka Creek Member of the Marcellus For-mation and theMahantango Formation are incorpo-rated into completion strategies. The areal extentof the Marcellus Formation with at least 50 ft(15 m) of a high API gamma-ray signal exceeds34 106 ac (Figure 2), which, given the typicalporosity and adsorption properties of Devonianizontal drilling penetrates a systematic fracture set(Curtis and Faure, 1997;Curtis, 2002; Law, 2002).In deeper, thermogenic black shale plays, the ex-tent to which productivity is enhanced by stimula-tion of natural fractures remains a question. Here,stimulation is understood to mean the linking ofnatural, higher permeability pathways to horizontalwellbores. Citing experience gained from the Bar-nett Shale, somehave questioned the paradigm thatnatural fractures are requisite for a successful play(Bowker, 2007). Indeed, questions remain regardingthe nature of fracturing of theMarcellus Shale in thesubsurface. The immediate purpose of this articleis to present new outcrop data showing that theMarcellus gas shale, at least in places in the northernAppalachianBasin, has the same fracturepattern thatmakes for successful production from the Huron(Dunkirk equivalent) black shale, even in the ab-sence of stimulation. These outcrop data are con-sistent with published data from the core that testify

black shale, translates to an excess of 1000 tcf andmaybe as high as 3500 tcf of gas in place (Figure 2).All indications are that the Marcellus gas shale willdevelop into a super giant gas field.

The extent to which the Appalachian shaleswere incorporated into the deformation of a conti-nent-continent collision, theCarboniferousPermianAlleghany orogeny, separates the Marcellus andother Devonian gas shales of the Appalachian Ba-sin fromgas shales in several otherNorthAmericanbasins (Rodgers, 1970; Hatcher et al., 1989). Al-though touched by Alleghanian tectonics, otherDevonian gas shales linked to the Acadian fore-land, including the New Albany and Antrim, werenot subjected to the amount of layer-parallel short-

ening (LPS) seen by theMarcellus (Engelder andEngelder, 1977;Craddock andvanderPluijm,1989).Layer-parallel shortening is a manifestation of de-tachmentwithin the Silurian Salina salt and ancil-lary blind thrusts beneath the Marcellus (WiltschkoandChapple, 1977;Davis andEngelder, 1985; Scan-lin and Engelder, 2003) (Figure 3). The DevonianMississippian Woodford gas shale in the Ouachitaforeland, Oklahoma, is structurally complex al-though the major Alleghanian detachment passesthrough the shallow Atoka, well above the Wood-ford (Durham, 2008). The Mississippian Barnett,Fayetteville, and Floyd gas shales in the Ouachitaforeland were not subjected to large-scale LPSduring the Alleghanian orogeny. The very eastern

Figure 2. The areal extent of the Marcellus black shale in the Appalachian Basin (modified from Milici, 2005). Isopach contours on theMarcellus are 50-ft (15-m) intervals. Stars indicate the distribution of J1 joints in the subsurface of the Appalachian Basin as reported tothe authors by Marcellus Shale operators sharing proprietary data. The location of several North American Devonian black shales plusthe Mississippian Barnett and Fayetteville and the Jurassic Haynesville is shown on the inset map.Engelder et al. 861

Figure 3. Two-dimensional seismic lines through the Laurel Hill and Fayette anticlines. The background is a geological map of the State of Pennsylvania showing the Alleghenystructural front (ASF, dashed), which is the boundary between the Appalachian Plateau to the northwest and the Appalachian Valley and Ridge to the southeast. The stratigraphicposition of the Marcellus is shown with heavy arrows. Seismic lines are modified from Scanlin and Engelder (2003). Refer to Figure 2 for the location of Pennsylvania.

862E&P

Note

Sykes, 1973; Engelder, 1982a; Zoback, 1992). The

J2 joints generally crosscut J1 joints when the twosets are found in the same bed (Figure 5). In theMarcellus and other black shale units of the Appa-lachian Basin, the J1 set is more closely spacedthan the J2 set (Figure 6). The affinity of J1 jointsfor organic-rich shale was recognized a centuryago (Sheldon, 1912); 30 yr lapsed before it waspart of the Floyd was incorporated in Alleghanianstructures as a mushwad (Thomas, 2001). SuchMesozoic gas shales of the Gulf Coast region asthe Haynesville-Bossier and the Pearsall were sub-jected to extensional tectonics with local salt move-ment forcing local but minor LPS.

Natural Fractures in the Appalachian Basin

Rock fracture occurs either by rupture in shear orrupture in tension. Rupture in shear yields faults(Handin and Hager, 1957), whereas rupture intension leads to the propagation of joints (Pollardand Aydin, 1988). At depth in basins where stressis compressive, tension is an effective stress andjoints are natural hydraulic fractures (Secor, 1965;Engelder and Lacazette, 1990). Faults are rarely sys-tematic and are invariably concentrated in zonesassociated with a master fault or fold (Aydin andJohnson, 1978), whereas joints are frequently sys-tematic and may be pervasive over large regions(Hodgson, 1961). Industry has established thatjoints of one systematic set in particular, the east-northeast set, are crucial to the success of horizon-tal completion techniques in the Appalachian Basin(Gerber, 2008). Systematic east-northeast jointshave been observed in the core recovered from theUpperDevonianHuron (Dunkirk equivalent) blackshale as part of the EGSP (Figure 4).

Outcrops of black shale within the Appalachianforeland commonly carry two systematic joint sets,herein referred to as the J1 and J2 sets (Sheldon,1912; Parker, 1942; Engelder and Geiser, 1980;Lash et al., 2004; Lash and Engelder, 2007, 2009).The J1 set is of particular interest because it strikesto the east-northeast and within a few degrees ofthemaximumhorizontal compressive stress, SHmax,of the contemporary tectonic stress field (Sbar anddemonstrated that J1 joints transect fold axes onthe Appalachian Plateau at low angles and are un-likely to be fold related (Parker, 1942). SystematicJ1 joints were initially assigned a subordinate rank,set III, mostly because they are not well developedin the more abundant outcrops of gray shale anddistal turbidites of the Catskill delta (Parker, 1942;Engelder and Geiser, 1980). Later still, J1 jointswere recognized in outcrops from Virginia to NewYork and interpreted to correlate with a strong east-northeasttrending coal cleat in Morrowan andDesmonian coal deposits scattered from Alabamato Pennsylvania (Engelder, 2004; Engelder andWhitaker, 2006).

The similar orientation of J1 joints in blackshales of the Appalachian Basin and the SHmax ofthe contemporary tectonic stress field lured manyauthors to the conclusion that all east-northeaststriking joints were de facto neotectonic and there-fore related in some way to processes involvingthe contemporary tectonic stress field, includingTertiary exhumation or Pleistocene glaciation(Clark, 1982; Engelder, 1982a; Dean et al., 1984;Hancock and Engelder, 1989;Gross and Engelder,1991; Engelder and Gross, 1993). These interpre-tations were challenged when it was noted that, insome places along the Appalachian Valley andRidge, the early J1 joints were folded along withbedding, a clear evidence of a pre- or early Al-leghanian propagation history (Engelder, 2004).Finally, we understood that a pre- or early Alle-ghanian joint set, the J1 set, occupies nearly thesame orientation as Appalachian neotectonic jointsand SHmax of the contemporary tectonic stress field(Pashin and Hinkle, 1997; Engelder, 2004; En-gelder and Whitaker, 2006; Lash and Engelder,2009).

The presence of early joints in black shale ofthe Appalachian Basin and their orientation, rela-tive to, first, the Alleghany orogeny tectonic stressfields and then the contemporary ones, gives riseto three geological conundrums. The first conun-drum involves the extrapolation of outcrop datafor the purpose of predicting the orientation ofunhealed joints within black shale at depth. Suchextrapolation assumes that the geologist is able todistinguish between joints that formed in the nearEngelder et al. 863

864 E&P Notesurface as geomorphic phenomena and joints thatformed close to or at maximum burial depth andpersisted through exhumation. Drawing a distinc-tion between early- and late-formed joints in theAppalachian Basin is complicated by the coinci-

dence of the strike of J1 joints and SHmax of thecontemporary tectonic stress field. Neotectonicjoints (J3) whose east-northeast strike is controlledby the contemporary tectonic stress field have amean orientation that falls within the same statistical

Figure 4. Rose plots showing the orientation of joints in the Dunkirkmiddle Huron interval of cores recovered during the EGSP (mod-ified from Cliffs Minerals, 1982). Well designations are those of the ESGP. Wells with less than three joints are omitted. Well WV-4 iscorrected for misalignment as indicated by induced fractures.

range as the J1 set (Hancock and Engelder, 1989;Lash and Engelder, 2009). The second conundrumstems from geological evidence that suggests thatthe J1 set propagated before Alleghany-orogeny-induced folding and concomitant LPS (Pashin

and Hinkle, 1997; Engelder, 2004). If the J1 setpredates Alleghanian LPS, then it survived thistectonic deformation in black shales as unhealedjoints with little modification during penetrativestrain despite being subnormal to the direction of

Figure 5. (A) Crosscutting J1 and J2 joints in the Marcellus black shale exposed in Oatka Creek, Le Roy, New York. View is to the east-northeast. (B) The J1, J2, and J3 joints in the Marcellus black shale within folded beds just south of Jacksons Corner, Pennsylvania. Theview is to the east-northeast showing south-dipping beds. Freiberger compass for scale. See Figure 2 for the locations.Engelder et al. 865

866 E&P Noteasmuch as 15%LPS (Engelder andEngelder, 1977;Geiser, 1988). Such a scenario presents a geologi-cal mystery that is difficult for some structuralgeologists to accept because it defies themore par-simonious interpretation that early joints can onlysurvive as veins, whereas pristine joints with cleansurfaces postdate penetrative deformation. Thethird conundrum arises from consideration ofthe origin of J2 joints. There remains the matterof reconciling the Andersonian stress state forforeland fold and thrust belts (i.e., the thrust-faultregime where the least principal stress, s3, is ver-tical) with the observed spectrum of vertical syn-tectonic J2 joints in the Appalachians and else-where (Anderson, 1951). Vertical joints implythat s3 was horizontal during fold and thrust tec-

tonics. We present data that address each of theseconundrums thereby permitting reasonable infer-ences regarding the extent to which the gas indus-try may expect unhealed joints in black shales suchas the Marcellus.

FIELD OBSERVATIONS

Two simple experiments will enable us to furtherexamine the aforementioned geological conun-drums and to address other questions regardingthe extent of joint development at depth in De-vonian black shales of the Appalachian Basin.Each experiment involves documenting joints inthe Marcellus Shale along two tracks directed at the

Figure 6. Representative box-and-whisker plots showing the joint density (orthogonal spacing) of J1 (black boxes) and J2 (gray boxes)joints at the base black shale units of the distal (southwest of LeRoy, New York) and more proximal (near Ithaca, New York) delta regions(see Figure 2 for the locations). Rocks of the proximal delta carry multiple J2 joint sets. The average strike of each joint set is listed on theordinate of each plot. The box encloses the interquartile range of the data set population; bounded on the left by the 25th percentile(lower quartile) and on the right by the 75th percentile (upper quartile). The vertical line through the box is the median value, and thewhiskers represent the extremes of the sample range. In the distal scan lines, the spacing on east-northeast joints reflects the late-stagereactivation of those joints (modified from Lash and Engelder, 2009).

Figure 7. The orientation of joints in the Marcellus and other Devonian black shales of the Appalachian Basin. Poles to joints are shown in the lower hemisphere stereographicprojection. The data from Antis Fort and Elimsport have been rotated back to their position relative to horizontal bedding.

Engelderet

al.867

different segments of the Valley and Ridge Appa-lachians. The common starting point is the FingerLakes region of New York where the orientationof approximately 100 J1 joints was measured ateach of 22 outcrops of black shale (Lash et al.,2004). Our first track extends east to outcrops ofthe Marcellus Shale along the Mohawk Valley andcrosses the ASF upon reaching the Hudson Valley(Figure 7). At each of these outcrops, as many as50 joints were measured, depending on the out-crop quality. These data are plotted as poles ona lower hemisphere stereonet projection and thenas rose diagrams. If a significant dip to bedding isobserved, then the poles are rotated to return bed-ding to horizontal before constructing the rosediagram. Our second track points south crossingthe ASF west of Williamsport, Pennsylvania. TheMarcellus is not present along the southern trackuntil the Valley and Ridge of Pennsylvania isreached.

Joint Characteristics that Carry across theAllegheny Structural Front

Several characteristics of joint development arecommon to both sides of the ASF. First, both J1and J2 joints carry through the ASF as planar setswith uniform spacing (Figures 5, 8). Second, J1and J2 joints propagated normal to bedding northof the ASF and remain normal to bedding whenlimb dips are modest (

Figure 8. (A) Moonshine Falls east of Aurora, New York. Joints plotted in present coordinates in a lower hemisphere stereonet and rose diagram (A-1). (B) Examples of joint devel-opment in the Oatka Creek Member of the Marcellus Formation at the Ed Snook Quarry along Old Fort Road off Route 44 west of Antis Fort, Pennsylvania. Bedding is overturned at N82E 75SE. The view looking north at the underside of overturned Marcellus beds. Joints plotted in present coordinates (B-1) and rotated to their position with horizontal bedding using afold axis plunging 05 toward 082E with a rotation of 105 (B-2). Locations are on Figure 7.

Engelderet

al.869

Figure 9. Regional map of J2 joints in the Brallier and Trimmers Rock formations of the Valley and Ridge Appalachians, Pennsylvania. When present, the dashed line shows theorientation of bedding.

870E&P

Note

Figure 10. (A) Union Springs Member of the Marcellus Formation at Marcellus, New York. Here, J1 joints propagate around but do not cleave concretions, a geological evidence fornatural hydraulic fracturing. Joints plotted in present coordinates in a lower hemisphere stereonet and rose diagram (C). (B) Examples of joint development in the Union SpringsMember at the Delmar Finck Quarry along Pikes Peak Road off of Route 44 east-northeast of Elimsport. Bedding is N75E 10SE. Joints plotted in present coordinates (D) and rotated totheir position with horizontal bedding using a fold axis plunging 05 toward 075 with a rotation of 10 (E).

Engelderet

al.871

Figure 11. (A) Skaneateles Formation at Moonshine Falls east of Aurora, New York. Looking south-southeast parallel to J2. (A-1, A-2) Neotectonic or exhumation joints parallel J2.(B) Union Springs Member of the Marcellus Formation at Union Springs, New York, looking parallel to J1 joints cutting vertically to the outcrop surface. (B-1) The J2 joints curving towardthe outcrop surface.

872E&P

Note

tion events as indicated by fringe cracks (Zhao andJacobi, 1997; Younes and Engelder, 1999). Fringecracks propagate both upward and downwardfrom parent joints (Figure 13).

Joint Characteristics that Differ across the ASF

Three characteristics of joint development are notshared in common across the ASF. First, J1 jointsare present in every studied Marcellus, Geneseo,and Middlesex black shale outcrop in the FingerLakes region as well as in other black formationsin the Lake Erie region (Figure 7). To the southof the ASF in Pennsylvania, J1 joints were identi-fied with certainty in less that 50% of the studiedMarcellus outcrops. The second characteristic thatdoes not find its way across the ASF is the degreeof clustering of poles of joints. On the foreland sideof theASF, J1 joints cleave outcrops as if propagatingin an isotropic homogeneous medium subjectedto a homogeneous stress field (Figure 7). All J1joints in one outcrop are parallel to a first approxi-mation. This behavior is recorded by the very tightcluster of poles to joints (Engelder and Whitaker,2006). The cluster of poles to J1 joints on the hin-terland side of the ASF is far looser (Figure 7).that flatten upward toward the top of bedrock(Figure 11B). These joints strike parallel to verticaljoints that carry right up to the bedrock-soil con-tact. Elsewhere, interspersed nonsystematic jointscurve to parallel planar neighbors (Figure 11A).

Finally, the plumose morphology that dec-orates the surfaces of joints hosted by siltstonesof theMiddle and Upper Devonian populates out-crops on both sides of the ASF (Figure 12). Plu-mose morphology is particularly valuable as a toolfor assessing relative joint propagation direction,velocity, and points of arrest (Bahat and Engelder,1984; Savalli and Engelder, 2005). Rupture accel-eration and then arrest are indicated by surfaceswith increasing roughness up to the point of arrest(Figure 12A). Smoother surfaces with irregularpropagation directions indicate slower subcriticalpropagation (Figure 12B). Joint propagation ex-tends over a long enough period that the regionalstress orientation driving the deformation duringthe Alleghany orogeny realigns between propaga-Adjacent J1 joints might be misaligned by 1 to3 (Figure 5B). The third characteristic that differsfrom one side of the ASF to the other is the rela-tionship between planar, neotectonic (J3) jointsand J1 joints. Here, we presumed that nonsystem-atic joints are all exhumation-related neotectonicjoints. Nonsystematic joints differ from the curving,systematic joints reported near faults (Rawnsleyet al., 1992). South of the ASF where J1 joints tiltwith bedding on the limbs of first-order folds, a ver-tical east-northeast set, J3, cuts bedding obliquely(Figure 5B). North of the ASF, J3 joints are mani-fested by reactivated J1 joints that abut J2 joints(Lash and Engelder, 2009).

Layer-Parallel Shortening in the Marcellus

Four structures show the extent to which theMar-cellus black shale was subjected to a penetrativestrain indicative of LPS during Alleghanian defor-mation. Worm borrows in shale within the transi-tion zone between the Marcellus and underlyingOnondaga Limestone have an elliptical shape withtheir long axes parallel with fold axes of the Valleyand Ridge in Pennsylvania (Figure 14A). Pencilcleavage with a lineation parallel to local fold axesis common in outcrops of theMarcellus in the Val-ley and Ridge of Pennsylvania (Figure 14B). Small-scale buckle folds are present in 2- to 3-mm-thick(0.070.11-in.-thick) silt layers within the Marcel-lus (Figure 14C). The axes of these buckle folds areparallel with first-order fold axes of the Pennsylva-nia Valley and Ridge. Finally, axial-planar stylolitesare found in limestone layers at the OnondagaMarcellus contact (Figure 14D).

Other structural elements common to theMarcellus include cleavage duplexes and small-scale faults that ramp toward the foreland. Theduplexes are commonly parallel to bedding (Nick-elsen, 1986). Bedding-parallel slip is also manifestby bedding-plane slickensides, which, in the Ma-hantango Formation, may be as closely spaced as510 cm (1.93.9 in.) on the flanks of first-orderfolds in the Valley and Ridge. Other bedding-parallel slip is distributed within Marcellus gougezones 23 cm (0.71.1 in.) thick (Figure 14D).Engelder et al. 873

874 E&P Note

cleavage duplexes of the Marcellus verges in the

latter contributes to the development of parallelsystematic joints. However, planarity in and of it-

folded succession are interpreted as postfolding

f a n[18.6er, 1ce oself does not permit one to distinguish late-formed,near-surface joints from early-formed, deep joints

Figure 12. (A) Discrete propagation events during the growth oout along Route 414 southwest of Watkins Glen, New York (30 kmtion mapped on this joint (modified from Lacazette and EngeldHuntingdon, Pennsylvania. A plumose pattern decorates the surfadirection of the foreland regardless of position rela-tive to local folding.

DISCUSSION

Conundrum 1: Recognition of Outcrop Jointsthat Formed at Depth

During the past half century, two important state-ments emerge from the observation of joints inoutcrop. First, it is unlikely that all joints arethe result of a single mechanism (Price, 1966,p. 110). Second, Fracture patterns are cumulativeand persistent. Cumulative implies several epi-sodes of fracturing Persistent means not easilyerased by later tectonic events (Nickelsen, 1976,p. 193). From these statements, it follows that theeffective tensile stress necessary for joint propaga-tion is attained at the end of several loading pathsinvolving stresses developed during burial, tectonicdeformation, and/or later exhumation (Engelder,1985). Furthermore, the energy necessary to sus-tain propagation of joints beyond initiation has atleast four loading configurations (Engelder andFischer, 1996). Given a complex matrix of loadingpaths and loading configurations, jumping fromfield observations to a unique genetic interpreta-tion is difficult.

Of all the properties of joints, the most strik-ing is their planarity. The mechanical explanationis that stress controls joint propagation and a recti-linear stress field serves to maintain in-plane jointgrowth (Pollard andAydin, 1988; Lawn, 1993). Thestructures. In the Appalachian Basin, verticaleast-northeast joints carried by tilted beds arecandidates for shallow neotectonic joints (J3)(Hancock and Engelder, 1989). Joints that prop-agated normal to bedding and remained in thatposition throughout subsequent folding are can-didates for early, prefolding joints that formed atdepth and persisted through exhumation to expo-sure at the Earths surface (Figure 11). These arethe J1 joints of the Appalachian Basin (Engelder,2004).

The upward growth of joints during exhuma-tion was a popular explanation of vertical joints inoutcrop (Nevin, 1931). The mechanical basis forthis hypothesis was that horizontal compressivestress decreases upward in the Earth as a conse-quence of a reduction in both temperature andthe gravitational component of the Earths stressfield (Price, 1966; Engelder, 1993). Fluid-drivenfractures would naturally climb because the incre-mental decrease in density-related fluid pressure isless than the incremental decrease in the gravita-tional component of the horizontal stress (Nunn,1996). However, the problem with the upwardgrowth explanation is that if stress in a unit vol-ume of rock is tracked during exhumation, the in-cremental decrease in horizontal stress in that unitcube does not keep up with the rate of verticalstress reduction (Brown and Hoek, 1978; Plumb,1994). This behavior is reflected by a stress statein near-surface rocks in which Shmin is greater thanSv (Nadan and Engelder, 2009). Indeed, some ofthe most common joints with a clear near-surfaceorigin are subhorizontal sheet fractures and exfo-liation joints (Holzhausen, 1989; Martel, 2006).Joints growing upward into a stress field whereShmin is greater than Sv should tilt and eventuallybecome horizontal as exemplified by the formation

atural hydraulic fracture (J2 joint) in the Ithaca siltstone croppingmi] west of Ithaca). The insert shows 68 increments of propaga-

992). (B) Joints in a bed of the Devonian Brallier Formation atf both a strike and J2 joint sets (modified from Ruf et al., 1998).The sense of slip on bedding-plane slickensides inthe Mahantango and the Marcellus gouge zones istop toward local anticlinal axes. Slip on the thicker

that persist through one or more tectonic cycles.In this regard, the attitude of planar joints relativeto bedding provides a clue. Vertical joints in aEngelder et al. 875

876 E&P Note

table. Evidence of upward growth of late joints inblack shale is seen in the tendency for late-formed

tributed to glacial loading and unloading cycles(Clark, 1982; Evans, 1989) as well as to the pre-

pagak, whis exis selogyjoints to curve parallel to the Earths surface underthe influence of a stress state where Sh is greaterthan Sv (Figure 11). Alternatively, vertical jointsthat extend to the bedrock-soil contact in black shaleoutcrops were generated at depth and persistedduring exhumation in their original orientation.

One common characteristic of joints in blackshale along both transects is their close spacingrelative to height (Figure 11). This observationis consistent with a deep-formed natural hydraulicfracture mechanism for driving vertical joints inblack shale (Fischer et al., 1995). Even whereblack shale carries closely spaced joints in both J1and J2 orientations (east-northeast and cross fold,respectively), late-formed joints are easily identi-fied in cross section view as they curve parallelwith the earlier, vertical planar joints (Figure 11).The same late-formed joints in plan view curve per-pendicular to the systematic joints and abut them,hence the name curving cross joints (Engelder andGross, 1993). The mechanical explanation for thisbehavior is that the crack-tip stress field of the late-formed joints is not transmitted across open joints,and without the benefit of a crack-tip stress field,late joints cannot jump across the earlier, open joint(Gross, 1993). Implicit in this explanation is therequirement that the older joint was open whenthe latter joint propagated. The persistence ofremnant horizontal compression during exhuma-tion means that curving cross joints are unique tobeds in the upper few meters in the AppalachianBasin. The mechanism for curving parallel joints

Figure 13. Abrupt twist hackles. (A) This set of fringe cracks prohosting the parent joint at Taughannock Falls State Park, New Yorpart of the Ithaca Formation. The sense of stress field rotation in th(31.-in.) base (modified from Younes and Engelder, 1999). (B) ThBrallier Formation at Huntingdon, Pennsylvania. Plumose morphothe layer below the scale marker.sumed brittle nature of the deformed rocks. Thethread that links both notions is the developmentof a flexural bulge accompanying glacial loadingthat may have generated a tensile stress in the elas-tically stiffest rocks entrained above the neutral fi-ber of the bulge. Although such joints have beendescribed from the Appalachian Basin (Lash andEngelder, 2007), such an explanation is not con-sistent with the regional distribution of systematicJ1 and J2 sets in Devonian black shales, which tendto be some of themost compliant, not stiffest, bedsof the Appalachian Basin. The evidence that sup-ports natural hydraulic fracturing for deep jointssimultaneously speaks against both the stiffnesshypothesis and the hypothesized function of glacialloading and unloading for the preferential jointingof the Marcellus and other Devonian black shale.

Geological CoincidenceThe J1 joint set and SHmax of the contemporarytectonic stress field are nearly parallel in easternNorthAmerica (Plumb andHickman, 1985; Zoback,1992). This correlation led to an early hypothesisthat the orientation of the J1 joints was controlledby the modern SHmax in the North Americanlithosphere (Engelder, 1982a). Indeed, J1 jointsin Devonian black shale have been explicitlycalled neotectonic joints (Hancock and Engelder,1989).We now know that the parallelism of SHmaxand the J1 set is a geological coincidence (Engelderand Whitaker, 2006). In the late Paleozoic, themodern eastern edge of North America (Laurentia)

ted downward into a thick shale bed from a thinner siltstone bedich is 12 km (7.4 mi) north of Ithaca, New York. These rocks areample is clockwise. The scale is a geologic compass with an 8-cmt of fringe cracks propagated upward into a siltstone layer of theshows the upward direction of propagation from a parent joint inof sheet fractures in exhumed granite bodies. Be-cause this model requires the existence of near-surface horizontal compressive stress, the drivingstress necessary for upward growth may require amodest tensile effective stress below the water

and curving perpendicular joints involves the dis-tortion of an otherwise symmetrical crack-tipstress field (Olson and Pollard, 1989; Lash andEngelder, 2009).

Near-surface joint propagation has been at-Engelder et al. 877

878 E&P Note

Figu

re14

.Fou

rdifferen

tstru

ctures

indicativ

eof

laye

r-parallelsho

rtening

with

intheMarcellu

sblack

shalene

arNe

wton-Ha

milton

,Pen

nsylvania.

(A)D

eformed

worm

tube

ssam

pled

attheMarcellu

s-Ono

ndag

aco

ntact.(B)Pe

ncilcle

avag

ein

theUn

ionSp

rings

Mem

berof

theMarcellu

sFo

rmation.

(C)Sm

all-scale

buckle

foldsin

asiltlay

erwith

intheUn

ionSp

rings

Mem

berof

theMarcellu

sFo

rmation.

Thepa

lmof

aha

ndis

seen

behind

thesamples.(

D)Th

eco

reof

theMarcellu

s-Ono

ndag

aco

ntact(

courtesy

ofSa

mson).T

heco

ntactisan

interla

yeredblackshalean

dlim

estone

(ligh

teru

nit).

Asty

lolitic

solutio

ncle

avag

eap

pearspa

ralleltotheaxisof

theco

re(vertical

arrows).A

smallb

edding

plan

ede

tach

men

tisseen

intheup

perpa

rtof

theco

re(arrow

sindicate

senseof

shea

r).was oriented about 45 clockwise from its presentorientation such that this same edge of Laurentiafaced south (modern coordinates). East-southeastto west-northwestdirected convergence of Gond-wana (Africa) against Laurentia (North America)generated a plate boundary system of dextral trans-form faults and concomitant SHmax that controlledthe orientation of J1 joint propagation. At the timeof their propagation, J1 joints were oriented east-southeast. However, post-Paleozoic continentaldrift carried these joints into their present orienta-tion parallel to the east-northeast orientation forSHmax of the contemporary tectonic stress field.A geological coincidence is the interpretation oflast resort and should be used sparingly, but thisis one such time when an alternative, parsimoniousinterpretation of J1 joints in black shale has failedto stand the test of time.We have presented exam-ples of other joints that are neotectonic (Hancockand Engelder, 1989).

Fractures in Eastern Gas Shales Project CoreOur observations and conclusions must be consis-tent with observations and conclusions based onthe EGSP core from the Appalachian Plateau ofNew York, Ohio, Pennsylvania, andWest Virginia(Cliffs Minerals, 1982). Fractures in ESGP core in-clude slickenside surfaces, coring-induced petal-centerline fractures, veins, and joints (Evans, 1994).Focusing just on joints and veins, the forelandward-most core (i.e., OH-4, OH-7, and PA-3) containsonly unmineralized joints, whereas fractures in thedeepest core (i.e., PA-2, PA-4, WV-6, WV-7, andWV-10) are nearly all mineralized veins. The corebetween these extremes is sparsely mineralized.Unmineralized joints in the foreland strike east-northeast and fall in either the J1 or J3 sets, whereasmineralized veins are more likely to belong to theJ2 set. The J2 veins in the EGSP core were inter-preted as Alleghanian, whereas the unmineralizedfractures were interpreted as neotectonic (Evans,1994). Our observations are consistent on twocounts. Veins are virtually absent on the forelandside of the ASF and common on the hinterlandside, particularly in the J2 orientation. The J1 jointspopulate all observed outcrops in the vicinity ofthe Finger Lakes region but are far less commonEngelder et al. 879

1vein filling, another form of healing, or a styloliticsurface. Furthermore, why do pervasively jointedDevonian gas shale outcrops carry so few veins?The preservation of unhealed joints is importantto gas production because healed fractures andveins would otherwise serve as barriers to gas flow.

This second conundrum is addressed by ana-lyzing the plumose structure on J1 and J2 jointsurfaces (Bahat and Engelder, 1984; Savalli andEngelder, 2005). Joint surface morphology, whichon the hinterland side of the ASF. To an approxi-mation, outcrops of the Valley and Ridge are aproxy for the deeper, hinterlandward EGSP core,whereas the outcrops in the Finger Lakes regionare a proxy for the shallower, forelandward EGSPcore. This would suggest that J1 joints are less com-mon to the Marcellus fairway than J2 joints. At thesame time, we doubt that any joints in EGSP core,however shallow, are neotectonic.

Conundrum 2: The Persistence of OpenJoints during Greater than 10%Layer-Parallel Shortening

The Appalachian Plateau detachment sheet wassubjected to greater than 10% LPS as indicatedby deformed fossils in the marine section of theCatskill delta throughout western New York andsouth to the ASF (Nickelsen, 1966; Engelder andEngelder, 1977; Geiser, 1988). This pattern ofLPS persists in sub-Marcellus rocks along the Hel-derburg Escarpment, which occurs north of theMarcellus Shale outcrop belt in western NewYork (Engelder, 1979b). There can be no doubtthat gas production comes from a black shale thatwas part of an allochthonous thrust sheet that ex-perienced greater than 10% LPS on the Appala-chian Plateau and as much as 50% in the Valleyand Ridge of Pennsylvania (Nickelsen, 1986). Evi-dence that the J1 set was present in the Marcellusbefore it was subjected to LPS is abundant (En-gelder and Whitaker, 2006). Still, surfaces of J1joints show little evidence of LPS, especially onthe foreland side of the Appalachian Basin. Atits core, conundrum 2 questions why a pervasivefabric produced by greater than 10% LPS does notmanifest itself on the surfaces of J joints as either880 E&P Noteis a product of the rupture process, is irregular on amicroscopic scale, whereas the overall surface of ajoint remains planar. The depth of the irregularityvaries with velocity of the rupture so that a contin-uous rupture may be distinguished from episodicjoint propagation (Figure 12). Joints driven by in-ternal pressure as natural hydraulic fractures leavea trail of incremental propagation and arrest events(Lacazette and Engelder, 1992). Gas productionand concomitant pressure buildup during burialmaturation of black shales is one of the mecha-nisms leading to natural hydraulic fracturing (Lashand Engelder, 2005).

Several lines of evidence point to the propa-gation of both J1 and J2 sets as natural hydraulicfractures. First, tensile joints cleave concretions,whereas the concretion acts as a barrier to naturalhydraulic fractures (McConaughy and Engelder,1999). A natural hydraulic fracture will propagatearound the concretion leaving the concretion intact(Figure 10). Second, episodic joint propagation isbest understood using Boyles Law for the behaviorof an ideal gas: P1V1 = P2V2, where P is pressureand V is volume (Lacazette and Engelder, 1992).The sudden rupturing of a joint and the conse-quent increase in volume cause the pressure with-in the joint to decrease thereby halting furtherpropagation. Evidence of incremental propagationindicates that pressure builds again until the rup-ture starts anew and that the source of fluid cannotfeed fluid to the growing joint at a speed sufficientto maintain continuous joint propagation. Fluid isfed to the joint volume through an interconnectedmatrix pore space on either side of the joint. The de-crease of pressure within the joint after each propa-gation cycle leads to an inward pressure gradientpromoting flow from the rock matrix to the openjoint. Renewed fluid flow into the joint causespressure to increase, which again elevates the stressintensity at the joint tip until another cycle of rup-ture commences, a process that repeats cycle aftercycle.

Role of MethaneIn addition to episodic propagation, joints hostedby rocks of the Devonian Catskill delta sequencedisplay progressively longer rupture increments

(Lacazette and Engelder, 1992). The rupture in-crement length scales with bed thickness, the ini-tial increments being shorter than bed thickness.After several dozen propagation episodes, rupturelength exceeds bed thickness. A gradual increasein increment length with joint growth is the man-ifestation of a compressible driving fluid, a prop-erty that water does not possess butmethane does.The long-term presence of methane also explainswhy so many joints observed in outcrops of theAppalachian Plateau show no indication of miner-alization. Methane within joints suppresses waterfilling and concomitant mineralization, thus pre-serving both J1 and J2 joints as unhealed, per-meable pathways in a tight-gas shale.

The major mechanisms for penetrative strainduring LPS include pressure solution andmechan-ical twinning of calcite (Engelder, 1979a, 1982b).Low-temperature, penetrative deformation is crit-ically dependent on water films at grain-grain con-tacts to allow the diffusion-mass transfer that en-ables pressure solution (Durney, 1972; Rutter,1983). Evidence for pressure solution within theMarcellus Shale of theValley andRidge is extensive,including pencil cleavage, deformedworm borrows,and stylolitic axial-planar cleavage (Figure 14). Thisdeformation mechanism also likely enabled bucklefolding within the Marcellus as well. Assuming theMarcellus entered the dry-gas window before orduring early folding, modern water saturation(Sw = 1020%) likely reflects Sw during folding.Such water saturation was apparently sufficientto allow pressure solution in the Marcellus Shalematrix during Alleghanian LPS. Equally, Sw alongmethane-filled, unhealed joints in the Marcelluswas likely insufficient to allow appreciable pres-sure solution and concomitant vein developmentat the scale of macroscopic joints. However, well-developed examples of pressure-solved joints areobserved in the Ordovician carbonates of the Val-ley and Ridge where Sw along joints must havebeen sufficient for the necessary diffusion masstransfer (Srivastava and Engelder, 1990). We con-clude that the ability of a methane fill to protectunhealed J1 joints from the ravages of diffusionmass transfer, including the deposition of veinmaterial, should not be underestimated.Alleghanian tectonics appears to have scat-tered joint planes on the hinterland side of theAFS, thereby affecting the tight cluster of polesto J1 joints described from stereographic projec-tions of data collected in the Finger Lakes region,New York (Figure 7). One possibility is that anLPS of 1015% just to the south of the ASF wassufficient to disrupt the otherwise well-aligned J1joints (Nickelsen, 1983). The problem with thishypothesis is that J1 joints on the AppalachianPlateau predate Alleghanian deformation with-out LPS having disrupted the tight cluster ofpoles to systematic joints (Figure 7). However,flexural folding may disrupt joint clustering inthe Marcellus to a much greater extent than LPSalone (Donath and Parker, 1964). Evidence ofthe two end members of flexural folding, flexural-slip folding, and flexural-flow folding is found inHamiltonGroup shales (Figure 14D). TheMarcel-lus Shale coarsens upward into the Mahantango,which carries bedding-parallel slickenside sur-faces spaced as closely as a few centimeters, amanifestation of flexural-slip folding. The Mar-cellus is more homogeneous and contains fewerdiscrete slip surfaces. However, axial-planar cleav-age is evident in lime-rich bedswithin theMarcellusShale (Figure 14D). The development of, and slipalong, the axial-planar cleavage is used to explainpassive folding (Alvarez et al., 1978), yet flexural-flow foldingmay also be an activemechanism duringfolding. Disruption of the early joint set and con-sequent scattering of poles to J1 joints in the Valleyand Ridgemay have been a consequence of flexural-flow folding. Furthermore, the sense of rotation forpoles to J1 joints is consistent with flexural flowpassively shearing bedding to rotate J1 joints awayfrom their original position normal to bedding,thereby leading to a joint population that has asteeper dip in upright bedding than would be thecase if joints tilted asmuch as bedding during folding(Figure 8B).

Unhealed Joints versus VeinsThe differences between J2 fracture developmentin deep EGSP core (nearly all mineralized; Evans,1994) and most of the unhealed joints in outcropon the hinterland side of the ASF (see Figure 9)Engelder et al. 881

present difficulties in interpretation. An outcropof Marcellus Shale along the Norfolk SouthernRailway line at Newton-Hamilton, Pennsylvania,contains a well-developed J2 joint set with lesscommon J2 veins. The same relative abundanceof J1 joints and veins has been observed at anynumber of outcrops of Middle and Upper Devo-nianmarine rocks in theValley andRidge. This ob-servation is common enough to suggest the possi-bility that unhealed (i.e., methane-filled) jointscan co-exist at depth with water-filled fracturesthat end up as veins. Here, the implication is thatwater invasion fails to drive gas from early joints,governed by a physical process similar to themecha-nism that allows the persistent water-over-gas con-tacts (Masters, 1984; Spencer, 1987).

The growth and preservation of unhealed J1joints are not consistent across the AppalachianBasin as indicated by comparison of their abun-dance in the Finger Lakes region relative to thatof the Pennsylvania Valley and Ridge (Figure 7).Throughout the northern Appalachian Basin,the J1 joints, where present, are mostly confinedto black shale. Where methane-driven joints prop-agated out of the black shale in the northern partof the basin, they did so as J2 joints, thus account-ing for the latters abundance in gray shale and silt-stone on either side of the ASF. This situation dif-fered along strike into the Virginia Valley andRidge where J1 joints appear in siltstone and finesandstones upsection from the nearest black shale(Engelder, 2004). In the Virginia part of the basin,methane-driven joints appear to first break outfrom organic-rich sections as J1 joints. This is alsotrue to the east of the Finger Lakes region where J1joints populate the more coarsely clastic part ofthe Genesee Group in a section of the Catskill del-ta that is inherently thicker (Younes and Engelder,1999).

Candidates for the early breakout of J1 jointsin the Pennsylvania Valley and Ridge have beendescribed from an outcrop of the Brallier Forma-tion of the Genesee Group at Huntingdon, Penn-sylvania (Ruf et al., 1998). Here, the Brallier con-tains both mineralized and unhealed J2 joints andunhealed strike joints that were originally inter-preted as fold related. Both unhealed joint sets882 E&P Noteare mostly confined to the silter beds in this distalturbidite section. The original and classic interpre-tation is that mineralized J2 joints formed first.However, abutting relationships suggest that theunhealed strike joints propagated before the un-healed J2 set. The strike joints fall within the rangeof strikes for J1 joints in the northern AppalachianMountains. This raises the possibility that thesestrike joints correlate with the J1 joints observedin coarser clastic rocks above black shale in Virginia.If so, the J1 joints in the Brallier at Huntingdon,Pennsylvania, display the characteristic geometryof J1 joints elsewhere in the Valley and RidgewhereJ1 joints fail to tilt over the requisite amount duringfolding to remain normal to bedding, an indicationof a small amount of flexural-flow folding (Rufet al., 1998). One interpretation is that unhealedJ1 joints were methane filled thereby preventingthe invasion of the water, a requisite for miner-alization at the time of propagation and mineralfilling of some J2 joints. Again, this raises the possi-bility that water-filled joints can invade a methane-saturated rock without displacing methane fromearlier methane-filled joints, a lesson that may ap-ply to a slickwater fracture stimulation of the Mar-cellus. Clearly, the degree of structural growthcorrelates with vein development in Devoniangas shales (Evans, 1994). Structural growth maypromote interformational fracturing that allowswater invasion and concomitant vein filling, butsuch water invasion is incapable of displacingmethane from early joints, a necessary conditionfor the persistence of joints through multiple tec-tonic cycles.

J3(?) in the Hudson ValleyThe orientation of east-northeast joints in theHudson Valley does not mimic the behavior ofJ1 joints in the Finger Lakes District (Figure 7).At Kingston, New York, where the lower part ofthe Bakoven (Union Springs equivalent) Shale ofthe Marcellus is exposed, the best developed jointset has a vectormean strike of 059. This set exhib-its four characteristics that may be more consis-tent with exhumation-related neotectonic jointsof the Appalachian Plateau. First, those joints taller

than ameter or two appear to have propagated outof plane. Second, the outcrop contains many shortjoints with vertical growth restricted to less than0.5 m (1.6 ft). Third, these joints do not clusterlike their counterparts in the Finger Lakes Districtof New York. Fourth, the vector mean strike ofthese joints falls within the range of the contem-porary tectonic stress field in the eastern UnitedStates (058069).

The pattern of east-northeast joints in the Ba-koven Shale at Catskill is closer to that at Kingstonthan found in black shale of the Finger Lakes Dis-trict. Although the outcrop gives the impressionof a robust vertical growth of joints, most jointshave visible top and/or bottom tips with verticalgrowth less than 2 m (6 ft). A weaker clusteringof poles may be amanifestation of curving growth.Moreover, growth is not orthogonal to bedding,an indication that joints were not tilted duringfolding (Figure 7). In places, the Bakoven containsbed-parallel veins against which the joints abut, asign that the joints postdate vein growth. Neitheroutcrop in the Hudson Valley displays a well-developed J2 set, evidence that a gas drive froma source rock may never have been present. Coin-cidentally, TOC of the Bakoven Shale of the Mar-cellus Formation is lower than its lateral equivalentin the distal foreland, the Union Springs Member,perhaps reflecting a greater dilution by clasticdetritus.

In summary, joints in the two best exposuresof theMarcellus along the Hudson Valley forelandfold and thrust belt give the impression of belong-ing to the J3 neotectonic set. If ever present, the J1set was overprinted by J2 during the strong LPSproduced by Hudson Valley foreland deformation(Marshak and Engelder, 1985). Alternatively,thermal maturation of the Marcellus during theDevonian Acadian orogeny in the Hudson Valleyforeland fold and thrust belt predated J1 jointing(Marshak and Tabor, 1989). Other outcrops ofthe Marcellus where J1 and J3 joints appear to-gether serve as the strongest evidence that J1 jointssurvive LPS in theValley and Ridge of Pennsylvania,so we conclude that J1 joints were never presentin the organically leaner Marcellus of the HudsonValley foreland (Figure 5B).Conundrum 3: Generating Vertical Joints in aThurst-Fault Stress Regime

Fold and thrust belts form mostly by the thruststacking of detached sections. The major stiff layerin the Appalachian Valley and Ridge is a greaterthan 2-km-thick CambrianOrdovician carbonatesection (Hatcher et al., 1989). Such thrust fault-ing is presumably characterized by a state of stressin which the least principal stress, s3, is vertical(Anderson, 1951). Such a stress regime is consis-tent with thrust ramp dips in the AppalachianValley and Ridge of 20 to 25. Yet, the major syn-tectonic joint set, J2, is vertical and in the cross-fold orientation. One of the striking features con-cerning the orientation of joints described fromcore recovered during the EGSP is the radial pat-tern that follows the oroclinal bend of the cen-tral Appalachian fold and thrust belt (Evans,1994). Likewise, outcrop mapping of the Devo-nian Brallier Formation along the Allegheny Frontreveals a similar radial pattern of cross-strike joints(Figure 9). This joint set may be the primary tar-get for a stimulation-driven fracture from hori-zontal wells within the Appalachian Plateau de-tachment sheet.

The horizontal s3 necessary to form verticaljoints within a thrust-fault regime may be ex-plained by the extension fracture hypothesis,which holds that joints are driven by the joint-parallel principal stress, SHmax (Lorenz et al.,1991). Extension fractures are the product of lab-oratory compression experiments at low confiningpressure where samples split end to end in the di-rection of maximum compression even when allmacroscopic principal stresses are compressive(Griggs, 1936). The paradox concerning extensionfractures is that they propagate in the absence ofmacroscopic tensile stress. However, there is con-vincing experimental evidence that the extensionfracturing of a brittle material is due to a wedgingaction such that a local tensile stress is developedat the point of the wedge (Griggs and Handin,1960, p. 351). Tension also develops as a conse-quence of slip on internal cracks subjected to shearstress, and such tension produces wing cracks(Brace and Bombolakis, 1963; Cruikshank et al.,Engelder et al. 883

1991). For most geologists, the allure of extensionfracturing does not serve as a model for verticaljoints because fractures originate in response tolocal tensile stresses around flaws or cracks on amicroscopic scale (Paterson, 1978, p. 19). Al-though the axial splitting mechanism is a perfectlyadequate mechanism for the origin of sheet frac-tures in the near surface, it fails to explain regional,large-scale joint propagation at depth in a sedi-mentary basin.

Crosscutting relationships (Srivastava and En-gelder, 1990) indicate that the final episode ofcrack propagation and cross-fold vein develop-ment in the carbonate thrust sheets of the Val-ley and Ridge Appalachians occurred after theCambrianOrdovician carbonates were emplacedonto the upper flat of thrust duplexes. The orien-tation of the veins indicates that s3 was parallel tothe strike of the host rock and local fold axes. Suchan orientation of s3 within the thrust-faultingstate of stress may reflect the influence of hori-zontal contraction caused by cooling and removalof overburden during syntectonic erosion. An-other mechanism for reducing strike-parallelstress is strike-parallel stretching to accommodatean increase in the radius of curvature around anoroclinal bend. Strike-parallel stretching is also re-quired to accommodate lateral ramps in a forelandfold and thrust belt.

HORIZONTAL VERSUS VERTICAL WELLS

The presence of systematic J1 joints in Marcellusoutcrops on either side of the deep central regionof the Appalachian Basin increases the probabil-ity that the J1 joint set will be found in the Mar-cellus at depth. Reports of J1 joints appearing inFormation MicroImager (FMI) images of recentwells penetrating the Marcellus confirm its pres-ence at drilling depths (Figure 2). Furthermore, J1joints are abundant in Huron (Dunkirk equivalent)black shale EGSP cores (Figure 4). The presence ofJ1 and J2 joints in the EGSP cores assures that, insome parts of the Appalachian Basin, joints of bothsets propagated at depths in excess of 2 km (CliffsMinerals, 1982; Evans, 1994).884 E&P NotePerhaps one of the earliest hints that the Mar-cellus contains unhealed joints at depth camefrom blowouts as early wells penetrated the Mar-cellus to exploit the gas found in the deeper De-vonian Oriskany Sandstone of the AppalachianBasin (Bradley and Pepper, 1938). A notable ex-ample is the April 3, 1940, Crandall Farm blow-out near Independence, New York, where gasproduction reached 60mmcf during the first eightdays of uncontrolled flow from the upper part ofthe Marcellus formation at a depth of 4800 ft(1463 m) (Taylor, 2009). Because this productioncame from an unstimulated vertical well thatlacked any evidence of faulting that could havetapped gas from the Oriskany some 120 ft (36 m)below, the interpretation is that the blowout wasfed by self-sourced joints within the MarcellusShale. Over the following half century, blowoutswere a common consequence of drilling verticalwells penetrating the Marcellus. The low perme-ability of the Marcellus Shale suggests that many,if not all, blowouts must have tapped a reservoirof interconnected natural fractures. In fact, blow-outs were one of the major attractions drawingRange resources to Washington County, Pennsyl-vania, where Range started targeting theMarcellusgas shale during 2004 (W. A. Zagorski, 2009, per-sonal communication).

Although both J1 and J2 joints in black shaleare natural hydraulic fractures and, consequently,virtually identical in terms of aperture and surfaceroughness, two important differences betweenthese joint sets are observed relative to engineer-ing and completion techniques necessary to max-imize the production of natural gas. First, unmin-eralized joints subjected to lower normal stresswill be more permeable (Kranz et al., 1979).The least horizontal normal stress, Shmin, in theAppalachian Basin is nearly perpendicular to J1joints, meaning that, all other things being equal,stress-controlled permeability favors J1 joints overJ2 joints. This may be the seminal characteristicfor unstimulated production from horizontalwells in the Huron (Dunkirk equivalent) blackshale of Kentucky. Second, J1 joints are better de-veloped and more closely spaced than J2 joints inorganic-rich rocks (Loewy, 1995; Lash et al.,

stresses in the contemporary tectonic stress field.Hence, a higher joint density and stress-induced2004). Thus, even if J1 and J2 joints have the samepermeability, the host black shale will exhibitgreater bulk permeability in the J1 direction.

The completion of vertical wells may involvelarge hydraulic fracture treatments. Induced hy-draulic fractures will propagate east-northeastwest-southwest, the direction of SHmax of thecontemporary tectonic stress field (Evans et al.,1989). Hydraulic fractures propagating in thisdirection may travel along J1 joints to intersectJ2 joints. In this case, the major drainage pathto the well is first along J2 joints and then alongthe east-northeasttrending hydraulic fracture,which may or may not have propagated alongpre-folding J1 joints. By intersecting J2 joints,hydraulic fracture treatments in vertical wellsare capable of taking advantage of neither thebulk permeability anisotropy of black shale northe normal-stress induced permeability anisot-ropy of J1 vs. J2 joints. Still, the ability of J2 jointsto deliver natural gas to the plane of an artificialhydraulic fracture should not be underestimatedfor those operators wishing to use the conventionalvertical completion techniques in the MarcellusShale.

Given that the contemporary tectonic stressfield controls the propagation direction of hydrau-lic fractures across J2 joints, the only practicalmeans of immediately communicating with themore permeable J1 joint set is by horizontal dril-ling in a north-northwest or south-southeast di-rection. In this case, artificial hydraulic fracturesgenerated in the horizontal part of the wellborewill propagate in the direction of SHmax, whichis parallel to J1 joints. The likely outcome of ahydraulic fracture treatment in a horizontal welldrilled to the west-northwest, for example, is thereopening of J1 instead of the fracturing of intactblack shale. Hence, horizontal drilling in Devo-nian black shale should be directed to the north-northwest, perpendicular to SHmax of the con-temporary tectonic stress field, to benefit fromboth the bulk permeability anisotropy from jointdevelopment and the normal-stress-induced per-meability anisotropy of these rocks. This is thepractice established by Marcellus Shale operatorsthrough early 2009.permeability anisotropy in Devonian black shalespeak to the advisability of horizontal drilling to-ward the north-northwest or south-southeast tocross the more densely formed J1 systematic jointset subjected to a lower normal stress in the con-temporary tectonic stress field.

The presence of systematic J1 joints in out-crops of the Marcellus Shale on either side ofthe deep central region of the Appalachian Basinincreases the probability that the J1 set will befound in the Marcellus at depth. A compilationof proprietary FMI images from recent wells pene-trating the Marcellus Shale confirms the presenceof the J1 set at depth (Figure 2).

REFERENCES CITED

Alvarez, W., T. Engelder, and P. A. Geiser, 1978, Classifica-tion of solution cleavage in pelagic limestones: Geology,v. 6, p. 263266, doi:10.1130/0091-7613(1978)62.0.CO;2.

Anderson, E. M., 1951, The dynamics of faulting and dykeformation with application to Britain, 2d ed.: London,Oliver and Boyd, 206 p.

Aydin, A., and A. M. Johnson, 1978, Development of faultsas zones of deformation bands and as slip surfaces insandstone: Pure and Applied Geophysics, v. 116,p. 931942, doi:10.1007/BF00876547.

Bahat, D., and T. Engelder, 1984, Surface morphology oncross-fold joints of the Appalachian Plateau, New Yorkand Pennsylvania: Tectonophysics, v. 104, p. 299313,doi:10.1016/0040-1951(84)90128-8.

Bowker, K. A., 2007, Barnett Shale gas production, FortCONCLUSIONS

Successful horizontal drilling of unconventionalreservoirs may depend on the presence of sys-tematic fractures. The J1 joints formed preferen-tially in Devonian black shale throughout the Ap-palachian Basin early in the Alleghanian tectoniccycle as a consequence of burial-related thermalmaturation of kerogen to hydrocarbons. This jointset, the most closely spaced in black shale, is noworiented parallel with SHmax of the contemporarystress field. Black shale also carries a less well-developed younger joint set (J2) which, by virtueof its orientation, is subjected to higher normalEngelder et al. 885

Worth Basin: Issues and discussion: AAPG Bulletin,v. 91, p. 523533, doi:10.1306/06190606018.

Brace, W. F., and E. G. Bombolakis, 1963, A note on brit-tle crack growth in compression: Journal of Geophys-ical Research, v. 68, p. 37093713, doi:10.1029/JZ068i012p03709.

Bradley, W. H., and J. F. Pepper, 1938, Structure and gaspossibilities of the Oriskany Sandstone in Steuben,Yates, and parts of the adjacent counties, New York:U.S. Geological Survey Bulletin, v. 899-A, p. 68.

Brown, E. T., and E. Hoek, 1978, Trends in relationships be-tween measured in-situ stresses and depth: InternationalJournal of Rock Mechanics and Mining Sciences andGeomechanics Abstracts, v. 15, p. 211215.

Browning, I. B., 1935, Relation of structure to shale gas accu-mulation: Devonian shalesA symposium by the Appa-lachian Geological Society: Charleston, West Virginia,Appalachian Geological Society, p. 1620.

Campbell, G., 1946, New Albany Shale (Indiana, Kentucky,Ohio, Tennessee, Alabama): Geological Society ofAmerica Bulletin, v. 57, p. 829908, doi:10.1130/0016-7606(1946)57[829:NAS]2.0.CO;2.

Clark, J. A., 1982, Glacial loading: A cause of natural frac-turing and a control of the present stress state in regionsof high Devonian shale gas: Unconventional Gas Recov-ery Symposium, Pittsburgh, Pennsylvania, May 1618,Society of Petroleum Engineers Paper No. 10798,p. 8891.

Cliffs Minerals, 1982, Analysis of the Devonian shales in theAppalachian Basin: Final report contract DE-AS21-80MC14693, in U.S. DOE, ed.: Washington, D.C.,Springfield Clearing House, p. 314.

Craddock, J. P., and B. A. van der Pluijm, 1989, Late Paleozoicdeformation of the cratonic carbonate cover of easternNorth America: Geology, v. 17, p. 416419, doi:10.1130/0091-7613(1989)0172.3.CO;2.

Cruikshank, K. M., G. Zhao, and A. M. Johnson, 1991, Anal-ysis of minor fractures associated with joints and faultedjoints: Journal of Structural Geology, v. 13, p. 865886,doi:10.1016/0191-8141(91)90083-U.

Curtis, J. B., 2002, Fractured shale-gas systems: AAPG Bul-letin, v. 86, p. 19211938.

Curtis, J. B., and G. Faure, 1997, Accumulation of organicmatter in the Rome trough of the Appalachian Basinand its subsequent thermal history: AAPG Bulletin,v. 81, p. 424437.

Davis, D. M., and T. Engelder, 1985, The role of salt in foldand thrust belts, in N. L. Carter and S. Uyeda, eds., Tec-tonophysics: Amsterdam, Elsevier, v. 119, p. 6788.

Dean, S., M. Baranoski, L. Bertoli, G. Kribbs, T. Stephens, B.Kulander, D. Sochman, and D. Mumpower, 1984, Re-gional fracture analysis in western Valley and Ridge andadjoining plateau, West Virginia and Maryland: AAPGBulletin, v. 67, p. 448.

Desantis, M. K., C. E. Brett, and C. A. ver Straeten, 2007,Persistent depositional sequences and bioevents in theEifelian (early Middle Devonian) of eastern Laurentia:North American evidence of the Kacak events?: Geolog-ical Society Special Publications 278, p. 83104.

Donath, F. A., and R. B. Parker, 1964, Folds and folding:886 E&P NoteGeological Society of America Bulletin, v. 75, p. 4562,doi:10.1130/0016-7606(1964)75[45:FAF]2.0.CO;2.

Durham, L. S., 2008, Guess what? Its complex: Woodfordjoins shale parade: AAPG Explorer, v. 29, p. 26.

Durney, D.W., 1972, Solution-transfer, an important geolog-ical deformation mechanism: Nature, v. 235, p. 315317, doi:10.1038/235315a0.

Engelder, T., 1979a, Mechanisms for strain within the UpperDevonian clastic sequence of the Appalachian Plateau,western New York: American Journal of Science,v. 279, p. 527542.

Engelder, T., 1979b, The nature of deformation within theouter limits of the central Appalachian foreland foldand thrust belt in New York state: Tectonophysics,v. 55, p. 289310, doi:10.1016/0040-1951(79)90181-1.

Engelder, T., 1982a, Is there a genetic relationship betweenselected regional joints and contemporary stress withinthe lithosphere of North America?: Tectonics, v. 1,p. 161177, doi:10.1029/TC001i002p00161.

Engelder, T., 1982b, A natural example of simultaneous op-eration of free-face dissolution and pressure solution:Geochemica et Cosmochimica Acta, v. 46, p. 6974,doi:10.1016/0016-7037(82)90291-5.

Engelder, T., 1985, Loading paths to joint propagation duringa tectonic cycle: An example from the Appalachian Pla-teau, U.S.A., in P. L. Hancock and C. M. Powell, eds.,Journal of Structural Geology: Amsterdam, Elsevier,v. 7, p. 459476.

Engelder, T., 1993, Stress regimes in the lithosphere: Prince-ton, New Jersey, Princeton Press, 451 p.

Engelder, T., 2004, Tectonic implications drawn from differ-ences in the surface morphology on two joint sets in theAppalachian Valley and Ridge, Virginia: Geology, v. 32,p. 413416, doi:10.1130/G20216.1.

Engelder, T., and R. Engelder, 1977, Fossil distortion and de-collement tectonics of the Appalachian Plateau: Geol-ogy, v. 5, p. 457460, doi:10.1130/0091-7613(1977)52.0.CO;2.

Engelder, T., andM. P. Fischer, 1996, Loading configurationsand driving mechanisms for joints based on the Griffithenergy-balance concept: Tectonophysics, v. 256,p. 253277, doi:10.1016/0040-1951(95)00169-7.

Engelder, T., and P. Geiser, 1980, On the use of regional jointsets as trajectories of paleostress fields during the devel-opment of the Appalachian Plateau, New York: Journalof Geophysical Research, v. 85, p. 63196341, doi:10.1029/JB085iB11p06319.

Engelder, T., and M. R. Gross, 1993, Curving cross joints andthe lithospheric stress field in eastern North America:Geology, v. 21, p. 817820, doi:10.1130/0091-7613(1993)0212.3.CO;2.

Engelder, T., and A. Lacazette, 1990, Natural hydraulic frac-turing, in N. Barton and O. Stephansson, eds., Rockjoints: Rotterdam, A.A. Balkema, p. 3544.

Engelder, T., and A. Whitaker, 2006, Early jointing in coaland black shale: Evidence for an Appalachian-wide stressfield as a prelude to the Alleghanian orogeny: Geology,v. 34, p. 581584, doi:10.1130/G22367.1.

Ettensohn, F. R., 1985, The Catskill delta complex and theAcadian orogeny: A model, in D. L. Woodrow and

W.D. Sevon, eds., The Catskill delta: Geological Societyof America Special Paper 201, p. 3949.

Evans, K. F., 1989, Appalachian stress study: 3. Regionalscale stress variations and their relation to structureand contemporary tectonics: Journal of Geophysical Re-search, v. 94, p. 17,61917,645.

Evans, M. A., 1994, Joints and decollement zones in MiddleDevonian shales; evidence for multiple deformationevents in the central Appalachian Plateau: GeologicalSociety of America Bulletin, v. 106, p. 447460,doi:10.1130/0016-7606(1994)1062.3.CO;2.

Evans, K. F., T. Engelder, and R. A. Plumb, 1989, Appala-chian stress study: 1. A detailed description of in situstress variations in Devonian shales of the AppalachianPlateau: Journal of Geophysical Research, v. 94,p. 71297154, doi:10.1029/JB094iB06p07129.

Fischer, M. P., M. R. Gross, T. Engelder, and R. J. Greenfield,1995, Finite-element analysis of the stress distributionaround a pressurized crack in a layered elastic medium;implications for the spacing of fluid-driven joints inbedded sedimentary rock: Tectonophysics, v. 247,p. 4964, doi:10.1016/0040-1951(94)00200-S.

Fontaine, J., N. Johnson, and D. Schoen, 2008, Design, exe-cution, and evaluation of a typical Marcellus Shaleslickwater stimulation: A case history: 2008 Eastern Re-gional/AAPG Eastern Section Joint Meeting, Pittsburgh,Pennsylvania, October 1115, Society of Petroleum En-gineers Paper No. 117772, 11 p.

Geiser, P. A., 1988, The role of kinematics in the construc-tion and analysis of geological cross sections in deformedterranes, inG.Mitra and S.Wojtal, eds., Geometries andmechanisms of thrusting: Geological Society of AmericaSpecial Paper 222, p. 4776.

Gerber, M., 2008, BMO capital markets 4th Annual Appala-chian E&P forum: http://audability.com/AudabilityAdmin/Clients/BMO/10566_110200880000AM/lobby.aspx?Event_ID=566 (accessed January 31, 2009).

Griggs, D. T., 1936, Deformation of rocks under high confin-ing pressures: Journal of Geology, v. 44, p. 541577.

Griggs, D. T., and J. Handin, 1960, Observations on fractureand a hypothesis of earthquakes, in D. T. Griggs and J.Handin, eds., Rock deformation: Geological Society ofAmerica Memoir 79, p. 247264.

Gross, M. R., 1993, The origin and spacing of cross joints; ex-amples from the Monterey Formation, Santa Barbaracoastline, California: Journal of Structural Geology,v. 15, p. 737751, doi:10.1016/0191-8141(93)90059-J.

Gross, M. R., and T. Engelder, 1991, A case for neotectonicjoints along the Niagara escarpment: Tectonics, v. 10,p. 631641, doi:10.1029/90TC02702.

Gwinn, V. E., 1964, Thin skinned tectonics in the Plateauand northwestern Valley and Ridge provinces of the cen-tral Appalachians: Geological Society of America Bulle-tin, v. 75, p. 863900, doi:10.1130/0016-7606(1964)75[863:TTITPA]2.0.CO;2.

Hall, J., 1843, Natural history of New York: IV. Comprisingthe survey of the Fourth Geological District: Albany,New York, Carrol & Cook, 525 p.

Hancock, P. L., and T. Engelder, 1989, Neotectonic joints:Geological Society of America Bulletin, v. 101,p. 11971208, doi:10.1130/0016-7606(1989)1012.3.CO;2.

Handin, J. W., and R. V. Hager Jr., 1957, Experimental de-formation of sedimentary rocks under confining pres-sure: Part 1. Tests at room temperature on dry samples:Part 2. Tests at high temperature: AAPG Bulletin, v. 41,p. 150.

Hatcher, R. D., W. A. Thomas, P. A. Geiser, A. W. Snoke, S.Mosher, and D. V. Wiltschko, 1989, Alleghanian oro-gen, in R. D. Hatcher, W. A. Thomas, and G. W. Viele,eds., The AppalachianOuachita orogen in the UnitedStates: Boulder, Colorado, The Geological Society ofAmerica, v. F-2, p. 233318.

Hodgson, R. A., 1961, Classification of structures on joint sur-faces: American Journal of Science, v. 259, p. 493502.

Holzhausen, G. R., 1989, Origin of sheet structures: 1. Mor-phology and boundary conditions: Engineering Geology,v. 27, p. 225278, doi:10.1016/0013-7952(89)90035-5.

Hopkins, C. W., R. L. Rosen, and D. G. Hill, 1998, Charac-terization of an induced hydraulic fracture completion ina naturally fractured Antrim Shale reservoir: Eastern Re-gionalMeeting, Pittsburgh, Pennsylvania, November 911,Society of PetroleumEngineers Paper No. 51068, p. 177185.

Hunter, C. D., and D. M. Young, 1953, Relationship of naturalgas occurrence and production in eastern Kentucky (BigSandy gas field) to joints and fractures in Devonian bitu-minous shale: AAPG Bulletin, v. 37, p. 282299.

Kaufmann, B., 2006, Calibrating the Devonian time scale: Asynthesis of U-Pb ID-TIMS ages and conodont stratigra-phy: Earth Science Reviews, v. 76, p. 175190.

Kranz, R. L., A. D. Frankel, T. Engelder, and C. H. Scholz,1979, Permeability of whole and jointed Barre granite:International Journal of Rock Mechanics and MiningSciences and Geomechanics Abstracts, v. 16, p. 225234.

Lacazette, A., and T. Engelder, 1992, Fluid-driven cyclic prop-agation of a joint in the Ithaca siltstone, Appalachian Ba-sin, New York, in B. Evans and T.-F. Wong, eds., Faultmechanics and transport properties of rocks: London,Academic Press, p. 297324.

Lash, G., S. Loewy, and T. Engelder, 2004, Preferential joint-ing of Upper Devonian black shale, Appalachian Pla-teau, U.S.A.: Evidence supporting hydrocarbon genera-tion as a joint-driving mechanism: Geological SocietySpecial Publications 231, p. 129151.

Lash, G. G., and T. Engelder, 2005, An analysis of horizontalmicrocracking during catagenesis: Example from theCatskill delta complex: AAPG Bulletin, v. 89, p. 14331449, doi:10.1306/05250504141.

Lash, G. G., and T. Engelder, 2007, Jointing within the outerarc of a forebulge at the onset of the Alleghanian orog-eny: Journal of Structural Geology, v. 29, p. 774786,doi:10.1016/j.jsg.2006.12.002.

Lash, G. G., and T. Engelder, 2009, Tracking the burial andtectonic history of Devonian shale of the AppalachianBasin by analysis of joint intersection style: GeologicalSociety of America Bulletin, v. 121, p. 265277,doi:10.1130/B26287.1.Engelder et al. 887

Law, B. E., 2002, Basin-centered gas systems: AAPG Bulletin,v. 86, p. 18911919.

Lawn, B., 1993, Fracture of brittle solids: Cambridge, Cam-bridge University Press, 378 p.

Loewy, S. L., 1995, The post-Alleghanian tectonic history ofthe Appalachian Basin based on joint patterns in Devo-nian black shales: M.S. thesis, Pennsylvania State Uni-versity, University Park, Pennsylvania, 179 p.

Lorenz, J. C., L. W. Teufel, and N. R. Warpinski, 1991, Re-gional fractures I: A mechanism for the formation of re-gional fractures at depth in flat-lying reservoirs: AAPGBulletin, v. 75, p. 17141737.

Marshak, S., and T. Engelder, 1985, Development of cleav-age in limestones of a fold-thrust belt in eastern NewYork, in P. L. Hancock and C. M. Powell, eds., Journal ofStructural Geology: Amsterdam, Elsevier, v. 7, p. 345359.

Marshak, S., and J. R. Tabor, 1989, Structure of the Kingstonorocline in the Appalachian fold-thrust belt, New York:Geological Society of America Bulletin, v. 101, p. 683701, doi:10.1130/0016-7606(1989)1012.3.CO;2.

Martel, S. J., 2006, Effect of topographic curvature on near-surface stresses and application to sheeting joints: Geo-physical Research Letters, v. 33, L01208, p. 5.

Masters, J. A., 1984, Lower Cretaceous oil and gas in westernCanada, in J. A. Masters, ed., Elmworth: Case study of adeep basin gas field: AAPG Memoir 38, p. 135.

McConaughy, D. T., and T. Engelder, 1999, Joint interac-tion with embedded concretions: joint loading config-urations inferred from propagation paths: Journal ofStructural Geology, v. 21, p. 16371652, doi:10.1016/S0191-8141(99)00106-6.

Milici, R. C., 2005, Assessment of undiscovered natural gasresources in Devonian black shales, Appalachian Basin,eastern U.S.A.: U.S. Geological Survey Open-File Re-port 2005-1268, v. 1.0, p. A93.

Nadan, B. J., and T. Engelder, 2009, Microcracks in NewEngland granitoids: A record of thermoelastic relaxationduring exhumation of intracontinental crust: GeologicalSociety of America Bulletin, v. 121, p. 8099,doi:10.1130/B26202.1.

Nevin, C. M., 1931, Principles of structural geology: NewYork, John Wiley & Sons, 303 p.

Nickelsen, R. P., 1966, Fossil distortion and penetrative rockdeformation in the Appalachian Plateau, Pennsylvania:Journal of Geology, v. 74, p. 924931.

Nickelsen, R. P., 1976, Early jointing and cumulative fracturepatterns: Utah Geological Association Publication, v. 5,p. 193199.

Nickelsen, R. P., 1983, Aspects of Alleghanian deformation,in R. P. Nickelsen and E. Cotter, eds., Silurian deposi-tional history and Alleghanian deformation in the Penn-sylvania Valley and Ridge: Guidebook for the 48th An-nual Field Conference of Pennsylvania Geologists,Pennsylvania: Harrisburg, Pennsylvania Geological Survey,p. 2939.

Nickelsen,R. P., 1986,Cleavageduplexes in theMarcellus Shaleof theAppalachian foreland: Journal of StructuralGeology,v. 8, p. 361371, doi:10.1016/0191-8141(86)90055-6.

Nickelsen, R. P., and V. N. D. Hough, 1967, Jointing in the888 E&P NoteAppalachian Plateau of Pennsylvania: Geological Societyof America Bulletin, v. 78, p. 609629, doi:10.1130/0016-7606(1967)78[609:JITAPO]2.0.CO;2.

Nunn, J. A., 1996, Buoyancy-driven propagation of isolatedfluid-filled fractures: Implications for fluid transport inGulf of Mexico geopressured sediments: Journal of Geo-physical Research, v. 101, p. 29632970, doi:10.1029/95JB03210.

Olson, J., and D. D. Pollard, 1989, Inferring paleostressesfrom natural fracture patterns: A new method: Geology,v. 17, p. 345348, doi:10.1130/0091-7613(1989)0172.3.CO;2.

Over, D. J., 2002, The Frasnian/Famennian boundary in cen-tral and eastern United States: Palaeogeography, Palaeo-climatology, Palaeoecology, v. 181, p. 153169,doi:10.1016/S0031-0182(01)00477-1.

Over, D. J., 2007, Conodont biostratigraphy of the Chatta-nooga Shale, Middle and Upper Devonian, southern Ap-palachian Basin, eastern United States: Journal of Paleon-tology, v. 81, p. 11941217, doi:10.1666/06-056R.1.

Over, D. J., R. Lazar, G. C. Baird, J. Schieber, and F. R.Ettensohn, 2009, Protosalvinia Dawson and associatedconodonts of the upper trachytera zone, Famennian,Upper Devonian, in the eastern United States: Journalof Paleontology, v. 83, p. 7079, doi:10.1666/08-058R.1.

Overby,W. K., A. B. Yost II, and D. A.Wilkins, 1988, Induc-ing multiple hydraulic fractures from a horizontal well-bore: 1988Annual Technical Conference andExhibition,Houston, Texas, October 25, Society of Petroleum En-gineers Paper No. 18249.

Parker III, J. M., 1942, Regional systematic jointing in slightlydeformed sedimentary rocks: Geological Society ofAmerica Bulletin, v. 53, p. 381408.

Pashin, J. C., and F. Hinkle, 1997, Coalbed methane inAlabama: Geological Survey of Alabama, CircularReport 192, p. 71.

Paterson, M. S., 1978, Experimental rock deformationThebrittle field: New York, Springer-Verlag, 254 p.

Plumb, R. A., 1994, Variations of the least horizontal stressmagnitude in sedimentary rocks, in P. P. Nelson and S. E.Laubach, eds., Rock mechanics models and measure-ments: Challenges from industry: Rotterdam, Balkema,p. 7178.

Plumb, R. A., and S. H. Hickman, 1985, Stress-induced bore-hole elongation: A comparison between the four-armdipmeter and the borehole televiewer in the Auburngeothermal well: Journal of Geophysical Research,v. 90, p. 55135521, doi:10.1029/JB090iB07p05513.

Pollard, D. D., and A. Aydin, 1988, Progress in understand-ing jointing over the past century: Geological Society ofAmerica Bulletin, v. 100, p. 11811204, doi:10.1130/0016-7606(1988)1002.3.CO;2.

Price, N. J., 1966, Fault and joint development in brittle andsemi-brittle rock: London, Pergamon Press, 176 p.

Rawnsley, K. D., T. Rives, J.-P. Petit, S. R. Hencher, andA. C. Lumsden, 1992, Joint development in perturbedst