Tectonophysics 482 (2010) 82–91

Contents lists available at ScienceDirect

Tectonophysics
j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / t e c t o

Localized rotation of principal stress around faults and fractures determined from
borehole breakouts in hole B of the Taiwan Chelungpu-fault Drilling Project (TCDP)
Weiren Lin a,⁎,1, En-Chao Yeh b,2, Jih-Hao Hung c, Bezalel Haimson d, Tetsuro Hirono e
a
Kochi Institute for Core Sample Research, Japan Agency for Marine-Earth Science and Technology, 200 Monobe-otsu, Nankoku, Kochi 783-8502, Japan
b
Kochi Institute for Core Sample Research, Japan Agency for Marine-Earth Science and Technology, Japan
c
Institute of Geophysics, National Central University, Chung-Li 32054, Taiwan
d
Department of Materials Science and Engineering and the Geological Engineering Program, University of Wisconsin, Madison, WI, 53706, USA
e
Graduate School of Science, Osaka University, Osaka 560-0043, Japan

a r t i c l e i n f o a b s t r a c t

Article history: To reveal details of stress perturbations associated with faults and fractures, we investigated the faults and
Received 12 September 2008 large fractures accompanied by stress-induced borehole breakouts or drilling-induced tensile fractures in
Received in revised form 10 April 2009 hole B of the Taiwan Chelungpu-fault Drilling Project (TCDP). Then, we determined the relationship between
Accepted 15 June 2009
the faults and fractures and stress orientation changes. We identified faults and fractures from electrical
Available online 23 June 2009
images of the borehole wall obtained by downhole logging but also from photographs and descriptions of
Keywords:
retrieved core samples, and measured the variations in the principal horizontal stress orientation ascertained
Stress orientation from borehole breakouts observed on the electrical images in the vicinity of the faults and fractures.
Stress perturbation Identification of geological structures (faults, fractures, and lithologic boundaries) by electrical images only is
Fault difficult and may sometimes yield incorrect results. In a novel approach, therefore, we used both the
Fracture electrical images and core photographs to identify geological structures. We found four patterns of stress
Borehole breakout orientation change, or no change, in the vicinity of faults and fractures in TCDP hole B: (i) abrupt
(discontinuous) rotation in the vicinity of faults or fractures; (ii) gradual rotation; (iii) suppression of
breakouts at faults, fractures, or lithologic boundaries; and (iv) no change in the stress orientation. We
recognized stress fluctuations, that is, heterogeneous mesoscale (≥ 10 cm) stress distributions with respect
to both stress orientation and magnitude. In addition, we found that stress state changes occurred frequently
in the vicinity of faults, fractures, and lithologic boundaries.
© 2009 Elsevier B.V. All rights reserved.

1. Introduction 2000; Ma et al., 2000; Shin and Teng, 2001; Wang et al., 2002; Ji et al.,
2003; Yue et al., 2005). To understand the physics of the earthquake and
It is well known that stress and earthquakes are interrelated: stress the mechanism of rupture propagation, the TCDP drilled two vertical
triggers earthquakes and earthquakes alter shear and normal stresses on holes 40 m apart (hole A to an approximate depth of 2000 m and hole B
fault planes (Stein,1999; Seeber and Armbruster, 2000; Hardebeck, 2004; to an approximate depth of 1350 m) about 2 km east of the surface
Ma et al., 2005). One effective and important approach to understanding rupture (Fig. 1b–d), near the town of Da-Keng (Ma et al., 2006). The
how stress and faulting are interrelated is to examine relationships Chelungpu fault dips gently to the east (30°), and slips principally within
between stress state changes and the presence of faults and fractures. In and parallel to the bedding of the Pliocene Chinshui Shale (Fig. 1c). The
this study, we focused on stress perturbations in the vicinity of faults and TCDP holes penetrate three major fault zones (Fig. 1d; e.g., Hirono et al.,
fractures by using data from the Taiwan Chelungpu-fault Drilling Project 2006) within the Chinshui Shale, which, despite its formal lithostrati-
(TCDP), similar to a previous study (Lin et al., 2007a). graphic name, in this area is composed mainly of siltstone (Lin et al.,
The large, destructive Chi-Chi earthquake (Mw 7.6) occurred in west- 2007c). Each of the three fault zones has components of wall rock,
central Taiwan on 21 September 1999 as a result of convergence between damage zone and fault core. By regional stratigraphic constraint,
the Philippine Sea and Eurasian plates (Fig. 1a and b; Kao and Chen, however, we believe they will merge into one below certain depth
since we did not find distinct stratigraphic offset on the surface geology.
A main objective of the TCDP was to determine the spatial distri-
⁎ Corresponding author. Fax: +81 88 878 2192. bution of the in situ stress and, in particular, to determine the stress state
E-mail address: lin@jamstec.go.jp (W. Lin).
1
Key Laboratory of Tectonics and Petroleum Resources of Ministry of Education,
on and around the fault plane before, during, and after the earthquake.
China University of Geosciences, Wuhan, China. Previous stress-related studies in the area have reported the focal
2
Now at Department of Geosciences, National Taiwan University, Taipei 106, Taiwan. mechanisms of earthquakes occurring before the 1999 Chi-Chi

0040-1951/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.tecto.2009.06.020
 W. Lin et al. / Tectonophysics 482 (2010) 82–91 83

Fig. 1. Schematic diagram illustrating the tectonics, regional structures, geological section and two drill holes. (a) Tectonic setting of Taiwan and location of the 1999 Chi-Chi
earthquake epicenter (after Lee et al., 2002). (b) Geological map showing the formation distribution and the several faults in the central portion of the western Taiwan. The
Chelungpu fault (red line) ruptured during the 1999 earthquake (after Yeh et al., 2007). The Taiwan Chelungpu-fault Drilling Project (TCDP) site is indicated by a red star. The focal
mechanism of the Chi-Chi main shock is located at the hypocenter of the Chi-Chi earthquake (Kao and Chen 2000). (c) Cross section through the drill site illustrates the relation
between formations and major fault zones (after Yeh et al., 2007). (d) Diagram shows the three major fault zones belonging to the Chelungpu fault system penetrated by the two
TCDP holes A and B. The dashed frames in (a) and (c) show the areas of (b) and (d) respectively; and the dashed line in (b) shows the orientation and range of the section map (c).

earthquake (Yeh et al., 1991), and a stress tensor inversion of the Chi-Chi orientations mainly from borehole breakouts in hole A and hole B,
earthquake sequence (Kao and Angelier, 2001; Wang and Chen, 2001; respectively; Lin et al. (2007b) published preliminary results of three-
Ma et al., 2005; Blenkinsop, 2006) revealed some regional stress dimensional stress determination by anelastic strain recovery (ASR) of
information in the Taiwan area. Moreover, as part of the TCDP, Hung et al. core samples from hole A; and Yabe et al. (2008) measured stresses by
(2009) determined stress magnitudes by hydraulic fracturing at four acoustic emission (AE) and deformation rate analysis (DRA) of core
depths in hole B; Wu et al. (2007) and Lin et al. (2007a) obtained stress samples. In addition, Haimson et al. (2009) have estimated the
84 W. Lin et al. / Tectonophysics 482 (2010) 82–91

maximum principal horizontal stress magnitude by using the combina- core samples at core depths of approximately 1136, 1194, and 1243 m
tion of borehole breakout width and rock strength data from true triaxial (Hirono et al., 2006). The depth differences arise from the drilling rod
compression tests, which are considered able to replicate in situ stress length and wire length of downhole logging tools which makes the
conditions reliably (see Haimson and Chang, 2002). logging depths usually shallower by approximately 3 m than the core
Lin et al. (2007a) found clearly recognizable principal stress rotations depths. In addition to the three major fault zones, 30 minor faults with
in the vicinity of the shallowest major fault zone, at 1133 m logging stress-induced breakouts or drilling-induced tensile fractures in their
depth in hole B, suggesting that this fault zone ruptured during the 1999 vicinity were recognized both in the FMI images and the core samples.
earthquake. Their study, however, focused on only this major fault zone. In this study, we define minor faults as those less than 0.5 m wide and
To reveal the details of stress perturbations associated not only with accompanied by fault gouge or showing displacement.
major fault zones but also with minor faults and fractures, in this study, Since both stress-induced borehole breakouts and drilling-induced
following Lin et al. (2007a), we investigated all of the faults and large tensile fractures depend on in situ stress conditions, we can use
fractures accompanied by stress-induced borehole breakouts or drilling- information on their geometry, observed in the borehole wall images, to
induced tensile fractures and their relationship with stress orientation in estimate orientations of in situ principal stresses in the plane perpendi-
TCDP hole B. We identified faults and fractures not only from Fullbore cular to the borehole axis without knowledge of any other parameter
Formation MicroImager (FMI) electrical images of the borehole walls values (Zoback et al., 1985; Barton et al., 1988; Moos and Zoback, 1990;
but also from photographs and descriptions of retrieved core samples, Vernik and Zoback, 1992; Zoback et al., 2003; Haimson, 2007). The
and then measured the variations in the principal horizontal stress azimuth of the breakout is the same as the azimuth of the minimum
orientations from the borehole breakouts identified in the electrical principal horizontal stress Shmin but differs by 90° from the azimuth of
images in the vicinity of the faults and fractures. Identification of maximum principal horizontal stress SHmax; whereas the azimuth of the
geological structures such as faults, fractures, and lithologic boundaries drilling-induced tensile fracture is the same as the azimuth of SHmax (e.g.,
from only electrical borehole images is difficult and may sometimes Zoback et al., 2003). Many borehole breakouts and a few tensile fractures
yield incorrect results. Therefore, we used both electrical images and were observed within the surveyed depth range in TCDP hole B. Breakouts
core photographs to identify the geological structures. This novel or tensile fractures are almost found in pairs, with one member of the pair
approach allowed us to examine stress perturbations resulting from the opposite the other in the borehole wall. Thus, the difference in orientation
presence of the geological structures. In TCDP hole B we found four types between the two positions is always approximately 180°. We recorded the
of relationship between faults and fractures and changes of borehole azimuth data of breakouts and tensile fractures according to the criteria
breakout and drilling-induced tensile fractures in their vicinity, described by Lin et al. (2007a), and plotted the azimuth profile of the
indicating principal horizontal stress orientation changes. From these SHmax (Fig. 2). Because we conducted the FMI logging in hole B
observations, we recognized heterogeneous mesoscale (≥10 cm) stress approximately 5 years and 7 months after the last slip event, the results
distributions with respect to both orientation and magnitude. Moreover, represent the post-seismic stress state at the time of drilling.
we found that stress state changes were frequent in the vicinity of faults, In general, the azimuths of SHmax were mostly distributed between
fractures, and lithologic boundaries. N105E and N155E at all surveyed depths, except at around 1133 m. We

2. Profile of current principal horizontal stress orientations in

Lin et al. (2007a) carried out a stress analysis in TCDP hole B using
both stress-induced compressive failures (borehole breakouts) and
drilling-induced tensile fractures, focusing on the stress state in the
vicinity of the shallowest major fault zone (1133 m logging depth). At
depths near this shallowest major fault zone, they compiled the stress
states, along with several independent stress measurements obtained
by the ASR method (Lin et al., 2007b) and hydraulic fracturing tests
(Hung et al., 2009). Importantly, the data suggest that the principal
horizontal stress orientations rotated abruptly by approximately 90° in
the vicinity of the major fault zone. This can be considered important
scientific evidence that the Chelungpu fault ruptured at this major fault
zone during the 1999 Chi-Chi earthquake. Furthermore, Lin et al.
(2007a) proposed a constraint of possible magnitude range for the
current principal horizontal stresses at depths near the fault zone.
In this study, we ascertained borehole breakouts, drilling-induced
tensile fractures and the orientation profile of the current principal
horizontal stresses, determined at 1-m intervals, in TCDP hole B from the
FMI images and calculated the statistics of the orientations. Then, we
examined the characteristics of the stress orientation changes at 25-cm
intervals in the vicinity of all minor faults and major fractures observed
in the depth range of the FMI survey and retrieved core samples.
We conducted downhole wireline FMI logging twice to obtain
electrical images of the borehole wall from approximately 930 to
Fig. 2. Profile of current principal horizontal stress orientations in TCDP hole B. The
1330 m depth in hole B immediately after the target drilling depth was azimuth distribution of the maximum principal horizontal stress SHmax determined
reached. In hole B, core samples were fully retrieved in the depth from borehole breakouts (△ and ▽) and drilling-induced tensile fractures (+ and ×).
range from 950 to 1350 m (e.g., Hirono et al., 2007). Therefore, we set The symbols △ and + and ▽ and × refer to data obtained during the first and second
the investigation depth range in this study between 950 m and FMI downhole logging runs (P1 and P2), respectively. The solid circles show the average
azimuths of SHmax from both breakouts and tensile fractures in each sub-range; vertical
1330 m, where both FMI images and core samples were available. bars show the depths of the sub-ranges; horizontal error bars show the standard
Three major fault zones were observed in the FMI electrical images at deviations. Arrows labeled Fig. 2 to Fig. 7 correspond to the locations shown in the
logging depths of approximately 1133, 1191, and 1240 m, and in the respective figures. Modified from Lin et al., (2007a).
 W. Lin et al. / Tectonophysics 482 (2010) 82–91 85

Table 1 (near 1170 m depth) to N133E (near 1300 m). This azimuth of N119–
Statistics of the azimuths of maximum principal horizontal stress in sub-ranges of the 133E is consistent with the downdip direction of the bedding planes of
survey depth range.
the Chinshui Shale, and is also the same as the azimuth of rupture of
Sub-range Depth range (m) Data number Average Standard the Chelungpu fault during the 1999 Chi-Chi earthquake (Lin et al.,
azimuth (°) deviation (°) 2007c) approximately. Moreover, this orientation is in good agreement
1 937–997 69 132 22 with the regional stress state at the TCDP drilling site obtained from
2 1001–1057 50 120 14
fault slip analyses of the 1999 Chi-Chi earthquake data (Kao and
3 1073–1119 63 133 16
4 1120–1139 14 212 22 Angelier, 2001; Blenkinsop, 2006) and tectonic stress data for central
5 1143–1189 20 119 19 Taiwan (Yeh et al., 1991). As mentioned above, the stress orientation
6 1215–1244 16 124 10 changed abruptly by approximately 90° in the vicinity of the fault zone
7 1258–1328 53 133 9 at 1133 m depth. Even disregarding this abrupt change, the azimuths of
SHmax were not constant but varied abruptly or gradually. Most of these
azimuthal variations were associated with faults, fractures, or litho-
divided the whole surveyed depth range into several sub-ranges logic boundaries.
(shown in Table 1 and Fig. 2 by vertical bars) according to the presence
or absence of breakouts and tensile fractures and also considering 3. Descriptions of stress orientation rotations and discussions
lithologic boundaries. Azimuthal standard deviations of all sub-ranges
were between 9° and 22°. The azimuthal standard deviations can be 3.1. Abrupt changes in stress orientation in the vicinity of faults and
considered to include both measurement errors and stress orientation fractures
fluctuations. Thus, the precision of the azimuth data obtained from the
borehole breakouts and the tensile fractures was generally less than Our first example of an abrupt, drastic change in the orientation of
the azimuthal standard deviations. The average azimuths of the sub- the principal horizontal stresses is in the vicinity of a minor fault at
ranges were concentrated within a relatively narrow range from N119E approximately 1119.7 m logging depth (Fig. 3). The minor fault, which

Fig. 3. An example of an abrupt stress orientation change in the vicinity of a minor fault at 1119.7 m logging depth. (a) FMI electrical image (unrolled electrical image of the borehole
wall in TCDP hole B). The left edge corresponds to north. Dark colors in the image represent conductive areas and light colors resistive areas. (b) An optical image of the flat surface of
a half-core split in the vertical plane approximately parallel to the downdip direction. The constriction ratio in the vertical and lateral directions is 2:1 in this figure and also in Figs. 3–
7; the core diameter in Figs. 2–7 is approximately 83 mm. (c) An optical photograph of the cylindrical core taken on the drilling rig floor immediately after retrieval. (d) Plot of the
SHmax azimuth over a depth range of several meters around this minor fault.
86 W. Lin et al. / Tectonophysics 482 (2010) 82–91

was interpreted as a left-lateral strike-slip fault with normal slip from the maximum principal horizontal stress was consistent with the
the core description, has a large dip angle (dip/dip orientation: 70/050 tectonic stress orientation before the earthquake. In addition, this
approximately). The azimuth of the breakout abruptly rotates by stress state was also consistent with the reverse (thrust) fault regime
approximately 90° (see Fig. 3a and d). Therefore, at the minor fault which agrees with fault rupturing type of the Chi-Chi earthquake. As a
depth, the orientation of SHmax changes roughly from the rupturing result of the stress dropping during the fault rupturing, it is reasonable
direction of the 1999 earthquake to one perpendicular to the to infer that the stress in the rupturing direction changed from the
rupturing direction. Within a narrow depth interval (centered at maximum principal horizontal stress into the minimum principal
around 1133 m) below the minor fault, the SHmax azimuth (N212E on horizontal stress through the earthquake; whereas the stress in the
average) differs by approximately 90° from its azimuth at other depths strike direction changed from the minimum horizontal stress into the
(Table 1 and Fig. 2). That is, the minor fault at 1119.7 m depth, where maximum horizontal stress in the vicinity of the fault zone ruptured.
the SHmax azimuth abruptly rotates, is the boundary between the Our second example (Fig. 4) is of an abrupt but minor orientation
normal stress orientation, one consistent with the regional stress rotation of the principal horizontal stress across a fracture at
orientation, and an anomalous stress orientation caused by the 1999 approximately 1077.4 m logging depth (corresponding to the optical
earthquake (Lin et al., 2007a). Moreover, the SHmax azimuth (N212E split core image in Fig. 4c). Across the fracture (60/050 approxi-
on average) at approximately 1133 m in hole B is consistent with the mately), the breakout (Shmin) azimuth abruptly rotates by approxi-
stress orientation estimated by the ASR method at the depth of mately 20°, but it remains almost constant above and below the
the shallowest major fault zone in hole A (Lin et al., 2007b). Yamashita fracture. In the vicinity of this fracture, the retrieved core samples
et al. (2004) reported a similar example, in which the Nojima fault slip were broken, so the fracture attitude could not be identified by the
during the 1995 Kobe earthquake induced rotation of the principal core samples, but correlation between FMI image and core images of
stress axis orientation. this fracture could be identified by taking into account the relation-
Due to the thrust fault rupturing of the 1999 earthquake, it could be ship between logging and core depths. Additionally, it is clear that the
considered that both the maximum and minimum principal horizontal fracture is at the lithologic boundary between a layer of alternating
stresses dropped in the vicinity of the shallowest major fault zone. sandstone and shale (upper side; Fig. 4a) and a layer of bioturbated
However, drops of the two principal horizontal stress magnitudes may sandstone (Fig. 4c). It could be inferred that the minor and abrupt
differ from each other. Because the stress in the rupturing direction stress rotation of across a fracture at approximately 1077.4 m logging
was the driving force, drop of the stress magnitude in this direction was depth may be caused by the mechanical properties such as Young's
more than that in the direction perpendicular to the rupturing modulus change associated with the lithology change.
direction (the fault strike direction). Therefore, it is possible that Another example of an abrupt rotation in stress orientation (Fig. 5,
magnitude of the stress in the rupturing direction was more than that in the vicinity of 1031.0 m) was identified on the basis of drilling-
of the stress in the strike direction before the fault rupturing. That is, induced tensile fractures. Across two fractures (logging depth, around

Fig. 4. An example of an abrupt but minor orientation change of the principal horizontal stress across a fracture (at 1077.4 m depth) and examples of no orientation change around
several fractures are shown in the optical image in (a) and the lower fracture in the optical image in (c). (b) FMI electrical image. (d) A plot of the SHmax azimuth over the
corresponding depth range.
 W. Lin et al. / Tectonophysics 482 (2010) 82–91 87

Fig. 5. Examples of an abrupt but minor stress orientation change across two fractures (around 1031.0 m) determined from drilling-induced tensile fractures and breakout
suppression at the lithologic boundary (1032.2 m depth). (a and c) Optical images of the flat surface of the half-core splits. (b) FMI electrical image.

1031.0 m; 30/120; approximately parallel to bedding of the Chinsui consistent with the regional stress orientation. Because the gradual
Shale), the locations of the tensile fractures, which are the same as the stress rotation is local (see the arrow labeled as Fig. 6 in Fig 2) and the
azimuths of SHmax (Zoback et al., 2003), abruptly rotated by 12° lithologies in hanging and foot walls are almost the same (Fig. 6a and
approximately. For this example, the main reason that induced abrupt b), the rotation might be interpreted due to the presence of the minor
stress rotation might be the presence of the fractures. fault.
Similarly, in the SAFOD (San Andreas Fault Zone Observatory at
3.2. Gradual change in stress orientation in the vicinity of faults and Depth) pilot hole, Hickman and Zoback (2004) observed gradual
fractures rotations of the localized principal stress orientation around fault
zones. For example, the apparent azimuth of SHmax just below a minor
Immediately below a minor fault (approximately 3 cm thick; fault gradually rotated by approximately 70° over an interval of only
logging depth, 975.4 m; 30/130 approximately) (Fig. 6a–c), breakouts 5 m. Moreover, Shamir and Zoback (1992) and Barton and Zoback
were identified that gradually rotate below the fault by 90° (1994) presented additional examples of stress rotation associated
approximately over an interval of approximately 2.7 m, until reaching with fault movements from the Cajon Pass drill hole near the San
another fracture at 978.0 m depth. Below this fracture (978.0 m), the Andreas fault and the KTB (Kontinentales Tiefbohprogramm der
breakouts do not rotate but maintain an almost constant azimuth Bundesrepublik) drill hole, respectively.
88 W. Lin et al. / Tectonophysics 482 (2010) 82–91

Fig. 6. Examples of a gradual change in the stress orientation below a minor fault (975.4 m), of breakout suppression at the same minor fault and of no change in the stress orientation
across a fracture (978.0 m). (a and f) Optical photographs of the cylindrical core. (b and e) Optical images of the flat surface of the half-core splits. (c) FMI electrical image. (d) A plot
of the SHmax azimuth over the corresponding depth interval.

3.3. Suppression of breakouts at faults or fractures or lithologic boundaries present in the foot wall of the fault (Fig. 6c). This suggests that
breakouts were suppressed at this minor fault. It is clear from the
Breakouts were absent in the hanging wall of the minor fault at optical core images that the lithology of the hanging and foot walls is
975.4 m logging depth described in Section 3.2, but breakouts were almost the same (Fig. 6a and b), suggesting that their compressive
 W. Lin et al. / Tectonophysics 482 (2010) 82–91 89

strengths are likely to be approximately equal. Thus, we can infer that sandstone with bioturbation is higher than that in the pure sandstone.
the stress state, especially the magnitude of the maximum principal If this interpretation is true, then Young's modulus of the sandstone
horizontal stress, differs greatly between the two parts. with bioturbation is much higher than that of the pure sandstone,
Another example of suppression of breakouts at a minor fault which suggests that the harder or stronger rock formation bears a high
(thickness, b5 mm with thin and soft gouge; dip, 40/260 approxi- load due to gravitational or tectonic loading. These observations again
mately; a right-lateral oblique strike-slip fault) similarly suggests that indicate that the stress distribution is not homogeneous in actual,
the magnitude of the maximum principal horizontal stress differs natural formations.
between the hanging and foot walls (Fig. 7). It is likely that the stress In addition, the azimuth of the breakout below the lithologic
magnitude changes stepwise across the fault. boundary differs by approximately 90° from that of the tensile
An example of breakout suppression at a lithologic boundary can be fractures above the boundary (Fig. 5b). That is, the orientation of S-
recognized at around 1032.2 m logging depth (Fig. 5b and c). Above the Hmax is the same above and below the boundary, although it fluctuates
boundary is pure sandstone (i.e. without bioturbation), and below it is around some fractures. Taken together, these observations of stress
sandstone with strong bioturbation. The porosity, measured from the orientation and our interpretation of a stress magnitude change,
core samples from TCDP hole B, of the pure sandstone is 22.7 ± 3.5% mentioned above, suggest that the magnitude of SHmax changed
(average ± standard deviation, n = 95) and that of the sandstone with stepwise, but its orientation remained the same, across this lithologic
bioturbation is 10.3 ± 3.6% (n = 98) (Lin et al., 2008). Because of the boundary.
higher porosity of the pure sandstone, it is reasonable to infer that its
compressive strength is lower than that of the sandstone with 3.4. No change in stress orientation across faults or fractures
bioturbation. As a suite of useful experimental data, Chen (2005)
conduct uniaxial compression tests with the TCDP hole A core samples Stress orientations do not always change around faults and
under water-saturated conditions. The author showed that the average fractures. For example, the breakouts remain at almost the same
porosity and uniaxial compressive strength of a pure sandstone azimuth in the vicinity of a minor fault (logging depth, 1088.3 m; 30/
specimen were 15.1% and 6.4 MPa, respectively; whereas the values 120 approximately) (Fig. 8) inferring SHmax orientation does not
of sandstone with bioturbation were 6.4% and 30.4 MPa (average, change. Other examples of an approximately constant stress orienta-
n = 3), respectively. Therefore, the prediction that the sandstone with tion across fractures occur around several fractures between 1075.9
bioturbation is stronger much than the pure sandstone can be and 1077.1 m depth (Fig. 4a) and around a fracture at 1077.8 m depth
considered to be reliable. Thus, the presence of breakouts in the (Fig. 4c), all of which are parallel to the formation bedding with 30/
stronger rock, and not in the weaker rock, is interesting, because in 120. Also, around a fracture at 978.0 m depth (30/120 approximately),
general breakouts occur easily in rock with lower strength if the stress no stress orientation change was recognized (Fig. 6c, e, and f). The lack
magnitudes are the same. Therefore, a possible interpretation is that of a stress orientation change suggests that the stress state remains
the magnitude of the maximum principal horizontal stress in the mostly continuous around faults and fractures if the lithologies of the

Fig. 7. An example of breakout suppression at a fracture (1083.2 m depth). (a) FMI electrical image. (b) Optical image of the flat surface of the half-core split. (c) Optical photograph of
the cylindrical core.
90 W. Lin et al. / Tectonophysics 482 (2010) 82–91

Fig. 8. Example of consistent stress orientation across a minor fault (1083.3 m). (a) FMI electrical image. (b) Optical image of the flat surface of the half-core split. (c) Optical
photograph of the cylindrical core.

hanging and foot walls are similar and if shear stresses on the faults or 5. Conclusions
fractures are less than their frictional strengths.
To understand stress perturbations associated with minor faults
4. Distribution and quantity of faults and fractures according to and fractures, we investigated all of the faults and large fractures
stress orientation change patterns accompanied by borehole breakouts or drilling-induced tensile
fractures and their relationship to stress changes in TCDP hole B. We
We determined the numbers of faults and fractures in the target identified faults and fractures not only from the FMI electrical images
depth range of 950–1330 m with respect to the four patterns of stress of the borehole wall but also from photographs and descriptions of
orientation change (Table 2). Only those faults and fractures drilling core samples, and measured the variations in the principal
accompanied by stress-induced borehole breakouts or drilling- horizontal stress orientation in the vicinity of the faults and fractures
induced tensile fractures, where it was possible to examine any ascertained from the breakouts identified in the electrical images.
rotations of stress orientation, were counted. Thus, not all faults and Identification of geological structures (faults, fractures, and lithologic
fractures in the depth range were included. boundaries) from electrical images alone is sometimes difficult.
The proportion of minor faults with stress changes, including Therefore, we used both electrical images and core photographs to
abrupt or gradual orientation rotations or suppression of breakouts, identify the geological structures.
was clearly higher than that of fractures (Table 2). Therefore, the stress
orientation more often rotated in the vicinity of minor faults than in
the vicinity of fractures. The reason may be that faults with gouge slip Table 2
more easily than fractures. In detail, the stress state can remain Minor faults and fractures accompanied by breakouts or drilling-induced tensile
continuous across a plane such as a fault or a fracture if the shear fractures.
stress on the plane is not larger than the product of the frictional Total number of Stress orientation change Number of faults Percentage
coefficient and the normal stress on the plane. In general, a fracture faults or fractures patternsa or fractures
surface is likely to be rougher than a fault surface; on the other hand, Minor faultsb 30c Abrupt change 8 27
fault gouge, which has a relatively low frictional coefficient, is usually Gradual change 4 13
present in a fault. Therefore, the frictional coefficient of a fracture Suppression of breakouts 6 20
No change 13 43
might be higher than that of a minor fault (Byerlee, 1978; Mizoguchi
Fractures 65 Abrupt change 9 14
et al., 2008). Consequently, a stress change is more likely in the vicinity Gradual change 2 3
of minor faults compared with fractures. Suppression of breakouts 16 25
Another interesting observation is that suppression of breakouts No change 38 58
was relatively common for both minor faults and fractures. Breakouts a
See the text for detailed definitions of the stress orientation change patterns.
b
often occurred and then disappeared at faults or fractures or lithologic The number of minor faults with stress-induced borehole breakouts or drilling-
boundaries. This situation might reflect a stress state change in the induced tensile fractures in their vicinity between approximately 950 and 1330 m depth
in TCDP hole B. Thus, not all minor faults in the depth range are included.
vicinity of faults and fractures. On the other hand, in addition to a c
The minor fault at approximately 975.4 m, shown in Fig. 5, was counted as
stress change, at a lithologic boundary, a strength change on either exhibiting both a gradual change and suppression of breakouts. Therefore, the sum of
side of the boundary might also be an important factor. the third column exceeds this total, and the sum of the percentages does not equal 100%.
 W. Lin et al. / Tectonophysics 482 (2010) 82–91 91

We found four patterns of stress changes in the vicinity of faults Ji, C., Helmberger, D.V., Wald, D.J., Ma, K.-F., 2003. Slip history and dynamic implications of
the 1999 Chi-Chi, Taiwan, earthquake. J. Geophys. Res. 108. doi:10.1029/2002JB001764.
and fractures in TCDP hole B: (i) the stress orientation (breakout Kao, H., Chen, W.-P., 2000. The Chi-Chi earthquake sequence: active out-of-sequence
azimuth) rotates abruptly (discontinuously) in the vicinity of the thrust faulting in Taiwan. Science 288, 2346–2349.
faults or fractures; (ii) the orientation rotates gradually; (iii) break- Kao, H., Angelier, J., 2001. Stress tensor inversion for the Chi-Chi earthquake sequence
and its implications on regional collision. Bull. Seismol. Soc. Am. 91, 1028–1040.
outs are suppressed at faults, fractures, or lithologic boundaries; or Lee, J.-C., Chu, H.-T., Angelier, J., Chan, Y.-C., Hu, J.-C., Lu, C.-Y., Rau, R.-J., 2002. Geometry
(iv) the orientation does not change across faults or fractures. We and structure of northern surface ruptures of the 1999 Mw = 7.6 Chi-Chi Taiwan
recognized stress fluctuations, that is, heterogeneous mesoscale earthquake: influence from inherited fold belt structures. J. Struct. Geol. 24,173–192.
Lin, W., Yeh, E.-C., Ito, H., Hung, J.-H., Hirono, T., Soh, W., Ma, K.-F., Kinoshita, M., Wang,
(N10 cm) stress distributions with regard to both stress orientation
C.-Y., Song, S.-R., 2007a. Current stress state and principal stress rotations in the
and magnitude. We also found that stress state changes were common vicinity of the Chelungpu fault induced by the 1999 Chi-Chi, Taiwan, earthquake.
in the vicinity of faults, fractures, and lithologic boundaries. As further Geophys. Res. Lett. 34, L16307. doi:10.1029/2007GL030515.
Lin, W., Yeh, E.-C., Ito, H., Hirono, T., Soh, W., Wang, C.-Y., Ma, K.-F., Hung, J.-H., Song, S.-R.,
research, the stress change in the vicinity of faults and fractures
2007b. Preliminary results of stress measurement by using drill cores of TCDP hole-A:
should be quantified and interpreted in detail to better understand the an application of anelastic strain recovery method to three-dimensional in-situ stress
relationships between stress changes and faults or fractures. determination. Terr. Atmos. Ocean. Sci. 18. doi:10.3319/TAO.2007.18.2.379(TCDP).
Lin, A.T., Wang, S.M., Hung, J.H., Wu, M.S., Liu, C.S., 2007c. Lithostratigraphy of the
Taiwan Chelungpu-fault Drilling Project-A borehole and its neighboring region,
Acknowledgments central Taiwan. Terr. Atmos. Ocean. Sci. 18. doi:10.3319/TAO.TCDP0706S03.
Lin, W., Matsubayashi, O., Yeh, E.-C., Hirono, T., Tanikawa, W., Soh, W., Wang, C.-Y., Song,
We greatly appreciate Mark Tingay, the guest editor and two S.-R., Murayama, M., 2008. Profiles of volumetric water content in fault zones
retrieved from hole B of the Taiwan Chelungpu-fault Drilling Project (TCDP).
reviewers (Birgit Müller and the other anonymous reviewer) for their Geophys. Res. Lett. 34, L01305. doi:10.1029/2007GL032158.
valuable and constructive comments and suggestions which helped us Ma, K.-F., Song, T.-R., Lee, S.-J., Wu, H.-I., 2000. Spatial slip distribution of the September
to improve our manuscript much. We gratefully acknowledge H. Ito, 20, 1999, Chi-Chi, Taiwan earthquake (Mw 7.6) — inverted from teleseismic data.
Geophys. Res. Lett. 27, 3417–3420.
W. Soh, M. Kinoshita, and S. Saito of the Japan Agency for Marine-Earth Ma, K.-F., Chan, C.-H., Stein, R.S., 2005. Response of seismicity to Coulomb stress triggers
Science and Technology (JAMSTEC) for useful discussions, C.-Y. Wang, and shadows of the 1999 Mw = 7.6 Chi-Chi, Taiwan, earthquake. J. Geophys. Res.
K.-F. Ma, S.-R. Song, and Y.-B. Tsai, who are the principal investigators 110, B05S19. doi:10.1029/2004JB003389.
Ma, K.-F., Tanaka, H., Song, S.-R., Wang, C.-Y., Hung, J.-H., Tsai, Y.-B., Mori, J., Song, Y.-F.,
of TCDP, Taiwan, and Y. Kawamura, Y. Sanada, and T. Moe of the Center Yeh, E.-C., Soh, W., Sone, H., Kuo, L.-W., Wu, H.-Y., 2006. Slip zone and energetics of a
for Deep Earth Exploration at JAMSTEC for helpful support during the large earthquake from the Taiwan Chelungpu-fault Drilling Project. Nature 444,
FMI logging in TCDP hole B. W. Lin thanks the Japan Society for the 473–476. doi:10.1038/nature05253.
Mizoguchi, K., Takahashi, M., Tanikawa, W., Masuda, K., Song, S.-R., Soh, W., 2008.
Promotion of Science (JSPS) for financial support (Grant-in-Aid for
Frictional strength of fault gouge in Taiwan Chelungpu fault obtained from TCDP
Scientific Research C: 19540453). hole B. Tectonophysics 460, 198–205. doi:10.1016/j.tecto.2008.08.009.
Moos, D., Zoback, M.D., 1990. Utilization of observations of well bore failure to constrain
References the orientation and magnitude of crustal stresses: application to continental, deep
sea drilling project, and ocean drilling program boreholes. J. Geophys. Res. 95,
Barton, C.A., Zoback, M.D., 1994. Stress perturbations associated with active faults 9305–9325.
penetrated by boreholes: possible evidence for near-complete stress drop and a Seeber, L., Armbruster, J.G., 2000. Earthquakes as beacons of stress change. Nature 407,
new technique for stress magnitude measurement. J. Geophys. Res. 99, 9373–9390. 69–72.
Barton, C.A., Zoback, M.D., Burns, K.L., 1988. In situ stress orientation and magnitudes at Shamir, G., Zoback, M.D., 1992. Stress orientation profile to 3.5 km depth near the San
the Fenton geothermal site, New Mexico, determined from wellbore breakouts. Andreas Fault at Cajon Pass, California. J. Geophys. Res. 97, 5059–5080.
Geophys. Res. Lett. 15, 467–470. Shin, T.-C., Teng, T.-L., 2001. An overview of the 1999 Chi-Chi, Taiwan, Earthquake. Bull.
Blenkinsop, T.G., 2006. Kinematic and dynamic fault slip analyses: implications from Seismol. Soc. Am. 91, 895–913.
the surface rupture of the 1999 Chi-Chi, Taiwan, earthquake. J. Struct. Geol. 28, Stein, R.S., 1999. The role of stress transfer in earthquake occurrence. Nature 402, 605–609.
1040–1050. Vernik, L., Zoback, M.D., 1992. Estimation of maximum horizontal principal stress
Byerlee, J.D., 1978. Friction of rocks. Pure Appl. Geophys. 116, 615–626. magnitude of from stress-induced well bore breakouts in the Cajon Pass Scientific
Chen, C.-W., 2005. The Study of the Mechanical Characteristics from the Host Rock of Research Borehole. J. Geophys. Res. 97, 5109–5119.
the Chelungpu Fault, M.S., Thesis, 169 pp, National Taiwan University, Taipei. (In Wang, C.-Y., Li, C.-L., Yeh, H.-Y., 2002. Mapping the northern portion of the Chelungpu
Chinese with English abstract). fault, Taiwan by shallow reflection seismics. Geophys. Res. Lett. 29. doi:10.1029/
Haimson, B., 2007. Micromechanisms of borehole instability leading to breakouts in 2001GL014496.
rocks. Int. J. Rock Mech. Min. Sci. 44, 157–173. Wang, W.-H., Chen, C.-H., 2001. Static stress transferred by the 1999 Chi-Chi, Taiwan,
Haimson, B.C., Chang, C., 2002. True triaxial strength of the KTB amphibolite under earthquake: effects on the stability of the surrounding fault systems and aftershock
borehole wall conditions and its use to estimate the maximum horizontal in situ triggering with a 3D fault-slip model. Bull. Seismol. Soc. Am. 91, 1041–1052.
stresses. J. Geophys. Res. 107, 2257. doi:10.1029/2001JB000647. Wu, H.-Y., Ma, K.-F., Zoback, M., Boness, N., Ito, H., Hung, J.-H., Hickman, S., 2007. Stress
Haimson, B., Lin, W., Oku, H., Song, S.-R., 2009. Integrating borehole breakout dimensions, orientations of Taiwan Chelungpu-fault Drilling Project (TCDO) hole-A as observed
strength criterion, and hydraulic fracturing to constrain the state of stress across the from geophysical logs. Geophys. Res. Lett. 34, L01303. doi:10.1029/2006GL028050.
Chelungpu Fault, Taiwan. Tectonophysics. doi:10.1016/j.tecto.2009.05.016. Yabe, Y., Song, S.-R., Wang, C.-Y., 2008. In-situ stress at the northern portion of the
Hardebeck, J.L., 2004. Stress triggering and earthquake probability estimates. J. Geophys. Chelungpu fault, Taiwan, estimated on boring cores recovered from a 2-km-deep
Res. 109, B04310. doi:10.1029/2003JB002437. hole of TCDP. Earth Planets Space 60, 809–819.
Hickman, S., Zoback, M., 2004. Stress orientations and magnitudes in the SAFOD pilot Yamashita, F., Fukuyama, E., Omura, K., 2004. Estimation of fault strength: reconstruc-
hole. Geophys. Res. Lett. 31, L15S12. doi:10.1029/2004GL020043. tion of stress before the 1995 Kobe earthquake. Science 306, 261–263.
Hirono, T., Lin, W., Yeh, E.-C., Soh, W., Hashimoto, Y., Sone, H., Matsubayashi, O., Aoike, K., Yeh, Y.-H., Barrier, E., Lin, C.-H., Angelier, J., 1991. Stress tensor analysis in the Taiwan
Ito, H., Kinoshita, M., Murayama, M., Song, S.-R., Ma, K.-F., Hung, J.-H., Wang, C.-Y., area from focal mechanisms of earthquakes. Tectonophysics 200, 267–280.
Tsai, Y.-B., 2006. High magnetic susceptibility of fault gouge within Taiwan Yeh, E.-C., Sone, H., Nakaya, T., Ian, K.-H., Song, S.-R., Hung, J.-H., Lin, W., Hirono, T.,
Chelungpu fault: nondestructive continuous measurements of physical and Wang, C.-Y., Ma, K.-F., Soh, W., Kinoshita, M., 2007. Core description and
chemical properties in fault rocks recovered from hole B, TCDP. Geophys. Res. Lett. characteristics of fault zones from hole-A of the Taiwan Chelungpu-fault Drilling
33, L15303. doi:10.1029/2006GL026133. Project. Terr. Atmos. Ocean. Sci. 18, 327–357. doi:10.3319/TAO.2007.18.2.327(TCDP).
Hirono, T., Yeh, E.-C., Lin, W., Sone, H., Soh, W., Hashimoto, Y., Matsubayashi, O., Aoike, K., Ito, Yue, L.-F., Suppe, J., Hung, J.-H., 2005. Structural geology of a classic thrust belt
H., Kinoshita, M., Murayama, M., Song, S.-R., Ma, K.-F., Hung, J.-H., Wang, C.-Y., Tsai, Y.-B., earthquake: the 1999 Chi-Chi earthquake Taiwan (Mw = 7.6). J. Struct. Geol. 27,
Kondo, T., Nishimura, M., Moriya, S., Tanaka, T., Fujiki, T., Maeda, R., Muraki, H., 2058–2083.
Kuramoto, T., Sugiyama, K., Sugawara, T., 2007. Nondestructive continuous physical Zoback, M.D., Moos, D., Mastin, L., Andersson, R.N., 1985. Well bore breakouts and in-
property measurements of core samples recovered from hole B, Taiwan Chelungpu- situ stress. J. Geophys. Res. 90, 5523–5530.
fault Drilling Project. J. Geophys. Res. 112, B07404. doi:10.1029/2006JB004738. Zoback, M.D., Barton, C.A, Brudy, M., Castillo, D.A., Finkbeiner, T., Grollimund, B.R., Moos,
Hung, J.-H., Ma, K.-F., Wu, Y.-H., Wu, H.-Y., Ito, H., Lin, W., Yeh, E.-C., 2009. Structure D.B., Peska, P., Ward, C.D., Wiprut, D.J., 2003. Determination of stress orientation
geology, physical property, fault zone characteristics and stress state in scientific and magnitude in deep wells. Int. J. Rock Mech. Min. Sci. 40, 1049–1076.
drill holes of Taiwan Chelungpu Fault Drilling Project. Tectonophysics 466, 307–321.
doi:10.1016/j.tecto.2007.11.014.