The Distribution of Natural Fractures and Joints at Depth in Crystalline Rock

Home , Fracture (geology), Joint (geology)

University of Nebraska - Lincoln DigitalCommons@University of Nebraska - Lincoln

USGS Staff -- Published Research US Geological Survey

1982

Donald A. Seeburger Stanford University

Mark D. Zoback U.S. Geological Survey, [email protected]

Follow this and additional works at: https://digitalcommons.unl.edu/usgsstaffpub

Part of the Earth Sciences Commons

Seeburger, Donald A. and Zoback, Mark D., "The Distribution of Natural Fractures and Joints at Depth in Crystalline Rock" (1982). USGS Staff -- Published Research. 454. https://digitalcommons.unl.edu/usgsstaffpub/454

This Article is brought to you for free and open access by the US Geological Survey at DigitalCommons@University of Nebraska - Lincoln. It has been accepted for inclusion in USGS Staff -- Published Research by an authorized administrator of DigitalCommons@University of Nebraska - Lincoln. JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 87, NO. B7, PAGES 5517-5534,JULY 10, 1982

The Distribution of Natural Fractures and Joints at Depth in Crystalline Rock

DONALD A. SEEBURGER

GeophysicsDepartment, Stanford University, Stanford, California 94305

MARK D. ZOBACK

U.S. GeologicalSurvey, Menlo Park, California 94025

This paperpresents the resultsof studiesof the natural fracturedistribution encountered in 10 testwells drilledin threeareas of the UnitedStates. Seven of the wellswere drilled to depthsof 200-250m, while threewere drilled to depthsof about 1 kin. Using an ultrasonicborehole televiewer, fracture depths, strikes, and dips were determined.Steeply dipping fractureswere found throughout each of the wells, and in general,few horizontalfractures were observed.Statistically significant fracture pole concentrationswere found for each well which were basicallyinvariant with depth, although some variation of fracture orientationwith depth was found in two wells.The significantfracture orientations were not found to be the samein wellsonly severalkilometers apart in a givenregion. In none of the wellsdid the numberof observablefractures decrease markedly with increasingdepth. No simplerelationship of fractureorientation or fracturedensity with major structuralfeatures such as the San Andreasfault wereobserved, and no simplerelation between the significant fracture orientations and either past or presentregional stress fields could be determined.

INTRODUCTION existence and the orientation of the fractures were determined; The upper crust of the earth is composedof fracturedrock. most boreholestudies have beenconcerned with only the pres- Fractures are found on all scales from microfractures with ence of fractures. dimensionsof microns ,to lineaments with dimensionsof kilo- In this paper we presentthe resultsof in situ fracturestudies meters.These fractures exert a profoundeffect upon the physi- performed in test wells drilled in three different regions' five cal propertiesof rock. Laboratory and theoretical fracture wells were in the Mojave Desert near Palmdale, California; studies have been extensive. Fractures have been found to affect threewere near Limekiln Valley, southof Hollister,California; the strengthof rock [e.g., Brace, 1960], the velocity of elastic and two wells were at Monticello Reservoir, northwest of waves[e.g., Nut, 1971;Hadley, 1976], and the effectiveelastic Columbia,South Carolina. The wellswere continuously logged moduli [e.g.,Walsh, 1965; Budiansky and O'Connell,1976]. The using an ultrasonicborehole televiewer (described below), and permeabilityand distribution ofpore fluids in crystallinerocks the fractureorientations and distribution as a functionof depth are primarilydetermined by the densityand distributionof were determined. This data was evaluated to determine (1) fractures[e.g., Snow, 1968; Brace, 1980]. A knowledgeof the in whether statisticallysignificant fracture orientationsare to be situ fracture distribution is thus of great importance for found in the total fracturepopulation, (2) whetherthe number characterizationof the upper crust. But systematicsubsurface of fractures or the orientation of significantconcentrations observations of fractures have been few. Most such observa- varieswith depth,(3) whetherthe fracturefrequency and orien- tionshave been made in tunnelsand mineswhere the stress tation vary from locationto locationin a givenregion, and (4) field and fracturepattern may have beenconsiderably altered what relation, if any, can be found betweenthe observedfrac- duringoperations. For example,McGarr [1971] and McGarr tures and what is known about the regionalstress field and the et al. [1979] studied the relation of fracture occurrenceand geologichistory. rock burstsin deepmines. Fracture orientations were observed TECHNIQUE to be affectedby the presenceof the tabular excavations,and burst fractures were observed to occur where natural fractures The boreholeteleviewcr (manufactured by SimplecManufac- were unusuallysparse. Overbey and Rough [1971] compared turing Co. under licenseof Mobil Oil Corp.) is a rotating fractureorientations measured at surfaceoutcrops and from acoustictransducer emitting pulsesfocused in a 3ø beam at a aerial photographswith those found in oriented cores and rate of 180 times per second[Zernanek et al., 1970]. The trans- impressionpacker surveysof hydraulicallyfractured intervals ducer rotates at three revolutionsper secondand moves verti- in easternOhio. They found a fair degreeof correlation be- cally in the hole at a speedof 2.5 cm/s.The amplitudeof the tween natural and induced fracturesat depth and major frac- reflectedsignal is plotted as brightnesson a three-axisoscillo- ture trends measuredat the surface.The data discussedby scopeas a functionof the beam azimuth and verticalposition in Overbey and Rough [1971] are significantin that both the the hole. The scopetrace is triggeredat magneticnorth by a flux gate magnetometerin the tool. Essentially,the smoothness of the boreholewall is mapped.Where the smoothnessof the • Now at ChevronOverseas Petroleum Inc., SanFrancisco, Califor- nia 94105. borehole wall is perturbed by a planar featuressuch as a fracture, a dark sinusoidalpattern is seen(see Figure 1). Re- This paperis not subjectto U.S. copyright.Published in 1982by the solution of the tool dependsupon hole diameter, wall con- AmericanGeophysical Union. ditions,reflectivity of the formation,and acousticimpedance of Papernumber 2B(gO3. the well bore fluid. The wall conditionis the most important 5517 5518 SEEBURGERAND ZOBACK: DISTRIBUTIONOF NATURAL FRACTURES

the surface.Only thosefeatures for whichthe sinusoidalsigna- ture could be resolvedwere picked as fractures. -D 2 There are two major limitationsin the analysisof fracture orientation from televiewerdata. First, with the televiewer,only the orientationof a smallportion of a fractureplane is actually observed.On a large scale,many fracturesmay appearplanar, but whenrestricted to viewingonly smallportions of the entire fracture,variations of strike and dip by as much as 10ø may be apparent.Thus, in tryingto determinepreferred fracture orien- hie tation, data from televiewersurveys may result in more scatter and lower levelsof statisticalsignificance for preferredorienta- tions than would be found in surveyswhich considerlarger portionsof the fractures.The secondproblem is that nearly vertical fractures are not often intersectedby vertical wells. Thus, there is a bias in the method as vertical fracturesare not N E S W N sampled. B H T V LOG To evaluate the distribution of fracture orientations,poles to

Strike: Orientation of midpoint fractureplanes for all dippingfractures in eachwell wereplot- between peak and trough ted on a lower hemisphere,equal-area projection. To obtain an (at h/2) estimateof the statisticalsignificance of thesepole groupings (and thus to arrive at an estimateof the preferredfracture

_ Dip: tan• h/d orientations), orientation-density diagrams were prepared usinga methoddescribed by Karnb[1959]. The pole densities Fig. 1. Isometricsketch of fractureor beddingplane intersecting were contoured in intervals of 2a, where a is the standard boreholeat moderatedip angle and correspondingBorehole Tele- viewerlog I-afterZemanek et al., 1970,Figure 7]. deviation of the total number of points in a given area under random sampling. The expecteddensity E for no preferred orientationis 3a. The standarddeviation and the samplingarea usedin preparingthese diagrams are both determinedby a factor as a rough well bore makes detection of fine features statisticalrelation basedupon the numberof polesplotted. The quite difficult. Except for highly fractured intervals,the con- numberof polesN, the samplearea A, givenas a fractureof the ditions in the Mojave and Monticello wells were nearly ideal, total area of the hemisphere,the expecteddensity E, and the and all fractureswere aperturesof more than a few millimeters standarddeviation a are givenfor eachplot. Observeddensities were probably detected.The conditionsin the Limekiln Valley that differ from E(3a) by more than 2 or 3 timesthe standard wells were highly variable. In heavily fractured,poor picture- deviation(i.e., >6a) are likely to be significant,particularly if quality intervals,only a subsetof the total fracturepopulation the higherdensities are clusteredin one sectionof the diagram. could be analyzed. All of the wells were drilled as undeviated, An exampleof the methodis presentedin Figure 3; the number vertical holes. Deviation surveysrun in the later wells con- of polesN is 165,the samplearea A is 0.05,the expecteddensity firmed that the wellswere essentially vertical. E is 8.6, and the standarddeviation a is 2.9. As shownin Figure As an example, Figure 2 presentsteleviewer pictures of a 3b, the only statistically significantpole concentrationsin 15-m vertical sectionof a well in the Mojave desert.The inter- Figure 3a are two clusterswith meanstrikes and dipsof about val from about 168 to 172 m in Figure 2 is one in which N20øW, 63øSW, and N52øW, 55øNE. numeroussingle subparallelfractures are found. From about 173 to 175 m there is a relativelyunfractured interval, although DATA AND INTERPRETATION severalsmall fracturesmight be presentnear 174.5 m. From about 175 to 178 m is an interval in which a few distinct, Mojave Desert Wells parallel fracturesare found with largeapparent aperture. From Five wells were drilled in the westernMojave desertas part 178 to 183 m, two more fracturessimilar to thosejust aboveare of an in situ stressmeasurement program along the locked seenas well as severalfiner scalefeatures which also appear to portion of the San Andreasfault in southernCalifornia. The be fractures. siteslie along the part of the fault that last rupturedin the 1857 Knowing the well diameterthe dip of thesefractures may be Fort Tejon earthquake. Throughout the western Mojave calculatedby measuringthe peak to trough amplitude of the desert,Tertiary formationsrest unconformablyupon a surface sinusoids. The fracture strike is taken to be in the direction of of pre-Tertiary crystallinerocks that underwentdeep erosion the midpoint betweenpeak and trough (Figure 1). The test duringlate Cretaceousand early Tertiary time [Dibblee,1967]. wells were drilled with a diameter of approximately15 cm so In middleto late Miocenetime the Mojave blockwas deformed that the circumference(horizontal scale)is about 50 cm. Thus, primarily by normal faulting along northwest trending faults there is greater than 3:1 horizontal exaggerationin the pic- [Dibblee, 1967]. Geologic and paleomagneticstudies indicate turesand as a result,fractures with dips of lessthan 5ø appear that as much as 10% north-south shortening of the wedge to be horizontal. Apparently low-dipping fractures (such as between the Garlock and San Andreas faults may have oc- those in the first few meters of Figure 2) actually dip at least curredduring the Plioceneand Quaternary [Ponti and Burke, 10ø-15 ø' 1979]. This shorteninghas been accommodatedby strike slip Televiewer surveyswere run in each well from total depth faulting on northwest striking faults, thrust faulting on east- (TD) to the top of the water column or the bottom of casing,in west trending faults, folding, and possibleblock rotation. The either case,usually to within severalmeters or tens of metersof nature of the deformationis consistentwith an applied north- SEEBURGER AND ZOBACK' DISTRIBUTION OF NATURAL FRACTURES 5519

168 m

176

169 .• •-• - ..-•-.•-• •.-•. .•--.•. --•. ;.

177

--.m. • • - • ...... 170 ß• • :-.•.:..• •:..--• .. • .•-:-• . - .• - •...... :•..•.• .... '•.•?•:••' $, ::..-... . ":-' •?.7 ' •••..-.....,..-.•-'• x:•:' •..•.•-.•.•----m:m•••. '•;•

1'78 • ':"'-"-----:::-" --'k::.:::::::-':" ':-; '- -•?':'•-:•i•'""• '"":•%::*-••.•:•" ...•.•b..-;,•, . -:..•r;..,• -..• ... ½.:-:.. ?',.•:•.-:..'.'"•', '•'.:-.--.."':¾.•L*>..'"'• ;7•:4.:•:r•::•• 7'?•: -:*-'•"-{...'• .... 'i•': .... ,."?-•:•.:;-.

•: .... • .... •• ., • :,-. -... '• •.•.-.:.. • ..";'•.'-:.•.:•:i-'-.• '.•'z'-"½"'.i=••' .-::::.,.•...... -$..•::.,•? •' -,;•.•,..:• • 171

• '::"•':':.--7-' ':!:%.:- -T .'.--"' 2 '•- -'*" .•."•:•'.• '-• 1:79 "' :?'•e•• ..:•:•....:• •:•,z•:•:•:.., ...... • ...... -.,...,:•,•...... :...•.•.r:•..,•...•:•;.::.- ...... ;•:•.. •.•...:-.-•,::.,..?..•-•...... •: ...,.? -'h.•..----'•...... 172

- -:• -•:.•:...•.::.•-- ...•.-- -•:...... < •.z: ..-.•..•- -.-• ....ß .. . "• :'? •-."::'"--:..½...,. ....•$..-'• :'"•.,:.*.. "i..:'•.,:-'•*-..'-:'-"--. •

. :..,...•...... •:-..•:•:...... •.::..: • -. .:..a•.:.•-.....•-•....•.,•.

½..-:...-•.•:•_.•.•:•::::.;-:.--•.....:.'...... }•':•'"'•"'•'•'•:•••':••••"• '"'*/'•)•'"'•:"'•?:•::z• "*'...... '::•;•:;.'•}u...• '-•:.-:--•.:?.•... •::r' ß :11.73-.. :s•$:..;:z•.•,.•:::•a•.;•"?•'•-.;:;•--•:. :.::*•.•.•:•::•.----:--::•..:.• •::•:

'"'"::'•-•....•;;- '•:• :;x---•.'.''"4...... ; .... '1'-81.....

•.•.:,.•-r':"...... , • -:. •-:-:•..-...... , ...-...w'"• :::,•:.•.•;.,:•::'•:.• -•..:•'v::.::.•'- ...... :•-----•/•"::-'"-•:--.•:•cc•;:•:...:::"•"•-%•./:•:•.•'-t• c.-.:.----:-•:•--:L .• q:-.:.:•:•:.•:.:.z... q:•::...'--:•'•::.c'•:::'•'•:':"•' .::::...-"'. :-:/ 2•" '---.;:'::•:::•xx•::.-.-•.-..--"::•*.•::'•.•:.:'•'-• •...... •--'-'•:.a:::•"r' .:;?'•;-•e•/.,• :•.•::-'•:.....--•%•:4•.-•.•'--,: 1.:..7.4.. .:':•.•.•.." '- .--'--:.-;•.--:.•::-: '•i-:•:-.':e:•..• •"::'•*::•:•-•.-½: ...... ' ...... -::'•-•..:•'•' •:.-::-•::..:--...... :.'-'.•:': .....'-• .-:'?-.":•::y-:•::-*-7•'.e:,:•...... -* ...... •:-:-".'•:*.':•. -'-"*...... '•'•:--:•::-¾:,.' ...... '% • ',• '--.7'':. "• ....-...:--•..•...•,•z/•..• r-• - ••••-•-.•••-••y:...-.•. ,7 • *•. .:.•:• ...... ?,...•::..:•.•••.•.• .....• --•,••.•••••.••.-••:•:... •...... ::. :':':'•i''i .?.:%c'•.:•'..' • '%"'.X:* ...... -•'-?'"?:':'•%•?'•'• •-•*':.'•..,..?'"':".'::.--,'•: •'.•'•;• :'"•:•:•;;.'"".• :.-•::}.•::•.,'--.•..'-. --.....-;•e ..... •:'":::•'-•••:: ":*...... :':':' • -•'•'" :1'82:- -':'--:*'"':.:' ': •':'•.•::-:•,.•.'...... ' . -.:•:?--• - '"-'.--•:"•. 7..;.•" •D .-•-.....:':'-;..:•* "--' -.:'%•" -. ' ."-- . .. •. :• .• "•:. ....-•-.. -•Z•'-•" ..:'?. • ..;, ...... :-::-..s . • .....:,.-•':-' ...... ?•-.-...... '• -•- -.' ..'..... :•,--T•.• .-v •.-:. .... ":...... :?. -- ,- ..•..4!.',•-. - ,,:-':..•'-' '.. --- "- -. ::'.:....•- * * ? ....•'•?•:½•::;:- ...... •..••:•...... •:.., ...::.... -.,•;• : -:•;:•-•..%....•:•-...... :,. ::.•...•••,;...... •::-::•c•;.,:;'•;-'•-. • ...... '...... -u ...... :,•-. '-'-&':".•..--:q•-- '":-:'...... ,•:eT-•:::.•. -'4" • -• • ::-:-'...... ::- ...... :'-; .i-•:.:..•;. :...... :.;;-:--""-••T•;! "'•"•"**•::;..•:.:- .•:...... "•'"::*-:•.: ':'-/• ...... '7• .... .,,•e...... -':-...... ' .?•f:;:::.[•.:•:•::•.-•½x$¾?/•"•'"'.;%• • %:::....•.::•::c:•;- ...... '•;¾."• •'.-.? '•-...... •:..;•s-';' ':,?"•'•'.'"'--- 17.=--.51 , ;;*'"::'":...:.:.....::..;...,• ...-.:•:..½:.•;..%..,%* ...... ;.:.:., -

N ...... E' S W" N ...... 1"8'3m ' N S

Fig. 2. Televiewerpictures for 15.2m of well in two 7.6-m sections.Top of sequenceis at upperleft, bottom at lower right. Numerousfractures are presentwhich appears as dark sinusoidalbands. 5520 SEEBURGERAND ZOBACK' DISTRIBUTION OF NATURAL FRACTURES

LKC N N b

E E W--

S E=8.6 o'= 2.9 Fig. 3. (a) Lowerhemisphere equal-area stereographic projection of fracturepoles encountered in well LKC. (b) The correspondingorientation-density diagram. The standarddeviation, a = 2.9, and the samplingarea, A = 0.05, are determinedby the numberof fractures,N = 165,using the methodof Karnb[1959]. A randomdistribution of poleswould yield pole densitiesbetween 2a and 4a. Pole densitiesgreater than 6a representstatistically significant preferred fracture orientations.Two significantpole clustersare shownhere: a north-northweststriking southweõtdipping cluster with maximumcontour of 10a and a northweststriking, northeast dipping duster with maximumcontour of 6a. south compression.Geodetic measurements[Savage and near Valyermo to Adobe Mountain in the westernMojave Prescott,1979; Lisowskiand Savage, 1979; Prescott and Savage, desert(Figure 4). 1976; Thatcher,1976] indicatethat presentdeformation in the Well 1 wasdrilled on the north limb of a largesyncline 2 km western Mojave near Palmdale has a direction of maximum southwest of the San Andreas fault and between the San An- shorteningof N10ø-lSøW and a directionof maximumexten- dreasand Punchbowlfaults (Figure 4). Well 1 is the onlywell in sion of about N75ø-80øE.This deformationfield is basically this studywhich did not penetratecrystalline rock. The data consistent with the in situ stress field orientations determined and analysisare presentedhere for completeness.The a•(isof by hydraulic fracturingin two of the wells consideredin this thesyndine strikes N70øW and plunges to thewest. At thewell study,Mojave 1 and 2 [Zoback et al., 1980]. sitethe dip of the northlimb wasapproximately 450-50 ø to the Four of the wells in the westernMojave desertcomprise a southwest.The well was drilled to a depthof 245 m, entirely north-northeasttrending profile running acrossthe San An- throughthe upper MiocenePunchbowl formation. Here the dreas fault from the foothills of the San Gabriel Mountains Punchbowlformation consistsof massive,cross-bedded, coarse

118'15' 117'15' 35'00' ' ' ow,.osAIR FORCE -,4 ':' ROSAt•MO ND BASE '• •,

LANCASTER MOJAVE •, '•'4 DESERT

PALMDALE

VE;TORVI LLE

.• MEASUREMENTSITES

0 IO 20 KM i i • • , 34015' Fig. 4. Map showinglocation of wellsin the westernMojave desert,California. Well 1 is about 2 km southwestof the SanAndreas fault and was drilled entirely through the Miocence Punchbowl sandstone. Wells 2, 4, 5, andXT[,R are 4, 22, 34, and 4 km, respectively,northeast of the SanAndreas fault. All waredrilled in Mesozoicquartz monzonite which is part of the suiteof Mesozoicintrusives which form the basementof muchof thewestern Mojave. SEEBURGER AND ZOBACK' DISTRIBUTION OF NATURAL FRACTURES 5521

terrestrial sandstonewith large lensesof pebble and cobble NUMBER OF FRACTURES 180 conglomerate[Noble, 1954]. At surfaceoutcrops the formation 00 I I 60 I I I 120 I I is relativelyunfractured, although there are beddingplane frac- turesspaced at 1- to 3-m intervalsand somevertical joints with ,\ locally varying attitudes,spaced at 3- to 7-m intervals [Sbaret al., 1979]. Wells 2, 4, 5, and XTLR were drilled at buttes located 4, 22, 5O 34, and 4 km, respectively,northeast of the San Andreasfault '\ : (Figure4). Thesewells were drilled entirely in quartz monzonite x\ to depthsof 257, 230, 239, and 869 m, respectively.The quartz monzonite buttes are part of the suite of Mesozoic intrusives which form the basementof much of the western Mojave. Horizontal and steeplydipping fracturesand joints were ob- I00 servedin outcrops at thesesites. The buttes are apparently composedof slightly more siliceousmaterial than the sur- roundingbasement, and are thus more resistantto erosion(D. B. Burke,personal communication, 1980). _ The rock penetratedby the Mojave wells was found to be 150 - highlyfractured. Figure 5 containsfrequency plots showing the _ "• xx5 number of fracturesper meter as a function of depth for each _ • i well, and Figure 6 containsplots of the cumulativenumber of fracturesin each well as a function of depth. The number of _ \ 4 fracturesobserved in well 1, drilledin sandstone,is significantly lessthan that observedin the wellsdrilled in quartz monzonite. The numberof fracturesper meter is not uniform throughout eachwell. Fracturesoften occurin relativelydense clusters, as at about 55 m and 115-125 m in Mojave 4 (Figure 5). These fracturedintervals appear in Figure 6 as steeperportions of the cumulative fracture curve. The cumulative fracture curve for :•õ0 - I I : Mojave 2 appearsto flatten somewhatat depthsgreater than Fig. 6. Cumulative number of fracturesplotted as a function of 160 m, implyinga decreasein the fracturedensity. In the other depth for the Mojave wells.The sandstoneencountered in Mojave 1 is wells, the curves indicate that the fracture density may be significantlyless fractured than the granite encounteredin wells 2, 4, decreasingwith depth below 150-200 m, but there is only a and 5. In general,only a slightdecrease in the numberof fractureswith very moderatetrend to lower fracturedensity with depth.Also, increasingdepth is observed. note that in the upper 150 m the number of fractures encounteredin Mojave 2, 4, and 5 are similar. The data show no tendencyfor the fracturedensity to increaseas the San Andreas fault is neared. Polesto all of the dippingfractures are shownin Figure 7a.

MOJAVE 1 MOJAVE 2 MOJAVE 4 MOJAVE 5 Figure 7b containsthe orientation-densityplots derivedfrom FRACTURES PER METER the pole distributions.At least two significantmaxima are 0 2 4 0 2 4 0 2 4 0 2 4 , ! , , .__L• i , I ! I , , observedin the densityplots for each well. For all wells, the __ secondarymaxima are significantat only the 6a level.For wells 2 and4 the primarymaxima also barely exceed the 6a level.

__ __ In Mojave 1 the predominantfracture set (> 10a) strikes

__ __ N20ø-30øW and dips 350-40ø to the southwest.A secondclus- __ __ ter (> 6a) with fracturesstriking slightly east of north is also present.Bedding planes, which would strike about N70øW and dip about45ø-50øSW, were not observed.In Mojave 2 a broad cluster (>6a) of northwesterlystriking, southwestdipping (-,, 50ø) fractures are found.In Mojave 4, fracturesets trending N35ø-55øE and dipping 500-60ø to the northwest and southeastare foundat the 6a level.Most fracturesin Mojave 2 150-__ and 4 appearto be randomlyoriented, with minor significant maxima.In contrast,in Mojave 5 a fractureset strikingabout N80øEand dipping steeply ( -,,60 ø) to the southeastis extremely pronounced(> 16a). A secondset of fractures(>6a) is also 200-

_ observedin Mojave 5 strikingnorthwest and dipping to the northwest.Note that the majorityof observedfractures dipped ---TD from 40ø to 70ø . Very few horizontalfractures were observedin TD- thesewells: 3 in Mojave 1, 2 in Mojave 2, 3 in Mojave 4, and 4 in Mojave 5. Fig. 5. Frequencyplot of the numberof fracturesper meter as a Fracturesoccurring below a depth of 137 m were analyzed functionof depthin the Mojave wells1, 2, 4, and 5. separatelyto test for depth dependencein the fractureorienta- 5522 SEEBURGERANDZOBACK' DISTRIBUTION OFNATURAL FRACTURES SEEBURGERAND ZOBACK: DISTRIBUTIONOF NATURAL FRACTURES 5523 tion. Polesand densityplots for the fracturesfound in these By mid-Triassicmuch of the Piedmonthad beenworn down intervalsare shownin Figure8. A comparisonof Figures7 and to a broad peneplainwhich truncatedthe exposedcomplex 8 showsthat each of the maxima presentbelow 137 m also rock structures.Subsequent re-elevation of the region was ac- appearswhen fractures in the entirewell are considered.Thus, companiedby the formationof northeast-southwesttrending the fracturepatterns in the deeperintervals are basicallysub- structuraltroughs. These troughs are generallyparallel to re- setsof the patternsfound throughout the wells. gional northeast-southwesttrending Appalachian structures. Well XTLR was drilled at the same site as Mojave 2. The Further uplift in Late Triassic or Early Jurassictime was resultsof fracturestudies performed in this well are shownin accompaniedby the emplacementof numerousnorthwest- Figures9 and 10. From Figure9 it canbe seenthat the quartz southeasttrending diabasedikes intruded along preexisting monzonitewas highlyfractured throughout. Note that only a fractures. After another erosional cycle, during which the slight decreaseof fracturingwith depth was observed.The Piedmontwas againreduced to a peneplain,continental uplift greatestdecrease in fracturedensity was found at about 140 m, elevatedthe region now occupiedby the Appalachiansand as in the shallow wells. There is, however, no strong trend adjacentPiedmont areas. Since the Jurassic,the region has toward unfracturedrock with depth. undergoneweathering, erosion, and depositionwithout any One significant(> 14a) clusterof poles was found in well major tectonicdisturbances. XTLR striking N5ø-25øW and dipping about 60øSW.A com- The two wells,Monticello 1 and 2, were drilled to depthsof parisonof Figures7 and 10 showsthat the fractureorientations 1100and 1203m, respectively,as part of a programto studythe found in Mojave 2 and XTLR are nearly the same.With the stressesassociated with induced seismicityat Monticello Re- increaseddepth and number of fracturesin XTLR, the level of servoir.Both wellswere drilled in intrusivegranodiorite bodies significanceof thispole cluster increased substantially. of late Paleozoicage. The drill sitefor Monticello 1 waslocated Deep measurementsof the stressfield in XTLR [Zoback et on the top of a broad ridge west of the centerof the reservoir. a!., 1980] indicate that the minimum horizontal stressis the Monticello 2 was drilled within 1 km downstream of the dam minimumprincipal stress below a depthof about300 m. Below which impoundsthe reservoir.The two wells are about 5 km this depth strike slip faulting would be preferred. To test apart. A northward projectionof severalkilometers places the whetherthe changefrom a thrust-typeenvironment to a strike north-south trending, steeply dipping, dip slip (down to the slip environment.affected the fractureorientations, the distri- east)Wateree Creek fault near and to the west of the location of butionof thosefractures found below 300 m in XTLR was Monticello 2 [Secor,1980]. considered(Figure 11). One significant(> 14a) cluster was Boreholeteleviewer records were obtained for the total depth found. The orientation of this clusteris indistinguishablefrom of each well. The well conditions were excellent,and the record- the preferredfracture orientation found above 300 m, in the ed data quality wasgenerally very good. entirewell, and from Mojave 2. The levelof significancefor this The resultsof the fracturesurveys are presentedin Figures13 fractureorientation is greatestwhen the deepfracture data are and 14. Figure 13 showsthe number of fracturesas a function included.The primary effectseen in the near-surfaceintervals of depth in eachwell. Figure 14 showspoles to fractureplanes seemsto bethe addition of randomly oriented fractures which and orientation-densityplots. servedto decr•sethe significance of the shallow interval clus- The data showthat the stateof natural fracturingin the two ters.Therefore, the changein minimum stressorientation prod- wells is significantlydifferent. The total number of fracturesin ucedno noticeablechange in the observedfracture distribution. Monticello 2 is approximately 3 times that in Monticello 1 (Figure 13). Fractures in Monticello 1 were found to occur Monticello Reservoir, South Carolina mostlyin discreteintervals, such as at 140, !90, 300, and 580 m Two approximately1-km deepwells were drilled near Mon- within relativelyunfractured rock. In contrast,the granodiorite ticello Reservoir,Fairfield County, South Carolina (Figure 12). encounteredin Monticello 2 was highly fractured,particularly Thewell sites are in an areaunderlain. by a complexseries of from the surface to about 275 m and from about 460 to 510 m. almandine-amPhibolitefacies, metamorphic rocks, and granitic Due to the highly fractured nature of the rock, small discrete intrusivesof the Charlotte Belt lithologiesof the Piedmont fracturezones, as found in Monticello 1, are not as apparentin Province [Damesand Moore, 1974]. These consistof interlay- Monticello 2. The lowest fracture density in Monticello 2 is eredand folded gneisses, amphibolite, and schist, all of which found from about 510 to 750 m and is similar to the fracture have beenintruded by plutonsof graniteto granodioritecom- densityfound in Monticello 1 at the samedepth. Below 750 m, position.The followingdescription of the geologichistory of a slight increasein the fracture densityis observedin Monti- the regionhas been synthesized from the work of Overstreetand cello 2. It is interestingto note that the slopeof the cumulative Bell [1965], Fisheret al. [1970], and Damesand Moore [!974]. fracture curve for Monticello 2 in the intervals 275-460 m and Mostof themetamorphic rocks presently exposed in the 750 m-TD are about the same and the segmentsare nearly Piedmont were originally depositedas a thick sequenceof colinear.In the intervalfrom 460 to 750 m, strainenergy release shale,siltstone, volcanic tuff, etc. in the 'Appalachiangeosyn- was apparently concentratedin the denselyfractured zone at cline' between250 and 600 m. y. ago. Episodesof folding, 460-510 m. As a result,the fracturedensity from 510 to 750 m is regionalmetamorphism, and igneousintrusion apparentlyfol- the lowestfound in this well. In neitherwell is there an appreci- lowed each of three long sedimentationintervals. At the end of abledecrease in the numberof observedfractures with depth. thelate Paleozoic, the final depositional period, the previously Twenty-six of the 147 fracturesin Monticello 1 and 65 of the formed metamorphic rocks and the unmetamorphosedsedi- 439 fracturesin Monticello 2 were horizontal.This is by far the ments experiencedregional metamorphism.During this time highestdensity of horizontal fracturesof any of the data sets. the Piedmont was uplifted, accompaniedby faulting, folding, Horizontal fractureswere found throughout each well; how- and the intrusion of discordant plutons. Also during this ever, about half of each total was found in the upper 300 m of period, northwestwarddirected overthrustingoccurred on a the well (Figure 15). In both wells,several horizontal fractures major scalein the southernAppalachians. werefound at depthsgreater than 1 km. In addition,the density 5524 SEEBURGERAND ZOBACK' DISTRIBUTION OFNATURAL FRACTURES

II II II II •J b

E z

II II z<• b

I Ez z (,/3

<[ Lu b

E z z.-• •'- ' _'2

II II z<• b

z (,/3

- SEEBURGERAND ZOBACK.'DISTRIBUTION OF NATURAL FRACTURES 5525

XTLR FRACTURES/METER CUMULATIVE FRACTURES

2 4 6 0 2OO 4OO 6OO I I ! I , i I I I I i

TD TD

lOO Fig. 9. Frequencyplot of numberof fractureplotted as a functionof depthfor well XTLR (left)and cumulativenumber of fracturesplotted as a functionof depthfor XTLR (right). of horizontalfractures is relativelyhigh in Monticello2 in the N50øW, dip 25øW. For Monticello 2, there is one extremely interval400-500 m. The presenceof thesehorizontal fractures densecluster of poles (maximum contour > 12•) with a strike may be due to decreasedconfining stresses as a result of the of about N5øE and dip 70øE. post-Jurassicerosional history of the area. In addition to the major differencesin fractureorientation The orientation-densityplot for Monticello 1 (Figure 14) found within a horizontal distance of only 5 km, there were showstwo smallsignificant maxima (contours > 6•) with ap- major differencesin fracture orientation vertically in each well. proximateorientations of strike N45øE, dip 60øE and strike In Monticello 1 this differenceis shown in Figure 16 in which

N XTLR N

w E W E

S S A = 0.016 E=9.0 0-=3.0 Fig. 10. Lowerhemisphere, equal-area plots of polesto all dippingfractures in wellXTLR (left)and orientation-density diagramfor all dippingfractures (right). The maximum contours (14a) indicate that the preferred fracture orientation strikes north-northwestand dipssteeply to the southwest.This poledistribution compares favorably, particularly in the deeper interval,with that foundin Mojave2, whichwas drilled at the samesite as XTLR. The strikeof thissignificant fracture orientationis parallel to thedirection of themaximum regional compressive stress, possibly implying that these fractures are beingheld openlike tensilefeatures. 5526 SEEBURGERAND ZOBACK: DISTRIBUTION OF NATURAL FRACTURES

XTLR

300 m-TD

E W E

S A = 0.02;5 E=9.0 o'= 3.0

Fig. 11. Lower hemisphere,equal-area stereographic pole diagrams for all dippingfractures below 300 m in XTLR. orientation-densityplots are shown for the intervals surface- 2, but as shown in Figure 16, this fracture set is most pro- 305m, 305-610m, and 610m-TD. In the upperzone (surface- nouncedat depthsgreater than 600 m. However, this orienta- 305 m) one significantcluster with northweststrike and south- tion is similar to that of the Wateree Creek fault, where studied west dip is apparent, a subsetof the cluster found when all several kilometers south of the Monticello 2 site. fractures encountered in the well were considered. In the middle It has been noted that the fracture distribution in the interval zone (305-610 m), a northeaststriking, southeast dipping clus- 305-610 m in Monticello 2 is markedly different than that ter is apparent.This clusteris also apparentwhen all fractures found aboveand below. The majority of the fracturesfound in are considered.In the pole densityplot for the bottom third of this interval werelocated in a highlyfractured interval centered the well (610 m-TD), the distribution is seento be essentially at a depth of about 500 m (Figure 13).Pore pressuremeasure- random.In Monticello 2 (Figure 17) the fracturesin the upper ments made in the well revealeda zone of artesian pressures zone form two significantclusters (> 6a): one strikingapproxi- from above 394 m to somewhereabove 591 m [Zoback and mately north-southand dippingto the east,and one striking Hickman,1982]. The presenceof the artesianzone suggests that eastnortheast and dippinggently to the southeast.In the lower impoundmenthas raised the pore pressurein an aquifer rep- interval (610 m-TD) there is a densecluster ( > 16a) of fractures resented,perhaps, by the distinct fracture set found in the strikingnorth-south and dippingto the east.This north-south interval 305-610 m. striking set is prominent when the fracturesas a whole are considered. The fracture cluster found in the interval 305-610 Limekiln Valley, California m hasa strike which differsby about 60ø from that found in the Fracture studieshave been performedin three wells drilled rest of the hole. This group of fracturesis the west-northwest near the creepingsection of the San Andreasfault in central trending, northeast dipping set responsiblefor the extended California.These wells were drilled to depthsof about 220 m in lobe on the densityplot for the entirewell (Figure 14). Cretaceousquartz monzoniteof the Gabilan Range.The wells Secor [1980] presentsthe resultsof joint studiesmade at were located 2, 4, and 14 km west of the San Andreas fault surfaceoutcrops near MonticelloReservoir. At outcropswithin (Figure 18). a few kilometers of the well sites,fracture distributions similar The Gabilan Range is part of the Coast Rangesprovince of to those found in both Monticello 1 and 2 were found. Secor centralCalifornia. (For a detailedsummary of the geologyof concluded that there was little regional consistencyto the this area,see Page [ 1966] and Compton[ 1966].)The structural orientation of the major joint sets.The marked differencewe and stratigraphichistory of the Coast Rangesprovince has have observedin the two wells seemsto support this con- beenquite complex. In the Tertiary,folding, high-angle reverse clusion. Surface fractures were also studied on the cleared faulting,thrusting, and strike slip faultingin the Coast Ranges bedrock surfaceat the site of the Virgil C. Summer Nuclear have been more or less simultaneous.All of these types of Station [Dames and Moore, 1974]. This site is about 6 km deformationhave affected Plio-Pleistocene deposits in one area southeast of the Monticello 1 site and 1 km east of Monticello or another,implying that all havebeen at leastlocally active in 2. Poles to 247 fractureswith no observedshear displacement the Quaternary.The Gabilan Rangeitself is a broad granitic across their faces and to 85 fractures which exhibited either complex enclosingnumerous large metamorphicrelics. Its shear displacementor hydrothermal alteration were plotted lengthis about75 km in a northwest-southeastdirection, with a [Damesand Moore, 1974]. In both cases,a pole densitymaxi- maximum width of about 15 kin. It is boundedon the east by mum occurredfor planesstriking about N44øE, dippingmore the San Andreas fault and on the north, west, and south by than 60øSE. This orientation is very similar to one of the Cenozoicclastic sedimentary rocks. As a whole, the Gabilan maxima found in Monticello 1. However, other maxima in the Range appearsrelatively undeformed,although there is evi- surfacedata, suchas a northeasttrending, northwest dipping denceof small-scalefolding and lineation. set, are not apparent in data from either well. There is no Geodetic measurements of strain across the San Andreas surface indication of the pronouncednorth-south trending, fault [Thatcher, 1979; Savageand Burford, 1973] have been eastwarddipping set of fractureswhich is found in Monticello modeledby the motion of relatively rigid blocks.There has SEEBURGERANDZOBACK: DISTRIBUTION OFNATURAL FRACTURES 5527

..... 4'-.,•'.• ,• MapLocation "I • N

BelairBelt i 1 kilometers Monticello

Monticello 1

Monticello

Reservoir

Monticel Io 2

0 5 i i I I [ kilometers Fig.12. Mapshowing location ofwells near Monticello Reservoir, Fairfield County, South Carolina. Both wells were drilledin intrusive granodiorite bodies oflate Paleozoic age in the Charlotte Belt. beenno evidenceof strainaccumulation on eitherside of the Televiewerpicture quality was generally good in wellsLKB faultsystem, so that the relative motion appears to beaccom- and LKC, but well LKD wasso intenselyfractured that the modatedby slip accumulation in a verynarrow zone centered overall data qualityis relativelypoor. The fracturedensity on the San Andreasfault. Such motion is consistentwith a (numberof fractures per meter) is greater in thesewells than in roughlynorth-south regional compression. However, an analy- anyof theother wells studied (compare Figures 6, 16,and 19). sisof foldtrends [Page, 1966] indicates that thedirection of LKC, the leastfractured of the Limekilnwells is moredensely maximumshortening is N35ø-50øE.Also, earthquake focal fracturedthan Mojave4 and Monticello2, the mostdensely mechanismsindicate general north-south compression but with fracturedwells in the othertwo data sets.Only in LKC is there considerablescatter rZoback and Zoback,1980]. Gawthrop a possibleindication that the number of fractures isdecreasing r1977], in a studyof the seismicity ofthe central coastal region withdepth. There is no relationbetween proximity to theSan of California,found that the leasthorizontal principal stress Andreasfault and the density of fracturingin thesewells. LKD, was orientedabout N60øW, This conclusionwas basedupon the well farthestfrom the fault, encounteredthe most fractures, anaverage of 30events, with approximately equal numbers of whileLKB, the wellclosest to thefault, lies between LKC and strikeslip and thrust events. The direction of maximumcom- LKD in fracture density. pressionvaried from N10øW to N60øE,with an averageof Polesto dippingfracture planes and orientation-density aboutN30øE, close to thedirection inferred from fold orienta- plotsfor eachwell are presented in Figure20. The fractures tions. encounteredin the Limekiln wells seem to belong to a few 5528 SEEBURGERANDZOBACK: DISTRIBUTION OFNATURAL FRACTURES

MONTICELLO 1 MONTICELLO 2 FRACTURES / METER CUMULATIVE FRACTURES I00 200 300 400 I I I I I I

200

4OO 2 E 50-

600

800

iooc

120C Fig.13. Frequencyplots of the number offractures per meter plotted as a functionofdepth for Monticello 1 and 2 plottedas functions ofdepth. The number of fractures encountered in Monticello 2 was about 3 timesthat encountered in Monticello1and there islittle indication inthese wells ofa decrease inthe number offractures with increasing depth. well-developedfracture systems as, in eachcase, the maximum cluster(> 10tr)has an approximatemean strike of N65øEand contouron thedensity plot is at least10it. In wellLKB, located dip of 55øE. There are also two clustersin the orientation- 2 km westof theSan Andreas fau;*, two significant clusters of densityplot for well LKC, which is located4 km from the San polesare observed. The most significant cluster (> l&r) hasa Andreasfault. The most significant cluster (> 10tr)has an ap- meanstrike of about N60øW and dip of 60øW.The secondary proximate mean strikeof N20øW and dip 65øW,while the

MONTICELLO 1 MONTICELLO 2

N N

E W

S S

W E W E

--r-- =0.07 s E s E=.O cr= 2.8 o'= 3.o Fig.14. Lower hemisphere, equal-area diagrams forMonticello 1 and2. (a) Poles toall dipping fractures. (b) Orientation-densitydiagrams forall dipping fractures. Although located only 5 kmapart, the fracture patterns found in thesewells are verydifferent. SEEBURGERAND ZOBACK' DISTRIBUTION OF NATURAL FRACTURES 5529

CUMULATIVE HORIZONTAL FRACTURES are apparent on the orientation-densityplot for well LKD, 20 40 60 ß 0¸ I I I I I located 14 km from the fault. The most significant cluster (> 14tr) has an approximatemean strike of N10øE and dip of 60øW.The two lesserclusters ( > 8tr) have approximateorienta- tions of N5øW, dip 65øE and N40øE, dip 65øE; these two clustersmay not be statisticallydistinct. Orientation-density 2OO plots for the upper 110 m and 110 m-TD for each well are shownin Figure 21. A comparisonof the fractureorientations in the shallow and deep sectionswith those of the entire well indicatesthat the orientationsare consistentthroughout each entire well. 400 FRACTURE ORIENTATION AND REGIONAL STRESSES Theoriesof natural fracturegenesis can be divided into four basic groups:(1) tensilefractures due to applied compressive stresses,e.g., longitudinal splitting [Jaeger and Cook, 1969] 6OO found in planes perpendicularto the direction of least compression[see Engelder and Gelset,1980], (2)shearfractures due MONTICELLO 1 to compressivestresses, (3) tensilefractures which accommo- date extension due to a net tensile stress or to the removal of confiningstresses, e.g., fracturing due to residualstresses [En- 800 gelder, 1979], and (4) natural hydraulic fracturing [Secor, 1965]. Conjugatepairs of shearfractures may develop,with an angle of about 30ø betweenthe direction of maximum com- pressivestress and the fracture planes;however, often only one of the conjugatepairs becomesa well-developedfracture set iooo - [Wilcox et al., 1973]. The measuredshear stresses in thesewells are currently low (lessthan 100 bars) [Zoback et al., 1980; Zoback and Hickman, 1982], even in unfracturedintervals. If such low shear stresses Fig. 15. Cumulative number of horizontal fractures plotted as a are representativeof the stressstate under which thesefracture functionof depth for Monticello 1 and 2. About half of the horizontal fractures in each well were found at depths above 300 m. Several setswere formed, extensivedevelopment of shear setswould horizontal fractureswere also found at depths of about 1 km in both not be expectedand most of the fractureswould be of tensile wells and in the highly fracturedintervals between about 400 and 500 origin. It is interestingto note, however,that in severalcases m in Monticello 2. (e.g., in LKC and LKD, Figure 20) the pole clustersin the orientation-densityplots could be interpretedas indicative of secondarycluster (>6a) has an approximate mean strike of the presenceof conjugate shear fracture sets. Whether these N50øW and dip 55øE.The fracturepopulation picked for LKD clustersare due to the fractureshaving formed in shear,to the is only a subset of the total population due to the intense resultsof shearmotion on favorably oriented,preexisting ten- fracturingand resultantdata quality problems.Three clusters sile fractures, or to processesnot involving shear stresses

M O N T ICE L L 0 I

0 m-505 m 505 m - 610 rn 610 m-TD

N N N

W E W EW E

N=25 A =0.15 ß A = 0.166 S E= 6.6 S E =7.6 S E = 7.5 0-=2.2 0-=2.5 0-=2.5

Fig. 16. Orientation-densitydiagrams for Monticello 1 for the intervals surface-305m, 305-610 m, and 610 m-TD. Different fracturedistributions are foundin eachof theseintervals: shallow dipping, northwest striking fractures in the upper interval; southeastdipping and northeaststriking fracturesin the intermediateinterval; and a random distribution in the lower interval. 5530 SEEBURGERAND ZOBACK: DISTRIBUTION OF NATURAL FRACTURES

MONTICELLO 2_

Om-305m 305 rn-610 rn 610 rn -TD N N N

W > E W EW

86 ?.(r - j -"-'-F--'- A - 0.054 A - 0.083 A - O.07Z S E -8.5 S E - 8.3 ' S E - 8.4 o- = 2.8 o- = 2.8 o- = 2.8

Fig. 17. Orientation-densitydiagrams for Monticello 2 for the intervalssurface-305 m, 305-610 m, and 610 m-TD. In the upper interval two significantpole maxima are found at the 6a level. One of the maxima is part of the north-south strikingset found when the fracturepopulation of the entire well was considered(Figure 14). The secondshallow dipping group showsup as a slight(>4a) concentrationin Figure 14. In the intermediateinterval a strong northweststriking, northeastdipping set is found. Most of the fracturesin this interval are found in the denselyfractured zone at about 500-m depth.In the bottom intervalthe north-southstriking fracture set is veryprominent; there is little evidencebelow 600 m for the northweststriking set of the intermediateinterval. cannot be determined here. From the televiewer data it is of fractureshad developed,with one set parallel to the fault impossibleto determine whether the fracture setsare of shear plane. He also concludedthat the abundanceof microfractures or tensileorigin or whetherany sheardeformation has occurred increaseswith proximity to the fault and is independentof the acrossthe fracturessubsequent to their creation. depth of burial. The Mojave and Limekiln wellsclosest to the The Mojave and Limekiln Valley wellswere locatedclose to San Andreasfault were 2 km away from it. It is possiblethat the San Andreasfault. The stressfield responsiblefor causing closerto the fault the fracture orientationsmay parallel the San displacementon the fault and the deformationassociated with Andreas(Friedman's [1969] data generallycame from within a faulting might be expectedto affect the fracture distributions. half kilometer of the mapped subsurfacelocation of the Oak For example, Friedman [1969] studied the macrofractureand Ridge fault). However,the data presentedhere showthat there microfracture distributions from cores recovered from wells is no tendencyfor the fractureplanes to align themselveswith drilled near the Oak Ridge fault, a major reversefault in the the San Andreas fault nor is there an increase in fracture Ventura basin. Friedman concludedthat conjugateshear sets densityas the fault is approached.

36 ø 52.5' Salt Bautista

CA L A VERA S FA UI T

LKB

Salinas

0 IO MEASUREMENT ki Iometers '•' SITE 36030.0 ' 121ø45.0' 121ø7.5'

Fig. 18. Map showinglocation of wells near Hollister, California. Wells LKB, LKC, and LKD are 2, 4, and 14 km southwestof the SanAndreas fault, respectively.These wells were drilled in Cretaceousquartz monzonite. SEEBURGERAND ZOBACK:DISTRIBUTION OF NATURALFRACTURES 5531

FRACTURES/METER FRACTURE / METER o 2 4 0 2 4 i I

i i i iiii ii

I I00 - E E

150-

LKB LKC 25O 250-

FRACTURES/METER CUMULATIVE FRACTURES 2 4 6 O IO 20 30 i i d !

o150 -

200-' 2OO- •'•LKC LKD

250. LKD 250- Fig.19. (a)-(c)Frequency plot of number offractures per meter plotted as a functionofdepth for wells LKB, LKC, and LKD,respectively. (d)Cumulative number of fractures plotted as a functionof depth.In general,these wells encountered moredensely fractured rock than any of the others in thisstudy. Again, there is only a slighttendency forthe number of fracturestodecrease wtih depth. Also, there is no tendency for the number of fractures toincrease near the San Andreas fault.

At Monticello Reservoir,the current local stressfield has at depthsof lessthan 1 km [Talwaniet al., 1978].The earth- beendetermined by in situstress measurements using the hy- quakesappear to occurin dustersand apparently do not define drofracturetechnique [Zoback and Hickman, 1982] and from linear fault planes.Events in the differentclusters all have earthquakefocal mechanisms [Talwani et al., 1978].In Monti- slightlydifferent composite focal mechanisms.Most earth- cello ! and 2, the differencein magnitudebetween the two quakesare of the thrusttype. A comparisonof focalmecha- horizontalprincipal stresses isrelatively small except at shallow nismsfor earthquakesclose to the Monticello! site(Figure 18) depthwhere the greatest horizontal principal stress is substan- with the shallow fracture orientationsin the well showsfairly tiallygreater than the vertical stress. As discussedby Zoback goodagreement of possiblefault planesand significantgroup- and Hickman,this verticalstress profile is indicativeof con- ingsof naturalfractures [Zoback and Hickman, 1982]. A com- ditionsconducive to thrust-typefaulting in the upper300 m or parisonof poledensities in the upperthird of Monticello2 soif appropriatelyoriented fracture planes are present.Seis- (34-305m) with compositefocal mechanisms for earthquakes mologicstudies confirm that the reservoir-induced earthquakes occurringnear that well site also shows a goodcorrelation with whichhave been recorded in thisarea are apparentlyoccurring oneof thepossible fault planes. Thus, the shallow earthquakes 5532 SEEBURGERAND ZOBACK: DISTRIBUTION OF NATURAL FRACTURES

LKB LKC LKD

E W E W E

S S S

N N N

E W E W E

= 83 A = 0.036 A = 0.052 A = 0.031 s s s E = 8.7 E = 8.6 E = 8.8 o': 2.9 cr: 2.9 cr= 2.9 Fig.20. Lowerhemisphere, equal-area diagrams for LKB,LKC, and LKD. (a) Poles to alldipping fractures. (b) Orientation-densitydiagram for all dipping fractions. In each well, dense pole concentrations werefound. Note that, again, thereis no significant fracture orientation common toeach of the populations. at Monticello Reservoir seem to be associated with shear orientationsof fracturesnear faults may change with timeand motion on preexistingfractures such as thoseencountered in deformation,and Hodgson[1971] found little correlationof' these wells. jointswith fold axes in hisfield mapping of theComb Ridge Asshown in Figures7, 14,and 20, the orientation of signifi- area of the Colorado Plateau. cantfracture clusters varies rapidly from well to wellin a given region.If thefractures in anyarea are the result of regionally SUMMARY AND CONCLUSIONS applied stresses,this difference of fracture orientations would In this paper, observationsof the natural fracture distri- implythat severalfracture mechanisms are involvedand/or butionsfound in 10wells from three different regions of North that extensivepostfracture deformation has occurred.Alter- Americahave been presented. All but one of thesewells were natively, the varied fracture orientationscould be the result of drilled in granitic rock. Numerous fractures were found localvariations or amplificationsof the stressfield. throughouteach well, and only a slightdecrease of fracture Furthercomplications are causedby lithologyvariations. densitywas observed with depth. Thus, from the data presented For example,the differencein fracturedensity observed in here,fractures in crystallinerock could probably be expected at Mojave1, drilledentirely in sandstone,and the other Mojave depthsfar in excessof 1 km. In somewells, fractures seemed to wells,drilled in crystallinerock, may be due in part to the be relativelyuniformly distributed with depth,while in others different thermal historiesof theserock units. theywere primarily concentrated in denselyfractured intervals. Clearly,the major complication in attempting to analyzethe At leastone statistically significant concentration of fracture mechanismof fractureformation in the uppercrust is the planeswas found in everywell. In the Mojaveand Limekiln uncertaintyofthe geologic history of the rock mass. Rocks may Valley wells,these significant fracture concentrations did not undergorepeated tectonic loadings with resultant folding, frac- varymuch with depth.However, in the Monticellowells, more turing,and faulting.Each tectoniccycle may add its own variationof fractureorientation with depth was found. In all of distinctfracture set, or preexistingfracture sets may be modi- thewells, most fractures were steeply digging, and few horizon- fied.The fractures or joints which are observed today reflect the tal fractures were observed. entiredeformational history of the rock.The existenceof frac- The fracturedensity in the wellswas found to varysignifi- ture setswhich cannot be explainedon the basisof the current cantlywithin a givenregion due, apparently, to bothdifferences stressfield is not unexpected.As examples,Tchalenko and in lithology(in the Mojave wells)and/or local structuralor Arnbraseys[1970], Freund [1974], and Wilcoxet al. [1973] stressconcentration effects (in the Monticellowells). The orien- foundfrom field mapping and clay model experiments that the tation of the mostsignificant fracture concentration was also SEEBURGERAND ZOBACK' DISTRIBUTION OF NATURAL FRACTURES 5533

LKB LKC LKD

O- I10 m O- I10 m 0-110 m N N N

w Ew ,•••-EW E

A= 0.052 S E = 8.5 S E = 8.3 S E= 8.6 o'= 2.8 ½=2.8 o'= 2.9 I10 m-TD IIOm-TD IlOm - TD N N N

N -- A = 0.083 A = 0.17 A = 0.072 S E=8.0 S E:7.5 S E= •8.2 cr = 2.7 o' = 2.5 o' = 2.7

, Fig.21. Lowerhemisphere, equal-area orientation-density diagrams forthe upper (top row) and lower (bottom row) sectionsof LKB,LKC, and LKD to showany variation of fractureorientation with depth. Little variation of fracture orientationwith depth is seen in thesewells.

foundto behighly variable within a region.No trendsrelated REFERENCES to suchnearby large-scale structural features as the San An- Brace,W. F., An extensionof Griffiththeory of fractureto rocks,J. dreasfault were observed. As a result,in no areacould there be Geophys.Res., 65, 3477-3480, 1960. Brace,W. F., Permeabilityof crystallineand argillaceous i'ocks' Status predicteda simple, genetic relationship between all ofthe frac- andproblems, lnt. J. Rock Mech. Min. Sci., 17' 241-251, 1980. turesets and a regionalstress field or structure.The observed Budiansky,B.,and R. J. O'Connell, Elastic moduli of cracked soiids, fracturesmay have formed in tensionor shear;the mod e of lnt. J. SolidsStruct., 12, 81-97, 1976. fractureformation cannot be predictedfrom the t½leviewer Compton,R.R., Granitic and metamorphic rocksof theSalinian block,California Coast Ranges, in Geologyof NorthernCalifornia, data.Many factors could be responsible for thecomplex frac- E.H. Bailey, ed., Bull. Calif. Div. Mines Geoli, 190, 277-287, 1966. turepatterns observed in eacharea. Damesand Moore, Supplemental geologic investigation virgil C. In general,little variation of fractureorientation with depth SummerNucl6ar Station--unit one Fairfield County, South Carol- was observedin the wellsstudied. Thus, surfacefracture pat- ina for South Carolina Electricand Gas Company,report, 61 pp., ternsare probablya goodindication of the orientationof Washington,D.C., 1974. Dibblee, T. W., Jr., Areal geologyof the westernMojave Desert, fracturesin theupper kilometer of the crust. However, from the California,U.S. Geol: $urv. Prof. Pap., 522, 153pp., 1967. observationspresented here, it appearsunlikely that the Engelder,T., Thenature of deformationwithin the outer limits of the number,orientatiøn, or typeof fracturesto be encounteredin centralAppalachian foreland fold and thrust belt in NewYork state, situcould have been predicted solely from studies of regional Tectonophysics,55,289-310, 1979. Engelder,T., andP. Geiser,On the useof regionaljoint setsas geologicstructures or the regionalstress field since (1) the trajectoriesofpalcostress fieldsduring the development ofthe APpa - number of fracturesand their orientationsvaried considerably lachianPlateau, New York, J. Geophys.Res., 85, 6319-6341, !980. from site to sitein eachOf the areasstudied and (2) no obvious Fisher,G. W.,F. J. Pettijohn,J. C. Reed,Jr., K. N. Wemer(EdS.), geneticrelationship was observed between the fractures and the StudiesofAppalachian Geology: Central and Southern, 460 pp., Inter- science,New York, 1970. geologicstructures or theregional stress field. Until our know- Freund,R., Kinematicsof transformand transcurrentfaults, Tec- ledgeof the localgeologic history and its effecton fracture tonophysics,21, 93-134, 1974. formationimproves, an understandingof the fracturesystems Friedman,M., Structurealanalysis of fracturesin coresfrom Saticoy commonlyfound in situwill generallyrequire detailed, local, field,Ventura County, California, Am. Assoc.Pet. Geol.Bull., 53, subsurface studies. 367-389, 1969. 5534 SEEBURGERAND ZOBACK.' DISTRIBUTION OF NATURAL FRACTURES

Gawthrop, W. H., Seismicityand tectonicsof the central California plate motion in central California, J. Geophys.Res., 78, 832-845, coastalregion, M.S. thesis,76 pp., Univ. of Colo.,Boulder, 1977, 1973. Hadley, K., Comparison of calculatedand observedcrack densities Savage,J. C., and W. H. Prescott,Geodimeter measurements of strain and seismicvelocities in Westerly granite, J. Geophys.Res., 81, during the southernCalifornia uplift, J. Geophys.Res., 84, 171-177, 3484-3494, 1976. 1979. Hodgson,R. A., Regionalstudy of jointing in Comb Ridge-Navajo Sbar, M. L., T. Engelder,R. Plumb, and S. Marshak, Stresspattern Mountain area, Arizona and Utah, Am. Assoc.Pet. Geol. Bull., 45, near the San Andreas fault, Palmdale, California, from near surface 1-38, 1961. in-situ measurements,J. Geophys.Res., 84, 156-164, 1979. Jaeger,J. C., and N. G. W. Cook, Fundamentalsof RockMechanics, 515 Secor,D. T., Role of fluid pressurein jointing,Am. J. Sci.,263, 633-646, pp., Chapman and Hill, London, 1969. 1965. Kamb, W. B., Ice petrofabricobservations from Blue Glacier, Wash- Secor, D. T., Geological studiesin an area of induced seismicityat ington, in relation to theory and experiment,J. Geophys.Res., 64, Monticello Reservoir,South Carolina, first technicalreport, contract 1891-1910, 1959. 14-08-0001-19124, U.S. Geol. Surv., Reston, Va., 1980. Lisowski,M., and J. C. Savage,Strain accumulation from 1964to 1977 Snow,D. T., Rock fracturespacings openings, and l•orosities,J. Soil near the epicentral zone of the 1976-1977 earthquake swarm Mech. Found. Div. Am. Soc. Civ. Eng., 94(SM1), 73-91, 1968. southeastof Palmdale,California, Bull. Seismol.Soc. Am., 69, 751- Talwani, P., D. Stevenson,J. Sanber,B. K. Rastogi,A. Drew, J. Chiang, 756, 1979. and D. Amick, Seventhtechnical report, contract 14-08,0001-14553, McGarr, A., Violent deformationof rock near deep-level,tabular U.S. Geol. Surv., Reston, Va., 1978. excavations-seismicevents, Bull. Seismol.Soc. Am., 61, 1453-1466, Tchalenko, J. S., and N. N. Ambraseys,Structureal analysis of the 1971. Dasht-e Bayaz (Iran) earthquakefractures, Geol. Soc.Am. Bull., 81, McGarr, A., D. D. Pollard, N. C. Gay, and W. D. Ortlepp, Observa- 41-60, 1970. tions and analysisof structuresin exhumed mine-inducedfaults, Thatcher, W., Episodic strain accumulation in southern California, Proceedingsof ConferenceVIII: Analysisof Actual Fault Zonesin Science,194, 691-695, 1976. Bedrock,U.S. Geol.Surv. Open File Rep.,79-1239, 101-120, 1979. Walsh, J. B., The effect of cracks on the compressibilityof rock, J. Noble, L. F., The SanAndreas fault zonefrom SoledadPass to Cajon Geophys.Res., 70, 381-389, 1965. Pass,California, Bull. Calif. Div. Mines Geol.,170, 37-38, 1954. Wilcox, R. E., T. P. Harding, and D. R. Seeley,Basic wrench tectonics, Nur, A., Effectsof stresson velocityanisotropy in rock with cracks,J. Am. Assoc.Pet. Geol. Bull.,57, 74-96, 1973. Geophys.Res., 76, 2022-2034, 1971. Zemanek, J., E. E. Glenn, L. J. Norton, and R. L. Caldwell, Formation Overbey,W. K., and R. L. Rough,Prediction of oil- and gas-bearing evaluationby inspectionwith the boreholeteleviewer, Geophysics, rock fracturesfrom surfacestructural features, Rep. Invest. U.S. Bur. 35, 254-269, 1970. Mines, 7500, 14 pp., 1971. Zoback, M.D., and S. Hickman, In situ study of the physicalmecha- Overstreet,W. C., and H. Bell III, The crystallinerocks of South nismscontrolling induced seismicity at Monticello Reservior,South Carolina, U.S. Geol.Surv. Bull., 1183, 126 pp., 1965. Carolina, J. Geophys.Res., in press, 1982. Page, B. M., Geologyof the Coast Rangesof California,Geology of Zoback, M.D., H. Tsukahara, and S. Hickman, Stressmeasurements Northern California,edited by E. H. Bailey,Bull. Calif. Div. Mines at depthin the vicinityof the SanAndreas fault: Implicationsfor the Geol., 190, 255-276, 1966. magnitudeof shearstress at depth,J. Geophys.Res., 85, 6157-6173, Ponti, D. J., and D. B. Burke, Ongoing crustal deformationin the 1980. westernMojave region of south-centralCalifornia (abstract),Geol. Zoback, M. L., and M.D. Zoback, State of stressin the coterminous Soc.Am. Abstr. Programs,11,497, 1979. United States,J. Geophys.Res., 85, 6113-6156, !980. Prescott,W. H., and J. C. Savage,Strain accumulationon the San Andreasfault near Palmdale,California, J. Geophys.Res., 81, 4901- (ReceivedMarch 24, 1981; 4908, 1976. revisedNovember 10, 1981; Savage, J. C., and R. O. Burford, Geodetic determination of relative acceptedMarch 12, 1982.)