Tectonophysics 601 (2013) 112–124

Contents lists available at SciVerse ScienceDirect

Tectonophysics

j ourna l homepage: www.e lsev ie r .com/ locate / tecto

Differences in magnetic properties of fragments and matrix of breccias from therupture of the 2008 Wenchuan earthquake, China: Relationship to faulting

Tao Yang a,⁎, Jianye Chen b, Xiaosong Yang b, Hongqiang Wang a, Haiqiang Jin a

a Key Laboratory of Seismic Observation and Geophysical Imaging, Institute of Geophysics, China Earthquake Administration, Beijing 100081, Chinab State Key Laboratory of Earthquake Dynamics, Institute of Geology, China Earthquake Administration, Beijing 100029, China

⁎ Corresponding author. Tel./fax: +86 10 6872 9195.E-mail address: yangtao@cea-igp.ac.cn (T. Yang).

0040-1951/$ – see front matter © 2013 Elsevier B.V. Allhttp://dx.doi.org/10.1016/j.tecto.2013.05.002

a b s t r a c t

a r t i c l e i n f o

Article history:Received 5 November 2012Received in revised form 28 April 2013Accepted 5 May 2013Available online 14 May 2013

Keywords:Magnetic propertiesFault brecciaSeismic faultingWenchuan earthquake

Rock magnetic and mineralogical analyses were performed on fragments and matrix of six fault breccias fromthe Zhaojiagou outcrop of Yingxiu–Beichuan Fault, which was the main fault ruptured during the 2008Wenchuan Mw 7.9 earthquake, at Leigu Town, Beichuan County, Sichuan Province (China). The matrix,which is generally enriched in dolomite, feldspar and clay minerals, but depleted in calcite, has much higherlow-field magnetic susceptibility and magnetization, and lower coercivity in comparison with fragments andbulk samples. Magnetic behavior of the bulk samples and fragments is dominated by dia-/paramagneticcomponents, in contrast, small amounts of partially oxidized magnetite and lepidocrocite are present inthe matrix. A simple conceptual model associating faulting-related effects was proposed to explain thesesignificant differences in magnetic properties of fragments and matrix. During coseismic slip, intense shearmay crush the pre-existing magnetic grains in fault rocks into finer ones; while stress and frictional heating,to a little extent, induce magnetic changes and thermochemical alterations of magnetic mineralogy in faultbreccias, respectively. During interseismic periods, meteoric fluids would infiltrate and percolate into faultzone, and cause dissolution, precipitation and recrystallization of Fe-bearingminerals. As the diverse permeabil-ity structure and grain size of fragments and matrix, these effects would modify their magnetic mineralogy atvarious levels. Consequently, faulting-related effects, especially the fluid movements, taking place over manyprevious earthquake cycles, would be the most likely reasons for the observed different magnetic propertiesin fragments and matrix. It further proposes that magnetic studies of fault breccias would provide clues tohelp understand seismic faulting and history of fault activity.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

The fault core, which is considered to represent fault zone in theseismogenic regime (Agosta and Aydin, 2006; Caine et al., 1996),generally evolves by grain comminution mostly consisting of earlybulk crushing and late abrasion of grains, ultimately leading to theconsumption of breccia materials in favor of fault gouge (Billi, 2005;Storti et al., 2003). This process thus usually results not only inphysical degradation of fault zone, but also in geochemical and min-eralogical alteration (Holland et al., 2006; Schulz and Evans, 1998).Consequently, it is pointed out that fault rocks record major episodesof fault movement, being one of the most important issues for under-standing not only the nature of fault zones, but also the entire historyof fault motion (Sibson, 1977; Tanaka et al., 2001). Fault rocks thushave been widely investigated in microstructure (Hausegger et al.,2010; Isaacs et al., 2007), grain size distribution (Billi, 2005; Hattoriand Yamamoto, 1999; Storti et al., 2003; Wilson et al., 2005), mineralassemblage (Chen et al., 2007; Isaacs et al., 2007; Matsuda et al.,

rights reserved.

2004), geochemical composition (Chen et al., 2007; Isaacs et al.,2007; Kolodny et al., 2005; Tanaka et al., 2001), and permeabilitystructure (Caine et al., 1996; Chen et al., 2011; Evans et al., 1997),to better understand their development and their physical/chemicalattributes.

Despite the wealth of data about fault cores and related rocks, todate, magnetic properties of fault rocks are still scarce (Ferré et al.,2005; Fukuchi, 2003; Hirono et al., 2006; Mishima et al., 2006;Tanikawa et al., 2008), although it is believed that magnetic proper-ties of fractured rocks are generally changed by the fragmentationand consequent alteration of magnetic minerals (Hailwood et al.,1992). Among fault rocks, fault breccia is a common product alongupper crustal fault zones, particularly in the top few kilometers ofcrust. It is made up of the survivor coarse grains (i.e., fragments)and the surrounded fine materials (namely, matrix) that are generallyproduced by local fragmentation of larger particles (Woodcock andMort, 2008). Thus, a comparative analysis of fragments and matrixin a breccia would yield information on the evolution of seismic fault.

In a previous paper, Yang et al. (2012) reported on differences insome magnetic properties comparing breccia fragments and matrixmaterial from the Zhaojiagou outcrop of the Yingxiu–Beichuan fault,

113T. Yang et al. / Tectonophysics 601 (2013) 112–124

which ruptured during the Wenchuan earthquake in 2008 (momentmagnitude Mw = 7.9). In the present paper, further mineralogicaland magnetic analyses (e.g., thermomagnetic and low-temperaturedemagnetization) of these fragments and matrix are conducted toextend the previous work and (1) to document detailed magneticproperties of fragments and matrix of fault breccias, (2) to revealthe reasons responsible for the differences in magnetic properties be-tween fragments and matrix, and (3) to infer their potential relation-ship with faulting-related effects.

2. Geologic settings and samples

The Longmenshan (LMS) thrust belt, about 500 km long and30–50 km wide, lies between the Songpan–Ganzi terrane (a Triassicorogenic belt) to the west and the Paleozoic–Cenozoic sediments ofthe Sichuan basin to the east. It is dominated by four majornortheast-trending thrusts: the Wenchuan–Maoxian fault (WMF),Yingxiu–Beichuan fault (YBF), Guanxian–Anxian fault (GAF), andPingwu–Qingchuan fault (PQF) (Fig. 1). The fault system cutsNeoproterozoic basement metamorphics marked by steeply west-ward dipping detachments (e.g., the Pengguan complex, PGC),which is overlapped on the Sichuan foreland basin (Xu et al., 2008).The YBF, trending N30–55°E and dipping 50–80° to the northwest,was the main fault that ruptured in the Wenchuan earthquake thatoccurred on May 12, 2008. A north–northeast-striking rupture witha length of ~270 km was produced along the YBF, with a maximumdisplacement amounting to 8.0–10.0 m and 5.0–6.0 m in the verticaland horizontal directions, respectively (Liu-Zeng et al., 2009). Fieldinvestigation showed that most of surface ruptures occurred along apre-existing shear zone in the LMS thrust belt. Generally, the mainco-seismic shear zone consists of a fault core that includes a narrowfault gouge zone of b15 cm in width (generally 1–2 cm), a fault brec-cia zone of b~3 m in width, and a wide damage zone of >5 m inwidth that is composed of cataclastic rocks (Lin et al., 2010).

At the Zhaojiagou village (31.81°N, 104.43°E), Leigu Town, BeichuanCounty, the fracture zone of YBF extends laterally about 8 m. The faultplane strikes N45°E and dips 64–78°NW, and a fault scarp with ~8 m ofvertical displacement was produced by the 2008 Wenchuan earthquake(Fig. 2a). The hanging wall protoliths consist of fine sandstones and

Fig. 1. Geological map of the central part of the Longmenshan (LMS) fault zone (compilWenchuan Mw 7.9 earthquake are indicated by the bold portions of the various fault traces.fault; PQF, Pingwu–Qingchuan fault.

carbonates, whereas the footwall is sandstone. The principal slip surfacecuts through the carbonate layers. Different types of fault rocks occuracross fault zone. From west to east, they are (1) shattered limestone,(2) fractured limestone, (3) crushed breccias, (4) gray and dark blackgouge, (5) crushed breccias, (6) shattered breccias, and (7) fracturedsandstone (Fig. 2b and c). The internal structure of the fault zone isrelatively simple. Only one principal slip zone can be found and nobranches exist. The width of the fault core is extremely narrow(~20 cm), consisting of ~2-cm-thick gray gouge in the center andweakly foliated yellowish crushed breccias besides. An even thinnerblack gouge layer (~3 mm) is developed adjacent to the gray gouge.The nonclay fault rocks are progressively cemented toward the faultcenter (Fig. 2c).

Microstructure of fault rocks is shown in Fig. 3. Dense X-shapedfractures are developed in breccias, and network of fractures increasetoward the gouge, while fragmented grains also get finer. Gouge isfoliated, and the formation of color bandings may be due to thedifferent mineral compositions or grain sizes (Fig. 3a). Multiple defor-mation events and healing processes are clearly evidenced by cross-cutting veins, aperture filling, and fragment cementation (Fig. 3b),and dilatant fractures, dissolution textures, as well as the precipitationof new phases (Fig. 3c). The cemented breccias generally have littlefractures (Fig. 3d).

Fifteen hand samples, including host rock, fault breccias andgouge, collected crossing the fault zone at this site, were magneticallystudied by Yang et al. (2012). In the present study, six breccia samplesof them were subjected to further magnetic and mineralogical analy-ses. Samples include shattered (ZJG-4) and crushed breccias (ZJG-6,-7, -8) next to the footwall damage zone, and crushed breccias(ZJG-12, -13) adjacent to the hanging-wall damaged zone.

The diagnostic character of fault breccias is the fragments-embedded-in-matrix texture (Fig. 4). The fragments are surroundedby a matrix of fine-grained comminuted material and clay minerals.The fragments show irregular, angular to slightly rounded, bound-aries, indicating fragment wear, attrition, and fragment rotation.

In the laboratory, fragments and matrix of these breccias wereseparated following the procedure of Kolodny et al. (2005). Sampleswere broken up by hammer, cutter and pliers to sufficientlysmall pieces so that fragments could be separated from matrix by

ed after Verberne et al., 2010). The coseismic surface ruptures induced by the 2008YBF, Yingxiu–Beichuan fault; WMF, Wenchuan–Maoxian fault; GAF, Guanxian–Anxian

Fig. 2. (a) Photograph of Zhaojiagou outcrop of Yingxiu–Beichuan Fault at Leigu Town, Beichuan County. The red line indicates the primary slip plane. (b) Sketch of (a) showing theoccurrence of a shear zone. (c) Close-up view of the slip zone shown in (b), the yellow lines indicate the boundaries of the fault core (modified from Yang et al., 2012).

114 T. Yang et al. / Tectonophysics 601 (2013) 112–124

hand-picking. The separation process was assisted by using sievewith standard mesh apertures. Material that passes through thesieves 0.125 mm in size is hereafter referred to as matrix and theresidue portion is referred to as fragments. It is still assumed that a

Fig. 3. Microstructure of fault rocks from the Zhaojiagou outcrop

certain amount of cross contamination might have occurred. Theweight ratios of matrix and fragments (M/F ratio) are given inTable 1, which generally increase notably toward the principal slipsurface (Fig. 5).

of Yingxiu–Beichuan Fault at Leigu Town, Beichuan County.

Fig. 4. Polished sample section of a typical cemented fault breccia, which is character-ized by the formation of host rock-derived fragments embedded within finer-grainedmatrix.

100

1

2

3

M/F

wei

ght r

atio

10 100

Distance normal to principal slip surface (cm)0

CB G CB SB

Fig. 5. Variation in weight ratio of matrix and fragments (M/F ratio) of fault brecciasagainst the distance normal to the principal slip surface (x = 0). SB—shatteredbreccias, CB—crushed breccias, and G—gouge.

115T. Yang et al. / Tectonophysics 601 (2013) 112–124

3. Methods

The low-field (200 A/m at 976 Hz) magnetic susceptibility (χlf)was measured using an AGICO MFK1-FA Kappabridge magneticsusceptibility meter at a frequency of 976 Hz and a field intensity of200 A/m, respectively. Magnetic hysteresis loops and back-field mag-netizations were measured to determine the hysteresis parameters,coercive force (Bc), remanence coercivity (Bcr), saturation remanence(Mrs) and saturation magnetization (Ms) using a MicroMag™ Model3900 vibrating sample magnetometer (VSM, Princeton Measure-ments Corp.); the maximum applied field was 1.0 T. High-fieldmagnetic susceptibility (χhf) was calculated as the high-field (above0.7 T) slope of each hysteresis loop.

Low-temperature magnetic measurements were performed using aMagnetic Property Measurement System (MPMS XL-7, QuantumDesign), to help for discriminating magnetic minerals. A low-temperature saturation isothermal remanent magnetization (LT-SIRM)was acquired in a 2.5 T magnetic field after cooling the sample fromroom temperature to 10 K in a zero magnetic field. Decay of thisLT-SIRM was monitored at intervals of 3 K below 150 K and 5 K from

Table 1Description and mineralogical compositions of bulk sample, fragments and matrix of fault b

Sample Description Distance normal toprincipal slipsurface (m)

Matrix/fweight

FootwallZJG-1 Sandstone (host rock) 5.00ZJG-4 Bulk Shattered breccia 2.05 0.25

FragmentMatrix

ZJG-6 Bulk Crushed breccia 0.13 0.15FragmentMatrix

ZJG-7 Bulk Crushed breccia 0.07 0.86FragmentMatrix

ZJG-8 Bulk Crushed breccia 0.03 2.69FragmentMatrix

ZJG-9 Black gouge 0

Hanging-wallZJG-12 Bulk Crushed breccia 0.04 0.15

FragmentMatrix

ZJG-13 Bulk Crushed breccia 0.07 0.07FragmentMatrix

ZJG-14 Limestone (host rock) 6.00

150 to 300 K during warming of the sample. Thermomagnetic analysiswas carried out using a MFK1-FA Kappabridge coupled with a CS-4high-temperature furnace operating at a field of 200 A/m and a fre-quency of 976 Hz. Samples were progressively heated to 700 °C underan argon atmosphere at a heating rate of approximately 11 °C/min,and subsequently cooled to room temperature.

Mineralogical composition of samples, including host rock, faultgouge, bulk breccia, fragments and matrix, was determined by X-raydiffraction (XRD). Unoriented, hand-powdered sample was scannedover a 2θ range of 3–70° with CuKα radiation (0.15418 nm), at ascan rate of 8° (2θ)/min with step width of 0.02°, using a Dmax12 kW X-ray power diffractometer (40 kV and 100 mA).

4. Results

4.1. Mineralogical composition

XRD analysis identifies quartz, dolomite, calcite, feldspar and clayminerals as major minerals in samples (Table 1). The matrix hashigher content of quartz, feldspar, dolomite and clay minerals, andlower content of calcite with respect to the fragments. Generally,the content of quartz and dolomite increases and that of calcite andfeldspar decreases toward the principal slip surface, where thecontent of clay minerals also increases remarkably, with peak values

reccias, host rock and gouge samples.

ragments'ratio

Quartz(%)

Feldspar(%)

Dolomite(%)

Calcite(%)

Clay minerals(%)

31 50 – 14 538.0 25.7 1.2 4.1 31.033 37 – 18 1245 22 – 3 3237.2 31.9 – 2.0 28.939 38 – 12 1037 43 2 8 1131.8 16.9 14.9 5.1 31.33 – 49 46 29 – 61 22 89.8 3.1 46.0 16.0 25.1

31 3 37 9 1931 5 33 6 2524.2 0.8 7.4 7.6 60.0

2.5 – 64.4 28.3 4.82 – 92 4 26 – 74 13 60.7 – 51.0 46.2 2.11 – – 94 44 – 2 85 9

b1 – – 98 2

116 T. Yang et al. / Tectonophysics 601 (2013) 112–124

of 60% for the gouge, 19% for the fragments, and 25% for the matrix,respectively (Fig. 6).

4.2. Magnetic susceptibility

Magnetic properties of the host rock, gouge, and six pairs of frag-ments and matrix are compiled in Table 2. It is evident that both of χlfand χhf for fragments are generally comparable with those of bulksamples and host rock. In contrast, the susceptibilities of the matrixsamples are higher, especially for those from the hanging-walldamaged zone (Table 2, Fig. 7a and b). The high positive χhf forbulk samples, fragments and matrix in the footwall damaged

0

10

20

30

40

50

Qua

rtz

(%)

0

10

20

30

40

50

0

10

20

30

40

50

Fel

dspa

r (%

)

0

10

20

30

40

50

40

60

80

100

Dol

omite

(%

)

0

20

40

60

80

100

0

20

40

60

80

100

Cal

cite

(%

)

0

20

40

60

80

100

0

20

40

60

Cla

y m

iner

als

(%)

0

20

40

60

1000 100 10 0 10 100 1000 1000 100 10

Distance normal to pr

HR HRCB G CB SB CB

Bulk Frag

Fig. 6. Variation in mineral composition for bulk sample, fragments and matrix against thbreccias, CB—crushed breccias, and G—gouge.

zone (Table 2 and Fig. 7b) indicates that they are dominated by para-magnetic minerals. In contrast, negative χhf values for those from thehanging-wall damaged zone reveal a strong contribution of diamag-netic minerals, however, matrix of these two breccias has positive χhf.

4.3. Magnetic hysteresis properties

Bulk samples and fragments from the footwall exhibit dominantlyparamagnetic behavior as indicated by χhf, namely, with a uniformslope of magnetization versus magnetizing field (Fig. 8a–d). However,loops for bulk samples and fragments from the hanging-wall dam-aged zone show a strong contribution of diamagnetic components

0

10

20

30

40

50

0

10

20

30

40

50

0

20

40

60

80

100

0

20

40

60

80

100

0

20

40

60

0 10 100 1000 1000 100 10 0 10 100 1000

incipal slip surface (cm)

CB SB CB CB

SB

ments Matrix

e distance normal to the principal slip surface (x = 0). HR—host rock, SB—shattered

Table 2Magnetic parameters of bulk sample, fragments and matrix of fault breccias, typical protolith and fault gouge samples.

Sample χlf (10−8 m3 kg−1) Ms (10−5 Am2 kg−1) Mrs (10−5 Am2 kg−1) Bc (mT) Bcr (mT) χhf (10−8 m3 kg−1) χAT/χ0

ZJG-1 7.94 91.14 7.15 7.21 37.62 3.93 44.7ZJG-4 Bulk 6.06 116.92 10.98 10.98 26.07 5.35 18.9

Fragments 7.26 54.92 3.63 10.02 32.89 4.96 30.2Matrix 9.76 120.46 8.72 8.81 22.69 5.73 15.9

ZJG-6 Bulk 6.04 141.31 11.04 9.88 28.07 6.20 16.0Fragments 8.00 59.15 3.77 7.00 35.38 5.26 22.1Matrix 10.91 354.73 31.23 8.95 21.29 6.73 15.6

ZJG-7 Bulk 7.30 142.33 26.31 12.22 26.98 4.79 16.3Fragments 3.85 68.40 21.23 10.56 47.77 1.34 54.4Matrix 13.33 711.90 74.05 7.83 17.98 3.83 26.0

ZJG-8 Bulk 7.78 192.15 32.89 11.85 30.92 4.85 36.0Fragments 6.39 73.68 10.04 24.65 34.00 4.08 46.0Matrix 15.57 378.43 49.68 9.41 23.20 7.93 30.0

ZJG-9 10.13 198.04 11.09 10.64 24.38 9.53 31.4ZJG-12 Bulk 2.17 85.40 21.97 11.18 40.56 −0.72 51.2

Fragments 3.41 69.42 7.78 13.13 45.24 −0.46 21.0Matrix 39.04 2474.85 377.87 7.80 16.61 2.05 11.8

ZJG-13 Bulk 3.29 174.23 25.99 8.13 23.00 −1.20 36.4Fragments 4.30 98.82 7.51 13.08 47.95 −0.73 14.8Matrix 37.10 2485.71 330.15 7.93 17.44 2.30 6.4

ZJG-14 4.07 62.68 21.66 17.20 42.35 −0.32 8.2

Notes: χlf—low-field magnetic susceptibility; χhf—high-field magnetic susceptibility; Ms—saturation magnetization; Mrs—saturation remanence; Bc—coercive force; Bcr—remanencecoercivity; χAT/χ0 indicates the ratios between the end magnetic susceptibility of the cooling runs (χAT) and their initial values at room temperature (χ0). Except sample ZJG-7 andχAT/χ0 values for the bulk, fragments and matrix of breccia samples, magnetic parameters are compiled from Yang et al. (2012).

117T. Yang et al. / Tectonophysics 601 (2013) 112–124

(Fig. 8e and f). In contrast, matrix shows a slight hysteresis closedbelow 0.1 T (Fig. 8).

Generally, the matrix samples show the most enhanced Ms and Mrs

values (Fig. 7c and d). This is consistent with the susceptibilities behav-ior (Fig. 7a and b), revealing an enhancement of magnetic minerals inthematrix of the breccias. In contrast, Bc and Bcr show opposite tenden-cies. Fragments generally show the highest Bc and Bcr, followed by the

10

χ lf(1

0-8 m

3 kg-1

)

40 (a)

2

-83

-1

100

1000

Ms(1

0-5 A

m2 k

g-1)

3000 (c)

20

5

10

15

20

25

Bc

(mT

)

(e)

Distance normal to principal slip surface (cm)

1000 100 10 0 10 100 1000- --

HR

CB

HRG CB SB

Fig. 7. Comparison of magnetic parameters for host rock, gouge, and bulk sample, matrixbreccias, and G—gouge. Magnetic parameters are compiled from Yang et al. (2012), except

bulk samples, and matrix has relatively lower ones (Table 2, Fig. 7e andf).

4.4. Thermomagnetic analysis

Upon heating, magnetic susceptibility of bulk samples and frag-ments starts to increase above 400 °C, reaches a maximum at about

Bulk

FragmentsMatrix

-2

0

2

4

6

8

10

χ hf(1

0 m

kg

)

(b)

10

100

Mrs

(10-5

Am

2 kg-1

)

400 (d)

2

10

20

30

40

50

Bc

(mT

)

(f)

Distance normal to principal slip surface (cm)

1000 100 10 0 10 100 1000- --

HR CBG

CB SB HR

and fragments of fault breccias. HR—host rock, SB—shattered breccias, CB—crushedsample ZJG-7.

-50

-25

0

25

50

M (

10-3 A

m2 k

g-1)

-50

-25

0

25

50

-50

-25

0

25

50

M (

10-3 A

m2 k

g-1)

-45

-30

-15

0

15

30

45

-42

-21

0

21

42

M (

10-3 A

m2 k

g-1)

-60

-30

0

30

60

-12

-6

0

6

12

-40

-20

0

20

40

-40

-20

0

20

40

M (

10-3 A

m2 k

g-1)

-36

-18

0

18

36

-70

-35

0

35

70

-50

-25

0

25

50Bulk Fragments Matrix

-6

-3

0

3

6

M (

10-3 A

m2 k

g-1)

-4

-2

0

2

4

-42

-21

0

21

42

-1 1

B (T)

-8

-4

0

4

8

M (

10-3 A

m2 k

g-1)

-1 1

B (T)

-6

-3

0

3

6

-1-0.5 0 0.5 -0.5 0 0.5 -0.5 0 0.5 1

B (T)

-48

-24

0

24

48

(a) ZJG-4

(b) ZJG-6

(c) ZJG-7

(d) ZJG-8

(e) ZJG-12

(f) ZJG-13

Fig. 8. Magnetic hysteresis loops for bulk sample, fragments and matrix of fault breccias before para-/diamagnetic correction, where B is the applied magnetic field, and M is themagnetization. The insets are loops between the field of ±0.1 T.

118 T. Yang et al. / Tectonophysics 601 (2013) 112–124

119T. Yang et al. / Tectonophysics 601 (2013) 112–124

500–550 °C, and then decreases to nearly zero at 580–600 °C (Fig. 9),which is consistent with the Curie temperature of magnetite (Dunlopand Özdemir, 1997). However, significant small rises between 250and 300 °C are observed on the heating run of matrix, other than

0

2

4

6

0

2

4

6

0

6

12

18

0

8

16

24

32

0

10

20

30

40

0

15

30

45

60

0

1

2

3

0

6

12

18

0

1

2

3χh (T)/χ0

0

2

4

6

8

10

0

10

20

30

40

0

12

24

36

0

2

4

6

8

0.4

0.6

0.8

1χh (T)/χ0

0

5

10

15

20

χc (T)/χ0

(a) Z

JG-4

Temperature ( oC)

0

10

20

30

40

0

15

30

45

Temperatu

0

5

10

15

0 200 400 600 0 200 4

Bulk Fragme

(b) Z

JG-6

(c) Z

JG-7

(d) Z

JG-8

(e) Z

JG-1

2(f

) ZJG

-13

Fig. 9. Curves of temperature-dependent magnetic susceptibility for bulk sample, fragmentnetic susceptibility at room temperature χ0. The inset shows the heating curve between rooruns, respectively.

the dominant magnetic susceptibility peaks over 400 °C (Fig. 9). Forall samples, the cooling run is well above the heating run, and thefinal magnetic susceptibilities are between 6 and 54 times higherthan their initial values at room temperature (Table 2). Generally,

0

0.4

0.8

1.2χh (T)/χ0

0

6

12

18χc (T)/χ0

0

6

12

18

24

0

14

28

42

56

70

0

4

8

12

16

0

12

24

36

0

1

2

3

4

0

10

20

30

0

1.2

2.4

3.6

0

6

12

18

0

10

20

30

40χc (T)/χ0

0

10

20

30

40

50

0 100 200 3000.8

1.1

1.4

0

6

12

18

24

0

5

10

15

0

5

10

15

20

re ( oC)

0

6

12

18

00 600 0 200 400 600Temperature ( oC)

0

2

4

6

8

0

3

6

9

nts Matrix

s and matrix of fault breccias. Each curve was normalized with its corresponding mag-m temperature and 300 °C. The red and blue lines denote heating (χh) and cooling (χc)

120 T. Yang et al. / Tectonophysics 601 (2013) 112–124

fragments in the footwall damaged zone and bulk breccias in thehanging-wall damaged zone have much higher increase in end mag-netic susceptibility, and the increase in end magnetic susceptibility ofmatrix is relatively lower.

4.5. Low-temperature thermal demagnetization

LT-SIRM demagnetization curves are shown in Fig. 10. Most ofthem display a steeper loss of remanence below 50 K, due to the ther-mal demagnetization of (super)paramagnetic minerals (Moskowitzet al., 1993). A subtle but recognizable drop is indentified near120 K on the demagnetization curves for matrix of breccia samplesZJG-7, ZJG-8, ZJG-12 and ZJG-13 (Fig. 10c–f), which is a weak indica-tion of the Verwey transition suggesting the presence of partiallyoxidized magnetite in the matrix of the breccias (Dunlop andÖzdemir, 1997). However, LT-SIRM demagnetization curves for bulksamples and fragments, and matrix of samples ZJG-4 and ZJG-6(Fig. 10) show no such Verwey transition near 120 K, indicating theabsence of stoichiometric magnetite.

5. Discussion

5.1. Magnetic minerals in fault breccias, fragments and matrix

As evident from the above magnetic measurements, bulk samplesand fragments from the footwall damaged zone are dominated by

0 50 100 150 200 250 3000

0.2

0.4

0.6

0.8

1

1.2

Nor

mal

ized

SIR

M

(a) ZJG-4

0 50 100 150 200 250 300Temperature (K)

0

0.2

0.4

0.6

0.8

1

Nor

mal

ized

SIR

M

(e) ZJG-12

0 50 100 150 200 250 3000

0.2

0.4

0.6

0.8

1

Nor

mal

ized

SIR

M

(c) ZJG-7

Tv

Tv

Tv

Bulk

Matrix

Fragments

Bulk

Matrix

Fragments

Matrix

Fragments

Bulk

Fig. 10. Low-temperature thermal demagnetization of an isothermal remanent magnetizattemperature of the Verwey transition.

paramagnetic components, and those in hanging-wall damagedzone have a strong diamagnetic contribution. In contrary, clearhysteresis closed below 0.1 T for matrix reveals the presence of ferri-magnetic components, which seems to be the most possible reasonfor the much higher χlf and magnetization, and lower coercivity(Table 2 and Fig. 7). It has been proven that comparing the Bcr withthe ratio of Mrs to χlf is one of the most useful exercises to discriminatebetween magnetic minerals (Peters and Dekkers, 2003). In Fig. 11, thesamples fall into three groups, suggesting the differences in theirmagnetic carriers. It is also interesting to note that fragments andhost rock, matrix and gouge occupy the same domains, respectively.

Temperature-dependent (high- and low-temperature) magneticmeasurements would provide further clues of magnetic mineralogy.Thermomagnetic analysis showed that magnetic susceptibility peaksbetween 420–450 °C and 560–580 °C dominate the heating run(Fig. 9). It could be attributed to neoformation of magnetite due tothermal alteration/decomposition of Fe-rich silicates/carbonates(Hoffmann et al., 1999; Zhang et al., 2012). Their paramagnetic natureis demonstrated by hysteresis loops (Fig. 8) and by the significantdecay of magnetization below 50 K on the LT-SIRM demagnetizationcurves (Fig. 10). The newly formed magnetite during the measure-ments may be a non-stoichiometric phase due to substitution ofFe2+ and Fe3+ ions by other elements (Kapička et al., 2001). It alsocan be confirmed by both the drop (between 550 and 590 °C) uponthe heating run and the increase from 590 to 540 °C during thecooling run (Fig. 9).

0 50 100 150 200 250 3000

0.2

0.4

0.6

0.8

1

0 50 100 150 200 250 3000

0.4

0.8

1.2 (b) ZJG-6

(d) ZJG-8

0 50 100 150 200 250 300Temperature (K)

0

0.2

0.4

0.6

0.8

1(f) ZJG-13

Tv

Tv

Tv

Bulk

Matrix

Fragments

Bulk

MatrixFragments

Matrix

Fragments

Bulk

ion imparted at 10 K for bulk sample, fragments and matrix of fault breccias. Tv is the

15 20 25 30 35 40 45 50Bcr (mT)

0

2

4

6

8

10

12M

rs/χ

(kA

/m)

Bulk

FragmentsMatrix

Gouge

Host rock

Fig. 11. Bivariate scatter-plots of remanence coercivity (Bcr) and the ratio of saturationremanence (Mrs) to low-field magnetic susceptibility (χlf).

121T. Yang et al. / Tectonophysics 601 (2013) 112–124

In contrast to bulk samples and fragments, matrix samples showsignificant rise and drop in magnetic susceptibility between 250 and350 °C upon the heating (e.g., Fig. 9c/e/f). This is likely due to thepresence of lepidocrocite (γ-FeOOH), which is a kind of rust in faultzone, rich in H2O and iron ions led by fault fluid flow (Fukuchi et al.,2005). The transformation of lepidocrocite into maghemite usuallystarts at about 250 °C and is completed around 300 °C (Gehring andHofmeister, 1994). In addition, the presence of small amounts of par-tially oxidized magnetite in most of the breccia matrix is revealed bythe weak Verwey transition on the LT-SIRM demagnetization curves(Fig. 10c–f).

Taken together, iron-bearing silicates and carbonates would bemajor magnetic carriers for the bulk breccias and fragments, whileonly very small amounts of ferrimagnetic minerals are present inthe matrix. The latter is confirmed by hysteresis loops (Fig. 8), muchhigher low-field magnetic susceptibility and magnetization, as wellas lower coercivity (Fig. 7 and Table 2). Lepidocrocite is identified in-directly, but only in the matrix samples. Fragments and matrix showsimilar magnetic behaviors in comparison with host rock, bulk brec-cias, and gouge, respectively (Fig. 11), revealing that magnetic prop-erties of breccias are likely predominated by the host rock-derivedfragment. It may be related to its large volume portion in the bulksamples, and further suggest that breccias are of fragmentation origin.

5.2. Magnetic differences and faulting-related effects

As mentioned above, magnetic measurements have revealed signif-icant differences in magnetic properties between breccia matrix andfragments, which are essentially indicative of the different formingmechanisms and/or sources of magnetic minerals, and further linkedto faulting-related effects. At present, mechanisms of magnetic changes

0 4χlf (10-8

0

20

40

60

Cla

y m

iner

als

(%)

2 4 6 8χlf (10-8 m3 kg-1)

0

1

2

3

M/F

rat

io

(b)(a)

Fig. 12. Bivariate scatter-plots of (a) low-field magnetic susceptibility (χlf) vs. weight ratio ofmagnetization (Ms) vs. content of clay minerals.

during fault rock formation are not yet fully understood; however, can-didate faulting-relatedmechanismswhichmight cause suchdifferencesand their possibilities are discussed below.

(1) Stress-induced magnetic changes. The available studies haveproved that the stress/strain could significantly affect rockmagnetic properties in a variety of ways (e.g., Gilder et al.,2004, 2006; Jackson et al., 1993; Kapička et al., 2006; Kean etal., 1976). For example, an experimental deformation on a setof synthetic “calcites sandstone” samples containing magnetiterevealed that ductile shortening strains of up to 25% irrevers-ibly increase their coercivity, and magnetic anisotropy, butdecrease their mean low-field susceptibility and the compo-nent of remanence parallel to shortening (Jackson et al.,1993). Generally, effect of stress/strain on the magnetic min-erals is complicated by many factors, such as magnetic carrier(including magnetic mineralogy, content, grain size, etc.),stress loading pattern and time, and so on (Gilder et al.,2006; Kapička, 1992; Martin, 1980). In the present study, stresswithin the subsurface fault zone is too low to induce significantmagnetic changes, for example, the effective vertical stress σV

and maximum horizontal stress σH at the 590 m fault zone,which was identified as the coseismic slip zone of Wenchuanearthquake within the pilot hole of the Wenchuan EarthquakeFault Scientific Drilling Program (WFSD-1), were estimated as9.25 and 18.5 MPa, respectively (Yang et al., 2012). On theother hand, iron-bearing silicates and carbonates are inferredas major magnetic carriers for the bulk breccias and fragments,and only very small amounts of partially oxidized magnetiteare found in the matrix. In such a case, the stress effects onmagnetic properties of breccias can be ignored, at least in thesubsurface zone.

(2) Crush of pre-existing magnetic grains into finer grains. It is be-lieved that the fault core evolves by grain comminution mostlyconsisting of early bulk fragmentation and late abrasion of grains(Billi, 2005; Hattori and Yamamoto, 1999). In our study, especial-ly for the fault core portion, the host rock is carbonate-rich andthus less resistant to mechanical damage. In such a case, it isconceivable for pervasive shearing to cause not only enrichmentof matrix, but also elevation of magnetic susceptibility, as thewidespread splitting of pre-existing ferrimagnetic grains intomuch finer size (Dearing, 1999). It is found that χlf significantlycorrelates with the weight ratio of matrix and fragments(Fig. 12a), which increases notably approaching the principalslip surface (Fig. 5). We currently lack evidence to decide towhat extent such a process is involved in or responsible for thisfinding, but consider this a likely possibility.

(3) Frictional heating induced thermochemical reactions. Frictionalheating, which constitutes the largest part of the total seismic en-ergy budget, is frequently generated by shear friction with a highslip rate during a large earthquake (Scholz, 2002). It will quickly

Ms (10-5 Am2 kg-1)

0

20

40

60

Cla

y m

iner

als

(%)

8 12

m3 kg-1)

50 100 150 200

(c)

matrix and fragments (M/F ratio), (b) χlf vs. content of clay minerals, and (c) saturation

-30 -20 -10 0 10 20 30Distance to principal slip surface (cm)

0

200

400

600

800

1000Te

mpe

ratu

re (

o C)

μ=0.05

μ=0.10

μ=0.15

Fig. 13. Evolution of the maximum temperature of frictional heat with the distancefrom the co-seismic slip plane (x = 0) at 590 m fault zone during the 2008 Wenchuanearthquake at various frictional coefficients reconstructed by a one-dimensional modelusing diffusion equations (Cardwell et al., 1978).

122 T. Yang et al. / Tectonophysics 601 (2013) 112–124

elevate the temperature within the fault zone (Rice, 2006), andinduces thermochemical reactions within the fault zone, such asmineral decomposition of clay minerals in fault rocks. Duringthis process, the newly formed ferrimagnetic minerals (e.g., mag-netite, maghemite and pyrrhotite) would increase its magnetiza-tion by a few orders of magnitude (Fukuchi et al., 2005; Han etal., 2007; Mishima et al., 2006; Tanikawa et al., 2008). Here the590 m fault zone is chosen as an example to examine the thermalhistory, including frictional heat generation and conduction, onthe basis of one-dimensional thermal diffusion model with a fi-nite fault thickness (Cardwell et al., 1978). The related calculationassumptions and parameters followed those in Yang et al.(2012). The model showed that frictional heating is mostlylocalized within coseismic slip zone, and the temperature decaysignificantly with the distance to the principal slip surface(Fig. 13). Therefore, the frictional heating is unlikely to act onfault breccias, or only on those close to the principal slip zone,to a little extent.

(4) Chemical alteration and neomineralization due to fluid move-ment within the fault zone. Breccia zone is generally consideredto be highly permeable because of fracturing and subsidiaryfaulting, and to serve as conduits for fluid flow (Vermilye andScholz, 1998). Short-term fluids promoted by earthquakes(Muir-Wood, 1994) and meteoric fluids during interseismicperiods generally infiltrate and percolate along dense micro- andmacro-cracks into the fault zone. The permeability (at 40 MPaeffective pressure) for the studied fault rocks is from 10−13 to10−17 m2 for different breccias, much higher than those forfresh gouges and country rocks (10−18 m2 to b10−19 m2)(Chen et al., 2011). Thus, when the fluids percolate in the faultzone, the fault parallel flow would be confined within the brecciazone enveloped by the impermeable fault gouge and host rock(Evans et al., 1997). As the fault core is carbonate-rich, they canbe easily disaggregated, dissolved, and transported away by fluidsduring long-term inter-seismic periods. The occurrence of widelyfluid movements at the present outcrop can be evidenced by thewidespread precipitation of the fluid-related mineral phase dolo-mite (Fig. 6, Table 1), the different healing textures in microscopicphotos (Fig. 3), as well as the distinct differences in stable isotopecompositions (δ13C and δ18O) between fragments, matrix, andvein material (Chen et al., 2013). Breccia matrix is more suscepti-ble to dissolution and chemical reactions due to its fine-grainedsize and thus large specific surface. Destabilization of iron-

bearing clay minerals in them by fault fluids will release Fe thateasily enters the fluid and is transported by it (Blumstein et al.,2004; Katz et al., 2000). The Fe-riched fluids are potential sourcesof higher crystallization of magnetic minerals (Pechersky andGenshaft, 2001 and references therein). Such a circulatingfluid may precipitate fine-grained iron-oxides (e.g., magnetite)or -hydroxides (e.g., goethite and lepidocrocite) under differentredox and pH conditions (Goldstein and Brown, 1988; Grosz etal., 2006; Guichet et al., 2006). For example, the presence of abun-dant magnetite in the matrix of breccia was reported from theWernecke Mountains, Yukon, Canada, due to the fluid precipita-tion (Hunt et al., 2005). Chou et al. (2012) found neoformedgoethite and partially altered magnetite in the fault zone ofthe Chi-Chi earthquake (Mw 7.6, 1999), both are related to co-seismic fluid. Although there is no such magnetic mineral identi-fied directly by XRD analysis, the matrix shows relatively highermagnetization, and its χlf and Ms significantly correlate with thecontent of clay minerals (Fig. 12b and c), which is one of themost important products of the fluid infiltration (Chen et al.,2013). It suggests that the enrichment of magnetic mineral inthe matrix is also related the formation of clay minerals, and fur-ther link to the fault fluid movements. With respect to the matrix,fragments generally survive or to a lesser extent affected by thefluid alteration effects, due to the much coarser grain and thuslower specific surface.

In summary, it is possible to propose a simple conceptual modelassociating the significant magnetic differences in the fragmentsand matrix of breccias. During coseismic slip events, intense shearmay crush the pre-existing magnetic grains in fault rocks into finergrains, meanwhile, stress/strain and frictional heating may affectmagnetic properties of fault breccias to a little extent. Duringinter-seismic periods, meteoric fluids infiltrate and percolate intofault zone and would cause dissolution, precipitation and recrystalli-zation of Fe-bearing minerals. Due to the different permeability struc-ture and grain size of breccia fragments and matrix, these alterationeffects would significantly modify magnetic mineralogy in matrix, toa much lesser extent in fragments. Consequently, here the magneticdifferences in the fragments and matrix may reflect a cumulative na-ture taking place over many previous earthquake cycles that occurredalong the LMS thrust belt zone (Zhang et al., 2010). Although theseco-/inter-seismic effects have to be well understood by furtherworks, this model predicts that similar effects should occur in otheractive fault zones, too. In addition, these observations suggest thatmagnetic study of fault breccias would provide clues to help under-stand seismic faulting and history of fault activity.

6. Conclusions

Rock magnetic and mineralogical analyses were conducted onpairs of fragments and matrix of fault breccias from the Zhaojiagououtcrop on the Yingxiu–Beichuan Fault, ruptured during the 2008Wenchuan Mw 7.9 earthquake (China). Results showed that magnet-ic behavior of the bulk samples and fragments is dominated by dia-/para-magnetic components, such as Fe-bearing silicates and carbon-ates, whereas minor amounts of partially oxidized magnetite andlepidocrocite are identified in matrix with higher low-field magneticsusceptibility and magnetization, and lower coercivity. Meanwhile,the matrix is generally enriched in dolomite, feldspar and clay min-erals, but depleted in calcite with respect to the fragments. These ob-servations are proposed to be associated with faulting-related effects.During coseismic slip, intense shear crushes pre-existing magneticgrains in fault rocks into finer grains; meanwhile the stress andfrictional heat, to a little extent, induce magnetic changes and ther-mochemical alterations of Fe-bearing minerals in fault breccias,respectively. During interseismic periods, meteoric fluids would

123T. Yang et al. / Tectonophysics 601 (2013) 112–124

infiltrate and percolate into fault zone and cause dissolution, precipi-tation and recrystallization of Fe-bearing minerals. As the differencesin permeability structure and grain size of fragments and matrix, thefluid movements would modify their magnetic mineralogy at variouslevels. These effects, especially the fluid infiltration, taking place overmany previous earthquake cycles, would be the most likely causes forthe observed different magnetic properties in fragments and matrix.These results suggest that magnetic studies of fault breccias wouldprovide clues to help understand seismic faulting and history offault activity.

Acknowledgments

This study was supported by the Wenchuan Earthquake FaultScientific Drilling Program (WFSD 0009), the State Key Laboratoryof Earthquake Dynamics, Institute of Geology, China EarthquakeAdministration, under grant LED2010A03, and the National NaturalScience Foundation of China (No. 41204062). The authors thankH.Y. Zhang at the Department of Physics, Tsinghua University forher generous help in the low-temperature magnetic measurements.Thanks are extended to J.X. Dang at IGCEA for his invaluable assis-tance during the fieldwork. We would like to thank Dr. Simo Spassovand an anonymous reviewer for their constructive comments on thismanuscript.

References

Agosta, F., Aydin, A., 2006. Architecture and deformation mechanism of a basinbounding normal fault in Mesozoic platform carbonates, central Italy. Journal ofStructural Geology 28, 1445–1467.

Billi, A., 2005. Grain size distribution and thickness of breccia and gouge zones fromthin (b1 m) strike-slip fault cores in limestone. Journal of Structural Geology 27(10), 1823–1837.

Blumstein, A.M., Elmore, R.D., Engel, M.H., Elliot, C., Basu, A., 2004. Paleomagnetic dat-ing of burial diagenesis in Mississippian carbonates, Utah. Journal of GeophysicalResearch 109, B04101. http://dx.doi.org/10.1029/2003JB002698.

Caine, J.S., Evans, J.P., Forster, C.B., 1996. Fault zone architecture and permeabilitystructure. Geology 24, 1025–1028.

Cardwell, R.K., Chinn, D.S., Moore, G.F., Turcotte, D.L., 1978. Frictional heating on a faultzone with finite thickness. Geophysical Journal of the Royal Astronomical Society52 (3), 525–530.

Chen, W.D., Tanaka, H., Huang, H., Lu, C., Lee, C., Wang, C., 2007. Fluid infiltration asso-ciated with seismic faulting: examining chemical and mineralogical compositionsof fault rocks from the active Chelungpu fault. Tectonophysics 443, 243–254.

Chen, J.Y., Yang, X.S., Dang, J.X., He, C.R., Zhou, Y.S., Ma, S.L., 2011. Internal structure andpermeability of Wenchuan earthquake fault. Chinese Journal of Geophysics 54 (7),1805–1816 (in Chinese with English abstract).

Chen, J.Y., Yang, X.S., Ma, S.L., Spiers, C.J., 2013. Mass removal and clay mineral dehy-dration/rehydration in carbonate-rich surface exposures of the 2008 Wenchuanearthquake fault: geochemical evidence and implications for fault zone evolutionand coseismic slip. Journal of Geophysical Research — Solid Earth 118. http://dx.doi.org/10.1002/jgrb.50089.

Chou, Y.M., Song, S.R., Aubourg, C., Song, Y.F., Boullier, A.M., Lee, T.Q., Evans, M., Yeh,E.C., Chen, Y.M., 2012. Pyrite alteration and neoformed magnetic minerals in thefault zone of the Chi-Chi earthquake (Mw 7.6, 1999): evidence for frictionalheating and co-seismic fluids. Geochemistry, Geophysics, Geosystems 13,Q08002. http://dx.doi.org/10.1029/2012GC004120.

Dearing, J., 1999. Magnetic susceptibility. In: Walden, J., Oldfield, F., Smith, J. (Eds.),Environmental Magnetism: A Practical Guide. : Technical Guide, No. 6. QuaternaryResearch Association, London, pp. 35–62.

Dunlop, D.J., Özdemir, Ö., 1997. Rock Magnetism: Fundamentals and Frontiers.Cambridge University Press, Cambridge.

Evans, J.P., Forster, C.B., Goddard, J.V., 1997. Permeability of fault-related rocks, andimplications for hydraulic structure of fault zones. Journal of Structural Geology19, 1391–1404.

Ferré, E.C., Zechmeister, M.S., Geissman, J.W., MathanaSekaran, N., Kocak, K., 2005. Theorigin of high magnetic remanence in fault pseudotachylites: theoretical consider-ations and implication for coseismic electrical currents. Tectonophysics 402 (1–4),125–139.

Fukuchi, T., 2003. Strong ferrimagnetic resonance signal and magnetic susceptibility ofthe Nojima pseudotachylyte in Japan and their implication for coseismic electromag-netic changes. Journal of Geophysical Research 108 (B6), 2312. http://dx.doi.org/10.1029/2002JB002007.

Fukuchi, T., Mizoguchi, K., Shimamoto, T., 2005. Ferrimagnetic resonance signal producedby frictional heating: a new indicator of paleoseismicity. Journal of GeophysicalResearch 110, B12404. http://dx.doi.org/10.1029/2004JB003485.

Gehring, A.U., Hofmeister, A.M., 1994. The transformation of lepidocrocite duringheating: a magnetic and spectroscopic study. Clays and Clay Minerals 42, 409–415.

Gilder, S.A., LeGoff, M., Chervin, J.-C., Peyronneau, J., 2004. Magnetic properties of singleand multi-domain magnetite under pressures from 0 to 6 GPa. GeophysicalResearch Letters 31, L10612. http://dx.doi.org/10.1029/2004GL019844.

Gilder, S.A., Le Goff, M., Chervin, J.-C., 2006. Static stress demagnetization of single andmultidomain magnetite with implications for meteorite impacts. High PressureResearch 26 (4), 539–547.

Goldstein, A.G., Brown, L.L., 1988. Magnetic susceptibility anisotropy of mylonites fromthe Brevard Zone, North Carolina, U.S.A. Physics of the Earth and PlanetaryInteriors 51, 290–300.

Grosz, S., Matthews, A., Ilani, S., Ayalon, A., Garfunkel, Z., 2006. Iron mineralization anddolomitization in the Paran Fault zone, Israel: implications for low-temperaturebasinal fluid processes near the Dead Sea Transform. Geofluids 6, 137–153.

Guichet, X., Jouniaux, L., Catel, N., 2006. Modification of streaming potential by precipita-tion of calcite in a sand–water system: laboratory measurements in the pH rangefrom 4 to 12. Geophysical Journal International 166, 445–460.

Hailwood, E.A., Maddock, R.H., Fung, T., Rutter, E.H., 1992. Paleomagnetic analysis offault gouge and dating fault movement, Anglesey, North Wales. Journal of theGeological Society of London 149, 273–284.

Han, R., Shimamoto, T., Ando, J., Ree, J.H., 2007. Seismic slip record in carbonate-bearingfault zones: an insight from high-velocity friction experiments on siderite gouge.Geology 35 (12), 1131–1134.

Hattori, I., Yamamoto, H., 1999. Rock fragmentation and particle size in crushed zonesby faulting. Journal of Geology 107, 209–222.

Hausegger, S., Kurz, W., Rabitsch, R., Kiechl, E., Brosch, F.-J., 2010. Analysis of the inter-nal structure of a carbonate damage zone: implications for the mechanisms of faultbreccia formation and fluid flow. Journal of Structural Geology 32, 1349–1362.

Hirono, T., Lin, W., Yeh, E.C., et al., 2006. High magnetic susceptibility of fault gougewithin Taiwan Chelungpu fault: nondestructive continuous measurements ofphysical and chemical properties in fault rocks recovered from Hole B, TCDP. Geo-physical Research Letters 33, L15303. http://dx.doi.org/10.1029/2006GL026133.

Hoffmann, V., Knab, M., Appel, E., 1999. Magnetic susceptibility mapping of roadsidepollution. Journal of Geochemical Exploration 66, 313–326.

Holland, M., Urai, J.L., van der Zee, W., Stanjek, H., Konstanty, J., 2006. Fault gouge evo-lution in highly overconsolidated claystones. Journal of Structural Geology 28,323–332.

Hunt, J., Baker, T., Thorkelson, D., 2005. Regional-scale Proterozoic IOCG-mineralized brec-cia systems: examples from the Wernecke Mountains, Yukon, Canada. MineraliumDeposita 40 (5), 492–514.

Isaacs, A.J., Evans, J.P., Song, S.-R., Kolesar, P.T., 2007. Structural, mineralogical, andgeochemical characterization of the Chelungpu Thrust Fault, Taiwan. Terrestrial,Atmospheric and Oceanic Sciences 18 (2), 183–221.

Jackson, M., Borradaile, G., Hudleston, P., Banerjee, S., 1993. Experimental deformationof synthetic magnetite-bearing calcite sandstones: effects on remanence, bulkmagnetic properties, and magnetic anisotropy. Journal of Geophysical Research98 (B1), 383–401.

Kapička, A., 1992. Magnetic susceptibility under hydrostatic pressure of syntheticmagnetite samples. Physics of the Earth and Planetary Interiors 70, 248–252.

Kapička, A., Jordanova, N., Petrovský, E., Ustjak, S., 2001. Effect of different soilconditions on magnetic parameters of power-plant fly ashes. Journal of AppliedGeophysics 48, 93–102.

Kapička, A., Hrouda, F., Petrovský, E., Poláčekc, J., 2006. Effect of plastic deformation inlaboratory conditions on magnetic anisotropy of sedimentary rocks. High PressureResearch 26 (4), 549–553.

Katz, B., Elmore, R.D., Cogoini, M., Engel, M.H., Ferry, S., 2000. Associations between burialdiagenesis of smectite, chemical remagnetization, and magnetite authigenesis in theVocontian trough, SE France. Journal of Geophysical Research 105 (B1), 851–868.

Kean, W.F., Day, R., Fuller, M., Schmidt, V.A., 1976. The effect of uniaxial compression onthe initial susceptibility of rocks as a function of grain size and composition of theirconstituent titanomagnetites. Journal of Geophysical Research 81, 861–872.

Kolodny, Y., Chaussidon, M., Katz, A., 2005. Geochemistry of a chert breccia.Geochimica et Cosmochimica Acta 69 (2), 427–439.

Lin, A., Ren, Z., Kumahara, Y., 2010. Structural analysis of the coseismic shear zone ofthe 2008 Mw 7.9 Wenchuan earthquake, China. Journal of Structural Geology 32(6), 781–791.

Liu-Zeng, J., Zhang, Z., Wen, L., Tapponnier, P., Sun, J., Xing, X., Hu, G., Xu, Q., Zeng, L.,Ding, L., Ji, C., Hudnut, K.W., van der Woerd, J., 2009. Co-seismic ruptures of the12 May 2008, Ms 8.0Wenchuan earthquake, Sichuan: east–west crustal shorteningon oblique, parallel thrusts along the eastern edge of Tibet. Earth and PlanetaryScience Letters 286 (3–4), 355–370.

Martin III, R.J., 1980. Is piezomagnetism influenced by microcracks during cyclicloading? Journal of Geomagnetism and Geoelectricity 32, 741–755.

Matsuda, T., Omura, K., Ikeda, R., Arai, T., Kobayashi, K., Shimada, K., Tanaka, H., Tomita,T., Hirano, S., 2004. Fracture-zone conditions on a recently active fault: insightsfrom mineralogical and geochemical analyses of the Hirabayashi NIED drill coreon the Nojima fault, southwest Japan, which ruptured in the 1995 Kobe earth-quake. Tectonophysics 378, 143–163.

Mishima, T., Hirono, T., Soh, W., Song, S.-R., 2006. Thermal history estimation of theTaiwan Chelungpu fault using rock-magnetic methods. Geophysical ResearchLetters 33, L23311. http://dx.doi.org/10.1029/2006GL028088.

Moskowitz, B.M., Frankel, R.B., Bazylinski, D.A., 1993. Rock magnetic criteria for the detec-tion of biogenic magnetite. Earth and Planetary Science Letters 120, 283–300.

Muir-Wood, R., 1994. Earthquakes, strain-cycling and themobilization of fluids. GeologicalSociety, London, Special Publications 78, 85–98.

Pechersky, D.M., Genshaft, Yu.S., 2001. Petromagnetism of the continental lithosphereand the origin of regional magnetic anomalies: a review. Russian Journal of EarthSciences 3 (2), 97–124.

124 T. Yang et al. / Tectonophysics 601 (2013) 112–124

Peters, C., Dekkers, M.J., 2003. Selected room temperature magnetic parameters as afunction of mineralogy, concentration and grain size. Physics and Chemistry ofthe Earth 28, 659–667.

Rice, J.R., 2006. Heating and weakening of faults during earthquake slip. Journal ofGeophysical Research 111, B05311. http://dx.doi.org/10.1029/2005JB004006.

Scholz, C.H., 2002. The Mechanisms of Earthquake Faulting. Cambridge UniversityPress, New York (496 pp.).

Schulz, S.E., Evans, J.P., 1998. Spatial variability in microscopic deformation and compo-sition of the Punchbowl fault, southern California: implications for mechanisms,fluid-rock interaction, and fault morphology. Tectonophysics 295 (1–2), 223–244.

Sibson, R.H., 1977. Fault rocks and fault mechanisms. Journal of the Geological Societyof London 133, 191–213.

Storti, F., Billi, A., Salvini, F., 2003. Particle size distributions in natural carbonate faultrocks: insights for non-self-similar cataclasis. Earth and Planetary Science Letters206, 173–186.

Tanaka, H., Fujimoto, K., Ohtani, T., Ito, H., 2001. Structural and chemical characteriza-tion of shear zones in the freshly activated Nojima fault, Awaji Island, southwestJapan. Journal of Geophysical Research 106 (B5), 8789–8810.

Tanikawa, W., Mishima, T., Hirono, T., Soh, W., Song, S.-R., 2008. High magnetic suscep-tibility produced by thermal decomposition of core samples from the Chelungpufault in Taiwan. Earth and Planetary Science Letters 272 (1–2), 372–381.

Verberne, B.A., He, C.R., Spiers, C.J., 2010. Frictional properties of sedimentary rocks andnatural fault gouge from the Longmen Shan fault zone, Sichuan, China. Bulletin ofthe Seismological Society of America 100 (5B), 2767–2790.

Vermilye, J.M., Scholz, C.H., 1998. The process zone: a microstructural view of faultgrowth. Journal of Geophysical Research 103 (B6), 12223–12237.

Wilson, B., Dewers, T., Reches, Z., Brune, J., 2005. Particle size and energetics of gougefrom earthquake rupture zones. Nature 434, 749–752.

Woodcock, N.H., Mort, K., 2008. Classification of fault breccias and related fault rocks.Geological Magazine 145 (3), 435–440.

Xu, Z.Q., Ji, S.C., Li, H.B., Hou, L.W., Fu, X.F., Cai, Z.H., 2008. Uplift of the Longmen Shanrange and the Wenchuan earthquake. Episodes 31 (3), 291–301.

Yang, T., Chen, J.Y., Wang, H.Q., Jin, H.Q., 2012. Rock magnetic properties of fault rocksfrom the rupture of the 2008 Wenchuan earthquake, China and their implications:preliminary results from the Zhaojiagou outcrop, Beichuan County (Sichuan).Tectonophysics 530–531, 331–341.

Zhang, P.Z., Wen, X.Z., Shen, Z.K., Chen, J.H., 2010. Oblique, high-angle, listric-reversefaulting and associated development of strain: the Wenchuan earthquake of May 12,2008, Sichuan, China. Annual Review of Earth and Planetary Sciences 38, 353–382.

Zhang, C.X., Paterson, G.A., Liu, Q.S., 2012. A newmechanism for the magnetic enhance-ment of hematite during heating: the role of clay minerals. Studia Geophysica etGeodaetica 56 (3), 845–860.