BasinResearch ChronologyandtectoniccontrolsofLateTertiary

Chronologyand tectonic controls of LateTertiarydeposition in the southwesternTian Shan foreland,NWChinaRichard V. Heermance,n Jie Chen,w Douglas W. Burbankn and Changsheng WangwnDepartment of Earth Sciences, University of California at Santa Barbara, Santa Barbara, CA, USAwStateKeyLaboratory of EarthquakeDynamics, Institute of Geology, ChinaEarthquakeAdministration, Beijing, China

ABSTRACT

Magnetostratigraphy from the Kashi foreland basin along the southern margin of theTian Shan inWestern China de¢nes the chronology of both sedimentation and the structural evolution of thiscollisional mountain belt. Eleven magnetostratigraphic sections representing �13 km of basin strataprovide a two- and three-dimensional record of continuous deposition since �18Ma.The distinctiveXiyu conglomerate makes up the uppermost strata in eight of11magnetostratigraphic sectionswithinthe foreland and forms awedge that thins southward.The basal age of the conglomerate varies from15.5� 0.5Ma at the northernmost part of the foreland, to 8.6� 0.1Ma in the central (medial) part ofthe foreland and to1.9� 0.2, �1.04 and 0.7� 0.1Ma along the southern deformation front of theforeland basin.These data indicate the Xiyu conglomerate is highly time-transgressive and hasprograded south since just after the initial uplift of the Kashi BasinThrust (KBT) at 18.9� 3.3Ma.Southward progradation occurred at an average rate of �3mmyear�1between15.5 and 2Ma, beforeaccelerating to �10mmyear�1. Abrupt changes in sediment-accumulation rates are observed at16.3and13.5Ma in the northern part of the foreland and are interpreted to correspond to southwardstepping deformation. A subtle decrease in the sedimentation rate above the Keketamu anticline isdetermined at �4.0Ma andwas synchronous with an increase in sedimentation rate further southabove the Atushi Anticline.Magnetostratigraphy also dates growth strata ato4.0, 1.4� 0.1 and1.4� 0.2Ma on the southern £anks the Keketamu, Atushi andKashi anticlines, respectively.Together, sedimentation rate changes and growth strata indicate stepped migration of deformationinto the Kashi foreland at least at 16.3, 13.5, 4.0 and1.4Ma. Progressive reconstruction of a seismicallycontrolled cross-section through the foreland produces total shortening of13^21km and migrationof the deformation front at 2.1^3.4mmyear�1between19 and13.5Ma, 1.4^1.6mmyear�1between13.5 and 4.0Ma and10mmyear�1 since 4.0Ma.Migration of deformation into the foreland generallycauses (1) uplift and reworking of basin-capping conglomerate, (2) a local decrease of accommodationspace above any active structure where uplift occurs, and hence a decrease in sedimentation rate and(3) an increase in accumulation on the margins of the structure due to increased subsidence and/orponding of sediment behind the growing folds. Since 5^6Ma, increased sediment-accumulation(�0.8mmyear�1) and gravel progradation (�10mmyear�1) rates appear linked to higherdeformation rates on the Keketamu, Atushi andKashi anticlines and increased subsidence due toloading from both theTian Shan and Pamir ranges, and possibly a change in climate causingaccelerated erosion.Whereas the rapid (�10mmyear�1) progradation of theXiyu conglomerate after4.0Ma may be promoted by global climate change, its overall progradation since15.5Ma is due to theprogressive encroachment of deformation into the foreland.

INTRODUCTION

Although many foreland basins contain upward-coarsen-ing stratigraphic successions (e.g. Covey, 1984; Burbank &

Raynolds,1988;Heller etal.,1988; Jordan etal.,1988;DeCel-les, 1994; Bullen et al., 2001; Homke et al., 2004; Jones et al.,2004; Uba et al., 2006), the controls on the coarseningtrends are commonly uncertain. An in£ux of gravel intosuch basins could result from (1) uplift and erosion of theadjacent mountain front (Armstrong & Oriel, 1965; Bur-bank et al., 1988), (2) enhanced precipitation or a more ero-sive climatic environment (Molnar, 2001; Zhang et al.,

Correspondence: Richard V.Heermance, USGS-GeologicDivi-sion,520N.ParkAve,Tucson,AZ85719,USA.E-mail: [email protected]

BasinResearch (2007) doi: 10.1111/j.1365-2117.2007.00339.x

r 2007 The Authors. Journal compilation r 2007 Blackwell Publishing Ltd 1

2001) that causes an increase in both the size and £ux ofsediment into a foreland, (3) continued erosion in the faceof tectonic quiescence and reduced subsidence within thebasin (Heller et al., 1988; Jordan et al., 1988; Flemings &Jordan, 1990; Burbank, 1992; Paola et al., 1992) or (4) achange in either the lithologic resistance of the source areaduring unroo¢ng or the distance of the depocentre fromthe active deformation front (DeCelles et al., 1991a; Paolaet al., 1992; Jones et al., 2004; Carroll et al., 2006). Addingto the complexity of sedimentation in foreland basins arethe multiple ways in which syntectonic folds and faultscut, uplift and often cause re-working of previously depos-ited basin sediments (Colombo, 1994). Such structuresreduce subsidence rates above actively growing folds, maycause ponding of Piggyback Basin (PBB) sediments ontheir hinterland £anks and create new source areas forbasin deposits (Burbank et al., 1996). Thus, lithofacieschanges observed at any single location within a basinmay only indicate local structural in£uences and notbasin-wide changes.

Within a foreland basin, conglomeratic facies are oftenlocalized near the mountain front where transversestreams debouch into the foreland basin and lose much oftheir sediment-transport capacity. In the simplest sce-nario, grain size should naturally decrease into the fore-land as a result of the decreasing river gradient within thebasin and as abrasion reduces the clast sizes (Paola, 1988;Sklar et al., 2006). However, variability in climate (dis-charge), subsidence rate or source lithology during theevolution of the foreland may cause the conglomeratefacies to prograde irregularly into the foreland on di¡erenttime scales (e.g. Paola et al., 1992). Thus, the geometry ofthe conglomerate facies is often described as a time-transgressive wedge that thins towards the foreland centreand may be perturbed by growth of structures within theforeland. If dispersion of gravel can be linked to initia-tion and growth of speci¢c structures, or if structuralshortening and gravel progradation rates match, they

suggest a tectonic coupling that is independent of climaticin£uences.

Both northern and southern margins of theTian Shanin Central Asia contain a late-stage basin-capping con-glomerate. In the southernTian Shan, the Xiyu conglom-erate caps the foreland sequence and has been inferred torepresent either the Plio-Pleistocene initiation of defor-mation of the southern Tian Shan (e.g. Huang & Chen,1981) or a change in either climate (Molnar et al., 1994) or aclimate-a¡ected erosion rate (Zhang etal., 2001). Althoughall of these factors may contribute to conglomerate aroundtheTian Shan, the age and primary depositional control ofthe Xiyu conglomerate remains controversial. Ages be-tween 4.8 and 1.0Ma have been assigned to the Xiyu con-glomerate at di¡erent localities around the Tian Shanbased on magnetostratigraphy, fossil assemblages and cli-matic interpretations for the time period (Huang &Chen,1981; Liu et al., 1996; Burch¢el et al., 1999; Chen et al., 2002,2007; Sun et al., 2004; Charreau et al., 2005).This age un-certainty of almost 4Myr obscures the exact timing of rockuplift or climate change, if any, that has driven conglomer-ate deposition. Moreover, a time-lag commonly exists be-tween a tectonic event in the hinterland and the arrival inthe foreland of coarse-grained facies related to that event(Jordan et al., 1988; Jones et al., 2004). In addition to thesimple rate of progradation, further lags can result fromdecreasing proportions of gravel fraction within the sedi-ment supply andvariable subsidence rates farther from thethrust front (Paola et al., 1992; Jones et al., 2004). As a con-sequence, the distribution of conglomerate facies acrossthe foreland is typically time-transgressive, and any singledate from an intrabasinal conglomerate is commonly apoor indicator of basin-wide history. Instead, a suite ofages that document gravel progradation are needed to testdi¡erent causative depositional models.

Located at the southwest corner of the Chinese TianShan, the Kashi Basin contains 46-km-thick Tertiarystrata (Fig. 1). The arid modern climate, when combined

Fig. 2aTarim basin

JunggarBasin

Bishkek

Urumqui

Kepintage Kuqa

70°E 45°N

90°E75°E 85°E80°E

95°E

(b)

40°N

Kashi

TarimBasin

Tian Shanian Shan

Tian Shan

Sanju

Kashi

India

Kasakh Platform

Tibet-QinghaiPlateau

0 5050 100100

km

0 50 100

km

(a)

Fig.1. Digital elevation model of theTian Shan region. (a) Shaded relief map of Central Asia showing the proximity of theintracontinental Tian Shanwith India and Tibet. (b) Shaded relief map with the Kashi Basin highlighted in white at the northwestcorner of theTarim basin.The study area (boxed region, Fig. 2a) and major geographic locations are labelled.

r 2007 The Authors. Journal compilation r 2007 Blackwell Publishing Ltd,Basin Research, 10.1111/j.1365-2117.2007.00339.x2

R.V. Heermance et al.

with active deformation along the southern margin of theChinese Tian Shan, has exposed numerous sections offoreland basin strata, making it an excellent locality to ex-amine relationships between sedimentation and tectonicdeformation.We studied14 sections throughout the basin,11ofwhich contain theXiyu conglomerate (Fig. 2a).We de-¢ne how the age, thickness and lithology vary at each sec-tion. We also use magnetostratigraphy to date growthstrata on the £anks of folds, thereby providing a compari-son of structural deformation with lithofacies variationand the appearance of conglomerate across the foreland.We use detailed mapping and magnetostratigraphy of theforeland stratawithin theKashiBasin to de¢ne: (1) the on-set of sedimentation; (2) the distribution of lithofacies;(3) the ages of lithostratigraphic units; (4) temporal andspatial relation of lithofacies to the initial formation andsubsequent growth of speci¢c structures within the fore-land and (5) the relationship of the Xiyu conglomerateprogradation to shortening along the southern marginof the Tian Shan. These conclusions are combined withpalinspastic reconstruction of a foreland cross-section tocompare the migration of the deformation front andstructural uplift with facies migration across the foreland.

BACKGROUND

Western China has a complex geologic history of tectonicsand erosion. Major collisional events occurred during theLate Devonian to Early Carboniferous and Late Carboni-ferous to early Permian, resulting in the complete amalga-mation of theTian Shan region of Central Asia (Windleyet al., 1990; Avouac et al., 1993; Carroll et al., 1995; Yin et al.,1998; Dumitru et al., 2001). At least three deformationalevents occurred duringMesozoic time as a result of the ac-cretion of continental blocks to the South Asian margin(Hendrix et al., 1992). From the late Cretaceous to the Oli-gocene, the regionwas an area of tectonic quiescence, withsporadic shallow-marine incursions (Allen et al., 1991). Aregional erosion surface was beveled across the Mesozoicand PalaeozoicTian Shan rocks during the earlyTertiary(Burbank et al., 1999; Abdrakhmatov et al., 2001), and thisunconformity is present in the Kashi Basin and across theKyrgyzTian Shan north of the study area.

Tertiary growth of theTian Shanwas a delayed responseto the continental collision between India and Eurasia(Molnar & Tapponnier, 1975; Burch¢el et al., 1999). Struc-tural analyses from di¡erent margins of the range have re-vealed that the magnitude of shortening varies spatiallyfrom the east to the west and from the north to the southacross the range, as do the timing and rates of deformation(Yin et al., 1998; Allen et al., 1999; Abdrakhmatov et al.,2001;Thompson et al., 2002; Scharer etal., 2004).The totalCenozoic shortening for both the eastern andwesternTianShan may be as much as 124 � 30 and 203 � 50km, re-spectively (Avouac et al., 1993). Recent geodetic (GPS)shortening rates indicate that the western Tian Shan isshortening at 21mmyear�1, while the eastern Tian Shan

is shortening at only 8mmyear�1 (Abdrakhmatov et al.,1996; Reigber et al., 2001; Wang et al., 2001). Much of theCenozoic deformation has likely localized along pre-exist-ing structural boundaries within the Palaeozoic and Me-sozoic rocks (Windley et al., 1990), and commencedduring the late Oligoceneêarly Miocene in the easternand southwestern Tian Shan (801^871 east longitude), asinferred from thermal-cooling data (Hendrix et al., 1994;Sobel & Dumitru, 1997; Sobel & Strecker, 2003), thecessation of marine sedimentation and in£uxes ofcoarse-grained strata (Yin et al., 1998). In contrast, initialdeformation in the KyrgyzTian Shan (�74^801 longitude)began in the middle Miocene (12^16)Ma based on coolinghistories, structural analysis and basin sedimentation(Sobel & Dumitru, 1997; Burbank et al., 1999; Bullen et al.,2001;Thompson et al., 2002).

KASHI BASIN STRATIGRAPHY

The Kashi foreland basin in NW China contains a thick(46 km) section of Tertiary strata that unconformablyoverlie the Palaeozoic and/or Cretaceous bedrock (Ballyet al., 1986) and provide a record of sedimentation sincethe onset of Neogene deformation in the Tian Shan.Although this entireTertiary section is not exposed at anyone location, we reconstruct the complete foreland strati-graphy using 14 sections (Fig. 2a) totalling more than15.5 km (�12.5 km of magnetostratigraphy) of stratigraphicthickness. Encroachment of deformation on the forelandhas caused uplift and erosion of the northernmost fore-land strata, resulting in exposure of the oldest syntectonicsediments in the north, whereas the youngest beds are stillaccumulating south of the active deformation front. Forsimplicity, we have separated the Kashi foreland into fourregions based on geographic location, bounding struc-tures and geologic similarities. From north to south, wedelineate the (1) hinterland, (2) northern foreland,(3) medial foreland and (4) southern foreland (Fig. 2a).The hinterland is de¢ned as the region north of the KashiBasin Thrust (KBT). The northern foreland consists ofthe region south of theKBTbut north of theTashipishakeAnticline. The medial and southern foreland are locatedsouth of the northern foreland basin and separated by thesyncline between the Keketamu and Atushi Anticlines(Fig. 2b). Stratigraphy from each region is synthesizedbased on (1) distinctive formation boundaries, (2) theunconformity along the base of the Tertiary strata and(3) magnetostratigraphic correlation between sections.

Stratigraphic thickness and descriptions, clast countsand palaeocurrent data were collected from the measuredforeland strata. Sections were measured with tape-and-compass or with the Jacob sta¡ and the Abney level. Clastcounts in conglomerates (matrix or clast lithology, colourand b-axis diameter,Table 1) were made every 10 cm alongtransects perpendicular to bedding to avoid bias due tosorting within individual beds. Data from �50 countsper site are shown as the percentage of limestone or

r 2007 The Authors. Journal compilation r 2007 Blackwell Publishing Ltd,Basin Research, 10.1111/j.1365-2117.2007.00339.x 3

Tian Shan foreland gravel progradation

0 5 10 15 20

Kilometers

measured sections (black=magnetostratigraphy, white=no magnetostratigraphy)

NORTH

Ahu Sisters East

Ahu Sisters West

Reservoir Bypass

Reservoir

Wenguri

AyakeqianaN. Qigailike N. Keketamu

Upper Boguzihe

Boguzihe

Ganhangou

TA9402Middle Atushi

Kashi TownKashi West

Kashi Anticline

Atushi Anticline

Keketamu AnticlineTashipishake

Anticline

Kashi BasinThrust

med

ial

sout

hern

nort

hern

fore

land

fore

land

fore

land

(a)

Qigailike

(b)

Muziduke Fault

Maidan Fault

Talas FerghanaFault

Kilometers

Kashi Basin Thrust

Keketamu Anticline

76°E 76.25°E 76.5°E75.5°E 75.75°E

40°N

39.75°N

39.5°N

40.25°N

40°N

39.75°N

39.5°N

40.25°N

Atushi Anticline

Kashi Anticline

Mutule Anticline

A

A’

Tashipishake Ant.

Mingyaole Anticline

Fig 2a

0 10 205

N

S

EW



sandstone clasts (Figs 3^5). Palaeocurrent measurementswere obtained from channel-axis directions, imbricatedclasts and cross-bedding.Typically, at each site, an averagedirection was calculated from 45 indicators. These dataare synthesized across a broad range of outcrops withineach stratigraphic lithofacies and are included with thestratigraphic logs (Figs 5^7).

Previous work from the western Tarim basin dividedthe Tertiary strata into more than nine formationsthat provided the basis for this study (Fig. 6a, Huang& Chen, 1981; Bally et al., 1986; Zhou & Chen, 1990; Liuet al., 1996; Jia et al., 2004). The stratigraphy within theKashi foreland,however, di¡ers geometrically and lithologi-cally from strata described elsewhere in the westernTarim, and therefore we avoid using the existing nomen-

clature unless a formation can be identi¢edwith certainty(e.g. Pakabulake Formation, Fig. 6). Moreover, this studydoes not attempt to describe the detailed sedimentologyof the Kashi Basin, although we do provide brieflithofacies descriptions and depositional environmentinterpretations (Table 2). Instead, we focus on the strati-graphic geometry and lithofacies boundaries, and placemagnetostratigraphic age limits on each formation.Overall, this study de¢nes seven Neogene and threepre-Neogene formations: undi¡erentiated Palaeozoic,Cretaceous, Palaeogene,Wuqia Group (Units A^C, Paka-bulake Formation), Atushi Formation, Xiyu conglomerateand PBB strata. These data provide a context for under-standing the basin evolution, particularly since the earlyNeogene.

Table1. Clast count data

SiteBasinlocation

Palaeomagneticsection

Stratigraphicheight (m)

Clast compositions

Totalclastcounts

Totalmatrixcounts

Meanclastsize(cm)

Standarddeviation

%Lime-stone

%Sand-stone %Other

‘Other’clasttype

2Y Northern Ayakeqiana 3 6 88 6 ch 32 16 2.0 2.04L Northern North Qigailike 230 17 84 0 30 15 1.8 1.61Y Northern Ayakeqiana 275 70 25 5 sh 44 24 3.3 2.95NK Northern North Keketamu 340 32 62 2 qtz 37 11 5.6 4.33Y Northern Ayakeqiana 700 80 6 14 sh, ba, ch, gr 50 18 3.9 4.26NK Northern North Keketamu 790 83 17 0 41 9 5.1 6.09K Medial Reservoir 1800 60 40 0 47 3 4.6 3.81K Medial Reservoir 2600 55 38 7 ch, ba 55 19 3.3 2.92K Medial Reservoir 2600 46 46 9 qtz, meta. 57 13 3.2 2.510K Medial Reservoir 2700 48 30 22 gr, sh 46 4 4.1 3.9600K Medial Reservoir Bypass 3400 60 40 0 75 25 2.6 3.34K Medial ASW 3440 86 14 0 44 6 2.4 1.215K Medial ASE 4600 93 7 0 45 6 2.6 2.816K Medial ASE 4600 91 7 2 gr 46 6 2.3 2.414K Medial ASE 4800 62 38 0 50 5 3.1 3.34A Southern Boguzihe 0 28 58 13 qtz, sh 53 30 1.8 1.23A Southern Boguzihe 200 12 72 16 ba, qtz, sh 81 42 2.6 1.61A Southern Boguzihe 500 51 40 9 ch, gr 35 9 2.5 1.82A Southern Boguzihe 500 25 54 21 gr 43 23 2.6 2.85A Southern Middle Atushi 550 52 38 10 ba, ch 42 9 2.1 1.613E Southern KashiWest 600 58 29 13 ch, sh, ba, gr 55 0 3.0 3.27A Southern Boguzihe 600 42 56 2 ch 52 5 4.5 3.311AC Southern Active channel NA 85 15 0 46 5 4.2 2.512AC Southern Active channel NA 76 24 0 50 0 4.3 3.42BU PBB PBB 300 38 72 0 34 13 3.5 4.81BU PBB PBB 320 51 44 5 gr 44 4 1.7 1.2

Clast type de¢nitions: qtz, quartz; sh, indurated shale; ch, chert; ba, basalt; gr, granite; meta., undi¡erentiated metamorphic; PBB, Piggyback Basin;ASW, Ahu SistersWest; ASE, Ahu Sisters East.

Fig. 2. Landsat image and geologic map of theKashi foreland. (a) Landsat image with the locations of measured stratigraphic sections(bold black andwhite line segments) within the Kashi foreland. In this study, the region is divided into northern, medial and southernbasins (shown on left of image) based on relative distances south of the Kashi BasinThrust. Seismic lineTA9402 (shown as dotted line)used for constructing the regional cross- section cutsNE^SWacross most of the structures.Outline of the image area is shown in (b). (b)Geologic map of the study area.Major faults and folds referred to in the text are indicated, and the distribution of theXiyu conglomeratethroughout the basin is shown.Location of cross-sectionAÂ0 (Fig.10c) is shown.Note the change fromprimarily south-vergent thrustfaults in the northern half of the study area to folds in the southern half of the region.



Pre-Miocene strata

Pre-Miocene strata in the study area are con¢ned to thehinterland, northern and locally in the medial basin re-gions (Fig 2a). These strata consist of Palaeozoic, Cretac-eous and locally Eocene sedimentary rocks and areeverywhere overlain unconformably by the Neogene stra-ta. This unconformity varies from an angular unconfor-mity between the Palaeozoic strata and overlyingCretaceous orTertiary rocks (Fig.7b), to a sub-parallel dis-conformity between the Cretaceous and/or Eocene rocksand the overlyingTertiary strata.This implies pre-Cretac-eous deformation in the region, but no deformation be-tween deposition of Cretaceous strata and the initiation ofNeogene uplift of theTian Shan. This unconformity is asharp contact that acts as a marker for correlation of basinstratawithin the foreland. Palaeozoic bedrock in theKashiregion consists primarily of Permian and Carboniferous

limestone and sandstone with rare conglomerate beds(Allen et al., 1999).Within the Kashi foreland, Palaeozoicstrata are exposed primarily north of the KBT (Figs 2band 7a).Where Palaeozoic rocks are exposed south of theKBT, they are intensely deformed (folded and faulted)and consist of me¤ lange blocks of limestone and sandstonein a shale matrix. Palaeozoic strata are interpreted as mar-ine rocks deposited in a regressive marine setting withinthe northern Kashi foreland (Carroll et al., 1995).

Cretaceous strata consist of distinctive pink, purple andorange siltstone, sandstone and conglomerate locally in-truded by dark-grey diorite of early Eocene age (Huanget al., 2006; E. Sobel, pers. comm.). These strata are�200m thick at the western edge of the study area but arenot preserved east of �75.71E.These strata are likely partof the Kezilesu Formation (Jia et al., 2004; Huanget al., 2006) and are interpreted as a braided river system(Miall, 1985; Zhou &Chen, 1990; Hendrix et al., 1992).

Pz unconf.

KBT

North Qigailike Section

North Keketamu Section

stra

tigra

phic

hei

ght (

m)

Ayakeqiana Section

0

−100

200

400

600

800

clay

f. sa

ndc.

san

dco

bble

clay

f. sa

ndc.

san

dco

bble

clay

f. sa

ndc.

san

dco

bble

KBT splay fault

n=8

n=13

Xiyu Conglomerate w/ interbedded shale

Cretaceous siltstone and sandstone

Paleozoic limestone and shale

thrust fault (KBT is indicated)

K

palaeocurrent data: unidirectional(gray) and bidirectional (black) current-direction indicators. n = number of data points per rose diagram.

KEY

K K

thrust fault

Cretaceousdisconformity

Cretaceousdisconformity

Paleozoicunconformity

Paleozoicunconformity

Pz unconf. stra

tigra

phic

hei

ght (

m)

limest

one

(ls)

sandst

one

(ss)

ss ls

clast composition

2Y

2Y

1Y4L

5NK

6NK

Wuqia Group Unit A

3Y1

0

200

400

600

800

50/50100%100%

Clast count data with point label. The graph at the right shows how composition generally changes up-section for the composite data from the northern foreland.

XiyuCongl .

XiyuCongl.

unit A unit Aunit A

Fig. 3. Northern foreland stratigraphy.The three sections are all located in the footwall of the Kashi BasinThrust (KBT) (Fig. 2a).A sharp unconformity is present between theTertiary and Cretaceous or Palaeozoic strata at the base of each section.The Xiyuconglomerate caps each section and is truncated on the north by theKBTatNorthQigailike andNorthKeketamu and by a splay fault ofthe KBTat the Ayakeqiana section (shown).Thus, these sections represent the minimum thickness for the Xiyu conglomerate at theselocations. Lacustrine and £uvial strata below the Xiyu conglomerate are interpreted as coarse-grained facies of Unit A in the lowerWuqia Group. Clast composition data are based on clast counts from conglomerate beds in each section, and show a compositeunroo¢ng sequence of sandstone- to limestone-dominated at each section (seeTable1 for clast count data).



In the west-central Kashi foreland �5 km south ofWenguri village (Fig. 2a), o40m of dark-green, ¢ssileshale and fossiliferous limestone crop out disconformablybetween theCretaceous andNeogene strata.The presenceof Ostrea (Turkostrea, Jia et al., 2004) within the limestoneplaces this unit within the Eocene, and it likely representspart of the Kalataer Formation (Fig. 6a, Yin et al., 2002;Jia et al., 2004). The base of this limestone is brecciatedand has sheared, foliated fabrics, suggesting that it mayhave been faulted into place.The dark-green, ¢ssile shalemay represent part of the marine Wulagen Formationfound �100 km south near Sanju (Fig.1a,Yin et al., 2002).

These strata, like the underlying Cretaceous units, areeverywhere sub-parallel but disconformable beneath theNeogene strata and are considered to have been depositedbefore Neogene foreland development in the region.

Wuqia Group

The Wuqia Group unconformably overlies the pre-Mio-cene strata and is separated into four distinct units withintheKashi foreland:Units A^C and the Pakabulake Forma-tion (Fig. 6a). These units are correlated stratigraphically

Ahu

Sis

ters

Wes

t

Ahu

Sis

ters

Eas

t R

eser

voir

Byp

ass

Res

ervo

ir

Qig

alik

e

n=9

n=13

n=13

Xiyu Conglomerate

Siltstone and sandstone

Gypsiferous Siltstone and Shale

Siltstone and Shale

Cretaceous Siltstone and Sandstone

Shale

0

200

400

600

800

1000

1200

1400

1600

1800

2000

2200

2400

2600

2800

3000

3200

3400

3600

3800

4000

4200

4400

4600

4800

5000st

ratig

raph

ic h

eigh

t (m

)

clay

fine

sand

coar

se

sand

cobb

le

clay

fine

sand

coar

se

sand

cobb

le

K

n=29

n=7

14K

16K15K

4K600K

10K

2K 1K

9K

0% 50% 100%

0−100

200

400

600

800

K palaeocurrent data: unidirectional (gray) and bidirectional (black) current-direction indicators. n is number of data points per rose diagram.

Unit BUnit

B

Unit C

Pakabulake Fm.

Xiyu Fm.

start growth strata

Keketamu compositeSection

3800–5000 m

3300–3800 m

2000–2600 m

1300 –1800 m

WuqiaGroup

lime

sto

ne

(ls)

san

dst

on

e(s

s)

ss ls

clast composition

50/50100%100%

Fig.4. Summary of medialforeland stratigraphy.The relativestratigraphic location andinterbasin stratigraphic overlap ofeach section are shown on the left(see Fig. 2a for section locations).All Tertiary strata below the Xiyuconglomerate are part of theMioceneWuqia Group, andformation names are listedadjacent to the strata column.Theboundary between the PakabulakeFormation and the Xiyuconglomerate is gradational andbegins �100m below the base ofthe massive conglomerate at2600m. Clast compositions showwell-mixed lithology in the lowerpart of the Xiyu conglomerate, butthe upper part is dominated bylimestone clasts. Count locationsare drawn at their appropriatestratigraphic levels within theXiyuconglomerate.



via tracing of marker beds between sections, and are brie£ydescribed here.

Unit A is con¢ned to the northern foreland and com-prises 300^350m of tabular siltstone and sandstone beds.

This unit is described from the Ayakeqiana, North Qigai-like andNorthKeketamu sections (Table 2, Fig. 3). Its baseis marked by a distinctive 2^10-m-thick orange-brown,well-rounded chert pebble-conglomerate and it grades

n=19

N=37

Middle Atushi (AM)

0

200

400

600

800

1000

1200

1400

1600

1800

2000

2200

2400

2600

2800

3000

3200

3400

3600

3800

3000

3200

3400

3600

3800

Boguzihe (A)

Ganhangou (GA)

0

200

400

800

1000

growthstrata

startactive fan

200

400

600

800

1000

Kashi-West (E)

KASHI ANTICLINE

ATUSHI ANTICLINE

Kashi-Town (T)

n=24

clay

f. sa

ndc.

san

dco

bble

clay

f. sa

ndc.

san

dco

bble

clay

f. sa

ndc.

san

dco

bble

clay

f. sa

ndc.

san

dco

bble

clay

f. sa

ndc.

san

dco

bble

600

stra

tigra

phic

hei

ght (

m)

stra

tigra

phic

hei

ght (

m)

0–3000m

growthstrata

growthstrata

growthstrata

correlationline

correlationline lim

est

one (

ls)

sandst

one (

ss)

ss ls

50/50

100%100%

13E

13E

5MA5MA7A7A

4A4A

3A3A

1A1A-2A 2A

clast composition(a)

(b)

(c)

Clast count data shown at their respective stratigraphic levels within each section and on the composite “clast composition section (c).

Atu

shi F

m.

n=13

Xiyu Conglomerate

Fluvial siltstone and sandstone

Gypsiferous Siltstone and Shale

Lacustrine Siltstone and Shale

Shale

palaeocurrent data: unidirectional (gray) and bidirectional (black) current-direction indicators. n is number of data points per rose diagram.

KEY

Fig. 5. Stratigraphy of the southern foreland. (a) Stratigraphy from three sections along the south £ank of theAtushi Anticline (Fig. 2a).Cross-bedding, ripple marks, channel-axis lineations and imbricated clasts indicate palaeocurrent directions change fromwest to eastin the £uvio-lacustrine Atushi Formation to southward in the Xiyu conglomerate in theMiddle Atushi section. (b) Stratigraphicsections correlated along the south £ank of the Kashi Anticline (Fig. 2a). Although the KashiWest section is located only13 kmwestalong the strike from the Kashi Town section, the Xiyu conglomerate is much thinner and located higher stratigraphically at KashiTown, illustrating the lateral variability of the conglomerate facies. (c) Clast composition data from point counts shown relative to othercounts collected at di¡erent stratigraphic levels. See text for details.



up section into 1^2-m-thick beds of siltstone and shalewith interbedded sandstone and conglomerate beds. Mudcracks and mottled pedogenic layers were observedwithinthe siltstone and shale. At the top of the section, Unit Agrades into massive cobble conglomerate.Together, thesestrata are interpreted as £uvial channel and overbank de-posits (e.g. Miall, 1996;McCarthy et al., 1997).

Unit B is described from only the medial foreland at theQigailike section (Fig. 4), but can be traced laterally formore than100 km across the entire study area.The base ismarked by the same o10-m-thick conglomerate unit asbelow Unit A, but quickly grades up into 300^400m oftablular,1^2-m-thick, darkbrown shale and siltstone beds.Bidirectional ripples and trace fossils are abundant. Up to5-m-thick gypsum beds occur locally near the base, andgypsum veins are present within the shale throughout thesection. Ripples and trace fossils, tabular siltstone bedsand the presence of gypsum are interpreted as recordinga lake fringe to shallow lacustrine environment in an aridsetting (Magee et al., 1995; Ort|¤ et al., 2003). These stratagrade laterally intoUnitA to the north, implying that theseunits were deposited contemporaneously, despite the dif-ferent grain size and depositional environment. Units Aand B may correlate with the Keziluoyi Formation de-scribed from other regions of theTian Shan (Fig. 6, Yinet al., 2002; Jia et al., 2004)

Unit C consists of 200m of interbedded shale and gyp-sum (Fig. 4) and is found in the medial foreland (Fig. 2a).

The lower100m consist of green and grey shale beds inter-beddedwith gypsum beds up to 0.5m thick. Shale beds aretypically ¢ssile and 0.5^1.5m thick. Overlying these strataare �100m of red and brown shale and gypsum, and bed-ding is typically o0.5m thick. These strata are laterallycontinuous and can be traced through the Qigailike andKeketamu sections (Fig. 6) as well as along strike for4100 km. These beds, however, are not found in thenorthern foreland, where their stratigraphic equivalent isthe Xiyu conglomerate. The Unit C lithofacies is similarto the Anjuan Formation within the Wuqia Group (Fig.6a, Bally et al., 1986; Zhou & Chen, 1990; Jia et al., 2004),but is only �200m thick in the Kashi foreland whereas itis described as 600m thick elsewhere in theTarim basin.These distinctive strata are used as a marker unit to corre-late laterally between sections within the medial foreland.The shale and gypsum beds are interpreted as ephemerallacustrine and playa deposits.The variation between gyp-sum and shale deposition is likely the response to changesin the lake level within the enclosed basin or distance fromthe lake edge (Vandervoort, 1997; Carroll & Bohacs, 1999),but the absence ofmud cracks and erosion surfaces suggestthat this formationwas deposited in a sub-aqueous envir-onment.

The Pakabulake Formation makes up the upper part ofthe Wuqia Group, and is the only unit that can be unam-biguously tied to the existing westernTarim basin nomen-clature (Figs 4 and 6, e.g. Bally et al., 1986).The formation

Artux Formation (<2000 m)

Wusu Conglomerate (<200 m)

Atushi Fm.(3500 m)

(b) KASHI FORELAND STRATA(a) WESTERN TARIM BASIN STRATA

XiyuConglomerate

PBB

hiat

us

Xiyu Conglomerate (0–2500 m)

Pakabulake Fm.(2000 m)

Anjuan Fm.(600 m)

Keziluoyi Fm.(430 m)

Bashibulake (300 m)Wulagen (100 m)

Kalatar (70 m)

PakabulakeFm. (2000m)

Unit C (200 m)

Unit B(350 m)

Unit A(270 m)

UndifferentiatedPaleozoic

UndifferentiatedPaleozoic

Cretaceous

Wuq

ia G

roup

Cretaceous

Pal

eoge

neN

eoge

ne

Miocene

OligoceneEocene

Paleocene

Pliocene

Pleistocene

Devonian-Permian

0

5

10

15

20

25

Wuq

ia G

roup

Wuq

ia G

roup

Kezilesu Fm (0-?)

~60

00 m

CO

RR

ELA

TIO

N O

F K

NO

WN

ST

RA

TIG

RA

PH

YW

ITH

TH

E K

AS

HI F

OR

ELA

ND

ST

RA

TIG

RA

PH

Y

age

(Ma)

~65

~250

Fig. 6. (a) Stratigraphy from the westernTarimBasin afterHuang&Chen (1981), Bally etal. (1986), Zhou&Chen (1990), Liu etal. (1996),Jia et al. (2004). (b) Stratigraphy of the Kashi foreland, and probable correlation withwesternTarim Basin stratigraphy. Formationthicknesses are based on mapping (this study).



consists of42000m of dark brown and grey siltstone andshale with 3^5-m-thick, tabular sandstone beds through-out the section. The contact with the underlying AnjuanFormation is de¢ned by the ¢rst distinctive 5^10-m-thick,¢ne-to-medium-grained sandstone bed above Unit C.Tabular sandstone beds contain 1-m-amplitude epsiloncross-beds. These sandstone layers within laminar silt-stone and shale are interpreted to represent £uvial chan-nels within an aggrading £oodplain (Miall, 1996;McCarthy et al., 1997; Bridge, 2003).

Atushi formation

The Atushi Formation conformably overlies the WuqiaGroup and consists ofo4000m of yellow^brown and tansandstone, siltstone, mudstone, gypsum (thin interbedso10 cm) and rare pebble conglomerate (increasing neartop of section). These strata are similar in appearance tothe Pakabulake Formation within the Wuqia Group, butare distinguishable by the lighter yellow^brown colour of

the silt and sand grains within the Atushi Formation (alsocalled Artux Formation at other locations around theTianShan). This formation is interpreted as a low-energymeandering stream depositional environment. One-metre-amplitude epsilon cross-beds within discontinu-ous tabular sandstone beds are channel facies that sweptacross the shale £oodplain deposits (Miall, 1996;McCarthy etal., 1997; Bridge, 2003).Mud cracks and inter-bedded gypsum and shale support sub-aerial depositionwith ephemeral lakes.The Atushi Formation is de¢ned asPlio-Pleistocene in age (Yin et al., 2002; Jia et al., 2004), buthas a gradational contact with the underlying PakabulakeFormation. The basal part of the unit is gypsiferous andmay act as the uppermost structural detachment layerwithin theTertiary strata.

Xiyu conglomerate

Eleven of 14 measured sections are capped or composedentirely of the Xiyu conglomerate (Fig. 2a).The Xiyu con-

NORTH SOUTHXiyu Congl.

PzPBB KBT

(a)

(b) (c)

(e)

unco

nfor

mity

bedding

bedding(d)

Wuqia Group

megabrecciaclast

Xiyu Congl.Atushi Fm.

70°unit APz shale

Xiyu Congl.

Keketamu Syncline

Keketamu Anticline

Xiyu Congl.unit CXiyu Congl.

Pakabulake Fm

Fig.7. Photos of theKashi Basin stratigraphy. All views are towards the east. (a) Proximal basin showing theKashi BasinThrust (KBT)placing the Palaeozoic strata above the Xiyu conglomerate.The thin piggyback Basin strata (PBB) lie unconformably above thePalaeozoic stratigraphy in theKBThanging wall. Field ofview is approximately 3 km. (b) Angular unconformity between the Palaeozoicshale andUnit A of theWuqiaGroup. Bedding is shown as the dashed lines. (c) Steeply dipping (701) outcrop ofXiyu conglomeratewitha limestone megaclast, which is �300m long � 50mwide.The clast is surrounded by proximal faciesXiyu conglomerate. Field ofviewis approximately1km. (d) Gradational contact of the Xiyu conglomerate with the Atushi Formation on the south £ank of the AtushiAnticline in the distal basin. Note the lighter-coloured siltstone interbeddedwith the darker conglomerate. (e) Panoramic view of themedial basin deposits.The base of the Keketamu magnetostratigraphic section starts near the core of theKeketamu anticline and runsnorth. Solid lines indicate formation contacts, and dashed lines are bedding planes.The resistant Xiyu conglomerate forms the hightopography in the core of the Keketamu syncline and along the southern £ank o¡ the Keketamu anticline.The stratigraphic distance(white arrows) betweenWuqiaGroupUnit C in the core of theKeketamu anticline to the base of theXiyu conglomerate is much greateralong the southern £ank than the northern £ank, demonstrating the wedge geometry of theXiyu conglomerate. Field ofview is �4 km.



Tab

le2.

KashiBasinStratigraphy

Form

ation(�thickn

ess)

Geographic

occurrence

Lithologicdescription

Interpretation

PiggybackBasin

strata:

PBBPiggyback

Basin:PBB(o

300m)

Hinterland,

northern

Shale,siltstoneand

conglomerate.Basalconglomerate10^20

mthick:angu

larclasts,clast

supp

orted.Light-brownandgreenshalewithinterbeddedsand

ston

ebeds.H

ardp

ancarbon

atelayersareinterbeddedin

thesemassive

siltston

ebeds.A

llbeds

arelaterally

discon

tinu

ousw

ithconglomeratelitho

facies.T

hetopofthesestrataaremassive

clast-supp

ortedpebb

leandcobb

leconglomerate

Fluvio-lacustrine,wedge-top

deposition.Braided

channeland

fan

depo

sition

withlocalized

lacustrine

environm

ents

Xiyuconglomerate

(0^3000m)

Northern,

medialand

southern

Massivepebble-cobbleconglomerate.Clast-sup

ported

andwell-sorted

grains.G

rain

size

decreasestow

ards

southinbasin.Trough

cross-beds

andim

bricated

clastsare

abun

dant.L

ensoidalconglomeratebeds

aretypically

2^10-m

-wide,o2-m-thick

and

trun

cateun

derly

ingstrata

Fluvial:braidedstream

andalluvialfandepo

sits

AtushiFormation

(3000^

4000

m)

Southern

Shaleand

siltstonew

ithinterbeddedsandstone.M

oderateyellowbrow

ntabu

larsiltston

ebeds.Interbedd

edgypsum

andripp

lesabun

dant.D

iscontinuo

ustabu

larsandstone

o5mthickcontains

1mam

plitud

ecross-beds

Fluvial:meand

eringstream

system

withinandaggrading

£oodplain.Mud

-cracksandgypsum

areevidence

for

depo

sition

inan

aridenvironm

ent

Pakabu

lake

Form

ation

(o2000

m)

Medial

Shaleand

siltstonew

ithinterbeddedsandstone.B

rownandgray

siltston

eandshale:1^2m

beds

withinterbeddedgypsum

,lam

inations,and

ripp

les.Upto4-m-thick

tabu

lar

sand

ston

ebeds

containcross-beds

attheirbase

andgradeup

intolaminarbeds.

Erosive

basewitho0.5mrelief

Fluvialand

Lacustrine:Fluvialchannelsystemwithinan

aggrading£o

odplain

Wuq

iaGroup

:UnitC

(250

m)

Medial

Gysiferousshale:Green-greyandredshalewith�40^50%

gypsum

.1^10cm

¢ssileshale

beds

interbeddedwith1^5-cm

-thick

gypsum

andsiltbeds.S

iltis¢n

ely(1mm)

laminated.L

owerhalfofform

ationisgreenish-greyandcontains

¢veto

eigh

tbedsof

gypsum

from

0.5to2.0mthick.Upp

erhalfisbrow

nish-red

andgypsum

isinterbeddedwithintheshalein1^2cm

beds,exceptfor

withinthelower10

mof

sectionwherethreeto

¢vebeds

o2.0-m-thick

areob

served

Shallowlacustrine/playadepositsw

ithinan

enclosed

basinand

aridenvironm

ent

Wuq

iaGroup

:UnitB

(300

m)

Medial

Fine-m

ediumsandstoneandsiltstone.Darkred-brow

nmicaceous

siltston

eandshale:

0.1^1.5m-thick,¢ssile,trace

fossils

abun

dant,con

tinu

ousp

lanarb

eds,minor

gypsum

,ripples

andmud

-cracksalon

gtopsurfaceofbeds

Lacustrine:Near-shore,shallowlacustrine

environm

ent

Wuq

iaGroup

:UnitA

(200^250

m)

Northern

Siltstoneandsandstone

w/in

terbeddedconglomerate:Po

orlyconsolidated,1^2-m

-thick

siltston

ebeds

withinterbeddedo1m

tabu

larsandstone

beds.O

ccasionallenticular

pebb

leconglomerate.Mud

cracks

and2^3cm

mottled

soilpeds

occurw

ithinthe

siltston

e

Fluvial:channeldepositsin

anaggrading£o

odplainor

ephemerallake

margin

Palaeogene

strata

Medial

Limestoneandshale:Fo

ssiliferous

(ostracods)lim

estone

brecciaun

conformablybelow

bright

green,¢ssileshale

Marine:shallowanddeep

marinefacies

Cretaceou

s:Kezilesu

(0^500

m)

Hinterland,

northern,

medial

Sandstone,siltstoneandpebbleconglomerate:Orang

eandred,0.5^1.0-m-thick

beds

ofsiltsone.L

enticular,discon

tinu

oussand

ston

eandconglomerate.Distinctiv

e,well-

roun

ded1^2cm

chertp

ebbles

arefoun

dwithintheconglomerate

Fluvial:Anastam

ozingstream

channelswithinan

aggrading

£oodplain

Und

i¡erentiated

Palaeozoicstrata

Hinterland,

northern

Limestone,sandstone,shaleandconglomerate.Fo

ssiliferous

dark-andlight-greymarine

limestone.Yellow-brown,well-indu

ratedsand

ston

ewithinterbeddedconglomerate.

Darkgreenandredshale.Occursas

anintenselydeform

edme¤ lange

Transgressive

marinestrata



glomerate makes up a distinct, mappable unit that can betraced across foreland structures from north to south,and laterally from east to west, across the basin (Fig. 2b).Stratigraphic thickness varies from at least 500m in thenorthern foreland, to42500m in the medial basin and de-creases to o100m in the southern foreland (Fig. 8a).Furthermore, lithofacies within the Xiyu conglomeratevary from north to south across the foreland. At its north-ern limit, theXiyu conglomerate contains large Palaeozoiclimestone megaclasts (o100m diameter, Fig. 7c) and thelithofacies are dominated by poorly sorted, matrix-sup-ported conglomerate with angular and sub-angular, peb-ble-to-boulder clasts (Fig 8b). In the medial basin,pebble- and cobble-sized clasts are typical, and bouldersare absent (Fig. 8c).The Xiyu conglomerate in the medialbasin contains well- sorted 0.5^1.0m lenticular beds that

are normally graded with abundant trough and planarcross-strati¢cation. A decrease in grain size and increasein rounding are observed in the southern foreland (Fig.8d), and the stratigraphic thickness varies from �500mat Middle Atushi (Fig. 5) to o100m along the southernedge of theKashi Anticline (Figs 5b and 8a). Furthermore,strata pinch out laterally in the southern foreland into the¢ne-grained Atushi Formation. At all locations, the con-tact between underlying strata and the Xiyu conglomerateis gradational, beginning with a few, 1^5-m-thick con-glomerate beds interbedded within ¢ner grained stratathat change within 100m to massive conglomerate withsparse siltstone lenses (Fig.7d).Conglomerate beds consistof pebble-sized, well-rounded, well- sorted, clast-sup-ported strata sorted into channels between 0.1 and 0.5mdeep (Fig. 8d).

50 cm50 cm

50 cm

Medial Foreland

Southern Foreland

Northern Foreland

stratanot-preserved

evaporites

Legend

elev

atio

n (k

m)

−5

−6

−7

−8

−4

−3

−2

−1

1

2

0

−4

−6

−7

−3

−2

−1

1

0

elev

atio

n (k

m)

−5

−6

−7

−8

−4

−3

−2

−1

1

2

0

PBB

southernforeland

medialforeland

northernforeland

base of Tertiary strata

~5x vertic

al exa

ggeration

gravelfront

−5

silt, sand, and evaporitesconglomerate depositionPiggy-back BasinPliocene Atushi Fm.

Xiyu ConglomerateMiocene Wuqia Grouppre-Tertiary strata

~25 km

N

top of

unit

BW

uqia

GroupAtus

hi Fm.

(a)

(b)

(c) (d)

Boguzi HeBashikeremu He

pre-Tertiarystrata

lateral pinch-out ofXiyu Conglomerate

Boguzihe

Kashi

PBB

Qigailike

Ayakeqiana

KeketamuComposite

Keke

tam

u A

nt.KBT

KBT

Atu

shi A

nt.

Kash

i Ant

.

Tash

ipish

ake

Ant.

Fig. 8. Stratigraphic architecture and Xiyu conglomerate lithofacies variability across the Kashi foreland. (a) Idealized block model ofthe Kashi foreland. Structural deformation has not been considered, and the stratigraphic sections are placed in their actual positionsrelative to each other along the east^west transect. Stratigraphic sections are pinned at the base of theTertiary strata for the northernand medial forelands, and at the surface in the southern foreland. Surface geology of the block model shows the generalized geology ofthe foreland, and the locations of gravel and evaporite deposition in the active alluvium. (b) Photo of typical Xiyu conglomerate withinthe northern foreland, with large, angular clasts. (c) Xiyu conglomerate outcrop in the medial basin showing 0.5^1.0m beds with sub-rounded cobbles. (d) Smaller clasts and improved clast sorting within the Xiyu conglomerate in the southern foreland.



TheXiyu conglomerate represents a time-transgressivewedge that prograded southward andwas derived from theuplifted range front in the north. The northern forelanddeposits are interpreted as proximal alluvial-fan deposits,with debris £ows and landslides interbeddedwith braidedchannel deposits, likely derived locally from the hangingwall of the KBT (e.g. DeCelles et al., 1991a, b). Uncommonlarge-scale, gravity-driven failures are interpreted to deli-ver these megaclasts into the proximal foreland from lo-cally high-relief topography (e.g. Blair & McPherson,1999). Clast composition in the northern foreland changesupsection from sandstone to limestone dominated (Fig. 3)and is interpreted to record the unroo¢ng of Cretaceousand Palaeozoic strata, beginning with the deposition ofrounded Cretaceous pebbles in the lower part of the con-glomerate and changing upward to limestone pebbles de-rived from the Palaeozoic strata (e.g. Colombo, 1994). Incontrast to the northern foreland, the thick pebble andcobble, clast- supported conglomerate that makes up themedial foreland is interpreted to encompass alluvial fanand braided stream facies that represent the downstreamequivalents of the northern basin proximal facies (Miall,1985). Clast composition varies from mixed (50% sand-stone, 50% limestone) near the base of the section to lime-stone-dominated near the top of the section (Fig. 4).Theconglomerate facies in the southern foreland are also in-terpreted as braided channel deposits; however, the smal-ler clast size and increased rounding suggest depositionmore distal from the mountain front (Paola et al., 1992).Clast composition changes up-section from 60 to 70%sandstone clasts at the base of the Xiyu to approximatelyequal proportions at the top of the section, in contrast tothe active channel of the Boguzihe River where the riverclasts are �75% limestone and 25% sandstone (Table 1).Interestingly, granite clasts are only found in the upper-most parts of the Xiyu conglomerate in the medial andsouthern foreland and within the active channel (Table 1),suggesting recent unroo¢ng of granitic basement withinthe source area.

PBB strata

Less than 400m of conglomerate, sandstone and siltstonelie unconformably on top of the Palaeozoic strata north ofthe KBT (Fig. 7a). Locally these strata overlie the steeplydipping Xiyu conglomerate in angular unconformity andare termed PBB strata. Green and light-brown siltstonein the lower half of the section grades up section into amassive pebble^cobble conglomerate (Fig.9).

The PBB strata are interpreted as restricted lacustrineand £uvial deposits. Lacustrine deposits are laterally dis-continuous, and in some places the entire section consistsof massive pebble and cobble conglomerate. These strataare unique in the Kashi Basin in that they constitute anunfaulted sequence of strata that post-dates most defor-mation in the northern foreland. Strata dip locally 4601north near their basal contact with the Palaeozoic bedrockin the south, but dips decline abruptly upward too101 at

the top of the section.The observed thin stratigraphic suc-cessions, local lacustrine facies and a buttress unconfor-mity that separates the PBB from underlying strata alongthe northern, more proximal basin margin, are consistentwith PBB sedimentation in a wedge-top setting (e.g. De-Celles &Giles, 1996). Clast compositionwithin the pebbleconglomerate is 40^50% limestone, 45^60% sandstoneand o5% granite (Table 1). These PBB conglomeratesmay be temporally related to the Pleistocene Wusu Con-glomerate that unconformably overlies theXiyu conglom-erate in other areas along the margins of theTarim basin(Sun et al., 2004; Sun & Liu, 2006). Sparse palaeocurrentdata indicate west-to-east current directions, parallel tostructural strike and contrastwith the north-to-south pa-laeocurrent directions within theXiyu conglomerate fromall other parts of the basin (Fig.9).

Basin geometry

Together, our stratigraphic descriptions from 14 sectionswithin the Kashi foreland allow us to construct a modelfor the basin’s sedimentary architecture (Fig 8a). By pin-ning the sections to the unconformity above the pre-Ter-tiary rocks at the base, and placing each section in its

n=13

fluvial conglomerate

KEY

lacustrine siltstone and sandstone

Paleozoic shale and limetsone

bi-directional paleocurrent data (ripples, cross beds, channel lineations). n is the number of data points per rose diagram.

clast count location

0

−100

200

400

clay

f. sa

nd

c. s

and

cobb

le

n=9

VGP Latitude−90 0 90st

ratig

raph

ic h

eigh

t (m

) 2BU1BU

Fig.9. Piggyback Basin stratigraphy and magnetostratigraphyfrom the Upper Boguzihe section (Fig. 2a).The section iscontinuous from the Palaeozoic angular unconformity at thebase to the top of the section, where the conglomerateonlaps the Palaeozoic basement at a buttress unconformity.Magnetostratigraphy shows only two polarity reversals:�200-m-thick reversed zone surrounded by normal polarity zones.



present-day position south of the KBT, the overall strati-graphic geometry becomes evident. The Wuqia Groupforms the stratigraphically lowest part of theTertiary stra-ta, and grades up-section into the Atushi Formation.TheXiyu conglomerate forms a time-transgressive, south-ward-prograding wedge that grades up-section into theactive alluvial fan surfaces or laterally into the southernforeland within the silt, sand and evaporites south of thedeformation front (Fig. 8a). By correlating the WuqiaGroup between the northern and medial basin sections,and by tying the Boguzihe and Keketamu sections to-gether by tracing bedding across structures, the combinedTertiary stratigraphic depth of the Kashi foreland is�6 km. Our observation of continuous, uninterruptedstratigraphic sections without obvious hiatuses from thebasal disconformity throughout the entirety of each sec-tion implies that these strata were deposited before defor-mation and are foredeep deposits. In contrast, the PBB arelikely wedge-top facies deposited after deformation of theforedeep strata.

STRUCTURE OF THE KASHI BASIN

The overall structural geometry of the Kashi Basin can bebroadly described as a ‘wedge thrust’ triangle zone (Banks& Warburton, 1986; Couzens & Wiltschko, 1996; Jones,1996; Chen et al., 2004), where back-thrusts and detach-ment folds form in front of foreland-migrating fault-bendfolds and duplex structures. The hinterland of the fold-and-thrust belt has been uplifted and thrust south abovethe more distal foreland features.This style of deformationhas caused the oldest and deepest part of the foreland stra-ta to be exposed along the northern limit, whereas theyoungest strata are observed along the £anks of the de-tachment folds con¢ned to theTertiary stratigraphy alongthe southern deformation front (Scharer et al., 2004; Chenet al., 2007). Although detailed structural descriptions oftheKashi foreland are beyond the scope of this paper, herewe summarize structures observed in the northern, med-ial and southern foreland as well as in the hinterland.

The South Tian Shan, Muziduke and Talas Ferghanafaults are all present north of the KBT (Fig. 2b), but arenot directly in contact withTertiary deposits in the Kashiarea.We focus our study on structures that deformTertiarystrata, although the presence of reset apatite ¢ssion-trackages within the hanging walls of the SouthTian Shan andMuziduke faults and their absence from the immediatefootwalls of these structures indicate these faults were ac-tive between 20 and 25Ma (Sobel et al., 2006).

The KBT is the major structure that cuts the northernforeland region. It juxtaposes Palaeozoic strata with theXiyu conglomerate along a sub-vertical thrust fault for atleast 80 km along strike (Fig. 2b). Thick (4500m) fore-land-basin strata are only found south of the KBT. Oneapatite ¢ssion-track cooling age from �100m north ofthe KBTsuggests initiation of faulting and rock uplift at18.9� 3.3Mawithin the hanging wall (Sobel et al., 2006).

TheTashipishake Anticline parallels the KBT �10 kmto the south.This anticline strikes east^west and has Pa-laeozoic strata exposed in its core and Cretaceous andNeogene strata along its £anks. The region between itssouthern £ank and theKBTde¢nes the northern foreland(Fig. 2a). The anticline is south vergent with a steep-to-overturned south limb and �501 -dipping north limb.We interpret theTashipishake Anticline as a fault-bend-fold (e.g. Suppe, 1983) with its upper detachment withinthe gypsiferous units at the base of the Tertiary. Farthersouth pre-Tertiary strata are rare, and the lower WuqiaGroup strata are exposed in the cores of anticlines, sug-gesting that the basal Tertiary gypsiferous horizons mayact as detachment levels within this thin-skinned style ofdeformation.

The distal basin contains the Atushi and Kashi anticli-nes: two active structures deforming Plio-Quaternarystrata in the southernmost part of the basin.TheKashi an-ticline is a well-developed detachment fold above a gyp-sum-rich detachment horizon at between 3 and 6 kmdepth (Scharer et al., 2004; Chen et al., 2007). The AtushiAnticline, located �10^15 km north of the Kashi anti-cline, is also a box-like anticline with steep limbs. Boththe Kashi and Atushi folds show a distinct north vergencewith steeper limbs on the north £anks: a striking contrastwith the south vergence of the faults and folds in the med-ial and proximal basin.

Surface mapping (Fig. 2b) is combinedwith seismic lineTA9402 (Figs 2a and10b) to construct a cross-section (AÂ0:Fig. 10c) across the central part of the Kashi foreland.Thetime-migrated seismic line image was provided to us byPetroChina Ltd. (Beijing, China), but due to its proprie-tary nature, we can only present a line-drawing here(Fig. 10b). Approximate depth conversions for the seismicsections rely on a velocity pro¢le for the Kashi forelandprovided by PetroChina (Fig. 10a). Velocities above anygiven travel time were averaged and then multiplied bythat time to determine the approximate depth. Depth-converted values were used to construct the balancedgeologic cross-section. Despite errors in our depth calcu-lations or in the velocity model, as well as unknown seis-mic migration errors over whichwe have no control, theseseismic lines provide valuable constraints on subsurfacegeometry that would be otherwise impossible to inferbased only on data at the surface.

Undeformed, sub-horizontal Tertiary strata are clearlyobserved below the southern half of seismic lineTA 9402(Fig. 10b). We interpret a strong re£ector at 4.8^5.0 s(�6.2^6.5 km depth) to represent the velocity contrast be-tween theweak, gypsiferous shale of the basalTertiary sec-tion and the underlying Palaeozoic limestone orCretaceous sandstone. The calculated depth �6.5 km tothis re£ector is consistent with our composite basin depthas determined from our measured sections (Fig.8a), givingus con¢dence that this horizon represents the base of theTertiary foredeep strata.

Line-length shortening across all the major mappedstructures yields a total, minimum shortening of 13^21km



(Heermance, 2007). This estimated shortening range isdue both to unknown geometry of the folds that have beeneroded and to a general lack of hanging wall cuto¡s acrossfaults. These estimates nevertheless allow us to recon-struct the basin architecture (Fig.10d).

MAGNETOSTRATIGRAPHY

We present magnetostratigraphic results from 11 of 14measured sections spaced throughout the Kashi Basinand spanning 412500m of stratigraphic thickness (Fig.2a, Table 3). New magnetostratigraphic data from sevensections [Qigailike, Reservoir, Reservoir Bypass, Ahu Sis-ters West (ASW), Ahu Sisters East (ASE), Ayakeqiana andPBB] and the lower1600m of the Boguzihe section are de-scribed here, and data from the Kashi West, Kashi Town,Ganhangou and the upper half (1800m) of the Boguzihesections are described in detail elsewhere (Chen et al.,

2002, 2007).The new magnetostratigraphic synthesis pre-sented here provides a detailed temporal framework inthree-dimensions for late Tertiary deposition and defor-mation of theKashi foreland basin.The spatially dense ar-ray of dated sections and the total stratigraphic thicknessspanned by them de¢ne this as one of the most extensivechronologic data sets for an actively deforming forelandbasin that has been published to date.

Samplingmethods and analysis

Palaeomagnetic samples were collected from1441sites (se-dimentary layers) spaced throughout the 11 di¡erent sec-tions (Table 3). Sample spacing averaged 10m overall andwas less where suitable sampling sites were present, suchas at Boguzihe (7.6m), Ayakeqiana (8.1m) and Qigailike(4.8m). Sites were more widely separatedwithin poorly in-durated and friable strata, e.g. the Reservoir section, andwithin massive conglomerates, e.g. Reservoir Bypass and

0−2−4−6−8−10

0−2−4−6−8

−10 Pre-Tertiary

STS

F

KBT

AtushiAnticline

KeketamuAnticline

Muz

iduk

e F.

.

5 km

dep

th b

elow

surface (km

)

Kashi Anticline.05 km

Total shorteningsouth of KBThanging wall:

13–21 km

Atushi Anticline4–5 km

Keketamu Anticline4–5 km

KBT2–7

Tashipishake Ant. & KBT imbricates

3–4 km

shorteningestimates

0−2−4−6−8

20−2−4−6−8

2 dep

th b

elow

surface (km

)

South TianShan Fault

Muziduke Fault

(e)

hinterland northern medial southern

Wuqia GroupAtushi Fm.

Xiyu Congl.

detachment foldsimbricate thrust faults

PBB

KBT

faultbase Tertiary

Xiyu Conglomerate

Atushi Fm.

Wuqia Group

magnetostratigraphicsection

5.04.03.02.01.0

0

6.05 km

base Tertiary

Pz strata

WuqiaGroup

AtushiFm.

Seismic line TA 9402

0

2.0

4.8

8.4

(a) Kashi Foreland velocity profile

(b) line drawing of seismic line TA 9402

(c)

(d)

~depth(km)

2-way travel time (seconds)

A A’

KashiAnticline

AtushiAnticline

KeketamuAnticline

TashipishakeAnticline Kashi

Anticline

section key

Fig.10. (a) Seismic velocity pro¢le from the Kashi Basin (courtesy of PetroChina). (b) Line drawing of the time-migrated PetroChinaseismic lineTA9402 shown in Fig. 2a. (c) Geologic cross-section AÂ0 based on ¢eld mapping and depth migration of seismic lineTA9402 using the velocity model for the Kashi Basin. Stratigraphic boundaries are based on outcrop locations within the basin, andinterpreted stratigraphic depths from the cumulative basin section.The locations of the magnetostratigraphic sections are shown as theboxes above the cross-section.The base of theTertiary shows up as a strong re£ector on the seismic line at �6.5 km depthwithin thebasin, consistentwith our totalledmeasured stratigraphic thickness forTertiarydeposits. (d) Palinspastic reconstruction of deformationwithin the Kashi foreland south of the Kashi BasinThrust (KBT). (e) Line-length shortening estimates based on surface mapping foreach structure observed south of and including the KBT. Shortening ranges are due to uncertainties in the fault cut-o¡s and in thesub-aereal geometry of the folds. Details on shortening calculations are in Heermance (2007).



Tab

le3.

Magnetostratigraphicdatafrom

theKashiBasin

Basin

location

Site

pmag

Total

stratigraphic

thickn

ess

(m)

Total

sites

collected

Total

quality

sites

Total

quality

samples

Average

sample

spacing

(m)

Sample

spacing

standard

deviation

(m)

Magneto-zon

es

Interpreted

timespan

(Myr)

Calculated

timespan

(Myr)Œ

Standard

error

(�2s

,Myr)

Sedim

entation

rate

(mMyr�1 )

Sedimentation

ratestandard

error

(�2s

,mMyr�1 )

Southernbasin

Boguzihe

Yes

3225

425

403

449

7.69.10

194.0

4.3

2.0

720^

820

20Ganhang

oun

Yes

815

4440

4120.4

NA

41.1

0.8

0.9

NA

NA

Kashi-W

est

Yes

1250

124

117

139

10.7

17.70

72.0

1.5

1.2

600

130

Kash-Tow

nYes

795

5251

6215.6

6.10

101.7

2.6

1.6

520

20MiddleAtushi

No

972.5

85NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

Reservoir(K

eketam

u)Yes

2328

247

202

286

11.5

6.3

396.8

10.1

2.9

130^

430

30

Medialbasin

ReservoirBypass

Yes

688

4638

4818.1

24.6

91.6

2.4

1.4

430

10Ahu

SistersW

est

Yes

573

7329

2919.8

185

1.3

1.00.9

430

10Ahu

SistersE

ast

Yes

1168

5328

2841.7

33.7

33.3

2.1

0.6

270

70Qigailike

Yes

905

215

189

189

4.8

1.922

4.7

5.0

2.1

90^250

20Keketam

uCom

posite

Yes

3800

544

421

515

9.810.5

6412.2

15.8

3.7

130^

430

Na

Northernbasin

North

Qigailike

No

480

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

North

Keketam

uNo

527

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

Ayakeqiana

Yes

655

8881

928.1

5.4

133.5

3.0

1.6

70^280

30Hinterland

Upp

erBoguzihe

Yes

290

7460

624.7

33

0.5^

4.0?

0.4

0.6

100^

800?

NA

Totals

PMAGonly

Yes

12692

1441

1238

1425

10.3

NA

134

NA

NA

NA

NA

NA

Colum

nheadingdescriptions

aredescribedhere.

nNew

datafrom

Boguziherepresentsthelower1600

mofthesection.The

upper1625m

ofstratigraphicthickn

essandtheGanhangou

dataarepresentedinChenetal.(2002).

w CalculatedafterJoh

nson

andMcG

ee(1983),average

timeofmagentozones;210ka

(Neogene);230ka

(Plio

cene).

NA,datano

tapp

licableor

available.

Basinlocation

,relativelocation

ofmagnetostratigraphicsectiontotheK

ashiBasinThrustasd

e¢nedinthetext;Site,Stratigraph

icsectionwherethesamplesarelocated;To

talstratigraph

icthickn

ess,stratigraphicthickn

ess

betweenthelowestand

highestm

agnetostratigraphicsamples.A

ctualm

easuredsectionisusually

longer

butthe

base

andtopofthesections

wereno

tsam

pled

dueto

either

faulting

orpo

orsamplinglocalities;To

talsites

collected,totalnu

mberofsites(bedd

inghorizons)sam

pled

with

inthesection;To

talqualitysites,totalnum

berofsitesyieldingclearp

olarity

directions

(described

intext);To

talsam

plesprocessed,The

totalnum

berofsam

ples

processedatthelaboratory;Totalqu

ality

samples,S

amples

thatyieldun

ambigu

ousno

rmalor

reversed

polaritydirections;A

verage

samplespacing,averagesamplespacingbetweenthestratigraphically

highestand

lowest

sampleswith

inthesection;Samplespacingstandard

deviation,on

estandard

deviationfrom

themeandistance

betweentwoadjacent

samplesspacingbasedon

thepo

pulation

ofdistancesb

etweenanytwoadjacent

samples;

Magnetozones,thetotalnum

berofm

agnetozones(or

partialm

agnetozonesiflocated

atthetopor

bottom

ofthesection)determ

ined

with

ineach

section;Correlatedtimespan,timespannedby

stratigraphicsectionbasedon

ourcorrelation

with

inthegeom

agneticpolaritytimescale(G

PTS)

(Lou

rens

etal.,2005);Calculatedtimespan,expected

timespan

calculated

foreachsectionbasedon

thestatisticalm

etho

dofJohn

son&McG

ee(1983).A

lldata

arecalculated

basedon

unifo

rmdistribu

tion

ofsamplesitesexcept

fortheReservoirBy

pass,A

huSistersE

ast,andAhu

SistersW

est(italicized)thatw

erebasedon

rand

omsampledistribu

tionsandtheirtime-span

wasde-

term

ined

graphically

(Johnson

&McG

ee,1983,Fig.6).T

heselatterthreesections

have

thehigheststand

arddeviationofsamplespacingdu

etothepo

orqu

ality

andirregularityofsamplesitesw

ithintheX

iyuconglomerate.The

KashiWestsection

also

hasahigh

standard

deviationofsamplespacingdu

etotwowidelyspaced

(�140m)sam

ples

atthevery

topofthesectionwith

intheXiyuconglomerate.Otherwise,allsection

s,includ

ingmostofthe

KashiWestsection,had

relativ

elyun

iform

samplespacingthroughout;S

tand

arderror,95%

con¢

dencelevelfor

thecalculated

timespan

ofeach

sectionafterJohnson

&McG

ee(1983);S

edim

entation

rate,N

on-com

paction

correctedsedimentation

ratecalculated

fora

sectionbasedon

thecorrelated

magnetostratigraphy

andstratigraphicthickn

ess(Fig.16);Se

dimentation

ratestandard

error,95%

con¢

denceintervalforthesedimentation

ratedatabasedon

theleastsqu

ares

linearregression

throughdatapo

intsshow

ninFig.16.W

heremorethan

oneregression

wasused

to¢t

non-lin

eard

ata,theerrorsrepo

rted

arethemaxim

umvaluebetweenthedi¡erent

regression

s.



ASWand ASE, where samples were collected opportunis-tically from irregularly spaced, thin-siltstone and ¢ne-sandstone beds.Two or three oriented cores per site wereobtained in situ using a gas-powered drill with a 2.5-cm-diameter core bit.The core-plate orientation and beddingattitude of the site were determined by a magnetic com-pass and, when possible, a sun-compass.

Sample spacing was deemed su⁄cient to capture mostof the palaeomagnetic reversal events preserved in theKashi Basin strata for the following reasons. Previousworkwithin the Kashi foreland documented Plio-Pleistocenesediment accumulation rates of 600^800mMyr�1 (Chenet al., 2002, 2007). Miocene and Pliocene sediment accu-mulation rates from elsewhere within the northernTarimBasin vary between 200 and 700mMyr�1 (Sun etal., 2004;Charreau et al., 2006; Huang et al., 2006), although the lateOligocene and earlyMiocene rates mayhave been as low as70 mMyr�1 (Huang et al., 2006).The average timespan ofa magnetozone during theNeogene is 0.21Myr, but rangesfrom 0.01to1.07Myr (calculated fromLourens etal., 2005).Thus, by assuming that most of our strata are con¢ned tothe Pliocene and Miocene and that deposition rates varybetween 200 and 800mMyr�1, the 10m average site spa-cing would span between 0.012 and 0.05Myr.This spacingwould sample at least four times an average magnetozonetimespan of 0.21Myr. Furthermore, all zones 40.1Myrshould be sampled at least twice.

Magnetic polarities were determined for each analysedsample using standard methods described in Appendix A.One specimen per sampling site or stratigraphic level wasinitially measured for natural remnant magnetization,then subjected to alternating- ¢eld (AF) demagnetizationup to 100G in 15 or 25G steps to remove low-coercivitymagnetizations, and subsequently treated with stepwisethermal demagnetization using 6^12 demagnetizationsteps between 50 and 680 1C, typically in (1) 150 1C stepsup to 300 1C, (2) 100 or 501 steps up to 500 or 550 1C and(3) 30 or 201 steps up to 680 or 690 1C (Fig.11). At each site,a second and third specimenwere demagnetized if the ¢rstyielded unstable or overprinted directions or if a magne-tostratigraphic interval of normal or reversed polarity wasde¢ned by only one specimen.Thus, all magnetozones arede¢ned by at least two samples with consistent polarity.

Many previous magnetostratigraphic studies employ amethod that ranks sites based on two or more samples persite (e.g. Johnson etal.,1982; Schlunegger etal.,1997;Kempfetal,1999;Ojha etal., 2000;Chen etal., 2002).We do not uti-lize this method for the following reasons: (1) all samplesare run through the magnetometer right-side up and up-side down for each thermal demagnetization step to deter-mine whether or not the sample gives consistent readings;(2) the high accuracy of the computer-controlled magnet-ometer and consistency of the sample changer at the Cal-tech lab reduce much of the laboratory uncertainty inmeasurements that plagued early studies; (3) our methodlikely sampled each magnetozone at least twice (but fouron average) and (4) all single-site reversals are checked forconsistencywith samples fromboth the same stratigraphic

level and the adjacent sites stratigraphically above and be-low. For these reasons, we did not consider it necessary torun multiple samples from each site unless the sitewas at areversal boundary or de¢ned its own magnetozone.

Magnetozones and correlationwith thegeomagnetic polarity timescale (GPTS)

Within the 11 sampled sections in the Kashi Basin, 134magnetozones were identi¢ed (Figs12 and13,Table 3). Re-versal boundarieswere drawn at the midpoint between twosuperposed sites with opposite polarities. Thus, onesource of error in the age assigned to strata associatedwithany individual reversal is a function of the sedimentationrate and of the stratigraphic distance between the sites de-¢ning the reversal. Correlation of the magnetostratigraphy[magnetic polarity timescale (MPTS)] of each section tothe GTPS of Lourens et al. (2005) was assisted by the con-struction of longer composite magnetostratigraphic sec-tions based on stratigraphic ties between individualsections.These composites provide distinctive magnetos-tratigraphic patterns that, given a few assumptions aboutthe age or timespan of the strata, can be correlated withthe GPTS.

Southern foreland results

Correlation of southern foreland deposits to the Plioceneand Pleistocene epochs is based on: (1) ostracods withinthe Boguzihe section that place them within the Pliocene(Jia et al., 2004); (2) the observation that the section gradesupwards into the active Quaternary fan surfaces withoutobvious major unconformities or hiatuses, implying thedeposits are young (Fig. 5) and (3) the Atushi Formation,which dominates the section stratigraphy, has traditionallybeen consideredPliocene (Sun&Liu, 2006). Following theprevious correlations ofChen etal. (2002) and using the ex-tended magnetostratigraphy of the 3700-m-thick AtushiAnticline section, we interpret the 23 magnetozones de-¢ned at the Atushi Anticline section to span from�5.5Ma to �1Ma (Fig.12).This correlation to theGPTSappears unambiguous except for short magnetozones thatare based on single sites, and thus ignored, in the upperpart of each section.

Strata exposed in the Kashi Anticline correlate viacross-section AÂ0 (Fig. 10c) with the upper part of theAtushi Anticline strata and are, therefore, Pliocene oryounger. The magnetostratigraphic sections from theKashiAnticline (KashiTown,KashiWest) are inter-corre-lated by using air photos to trace beds between them(Fig.14, Chen etal., 2007: Fig. 2a).These sections are domi-nated by reversed polarities and their upper parts corre-late with the reversed zone between1.6 and 0.78Ma on theGPTS (Lourens et al., 2005). The base of the Kashi Westsection extends to �3Ma. The Kashi Town section wassampled up to the active fan surface, where its magnetos-tratigraphy is interpreted to terminate in the Brunheschron ato0.78Ma (Fig.12).



Progressive limb rotation of the £anks of the Atushi andKashi anticlines during deformation produced growthstrata that display fanning dips and angular unconformities(Scharer et al., 2006; Chen et al., 2007) that are dated at 1.4(KashiWest) and1.07Ma (KashiTown) along theKashiAn-ticline, and at1.4 (Boguzihe) and1.2 (Ganhangou)Ma alongthe Atushi Anticline (Fig. 12). Two ages of growth strata

from the Kashi Anticline provide evidence for lateral foldpropagation at 40mmyear�1 (Chen et al., 2007).Moreover,these ages only apply to the exposed and dated parts of thefold.The growth-strata ages used to de¢ne fold initiationfor both the Atushi and Kashi anticlines are the older oftwo separate such ages for each fold; only one growth-strataage, however, was obtained from the Keketamu anticline.

1

2

4

3

magnetizing field (Oe)0 1000 2000 3000 4000

magnetizing field (Oe)0 1000 2000 3000 4000

(e) K70.3: IRM

IRM

(10

−03

G)

IRM

(10

−03

G)

1

2

4

3

(c) K70.3 (746.1m): Pakabulake

REVERSEDDecl: 179.2Incl: −24.1

NORMALDecl: 331.2Incl: 47.9

NORMALDecl: 318.3Incl: 58

NORMALDecl: 343Incl: 49.9

NORTH

NORTH

EAST

SOUTH

SOUTH

SOUTH

WEST

(a) L32.1 (248.1 m): Bashibulake (b) K829 (199.1 m): Xiyu Conglomerate

(f) K188.1: IRM

(d) K188.1 (1918.4m): Pakabulake

WEST

NORTHSOUTH

WESTWEST

TC HTC

Fig.11. Examples of palaeomagnetic results. (a^d) Results from stepwise alternating- ¢eld and thermal demagnetization of selectedoriented samples from the Keketamu composite (b^d) and Qigailike (a) sections. Zijderfeld and equal area plots are shown side by side,with some of the major steps labelled. Examples of the low-temperature component (LTC) and high-temperature component (HTC)ofthe characteristic magnetic remanence (ChRM) are shown in (a). (e^f) Isothermal remanent magnetism (IRM) within an increasingmagnetic ¢eld of the samples shown in (c) and (d), respectively. Both magnetite and hematite are interpreted as the carriers of ChRM forboth samples.



Medial foreland results

Magnetostratigraphy from the medial basin is based on¢ve sections: Qigailike, Reservoir, Reservoir Bypass,ASW and ASE (Fig. 13). Despite 78 total magnetozones

de¢ned by these sections, correlation of any individual(o1000m) section with the GPTS is ambiguous due to alack of radiogenic ages or biostratigraphy to pin themagnetostratigraphy to a speci¢c reversal sequence.To help remedy this ambiguity, a 3800-m composite

age (Ma)−90 0 90

−90 0 90−90 0 90

−90 0 90

0

1000

2000

3000

3600B

oguz

ihe

Sec

tion

Gan

hang

ou S

ectio

n

VGP Latitude (°)

VGP Latitude (°)

VGP Latitude (°)

VGP Latitude (°)

GPTS(Lourens et al., 2005)

?

Kashi West Section

Xiyu Conglomerate

Atushi Fm. siltstoneand sandstone

magnetostratigraphic correlation(normal polarity zones shaded)

KEY

Kashi Town Section

KASHI ANTICLINEATUSHI ANTICLINE

lithology logs

magnetostratigraphy

growthstrata

southern foreland magnetostratigraphy

lithology

lithology

lithology

0

1

2

3

4

5

6

0 10kmKashi Town

BoguziheMiddle Atushi

Dis

tal B

asin

Pro

xim

alB

asin Medial

Basin

Ganhangou

Kashi West

Kashi Anticline

Atushi Anticline

Keketamu

Anticline

KBT

stra

tigra

phic

hei

ght (

m)

A1n

T3n

T4n

T6n

T4n

T3nT2n

T5nT5n

T1n

A2nA2n

A3n

A4n

A5n

A6n

A7n

A8n

A9n

A10n

growth strata

~1.2 Ma

~1.4 Ma

1.4 Ma

1.07 Ma

Fig.12. Correlation of the southern foreland magnetic polarity timescale (MPTS) with the Pliocene Epoch on the geomagneticpolarity timescale (GPTS) of Lourens etal. (2005).Data from the upper1600m of theAtushi Anticline and from theKashi Anticline arefromChen etal. (2002, 2007), respectively. Individual sample-site coordinates are listed inHeermance (2007).Only zoneswith more thanone sample are used to de¢ne a zone, and black andwhite zones indicate normal and reversed polarity, respectively. Inset shaded reliefmap shows the section locations within the southern foreland region (highlighted).The contact between the Xiyu conglomerate andAtushi Formation, as well as growth-strata locations, are shownwithin the lithology logs adjacent to each plot of magnetostratigraphy.



magnetostratigraphy containing 64 of the 78 magneto-zones was constructed by tracing and correlating beds be-tween the Reservoir, Reservoir Bypass and the ASWsection (Fig. 14). The Qigailike data are correlated withthe base of the Reservoir section by the distinctive markersandstone bed at the base of the Pakabulake Formation.The Qigailike data are used in lieu of the Reservoir datahere because they have a denser (�5 vs. 12m) sample-sitespacing and extend to the base of the Tertiary strata.Although the stratigraphic position of the ASE sectionthat caps the medial basin sequence is well de¢ned(Fig. 14), it is not included in the composite section(Fig.15) because the sample spacing through the conglom-erates is too large.

Correlation of the medial basin composite with theGPTS of Lourens et al. (2005) is based on the followingthree criteria: (1) The Anjuan and Pakabulake Formationswithin theWuqia Group have previously been assigned tothe middle and late Miocene (Yin et al., 2002; Jia et al.,2004).This interpretation is consistentwith their positionunconformably above the biostratigraphically dated LateEocene Kalataer Formation. (2) The uppermost medialbasin composite section may be traced laterally acrossfolds into the distal basin deposits at the Boguzihe section(Fig. 15a), making them Pliocene in age (this study, Chenetal., 2002).The top of the medial basin composite section,however, must be signi¢cantly older than Pleistocene, be-cause the 1000-m-thick ASE section is located above the

0

1

2

3

4

5

6

7

8

9

1010

1111

1212

1313

1414

1515

1616

1717

1818

1919

2020

age (Ma)

GP

TS

(Lo

uren

s et

al.,

200

5)

VGP Latitude

Lithology

Lithology

VGP Latitude

VGP Latitude

widely-spaced sites

growth strata

Xiyu Conglomerate

Pakabulake Fm.

Wuqia Unit C

Wuqia Unit A or B

corr

elat

ion

stratigraphic

magnetostratigraphic

0

200

400

600

800

1000

1200

1400

1600

1800

2000

2200

2400

2600

2800

3000

3200

3400

3600

3800

4000

4200

4400

4600

4800

5000

AS

Wes

t

Ahu

Sis

ters

(A

S)

Eas

t R

eser

voir

Byp

ass

Res

ervo

ir

stra

tigra

phic

hei

ght (

m)

−90 0 90

−90 0 90

−90 0 90

-90 0 90

?

?

KEYlit

holo

gy

~2.5 Ma

~8.7 Ma

−90 0 90

L1nL2nL3nL4n

L5n

L10n

L7n

Qigailike

VGP Latitude VGP Latitude-90 0 90

Y1nY2nY3n

Y3rY4nY7n

Ayakeqiana

A

K1nK0r

K2n

K3n

K7n

K10nK11nK12nK13nK14nK15n

K16n

K17n

K18n

K19nK20nK21nK23nK24n

K26n

K8n

K4n

K5nK6nK7n

C5n

Fig.13. Proximal andmedial basin magneticpolarity timescales (MPTS)and correlation with thegeomagnetic polaritytimescale (GPTS) ofLourens et al. (2005).Section locations are shownin highlighted region of theinset digital elevation model(A).Virtual geomagneticpole latitudes (VGP) for allsamples are shown relativeto their stratigraphicposition and indicatenormal and reversedpolarity for each sample.Site coordinates are listed inHeermance (2007).Interpreted ages for the baseof the Xiyu conglomerate inthe Reservoir andAyakeqiana sections and theage of growth strata areshown. Stratigraphiccorrelations tie together thesections for a completechronology of the medialand northern foreland.Sampled lithology is shownadjacent to eachVGP plot.



medial basin composite MPTS, does not include theBrunhes normal chron, is dominated by reversed polarityand contains at least two, but likely ¢ve, reversals (Fig.13).(3) The statistical method of Johnson & McGee (1983),described below, suggests that the section is likely to span16 � 4Myr, predicting an average sedimentation rate of0.22mmyear�1.This rate is consistent with rates of 0.13^0.43mmyear�1 inferred for the middle and late Miocenefrom the northern and southern Tian Shan £anks (Char-reau et al., 2005, 2006; Huang et al., 2006). This range ofrates suggests that the total time span for the 3800-m-thick medial basin composite MPTS is between 30 and10Myr (Fig. 15b). In sum, given the vertical frequency ofsamples, the pattern of reversals and an estimated dura-tion of 12^20Myr, we conclude that a convincing matchcan be made to only one part of the GPTS between 35 and1Ma: between �18 and 5.5Ma (Fig.15c).

The best- ¢t correlation ties magnetozoneK16n (Figs13and 15c), an anomalously long normal polarity zone thatspans �400m, with the longest normal zone in the Neo-gene between 10.0 and 11.0Ma. Based on this tie point, areasonable correlation is made between �18.3 and 5.5Maby simply matching magnetozones between the MPTSand GPTS. The long reversed magnetozone (�100m

thick) above L7n at Qigailike correlates best with the longreversed chron from �16^15Ma in the GPTS. Althoughthis preferred correlation of theMPTS to theGPTS (Figs13 and15c) appears reasonable and the number and spacingof reversals are consistent between the two, some mis-matches exist. For example, an ‘extra’ reversed magneto-zone below K10n (Fig. 13) does not appear to correlatewith the GPTS, and may result from unremoved rema-nence that added an extra normal polarity zone at K11n.The most uncertain correlations are for the base of theQi-gailike section where many reversals within a short strati-graphic section imply slow sedimentation rates, similar toprevious studies for the earlyMiocene along the southernTian Shan foreland1000 km east in theKuqa Basin (Fig.1,Huang et al., 2006). Our preferred correlation extends thebase of the Qigailike section to17.5^18Ma, and produces amatch with the GPTS that shows less time (�5Myr)spanned by the upper half of the section than the lowerhalf (�8Myr), consistent with increasing sedimentationrates since the earlyMiocene (Huang et al., 2006).

Correlation of theASE section located above the medialbasin composite is more ambiguous due to widely spacedsites (average spacing: 40m) within the Xiyu conglomer-ate, such that some reversals may not have been sampled.

0 1,000 2,000 3,000 4,000 5,000500

Meters

Keketamu Anticline

Reservoir Bypass

Ahu Sisters West

Ahu Sisters East

correlation line

growthgrowth

stratastratagrowth

strata

Reservoir Section

Fig.14. Aereal photo showing sample-site locations and the correlation between sections from the Keketamu compositemagnetostratigraphy. Correlation between sections was completed by tracing beds along strike between sections (white dashed lines).Normal (black) and reversed (white) polarity sites are shown as the circles shown on the photo. Only some of the sites are numberedfor clarity.



Moreover, three of the magnetozones are based on singlesites (grey bars, Fig.13) and are not included in our o⁄cialanalysis. Thus, several di¡erent correlations are possible.One correlation places the top of the ASE section, whichcorresponds to the location of the growth strata, at�2.5Ma (Fig. 13). If we allow the sediment accumulationrates to double to 800mMyr�1 after �6Ma (Chen et al.,2002), the top of the ASE section and the growth stratawould occur at �4Ma.This span of 2.5^4Ma set reason-able limits on the likely age of the top of the section in themedial basin.

Alternative correlations of the composite magnetic sec-tion (Fig. 15d and e) are limited by assumptions for themaximum and minimum ages interpreted for the MPTS.Magnetostratigraphic studies of middle and upper Mio-

cene strata elsewhere in theTian Shan foreland suggest ac-cumulation rates could range from 0.13 to 0.43mmyear�1

(Charreau et al., 2006; Huang et al., 2006). Applying theseend-member rates to our data (Fig.15b) predicts too manyor too few reversals within the calculated time to becon¢dently correlated anywhere to the GPTS within theMio-Pliocene, thereby suggesting an average rate midwaybetween these endmembers.Comparedwith our preferredmatch to the GPTS, attempts to produce either a youngeror older correlation yield less- satisfactory matches. Alter-native correlationA (Fig.15d) places the medial basin com-posite section as young as possible. As discussed above,chron K4n cannot be younger than 1.78Ma because theASE section above the medial basin composite (Fig. 14)contains at least one normal polarity zone with more than

~ 2

9 m

.y. a

t 0.1

3 m

m/y

r (H

uang

et a

l., 2

006)

~ 1

6 m

.y. a

t 0.2

2 m

m/y

r (C

harr

eau,

200

5, 2

006,

an

d ca

lcul

ated

afte

r Jo

hnso

n an

d M

cGee

, 198

3)

~9

m.y

. at 0

.43

mm

/yr

(Cha

rrea

u et

al.,

200

6)

(a)

BoguziheKeketamucomposite

(e) Alternative B

(c) Preferred correlation

(d) Alternative A

(b) Hypothetical magnetostratigraphic time span GPTS (Ma)

GPTS (Ma)

Boguzihe0 20

kilometersBoguzihe0 m

0 m

3800 m

K7n

K10n

K2n

K4n

K14n

K16n

K18n

K20n

K8n

L1n

L4n

L7n

L10n

0 m

K7n

K10n

K2n

K4n

K14n

K16n

K18n

K20n

K8n

K7n

K10n

K2n

K4n

K14n

K16n

K18n

K20n

K8n

L1n

L4n

L7n

L10n

L1n

L4n

L7n

L10n

0 m

Qig

ailik

e

Res

ervo

ir

Byp

ass

AS

Wes

t

Medial Basin

Composite

Magnetostratigraphy

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

3800 m3800 m

3800 m

Fig.15. Correlation for themedial basin compositestratigraphy. (a) Landsatimage showing correlation ofthe top of the medial basinwith the distal basinmagnetostratigraphy. (b)Hypothetical length of themagnetostratigraphy basedon assumed slow(0.13mmyear�1, left) and fast(0.4mmyear�1, right)sedimentation rates.(c) Preferredmagnetostratigraphy thatspans �16Ma, and providesthe best ¢t to thegeomagnetic polaritytimescale (GPTS). See textfor complete explanation.(d, e) Alternative correlationsof the medial basincomposite sectionwith theGPTS.



two sites, and the top of the ASE cannot include theBrunhes chron younger than 0.78Ma. Correlation of themagnetostratigraphic pattern back in time from 1.77Marequires that observed magnetozones be ignored (e.g.K9n, K11n and K13n), whereas other subchrons in thetimescale are not observed in the local magnetostratigra-phy. Although the same basal age of 17.5^18Ma as in ourpreferred correlation is predicted for this alternative cor-relation, the mismatches with the GPTS suggest that thisalternative correlation is less likely.

Alternative correlation B (Fig.15e) attempts to correlatethe base of the medial basin composite to the late Oligo-cene at �25Ma, which is the estimated age of uplift forthe southern £ank of the Tian Shan (Sobel & Dumitru,1997; Yin et al., 1998). This correlation spans from�30Ma to �11.5Ma and matches the GPTS quite well.This correlation predicts, however, that the top of themedial basin composite (Fig.15e) would be �11Ma,mucholder than the same stratigraphic level within the distalbasin at Boguzihe (Fig. 15a). The absence of any obviousdepositional hiatuses within the sections combined withour ability to trace bedding between structures strongly ar-gues that the top of the medial basin composite corre-sponds with the lower Pliocene and not older, as analternative correlationwould suggest.Thus, we dismiss al-ternative B as a probable correlation.

Northern foreland results

Similar to the Qigailike section, the Ayakeqiana section inthe proximal basin lies unconformably on Cretaceousstrata. This unconformity serves as the stratigraphicmarker for correlating the sections and tying the northernforeland magnetostratigraphy with the GPTS.Within the13 magnetozones at Ayakeqiana (Fig. 13), the prominentreversed magnetozone (Y3r) is interpreted to correlatewith the reversed chron between 15.1 and 16Ma on theGPTS based on stratigraphic height above the basal un-conformity (similar to the correlation atQigailike).Match-ing of magnetozones below Y3r suggests that the base ofthe section extends to �17.5Ma.

PBB results

The PBB is distinct from the Xiyu conglomerate in that itlies unconformably over the sub-vertical, middleMioceneXiyu conglomerate and is preserved only north of theTa-shipishake Anticline (Fig. 2b). Only two reversals were ob-served in PBB magnetostratigraphy within the relativelyshort 400-m-thick section (Fig. 9), making correlationwith the GPTS ambiguous. The PBB must be youngerthan �14Ma, the age of theXiyu conglomerate that it un-conformably overlies, but older than the bottom of theBrunhes chron at 0.78Ma, because of the reversed polarityzone observedwithin the section (Lourens et al., 2005).

Expected time span of section correlations

To assess our correlations against the expected duration ofeach section, we used the statistical method of Johnson &McGee (1983).This method uses (i) the number of rever-sals discovered in a stratigraphic section; (ii) the numberof sample sites within which those reversals were found;(iii) a time estimate for the average duration of polarity in-tervals during the Neogene (210 kyr.) or Plio-Pleistocene(220 kyr) and (iv) a choice of equal or random site spacingto calculate the expected time spanned by each section.The results (Table 3) indicate excellent agreement (withinerror) between the statistically estimated duration of eachsection (Johnson & McGee, 1983) with that predicted byour correlation with the GPTS (Lourens et al., 2005) forall sections except for the Reservoir section. The calcu-lated time spanned by the Reservoir section is 10Myr,which is �3Myr more than our correlated result. Thismismatch may be attributed to the fact that the frequencyof reversals between 15 and 8Ma is greater than that as-sumed in the Johnson & McGee (1983) model. Despitethe Reservoir section disparity, this statistical model,in combination with microfossil constraints from theBoguzihe section and tying theKeketamu section to chronC5n within the Miocene, suggests our correlations ofthese magnetostratigraphic sections to the GPTS arereasonable.

Sediment-accumulation rates

This extensive magnetostratigraphy de¢nes sediment-ac-cumulation rates for the sampled localities in the KashiBasin (Fig. 16). The accumulation rates are not correctedfor post-depositional compaction, in part because of theunknown compaction history.Given the prevalence of car-bonate in the source area and the arid to semi-arid condi-tions, early cementation is likely, such that standardporosity-vs.-depth curves would be inapplicable. Ob-served rate changes do not correlate with lithologicboundaries, and thus we interpret them as independentof compaction or lithology. Nonetheless, rates presentedhere should be minima that showhow sedimentation rateschange at speci¢c localities relative to basin structures andgeometry. From these data, we interpret how tectonicloading and deformation relate to sedimentation withinthe basin.

Overall, accumulation rates are fastest for younger stra-ta such that at any given position, rates tend to acceleratefrom older to younger. The oldest Tertiary strata (WuqiaGroup), however, have generally the ¢nest grain size (shale)that may be more compacted and thus exhibit a slower ap-parent accumulation rate. Within any section, the ratesmay have remained steady for as much as 4Myr. Rates be-fore 16.5Ma in both the medial and southern foreland arecomparable and represent the slowest rates in this study(�80mMyr�1). After 16Ma, the rates accelerated bymore than three-fold, but are higher in the proximalbasin where the Xiyu conglomerate ¢rst appears at�15.5� 0.5Ma (Fig. 16). Another three-fold increase in



sediment accumulation rate from 130 to 430mMyr�1 oc-curred at 13.5Ma at the Reservoir Section.

Pliocene accumulation rates within the Atushi Forma-tion in the distal basin were up to twice as rapid as Mio-cene rates in the medial basin, despite less conglomeratedeposition (Fig. 16), but are consistent with a regionalincrease in sedimentation rate inferred for the Pliocene(Me¤ tivieretal.,1999;Zhangetal., 2001).Rates for theAtushiand Kashi anticlines sections average 700mMyr�1, butappear dependent on positionwithin the basin: faster ratesprevail in more northerly and westerly sections. Regionalsubsidence during the Pliocene could have provided ac-commodation space for climatically in£uenced rapidlyeroding mountains surrounding the basin during thePliocene (Molnar, 2001; Zhang etal., 2001).The £exural ri-gidity of theTarim platform is interpreted to be �1024Nm(Burov &Diamont, 1992) and as such the forebulge shouldbe present at least 200 km from the tectonic load (Beau-mont, 1981). However, the Kashi Basin lies within 50 km

of theTian Shan mountain front and within 100 km of thePamir mountain front, and thus is located within the £ex-ural foredeep of two active, converging mountain ranges.Hence, the Kashi Basin is likely becoming deeper as theTarim basin closes up at this location, providing accom-modation space for an increased sediment supply duringthe Pliocene and Pleistocene.

DISCUSSION

Stratigraphic architecture and faciesdevelopment of the Xiyu conglomerate

The Xiyu conglomerate forms a southward-thinningwedge that capsTertiary foreland strata at almost every lo-cation within the foreland. The conglomerate can bemapped as a continuous unit (i.e. formation) across all thestructures south of the KBT (Fig. 17), despite its occur-

430±10

270±70

90±10

70±15

280±30250±20

130±30520±20

820±20

720±25

600±130

1

2

3

4

5

02468101214161820

age (m.y.)

stra

tigra

phic

hei

ght (

km)

16.3 13.5 4.0

Ayakeqiuana (Y)

Qigailike (L)

Reservoir

Reservoir Bypass

Ahu Sisters West BoguziheGanhangou

Kashi Town

Kashi WestAhu Sisters East

age and stratigraphic height of Xiyu Conglomerate.

ST

RA

TA Atushi Fm. (5.3–0 Ma)

Wuqia Group: Pakabulake Fm. (13.7–5.3 Ma)

Wuqia Unit A&B (17.5–14.9 Ma)

SEDIMENTATION RATE CHANGES

Wuqia Unit C (14.9–13.7 Ma)

SEDIMENTATIONRATES (m/m.y.)

15.5 Ma

8.6 Ma

1.9 Ma

1.9 Ma

0.8 Ma

1.4 Ma

Fig.16. Plot of stratigraphic depth vs. age based on magnetostratigraphic correlations for each measured section (locations shown ininset map). For visual clarity and to avoid overlapping data, stratigraphic depth measurements are relative to only the section measuredand should not be used to correlate between sections. Sedimentation rates are calculated based on a minimum of ¢ve reversals fromwithin each section. Speci¢ed uncertainties on sedimentation rates are 2-s values from least-squares linear regression through all dataas shown by the lines behind the data points. Sections have not been corrected for compaction e¡ects, and thus the slopes calculated foreach represent minimum sedimentation rates at each section.The stratigraphic location of the Xiyu conglomerate in each section isindicatedwith the correlated age for the base of the conglomerate within each section. Age correlation of basin strata are shown alongthe base of the graph.



rence at di¡erent stratigraphic levels throughout the fore-land. At its northern limit, the conglomerate appearso200m above the base of the Tertiary and is 4400mthick (Fig. 3). Here the conglomerate is truncated by theKBTand thus represents its minimum thickness at thislocation.Within the medial foreland, the conglomerate is42500m thick, and occurs approximately 2000m abovethe base of theTertiary (Fig. 4).Within the southern fore-land along the £anks of the Atushi and Kashi anticlines,the Xiyu conglomerate forms a thin (o300m) veneerabove siltstone and sandstone of the Atushi Formation(Fig. 5). Here, the conglomerate forms digitate lobes thatjut into the youngest foreland strata. Furthermore, the lo-cally thickestXiyu conglomerate at these distal locations islocatedwhere the active rivers cross folds, and these pinchout within a few kilometres laterally (e.g. Fig. 8a).

Combining all of the lithofacies and chronostrati-graphic data, the picture that emerges is of laterally con-

tinuous conglomerate deposition that prograded southover time. The coarse-grained facies are localized alongthe mountain front, whereas at more distal locations theconglomerate lithofacies are restricted to palaeovalleysthat cut southward across the basin (e.g. Lo¤ pez-Blanco,2002). The persistence of the active water gaps localizedthrough the thickest, and more resistant, conglomerate fa-cies along the southern foreland suggests that incision hasoutpaced uplift, and that rivers have been long lived (41^2Myr) across the landscape. Conglomerate deposition inthe northern foreland was contemporaneous with evapor-ite deposition farther south (Fig. 8a), similar to the presentconditions where gravel deposited within the Baishikere-mu He (He is Mandarin for ‘river’) channel south of theKashi anticline occurs only within �6 km of the southernexposed fold £anks. South of this ‘gravel front’are areas ofstrictly evaporite, ¢ne-grained £uvial and aeolian deposits(Fig. 8a).

AtushiAnticline

(~1.4 Ma)

KashiAnticline

(~1.4 Ma)

KeketamuAnticline(<4 Ma)

TashipishakeAnticline (~13.5 Ma)

Kashi Basin Architecture

matrix10

15

9

8

7

6

5

4

3

2

1

0st

ratig

raph

ic d

epth

(km

)10 km

5x vert. exag.

NORTH SOUTH

0

2

4

6

8

10

12

14

16

18

20

GPTSage (Ma)

Kas

hi

Atu

shi

Kek

etam

uA

yak.

Qig

alik

e

Kashi Basin

Thrust (16.3 Ma)

olderolder youngeryounger

matrix

0 30 40 50 602010

gypsiferous

Xiyu Congl.

magneto-stratigraphiccorrelation

~13.5

~13.5

~4.0

southernmedialnorthern

Unit C (14.9–13.7 Ma)

Pakabulake Fm.(13.7–5.3Ma)

Unit A & B (18-14.9 Ma)~16.3

3.0

5.6

(5.65)

1.0

14.7

~2.5

0.7

QigailikeA

yakeqiana

Kashi

Keketam

u

distance from Kashi Basin thrust (km)

PALEOZOIC AND

CRE TACEOUS STR ATA

WUQI A GROUP

Atushi Fm.(5.3–0 Ma)

12.814.5

17.5–18

Atushi

sediment accumulationrate change (Ma)

end of section age (Ma)

key

structureinitiation age

Fig.17. Schematic geometry of the Kashi foreland showing the magnetostratigarphic sections at their reconstructed pre-deforma-tion basin locations with accurate stratigraphic thickness.The magnetostratigraphic correlation on the right shows the correlationwith the geomagnetic polarity timescale of Lourens et al. (2005). Ages for the foreland stratigraphy, magnetostratgraphic sections andsediment accumulation rate changes are indicated in the ¢gure. Spatial location and the interpreted initiation age for the majorbasin structures are indicated along the base of the ¢gure. Ages for the top and bottom of each magnetostratigraphic section, andinterpreted changes in sediment accummulation rate, are shown in the ¢gure.



Time-transgressive Xiyu conglomerateprogradation

Five magnetostratigraphic age constraints combined withthe location of the southern edge of active gravel deposi-tion provide six basal ages for the Xiyu conglomeratewedge across the Kashi foreland. These ages of15.5 � 0.5Ma (northern foreland), 8.6 � 0.1Ma (medialforeland), 1.9� 0.2Ma (distal foreland), �1.4Ma (notbounded by reversals, distal foreland), 0.7 � 0.1Ma (distalforeland) and 0Ma allow us to illustrate the profoundlytime-transgressive nature and provide insights on the ap-parent disparity of ages for the Xiyu conglomerate foundaround theTian Shan andTarim basin (Fig.18a). Errors as-sociated with each age designation are based on the timespan of each reversal in which the base of the Xiyu con-glomerate was observed. Only four previous magnetostra-tigraphic studies provide absolute dates on the age of theXiyu conglomerate near the Tian Shan. Charreau et al.(2005) and Sun et al. (2004) interpreted di¡erent ages forthe Xiyu conglomerate at the Dushanzi section on thenorthern £ank of the ChineseTian Shan at 4.8 and 2.7Ma.Magnetostratigraphy from the southern margin of theTar-im Basin provide ages of 3.5Ma at Yecheng (Zheng et al.,2000) and 3.0Ma (Sun & Liu, 2006) at Sanju, �300 kmsoutheast of theKashi Basin. However, all of these magne-tostratigraphies are based on one measured sectionwithineach foreland region, or two sections correlated alongstrike within the same part of the basin relative to themountain front. Other ages given to the Xiyu conglomer-ate in the Tian Shan are based on assumed correlationswith global climate cooling at 2.5Ma (Avouac et al., 1993;Burch¢el et al., 1999) or its stratigraphic position abovePliocene strata (Huang & Chen, 1981; Yin et al., 1998).Furthermore, theXiyu conglomerate has been interpretedas a chronostratigraphic unit around the Tian Shan(Huang & Chen, 1981; Liu et al., 1996; Burch¢el et al.,1999). Although these studies do not systematically studythe spatial or temporal variability of the Xiyu conglomer-ate within any individual basin, the individual, site- speci-¢c ages are commonly used as evidence for either tectonicor climatic events since the Pliocene. In contrast, we de-monstrate here that the Xiyu conglomerate in the Kashiregion is a highly time-transgressive lithofacies that hascontinuously prograded south over415Ma.

The depositional ages for the arrival of gravel at di¡er-ent locationswithin theKashi foreland, combinedwith re-constructed shortening estimates across individual basinstructures (Fig.18b), allow us to reconstruct pre-deforma-tion gravel progradation rates into the foreland (Fig. 18c).Reconstructed distances along cross-section AÂ0 fromthe initial appearance of theXiyu conglomerate at theAya-keqiana section to theKeketamu andAtushi Anticlines are16^22 and 37^44 km, respectively.Based on the appearanceof the conglomerate at15.5Ma atAyakeqiana,8.6Ma atKe-ketamu and 1.9Ma at Atushi, a fairly constant rate (basedon these three points) of 2.5^3.5mmyear�1 is calculated(Fig. 18c). In contrast, using the reconstructed distances

between the Atushi and Kashi Anticlines and the furthestsouth location of gravel within the active river channels,and ages of 1.9, 0.7 and 0Ma, respectively, the gravel frontmigrated at 10^11mmyear�1 since 1.9Ma (Fig. 18c).Furthermore, west to east migration of the gravel frontalong the southern £ank of the Kashi Anticline occurredat �18mmyear�1 since 1.4Ma, suggesting that gravelprogradation rates are fastest towards the southeast. Thisthree- to four-fold increase in southern gravel prograda-tion rates occurred after �2Ma. Overall, the Xiyu con-glomerate has prograded 54^62 km south since its initialdeposition at 15.5Ma for an overall rate of 3.5^4mmyear�1, although higher gravel progradation ratesmay be southeasterly towards theTarim Basin centre.

Implications of observed changes insedimentation rate

At least three distinct changes in sediment accumulationrates at �16.3, 13.5 and �4Ma are interpreted withinthe Kashi foreland, and these ages are independent oflithofacies changes and formation boundaries thatcould be attributed to a climate shift (Fig. 16). Sediment-accumulation rates within terrestrial basins are controlledprimarily by rates of sediment supply and accommodationspace. Sediment supply is a function of climate that mayincrease the runo¡ and thus available streampower for £u-vial erosion (Molnar, 2004), source-rock lithology that iseither easily denuded and re-deposited or resistant andthus not as readily available for re-deposition (Carrollet al., 2006), and the proximity of the sediment source(Burbank & Beck, 1991). Accommodation space is createdeither by tectonically induced subsidence beneath an up-lifted block (Beaumont,1981; Jordan,1981) or by uplift thattraps sedimentation within a PBB (DeCelles & Giles,1996).Thus, an abrupt change in the sediment-accumula-tion rate could be the result of either a change in the £ux ofsediment (Molnar, 2001; Zhang etal., 2001) and/or tectonicsubsidence (Paola et al., 1992). The increased sedimenta-tion rates observed at16.3 and13.5Ma occur in the footwallof the KBTand south of theTashipishake Anticline, bothlocations thatwould experience tectonic subsidence due toshortening across each structure (Fig. 10c). Furthermore,uplift above theKeketamu anticline likely reduced accom-modation space, and thus sedimentation rate, while pro-viding a proximal source of sediment that contributed tothe increased rates above theAtushiAnticline13 km south.Based on these observations, we interpret the changes insedimentation rates at16.3,13.5 and 4Ma to be tectonicallyin£uenced phenomena due to uplift above theKBT,Tashi-pishake anticline and Keketamu anticlines, respectively.

Structural control on conglomerateprogradation

The initiation of rock uplift above speci¢c structures,combined with shortening estimates and location of eachstructure within the basin, permits calculation of the mi-gration rate of deformation into the foreland over time.



Deformation has proceeded south from the KBT since�16.3Ma. By pinning our basin location at the KBT,and adding the shortening range from each structure toits present-day distance south of the KBT, we can plotthe pre-deformation location of incipient structuresagainst their time of initiation. The slope of the line be-tween any two locations within the basin is the southwardmigration rate of the deformation front (Fig. 18c). Errorswithin the calculated rates are based on the range of short-ening values interpreted for each structure (Fig.10e). After

an initial uplift above the KBT at 16.3Ma, deformationjumped south to theTashipishakeAnticline and structuresin between at 13.5Ma at a rate of 2.1^3.4mmyear�1,but slowed after that time to 1.4^1.6mmyear�1 as defor-mation propagated south to the Keketamu anticlinesby 4Ma. After �4Ma, deformation propagated southat410mmyear�1 to the Atushi and Kashi Anticlines.

We suggest that localized upliftwithin the northern andmedial foreland caused steady progradation of the gravelfront towards the Boguzihe section until 1.9Ma. Changes

distance from Kashi Basin Thrust (km)

age

(Ma)

KashiBasin Thrust

reco

nstr

ucte

d po

sitio

n of

pr

esen

t-da

y gr

avel

fron

t

0

2

6

8

10

12

14

16

0 20 40 60

Atushi Anticline

2.1–3.4mm/yr

1.4–1.6mm/yr

>10mm/yr 10–11

2.5–3.5mm/yr

deformation frontmigration rate

gravel frontmigration rate

Tashipishake Anticline

Ayakequiana

KashiAnticline

Pamir foothills

Pamir foothills

High High Tian Shanian Shan

Pamir foothills

High Tian Shan

15.5 Ma8.6 Ma

1.9 Ma

1.4 Ma 0.7 Ma

limit of graveldeposition inforeland

Kashi Anticline

Atushi Anticline

Kashi Basin ThrustKashi Basin ThrustKashi Basin Thrust

0 Ma

Northern Foreland

Medial Foreland

Limits of Distal Foreland

A

A’

19

373724373724

(a)

(b)

(c)

Xiyu Conglomerateprogradation

deformationfront

migration

4

Keketamu Anticline

76.5°E76.0°E75.5°E75.0°E

40.0°N

39.5°N

NORTH

location of basal Xiyu Conglomerate Age

0 10 20 30 40 50

kilometers

Kashi Basin Thrust (KBT) Ayakeqiana

Keketamu (Keketamu Ant.)

Boguzihe (Atushi Ant.)

Kashi Anticline

active gravel front

current separation (km from KBT)

0 1 9 22 37 41

pre-deformation distance from KBT (km)

0 3–8 16–22 37–44 50–58 54–62

initiation of deformation (Ma)

16.3 13.5 4 1.4 1.4 0

initial deposition of conglomerate (Ma)

- 15.5 8.6 1.9 1.4-0.7 0

373724Average paleocurrent direction. Number is total measured indicators.

Fig.18. (a) LandSat photoof the Kashi foreland thatshows the spatial variation ofbasal ages from the Xiyuconglomerate. Palaeocurrentdirections for theconglomerate facies areindicated by the blackarrows.The numbers withinthe arrows are the number ofdata points used forpalaeocurrent average. Bi-directional white arrows areaverage palaeocurrentdirections for the siltstonefacies below theconglomerate.This ¢gureillustrates howpalaeocurrents change fromstructure parallel totransverse as the Xiyuconglomerate progradessouthward over time. (b) Preand post-deformationdistances of the structuresand base of conglomerate areindicated, as well as theinferred ages for theconglomerate facies andstructure initiation. (c) Plotof migration rates of thedeformation front and thegravel front over time.Distance from the KashiBasinThrust (KBT) (x-axis)are based on reconstructed,undeformed distances toeach structure. Error bars(deformation front) or circles(Xiyu Conglomerate) de¢nethe uncertainty in thereconstruction (describedin text).



in gravel progradation rate can be in£uenced by a changein climate (Molnar, 2001), a decrease in tectonic subsi-dence (Heller et al., 1988; Paola et al., 1992), an increase inrock uplift rate of the gravel source (Burbank et al., 1988)or a change in source area lithology (DeCelles et al., 1991b;Carroll et al., 2006). A lack of stratigraphic formationboundaries or lithofacies associated with changes in sedi-ment accumulation rate or in gravel progradation suggeststhat climate perturbations are unlikely to control the grav-el progradation observed. Because the thrusts are south-vergent, the gravel source might move south at a ratecomparable with shortening across the fault. Therefore,in the absence of climate perturbations or changes intectonic rates across the basin, the gravel progradation rateis likely to match shortening rates.

Deformation jumped south to the Keketamu anticlineat �4Ma and to the Atushi and Kashi anticlines at1.4Ma. After �4Ma, uplift above the distal structures re-duced subsidence in the area of deformation, driving theconglomerate front basinward at rates comparable withthe rate of structure migration.The �2.0Myrdiscrepancybetween deformation at Keketamu and the appearance ofgravel 14.8 km south at Boguzihe are most likely the resultof a lag time for prograding gravels to arrive at distal loca-tions (Jordan et al., 1988; Jones et al., 2004). Observedsouthward progradation rates for the Xiyu conglomerateare generally slower than syn-tectonic gravels in the Siwa-lik foothills (35^45mmyear�1) in the southernHimalayanforeland (Burbank et al., 1988) or in the Sevier foreland(20mmyear�1) in western North America (Horton et al.,2004).We suggest our slower rates may be due to lower sus-tained shortening rates across the Kashi foreland, or lowerstream power due to a more arid climate, than for the otherregions.

Good correlation between the structural propagationrates andgravel progradation rates suggests a causative linkbetween the two in this foreland basin. Between 15.5 and�4Ma, the gravel prograded south at 2.5^3.5mmyear�1,similar to the 1.4^3.4mmyear�1 for structural migrationinto the basin (Fig. 18c). Furthermore, a sharp increase inthe gravel progradation rate to 410mmyear�1 is ob-served after 2Ma, and correlates with the same increasein structural migration rate after 4Ma. The proximity ofthe gravel front to the deformation front, however, appearsinversely related to the activity of speci¢c structures.During times of more rapid shortening between 16.3 and13.5Ma, and after 2Ma, the gravel frontwas locatedwithina few kilometres of the active deformation front (Fig. 18c).In contrast, the gravel front migrated to over 20 km aheadof the deformation front between 13.5 and 4Ma, whenshortening rates were slower (Fig.18c).Thus, higher short-ening on speci¢c structures at the deformation front areassociated with more proximal gravel fronts, and lowrates or tectonic quiescence may cause the gravel front tomigrate away from the deformation front within a fore-land.Whereas a climate change between 2 and 4Ma mighthave accelerated prevailing rates of gravel progradation(Molnar, 2001; Zhang et al., 2001), it is unlikely that it

would also a¡ect the rate atwhich deformation propagatedacross the basin. Hence, we view climate as a secondaryin£uence on gravel progradation, and uplift above speci¢cstructures increased local slopes such that gravel waspushed farther into the foreland.

Basin evolution

A three-fold increase in sedimentation rate is observed at16.3Ma at the Ayakeqiana and Qigailike sections within1and 6km, respectively, south of the KBT (Figs 2a and16).This age overlaps with the 18.9� 3.3Ma apatite ¢ssion-track cooling from the KBT hanging wall (Sobel et al.,2006). Furthermore, the coarse-grained, debris- £ow fa-cies Xiyu conglomerate appears at Ayakeqiana at 15.5Ma.Together, these three ages suggest that uplift of the KBTat �16.3Ma induced subsidence in the region of Ayakeqi-ana at that time. By15.5Ma, su⁄cient relief was producedin the hanging wall to shed debris £ows into the basin thatproduced the proximal, coarse-grained Xiyu conglomer-ate. Before 16.3Ma, deposition of £uvio-lacustrineWuqiaGroup strata resulted from deposition 20^30 km southof the Maidan and Muziduke faults that were active by20^25Ma (Sobel et al., 2006).

These new chronologic data indicate that the overallpattern of deformation and the basin depocentre have mi-grated southward within the basin since at least 16.3Ma(Fig. 18a). This age is 4^5Myr earlier than the bedrock-involved thrusting that de¢nes initial growth of theKyrgyzRange on the north side of the Tian Shan (Bullen et al.,2001, 2003), but is consistentwith other estimates for initi-al deformation in the southern Tian Shan (Hendrix et al.,1994; Sobel & Dumitru, 1997; Yin et al., 1998; Sobel &Strecker, 2003). Moreover, near the study area, deforma-tion likely began earlier along the Maidan and Muzidukefaults (Fig. 2b) to the north (Sobel et al., 2006). Abruptchanges in sediment-accumulation rates at 16.3 and13.5Ma are interpreted as responses to tectonic loadingdue to thrusting above the KBTand Tashipishake Anti-cline. Steady sediment-accumulation and gravel-progra-dation rates suggest modest but continual thrustingbetween the KBT and Tashipishake Anticline, but nomajor southward steps of deformation until 4.0Ma whenthe Keketamu anticline began to deform. After �2Myrof deformation focused on the Keketamu fold, fault-related folding stepped south again to the Atushi andKashi Anticlines in the distal basin (Fig.18c).

CONCLUSIONS

Although an overall basinward progression of deformationcharacterizes most forelands, the speci¢c timing of initia-tion of individual thrust faults and folds is commonly un-known.New chronologic data from theTian Shan forelandyield one of the densest magnetostragraphic studies of anevolving terrestrial foreland basin system.These data pro-vide a detailed temporal context within which to quantify



the history of sedimentation and deformation of theKashiBasin. New magnetostratigraphies from 11measured sec-tions de¢ne a continuous 18Myr record of deposition,where rapid lateral and vertical facies changes occur dueto pulsed episodes of southward-migrating deformationinto the foreland. Using the extensive magnetostrati-graphic data in theKashi foreland,we interpret the timingof deformation based on accelerated accumulation rates inthe footwalls of thrust faults and both decelerating accu-mulation rates and growth strata in their hanging walls.

Data from the Kashi Basin show that the Xiyu con-glomerate is highly time-transgressive and spans from atleast 15.5Ma to the present. This ¢nding contrasts withprevious studies that assume the Xiyu conglomerate iseither Plio-Pleistocene in age or represents a chronostra-tigraphic unit. Distribution of the Xiyu conglomerate iscontrolled by river drainage patterns over millions ofyears, by the proximity of conglomerate source area andby the growth of new folds that promote sedimentary by-passing. Along with bypassing, reworking of theXiyu con-glomerate from hinterland uplifts has pushed the gravelfront out into the basin, whereas these same uplifts haveponded separate PBB strata behind their uplifted hangingwalls. Because the age and migration rate of a single con-glomerate lithofacies have rarely been so extensively con-strained within a single foreland, the punctuated butsustained and long-term conglomeratic progradation seenhere may serve as a catalyst for careful assessment of simi-lar lithofacies in other basins. Rather than a steady migra-tion of the deformation front, our data de¢ne a tectonichistory that is punctuated by southward jumps in the locusof fault-related folding. Gravel progradation rates are si-milar to structural shortening rates, suggesting a causallinkage.The distal limit of gravel progradation lies withina few kilometres of the structural deformation front dur-ing times of active, relatively rapid shortening, but mi-grates basinward, away from the locus of deformationduring times of little or no tectonic activity.

Our results indicate that the overall stratigraphic geo-metry was controlled by a long-term (415Myr) wave ofdeformation that swept across the foreland andmodulatedpatterns of facies migration.Thus, within the Kashi fore-land, tectonics is the fundamental control on conglomer-ate distribution, although climate likely plays a role atshorter time scales. Such ¢ndings provide an explanationfor the variable timing of conglomerate deposition aroundtheTian Shan, illustrate the relationship of facies patternsto spatial and temporal changes in rates of subsidence andprovide a well-constrained conceptual model for inter-preting upward-coarsening foreland sequences fromother terrestrial foreland basins.

ACKNOWLEDGEMENTS

We thank Sun Dongjiang, Feng Wenquan, Kate Scharerand Patricia Ruano for ¢eld assistance and PetroChina foraccess to seismic line TA 9402. Discussions with Bruce

Luyendyk, Joseph Goode and Joel Saylor were helpful andappreciated. Thoughtful and thorough reviews by PaulHeller, Mark Hendrix and Peter DeCelles profoundly im-proved this manuscript.We are especially grateful to J. L.Kirschvink and S. Bogue for use of and assistance in theCaltech and Occidental College palaeomagnetics labora-tories, respectively, and to B. Kopp for patience and assis-tance at Caltech. This research was supported by USNational Science Foundation grant EAR-0230403, Na-tional Science Foundation of China grant 40372081 and40672109. Software by Craig Jones was used for analysisand presentation of the palaeomagnetic data.We give spe-cial appreciation to the late Prof. S. Z.Wu who helped usdecipher the Palaeozoic stratigraphy during the start ofthis research.

REFERENCES

Abdrakhmatov, K., Weldon, R.J., Thompson, S.C., Bur-bank,D.W., Rubin, C.,Miller,M.M. &Molnar, P. (2001)Onset, style and current rate of shortening in the central TienShan, Kyrgyz Republic.Russ. Geol. Geophys., 42, 1585^1609.

Abdrakhmatov,K.Y.,Aldazhanov, S.A.,Hager, B.H.,Ham-

burger, M.W., Herring,T.A., Kalabaev, K.B., Makarov,V.I., Molnar, P., Panasyuk, S.B., Prilepin, M.T., Reilin-ger, R.E., Sadybakasov, I.S., Souter, B.J., Trapesnikov,Y.A,Tsurkov,V.Y. & Zubovich, A.V. (1996) Relatively recentconstruction of the Tien Shan inferred from GPS measure-ments of present-day crustal deformation rates. Nature, 384,450^453.

Allen, M.B.,Vincent, S. & Wheeler, P.J. (1999) Late Ceno-zoic tectonics of the Kepingtage thrust zone: interactions oftheTien Shan and Tarim Basin, northwest China.Tectonics, 18,639^654.

Allen, M.B., Windley, B.F., Chi, Z., Zhong-yan, Z. &Guang-Rei,W. (1991) Basin evolution within and adjacent tothe Tien Shan Range, NW China. J. Geol. Soc. London, 148,369^378.

Armstrong,F.C.&Oriel,S.S. (1965) Tectonic development ofIdaho-Wyoming thrust belt.AAPG. Bull., 49, 1847^1866.

Avouac, J.-P.,Tapponnier, P., Bai, M., You, H. & Wang, G.

(1993) Active thrusting and folding along the northern TienShan and late Cenozoic rotation of theTarim relative toDzun-garia and Kazakhstan. J. Geophys. Res., 98, 6755^6804.

Bally, A.W., Ming Chou, I., Clayton, R., Heugster, H.P.,Kidwell, S., Meckel, L.D., Ryder, R.T., Watts, A.B. &Wilson, A.A. (1986) Notes on sedimentary basins in China,Report of the American Sedimentary Basins Delegate to thePeople’s Republic of China, August 17^September 8, 1985.U.S.Geol. Surv. Open File Rep., 86^327, 108.

Banks,C.J.&Warburton, J. (1986) ‘‘Passive-roof ’’duplex geo-metry in the frontal structures of the Kirthar and Sulaimanmountain belts, Pakistan. J. Struct. Geol., 8, 229^237.

Beaumont, C. (1981) Foreland basins. Geophys. J. Roy. Astr. Soc.,65, 291^329.

Blair,T.C. &McPherson, J.G. (1999) Grain-size and texturalclassi¢cation of coarse sedimentary particles. J. Sediment. Res.,69, 6^19.



Bridge, J.S. (2003) Rivers and £oodplains: forms, processes, and sedi-mentary record. Oxford, UK, Malden, MA, USA: BlackwellPub., 494pp.

Bullen,M.E., Burbank,D.W., Abdrakhmatov,K.Y. &Gar-

ver, J. (2001) Late Cenozoic tectonic evolution of the north-western Tien Shan: constraints from magnetostratigraphy,detrital ¢ssion track, and basin analysis. Geol. Soc. Amer. Bull.,113, 1544^1559.

Bullen, M.E., Burbank, D.W. & Garver, J.I. (2003) Buildingthe northern Tien Shan: integrated thermal, structural, andtopographic constraints. J. Geol., 111, 149^165.

Burbank, D.W. (1992) Causes of recent Himalayan uplift de-duced from deposited patterns in the Ganges basin. Nature,357, 680^682.

Burbank,D.W. & Beck, R.A. (1991) Models of aggradation ver-sus progradation in theHimalayan Foreland.Geolog. Rund., 80,623^638.

Burbank,D.W., Beck, R.A., Raynolds, R.G.H.,Hobbs, R. &Tahirkheli, R.A.K. (1988) Thrusting and gravel prograda-tion in foreland basins: a test of post-thrusting gravel disper-sal.Geology, 16, 1143^1146.

Burbank,D.W.,McLean, J.K., Bullen,M., Abdrakhmatov,K.Y.&Miller,M.G. (1999) Partitioning of intermontane ba-sins by thrust-related folding, Tien Shan, Kyrgyzstan. BasinRes., 11, 75^92.

Burbank, D.W.,Meigs, A. & BrozovicŁ , N. (1996) Interactionsof growing folds and coeval depositional systems.Basin Res., 8,199^223.

Burbank,D.W. & Raynolds, R.G.H. (1988) Stratigraphic keysto the timing of thrusting in terrestrial foreland basins: appli-cation to the northwest Himalaya. In:New Perspectives in BasinAnalysis (Ed. by K.L. Kleinspehn & C. Paola), pp. 331^351.Springer-Verlag, NewYork.

Burchfiel, B.C., Brown, E.T., Qidong, D., Xianyue, F.,Jun, L., Molnar, P., Jianbang, S., Zhangming, W.

& Huichuan, Y. (1999) Crustal shortening on the marginsof the Tien Shan, Xinjiang, China. Int. Geol. Rev., 41,665^700.

Burov, E.B. & Diamont, M. (1992) Flexure of the continentallighosphere with multilayered rheology. Geophys. J. Int., 109,449^468.

Carroll,A.R.&Bohacs,K.M. (1999) Stratigraphic classi¢ca-tion of ancient lakes: Balancing tectonic and climatic controls.Geology, 27, 99^102.

Carroll, A.R., Chetel, L.M. & Smith, M.E. (2006) Feast tofamine: sediment supply control on Laramide basin ¢ll.Geol-ogy, 34, 197^200.

Carroll, A.R., Graham, S.A., Hendrix, M.S., Ying, D. &Zhou, D. (1995) Late Paleozoic tectonic amalgamation ofnorthwestern China; sedimentary record of the northernTarim, northwestern Turpan, and southern Junggar basins.Geol. Soc. Am. Bull., 107, 571^594.

Charreau, J., Chen,Y.,Gilder, S.,Dominguez, S., Avouac,J., Sen, S., Sun, D.J., Li,Y.A. &Wang,M.W. (2005) Magne-tostratigraphy and rock magnetism of the Neogene KuitunHe section (northwest China): implications for late Cenozoicuplift of theTianshan mountains. Earth Planet. Sci. Lett., 230,177^192.

Charreau, J.,Gilder, S., Chen,Y.,Dominguez, S., Avouac,J.-P., Sen, S., Jolivet,M., Li,Y. &Wang,W. (2006) Magne-tostratigraphy of theYaha section,Tarim Basin (China); 11Maacceleration in erosion and uplift of theTian Shan mountains.Geology (Boulder), 34, 181^184.

Chen, J., Burbank, D.W., Scharer, K.M., Sobel, E., Yin, J.,Rubin, C. & Zhao, R. (2002) Magnetostratigraphy of theUpper Cenozoic strata in the Southwestern Chinese TianShan: rates of Pleistocene folding and thrusting. Earth Planet.Sci. Lett., 195, 113^130.

Chen, J., Heermance, R., Scharer, K.M., Burbank, D.W.,Jijun,M. &Wang, C.S. (2007) Quanti¢cation of growth andlateral propagation of the Kashi anticline, Southwest ChineseTian Shan. J. Geophys. Res., 112, doi:10.1029/2006JB004345.

Chen,S.L.,Tang,L., Jin,Z., Jia,C.&Pi,X. (2004) Thrust andfold tectonics and the role of evaporites in deformation in theWestern Kuqa Foreland of Tarim Basin, Northwest China.Marine and Petrol. Geol., 21, 1027^1042.

Couzens, B.A. & Wiltschko, D.V. (1996) The control of me-chanical stratigraphy on the formation of triangle zones. Bull.Can. Petrol. Geol., 44, 165^179.

Covey,M. (1984)Lithofacies Analysis andBasinReconstruction,Plio -PleistoceneWesternTaiwanForedeep.Petrol.Geol.Taiwan,20, 53^83.

DeCelles, P.G. (1994) Late Cretaceous^Paleocene synorogenicsedimentation and history of Sevier thrust belt, northeastUtah and southwest Wyoming. Geol. Soc. Am. Bull., 106,32^56.

DeCelles, P.G. &Giles,K.A. (1996) Foreland basins systems.Basin Res., 8, 105^123.

DeCelles, P.G.,Gray,M.B.,Ridgway,K.D.,Cole,R.B., Piv-nik,D.A., Pequera,N. & Srivastava, P. (1991b) Controls onsynorogenic alluvial-fan architecture, Beartooth Conglomer-ate (Palaeocene), Wyoming and Montana. Sedimentology, 38,567^590.

DeCelles, P., Gray, M.B., Ridgway, K.D., Cole, R.B., Sri-vastava, P., Pequera, N. & Pivnik, D.A. (1991a) Kinematichistory of a foreland basin uplift from Paleogene synorogenicconglomerate, BeartoothRange,Wyoming andMontana.Geol.Soc. Am. Bull., 103, 1458^1475.

Dumitru,T.A., Zhou,D., Chang, E.Z.,Graham, S.A.,Hen-

drix,M.S., Sobel, E. & Carroll, A.R. (2001) Uplift, exhu-mation, and deformation in the Chinese Tian Shan. In:Paleozoic and Mesozoic Tectonic Evolution of Central and EasternAsia: From Continental Assembly to Intra-continental Deformation(Ed. by M.S. Hendrix & G.A. Davis), Geol. Soc. Am. Mem., 194,71^99.

Enkin,R.J.,Yang,Z.,Chen,Y.&Courtillot,V. (1992) Paleo-magnetic contraints on the geodynamic history of the majorblocks of China from the Permian to the present. J. Geohpys.Res., 97, 13953^13989.

Flemings,P.B.& Jordan,T.E. (1990) Stratigraphic modeling offoreland basins: Interpreting thrust deformation and litho-sphere rheology.Geology, 18, 430^434.

Heermance, R. (2007) The structural and stratigraphic evolu-tion of the Neogene Kashi foreland basin, northwest China,PhD. Thesis, University of California, Santa Barbara, SantaBarbara, 268pp.

Heller, P.L., Angevine, C.L.,Winslow, N.S. & Paola, C.(1988) Two-phase stratigraphic model of foreland-basinsequences.Geology, 16, 501^504.

Hendrix, M.S., Dumitru, T.A. & Graham, S.A. (1994) LateOligoceneêarly Miocene unroo¢ng in the Chinese TianShan: an early e¡ect of the IndiaÂsia collision. Geology, 22,487^490.

Hendrix, M.S., Graham, S.A., Carroll, A.R., Sobel, E.,McNight,C.,Schulein,B.&Wang,Z. (1992) Sedimentaryrecord and climatic implications of recurrent deformation in



the Tian Shan: evidence from Mesozoic strata of the NorthTarim, South Junggar and Turpan Basins, Northwest China.Geol. Soc. Am. Bull., 104, 53^79.

Homke, S.,Verge¤ s, J., Garce¤ s, M., Emami, H. & Karpuz, R.(2004) Magnetostratigraphy of the Miocene^Pliocene Zagrosforeland deposits in the front of the Push-e Kush Arc(Lurestan Province, Iran). Earth Planet. Sci. Lett., 225,397^410.

Horton, B.K., Constenius, K.N. & DeCelles, P.G. (2004)Tectonic control on coarse-grained foreland-basin sequences:an example from theCordilleran foreland basin,Utah.Geology,32, 637^640.

Huang, J.&Chen,B.-W. (1981)Molasses in theTethys^Himala-yan tectonic domain and its relationwith the Indian PlateMo-tion. Sci. Pap. Geol. Int. Exchange, 1, 1^14 (in chinese, Englishabstract).

Huang, B., Piper, J.D.A., Peng, S., Liu,T., Li, Z.,Wang, Q. &Zhu, R. (2006) Magnetostratigraphic study of the KucheDepression, Tarim Basin, and Cenozoic uplift of the TianShan Range, Western China. Earth Planet. Sci. Lett., 251,346^364.

Jia, C., Zhang, S. &Wu, S. (2004) Stratigraphy of theTarim Basinand Adjacent Areas. Science Press, Beijing (in Chinese withEnglish abstract).

Johnson,N.M. &Mcgee,V.E. (1983) Magnetic polarity strati-graphy: Stochastic properties of data, sampling problems,and the evaluation of interpretations. J. Geophys. Res., 88,1213^1221.

Johnson,N.M.,Opdyke,N.D., Johnson,G.D.,Lindsay,E.H.

& Tahirkheli, R.A.K. (1982) Magnetic polarity stratigraphyand ages of Siwalik group rocks of the Potwar Plateau, Pakistan.Palaeogeogr. Palaeoclimatol. Palaeoecol., 37, 17^42.

Jones, C.H. (2002) User-driven integrated software lives: ‘‘Pa-leoMag’’ paleomagnetics analysis on the Macintosh. Comput.Geosci., 28, 1145^1151.

Jones,M.A.,Heller, P.L., Roca, E.,Garce¤ s,M. & Cabrera,L. (2004)Time lag of syntectonic sedimenation across an allu-vial basin: theory and example from the Ebro Basin, Spain.Basin Res., 16, 467^488.

Jones, P.B. (1996) Triangle zone geometry, terminology and ki-nematics.Bull. Can. Petrol. Geol., 44, 139^152.

Jordan,T.E. (1981) Thrust loads and foreland basin evolution,Cretaceous, western United States. Am. Assoc. Petrol. Geol.Bull., 65, 2506^2520.

Jordan, T.E., Flemings, P.B. & Beers, J.A. (1988) Datingthrust-fault activity by use of foreland-basin strata. In: NewPerspectives in BasinAnalysis (Ed. byK.L.Kleinspehn &C. Pao-la), pp. 307^330. Springer-Verlag, NewYork.

Kempf,O.,Matter,A.,Burbank,D.&Mange,M. (1999)De-positional and structural evolution of a foreland basin marginin a magnetostratigraphic framework: the eastern Swiss Mo-lasse basin. Int. J. Earth Sci., 88, 253^275.

Kirschvink, J.L. (1980) The least-squares line and plane andthe analysis of paleomagnetic data.Geophys. J. R. Astr. Soc., 62,699^718.

Liu,T.S.,Ding,M.G.&Derbyshire,E. (1996)Gravel depositson the margins of theQinghai^Xizang plateau, and their envir-onmental signi¢cance. Palaeogeogr. Palaeoclimatol. Palaeoecol.,120, 159^170.

Lourens, L.J., Hilgen, F.J., Laskar, J., Shackelton, N.J. &Wilson, D. (2005) The Neogene period. In: Geological TimeScale (Ed. by F.M. Gradstein, J.G. Ogg & A. Smith), pp. 409^440. Cambridge University Press, Cambridge.

Lo¤ pez-Blanco, M. (2002) Sedimentary response to thrustingand fold growing on the SE margin of the Ebro basin (Paleo-gene, NE Spain). Sediment. Geol., 146, 133^154.

Magee, J.W., Bowler, J.M., Miller, G.H. & Williams,D.L.G. (1995) Stratigraphy, sedimentology, chronology andpalaeohydrology ofQuaternary lacustrine deposits atMadiganGulf, Lake Eyre, South Australia. Palaeogeogr. Palaeoclimatol.Palaeoecol., 113, 3^42, doi:10.1016/0031-0182(1095)00060-Y.

McElhinny,M.W. (1964) Statistical signi¢cance of the fold testin paleomagnetism.Geophys. J. Roy. Astr. Soc., 8, 328^340.

McFadden,P.L.& Jones,D.L. (1981)The fold test in paleomag-netism.Geophys. J. Roy. Astr. Soc., 67, 53^58.

McFadden, P.L. & Mcelhinny, M.W. (1990) Classi¢cation ofthe reversal test in paleomagnetism. Geophys. J. Int., 103, 725^729.

Mccarthy, P.J.,Martini, I.P. & Leckie,D.A. (1997) Anatomyand evolution of a Lower Cretaceous alluvial plain: sedimen-tology and palaeosols in the upper Blairmore Group, south-western Alberta, Canada. Sedimentology, 44, 197^220.

Miall,A.D. (1985) Architectural-element analysis: a new meth-od of facies analysis applied to £uvial deposits. Earth Sci. Rev.,22, 261^308.

Miall, A.D. (1996) The Geology of Fluvial Deposits: SedimentaryFacies,BasinAnalysis, andPetroleumGeology. Springer,NewYork,582pp.

Molnar, P. (2001) Climate change, £ooding in arid environ-ments, and erosion rates.Geology, 29, 1071^1074.

Molnar,P. (2004) LateCenozoic increase in accumulation ratesof terrestrial sediment: how might climate change have af-fected erosion rates?Annu. Rev. Earth Planet. Sci., 32, 67^89.

Molnar, P., Brown, E.T., Burchfiel, B.C., Deng, Q., Feng,X., Li, J., Raisbeck, G.M., Shi, J.,Wu, Z., Yiou, F. & You,H. (1994) Quaternary climate change and the formation of riv-er terraces across growing anticlines on the north £ank of theTien Shan, China. J. Geol., 102, 583^602.

Molnar,P.&Tapponnier,P. (1975)CenozoicTectonics ofAsia:e¡ects of a continental collision. Science, 189, 419^426.

Me¤ tivier,F.,Guademer,Y.,Tapponier,P.&Klein,M. (1999)Mass accumulation rates inAsia during theCenozoic.Geophys.J. Int., 137, 280^318.

Ojha,T., Butler, R.F.,Quade, J.,Decelles, P.G., Richards,D. & Upreti, B.N. (2000) Magnetic Polarity Stratigraphy ofthe Neogene Siwalik Group at Khutia Khola, Far WesternNepal.Geol. Soc. Am. Bull., 112, 424^424.

Ort|¤ ,F.,Rosell,L.&Anado¤ n,P. (2003)Deep to shallow lacus-trine evaporites in the Libros gypsum (SouthernTeruel Basin,Miocene,NeSpain): an occurrence of PelletalGypsumRhyth-mites. Sedimentology, 50, 361^386, doi:310.1046/j.1365-3091.2003.00558.x.

Paola, C. (1988) Subsidence and gravel transport in alluvial ba-sins. In:NewPerspectives inBasinAnalysis (Ed. byK.Kleinspehn& C. Paola), pp. 231^244. Springer-Verlag, NewYork.

Paola, C., Heller, P.L. & Angevine, C.L. (1992) Large-scaledynamics of grain-size variation in alluvial basins, 1: theory.Basin Res., 4, 73^90.

Quidelleur,X. & Courtillot,V. (1996) On low-degree sphe-rical harmonic models of paleosecular variation. Phys. EarthPlanet. Interiors, 95, 55^77.

Reigber, C., Michel, G.W., Galas, R., Angermann, D.,Klotz, J.,Chen, J.Y., Papschev,A.,Arslanov,R.,Tzurkov,V.E. & Ishanov,M. (2001) New space geodetic constraints onthe distribution of deformation in Central Asia. Earth Planet.Sci. Lett., 191, 157^165.



Scharer, K., Burbank, D.W., Chen, J. & Weldon, R. (2006)Kinematic models of £uvial terraces over active detachmentfold: constraints on the growth mechanism of the Kashi-Atushi fold system, ChineseTian Shan. Geol. Soc. Amer. Bull.,118, 1006^1021.

Scharer,K.M., Burbank,D.W., Chen, J.,Weldon, R.J., Ru-bin,C.,Zhao,R.&Shen, J. (2004)Detachment folding in theSouthwestern Tian Shan -Tarim foreland, China; shorteningestimates and rates. J. Struct. Geol., 26, 2119^2137.

Schlunegger, F., Matter, A., Burbank, D.W. & Klaper,E.M. (1997) Magnetostratigraphic constraints on relation-ships between evolution of the Central Swiss MolasseBasin and Alpine Orogenic Events. Geol. Soc. Am. Bull., 109,225^241.

Sklar, L., Dietrich, W.E., Foufuoula-Georgiou, E., La-shemes,B.& Bellugi,D. (2006) Do gravel bed river size dis-tributions record channel network structure? Water Resourc.Res., 42,W06D18, doi:10.1029/2006WR005035.

Sobel,E.,Chen, J.&Heermance,R.V. (2006) LateOligoceneÊarly Miocene initiation of shortening in the SouthwesternChineseTian Shan: implications for Neogene shortening ratevariations. Earth Planet. Sci. Lett., 247, 70^81.

Sobel, E.R. & Dumitru,T. (1997) Thrusting and exhumationaround the margins of the westernTarim basin during the In-diaÂsia collision. J. Geophys. Res., 102, 5043^5064.

Sobel, E.R. & Strecker, M.R. (2003) Uplift, exhumation andprecipitation: tectonic and climatic control of Late Cenozoiclandscape evolution in the northern Sierras Pampeanas, Ar-gentina.Basin Res, 15, 431^451.

Sun, J. & Liu, T. (2006) The age of the Taklimakan Desert.Science, 312, 1621.

Sun, J., Zhu, R.X. & Bowler, J. (2004) Timing of theTianshanMountains uplift constratined by magnetostratigraphic analy-sis of molasse deposits. Earth Planet. Sci. Lett., 219, 239^253.

Suppe, J. (1983) Geometry and kinematics of fault bend folding.Amer. J. Sci., 283, 648^721.

Thompson, S.C.,Weldon, R.J., Rubin, C.M., Abdrakhma-tov,K.,Molnar, P. & Berger,G.W. (2002) Late Quaternaryslip rates across the central Tien Shan, Kyrgyzstan, centralAsia. J. Geophys. Res., 107, ETG7.1ÊTG7.32.

Uba, C.E., Heubeck, C. & Hulka, C. (2006) Evolution of thelate Cenozoic Chaco foreland basin, Southern Bolivia. BasinRes., 18, 145^170.

Vandervoort,D.S. (1997) Stratigraphic response to saline lake-level £uctuations and the origin of cyclic nonmarine evaporitedeposits: the Pleistocene Blanca Lila Formation, northwestArgentina.Geol. Soc. Am. Bull., 109, 210^224.

Wang,Q., Zhang, P.-Z., Freymueller, J.T.,Bilham,R.,Lar-son,K.M., Lai,X.a.,You,X.,Niu, Z., Jianchun,W., Li,Y.,Liu, J. & Yang, Z. (2001) Present-day crustal deformation inChina constrained by global positioning system measure-ments. Science, 294, 574^577.

Ward, P.D.,Garrison,G.,Botha, J.,Buick,R., Erwin,D.H.,Kirschvink, J.L., De Kock, M.O. & Smith, R. (2005)Abrupt and gradual extinction among land Vertebrates in theKaroo Basin, South Africa. Science, 307, 709^714.

Windley,B.F.,Allen,M.B.,Zhang,C.,Zhao,Z.-Y.&Wang,G.-R. (1990) Paleozoic accretion and Cenozoic redeformationof the Chinese Tien Shan Range, central Asia. Geology, 18,128^131.

Yin, A., Nie, S., Craig, P., Harrison, T.M., Ryerson, F.J.,Qian,X. & Yang,G. (1998) Late Cenozoic tectonic evolutionof the southern ChineseTian Shan.Tectonics, 17, 1^17.

Yin, A., Rumelhart, P.E., Butler, R., Cowgill, E., Harri-

son,T.M.,Foster,D.A., Ingersoll,R.V.,Zhang,Q.,Zhou,X.Q.,Wang, X.F., Hanson, A. & Raza, A. (2002) Tectonichistory of the Altyn Tagh fault system in northern Tibet in-ferred from Cenozoic sedimentation. Geol. Soc. Am. Bull., 114,1257^1295.

Zhang, P.,Molnar, P. & Downs,W.R. (2001) Increased sedi-mentation rates and grain sizes 2^4Myr ago due to the in£u-ence of climate change on erosion rates.Nature, 410, 891^897.

Zheng,H.,Powell,C.M.,An,Z.,Zhou, J.&Dong,G. (2000)Pliocene uplift of the northern Tibetan Plateau. Geology, 28,715^718.

Zhou,Z.Y.&Chen, P.J. (1990)Biostratigraphy andGeological Evo-lution in theTarimBasin. Geological Press, Beijing.

Colombo, F. (1994) Normal and reverse unroo¢ng sequences insyntectonic conglomerates as evidence of progressive basin-ward deformation.Geology, 22, 235^238.

Manuscript received 10 August 2006; Manuscript accepted 26August, 2007.

APPENDIX A

Palaeomagnetic analysis

Specimens from the northern and medial foreland, as wellas from theAtushiAnticline andKashiTown section,wereanalysed at the California Institute of Technology and Oc-cidental College palaeomagnetic laboratories, both lo-cated in Los Angeles, CA. Both laboratories haveidentical sample changer and magnetometer set-ups. Re-manent magnetization was measured using a three-axisDCSQUID moment magnetometer system housed in amagnetically shielded m-metal room. The magnetometerhas a background noise ofo1pAm2 and is equippedwitha vacuum pick-and-put, computer-controlled samplehandling system, which can measure up to 180 samplesautomatically (Ward et al., 2005). Alternating ¢eld (AF) de-magnetization was performed with a computer-con-trolled, three-axis coil system. Thermal demagnetizationwas performed in a commercially built, magneticallyshielded furnace.

The Kashi West and the Upper Boguzihe specimenswere measured at the Laboratory for Palaeo- and RockMagnetism, GeoForschungsZentrum Potsdam using afully automated 2G Enterprises 755 SRM DC-SQUIDcryogenic long-core magnetometer with a backgroundnoise level of �300^400 pAm2. Despite higher back-ground noise levels than the California labs, the samplesprocessed in Potsdamyielded clear remanence directions.

The intensity of the natural remanent magnetism(NRM) for these specimenswas �10� 5^10� 6 emu cm� 3.Progressive demagnetization successfully resolved multi-ple components of magnetization. For most specimens,equal-area projection and orthogonal vector plots revealtwo clear components: a high-temperature component(HTC) and a low-temperature component (LTC) (Fig.11).TheLTC,which is a low-coercivity, low-blocking tem-



perature component, is removed by AF demagnetizationat 30 or 60G and thermal demagnetization at 150 or300 1C (e.g. Fig. 11b). This LTC has an in situ declinationand inclination similar to the present day magnetic ¢elddirection of 3.561E inclined 58.71 (Fig. A1i and j), althoughsigni¢cant scatter observed in the LTC directions implies

a more complex overprint than simply the present-day¢eld.Typically, the LTC does not decay towards the origin(Fig.11a^f).TheHTCwas completely unblocked by680 1C(Fig.11a, b and f), indicating a hematite carrier of the mag-netic remanance, although magnetite is likely present aswell. At times the HTC would cluster in either a normal

QIGAILIKE & AYAKEQIANAKEKETAMU-COMPOSITE

Geographic

OVERPRINTGeographic Geographic

Tilt-corrected

Geographic

Geographic Tilt-corrected

(a) (b)

(c) (d)

(f)

(i) (j)

Tilt-corrected

(g) (h)

QIGAILIKE

AYAKEQIANA

FOLD TEST

Tilt-corrected

Geographic

KEKETAMU-COMPOSITE

KC

KC

Y

L

L

Y

KC

YL

KCY

L

(e)

Fig. A1. Stereonet plots ofpalaeomagnetic data. Grey(inclined down) andwhite(inclined up) circlesindicate the a95 erroraround the Fisher mean.Fisher mean data are listedadjacent to each plot.Thegrey star is the location ofthe present-day magnetic¢eld (decl: 3.561, incl:58.71) for 39.61N75.91E.(a^f) Geographic and tilt-corrected least- squares ¢tsfor the Keketamucomposite (KC), Qigailike(L) and Ayakeqiana (Y)sections. (g, h) Fold test(McFadden & Jones, 1981)data showing the pre- andpost-tilt correctiongroupings for the normaland reversed zones forKC, L and Y. (i, j) Non-tilt-corrected secondaryremanence directionsfor the KC and thecombined Yand L.Thesesecondary overprintdirections are similar tothe present-day ¢elddirections, and representthe viscous overprint ofthe deformed strata.



(Fig. 11c) or reversed position before decaying to theorigin.

Coercivity spectrum analysis of isothermal remanentmagnetism (IRM) for typical samples indicates that mostIRM for the samples is acquired at low magnetizing ¢eldstrengths (o2000Oe), consistentwith magnetite and tita-nomagnetite as the primary carrier for the magnetism (Fig11e and f). However, IRM continues to be acquired athigher ¢eld strengths, consistent with a hematite contri-bution to the characteristic remanence (ChRM) as well.For example, the rapid acquisition of most IRM (450%)by 1000Oe in sample K188.1 (Fig. 11h) is consistent withprimarily magnetite as the carrier. In contrast, sampleK70.3 (Fig.11e) has a much £atter curve and a smaller per-centage (o50%) of their IRM is acquired by 1000Oe,consistent with observed complete unblocking tempera-ture of 680 1C and a hematite and magnetite HTC carrierfor these samples (Fig.11d and f ).

To determine ChRM directions, principal componentanalysis (Kirschvink, 1980) using least- squares ¢ts wasperformed for each specimen based on data from aminimum of four, and more typically ¢ve to eight, tem-perature steps. The data were analysed using palaeomag-netic software of Jones (2002) and N. R. Nowaczyk (pers.comm., 2004).The origin was includedwhen stepwise de-magnetization showed a progression towards it (Fig. 11cand d). Typical maximum angular deviation (MAD)was between 1 and 101, but a MAD of 151 was accepted, ifan unambiguous north or south polarity could be inter-preted.When HTC components clustered in a normal orreversed polarity position, the ChRMdirection was deter-mined by forcing a line from the cluster through the origin(Fig.11c).

Analytical tests of the palaeomagnetic data

The reversal test (McFadden and McElhinny, 1990) wasused to assess whether the ChRM had likely been isolatedfor each site. Both the Ayakeqiana (Fig. A1f) and Qigailike(Fig. A1d) sections pass this test at C quality, but the Keke-tamu composite (comprising data from the Reservoir, Re-servoir Bypass, ASE, and ASW) section does not (Fig.

A1b). The failing test results from variation of 17 and 131for both inclination and declination, respectively (Fig.A1b). Similar contrasts in inclination were observed at theAyakeqiana (Fig. A1f), Qigailike (Fig. A1d), KashiWest andKashiTown sections (Chen etal., 2007), and likely were thecause of theC-quality tests, rather thanAorBquality.Thisdi¡erence is most likely due to an unremoved overprint(McElhinny,1964;Quidelleur&Courtillot,1996;Charreauet al., 2005) that, given the folding geometry and presentmagnetic ¢eld direction, can steepen or £atten the normalpolarity directions with respect to the reversed polarity di-rections. Given that the normal and reversed polarity di-rections all plot within their respective hemispheres, thatthe sample populations are qualitatively antipodal and ac-knowledging that some overprint component has not beenremoved, we accept our polarity determinations despitethe failure of the reversal test for the Reservoir section.

In the fold test, we utilize the multiple-limb techniquefrom McFadden & Jones (1981) and compare the normaland reversed polarity zones from three separate fold limbs:north limb of theTashipishake Anticline, north and southlimbs of the Keketamu anticline (Fig. 2a).We kept the nor-mal and reversed zones separate to minimize the e¡ect ofthe unremoved overprint on the outcome of the test. Stra-ta that span approximately the same time, the base of theReservoir section (400^1400m stratigraphic height), theentire Ayakeqiana section, and Qigailike sections (Fig.A1g^h), pass the fold test at 95% level of con¢dence withan observed F statistic of 2.37 and 2.28 for normal and re-versed polarities, respectively (fcrit5 2.38, N5 382).Furthermore, samples from the same stratigraphic level(bed) always had consistent polarity directions, despitetheir collection o5m from each other, indicating themagnetozones were layer bound.The combination of thepositive results from the fold and reversal tests, as well asthe presence of layer-bound magnetic polarity zones indi-cate that these ChRM directions were acquired by the se-diments at or soon after deposition. The expecteddeclination and inclination for normal polarity samplesfrom the late Miocene (7.9Ma) is D5 7.51 I5 601 (Enkinet al., 1992), similar to the present-day ¢eld of D5 3.561,I5 58.71.



Documents

BasinResearch ChronologyandtectoniccontrolsofLateTertiary