Growth and linkage of a complex oblique-slip fault zone in the Pearl River Mouth Basin, northern South China Sea

Growth and linkage of a complex oblique-slip fault zone in the Pearl River Mouth Basin, northern South China Sea

Accepted Manuscript Growth and linkage of a complex oblique-slip fault zone in the Pearl River Mouth Basin, northern South China Sea Ke Huang, Guangfa...

86MB Sizes 0 Downloads 17 Views

Accepted Manuscript Growth and linkage of a complex oblique-slip fault zone in the Pearl River Mouth Basin, northern South China Sea Ke Huang, Guangfa Zhong, Min He, Lihua Liu, Zhe Wu, Xuefeng Liu PII:

S0191-8141(18)30442-5

DOI:

10.1016/j.jsg.2018.09.002

Reference:

SG 3734

To appear in:

Journal of Structural Geology

Received Date: 29 November 2017 Revised Date:

1 September 2018

Accepted Date: 4 September 2018

Please cite this article as: Huang, K., Zhong, G., He, M., Liu, L., Wu, Z., Liu, X., Growth and linkage of a complex oblique-slip fault zone in the Pearl River Mouth Basin, northern South China Sea, Journal of Structural Geology (2018), doi: 10.1016/j.jsg.2018.09.002. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT 1

Growth and linkage of a complex oblique-slip fault zone in the Pearl River

2

Mouth Basin, northern South China Sea

3

Ke Huanga, Guangfa Zhonga, *, Min Hea, b, Lihua Liub, Zhe Wub, Xuefeng Liuc

RI PT

4 5 6

a

7

Shanghai 200092, China

8

b

9

Road, Shenzhen 518054, China

SC

State Key Laboratory of Marine Geology, Tongji University, 1239 Siping Road,

M AN U

Research Institute of Shenzhen Branch, CNOOC China Limited, 3168 Houhaibin

10

c

11

Shangda Road, Shanghai 200444, China.

School of Communication and Information Engineering, Shanghai University, 99

13

TE D

12

* Corresponding author. Tel.: +86 21 65982784. E-mail address: [email protected]

14

E-mail addresses: [email protected] (K. Huang), [email protected] (G.

16

Zhong),

17

[email protected] (Z. Wu), [email protected] (X. Liu).

19

[email protected]

(M.

He),

[email protected]

(L.

Liu),

AC C

18

EP

15

20

Key words: Oblique-slip normal fault, fault growth pattern, strike and dip linkage,

21

Three-dimensional seismic, Pearl River Mouth Basin, South China Sea

22

ACCEPTED MANUSCRIPT 23

ABSTRACT

24

Oblique-slip faults are ubiquitous in sedimentary basins, but have been largely

26

neglected relative to their dip-slip counterpart. Case studies of oblique slip faults are

27

very limited. In this study, we used 3D seismic data, tied to well control, to investigate

28

the geometry, kinematic characteristics, growth and linkage, and tempo-spatial

29

evolution of a complex oblique-slip growth fault zone in the Pearl River Mouth Basin

30

(PRMB) of the northern continental margin, South China Sea (SCS). The

31

E-W-oriented 30.5-km-long fault zone consists of two superimposed oblique-slip fault

32

systems. The lower fault system is an integrated zig-zag fault consisting of three

33

ENE-WSW- to E-W-striking, right-stepping, sinistral en-echelon fault segments,

34

which were hard-linked by two NW-SE-trending fault segments; while the upper fault

35

system is composed of nine WNW-ESE-striking, left-stepping, dextral en-echelon

36

component faults, which had overlapping fault tips and were partly hard-linked by

37

breaching of the relay zones. The lower and upper fault systems were activated in the

38

Eocene and the Oligocene to Pliocene, which respectively correspond to the syn-rift

39

and post-rift stages in the PRMB. The fault throws (up to 288 m) and activity rates

40

(up to 34.5 m/Ma) of the upper fault system are overall much less than those of its

41

lower counterpart. Reactivation of the lower syn-rift fault system shows clear

42

influence on the development of post-rift en-echelon fault segments. A four-stage

43

evolution model of the oblique-slip fault zone was presented.

44

AC C

EP

TE D

M AN U

SC

RI PT

25

ACCEPTED MANUSCRIPT 1. Introduction

46

Accumulated evidence indicates that most of the faults in rifted basins are difficult to

47

describe by using a simple dip-slip fault model (Crider, 2001; Morley et al., 2004,

48

2007; Morley, 2007; Whipp et al., 2014; Duffy et al., 2015; Henstra et al., 2015;

49

Reeve et al., 2015; Morley, 2016). In rift basins, oblique slip faults should be

50

ubiquitous, but have been largely ignored in literature relative to their dip-slip

51

counterpart. It is suggested that the formation of oblique-slip faults in rifted basins

52

could be associated with reactivation of pre-existing faults or crustal weakness zones

53

trending oblique to the regional extension direction (Bellahsen and Daniel, 2005;

54

Henza et al., 2010, 2011; Chattopadhyay and Chakra, 2013). Variations in the

55

direction of regional extension may provide favorable conditions for the development

56

of oblique-slip fault systems in rifted basins (Bonini et al., 1997; Clifton et al., 2000;

57

Morley et al., 2004, 2007; Morley, 2007; Tingay et al., 2010; Paredes et al., 2013;

58

Whipp et al., 2014; Morley, 2016).

59

Numerous sand-box modeling experiments have been carried out to investigate the

60

origin of oblique-slip faults in extensional settings (Withjack and Jamison, 1986; Tron

61

and Brun, 1991; McClay and White, 1995; Bonini et al., 1997; Higgins and Harris,

62

1997; Keep and McClay, 1997; Clifton et al., 2000; McClay et al., 2002; Schlische et

63

al., 2002; Henza et al., 2010, 2011). Results indicate that formation of an oblique-slip

64

fault is controlled simultaneously by both the extension displacement perpendicular to

65

and the shear displacement parallel to the trend of the host rift. The relative amounts

66

of extension and shear components depend on the acute angle, α, between the rift

AC C

EP

TE D

M AN U

SC

RI PT

45

ACCEPTED MANUSCRIPT trend and the relative displacement direction on opposite sides of the rift (Withjack

68

and Jamison, 1986). Depending on the angle (α), oblique-slip faults may be dip-slip or

69

strike-slip dominated, that is, fault segments that lie at a relatively low angle to the

70

extensional direction will display smaller dip-slip and larger strike-slip displacement

71

components (Morley et al., 2004).

72

Research on fault growth and linkage is a hot topic in structural geology during the

73

past decades (Cartwright et al., 1995; Morley, 1999; Crider, 2001; Peacock, 2002;

74

Walsh et al., 2003; Kim et al., 2005; Baudon and Cartwright, 2008a, 2008b, 2008c;

75

Frankowicz and McClay, 2010; Giba et al., 2012; Jackson and Rotevatn, 2013;

76

Paredes et al., 2013; Tvedt et al., 2013; Wilson et al., 2013; Fazli Khani and Back,

77

2015; Robson et al., 2016, 2017). Available references, however, have been mostly

78

focused on dip-slip rather than oblique-slip normal faults (Cartwright et al., 1995;

79

Morley, 1999; Peacock, 2002; Walsh et al., 2003; Kim et al., 2005; Baudon and

80

Cartwright, 2008a, 2008b, 2008c; Paredes et al., 2013; Tvedt et al., 2013; Wilson et

81

al., 2013; Fazli Khani and Back, 2015; Robson et al., 2016, 2017). It has been

82

confirmed that normal faults are mostly segmented in both map view and section view

83

(Cartwright et al., 1995; Childs et al., 1995; Peacock, 2002; Walsh et al., 2003;

84

Frankowicz and McClay, 2010). Growth of a segmented normal fault can be

85

explained by two models, the isolated fault model (Walsh and Watterson, 1988;

86

Cartwright et al., 1995; Walsh et al., 2003; Giba et al., 2012; Jackson and Rotevatn,

87

2013; Jackson et al., 2016) and the constant length model (Morley, 1999; Walsh et al.,

88

2002, 2003; Giba et al., 2012; Jackson and Rotevatn, 2013; Jackson et al., 2016; Nicol

AC C

EP

TE D

M AN U

SC

RI PT

67

ACCEPTED MANUSCRIPT et al., 2016). According to the isolated fault model, fault segments in a fault zone tend

90

to behave independently, and grow by a simultaneous increase in both displacement

91

and length (Walsh and Watterson, 1988; Cartwright et al., 1995; Walsh et al., 2003;

92

Giba et al., 2012; Jackson and Rotevatn, 2013; Jackson et al., 2016). A isolated fault is

93

characterized by decrease or deficit of summed displacement in the overlap of two

94

adjacent fault segments (Walsh et al., 2002). In contrast, fault segments comprising a

95

constant length fault zone tend to rapidly attain their near-final lengths in the early of

96

their slip history, with fault displacement mostly built after their fault length has been

97

established (Morley, 1999; Walsh et al., 2002). Hence a constant length fault tends to

98

be under-displaced for its length during much of its life. Low or no displacement

99

deficits can be recognized in the areas of overlapping faults (Morley, 1999; Walsh et

100

al., 2002, 2003; Giba et al., 2012; Jackson and Rotevatn, 2013; Jackson et al., 2016;

101

Nicol et al., 2016). In addition, fault linkage can occur through soft or hard linkage

102

(Peacock, 2002; Kim et al., 2005). Soft linkage describes faults that are only

103

kinetically linked, while hard linkage describes fault segments that are physically

104

linked by e.g. breaching of the relay zones (Peacock, 2002; Kim et al., 2005).

105

In this study, we applied 3D seismic data to investigate the characteristics, growth and

106

linkage, and tempo-spatial evolution of a complex oblique-slip fault zone in the Pearl

107

River Mouth Basin (PRMB), northern South China Sea (SCS) margin. The fault zone

108

consists of two superimposed en-echelon oblique-slip fault systems, which developed

109

in the syn-rift and post-rift stages, respectively. Both the syn-rift and post-rift

110

oblique-slip fault systems contain several fault segments showing different step and

AC C

EP

TE D

M AN U

SC

RI PT

89

ACCEPTED MANUSCRIPT slip senses, suggesting that the fault systems formed under different tectonic stress

112

environments. Moreover, the syn-rift oblique-slip fault system formed earlier shows

113

significant control on the development of the post-rift fault system. Our research has

114

implications for understanding the complexity and diversity of faults in rifted basins,

115

and further for understanding the geological evolution of the basins, in view of the

116

pervasiveness of oblique-slip faults in rifted basins and the scarcity of related case

117

studies in literature. Application of oblique-slip fault model into the exploration

118

practice will undoubtedly broaden the horizon of searching for petroleum traps and

119

reservoirs and improve the success rate of petroleum exploration in the rifted basins.

120

In addition, as far as our knowledge, no literature has addressed the issue of

121

oblique-slip faults in the PRMB, therefore our research is of significance for more

122

accurately addressing the regional tectonic evolution of the basin.

123

2. Geologic background

124

The PRMB is located in the middle of the northern SCS continental margin and

125

comprises the largest Cenozoic hydrocarbon-bearing sedimentary basin in the margin.

126

The basin trends in a ENE-WSW direction parallel to its host continental margin and

127

is subdivided into five secondary tectonic units of the same strike, which from

128

north-northwest to south-southeast are the northern uplift, northern depression

129

(including the Zhu-I and Zhu-III depressions), central uplift (consisting of the Shenhu,

130

Panyu, and Dongsha uplifts), southern depression (consisting of the Zhu-II and

131

Chaoshan depressions), and southern uplift zones (Fig. 1) (Li, 1993; Chen et al., 2003;

AC C

EP

TE D

M AN U

SC

RI PT

111

ACCEPTED MANUSCRIPT Shi et al., 2014).

133

The Zhu-I depression occupies the middle to east part of the northern depression zone.

134

It consists of five sub-tectonic units, named ‘sags’ by local geologists, which from

135

southwest to northeast are the Enping, Xijiang, Huizhou, Lufeng and Hanjiang sags

136

(Fig. 1A). Each sag contains several grabens or half grabens separated by low-relief

137

tectonic highs.

138

The fault zone investigated here, named the EP18 fault zone, lies in the northeastern

139

Enping sag and comprises the northern boundary of a Paleogene half graben to the

140

south (EP 18 half graben in Fig. 1B). The Enping sag, covering an area of

141

approximately 5000 km2, comprises one of the major hydrocarbon-producing sags in

142

the Zhu-I depression (Xu et al., 2014; Fig. 1A). Similar to the other sags or

143

sub-tectonic units in the PRMB, the Enping sag has experienced two stages of

144

geological evolution, which are the Eocene syn-rift and the Oligocene to Recent

145

post-rift stages (Li, 1993; Chen et al., 2003). Horizon T70 is interpreted as the

146

breakup unconformity (corresponding to regional Nanhai tectonic event, Fig. 2),

147

which separates the underlain syn-rift from the overlying post-rift successions. Other

148

important syn-rift and post-rift unconformities include horizons T80, T60, and T32

149

and T30, which correspond to Zhuqiong-II, Baiyun, and Dongsha tectonic events,

150

respectively (Fig. 2). The stratigraphy of the Enping sag consists, from bottom to top,

151

of semi-deep to deep lacustrine shales with thin sandstone interbeds in the lower

152

Eocene Wenchang formations; fluvial-lacustrine, swamp and deltaic sandy shales with

153

thin coal interbeds in the upper Eocene Enping Formation; littoral sandy shales in the

AC C

EP

TE D

M AN U

SC

RI PT

132

ACCEPTED MANUSCRIPT Oligocene Zhuhai Formation; and marine sandstones, siltstones, and shales in the

155

lower Miocene Zhujiang Formation, the middle Miocene Hanjiang Formation, the

156

upper Miocene Yuehai Formation, and the Pliocene Wanshan Formation, as well as

157

the Quaternary layers (Fig. 2; Chen et al., 2003).

158

Several sets of faults, trending in the WNW to NW, E-W to ENE-WSW, and NE-SW

159

directions, have been identified in the PRMB (Li, 1993; Chen et al., 2003). The

160

NE-SW-striking faults primarily formed during the syn-rift stage (Eocene, Tg-T70) in

161

response to NW-SE-orientated extension (Lüdmann and Wong, 1999; Wang et al.,

162

2011; Xu et al., 2014; Leyla et al., 2015; Wu et al., 2015; Hu et al., 2016a, 2016b); the

163

E-W- to ENE-WSW-trending faults were mostly developed during the early post-rift

164

stage (Oligocene to early Early Miocene, T70-T50) in response to roughly

165

N-S-orientated extension (Lüdmann and Wong, 1999; Chan et al., 2010; Wang et al.,

166

2011; Wu et al., 2015; Hu et al., 2016a, 2016b); and the WNW- to NW-trending faults

167

predominantly formed during the middle to late post-rift stage (late Early Miocene to

168

Pliocene, T50-T20) associated with NNE-SSW-orientated extension (Chan et al.,

169

2010; Wu et al., 2015; Hu et al., 2016a, 2016b).

170

3. Data and methods

171

A 1300 km² three-dimensional prestack time-migrated seismic data volume in the

172

Enping sag was available, covering the fault zone studied here (Fig. 1B). Our target

173

interval includes the whole Cenozoic sedimentary succession in the basin, with depths

174

in our study area ranging from seafloor at approximately 110 m deep to the

AC C

EP

TE D

M AN U

SC

RI PT

154

ACCEPTED MANUSCRIPT pre-Cenozoic basement of about 7000 m. Corresponding interval velocity increases

176

from 1700 m/s near seafloor to about 4500 m/s near the basement, as revealed by

177

drilling data. The dominant frequency of our seismic data varies from 45 Hz in the

178

shallower interval to about 25 Hz in the deeper interval close to the basement.

179

Therefore, the vertical resolution of our seismic data in terms of tuning thickness is

180

variable, ranging from approximately 9 m in shallower interval to 45 m in deeper

181

interval. In addition, drilling data from one nearby exploration well (well A in Fig. 1B)

182

were used to make a synthetic seismogram for horizon calibration (Fig. 3).

183

The fault zone was characterized by integrating profile- and horizon-slice-based

184

interpretation (Figs. 4-5). Through interpreting the seismic profiles, fault elements

185

including morphology and occurrence of the fault zone were determined, and fault

186

throws were quantitatively measured (Fig. 4). Horizon slices with seismic coherence

187

attribute were used to describe the plane-view characteristics of the fault zone (Fig. 5).

188

In this study, the coherence attribute was built on the amplitude variance algorithm

189

(Bahorich and Farmer, 1995).

190

In order to quantitatively describe the fault zone, we measured fault throws through a

191

total of 151 seismic profiles oriented perpendicular to the studied fault zone at an

192

average spacing of 200 m. Here, a fault throw in time difference (ms TWT) was

193

measured between the corresponding footwall and hanging wall cut-offs of a horizon

194

on the seismic profiles, which was subsequently converted from time (ms TWT) to

195

depth (m) using a velocity model derived from the time-depth relationships from

196

well-logging data. These fault throws were further used to calculate parameters

AC C

EP

TE D

M AN U

SC

RI PT

175

ACCEPTED MANUSCRIPT depicting the activity and growth history of the fault zone, including expansion index

198

(EI) and fault activity rate (FAR). EI, defined as the ratio of hanging-wall to footwall

199

thicknesses of a given layer, is a traditional and widely-used parameter for kinematic

200

analysis of growth faults (Thorsen, 1963; Jackson and Rotevatn, 2013; Tvedt et al.,

201

2013; Jackson et al., 2016). FAR is calculated by dividing thickness difference

202

between the hanging-wall and footwall for a given layer by the corresponding

203

deposition time. In making these measurements and calculations, ten age-dated

204

regional seismic horizons were used, which from bottom to top are Tg (base of of the

205

Cenozoic basin, ~66 Ma), T80 (base of the Upper Eocene, ~38 Ma), T70 (base of the

206

Oligocene, ~33.9 Ma), T60 (base of the Lower Miocene, ~23.03 Ma), T50 (boundary

207

between the lower and upper portions of the Lower Miocene, ~19.1Ma), T40 (base of

208

the Middle Miocene, ~15.97 Ma), T35 (boundary between the lower and upper

209

portions of the Middle Miocene, ~13.82 Ma), T32 (base of the Late Miocene, ~10

210

Ma), T30 (base of the Pliocene, ~5.33 Ma) and T20 (base of the Quaternary, ~2.59 Ma)

211

(Figs. 2-3).

212

Several types of plots, including T-x plots (Figs. 6-7), T-z plots (Fig. 8), EI and FAR

213

plots (Fig. 8), and throw-strike projections (Fig. 9), as well as time thickness maps

214

(Fig. 10), were compiled to analyze the temporal and spatial distribution

215

characteristics of the fault throws, from which the growth and linkage history of the

216

fault segments were inferred. A T-x plot in which fault throws are plotted against the

217

distance along the strike of the fault depicts the lateral variations of fault throws along

218

the strike of a fault. It is used to analyze the segmentation and lateral linkage of faults

AC C

EP

TE D

M AN U

SC

RI PT

197

ACCEPTED MANUSCRIPT in map view. Throw minima on a T-x plot represent the position of fault segment

220

interconnections, and throw maxima represent the center of each fault segment

221

(Cartwright et. al., 1995, 1996; Jackson and Rotevatn, 2013; Tvedt et al., 2013). A T-z

222

plot in which fault throws are plotted against the depth to the midpoint between the

223

respective hanging wall and footwall cut-offs displays vertical throw variations along

224

a fault plane. It is used to analyze the vertical segmentation and dip-linkage of a fault

225

and to reconstruct its growth history (Mansfield and Cartwright, 1996; Nicol et al.,

226

1996; Meyer et al., 2002; Childs et al., 1996, 2003; Ge and Anderson, 2007; Baudon

227

and Cartwright, 2008a, 2008b, 2008c; Jackson and Rotevatn, 2013; Robson et al.,

228

2016, 2017). A throw-strike projection in which throw contours are plotted on a

229

vertical fault plane projection is used to illustrate throw distribution across fault

230

surfaces and provide insights into the growth and linkage history of segmented faults

231

in three-dimensional fault surfaces (Chapman and Meneilly, 1991; Walsh and

232

Watterson, 1991; Childs et al., 1993; Tvedt et al., 2013). The EI and FAR plots in

233

which EI or FAR values are plotted against the corresponding strata depth or

234

deposition time quantitatively reveal the variations in fault activity intensity with time.

235

Time thickness maps of major stratigraphic intervals within the growth sequence

236

reveal the variation of fault-controlled depocentres with time, therefore indirectly

237

reflect the growth history of the bounding faults (Jackson et al., 2016).

238

It should be pointed out that potential uncertainties may exist in our methodology,

239

including those associated with data resolution, fault and horizon picking, treatment

240

of fault dragging, and compaction effect. Firstly, as decreasing of data resolution with

AC C

EP

TE D

M AN U

SC

RI PT

219

ACCEPTED MANUSCRIPT depth, uncertainty in fault throw measurements will increase. Secondly, errors in

242

manual fault and horizon picking will inevitably cause uncertainty in measurement of

243

fault throws. Fortunately, this uncertainty was minimized by our reliable fault and

244

horizon picking on the high-quality 3D seismic data volume. Thirdly, improper

245

treatment of drag folds may cause certain uncertainty in throw measurements. We

246

followed the traditional method in structural geology to exclude the drag-folding

247

effect by defining the cut-offs using an extrapolated line that follows the regional

248

trend of the horizon prior to folding (e.g. Chapman and Meneilly, 1991; Mansfield

249

and Cartwright, 1996; Wilson et al., 2013; Duffy et al., 2015). Finally, we did not

250

make a decompaction correction on the measured fault throws being lack of reliable

251

local lithology information, which may result in potential uncertainty of throw

252

measurements. This uncertainty increases with burial depth (Taylor et al., 2008).

253

According to the porosity-depth equations in the PRMB (Equations 4 and 5 in He et

254

al., 2017), for an assumed sand-mud interbedded unit consisting of 50% mudstone

255

and 50% sandstone, compaction-caused thickness reduction will be up to 20% at 1000

256

m burial depth, 41% at 3000 m, 49% at 5000 m. This compaction-caused thickness

257

reduction will give rise to underestimating the fault throws. The deeper the burial

258

depth, the lager the degree of this underestimation.

259

4. Results

260

The EP18 fault zone is approximately 30.5 km long and has an approximately E-W

261

strike (Figs. 1 and 5). Coherence horizon slices indicate that the fault zone consists of

AC C

EP

TE D

M AN U

SC

RI PT

241

ACCEPTED MANUSCRIPT 5 to 9 fault segments at different horizons (or depths). This fault zone is evidently

263

composed of two tiers of fault systems separated by horizon T70, which are named

264

the lower and upper fault systems, respectively (Fig. 5). This vertical zonation is also

265

reflected in the fault parameter plots including the fault T-z plots (Fig. 8), EI and FAR

266

plots (Fig. 8), and throw-strike projections (Fig. 9).

267

4.1. The lower fault system

268

4.1.1. Geometry

269

As shown in the coherence horizon slices at horizons T70 and T80, the lower fault

270

system consists of three right-stepping, sinistral ENE-WSW- to E-W-striking fault

271

segments, named F1 to F3 from west to east, which are linked by two NW-SE-striking

272

short fault segments, named A and B (Figs. 5I-5L). These fault segments were

273

connected end-to-end to form a zigzag E-W-striking fault system. Noted that there is a

274

WNW-ESE-striking fault segment, named C, to the south of fault segment F3,

275

approaching westwards to but not linked with the E-W-striking fault system (Figs.

276

5I-5L).

277

In cross sections, the lower fault system offset downward into the basement (below

278

horizon Tg) and upward to horizon T70 at the Eocene-Oligocene boundary; and it

279

shows a listric fault surface geometry, with dips ranging from 35° to 48° and

280

decreasing downwards (Fig. 4). The fault system displays obvious growth

281

characteristics (Fig. 4), with a maximum thickness of syn-rift growth strata in the

282

hanging wall up to 4500 m (Figs. 4 and 10D).

AC C

EP

TE D

M AN U

SC

RI PT

262

ACCEPTED MANUSCRIPT 4.1.2. Throw distribution

284

As shown in the T-x plot (Fig. 6B), throws of the lower fault system are highly

285

variable, with the maximum found at horizon Tg in fault segment F2. Overall, fault

286

throws are larger at the center of the fault system, and decrease toward both the east

287

and west ends. Although the lower fault system consists of five fault segments as

288

shown in the horizon slices (Figs. 5I-5L), no obvious segmentation characteristics can

289

be identified in the T-x plots (Fig. 6B). Nevertheless, an abnormal throw decrease at a

290

distance of ~24 km is noted in the T-x plot (Fig. 6B).

291

The lower fault system shows an asymmetric pattern of throw distribution in the T-z

292

plots (Figs. 8A-E) in which throws decrease from horizon Tg upward to horizon T70.

293

Neglecting possible erosion (especially in the footwall) and compaction effects, the

294

maximum cumulative throw of the lower fault system is up to 5000 m (Figs. 6B and

295

8A-E). The fault system is characterized by tight horizontal contours on the

296

throw-strike projections (Fig. 9).

297

4.1.3. Fault activity

298

The lower fault system offset horizons between Tg and T70 during the Eocene (Figs.

299

4-6, 8-10). EI values may be up to 9.5, and are normally larger in the center and

300

decreases toward both ends (Figs. 8A-E). FAR values varied between 10.1 and 70.1

301

m/Ma during the Early Eocene (Tg-T80) and between 53.6 and 509.3 m/Ma during

302

the Late Eocene (T80-T70) (Figs. 8A-E). It should be noted that no EI values

303

available for the Early Eocene interval between horizons Tg and T80 in most portions

AC C

EP

TE D

M AN U

SC

RI PT

283

ACCEPTED MANUSCRIPT of the lower fault system due to no net depositional thickness on the footwall, leading

305

to the EI values to be infinite (Figs. 4 and 8B-E).

306

4.2. The upper fault system

307

4.2.1. General characteristics

308

The upper fault system developed over the lower one, and shows a similar

309

E-W-trending strike and approximately equal trace lengths to its lower counterpart

310

(Fig. 5). The upper fault system contains nine separate to partially connected fault

311

segments, which are named fault segments f1 to f9 from west to east, respectively

312

(Figs. 5A-H). Each fault segment is from 1.2 to 8.5 km in length. These fault

313

segments trend in a WNW-ESE direction and are arranged in a left-stepping, dextral,

314

en-echelon pattern (Figs. 5A-H).

315

All the fault segments comprising the upper fault system are normal growth faults

316

characterized by decreasing-upward fault throws (Figs. 4 and 8). Compared to their

317

lower counterparts, these fault segments are more or less planar and show obviously

318

steeper dips (between 59° and 65°) in cross-section view (Fig. 4). These fault

319

segments were physically linked in their central portions to the underlain syn-rift fault

320

system (Fig 4). At the linked portions, the fault planes are of a listric geometry,

321

consisting of a steeper-dipping upper portion (the upper fault system) and a

322

gently-dipping lower portion (the lower fault system) (Figs. 4A, E-I). As away from

323

their central portions, the upper fault segments were linked through a gently dipping

324

connecting segment with the underlain lower fault system, making the linked fault

AC C

EP

TE D

M AN U

SC

RI PT

304

ACCEPTED MANUSCRIPT 325

plane in a sigmoid shape (Figs.

326

segments were mostly not physically linked with the underlain lower fault system

327

(Figs. 4C, H).

328

Overall, the upper fault segments show much smaller throws and much lower EI and

329

FAR values than their lower counterparts (Figs. 4, 8-9).

330

4.2.2. Fault segment f1

331

Fault segment f1 lies at the western end of the upper fault system and approximately

332

over the western portion of fault segment F1 (Fig. 5). The fault segment strikes in a

333

WNW-ESE direction, dips SSW, and is up to 6.7 km in length, with fault length

334

gradually decreasing upward (Figs. 5A-H). In cross-sections, fault segment f1 shows

335

evident growth characteristics, with stratal thickness in its hanging wall obviously

336

greater than that in its footwall (Fig. 4). The fault segment was physically linked with

337

the underlain fault segment F1 in its center; but toward fault tips it failed to offset

338

downward the horizons below T60, and therefore was not physically linked with the

339

lower fault system (Fig. 4C).

340

The T-z plots (Figs. 8A, F) taken at the central portion of fault segment f1 show an

341

asymmetric pattern of throw distribution, with throws decreasing from horizon T70

342

upward to horizon T20. On the throw-strike projection (Fig. 9A), fault segment f1 is

343

characterized by semielliptical throw contours, with a maximum throw at horizon T70

344

in its central portion, from which throws decrease upward and toward tip-lines. The

345

fault segment shows much smaller throws (up to 230 m) than its underlain counterpart,

AC C

EP

TE D

M AN U

SC

RI PT

4B-D). Further toward their tips, the upper fault

ACCEPTED MANUSCRIPT fault segment F1 (up to 1500 m, Fig. 9A).

347

Fault segment f1 offset horizons downward to T70 and upward to T20 (Figs. 4-5). It

348

achieves a maximum throw of up to 235 m at horizon T70, (Fig. 7B), from which

349

fault throw decreases upward. Laterally, the fault segment shows larger throws at its

350

center, from which throws decrease toward both ends (Fig. 7B). Throws of the fault

351

segment decrease upwards from horizons T70 to T20 (Fig. 8F). EI and FAR values are

352

variable, which are relatively higher in the T60-T50 (early Early Miocene) and

353

T35-T20 (late Middle Miocene to Pliocene) intervals (Fig. 8F).

354

4.2.3. Fault segments f2 to f6

355

Fault segments f2 to f6 occupy the middle portion of the upper fault system. They are

356

arranged in an E-W-orientated en-echelon pattern in map view, tracing the

357

corresponding portions of the underlain lower fault system, including fault segments

358

eastern F1, A, F2 and western F3. Each of the fault segments has a WNW-ESE strike.

359

These fault segments laterally overlapped and were connected each other (Figs.

360

5A-H). In addition, fault segments f2 and f3 are linked by an additional small

361

E-W-striking fault (fault s in Fig. 7A). The f2 to f6 fault segments show their greatest

362

lengths at horizon T50 (Fig. 5M), from which the trace length of each fault segment

363

decreases both downward to horizon T70 and upward to horizon T20.

364

Fault segment f2 occurs directly over and follows the traces of the eastern F1, A, and

365

F2 fault segments of the lower fault system (Fig. 5). It is approximately 8.5 km,

366

comprising the longest fault segments in the upper fault system.

AC C

EP

TE D

M AN U

SC

RI PT

346

ACCEPTED MANUSCRIPT Fault segment f3 occupies a horizontal position similar to fault segment B, but the

368

strike of the fault segment shows a minor counterclockwise rotation from NW-SE to

369

WNW-ESE (Fig. 5). The fault segment is approximately 5.5 km long, which is

370

evidently longer than its underlying counterpart, fault segment B (about 3 km).

371

Fault segments f4 to f6 in general follow fault segment F3, but each of the fault

372

segments has a WNW-ESE strike, which is obviously different from the underlain

373

ENE-WSW-extending fault segment (Fig. 5).

374

As shown in the T-z plots (Figs. 8B-E) taken at the center of each fault segment, fault

375

segments f2 to f5 show a throw maxima zone at horizon T50, which is separated by a

376

throw minima zone near horizon T70 from the underlain lower fault system. From the

377

throw-strike projections (Figs. 9B-D), fault segments f2 to f5 are characterized by an

378

approximately symmetric distribution of throws and elongated, elliptical throw

379

contours, which are centered in the throw maxima zone (near horizon T50). The

380

throws observed between the T70-T50 strata decrease from horizon T50 downward to

381

T70 (Figs. 9B-D).

382

Fault segments f2 to f6 offset horizons from T50 upward to T20 (Figs. 4-5 and 8G-J),

383

and achieve their maximum fault throws at horizon T50, from which the throws in

384

general decrease downward to T70 and upward to T20 (Figs. 4-5, 7B and 8G-J).

385

Laterally, throws are highly variable along these fault segments, with a maximum of

386

288 m at the center of fault segment f3. For each fault segment, throws are normally

387

greatest at its center and decrease toward fault tips (Figs. 7B and 9B-D). EI and FAR

388

values are relatively higher in the T40-T35 (early Middle Miocene) and T32-T30

AC C

EP

TE D

M AN U

SC

RI PT

367

ACCEPTED MANUSCRIPT (Late Miocene) intervals (Figs. 8G-J).

390

4.2.4. Fault segments f7 to f9

391

Fault segments f7 to f9 lie in the east of the upper fault system, over the eastern

392

portion of fault segment F3 (Figs. 5E-H). Trending in a WNW-ESE orientation, these

393

fault segments arrange in an en-echelon pattern in which each fault segment is

394

separated by an unbroken relay zone (Figs. 5E-H; Peacock, 2002; Kim et al., 2005).

395

These fault segments are evidently shorter in length (less than 2 km) compared to the

396

other fault segments (f1 to f6, 2-8.5 km) in the upper fault system (Figs. 5E-H).

397

Similar to fault segments f2 to f6, fault segments f7 to f9 are physically linked in their

398

central portions with the underlain syn-rift fault segment F3 (Fig. 4I).

399

The fault segments offset the interval between horizons T70 to T50, with fault throws

400

gradually decreasing upward (Fig. 4I). Trace length of each fault segment shows a

401

similar upward-decreasing pattern (Figs. 5E-H).

402

5. Discussion

403

5.1. Oblique-slip nature of the fault zone

404

As described above, the fault segments comprising both the lower and upper fault

405

systems are all normal growth faults (Fig. 4). The lower fault system consists of two

406

sets of fault segments of different strikes, i.e. the ENE-WSW- to E-W-striking F1 to

407

F3 and the NW-SE-striking A and B fault segments. These fault segments were

408

physically linked in an end-to-end pattern to form a zigzag E-W-striking fault system

AC C

EP

TE D

M AN U

SC

RI PT

389

ACCEPTED MANUSCRIPT (Figs. 5-6). The upper fault system consists of nine WNW-ESE-striking, left-stepping,

410

en-echelon arranged fault segments (Figs. 5 and 7). All these indicate that both the

411

lower and upper fault systems were formed in an oblique-slip extension stress

412

environment. This argument is also supported by the findings that the E-W-oriented

413

fault zone was oblique to the simultaneous regional stress field direction. It is

414

suggested that the regional stress field orientation in the PRMB experienced clear

415

clockwise rotation from NW-SE extension during the syn-rift stage (Eocene, Tg to

416

T70) to N-S extension during the early to middle post-rift stage (Oligocene to early

417

Early Miocene, T70 to T50), and further to NNE-SSW extension during the middle to

418

late post-rift stage (late Early Miocene to Pliocene, T50 to T20) (Lüdmann and Wong,

419

1999; Chan et al., 2010; Wang et al., 2011; Wu et al., 2015; Hu et al., 2016a, 2016b).

420

The regional NW-SE-orientated extension stress field acting on the E-W-trending

421

fault zone might cause the formation of the right-stepping, sinistral en-echelon lower

422

fault system during the Eocene ages. Similarly, regional N-S to NNE-SSW orientated

423

extension acting on the E-W-trending fault zone might give rise to the development of

424

the left-stepping, dextral en-echelon upper fault system during the Oligocene to

425

Pliocene ages. It is suggested that step sense is directly related to the slip sense of an

426

en-echelon fault zone, that is, left-stepping fault segments are indicative of dextral

427

displacement, whereas right-stepping segments indicate sinistral displacement

428

(Tchalenko, 1970; Sylvester, 1988; Richard, 1991; Smith and Durney, 1992; Crider,

429

2001; Cembrano et al., 2005; Swanson, 2006; Ghosh and Chattopadhyay, 2008;

430

Dooley and Schreurs, 2012).

AC C

EP

TE D

M AN U

SC

RI PT

409

ACCEPTED MANUSCRIPT 5.2. Time and intensity of fault activity

432

As previously described, the lower fault system was active during the Eocene

433

(Tg-T70) with major activity during the late Eocene (T80-T70) (Fig. 8). In general,

434

the central portion of the lower fault system has greater throws and activity rates than

435

those at both ends (Figs. 6 and 8), indicating that the lower fault system was more

436

intensely active in its central portion.

437

The upper fault system was overall active during the Oligocene to Pliocene ages (Figs.

438

5 and 8). Nevertheless, certain difference in fault-activity time of individual

439

component fault segments is noted: fault segment f1 was active during the Oligocene

440

to Pliocene (T70-T20), fault segments f7 to f9 during the Oligocene to early Early

441

Miocene (T70-T50), and fault segments f2 to f6 during the late Early Miocene to

442

Pliocene (T50-T20). Also, different fault segments achieved their maximum throws

443

and trace lengths at different horizons: fault segments f1 and f7 to f9 at horizon T70

444

(Figs.

445

4M, 7B-E and 8B-D). This indicates that different segments were nucleated at

446

different times: fault segments f1 and f7 to f9 at the western and eastern ends of the

447

upper fault system nucleated earlier (during the beginning of the Oligocene), while

448

fault segments f2 to f6 in the middle nucleated later (during the late Early Miocene).

449

In addition, in terms of fault activity, fault segments f2 to f4 in the center of the upper

450

fault system normally show greater throws and higher EI and FAR values than those

451

of the other fault segments on both sides (Figs. 7-9), which indicates that the central

452

portion of the upper fault system had more intense activity than that at both ends.

TE D

M AN U

SC

RI PT

431

AC C

EP

2D, 2H, 4M, 7A and 8A), and fault segments f2 to f6 at horizon T50 (Figs.

ACCEPTED MANUSCRIPT In short, the lower and upper fault systems were active during the Eocene and the

454

Oligocene to Pliocene ages, which rightly correspond to the syn-rift and post-rift

455

stages, respectively (Fig. 2). Therefore, we interpreted the lower and fault systems as

456

syn-rift and post-rift fault systems, respectively. In addition, the lower fault system

457

shows much larger fault throws and higher EI and FAR values than those in the upper

458

fault system, which indicates that activity of the EP18 fault system is much stronger

459

in the syn-rift stage than that in the post-rift stage.

460

5.3. Growth pattern and strike linkage of the fault systems

461

The lower fault system is an integrated zig-zig fault, which consist of five fault

462

segments with two sets of strike: E-W to ENE-WSW (fault segments F1 to F3) and

463

the NW-SE (fault segments A and B) (Figs. 5I-L). These fault segments connected

464

end by end, and physically linked (hard-linked) each other. However, no evident

465

throw decreases in the linkage zones of adjacent fault segments can be identified on

466

the T-x plots (Fig. 6B), which could indicate that these fault segments might be linked

467

very early, and rapidly grow together as an integrated fault. Moreover, as shown on

468

horizon slices, the length of the lower fault system is almost unchanged throughout its

469

growth history (Figs. 5I-L). Therefore, we speculate that the lower fault system might

470

grow in accordance with the constant-length fault model (Morley, 1999; Walsh et al.,

471

2002, 2003; Giba et al., 2012; Jackson and Rotevatn, 2013; Jackson et al., 2016; Nicol

472

et al., 2016). It means that the fault segments comprising the lower fault system might

473

attain their near-final lengths very early, and be quickly linked each other to form an

AC C

EP

TE D

M AN U

SC

RI PT

453

ACCEPTED MANUSCRIPT integrated fault, and then get accumulation in throws without further significant lateral

475

propagation. Similar examples of oblique-slip faults have been reported in rifted

476

basins like the Taranaki Basin in New Zealand (Giba et al., 2012), the Suez Rift in

477

Egypt (Jackson and Rotevatn, 2013), and the Egersund Basin offshore Norway (Tvedt

478

et al., 2013).

479

The upper fault system consists of nine en-echelon aligned growth fault segments

480

(Figs. 5A-H). The fault segments in the center, f2 to f6, were partially hard-linked,

481

especially in the H50-H35 interval (Figs. 5C-F); while those on both ends of the fault

482

system, including fault segments f1 in the west end and fault segments f7 to f9 in the

483

east end, were not physically linked with adjacent fault segments as indicated by the

484

presence of the unbroken relay zones between the fault segments (Figs. 5A-H). This

485

could indicate that activity in the middle portion of the upper fault system was overall

486

more intense than in the west and east ends of the same fault system.

487

Throws of the fault segments in the upper fault system are normally greater at their

488

center and decrease toward their lateral tips (Fig. 7B). Evident throw decreases are

489

noted in the overlapping areas of adjacent fault segments (Fig. 7B). We suggest that

490

the growth of single fault segments in the upper fault system might be consistent with

491

the isolated fault model (Barnett et al., 1987), in which individual fault segments

492

could be nucleated in isolation from other fault segments, grow with a simultaneous

493

increase in both displacement and fault length by radial propagation of tip-lines

494

(Walsh and Watterson, 1988; Cartwright et al., 1995; Walsh et al., 2003; Giba et al.,

495

2012; Jackson and Rotevatn, 2013; Jackson et al., 2016). Similar literature examples

AC C

EP

TE D

M AN U

SC

RI PT

474

ACCEPTED MANUSCRIPT come from the Canyonlands Grabens of SE Utah (Cartwright et al., 1995), the Levant

497

Basin in the eastern Mediterranean (Baudon and Cartwright, 2008a), and the

498

Egersund Basin offshore Norway (Tvedt et al., 2013).

499

5.4. Dip linkage of both fault systems

500

Dip linkage of the upper fault segments with the underlain lower fault system shows

501

obvious segmentation characteristics.

502

Fault segment f1 at the western end of the post-rift fault system show similar

503

WNW-ESE strike with the underlying syn-rift fault segment F1 (Fig. 5M). Both fault

504

segments (f1 and western F1) are almost completely hard-linked (Fig. 9A). Throws of

505

fault segment f1 gradually decrease from horizon T70 upward. We infer that fault

506

segment f1 could be resulted from the upward propagation of the lower fault segment

507

(F1). Nevertheless, both ends of fault segment f1 were not linked with the underlying

508

F1 fault segment (Fig. 4C), which could represent a lateral propagation of this fault

509

segment.

510

Fault segments f7 to f9 at the eastern end of the post-rift fault system were only

511

physically linked in their middle portions with the underlain syn-rift fault system.

512

Similar to fault segment f1, throws of these fault segments (f7 to f9) decrease from

513

horizon T70 upward, which could be the result of upward propagation of

514

corresponding syn-rift fault segments (eastern F3). However, fault segments f7 to f9

515

take

516

ENE-WSW-extending fault segment (eastern F3). The latter might be associated with

AC C

EP

TE D

M AN U

SC

RI PT

496

a

WNW-WSW

strike,

which

is

different

from

the

underlying

ACCEPTED MANUSCRIPT the lateral propagation of these fault segments themselves.

518

Fault segments f2 to f6 in the central portion of the post-rift fault system show a

519

different fault-throw distribution pattern from the f1 and f7 to f9 fault segments at

520

both ends of the same fault system. These fault segments (f2-f6) achieve a throw

521

maximum at horizon T50, from which fault throws decrease downward to T70 (Figs.

522

8B-E, 9B-D). There is a throw minimum zone near horizon T70, which is separated

523

the upper fault segments from the underlain syn-rift fault system (Figs. 8B-E and

524

9B-D). Fault segments f2 to f6 were physically linked with the lower fault system

525

only at their central portions (Fig. 4). We infer that these linkages could be associated

526

with downward propagation of the throws of the fault segments. Toward both ends,

527

these fault segments are deviated from and not linked to the underlying syn-rift fault

528

system (Fig. 4H). This deviation in the strike of the fault segments may be associated

529

with a contemporaneous tectonic stress field different from that during the formation

530

of the previous lower fault system.

531

In summary, there are two dip linkage models between the post-rift fault segments and

532

the underlain lower fault system. One is the upward propagation of the lower fault

533

segments, and another is the downward propagation of the upper fault segments. Both

534

linkage models can be distinguished by their different patterns of throw distribution.

535

Throw distribution in the upward propagation model is characterized by a gradual

536

upward decrease in fault throws from the lower to the upper fault systems, whereas

537

throw distribution in the downward propagation model is featured by a throw

538

minimum zone between the lower and upper fault systems (Baudon and Cartwright,

AC C

EP

TE D

M AN U

SC

RI PT

517

ACCEPTED MANUSCRIPT 2008b; Jackson and Rotevatn, 2013; Tvedt et al., 2013). The upward propagation

540

model occurs in the west and east ends of the fault zone, between fault segments f1

541

and western F1 in the west end, and fault segments f7 to f9 and eastern F3 in the east

542

end. The downward propagation model is observed in the central portion of the fault

543

zone, between fault segments f2 to f6 and the underlying syn-rift fault system.

544

5.5. Influence of the pre-existing fault system

545

The pre-existing lower fault system might exert an important influence on the

546

development of the upper fault system. The major evidence is as follows: (1) Most of

547

the fault segments in the upper fault system were physically linked in their central

548

parts with the underlying lower fault system. In these linked portions, the upper fault

549

segments normally achieved their greatest throws. We infer that the upper fault

550

segments could have nucleated at the linked central portions over the lower fault

551

system. In other words, the pre-existing lower fault system controlled the nucleation

552

sites of the post-rift fault segments (Henza et al., 2010, 2011). (2) the upper fault

553

system is overall superimposed on the lower fault system. It shows a similar

554

E-W-trending strike and has an approximately equal extending length to the lower

555

fault system. This indicates that the pre-existing lower fault system could control the

556

position, strike, and scale of the upper fault system. Similar phenomena have been

557

observed in the Tertiary rift basins of Thailand, where late-stage faults followed

558

pre-existing faults, showing evident control from the pre-existing faults (Kornsawan

559

and Morley, 2002; Morley et al., 2004).

AC C

EP

TE D

M AN U

SC

RI PT

539

ACCEPTED MANUSCRIPT Difference in nucleation horizons (and therefore geological time) between the central

561

fault segments (f2 to f6) and the end fault segments (f1 and f7 to f9) in the upper

562

system is noted: the central fault segments (f2 to f6) at horizon T50, and the end fault

563

segments (f1 and f7 to f9) at horizon T70. It means that only the west and east ends of

564

pre-existing lower fault system were active during the early post-rift stage (T70-T50).

565

This could indicate that regional tectonic stress was relatively weak in the early

566

post-rift stage, and therefore only the tips of the pre-existing lower syn-rift fault

567

system reactivated, possibly due to the effect of stress concentration, which led to the

568

nucleation of fault segments f1 and f7 to f9 at the west and east ends of the fault zone.

569

Nucleation at horizon T50 of fault segments f2 to f6 at the central portion of the upper

570

fault system could be associated with local increase of tectonic stress. Similar

571

selective or segmented reactivation of pre-existing faults have been reported by

572

Baudon and Cartwright (2008b).

573

Reactivation of pre-existing faults resulting in the formation of en-echelon

574

oblique-slip fault segments has been reported from several rifted basins that

575

experienced multiple phases and different directions of extension (Morley et al., 2004;

576

Frankowicz and McClay, 2010; Giba et al., 2012; Jackson and Rotevatn, 2013). It

577

may represent an important formation mechanism for the formation of late-stage

578

en-echelon fault segments. In previous case studies, pre-existing faults are normally

579

dip-slip normal faults, and only the late-stage faults are en-echelon oblique-slip fault

580

segments (Morley et al., 2004; Frankowicz and McClay, 2010; Giba et al., 2012;

581

Jackson and Rotevatn, 2013). In our study, both the pre-existing lower and the

AC C

EP

TE D

M AN U

SC

RI PT

560

ACCEPTED MANUSCRIPT resulting upper fault systems are all composed of several en-echelon oblique-slip fault

583

segments (Fig. 5), and the lower and upper fault segments show different strikes and

584

opposite step senses: the lower fault system consists of ENE-WSW- to E-W-striking,

585

right-stepping en-echelon fault segments, while the upper fault system has

586

WNW-ESE-striking, left-stepping en-echelon fault segments. Formation of both the

587

lower and upper oblique-slip fault systems could be associated with the variations in

588

extension direction. Regional geological studies indicate that the stress field

589

orientation in the PRMB experienced clear clockwise rotation from NW-SE extension

590

during the syn-rift stage (Eocene, Tg-T70) to N-S extension during the early post-rift

591

stage (Oligocene to early Early Miocene, T70 to T50), and further to NNE-SSW

592

extension during the middle to late post-rift stage (late Early Miocene to Pliocene,

593

T50 to T20) (Lüdmann and Wong, 1999; Chan et al., 2010; Wang et al., 2011; Wu et

594

al., 2015; Hu et al., 2016a, 2016b).

595

In addition, the syn-rift fault system studied here strikes in an overall E-W direction,

596

which is different from the simultaneous NE-SW-orientated faults formed during the

597

syn-rift stage in the PRMB (Wang et al., 2011; Wu et al., 2015; Hu et al., 2016a,

598

2016b). We speculate that an E-W-trending crustal weakness zone could exist before

599

the syn-rift extension, which controlled the location and strike of the lower fault

600

system formed later. Along this weakness zone, the right-stepping en-echelon syn-rift

601

fault system developed under a local sinistral shear stress field caused by the

602

NW-SE-orientated extension acting on the E-W-trending crustal weakness zone. This

603

hypothesis is also supported by the following observations: (1) Although the lower

AC C

EP

TE D

M AN U

SC

RI PT

582

ACCEPTED MANUSCRIPT fault system shows evident segmentation characteristics in the horizon slices (Figs.

605

5I-L), its throw distribution (greater in the center, and decreasing towards lateral tips,

606

Fig. 6B) is more similar to a coherent fault system, indicating possible influence of a

607

pre-existing fault; (2) As discussed above, the length of the fault system was almost

608

unchanged during its life (Figs. 5I-L); therefore, the lower fault system could grow in

609

accordance with the so-called constant-length model (Morley, 1999; Walsh et al.,

610

2002, 2003; Giba et al., 2012; Jackson and Rotevatn, 2013; Jackson et al., 2016; Nicol

611

et al., 2016). Several studies indicate that under the influence of a pre-existing fault,

612

growth of later formed fault tends to be consistent with the constant-length model

613

(Morley, 1999; Giba et al., 2012; Jackson and Rotevatn, 2013).

614

5.6. The evolution model of the fault zone

615

Based on the growth and linkage characteristics of the syn-rift and post-rift fault

616

systems and regional stress field background, the evolution of the fault zone studied

617

here can be summarized in four stages as follows.

618

(1) Syn-rift stage: the growth of the E-W-striking lower oblique-slip normal fault

619

system during the Eocene (Tg-T70) in response to a regional NW-SE-orientated

620

extension (Fig. 11A). As previously discussed, the lower fault system could be built

621

on a pre-existing E-W-trending crustal weakness zone, which was oblique to the

622

regional NW-SE extension during the syn-rift stage, causing a local sinistral shear

623

stress field. The latter resulted in the formation of the syn-rift oblique-slip fault

624

system. The syn-rift fault system consists of three right-stepping, sinistral,

AC C

EP

TE D

M AN U

SC

RI PT

604

ACCEPTED MANUSCRIPT ENE-WSW- to E-W-striking fault segments (fault segments F1, F2 and F3) that were

626

hard-linked by two NW-SE-striking fault segments (fault segments A and B, Figs.

627

5I-L). These syn-rift fault segments could have attained their near-final lengths very

628

early and quickly linked to each other to form a continuous integrated fault. Then the

629

integrated fault accumulated throw without further significant lateral propagation.

630

Therefore, the growth pattern of the syn-rift fault system is overall in accordance with

631

the so-called constant length model (Childs et al., 1995; Walsh et al., 2002, 2003;

632

Giba et al., 2012; Jackson and Rotevatn, 2013; Jackson et al., 2016; Nicol et al., 2016).

633

Serving as the boundary fault of the EP18 half graben, the syn-rift fault system

634

controlled the sedimentary evolution of the half graben, leading to the formation of

635

thick syn-rift growth strata consisting of the Wenchang and Enping formations (Fig.

636

10D).

637

(2) Early to middle post-rift stage: a relatively dormant stage of fault activity during

638

the Oligocene to early Early Miocene (T70-T50) (Fig. 11B). Only the f1 and f7 to f9

639

fault segments at the western and eastern ends of the fault zone were active during

640

this early post-rift stage (Figs. 8 and 10C), which could have been associated with

641

stress concentration on both ends of the pre-existing lower fault system. Under the

642

action of a regional N-S-orientated extension, the E-W-trending western end of the

643

fault zone expressed dip-slip activity, resulting in the formation of the post-rift f1 fault

644

segment, while the ENE-WSW-orientated eastern portion of the fault zone displayed

645

an oblique-slip activity, causing the development of the en-echelon aligned post-rift

646

fault segments, f7 to f9.

AC C

EP

TE D

M AN U

SC

RI PT

625

ACCEPTED MANUSCRIPT (3) Middle post-rift stage: nucleation and growth of the central post-rift oblique slip

648

fault segments (f2 to f6) during the late Early Miocene to late Middle Miocene

649

(T50-T32), in response to a regional NNE-SSW-orientated extension (Fig. 11C).

650

During this stage, fault segment f1 persistently developed, and growth of fault

651

segments f7 to f9 terminated. Fault segments f2 to f6 nucleated and grew in their

652

central portions (Fig. 10B). These fault segments grew by radial propagation,

653

overlapping and interacting with each other (Fig. 11C) during this stage. The growth

654

pattern of the post-rift fault segments (f2 to f6) is in accordance with the isolated fault

655

model (Walsh and Watterson, 1988; Cartwright et al., 1995; Walsh et al., 2003; Giba

656

et al., 2012; Jackson and Rotevatn, 2013). Both ends of fault segments f2 to f6 are

657

characterized by elliptical throw contours on the throw-strike projections in the

658

T50-T32 (late Early Miocene to late Middle Miocene) interval (Figs. 10B-D), which

659

could indicate that these fault segments were still increasing their lengths and not

660

linked at that time.

TE D

M AN U

SC

RI PT

647

(4) Late post-rift stage: continued growth of the post-rift fault segments, f1 to f6,

662

during the Late Miocene to Pliocene (T32-T20, Figs. 10A and 11D). Both the west

663

and east ends of fault segments f2 to f6 are characterized by horizontal throw contours

664

in this stage in the throw-strike projections (Figs. 10B-D), which may suggest that

665

these fault segments were linked together. The growth of the fault zone overall

666

terminated at the end of the Pliocene (Fig. 4).

667

AC C

EP

661

ACCEPTED MANUSCRIPT 6. Conclusions

669

(1) The EP18 fault zone, constituting the northern boundary of a half graben in the

670

Enping sag of the northern PRMB, is a complex oblique-slip fault zone consisting of

671

two tiers of oblique-slip fault systems: a lower (syn-rift) and an upper (post-rift)

672

oblique slip fault systems, which were active during the Eocene (Tg-T70) and the late

673

Early Miocene to Pliocene (T50-T20) ages, respectively.

674

(2) The lower (syn-rift) fault system is an integrated zig-zig fault consisting of three

675

right-stepping, sinistral, ENE-WSW- to E-W-striking fault segments (F1, F2 and F3)

676

that are hard-linked by two NW-SE-striking fault segments (A and B). These fault

677

segments might have rapidly attained their near-final lengths very early and quickly

678

linked together as the kinematically related components of a single coherent fault

679

system. Then, the throws of the fault system accumulated without significant further

680

lateral propagation, and therefore, the growth pattern of the fault system is in

681

accordance with the constant length model.

682

(3) The upper (post-rift) fault system is composed of nine WNW-ESE-oriented,

683

left-stepping, dextral, en-echelon fault segments (f1 to f9), which show much smaller

684

throws and less intense activity than that of their lower syn-rift counterparts. Fault

685

segments f1 and f7 to f9 at the western and eastern ends of the fault system formed by

686

upward propagation of the underlain lower fault system, while fault segments f2 to f6

687

in the middle nucleated at horizon T50 over the pre-exiting lower fault system and

688

then propagated downward and linked with the underlain lower fault system. Growth

689

of the upper fault segments is overall consistent with the isolated fault model in which

AC C

EP

TE D

M AN U

SC

RI PT

668

ACCEPTED MANUSCRIPT individual fault segments nucleated in isolation and grew by radial propagation of

691

tip-lines. Most of the fault segments laterally overlapped and interacted with each

692

other, and were soft-linked (fault segments f1 to f2 and f7 to f9) or hard-linked (fault

693

segments f2 to f6).

RI PT

690

(4) The pre-existing lower (syn-rift) fault system shows clear influence on the

695

formation of the upper (post-rift) fault segments by controlling the position, strike,

696

and scale of the post-rift fault system and the nucleation sites of the post-rift fault

697

segments. (5) A conceptual model for the evolution of the EP18 fault zone is

698

presented. Four evolution stages are inferred, which are the formation of the

699

E-W-striking syn-rift oblique-slip normal fault system during the Eocene (Tg-T70) in

700

response to a regional NW-SE-orientated extension, a relatively dormant stage of fault

701

activity during the Oligocene to early Early Miocene (T70-T50), the nucleation and

702

growth of the post-rift oblique-slip normal fault system during the late Early Miocene

703

to late Middle Miocene (T50-T32) in response to a regional NNE-SSW-orientated

704

extension, and the continued growth of the post-rift fault segments during the Late

705

Miocene to Pliocene (T32-T20).

M AN U

TE D

EP

AC C

706

SC

694

707

Acknowledgments

708

We would like to thank the Shenzhen Branch of CNOOC Ltd. for providing the data

709

set and permitting us to publish the results. We are grateful to Christopher Morley, an

710

anonymous reviewer, and the editor Ian Alsop, for their thorough and constructive

711

reviews, which have substantially improved the manuscript. This work was partly

ACCEPTED MANUSCRIPT 712

funded by the National Science Foundation of China [grant numbers 91428039,

713

41676029, and 91028003].

714

RI PT

715 716 717

References

SC

718

720 721

M AN U

719

Bahorich, M., Farmer, S., 1995. 3-D seismic discontinuity for faults and stratigraphic features: the coherence cube. The Leading Edge 14, 1053-1058. Barnett, J., Mortimer, J., Rippon, J., Walsh, J., Watterson, J., 1987. Displacement

723

geometry in the volume containing a single normal fault. AAPG Bulletin 71,

724

925-937.

TE D

722

Baudon, C., Cartwright, J.A., 2008a. Early stage evolution of growth faults: 3D

726

seismic insights from the Levant Basin, Eastern Mediterranean. Journal of

727

Structural Geology 30, 888-898.

AC C

EP

725

728

Baudon, C., Cartwright, J.A., 2008b. The kinematics of reactivation of normal faults

729

using high resolution throw mapping. Journal of Structural Geology 30,

730

1072-1084.

731

Baudon, C., Cartwright, J.A., 2008c. 3D seismic characterisation of an array of blind

732

normal faults in the Levant Basin, Eastern Mediterranean. Journal of Structural

733

Geology 30, 746-760.

ACCEPTED MANUSCRIPT 734 735

Bellahsen, N., Daniel, J.M., 2005. Fault reactivation control on normal fault growth: an experimental study. Journal of Structural Geology 27, 769-780. Bonini, M., Souriot, T., Boccaletti, M., Brun, J.P., 1997. Successive orthogonal and

737

oblique extension episodes in a rift zone: Laboratory experiments with

738

application to the Ethiopian Rift. Tectonics 16, 347-362.

RI PT

736

Cartwright, J.A., Mansfield, C., Trudgill, B., 1996. The growth of normal faults by

740

segment linkage. Geological Society of London Special Publications 99,

741

163-177.

M AN U

SC

739

Cartwright, J.A., Trudgill, B.D., Mansfield, C.S., 1995. Fault growth by segment

743

linkage: an explanation for scatter in maximum displacement and trace length

744

data from the Canyonlands Grabens of SE Utah. Journal of Structural Geology

745

17, 1319-1326.

TE D

742

Cembrano, J., González, G., Arancibia, G., Ahumada, I., Olivares, V., Herrera, V.,

747

2005. Fault zone development and strain partitioning in an extensional strike-slip

748

duplex: A case study from the Mesozoic Atacama fault system, Northern Chile.

749

Tectonophysics 400, 105-125.

AC C

EP

746

750

Chan, L.S., Shen, W., Pubellier, M, 2010. Polyphase rifting of greater Pearl River

751

Delta region (South China): Evidence for possible rapid changes in regional

752

stress configuration. Journal of Structural Geology 32, 746-754.

753

Chapman, T., Meneilly, A., 1991. The displacement patterns associated with a

754

reverse-reactivated, normal growth fault. Geological Society, London, Special

755

Publications 56, 183-191.

ACCEPTED MANUSCRIPT 756

Chattopadhyay, A., Chakra, M., 2013. Influence of pre-existing pervasive fabrics on

757

fault patterns during orthogonal and oblique rifting: An experimental approach.

758

Marine and Petroleum Geology 39, 74-91.

760

Chen, C., Shi, H., Xu S., 2003. Hydrocarbon accumulation condition in Tertiary in the

RI PT

759

Eastern Pear River Mouth Basin [M]. Science Press, Beijing, pp. 1-266.

Childs, C., Easton, S., Vendeville, B., Jackson, M., Lin, S., Walsh, J., Watterson, J.,

762

1993. Kinematic analysis of faults in a physical model of growth faulting above a

763

viscous salt analogue. Tectonophysics 228, 313-329.

766 767 768 769

M AN U

765

Childs, C., Nicol, A., Walsh, J.J., Watterson, J., 1996. Growth of vertically segmented normal faults. Journal of Structural Geology 18, 1389-1397. Childs, C., Nicol, A., Walsh, J.J., Watterson, J., 2003. The growth and propagation of synsedimentary faults. Journal of Structural Geology 25, 633-648.

TE D

764

SC

761

Childs, C., Watterson, J., Walsh, J., 1995. Fault overlap zones within developing normal fault systems. Journal of the Geological Society 152, 535-549. Clifton, A.E., Schlische, R.W., Withjack, M.O., Ackermann, R.V., 2000. Influence of

771

rift obliquity on fault-population systematics: results of experimental clay

AC C

772

EP

770

models. Journal of Structural Geology 22, 1491-1509.

773

Crider, J.G., 2001. Oblique slip and the geometry of normal-fault linkage: mechanics

774

and a case study from the Basin and Range in Oregon. Journal of Structural

775

Geology 23, 1997-2009.

776

Dooley, T.P., Schreurs, G., 2012. Analogue modelling of intraplate strike-slip

777

tectonics: A review and new experimental results. Tectonophysics 574, 1-71.

ACCEPTED MANUSCRIPT 778

Duffy, O.B., Bell, R.E., Jackson, A.L., Gawthorpe, R.L., Whipp, P.S., 2015. Fault

779

growth and interactions in a multiphase rift fault network: Horda Platform,

780

Norwegian North Sea. Journal of Structural Geology 80, 99-119. Fazli Khani, H., Back, S., 2015. The influence of pre-existing structure on the growth

782

of syn-sedimentary normal faults in a deltaic setting, Niger Delta. Journal of

783

Structural Geology 73, 18-32.

RI PT

781

Frankowicz, E., McClay, K., 2010. Extensional fault segmentation and linkages,

785

Bonaparte Basin, outer North west shelf, Australia. AAPG Bulletin 94,

786

977-1010.

788

M AN U

787

SC

784

Ge, H., Anderson, J.K., 2007. Fault throw profile and kinematics of Normal fault: conceptual models and geologic examples. Geol. J. China Univ 13, 75-88. Ghosh, N., Chattopadhyay, A., 2008. The initiation and linkage of surface fractures

790

above a buried strike-slip fault: An experimental approach. Journal of Earth

791

System Science 117, 23-32.

EP

793

Giba, M., Walsh, J.J., Nicol, A., 2012. Segmentation and growth of an obliquely reactivated normal fault. Journal of Structural Geology 39, 253-267.

AC C

792

TE D

789

794

He, M., Zhong, G., Liu, X., Liu, L., Shen, X., Wu, Z., Huang, K., 2017. Rapid

795

post-rift tectonic subsidence events in the Pearl River Mouth Basin, northern

796

South China Sea margin. Journal of Asian Earth Sciences 147, 271-283.

797

Henstra, G.A., Rotevatn, A., Gawthorpe, R.L., Ravnås, R., 2015. Evolution of a major

798

segmented normal fault during multiphase rifting: The origin of plan-view zigzag

799

geometry. Journal of Structural Geology 74, 45-63.

ACCEPTED MANUSCRIPT 800

Henza, A.A., Withjack, M.O., Schlische, R.W., 2010. Normal-fault development

801

during two phases of non-coaxial extension: An experimental study. Journal of

802

Structural Geology 32, 1656-1667. Henza, A.A., Withjack, M.O., Schlische, R.W., 2011. How do the properties of a

804

pre-existing normal-fault population influence fault development during a

805

subsequent phase of extension? Journal of Structural Geology 33, 1312-1324.

806

Higgins, R., Harris, L., 1997. The effect of cover composition on extensional faulting

807

above re-activated basement faults: results from analogue modelling. Journal of

808

Structural Geology 19, 89-98.

M AN U

SC

RI PT

803

Hu, Y., Wu, Z., Zhong, Z., Zhang, J., Yu, W., Wang, G., Liu, Y., Wang, S., 2016a.

810

Characterization and genesis of the Middle and Late Eocene tectonic changes in

811

Zhu 1 Depression of Pearl River Mouth Basin. Oil & Gas Geology 37, 779-785

812

(in Chinese with English abstract).

TE D

809

Hu, Y., Wu, Z., Zhong, Z., Zhang, J., Yu, W., Wang, G., Liu, Y., Xie, F., 2016b.

814

Cenozoic Characteristics and transformation mechanism of fault system in Zhu-1

815

depression, South China. Chinese Journal of Geology 51, 494-509 (in Chinese

AC C

816

EP

813

with English abstract).

817

Jackson, C.A.-L., Bell, R.E., Rotevatn, A., Tvedt, A.B.M., 2016. Techniques to

818

determine the kinematics of synsedimentary normal faults and implications for

819

fault growth models. Geological Society London Special Publications 439, 187–

820

217.

821

Jackson, C.A.-L., Rotevatn, A., 2013. 3D seismic analysis of the structure and

ACCEPTED MANUSCRIPT 822

evolution of a salt-influenced normal fault zone: a test of competing fault growth

823

models. Journal of Structural Geology 54, 215-234.

825 826 827

Keep, M., McClay, K., 1997. Analogue modelling of multiphase rift systems. Tectonophysics 273, 239-270.

RI PT

824

Kim, Y.-S., Sanderson, D.J., 2005. The relationship between displacement and length of faults: a review. Earth-Science Reviews 68, 317-334.

Kornsawan, A., Morley, C.K., 2002. The origin and evolution of complex transfer

829

zones (graben shifts) in conjugate fault systems around the Funan Field, Pattani

830

Basin, Gulf of Thailand. Journal of Structural Geology 24, 435-449.

M AN U

SC

828

Leyla, B., Ren, J., Zhang, J., Lei, C., 2015. En Echelon Faults and Basin Structure in

832

Huizhou Sag, South China Sea: Implications for the Tectonics of the SE Asia.

833

Journal of Earth Science 26(05), 690-699.

836 837

Offshore Oil and Gas (Geology) 7(6), 11-17 (in Chinese with English abstract). Lüdmann, T., Wong, H.K., 1999. Neotectonic regime on the passive continental

EP

835

Li, P., 1993. Cenozoic tectonic movement in the Pearl River Mouth Basin. China

margin of the northern South China Sea. Tectonophysics 311, 113-138.

AC C

834

TE D

831

838

Mansfield, C., Cartwright, J., 1996. High resolution fault displacement mapping from

839

three-dimensional seismic data: evidence for dip linkage during fault growth.

840 841 842 843

Journal of Structural Geology 18, 249-263.

McClay, K., Dooley, T., Whitehouse, P., Mills, M., 2002. 4-D evolution of rift systems: Insights from scaled physical models. AAPG Bulletin 86, 935-960. McClay, K., White, M., 1995. Analogue modelling of orthogonal and oblique rifting.

ACCEPTED MANUSCRIPT 844

Marine and Petroleum Geology 12, 137-151. Meyer, V., Nicol, A., Childs, C., Walsh, J., Watterson, J., 2002. Progressive

846

localisation of strain during the evolution of a normal fault population. Journal of

847

Structural Geology 24, 1215-1231.

RI PT

845

Morley, C.K., 1999. Patterns of displacement along large normal faults : Implications

849

for basin evolution and fault propagation, based on examples from East Africa.

850

Medical Care Review 49, 233-255. Morley,

C.K.,

2007.

Variations

in

late

Cenozoic–Recent

strike-slip

and

M AN U

851

SC

848

852

oblique-extensional geometries, within Indochina: The influence of pre-existing

853

fabrics. Journal of Structural Geology 29, 36-58.

Morley, C.K., 2016. The impact of multiple extension events, stress rotation and

855

inherited fabrics on normal fault geometries and evolution in the Cenozoic rift

856

basins of Thailand. Geological Society of London Special Publications 439, 413–

857

445.

TE D

854

Morley, C.K., Gabdi, S., Seusutthiya, K., 2007. Fault superimposition and linkage

859

resulting from stress changes during rifting: Examples from 3D seismic data,

AC C

860

EP

858

Phitsanulok Basin, Thailand. Journal of Structural Geology 29, 646-663.

861

Morley, C.K., Haranya, C., Phoosongsee, W., Pongwapee, S., Kornsawan, A.,

862

Wonganan, N., 2004. Activation of rift oblique and rift parallel pre-existing

863

fabrics during extension and their effect on deformation style: examples from the

864

rifts of Thailand. Journal of Structural Geology 26, 1803-1829.

865

Nicol, A., Childs, C., Walsh, J.J., Manzocchi, T., Pfer, M.P.J.S., 2016. Interactions and

ACCEPTED MANUSCRIPT 866

growth of faults in an outcrop-scale system. Geological Society London Special

867

Publications 439, 23–39. Nicol, A., Watterson, J., Walsh, J., Childs, C., 1996. The shapes, major axis

869

orientations and displacement patterns of fault surfaces. Journal of Structural

870

Geology 18, 235-248.

RI PT

868

Paredes, J.M., Plazibat, S., Crovetto, C., Stein, J., Cayo, E., Schiuma, A., 2013. Fault

872

kinematics and depocenter evolution of oil-bearing, continental successions of

873

the Mina del Carmen Formation (Albian) in the Golfo San Jorge basin, Argentina.

874

Journal of South American Earth Sciences 46, 63-79.

876

M AN U

875

SC

871

Peacock, D., 2002. Propagation, interaction and linkage in normal fault systems. Earth-Science Reviews 58, 121-142.

Reeve, M.T., Bell, R.E., Duffy, O.B., Jackson, A.L., Sansom, E., 2015. The growth of

878

non-colinear normal fault systems; What can we learn from 3D seismic

879

reflection data? Journal of Structural Geology 70, 141-155.

TE D

877

Richard, P., 1991. Experiments on faulting in a two-layer cover sequence overlying a

881

reactivated basement fault with oblique-slip. Journal of Structural Geology 13,

AC C

882

EP

880

459-469.

883

Robson, A.G., King, R.C., Holford, S.P., 2016. 3D seismic analysis of gravity-driven

884

and basement influenced normal fault growth in the deepwater Otway Basin,

885

Australia. Journal of Structural Geology 89, 74-87.

886

Robson, A.G., King, R.C., Holford, S.P., 2017. Structural evolution of a

887

gravitationally detached normal fault array: analysis of 3D seismic data from the

ACCEPTED MANUSCRIPT 888

Ceduna Sub Basin, Great Australian Bight. Basin Research 29, 605-624. Schlische, R.W., Withjack, M.O., Eisenstadt, G., 2002. An experimental study of the

890

secondary deformation produced by oblique-slip normal faulting. AAPG Bulletin

891

86, 885-906.

RI PT

889

Shi, H., He, M., Zhang, L., Yu, Q., Pang, X., Zhong, Z., Liu, L., 2014. Hydrocarbon

893

geology, accumulation pattern and the next exploration strategy in the eastern

894

Pearl River Mouth Basin. China Offshore Oil and Gas 26 (3), 11-22 (in Chinese

895

with English abstract).

900 901 902 903 904 905 906 907

M AN U

899

Swanson, M.T., 2006. Late Paleozoic strike-slip faults and related vein arrays of Cape Elizabeth, Maine. Journal of Structural Geology 28, 456-473.

TE D

898

in oblique divergence. Tectonophysics 216, 235-253.

Sylvester, A.G., 1988. Strike-slip faults. Geological Society of America Bulletin 100, 1666-1703.

Taylor, S.K., Nicol, A., Walsh, J.J., 2008. Displacement loss on growth faults due to

EP

897

Smith, J., Durney, D., 1992. Experimental formation of brittle structural assemblages

sediment compaction. Journal of Structural Geology 30, 394-405.

AC C

896

SC

892

Tchalenko, J., 1970. Similarities between shear zones of different magnitudes. Geological Society of America Bulletin 81, 1625-1640.

Thorsen, C.E., 1963. Age of growth faulting in the southeast Louisiana. Transactions of the Gulf Coast Association of Geological Societies 13, 103-110.

908

Tingay, M.R.P., Morley, C.K., Hillis, R.R., Meyer, J., 2010. Present-day stress

909

orientation in Thailand's basins. Journal of Structural Geology 32, 235-248.

ACCEPTED MANUSCRIPT 910 911

Tron, V., Brun, J.-P., 1991. Experiments on oblique rifting in brittle-ductile systems. Tectonophysics 188, 71-84. Tvedt, A.B., Rotevatn, A., Jackson, C.A.-L., Fossen, H., Gawthorpe, R.L., 2013.

913

Growth of normal faults in multilayer sequences: a 3D seismic case study from

914

the Egersund Basin, Norwegian North Sea. Journal of Structural Geology 55,

915

1-20.

RI PT

912

Walsh, J.J., Bailey, W., Childs, C., Nicol, A., Bonson, C., 2003. Formation of

917

segmented normal faults: a 3-D perspective. Journal of Structural Geology 25,

918

1251-1262.

921 922

M AN U

920

Walsh, J.J., Nicol, A., Childs, C., 2002. An alternative model for the growth of faults. Journal of Structural Geology 24, 1669-1675.

Walsh, J.J., Watterson, J., 1988. Analysis of the relationship between displacements

TE D

919

SC

916

and dimensions of faults. Journal of Structural Geology 10, 239-247. Walsh, J.J., Watterson, J., 1991. Geometric and kinematic coherence and scale effects

924

in normal fault systems. Geological Society, London, Special Publications 56,

925

193-203.

AC C

EP

923

926

Wang, J., Liu, L., Chen, S., Shang, Y., 2011. Tectonic-sedimentary responses to the

927

second episode of the Zhu-Qiong movement in the Enping Depression, Pearl

928 929

River Mouth Basin and its regional tectonic significance. Acta Petrolei Sinica/ Shiyou Xuebao 32, 588-595 (in Chinese with English abstract).

930

Whipp, P., Jackson, C., Gawthorpe, R., Dreyer, T., Quinn, D., 2014. Normal fault

931

array evolution above a reactivated rift fabric; a subsurface example from the

ACCEPTED MANUSCRIPT 932

northern Horda Platform, Norwegian North Sea. Basin Research 26, 523-549. Wilson, P., Elliott, G.M., Gawthorpe, R.L., Jackson, A.L., Michelsen, L., Sharp, I.R.,

934

2013. Geometry and segmentation of an evaporite-detached normal fault array:

935

3D seismic analysis of the southern Bremstein Fault Complex, offshore

936

mid-Norway. Journal of Structural Geology 51, 74-91.

938

Withjack, M.O., Jamison, W.R., 1986. Deformation produced by oblique rifting. Tectonophysics 126, 99-124.

SC

937

RI PT

933

Wu, Z., Hu, Y., Zhong, Z., 2015. Cenozoic faults characteristics and regional dynamic

940

background of Panyu 4 sub-sag, Zhu- Depression. Journal of China University

941

of Petroleum 39, 1-9 (in Chinese with English abstract).

M AN U

939

Xu, X., Chen, S., Wang, F., Hu, K., Yu, S., Wang, X., Gao, Z., Liu, X., 2014.

943

Structural features and its impacts on hydrocarbon accumulation of Neogene in

944

Enping sag, Pearl River Mouth Basin. Geoscience 28, 543-550 (in Chinese with

945

English abstract).

948 949 950 951 952 953

EP

947

AC C

946

TE D

942

ACCEPTED MANUSCRIPT 954 955 956

RI PT

957 958 959

SC

960

M AN U

961 962 963

Figure captions

964

Figure 1. (A) Sketch map showing the location of the PRMB in the northern SCS

966

margin and subdivisions of structural elements in the basin. Image is from Google

967

Earth (accessed on 27 Nov 2017). EP, Enping sag; XJ, Xijiang sag; HZ, Huizhou sag;

968

LF, Lufeng sag; HJ, Hanjiang sag; SH, Shenhu uplift; PY, Panyu uplift; SD, Shunde

969

sag; BY, Baiyun sag. (B) Enlarged map showing the major half grabens and

970

associated boundary faults (red line segments) in the Enping sag (light blue-circled

971

area). Also shown are the locations of the 3D seismic horizon slices shown in Figs.

972

5-7 (black solid pentagon), and drill well A for synthetic seismogram in Fig. 3. NUZ,

973

Northern Uplift Zone; EX, Enxi low- relief tectonic high; PY, Panyu uplift.

AC C

EP

TE D

965

974 975

Figure 2. Simplified tectono-stratigraphic column in the PRMB (modified from He et

ACCEPTED MANUSCRIPT 976

al., 2017), with key horizons and seismic stratal units used in this study.

977

Figure 3. Synthetic seismogram of well A in the Enping Sag for horizon calibration.

979

See Fig. 1B for location of the well.

RI PT

978

980

Figure 4. Seismic profiles showing along-strike variations of cross-sectional

982

characteristics of the lower or syn-rift (bold red lines) and upper or post-rift (thin red

983

lines) fault segments. Index map at the bottom shows the locations of the seismic

984

profiles.

M AN U

SC

981

985

Figure 5. Coherence horizon slices at (A) horizon T30, (C) horizon T35, (E) horizon

987

T50, (G) horizon T60, (I) horizon T70, (K) horizon T80 and their corresponding line

988

drawings (B, D, F, H, J, and L) showing the spatial distribution and temporal

989

evolution of the EP18 fault zone. URZs, unbroken relay zones. Sub-figure (M)

990

assembles all of the fault traces in sub-figures B, D, F, H, J, and L together to show

991

the tempo-spatial distribution of the fault zone.

EP

AC C

992

TE D

986

993

Figure 6. (A) The coherence horizon slice at horizon T80 showing the characteristics

994

of the lower (syn-rift) fault system. The dark dots along the fault traces indicate the

995

intersections of the selected seismic profiles for throw measurements within the fault

996

zone. (B) The throw-distance (T-x) plots at horizons Tg, T80 and T70 showing the

997

lateral variations of fault throws of the lower (syn-rift) fault system.

ACCEPTED MANUSCRIPT 998

Figure 7. (A) The coherence horizon slice at horizon T50 showing the characteristics

1000

of the upper (post-rift) fault segments (f1 to f5). The dark dots along the fault traces

1001

indicate the intersections of the selected seismic profiles for throw measurements with

1002

the fault zone. (B) The throw-distance (T-x) plots at horizons T70 (triangles), T60

1003

(circles) and T50 (stars) showing the lateral variations of fault throws of the upper

1004

(post-rift) fault segments. Also shown are the locations of T-z plots for fault segments

1005

f1~f5 in Fig. 8.

M AN U

SC

RI PT

999

1006

Figure 8. Representative throw-depth (T-z) (black broken lines), EI and FAR (color

1008

histograms) plots in the centers of the post-rift fault segments, showing the

1009

throw-depth, EI and FAR variations of the post-rift fault segments: (A) and (F) fault

1010

segment f1; (B) and (G) fault segment f2; (C) and (H) fault segment f3; (D) and (I)

1011

fault segment f4; and (E) and (J) fault segment f5. Sub-figures (F) to (J) in the lower

1012

panel are the enlargements of sub-figures (A) to (E) in the upper panel. EI, expansion

1013

index. FAR, fault activity rate. See Fig. 7B for locations of these plots.

EP

AC C

1014

TE D

1007

1015

Figure 9. Throw-strike projections of upper (post-rift) fault segments f1 (A), f2 (B),

1016

f3 (C), and f4 (D). The cross marks denote the projected locations on the

1017

corresponding fault surfaces of the throw data points used for the contouring. Notice

1018

that the upper (post-rift) fault segments are linked downward with the underlain lower

1019

(syn-rift) fault system in their middle portions. The latter is indicated by the tightly

ACCEPTED MANUSCRIPT horizontal throw contours at the bottom of each sub-figure. Fault segments f2 to f4

1021

show horizontally elongated, elliptical throw contours centered on the throw maxima

1022

zone at horizon T50, suggesting that these upper (post-rift) fault segments nucleated

1023

at horizon T50 and grew by radial propagation of their tips. In contrast, fault segment

1024

f1 shows an asymmetric throw distribution with throws decreasing from horizon T70

1025

upward to horizon T20, indicating it was activated during T70-T20.

RI PT

1020

SC

1026

Figure 10. Time thickness maps of major stratigraphic intervals adjacent to the EP18

1028

fault zone. (A) the Late Miocene to Pliocene (T32-T20), (B) the late Early Miocene to

1029

late Middle Miocene (T50-T32), (C) the Oligocene to early Early Miocene (T70-T50),

1030

and (D) the Eocene (Tg-T70).

TE D

1031

M AN U

1027

Figure 11. A conceptual model for the evolution of the EP18 fault zone: (A) the

1033

growth of the lower (syn-rift) oblique-slip normal fault segments during the Eocene

1034

(Tg-T70) in response to the regional NW-SE extension; (B) a relatively dormant stage

1035

of fault activity during the Oligocene to early Early Miocene (T70-T50), with fault

1036

activity only seen at the western and eastern ends of the fault zone, where the upper

1037

(post-rift) fault segments, f1 and f7 to f9, developed in response to the regional N-S

1038

extension; (C) fault segment f1 continuously growing upward, and fault segments f2

1039

to f6 nucleated and propagated downward and upward during the late Early Miocene

1040

to late Middle Miocene (T50-T32) in response to the regional NNE-SSW extension;

1041

(D) continued growth of the upper (post-rift) fault system during the Late Miocene to

AC C

EP

1032

ACCEPTED MANUSCRIPT

EP

TE D

M AN U

SC

RI PT

Pliocene (T32-T20).

AC C

1042

CE ED

PT M AN US

CE

ED

PT

M AN US C R

CE

ED

PT

M AN US C R

AC C EP TE D M AN US C

RI PT

AC C

EP

TE D M AN US C

RI PT

CE ED

PT

M AN US

CC EP TE D

M AN US C

AC C EP TE D M AN US C

RI PT

EP TE

D

M AN U

AC C

EP

TE

D M AN US C

RI P

AC C EP TE D

M AN US C

RI PT

ACCEPTED MANUSCRIPT Highlights: 

A complex oblique-slip fault zone consisting of superimposed syn-rift and post-rift fault systems is investigated. The pre-existing syn-rift fault system shows a clear influence on the post-rift

RI PT



system.

Both fault systems show complicated growth and linkage characteristics.



A four-stage evolution model of the oblique-slip fault zone is presented.

AC C

EP

TE D

M AN U

SC