effect of anisotropic borehole wall failures when estimating in situ stresses: a case study in the...

12
Effect of anisotropic borehole wall failures when estimating in situ stresses: A case study in the Nankai accretionary wedge Hikweon Lee a , Chandong Chang b, * , See Hong Ong c , Insun Song a a Geological Environmental Division, Korea Institute of Geoscience and Mineral Resources, Daejeon 305-350, South Korea b Department of Geology, Chungnam National University, Daejeon 305-764, South Korea c Baker Hughes, 2929 Allen Parkway, Suite 2100, Houston, TX 77019-2118, USA article info Article history: Received 3 May 2013 Received in revised form 5 September 2013 Accepted 8 September 2013 Available online 20 September 2013 Keywords: Borehole stability Breakouts In situ stress Nankai accretionary prism Anisotropic rock strength Weak-plane failure abstract Breakouts observed in a vertical borehole (C0002A) drilled through two major tectonic sedimentary formations consisting of forearc basin (upper) and accretionary prism (lower) sediments in the Nankai accretionary wedge, Japan, exhibit distinctive geometric features in respective formations. Breakouts in the lower accretionary prism sediments are markedly wider than those in the forearc basin sediments, and breakout azimuths in the two units are horizontally rotated relative to one another. Breakout azi- muths are widely used as a proxy for the determination of principal stress directions. However, strength anisotropies related to the presence of bedding planes may affect both breakout azimuths and widths, which can result in misleading in situ stress interpretations. While thinly bedded mudstones are the dominant lithology in both the forearc basin and accretionary prism sediments, bedding planes in the accretionary prism sediments are relatively steeper than those in the forearc basin sediments, with possible implications for breakout geometry and interpretations of principal stress directions. To investigate the effects of bedding planes on breakout geometry (azimuth and width), we conducted borehole wall failure analyses using a weak-plane failure model that incorporates material strength anisotropies. The model results show that breakout widths and orientations are strongly affected by steeply dipping (>40 ) bedding planes in cases where dip directions are unaligned with the principal stress orientation. Our theoretical results suggest that variations in breakout azimuths in the C0002A site may be associated with bedding plane related strength anisotropy, and not associated with the rotation of stress eld. That is, stress orientation is consistent throughout the borehole (down to the bottom-hole depth of 1495 m below sea oor). In addition, disregarding the effects of bedding planes tends to yield an overestimation of in situ stress magnitude. Ó 2013 Elsevier Ltd. All rights reserved. 1. Introduction When drilling a hole into the earth crust subjected to in situ stresses, rock failures occur around the borehole if the local stresses induced there exceed rock strength (Haimson and Herrick, 1985; Zoback et al., 1985; Zheng et al., 1989). Two representative types of drilling-induced borehole wall failures observed in the eld are breakouts (zones of compressive rock failure) and drilling- induced tensile fractures. In a vertical hole drilled into an isotropic and homogeneous rock formation, breakouts, which occur at dia- metrically opposite zones around the borehole, tend to be aligned with the direction of the minimum horizontal principal stress (s h ), whereas the tensile fractures, which also occur at diametrically opposite sides of the borehole wall, tend to be aligned with the direction of the maximum horizontal principal stress (s H ). Thus, these two types of drilling-induced failures have been widely and reliably used as indicators of the principal stress orientations in rocks (Bell and Gough,1979; Hickman et al., 1985; Plumb and Cox, 1987; Shamir and Zoback, 1992; Brudy and Zoback, 1999). In addition to the determination of principal stress axis orien- tations, efforts to estimate in situ stress magnitudes have also been attempted using both theoretical (Zoback et al., 1985; Zheng et al., 1989) and experimental (Haimson and Herrick, 1986; Haimson and Song, 1993) approaches. These approaches, as well as other related research (Herrick and Haimson, 1994; Haimson and Lee, 2004), have consistently demonstrated that breakout dimensions (spe- cically, breakout widths at the borehole wall) depend on far-eld stress magnitudes. As a result, breakout widths, which can often be measured using borehole imaging tools, have been used to estimate s H magnitudes (Moos and Zoback, 1990; Morin et al., 1990; Vernik * Corresponding author. Tel.: þ82 42 821 6430; fax: þ82 42 822 7661. E-mail address: [email protected] (C. Chang). Contents lists available at ScienceDirect Marine and Petroleum Geology journal homepage: www.elsevier.com/locate/marpetgeo 0264-8172/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.marpetgeo.2013.09.004 Marine and Petroleum Geology 48 (2013) 411e422

Upload: insun

Post on 31-Dec-2016

215 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: Effect of anisotropic borehole wall failures when estimating in situ stresses: A case study in the Nankai accretionary wedge

lable at ScienceDirect

Marine and Petroleum Geology 48 (2013) 411e422

Contents lists avai

Marine and Petroleum Geology

journal homepage: www.elsevier .com/locate/marpetgeo

Effect of anisotropic borehole wall failures when estimating in situstresses: A case study in the Nankai accretionary wedge

Hikweon Lee a, Chandong Chang b, *, See Hong Ong c, Insun Song a

a Geological Environmental Division, Korea Institute of Geoscience and Mineral Resources, Daejeon 305-350, South Koreab Department of Geology, Chungnam National University, Daejeon 305-764, South Koreac Baker Hughes, 2929 Allen Parkway, Suite 2100, Houston, TX 77019-2118, USA

a r t i c l e i n f o

Article history:Received 3 May 2013Received in revised form5 September 2013Accepted 8 September 2013Available online 20 September 2013

Keywords:Borehole stabilityBreakoutsIn situ stressNankai accretionary prismAnisotropic rock strengthWeak-plane failure

* Corresponding author. Tel.: þ82 42 821 6430; faxE-mail address: [email protected] (C. Chang).

0264-8172/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.marpetgeo.2013.09.004

a b s t r a c t

Breakouts observed in a vertical borehole (C0002A) drilled through two major tectonic sedimentaryformations consisting of forearc basin (upper) and accretionary prism (lower) sediments in the Nankaiaccretionary wedge, Japan, exhibit distinctive geometric features in respective formations. Breakouts inthe lower accretionary prism sediments are markedly wider than those in the forearc basin sediments,and breakout azimuths in the two units are horizontally rotated relative to one another. Breakout azi-muths are widely used as a proxy for the determination of principal stress directions. However, strengthanisotropies related to the presence of bedding planes may affect both breakout azimuths and widths,which can result in misleading in situ stress interpretations. While thinly bedded mudstones are thedominant lithology in both the forearc basin and accretionary prism sediments, bedding planes in theaccretionary prism sediments are relatively steeper than those in the forearc basin sediments, withpossible implications for breakout geometry and interpretations of principal stress directions. Toinvestigate the effects of bedding planes on breakout geometry (azimuth and width), we conductedborehole wall failure analyses using a weak-plane failure model that incorporates material strengthanisotropies. The model results show that breakout widths and orientations are strongly affected bysteeply dipping (>40�) bedding planes in cases where dip directions are unaligned with the principalstress orientation. Our theoretical results suggest that variations in breakout azimuths in the C0002A sitemay be associated with bedding plane related strength anisotropy, and not associated with the rotationof stress field. That is, stress orientation is consistent throughout the borehole (down to the bottom-holedepth of 1495 m below sea floor). In addition, disregarding the effects of bedding planes tends to yield anoverestimation of in situ stress magnitude.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

When drilling a hole into the earth crust subjected to in situstresses, rock failures occur around the borehole if the localstresses induced there exceed rock strength (Haimson and Herrick,1985; Zoback et al., 1985; Zheng et al., 1989). Two representativetypes of drilling-induced borehole wall failures observed in thefield are breakouts (zones of compressive rock failure) and drilling-induced tensile fractures. In a vertical hole drilled into an isotropicand homogeneous rock formation, breakouts, which occur at dia-metrically opposite zones around the borehole, tend to be alignedwith the direction of the minimum horizontal principal stress (sh),whereas the tensile fractures, which also occur at diametrically

: þ82 42 822 7661.

All rights reserved.

opposite sides of the borehole wall, tend to be aligned with thedirection of the maximum horizontal principal stress (sH). Thus,these two types of drilling-induced failures have been widely andreliably used as indicators of the principal stress orientations inrocks (Bell and Gough, 1979; Hickman et al., 1985; Plumb and Cox,1987; Shamir and Zoback, 1992; Brudy and Zoback, 1999).

In addition to the determination of principal stress axis orien-tations, efforts to estimate in situ stress magnitudes have also beenattempted using both theoretical (Zoback et al., 1985; Zheng et al.,1989) and experimental (Haimson and Herrick, 1986; Haimson andSong, 1993) approaches. These approaches, as well as other relatedresearch (Herrick and Haimson, 1994; Haimson and Lee, 2004),have consistently demonstrated that breakout dimensions (spe-cifically, breakout widths at the borehole wall) depend on far-fieldstress magnitudes. As a result, breakout widths, which can often bemeasured using borehole imaging tools, have been used to estimatesH magnitudes (Moos and Zoback, 1990; Morin et al., 1990; Vernik

Page 2: Effect of anisotropic borehole wall failures when estimating in situ stresses: A case study in the Nankai accretionary wedge

Figure 1. Map of the study area. (a) Locations of selected drilling sites of the Integrated Ocean Drilling Project (IODP) NanTroSEIZE Expedition 314 in the Nankai subduction zone offthe coast of southwest Japan (after Tobin et al., 2009). Yellow arrows and labels show the directions and rates of convergence, respectively, between the Philippine Sea plate andJapan, which is on the Eurasian plate (Seno et al., 1993; Miyazaki and Heki, 2001; Zang et al., 2002). The rectangular outline shows the area of the 3D seismic survey. (b) Seismicreflection profile along the 2D survey line labeled “Line 5” in (a), and the locations of drilling sites (after Louis et al., 2012). (For interpretation of the references to colour in this figurelegend, the reader is referred to the web version of this article.)

H. Lee et al. / Marine and Petroleum Geology 48 (2013) 411e422412

and Zoback, 1990; Brudy et al., 1997; Brudy and Zoback, 1999;Haimson and Chang, 2002; Hickman and Zoback, 2004).

The use of borehole breakouts to estimate in situ stress orien-tations and magnitudes is relatively straightforward in isotropicand homogeneous rock materials; however, in mechanicallyanisotropic rock units, which is often caused by rock fabrics such asbeddings and foliation, the resultant breakout patterns can becomplex (Vernik and Zoback, 1990; Mastin et al., 1991; Aoki et al.,1993; Zou et al., 1996; Okland and Cook, 1998). Vernik andZoback (1990) and Mastin et al. (1991) found that breakoutshapes and orientations observed from scientific drillholes (e.g.KTB, Cajon Pass) are strongly altered by the degree of rock anisot-ropy and the relative orientations among the dip direction ofweakness planes, borehole axis, and in situ stress orientation. Theyobserved that breakouts created in anisotropic formations withsteeply dipping weakness planes (�40�) were much wider at theborehole wall thanwere those in formations with sub-horizontal tomoderately dipping foliations (<40�), and also that the orientationsof breakouts in such anisotropic formations do not match withregional stress orientations. Similarly, Aoki et al. (1993) and Zouet al. (1996), in studies of borehole failures in thinly laminated (orbedded) formations, found that shear failures within the rockmatrix and failures along bedding planes both affected breakoutorientation and size.

The effects of weak planes on breakout patterns are generallyattributed to anisotropic failure mechanisms in rocks, the princi-ples of which are well established thanks to many researchers(Chenevert and Gatlin, 1965; Donath, 1964; McLamore and Gary,1967; Ramamurthy et al., 1993; Niando et al., 1997; Ajalloeian andLashkaripour, 2000; Tien et al., 2006). Their works have demon-strated that failure strength varies as a function of the angle (j)between weakness planes and the loading axis of the major prin-cipal stress. Typically, the failure strength is a maximum at eitherj¼ 0� or 90�, and aminimum at jz 30�. The characteristics of rockfailure modes also depend on relative loading directions. Shearfailures within intact rock matrix are dominant in rocks loaded atangles of j ¼ 0� or 90�, whereas slippage failure along weaknessplanes is more pronounced in rocks loaded at j ¼ 20e40�.

We employ the concepts of anisotropic rock strength and failurebehavior to investigate the significance of anisotropic effect on theborehole breakout formation and also on resulting in situ stressestimations based on the altered breakout geometry. We choosethe Nankai accretionary complex, southwest Japan, as a study areabecause of its tectonic significance, as well as the observed featuresof borehole wall failures represented by widening and rotation ofbreakouts (Chang et al., 2010), which might result from the aniso-tropic effect. Here, we introduce a semi-analytical weak-planeborehole instability model that incorporates rock strength

Page 3: Effect of anisotropic borehole wall failures when estimating in situ stresses: A case study in the Nankai accretionary wedge

Figure 2. Borehole resistivity image, borehole breakout geometry (azimuth and width), and bedding orientations in borehole C0002A. (a) Circumferential resistivity image showingbreakouts (marked by diametrically opposing dark zones). Borehole breakouts are continuously detected below w200 mbsf; no drilling-induced tensile fractures are observed inthe hole. The log-based units are marked by roman numerals (Units IeIV). (b) Azimuths of borehole breakouts averaged per 30-m depth interval. (c) Close-up view of the boxed areain (b) showing the gradual rotation of breakout azimuths with depth. (d) Widths of borehole breakouts averaged per 30-m depth interval, showing enlarged breakout widths in theaccretionary prism sediments. Horizontal error bars in (b)e(d) indicate standard deviations of the values of the azimuth and width. (e) Bedding dip angles. Dip angles are <10� inUnits IeII, gradually increase in Unit III, and vary in the range of 30�e60� in Unit IV. (f) Dip directions. (g) Equal area lower hemisphere stereographic projections of poles to beddingplanes in each of the units. (h) Rose diagrams showing strikes (¼ dip direction e 90�) of bedding planes in each of the units. A comparison of the breakout width (d) and bedding dip(e) indicates that breakouts associated with steeply dipping bedding planes are greater than those associated with horizontal or gently dipping bedding planes.

H. Lee et al. / Marine and Petroleum Geology 48 (2013) 411e422 413

anisotropy. Using the model, we conduct a series of borehole wallfailure analyses to model breakout geometry observed in a verticalborehole (C0002A) drilled into inclined bedded formations, so as toinvestigate the significance of the weak-plane effect in determiningthe orientations and magnitudes of in situ stresses.

2. Geological setting and study borehole (C0002A)

In the Nankai subduction zone off the coast of southwestJapan (Fig. 1), the Philippine Sea plate is subducting beneath theEurasian plate at a rate of 4.1e6.5 cm/yr along an azimuth of300�e315� (Seno et al., 1993; Miyazaki and Heki, 2001; Zanget al., 2002). The upper sediment layers in the Shikoku basin,located on the subducting oceanic plate, are being scraped offand accreted to the margin of the overriding plate, forming anaccretionary prism complex. Locally, sediments are accumulatingon the seaward-dipping slope of the younger accretionary prismsediments. A family of large-scale thrust faults is developedwithin the younger accretionary prism sediments at the base ofthe seaward-dipping slope (Moore et al., 2007). One major out-of-sequence thrust, referred to as the megasplay fault,branches from the main decollement and emerges along theinner trench slope (Moore et al., 2007). The Kumano forearcbasin, in which sediments are accumulating on top of olderaccretionary prism sediments, is located landward of the meg-asplay fault. Seismic reflection data suggest that the forearc basinhas developed and extended by continuous uplift of the

megasplay fault (Park et al., 2002; Kimura et al., 2007). Theseismic data also indicate the presence of numerous small-scalenormal faults in the forearc basin, trending parallel to the trench(Gulick et al., 2010).

Large M8-class earthquakes, which might be associated withmegasplay fault slip, have occurred repeatedly in the Nankaisubduction zone. A comprehensive drilling project, the NankaiTrough Seismogenic Zone Experiment (NanTroSEIZE), a part of theIntegrated Ocean Drilling Program (IODP), is currently underwayto explore the physics of fault and earthquake mechanisms in thesubduction zone. A number of offshore boreholes were drilled atselected sites (Fig. 1) during the first expedition of NanTroSEIZE(Expedition 314) (Kinoshita et al., 2008). Most of the boreholeswere drilled in riserless drilling mode and logged with logging-while-drilling (LWD) and measurement-while-drilling (MWD)technologies to obtain a comprehensive suite of boreholegeophysical data and to monitor drilling parameters, includingtorque, annular pressure while drilling (APWD), and other pa-rameters. Of the boreholes, a vertical hole drilled at site C0002(C0002A), located on the seaward margin of the Kumano forearcbasin, is chosen as our study hole (Fig. 1). The water depth at thedrilling site is 1936 m. The hole was drilled through the forearcbasin sediments and into the upper part of the underlyingaccretionary prism sediments. An unconformity, located at adepth of w935 m below the sea floor (mbsf), stratigraphicallyseparates the forearc basin from the older accretionary prismsediments.

Page 4: Effect of anisotropic borehole wall failures when estimating in situ stresses: A case study in the Nankai accretionary wedge

H

Pp

v

h

Pm

Bedding planes

Figure 3. Schematic illustration of a vertical borehole drilled with borehole pressurePm into a thinly bedded rock formation subjected to boundary stresses sv, sH, and sh,and pore pressure Pp.

H. Lee et al. / Marine and Petroleum Geology 48 (2013) 411e422414

3. Borehole wall failures and geological structures observedalong hole C0002A

A resistivity image log was used to investigate borehole wallfailures in C0002A (Fig. 2a). Borehole breakouts appear in theunwrapped wall image as two diametrically opposing dark bands;the zones are recognized by their relatively low resistivities, whichresult from the replacement of conductive drilling fluid (sea water)in the enlarged zone between the tool pad and the damagedborehole wall. The image log shows that breakouts are almostcontinuously developed from a depth of w200 mbsf to the bottomof the borehole; on the other hand, no drilling-induced tensilefractures were identified in the hole. Chang et al. (2010) analyzedbreakouts to determine their average azimuths (Fig. 2b) and widths(Fig. 2d) in the borehole per 30-m depth intervals. As shown inFigure 2b, the breakouts are consistently oriented throughout thehole, with average azimuths of 134��13� and 310 � 10�, and thusprovide a good estimate of the orientation of the regional shdirection.

A more rigorous inspection of the data shows two notablecharacteristics of the breakouts, their orientation and size, both ofwhich are strongly influenced by in situ stress conditions. Averagebreakout azimuths show a slight clockwise rotationwith increasingdepth, from w130� (200e1200 mbsf) to w147� (1200e1400 mbsf)(Fig. 2c). If the borehole were drilled through isotropic rock for-mations, the observed rotation would indicate a correspondingchange in the orientation of the sh direction by w17� withincreasing depth. However, as we will show later, the verticalborehole encountered severely dipping bedding, which might haveinfluenced breakout orientations. Thus, we question whether therotation of breakouts is due to an actual rotation of in situ stressesor to mechanical anisotropic effects arising from the presence ofinclined bedding, or a combination of the two.

Regarding breakout widths observed in the borehole wall,marked increases in breakout width are observed below the un-conformity (w935 mbsf) (Fig. 2d). Average breakout widths in theforearc basin (logging units IeIII) are 40� � 18�, while those in theaccretionary prism (logging unit IV) are substantially wider(85� � 25�). In a mechanically isotropic rock medium, two possiblemechanismsmight cause the observed breakout widening. The firstpossible cause is a reduction in the rock strength of the accretionaryprism unit (Unit IV, Fig. 2) relative to that of the overlying units,such that for given stress conditions, borehole wall failure is moresevere in the weaker accretionary prism unit than in the forearcbasin units, thus leading to wider breakouts. However, geophysicalborehole logging data show unambiguously that Unit IV rocks aremore consolidated than the rocks of overlying units (indicated bythe lower porosities and resistivities, and higher sonic velocities, ofUnit IV rocks; Figure F1 of Expedition 314 Scientists, 2009); thus,the hypothesis of a strength reduction of rocks at depth seemsunlikely, as the higher degree of consolidation in the older accre-tionary prism sediments implies the opposite (i.e., a probablyhigher rock strength). The second possible cause is that, if the rockis isotropic, the difference in magnitudes between two horizontalprincipal stresses in the accretionary prism is larger than in theoverlying units. Larger differential stresses tend to induce a widerfailure zone around the borehole wall, leading to wider breakouts(Haimson and Herrick, 1985; Haimson and Song, 1993; Haimsonand Lee, 2004).

The prescribed descriptions of the characteristics of boreholebreakouts in terms of azimuth rotations and size variations arebased on the assumption that rock formations are mechanicallyisotropic, such that borehole wall failures are unaffected by rockfabric anisotropies. However, the image log from borehole C0002Ashows that anisotropic geologic features, such as natural fractures

and bedding planes, are pervasive throughout the hole (Kinoshitaet al., 2008). Bedding planes are sub-horizontal (<10�) from thesea floor to 860 mbsf (logging Units I and II), but at depths below860 mbsf, the bedding planes begin to dip. Bedding dips in theaccretionary prism sediments (logging Unit IV) vary significantly, inthe range of 30�e70� (Fig. 2e). Bedding orientations in the loggingunits are shown on stereographic projections and rose diagrams inFigure 2g and h, respectively. The poles to the bedding planes inforearc basin sediments (logging Units IeIII) cluster at the center ofthe stereonet, indicating mostly shallow dips, but those in theaccretionary prism sediments (logging Unit IV) cluster in two mainorientations: in upper Unit IV sediments, the prevailing dip of bedsis to the SSE, while in lower Unit IV sediments, the prevailing dip isto the NW (Fig. 2feh).

The data in Figure 2cee show that changes in breakout orien-tation and width occur primarily in layers with steeply inclinedbedding planes. Thus, there is a possibility that anisotropic strengthcharacteristics associatedwith inclined bedding planesmight causevariations in breakout geometries observed in the borehole. Slip-page along weak bedding planes can contribute to the failure be-haviors of rocks at their intersectionwith the borehole wall, as wellas alter the geometry of borehole breakouts, as observed in otherscientific boreholes (Vernik and Zoback,1990; Mastin et al., 1991). Itwas this observation that motivated the present study. In subse-quent sections, we review a theoretical model on anisotropicborehole wall failure modes that might have occurred in the olderaccretionary prism unit, and then apply the concept to the C0002Aborehole to assess the significance of anisotropic effects resultingfrom intrinsic rock fabric on borehole breakouts, and to explain themechanisms of breakout rotation and enlargement. Such anattempt may be applied further to adjust the in situ stress condi-tions at the site.

4. Borehole stability model for anisotropic formations

The strength characteristics of sedimentary rocks are normallyanisotropic on account of depositional features and mineralogicalvariations between different layers. Mudstones (including shales),

Page 5: Effect of anisotropic borehole wall failures when estimating in situ stresses: A case study in the Nankai accretionary wedge

Slip failurealongweak

planes?

Transform in-situ principal stresses from ICS toECS and then to BCS

Calculate stresses ( rr, r , rz, , z, and zz) at a point ( = 0 to 180) along and beyond the borehole wall in CCS

Shear failureacross intact rock matrix?

Calculate principal stresses ( 1, 2 , 3 ) at the point

Calculate normal and shear stress acting on the plane of weakness

Read inputs

Display failure zone around the borehole

Go to next point

Store the point where failure occurs

Go to next point

Store the point where failure occurs

Figure 4. Flow chart showing the computational algorithms for the determination offailure zones (i.e., breakouts) around a vertical borehole drilled into a thinly beddedformation. The required inputs are listed in Table 1. ICS, in situ stress coordinate sys-tem; ECS, earth coordinate system; BCS, borehole coordinate system; and CCS, cylin-drical coordinate system. See Lee et al. (2012) for details of the reference coordinatesystems.

H. Lee et al. / Marine and Petroleum Geology 48 (2013) 411e422 415

which are the dominant lithology in the older accretionary prismsediments (logging Unit IV), exhibit especially strong strengthanisotropy due to the presence of platy clay minerals that arepreferentially aligned parallel to mechanically weak bedding planesurfaces (Chenevert and Gatlin, 1965; Goodman, 1989; Ajalloeianand Lashkaripour, 2000). To investigate the influence of beddingorientation on borehole wall failures, we employ a borehole sta-bility model developed by Lee et al. (2012), which takes into ac-count the anisotropic strength properties of rock materials. A briefintroduction to the model and its computational algorithms are asfollows (details of the model can be found in Lee et al., 2012).

The borehole stability model can incorporate any arbitrary set oforientations of planes of weakness, in situ stresses, and boreholeorientations, thus allowing one to simulate a particular set of fieldconditions. However, we here focus on the problem of a verticalborehole drilled into a thinly bedded formation subject to in situprincipal stresses described by one vertical and two horizontalcomponents, which represents the situation in borehole C0002A(Fig. 3). The computational algorithms developed for the aniso-tropic borehole instability analysis are shown in Figure 4. Therequired input parameters include the orientation of weaknessplanes, the magnitudes and orientations of in situ principalstresses, pore pressure, borehole pressure, and the material prop-erties of the formation(s). All in situ stress tensors in the in situcoordinate system (ICS) are first transferred to the earth coordinatesystem (ECS) and then to the borehole coordinate system (BCS).Then, the local stresses at points around the borehole are calculatedin a cylindrical coordinate system (CCS) using Kirsch’s equation(Fairhurst, 1968; Bradely, 1979). Because the formation contains aset of weakness planes, shear failure can occur both within the rockmatrix and along weakness planes, depending on the relative anglebetween the loading axis and the weakness planes (Donath, 1964;Chenevert and Gatlin, 1965; McLamore and Gary, 1967;Ramamurthy et al., 1993; Niando et al., 1997; Ajalloeian andLashkaripour, 2000; Tien et al., 2006). Hence, we employ twofailure criteria: one for the failure of intact rock matrix and theother for slip failure along bedding planes.

For the examination of shear failure of the rock matrix, stressesat points around the borehole are converted into three principalstresses (s1, s2, and s3), which are then compared with the for-mation strength using the modified Wiebols & Cook criterion(Zhou, 1994; Colmenares and Zoback, 2002):

J1=22 � A� BJ1 � CJ21 ¼ 0 (1)

where J1 ¼ (s1 þ s2 þ s3)/3, J2 ¼ ((s1 � s2)2 þ(s2 � s3)2 þ (s3 � s1)2)/6, A ¼ ðCi=

ffiffiffi3

pÞ � ððCi=3ÞBÞ � ðCi=3Þ2C,

B ¼ffiffiffi3

pðq� 1Þ=qþ 2� ðC=3Þð2Ci þ ðqþ 2Þs3Þ, and C ¼ ð

ffiffiffiffiffiffi27

p=

ð2C1 þ ðq� 1Þs3 � CiÞÞððC1 þ ðq� 1Þs3 � Ci=2C1 þ ð2qþ 1Þs3 � CiÞ �ðq� 1=qþ 2ÞÞ, where C1¼ (1þ0.6mi)Ci, q ¼ ð

ffiffiffiffiffiffiffiffiffiffiffiffiffiffim2i þ 1

qþmiÞ2, Ci is the

uniaxial compressive strength (¼2Siffiffiffiq

p), Si is the cohesion of intact

rock matrix, and mi is the coefficient of internal friction.To examine slippage along bedding plane, the local stress state is

projected onto the plane of weakness so as to calculate the normal(sn) and shear (s) stresses, which are then compared with thecritical shear stress of a weakness plane (sw) defined by theCoulomb shear failure criterion (Jaeger and Cook, 1979):

sw ¼ Sw þ mwsn (2)

where Sw is the intrinsic shear strength (cohesion) along a plane ofweakness and mw is the friction coefficient of a weakness plane. Asshown in Eqs. (1) and (2), the individual strength criteria areparameterized using the material properties Si and mi for intact rockmatrix, and Sw and mw for planes of weakness. These strength pa-rameters can normally be determined from a series of triaxialcompression tests conducted at varying confining pressures onrecovered cores with varying orientations of weakness planes withrespect to the loading direction. However, because of the paucity ofavailable cores and test data, we estimate these parametersempirically, as described in the next section.

A computation loop, which transfers the stress tensors andcompares them against the failure criteria, is performed until all ofthe given points around the borehole have been checked for failure,after which the locations of failure regions around the borehole canbe predicted. It should be noted that the predicted failure zones arepotential areas of initial failure that form instantaneously afterdrilling. Breakouts may be widened and deepened by continualspalling of broken materials related to inelastic deformation andtime-dependent failure processes (Moore et al., 2011). In the pre-sent model, such processes are not considered.

5. Anisotropic borehole wall failure analysis

5.1. Effects of bedding planes on breakout geometry

Borehole stability analyses using the geomechanical modelrequire input data that include the mechanical properties of intactrock matrix (represented by Si and mi) and weakness planes (rep-resented by Sw and mw), the orientations and magnitudes of in situprincipal stresses, pore-pressure, drilling fluid pressure, and theorientation (dip and dip direction) of weakness planes. To obtainthe strength parameters of intact rock matrix, we use a continuousrock strength log for the C0002A borehole, determined fromempirical relationships between rock strength and sonic velocity(Vp), which were calibrated to actual rock strength data (Changet al., 2010). Such an approach is routinely used in the petroleumindustry, as obtaining rock strengths from core measurements at,for example, m-long intervals in km-deep cores, is not economi-cally feasible and virtually impossible. We use a constant mi value of0.4, following Chang et al. (2010).

Page 6: Effect of anisotropic borehole wall failures when estimating in situ stresses: A case study in the Nankai accretionary wedge

Table 1Mechanical rock properties collected from existing literature.

Shales Rock matrix (j ¼ 90�) Weak plane (j ¼ 30�) References

Ci(MPa)

Si(MPa)

mi Cw(MPa)

Sw(MPa)

mw

Tournemire 43.98a 14.92b 0.40a 18.00a 6.10b 0.40a Niandoet al., 1997

Pierre 1 13.79a 4.39b 0.47a 8.62a 3.07b 0.34a Willsonet al., 2007

Trafalgar 20.32b 10.00a 0.91a 10.74b 5.30a 0.76a Aokiet al., 1993

Pedernales 28.97a 8.19b 0.60a 6.70b 2.07a 0.50a Willsonet al., 1999

Permian 131.08a 34.48b 0.69a 64.60a 18.28b 0.60a Chenevert andGatlin, 1965

Cw is the uniaxial compressive strength of the weak plane.a Value given in the literature.b Value calculated using the MohreCoulomb failure criterion.

0

10

20

30

40

0 10 20 30 40

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Sw, ψ=30 = 0.50 Si, ψ=90 (R2=0.96)

S w, ψ

=30

(MPa

)

Si, ψ=90 (MPa)

μw,ψ=30 = 0.843 μi,ψ=90 (R2=0.93)

μ w, ψ

=30

μi, ψ=90

(a)

(b)

Figure 5. Correlations between the failure properties (cohesion, S, and frictional co-efficients, m) of shalely rocks in intact rock matrix (subscript i) and weakness planes(subscript w). (a) Relationship between the cohesion of weakness planes (Sw) and thatof intact rock matrix (Si). (b) Relationship between the frictional coefficient of weakplanes (mw) and the internal friction coefficient (mi) (b). The parameter j is the anglebetween weakness planes and the loading axis of the major principal stress.

H. Lee et al. / Marine and Petroleum Geology 48 (2013) 411e422416

Systematic core measurements, in which bedding planes withvarying dip angles are sheared, are generally required to obtainreliable measures of slip properties along bedding planes. Unfor-tunately, such systematic measurements are not available for thecore samples from site C0002, which were extracted from bore-holes C0002B, C and D during the Expedition 315. The p-wave ve-locity anisotropies determined from core measurements were onthe order of 5%e10% (Expedition 315 Scientists, 2009); however,these data are not directly related to strength anisotropies associ-ated with shear along bedding planes. Instead, we attempt todevelop a method for indirectly estimating strength anisotropy dueto the presence of bedding based on existing experimental data forsedimentary rocks (Chenevert and Gatlin, 1965; Aoki et al., 1993;Niando et al., 1997; Willson et al., 1999; Ajalloeian andLashkaripour, 2000 Willson et al., 2007). Mechanical rockstrength properties collected from the literature are listed inTable 1. Figure 5 compares collected data on the failure propertiesof rock matrix (as determined from laboratory triaxial compressiontests on cores with bedding planes normal to the axial loading di-rection, j ¼ 90�) and those of slip failure properties along beddingplanes (inclined at j ¼ 30�). These collected data are for poorlyconsolidated shalely rocks from oil wells, similar to those found inthe accretionary prism unit in borehole C0002A. These datademonstrate fairly good correlations between failure properties ofrock matrix and the slip properties of bedding planes. Interestingly,the data comparisons are for measures of a different nature (one forintact rocks and one for discontinuities), and it is generally knownthat the strength properties of discontinuities are independent ofthe failure properties of intact rockmatrix; thus, such a comparisonmay appear counterintuitive. Nonetheless, the results in Figure 5suggest that consolidation and diagenetic processes in weak sedi-mentary rocks tend to harden and strengthen both intact matrixand bedding planes concurrently. Based on these comparisons ofthe failure properties of rock matrix and bedding planes, we esti-mate the strength parameters of the bedding planes from those ofthe intact rock matrix.

In addition to strength parameters, the model requires inputs ofthe strike and dip (bw) of weakness planes, the magnitudes andorientations of in situ principal stresses, pore pressures, and bore-hole pressures. The vertical stress (sv) was determined by inte-grating the density log over depth. Pore pressure (Pp) was assumedto be hydrostatic. The annular pressure while drilling (APWD) wasused as a measure of borehole pressure (Pm). A leak-off test (LOT)conducted at 875.5 mbsf in borehole C0002F (located adjacent toC0002A) yielded a minimum horizontal principal stress of 32 MPa(Moore et al., 2013); this value is close to the lower limit of shestimated by Chang et al. (2010) at the same depth in C0002A, and

the difference between the two is w1 MPa. Thus, we simply raisethe lower limit profile of sh by 1MPa so as tomatch the LOT value atthe 875.5-mbsf depth, and use the shifted sh profile as an input tothe model.

Prior to the borehole wall failure analysis in C0002A, we firstexamine the influence of the orientations (dip and dip direction) ofweakness planes on the geometry of the induced breakouts. Theinput parameters for the model are listed in Table 2. To focus on thegeometries of induced breakouts for various cases of bedding ori-entations, we change bedding orientations while maintaining theorientation of sH in EeW. We vary dip angles of bedding planesfrom 0� to 70�, while maintaining their strikes either parallel to orat an angle of 45� from the sH direction. For comparison, we alsomodel the conditions for the case of isotropic strength.

Figure 6a shows breakout geometries for the case of isotropicstrength. As expected, the breakouts develop in a NeS direction,

Page 7: Effect of anisotropic borehole wall failures when estimating in situ stresses: A case study in the Nankai accretionary wedge

Table 2Inputs for a hypothetical borehole stability analysis.

Parameter Value

Depth (m) 2438.4Overburden stress (sv, MPa) 49.65Pore pressure (Pp, MPa) 25.38Minimum horizontal principle stress (sh, MPa) 41.4Maximum horizontal principle stress (sH, MPa) 44.1Azimuth of sH (degrees) N90EWellbore pressure (Pm, MPa) 29.9Cohesion of rock matrix (Si, MPa) 5Friction coefficient of rock matrix (mi, degrees) 0.4Poisson’s ratio (y) 0.35Biot’s parameter (a) 1Cohesion of weak plane (Sw, MPa) 2.5Friction coefficient of weak planes (4w, degrees) 0.337Dip direction of weak plane (aw, degrees) S, S45EDip angle of weak plane (bw, degrees) Varied (0e70)

(a)

σh

N

S

βw = 0o

βw = 10o

βw = 40o

βw = 30o

βw = 60o

βw = 70o

βw = 50o

bedding plane

(c)

bedding dip direction

(b)

σH

Figure 6. Failure regions around a vertical borehole, for (a) materials with isotropicstrength, and (b) and (c) materials with anisotropic strengths, for dip angles of beddingbw of 0�e70� and with dip directions toward the south or S45�E. The analyses wereconducted using the input parameters listed in Table 1. At lower bedding-plane dipangles (bw < 40�), the breakouts formed in materials with anisotropic strengthproperties were identical to those formed in materials of isotropic strength. However,at greater dip angles (bw � 40�), breakouts become wider and deeper. The breakoutazimuth is unchanged when the bedding dip direction is parallel to the sh direction;

H. Lee et al. / Marine and Petroleum Geology 48 (2013) 411e422 417

with a width of w89�. For gently dipping weakness planes(bw � 30�) (see Fig. 6b and c), breakout orientations and widthsremain the same as those for isotropic material strengths, indi-cating that bedding plane has no effect on the geometry of inducedbreakouts. This demonstrates that borehole wall failures observedat depths shallower than 860 mbsf, where the dips of beddingplanes are <10�, are not influenced by the presence of beddingplanes.

When the dip angle is greater than 40�, however, the beddingplane strongly affect breakout widths, both when bedding planestrikes are parallel and at 45� to the principal stress direction. Forbedding planes striking parallel to the sH direction, breakoutwidths tend to grow with increasing dip angles, whereas centralbreakout azimuths remain unchanged. For example, for bw ¼ 50�,the width of induced breakouts is w101�, which is 12� wider thanthat for isotropic materials, and the central azimuth is 180�, whichis exactly alignedwith the sh direction.When bedding strikes are atan angle to the sH direction, breakout widths increase withincreasing dip angles, and their azimuths rotate towards the di-rection of the bedding strike. For example, for bw ¼ 50�, thebreakout width isw135�, which is 46� wider than that for isotropicmaterials, and the central azimuth is w157�, which represents arotation of 23� from the far-field sh direction. At bw ¼ 70�, theenlarged breakout region takes on the two-humped shape of acamel back; this two-lobed breakout geometry is suggestive of a‘square borehole’, as observed in laboratory borehole wall failureexperiments in thinly laminated shale (Okland and Cook, 1998) andin the field (Brehm et al., 2006). Overall, the results shown inFigure 6 indicate that steep bedding plane dips contribute tobreakout widening, and that bedding strike tends to controlbreakout azimuths, especially when the strike is at an angle to thestress direction. Thus, we expect that the geometries of boreholebreakouts (width and azimuth) in the older accretionary prismsediments with steeply dipping bedding planes and strikes that arenot necessarily parallel to in situ stress orientations might beaffected by the presence of bedding plane weaknesses. If that is thecase, then the in situ stresses determined on the basis of theassumption of isotropic strength (i.e., disregarding the effects ofweakness planes) might yield incorrect orientations and magni-tudes of the principal stresses.

however, breakouts rotate towards the sH direction when the bedding dip angle is 45�

from the sh direction.

5.2. Model application to borehole C0002A

To investigate the significance of bedding plane weaknesses onbreakout formation in the accretionary prism sediments, weapplied a borehole wall failure model for the depth interval 860e1340 mbsf, where bedding planes are moderately to steeply

dipping (see Figs. 2 and 7). As observed in Figure 7, the beddingplanes are dipping either to the SE or NW at a wide range of dipangles (10�e60�). We divided the depth interval from 860 to1340 mbsf into five sub-intervals, based on average dip angles and

Page 8: Effect of anisotropic borehole wall failures when estimating in situ stresses: A case study in the Nankai accretionary wedge

800

900

1000

1100

1200

1300

1400

0 30 60 90 0 60 120 180 240 300 360

II

Dep

th (m

bsf)

Dip angle (deg) Dip direction (deg)

Figure 7. Dip angles (a) and dip directions (b) of bedding planes observed in the depthinterval 800e1400 mbsf in borehole C0002A. The log-based units (Units IeIV) andsubsections (sections AeE) are indicated. Bedding dips vary widely, in the range of10�e60�; however, dip directions are constrained to either the southeast or northwest.

σh direction

800

900

1000

1100

1200

1300

1400

90 120 150 180

Dep

th(m

bsf)

Breakout azimuth (deg)

Figure 8. Azimuths of observed (closed diamonds) and modeled (open squares)breakouts at depths below 860 mbsf, showing that the azimuths of simulated break-outs agree well with those of observed breakouts.

H. Lee et al. / Marine and Petroleum Geology 48 (2013) 411e422418

dip directions: section A (dip angle/dip direction ¼ 14�/120�),section B (36�/128�), section C (39�/340�), section D (53�/156�), andsection E (49�/340�).

We first model borehole wall failures assuming a fixed in situ shdirection of N130�E, which is the average azimuth of breakoutsobserved above 860 mbsf. The center azimuths of the modeledbreakouts in the 860e1200 mbsf interval, where bedding dips aremoderate to gentle (<40�), remained nearly constant at N130�E,indicating that gentle bedding plane orientations did not influencethe breakout azimuth (Fig. 8). However, below 1200 mbsf, wheredip angles are greater thanw40�, the center azimuths of breakoutstend to rotate gradually with depth, from approximately N130�E toN150�E. The modeled and observed breakout azimuths are inremarkable agreement, demonstrating that the orientations ofbreakouts can rotate clockwise by several tens of degrees on ac-count of the presence of steeply dipping bedding planes weak-nesses, without an actual change in the in situ stress direction.Although the rotation of breakouts with depth can occur as a resultof several other mechanisms, such as by the influence of majorfaults (Hickman and Zoback, 2004; Camac et al., 2006; Lin et al.,2010) or the local rotation of the stress field itself, our simulationsuggests that slip failure along weak bedding planes is a plausiblemechanism for the observed rotation of breakouts with depth inborehole C0002A, and that the rotation in breakout azimuths is nota result of the rotation of the in situ stress field itself. In other

words, the orientations of horizontal principal stresses may beconsistent throughout borehole C0002A.

Having the constrained information of the direction of in situhorizontal principal stresses, we then examine breakout widths asa sole function of bedding plane orientation. Because breakoutwidth depends on the magnitude of sH, we model breakout ge-ometry as a sH-dependent function. Two examples are presented:one corresponding to the 1051-mbsf depth, where the strike ofbedding is nearly parallel to the in situ sH direction (Fig. 9), and onecorresponding to the 1295-mbsf depth, where the strike of beddingis at an angle of w30� to the sH direction (Fig. 10). Table 3 gives theinput parameters used in the models. To visualize the effects ofbedding plane orientation on the geometry of breakouts, wecompare the shapes of the induced breakouts assuming that rockstrength is either isotropic (Figs. 9a and 10a) or anisotropic (Figs. 9band 10b).

For the case in which bedding planes strike parallel to the sHdirection (Fig. 9), the breakouts become wider and deeper withincreasing sH for both isotropic and anisotropic rock strengths(Fig. 9a, b, and d), while azimuths are maintained at w130�

regardless of the magnitude of sH (Fig. 9c). However, the resultsshow that breakout widths can vary depending on the weak planeeffect (Fig. 9d). For cases in which the sH/sh ratio is low (<1.15),modeled breakout widths in materials with anisotropic strengthsare up to 10� wider than in materials with isotropic strength.However, when the sH/sh ratio is >1.15, breakout widths in mate-rials with isotropic and anisotropic strengths are the same. Thebreakout widths observed at this depth (1051 mbsf) were 80�,which indicates that the sH/sh ratio should be w1.16 for bothisotropic and anisotropic cases.

The overall resemblance in the breakout geometries in forma-tions with both isotropic and anisotropic strengths is attributed to

Page 9: Effect of anisotropic borehole wall failures when estimating in situ stresses: A case study in the Nankai accretionary wedge

σh

(a) (b)

σH

σH/σh = 1.10

σH/σh = 1.20

σH/σh = 1.15

(d)

(c)

120

125

130

135

140

145

150

1.1 1.12 1.14 1.16 1.18 1.2

Anisotropic strength caseIsotropic strength case

σH

/ σh

55

60

65

70

75

80

85

90

1.1 1.12 1.14 1.16 1.18 1.2σ

H/ σ

h

Brea

kout

azi

mut

h (d

eg)

Brea

kout

wid

th (d

eg)

Figure 9. Schematics of borehole cross-sections showing breakouts for cases of materials with isotropic (a) and anisotropic (b) strength, at a depth of 1051 mbsf. The relationshipsbetween the horizontal stress ratios and the breakout azimuth (c) and width (d) are also shown. The breakout azimuths in the two cases are nearly the same.

H. Lee et al. / Marine and Petroleum Geology 48 (2013) 411e422 419

the fact that bedding plane dips are moderate to gentle (<40�) andstrikes are parallel to one of the horizontal principal stress di-rections. Within the failed breakout zones, the orientations of themaximum principal stress induced locally around the borehole areessentially parallel or subparallel to the strike of bedding, and thusslip failure along bedding planes is minimized. The results shown inFigure 9 indicate that moderately dipping beds (sections AeC) havelittle influence on breakout formation; thus, the marked enlarge-ment of breakouts below the unconformity at w935 m is aconsequence of the relatively large difference between the mag-nitudes of the two horizontal principal stresses as compared withtheir magnitudes in the overlying forearc basin sediments.

The effect of bedding plane weaknesses is apparent in areaswhere bedding plane dips are >40�, as is the case at a depth of1294.8 mbsf (section E) (Fig. 10). At this depth, the bedding planesdip 49� toward N20�W (340�). The difference between the beddingplane strike and the sH direction is w30�. The shapes, dimensions,and locations of the modeled breakouts in formations with aniso-tropic strengths are quite different than those in formations withisotropic strength. Breakout shapes in formations with anisotropicstrengths are asymmetric, as slip failure along bedding planes oc-curs more preferably in one side where bedding planes are morepreferentially oriented for slip with respect to local stress

directions. The dimensions of breakouts are noticeably larger whenthe effects of bedding plane dips are considered. Given a constanthorizontal stress ratio, breakouts in formations of anisotropicstrengths are wider (by approximately 15�e20�) and penetratedeeper into the rock than are those in materials of isotropicstrength. At 1294.8 mbsf, the breakout width is 88�. Assuminganisotropic material strengths, the resulting sH/sh ratio is 1.18,which is somewhat lower than sH/sh ¼ 1.28, the value which isderived if anisotropic strength effects are ignored. In other words,the stress states represented by horizontal stress ratios are over-estimated if the effects of slip failure are ignored (i.e., if the strengthproperties of borehole wall rocks are assumed to be isotropic). Inaddition, the center azimuth of the breakouts rotates clockwise (by7�e13�) relative to the sh direction when slip failure along beddingis considered (as we already demonstrated in Fig. 8).

Following the approach illustrated in the two examples above,we calculate the stress magnitudes in C0002A as a function ofdepth, taking into account bedding plane orientations. We performmultiple simulations for a given depth, by applying a range ofhorizontal stress ratios to determine the stress ratio that results in abreakout width observed in the borehole image. Again, we use theshifted lower bound of the range of sh provided by Chang et al.(2010) to calculate sH. Figure 11 summarizes the sH profile

Page 10: Effect of anisotropic borehole wall failures when estimating in situ stresses: A case study in the Nankai accretionary wedge

σH/σh = 1.15

σH/σh = 1.17

σH/σh = 1.20

(a) (b)

σhσH

(d)

Anisotropic strength caseIsotropic strength case

(c)

120

125

130

135

140

145

150

1.1 1.15 1.2 1.25 1.3σ

H / σ

h

50

60

70

80

90

100

1.1 1.15 1.2 1.25 1.3σ

H / σ

h

Brea

kout

azi

mut

h (d

eg)

Brea

kout

wid

th (d

eg)

Figure 10. Schematics of borehole cross-sections showing failure zones for cases of materials with isotropic (a) and anisotropic (b) strength, at a depth of 1294.8 mbsf. The re-lationships between the horizontal stress ratios and the breakout azimuth (c) and width (d) are also shown. The breakout geometries (particularly widths) and azimuths of the twocases are markedly different; the breakouts in cases of anisotropic strength are wider and rotated with respect to the sh direction as compared with the breakouts in cases ofisotropic strength.

H. Lee et al. / Marine and Petroleum Geology 48 (2013) 411e422420

determined from the anisotropic borehole stability model, as wellas model inputs of other stress values. For comparison, we alsoshow sH values obtained using the isotropic stability model. Asbedding planes in the forearc basin sediments are gently inclined

Table 3Inputs for a borehole failure analysis at depths of 1051.0 and 1294.8 mbsf.

Depth (mbsf) 1051.0 1294.8

Water depth (m) 1936 1936Overburden stress (sv, MPa) 38.3 42.9Pore pressure (Pp, MPa) 30.0 32.4Minimum horizontal stress (sh, MPa) 33.2 38.3Maximum horizontal stress (sH, MPa) Varied VariedAzimuth of sh (degrees) 130 130Wellbore pressure (Pm, MPa) 30.8 34.2Cohesion of rock matrix (Si, MPa) 2.63 3.98Friction coefficient of rock matrix (mi, degrees) 0.4 0.4Poisson’s ratio (y) 0.4 0.4Biot’s parameter (a) 1.0 1.0Cohesion of weak plane (Sw, MPa) 1.31 1.99Friction coefficient of weak planes (mw, degrees) 0.336 0.336Dip direction of bedding plane (aw, degrees) 128 340Dip angle of bedding plane (bw, degrees) 36 49

(<10�), the calculated magnitudes of sH should be independent ofwhether material strengths are isotropic or anisotropic; this iscorroborated by sH values obtained using both isotropic andanisotropic strength assumptions at depths just above the uncon-formity, as well as in sections B and C where bedding dips are<40�.Bedding dips in sections B and C have little effect on breakout di-mensions, and consequently on estimated sH magnitudes. How-ever, differences between the isotropic and anisotropic strengthmodels on the magnitudes of sH are notable in sections D and E,where the average dips of bedding planes are 53� and 49�,respectively, and where the predicted sH magnitudes estimatedbased on the anisotropic strength model are 4e6 MPa less thanthose obtained using the isotropic strength model.

The sH profile obtained by incorporating bedding plane weak-nesses into the model shows that the relative magnitudes of thethree in situ principal stresses are primarily in the order ofsh < sv < sH, indicating a stress regime that favors strike-slipfaulting. The sH magnitudes tend to increase rapidly with depthin the depth interval between the unconformity (w935 m) andw1100 mbsf, which changes the stress regime from a normalfaulting to a strike-slip faulting regime. The relatively large differ-ential stress (sH e sh) is responsible for the enlarged breakouts

Page 11: Effect of anisotropic borehole wall failures when estimating in situ stresses: A case study in the Nankai accretionary wedge

200

400

600

800

1000

1200

140020 25 30 35 40 45 50 55

Dep

th(m

bsf)

Stress (MPa)

Pp

σv

Pm

LOT at 875.5 mbsfin C0002F

from Exp. 338

σh σv

Unconformity

σH

Figure 11. Constrained in situ sH profiles derived using the anisotropic (closed circles)and isotropic (open triangles) borehole wall failure models. The stress regime in theforearc basin sediments predominantly favors normal faulting. By contrast, the state ofstress in the accretionary prism sediments favors normal faulting in the uppermostprism, transitioning to strike-slip faulting in the lower prism.

H. Lee et al. / Marine and Petroleum Geology 48 (2013) 411e422 421

observed in this depth interval. Below w1100 mbsf, the sHmagnitude profile is variable; in this region, sH magnitudes are 0e4 MPa greater than sv magnitudes.

6. Conclusions

Borehole breakout geometries are often utilized as direct in-dicators of the in situ stress orientations and magnitudes in bore-holes. However, the present results show that weakness planes,such as bedding, can affect the geometry of borehole breakouts. Ourstudy shows that breakout width can be significantly affectedwhenthe dips of weakness planes are greater than w40�, and thatbreakout azimuths can rotate if bedding strikes are at an obliqueangle to in situ stress orientations, againwhen the dips of weaknessplanes are greater than w40�. Such alterations in breakout geom-etry can lead to erroneous interpretations of in situ stresses, if theanisotropic strength effects due to slip failure along weaknessplanes are not taken into account.

The application of the weak plane failure model to boreholeC0002A shows that steeply inclined bedding is a possible cause forobserved breakout rotations with depth, suggesting that in situ shorientations in the borehole are consistently oriented at N130�Ethroughout the hole. In addition, at depths where bedding planes

are steeply inclined (>40�), the magnitudes of sH are lower by 4e6 MPa (approximately 10% of the estimated stress) than those thatwould be estimated if anisotropic strength had not been consid-ered, thus suggesting the importance of anisotropic effects on themagnitudes and orientations of the principal stresses.

The anisotropic effects of rock materials on stress estimationsmay not appear to be substantial. Based on this study, however, weinfer that anisotropic strength effects can be substantive in certainsituations, such as the case of large differences between rockmatrixstrengths and shear strengths of weakness planes. In such cases,slip failure would dominate rock matrix failure under a given stresscondition. As sedimentary rocks become consolidated, the strengthparameters of both matrix and weakness planes evolve, as wasshown above on the basis of available data (Fig. 5). However, whenthe strength parameter of the rock matrix is substantially greaterthan that of weakness planes, the weakness planes will be moresusceptible to slip. Another possible situation in which anisotropiceffects may be important is that of strongly dipping bedding planes.As shown in Figure 6, the geometry of breakouts is strongly affectedby highly dipping bedding planes. These two possible situations arelikely to be encountered in the deeper portions of accretionaryprism sediments in which rocks have undergone a high degree ofconsolidation, as well as severe folding and thrusting. In such cir-cumstances, in situ stress estimations based on breakout analysesmay require careful inspection regarding the effects of anisotropicmaterial strength.

Acknowledgments

Borehole geophysical logging data and resistivity image data inborehole C0002A were provided by the Integrated Ocean DrillingProgram (IODP). This research was a part of the IODP funded by theMinistry of Land, Transport and Maritime Affairs, Korea. Thisresearch was also partially supported by the Basic Research Projectof the Korea Institute of Geoscience and Mineral Resources(KIGAM). We appreciate comments and suggestions by Weiren Linand an anonymous reviewer.

References

Ajalloeian, R., Lashkaripour, G.R., 2000. Strength anisotropies in mud rocks. Bull.Eng. Geol. Env. 59, 195e199.

Aoki, T., Tan, C.P., Bamford, W.E., 1993. Effects of deformation and strength anisot-ropy on borehole failure in saturated shales. Int. J. Rock Mech. Min. Sci. Geo-mech. Abstr. 30 (7), 1035e1038.

Bell, J.S., Gough, D.I., 1979. Northeast southwest compressive stress in Alberta:evidence from oil wells. Earth Planet. Sci. Lett. 45, 475e482.

Bradely, W.B., 1979. Failure of inclined boreholes. J. Energy Resour. Technol. Trans.ASME 101, 232e239.

Brehm, A., Ward, C., Bradford, D., Riddle, G., 2006. Optimizing a deepwater subsaltdrilling program by evaluating anisotropic rock strength effects on wellborestability and near-wellbore stress on the fracture gradient. In: IADC/SPE 98227.IADC/SPE Drilling Conf. Miami, FL.

Brudy, M., Zoback, M.D., Fuchs, K., Rummel, F., Baumgartner, J., 1997. Estimation ofthe complete stress tensor to 8 km depth in the KTB scientific drill holes: im-plications for crustal strength. J. Geophys. Res. 102, 18453e18475.

Brudy, M., Zoback, M.D., 1999. Drilling-induced tensile wall-fractures: implicationsfor determination of in situ stress orientation and magnitude. Int. J. Rock Mech.Min. Sci. 36, 191e215.

Camac, B.A., Hunt, S.P., Boult, P.J., 2006. Local rotations in borehole breakouts-Observed and modeled stress field rotations and their implications for thepetroleum industry. Int. J. Geomech. 6 (6), 399e410.

Chang, C.,McNeill, L.C.,Moore, J.C., Lin,W.,Conin,M., Yamada,Y., 2010. In situstress statein the Nankai accretionarywedge estimated fromboreholewall failures. Geochem.Geophys. Geosyst. 11, Q0AD04. http://dx.doi.org/10.1029/2010GC003261.

Chenevert, M.E., Gatlin, C., 1965. Mechanical anisotropies of laminated sedimentaryrocks. In: SPE 890, 39th Ann. Fall Meeting of SPE. Houston, TX, pp. 67e77.

Colmenares, L.B., Zoback, M.D., 2002. A statistical evaluation of intact rock failurecriteria constrained by polyaxial test data for five different rocks. Int. J. RockMech. Min. Sci. 39, 695e729.

Donath, F.A., 1964. Strength Variation and Deformational Behavior of AnisotropicRocks. State of Stress in the Earth’s Crust. Elsevier, New York, N. Y., pp. 281e298

Page 12: Effect of anisotropic borehole wall failures when estimating in situ stresses: A case study in the Nankai accretionary wedge

H. Lee et al. / Marine and Petroleum Geology 48 (2013) 411e422422

Expedition 314 Scientists, 2009. Expedition 314 site C0002. In: Kinoshita, M.,Tobin, H., Ashi, J., Kimura, G., Lallemant, S., Screaton, E.J., Curewitz, D.,Masago, H., Moe, K.T., The Expedition 314/315/316 Scientists (Eds.), Proc. IODP,314/315/316. Integrated Ocean Drilling Program Management International,Inc., Washington, DC http://dx.doi.org/10.2204/iodp.proc.314315316.114.2009.

Expedition 315 Scientists, 2009. Expedition 315 site C0002. In: Kinoshita, M.,Tobin, H., Ashi, J., Kimura, G., Lallemant, S., Screaton, E.J., Curewitz, D.,Masago, H., Moe, K.T., the Expedition 314/315/316 Scientists (Eds.), Proc. IODP,314/315/316. Integrated Ocean Drilling Program Management International,Inc., Washington, DC http://dx.doi.org/10.2204/iodp.proc.314315316.124.2009.

Fairhurst, C., 1968. Methods of Determining in Situ Rock Stresses at Great Depth(Tech. Rep. I-86). Missouri River Div., U.S. Army Corps of Engineers.

Goodman, R.E., 1989. Introduction to Rock Mechanics, second ed. John Wiley andSons, p. 562.

Gulick, S.P.S., Bangs, N.L., Moore, G.F., Ashi, J., Martin, K.M., Sawyer, D.S., Tobin, H.J.,Kuramoto, S., Taira, A., 2010. Rapid forearc basin uplift and megasplay faultdevelopment from 3D seismic images of Nankai margin off Kii Peninsula, Japan.Earth Planet. Sci. Lett. 300, 55e62. http://dx.doi.org/10.1016/j.epsl.2010.09.034.

Haimson, B.C., Herrick, C.G., 1985. In situ stress evaluation from borehole breakoutse experimental studies. In: Research and Engineering Application in Rock Mass,Proc. 26th U.S. Symp. on Rock Mech., Rapid City. A. A. Balkema, Rotterdam,pp. 1207e1218.

Haimson, B.C., Herrick, C.G., 1986. Borehole breakouts e a new tool for estimating insitu stress? In: Stephansson, O. (Ed.), Rock Stress and Rock Stress Measurement,Proc. Int. Symp. on Rock Stress and Rock Stress Meas, pp. 271e280.

Haimson, B.C., Song, I., 1993. Laboratory study of borehole breakouts in CordovaCream: a case of shear failure mechanism. Int. J. Rock Mech. Min. Sci. Geomech.Abstr. 30, 1047e1056.

Haimson, B.C., Chang, C., 2002. True triaxial strength of the KTB amphibolite underborehole wall conditions and its use to estimate the maximum horizontal insitu stress. J. Geophys. Res. 107, 2257e2270.

Haimson, B.C., Lee, H., 2004. Borehole breakouts and compaction bands in twohigh-porosity sandstones. Int. J. Rock Mech. Min. Sci. 41, 287e301.

Herrick, C.G., Haimson, B.C., 1994. Modeling of episodic failure leading to bore-hole breakouts in Alabama limestone. In: Nelson, P., Laubach, S. (Eds.), RockMechanics: Models and Measurements Austin. Balkema, Rotterdam, pp. 217e224.

Hickman, S.H., Healy, J.H., Zoback, M.D., 1985. In situ stress, natural fracture dis-tribution and borehole elongation in Auburn Geothermal Well, Auburn, NewYork. J. Geophys. Res. 90, 5497e5512.

Hickman, S., Zoback, M.D., 2004. Stress orientation and magnitudes in the SAFODpilot hole. Geophys. Res. Lett. 31, L15S12. http://dx.doi.org/10.1029/2004GL020043.

Jaeger, J.C., Cook, N.G.W., 1979. Fundamental of Rock Mechanics, third ed. Chapmanand Hall, p. 515.

Kimura, G., Screaton, E.J., Curewitz, D., 2007. NanTroSEIZE Stage 1: NanTroSEIZEmegasplay and frontal thrusts. IODP Sci. Prosp. 316. http://dx.doi.org/10.2204/iodp.sp.316.2007.

Kinoshita, M., Tobin, H., Thu, M.K., Gaillot, P., Expedition 314 Scientists, 2008.NanTroSEIZE Stage IA: NanTroSEIZE LWD Transect. Integrated Ocean Drill.Program Prelim. Rep., 314, pp. 1e63.

Lee, H., Ong, S., Azeemuddin, M., Goodman, H., 2012. A wellbore stability model forformations with anisotropic rock strengths. J. Pet. Sci. Eng. 96-97, 109e119.

Lin, W., Yeh, E., Hung, J., Haimson, B., Hirono, T., 2010. Localized rotation of principalstress around faults and fractures determined from borehole breakouts in holeB of the Taiwan Chelungpu-fault drilling project (TCDP). Tectonophysics 482,82e91.

Louis, L., Humbert, F., Robion, P., Henry, P., Knuth, M., Likos, W., 2012. Data report:joint analysis of acoustic and magnetic susceptibility anisotropies in the Nankaiaccretionary prism. In: Kinoshita, M., Tobin, H., Ashi, J., Kimura, G., Lallemant, S.,Screaton, E.J., Curewitz, D., Masago, H., Moe, K.T., the Expedition 314/315/316Scientists (Eds.), Proc. IODP, 314/315/316. Integrated Ocean Drilling ProgramManagement International, Inc., Washington, DC http://dx.doi.org/10.2204/iodp.proc.314315316.216.2012.

Mastin, L.G., Heinemann, B., Krammer, A., Fuchs, K., Zoback, M.D., 1991. Stressorientation in the KTB pilot hole determined from wellbore breakouts. Scient.Drill. 2, 1e12.

McLamore, R., Gary, K.E., 1967. The mechanical behavior of anisotropic sedimentaryrocks. J. Eng. Ind. Trans. ASME 89, 62e73.

Miyazaki, S., Heki, K., 2001. Crustal velocity field of southwest Japan: subductionand arcearc collision. J. Geophys. Res. 106, 4305e4326. http://dx.doi.org/10.1029/2000JB900312.

Moore, G.F., et al., 2007. Three-dimensional splay fault geometry and implicationsfor tsunami generation. Science 318, 1128e1131. http://dx.doi.org/10.1126/science.1147195.

Moore, G., Kanagawa, K., Strasser, M., Dugan, B., Maeda, L., Toczko, S., the Expedition338 Scientists, 2013. NanTroSEIZE Stage 3: NanTroSEIZE Plate Boundary DeepRiser 2 (IODP Prel. Rept., 338). http://dx.doi.org/10.2204/iodp.pr.338.2013.

Moore, J.C., Chang, C., McNeill, L., Thu, M.K., Yamada, Y., Huftile, G., 2011. Growth ofborehole breakouts with time after drilling: implications for state stress, Nan-TroSEIZE transect, SW Japan. Geochem. Geophys. Geosyst. 12 (4), Q04D09.http://dx.doi.org/10.1029/2010GC003417.

Moos, D., Zoback, M.D., 1990. Utilization of observations of well bore breakouts toconstrain the orientation and magnitude of crustal stress: applications tocontinental, Deep Sea Drilling Project, and Ocean Drilling Program boreholes.J. Geophys. Res. 95, 9305e9325.

Morin, R.H., Newmark, L.R.L., Barton, C.A., Anderson, R.N., 1990. State of lithosphericstress and borehole stability at deep sea drilling project site 504B, Easternequatorial pacific. J. Geophys. Res. 95 (B6), 9293e9303.

Niando, H., Shao, J.F., Henry, J.P., Fourmaintraux, D., 1997. Laboratory investigation ofthe mechanical behaviour of Tournemire shale. Int. J. Rock Mech. Min. Sci. 34(1), 3e16.

Okland, D., Cook, J.M., 1998. Bedding related borehole instability in high-anglewells. In: SPE 47285, SPE/ISRM Conference on Rock Mechanics in PetroleumEngineering, Eurock’ 98, Trondheim, pp. 413e422.

Park, J.-O., Tsuru, T., Kodaira, S., Cummins, P.R., Kaneda, Y., 2002. Splay faultbranching along the Nankai subduction zone. Science 297 (5584), 1157e1160.http://dx.doi.org/10.1126/science.1074111.

Plumb, R.A., Cox, J.W., 1987. Stress directions in Eastern North America determinedto 4.5 km from borehole elongation measurements. J. Geophys. Res. 92 (B6),4805e4816.

Ramamurthy, T., Rao, G.V., Singh, J., 1993. Engineering behaviour of phyllites. Eng.Geol. 33, 209e225.

Seno, T., Stein, S., Gripp, A.E., 1993. A model for the motion of the Philippine Seaplate consistent with NUVEL-1 and geological data. J. Geophys. Res. 98 (17),17941e17948. http://dx.doi.org/10.1029/93JB00782.

Shamir, G., Zoback, M.D., 1992. Stress orientation profile to 3.5 km depth near theSan Andreas Fault at Cajon Pass, California. J. Geophys. Res. 97 (B4), 5059e5080.

Tien, Y.M., Kuo, M.C., Juang, C.H., 2006. An experimental investigation of the failuremechanism of simulated transversely isotropic rocks. Int. J. Rock Mech. Min. Sci.43, 1163e1181.

Tobin, H., Kinoshita, M., Ashi, J., Lallemant, S., Kimura, G., Screaton, E., Moe, K.-T.,Masago, H., Curewitz, D., the Expedition 314/315/316 Scientists, 2009. Nan-TroSEIZE Stage 1 expeditions: introduction and synthesis of key results. In:NanTroSEIZE Stage 1: Investigations of Seismogenesis, Nankai Trough, Japan[online], Proc. Integr. Ocean Drill. Program, 314/315/316. http://dx.doi.org/10.2204/iodp.proc.314315316.101.

Vernik, L., Zoback, M.D., 1990. Strength anisotropy in crystalline rock: implicationsfor assessment of in situ stresses from wellbore breakouts. In: Rock MechanicsContributions and Challenges, Proc. 31st US Symp. Balkema, Rotterdam,pp. 841e848.

Willson, S.M., Last, N.C., Zoback, M.D., Moos, D., 1999. Drilling in South America:stability approach for complex geological conditions. In: SPE 53940, SPE LatinAmerican & Caribbean Petrol. Eng. Conf. Caracas.

Willson, S.M., Edwards, S.T., Crook, A., Bere, A., Moos, D., Peska, P., Last, N., 2007.Assuring stability in extended-reach wells e analyses, practices, and mitigation.In: SPE/IADC 105405, SPE/IADC Drilling Conference, Amsterdam.

Zang, S.X., Chen, Q.Y., Ning, J.Y., Shen, Z.K., Liu, Y.G., 2002. Motion of the PhilippineSea plate consistent with the NUVEL-1A model. Geophys. J. Int. 150 (3), 809e819. http://dx.doi.org/10.1046/j.1365-246X.2002.01744.x.

Zheng, Z., Kemeny, J., Cook, N.G.W., 1989. Analysis of borehole breakouts. J. Geophys.Res. 94, 7171e7182.

Zhou, S., 1994. A program to model the initial shape and extent of boreholebreakout. Comput. Geosci. 20 (7/8), 1143e1160.

Zoback, M.D., Moose, D., Mastin, L., Anderson, R.N., 1985. Well-bore breakouts andin situ stress. J. Geophys. Res. 90, 5523e5538.

Zou, Y., Taylor, W.E.G., Heath, D.J., 1996. A numerical model for borehole breakouts.Int. J. Rock Mech. Min. Sci. Geomech. Abstr. 33 (1), 103e109.