ABSTRACT

Sedimentological and provenance datafrom the lower Miocene–Pliocene Dumri For-mation and Siwalik Group in western Nepalprovide new information about the timing ofthrust faulting and the links between erosionalunroofing of the Himalaya and the Cenozoic87Sr/86Sr record of the ocean. In westernNepal, the Dumri Formation is an ~750–1300-m-thick fluvial sandstone and overbank mud-stone unit. The Siwalik Group is >4200 mthick and consists of a lower member (>850m) of 2–12-m-thick fluvial channel sandstonesand oxidized calcareous paleosols, a middlemember (>2400 m) of very thick (>20 m)channel sandstones and mainly organic-richHistosols, and an upper member (>1000 m)composed of gravelly braided river deposits.Paleocurrent data indicate that middleMiocene–Pliocene rivers in western Nepalflowed southward, transverse to the thrustbelt, throughout deposition of the SiwalikGroup. No evidence was found for an axialfluvial trunk system (i.e., the paleo-GangesRiver) in Siwalik Group sandstones. A majorincrease in fluvial channel size is recorded bythe transition from lower to middle Siwalikmembers at ~10.8 Ma, probably in response toan increase in seasonal discharge.

Modal petrographic data from sandstonesin the Dumri Formation and the Siwalik

Group manifest an upsection enrichment inpotassium feldspar, carbonate lithic frag-ments, and high-grade metamorphic miner-als. Modal petrographic analyses of modernriver sands provide some control on potentialsource terranes for the Miocene–Pliocenesandstones. The Dumri Formation was mostlikely derived from erosion of sedimentaryand low-grade metasedimentary rocks in theTibetan (Tethyan) Himalayan zone duringearly Miocene emplacement of the Main Cen-tral thrust. The presence in Dumri sandstonesof plagioclase grains suggests exposure ofcrystalline rocks of the Greater Himalayanzone, perhaps in response to tectonic unroof-ing by extensional detachment faults of theSouth Tibetan detachment system. Duringdeposition of the lower Siwalik Group(~15–11 Ma), emplacement of the Dadel-dhura thrust sheet (one of the synformal crys-talline thrust sheets of the southern Himalaya)on top of the Dumri Formation suppliedabundant metasedimentary lithic fragmentsto the foreland basin. A steady supply of pla-gioclase grains and high-grade minerals wasmaintained by deeper erosion into the MainCentral thrust sheet. From ~11 Ma to thepresent, K-feldspar sand increased steadily,suggesting that granitic source rocks becamewidely exposed during deposition of the up-per part of the lower Siwalik Group. Thisprovenance change was caused by erosion ofpassively uplifted granites and granitic ortho-gneisses in the Dadeldhura thrust sheet abovea large duplex in the Lesser Himalayan rocks.Since the onset of deposition of the conglom-eratic upper Siwalik Group (~4–5 Ma), fault

slip in this duplex has been fed updip andsouthward into the Main Boundary and MainFrontal thrust systems.

We obtained 113 U-Pb ages on detrital zir-cons from modern rivers and Siwalik Groupsandstones that cluster at 460–530 Ma,~850–1200 Ma, ~1.8–2.0 Ga, and ~2.5 Ga. Anabundance of Cambrian–Ordovician grainsin the Siwalik Group suggests sources of Siwa-lik detritus in the granites of the Dadeldhurathrust sheet and possibly the Greater Hima-layan orthogneisses. The older ages are consis-tent with sources in the Greater and LesserHimalayan zones. An overall upsection in-crease in zircons older than 1.7 Ga suggests in-creasing aerial exposure of Lesser Himalayanrocks. None of the detrital zircons (even in themodern river samples) yielded a Cenozoic agethat might suggest derivation from the Ceno-zoic Greater Himalayan leucogranites, butthis may be attributable to the inheritanceproblems that characterize the U-Pb geo-chronology of the leucogranites.

When compared with recent studies of the87Sr/86Sr composition of paleosol carbonatenodules and detrital carbonate in paleosolsfrom the Siwalik Group, the provenance datasuggest that erosion and weathering of meta-morphosed carbonate rocks in the LesserHimalayan zone and Cambrian–Ordoviciangranitic rocks of the crystalline thrust sheetsin central and eastern Nepal may haveplayed a significant role in elevating the87Sr/86Sr ratio of middle Miocene synoro-genic sediments in the Indo-Gangetic fore-land basin and the Bengal fan, as well asglobal seawater.

2

Neogene foreland basin deposits, erosional unroofing, and the kinematichistory of the Himalayan fold-thrust belt, western Nepal

P. G. DeCelles*G. E. GehrelsJ. QuadeT. P. OjhaP. A. Kapp†

} Department of Geosciences, University of Arizona, Tucson, Arizona 85721

B. N. Upreti Department of Geology, Tribhuvan University, Tri-Chandra Campus, Ghantaghar, Kathmandu, Nepal

GSA Bulletin;January 1998; v. 110; no. 1; p. 2–21; 15 figures; 2 tables.

*e-mail: decelles@geo.arizona.edu†Present address: Department of Earth and Space

Sciences, University of California, Los Angeles, Cali-fornia 90024.

Data Repository item 9805 contains additional material related to this article.

INTRODUCTION

Many studies have linked the uplift of theHimalaya and Tibetan Plateau during Cenozoictime to regional and global climate changes andmajor changes in the oceanic 87Sr/86Sr record(Ruddiman and Raymo, 1988; Quade et al.,1989; Richter et al., 1992; Edmond, 1992).Weathering of metamorphic and igneous rocks,rich in radiogenic Sr and widely exposed in thehigher parts of the Himalaya, is proposed to haveconsumed atmospheric CO2 and thus contributedto global temperature decline (Raymo andRuddiman, 1992). The same process is thought tohave been responsible for the rise in oceanic87Sr/86Sr that has occurred since ~40 Ma (Hodellet al., 1989; Richter et al., 1992; Hodell andWoodruff, 1994). Geochemical data sets are nowavailable from several regions of the Himalayaand from the erosional detritus stored in the fore-land basin sediments of the Miocene SiwalikGroup along the southern flank of the orogenicbelt and in the Bengal fan (Quade, 1993; France-Lanord et al., 1993; Harrison et al., 1993; Quadeet al., 1995, 1997; Derry and France-Lanord,1996). However, geochemical models for un-roofing of the Himalaya have not been testedwith even the most basic provenance and sedi-mentological data from the foreland basin depos-its or by considerations of the regional structuraldevelopment of the Himalayan thrust belt.

In order to help fill the gap between geo-chemical and geological evidence, this paperpresents data from the lower Miocene(?) DumriFormation and Miocene–Pliocene SiwalikGroup in western Nepal (Fig. 1). The databaseconsists of standard sedimentological analysisof remarkably complete sections of the SiwalikGroup, including more than 1200 paleocurrentmeasurements, standard modal petrographicanalyses of Dumri and Siwalik sandstone sam-ples, and U-Pb dates from single zircon grainsfrom Siwalik sandstones. Modal petrographicdata and U-Pb zircon dates from modern riversands are presented as a means of comparingcompositions of sands from known source ter-ranes with those of the ancient sediments. Thesedimentological and provenance data are con-sidered within the context of a new regional bal-anced cross section, and a preliminary assess-ment of the kinematic history of thrusting since~22 Ma is presented. Our purpose is to shedsome light on the locations and timing of expo-sure of source terranes for the Dumri Formationand the Siwalik Group, in an effort to help con-strain models for the geodynamics of the NepalHimalaya and to provide information about po-tential sources of highly radiogenic Sr thatmight have contributed to the abrupt Cenozoicrise in oceanic 87Sr/86Sr.

REGIONAL STRUCTURE AND SOURCETERRANE COMPOSITIONS

Regional Geology and Structure

In Nepal and northern India, the Himalaya isdivided into four lithotectonic zones that are sep-arated by major structural discontinuities. Fromnorth to south, these are the Tibetan (or Tethyan)zone, the Greater Himalayan zone, the LesserHimalayan zone, and the Subhimalayan zone(e.g., Gansser, 1964; Valdiya, 1980; Fig. 1). In or-der to better understand the relationships betweenunroofing of these lithotectonic zones and theforeland basin deposits, we have constructed abalanced regional cross section on the basis of the1:250 000 geologic map compiled by Shrestha etal. (1987) and our own field data from a traversealong the Dadeldhura-Baitadi road (Figs. 1 and2). The cross section is line-length balanced, butno attempt was made to incorporate the locallyintense small-scale deformation that character-izes some of the stratigraphic units (particularlythe Galyang Formation). No subsurface data areavailable, the stratigraphy of the Dadeldhurathrust sheet is poorly known, and hanging-wallcut-offs are generally not preserved. Therefore,this cross section should be viewed only as a first-order approximation that will undoubtedly bechanged as new data become available. Never-theless, the cross section honors known surfacegeologic relationships, and the most importantfeatures, such as the synformal Dadeldhura thrustsheet and the large hinterland-dipping duplex toits north, are supported by thrust branching pat-terns and bedding and foliation dip data. Similarlarge-scale structures are present on the cross sec-tion of Schelling (1992) for eastern Nepal andthat of Srivastava and Mitra (1994) for northernIndia. Brief descriptions of each of the lithotec-tonic zones are provided in the following, in ad-dition to a summary of available structural andkinematic information.

The Tibetan zone consists of Cambrian toEocene marine sedimentary to low-grade meta-sedimentary rocks (mainly phyllite, quartzite,limestone, and local volcanic rocks) in a generallysouthward verging thrust belt that has a minimumof ~135 km of shortening in central-southernTibet (Burg and Chen, 1984; Searle, 1986;Ratschbacher et al., 1994). These rocks were de-posited along the precollisional northern marginof the Indian plate. The southern boundary of theTibetan zone is marked by the South Tibetandetachment system, a complex of northward-dipping normal detachment faults that has ~10 kmof structural relief and several tens of kilometersof top-to-the-north displacement (Burg and Chen,1984; Burchfiel et al., 1992; Coleman, 1996).

The Greater Himalayan zone consists of a

5–20-km-thick sheet of amphibolite-grade(kyanite and sillimanite bearing) schist, para-gneiss, and orthogneiss (Fig. 3; Hodges andSilverberg, 1988; Pêcher, 1989; Schelling, 1992;Macfarlane et al., 1992; Hodges et al., 1996;Coleman, 1996; and many others). In westernNepal these rocks are referred to as the HimalGroup. Large tourmaline-bearing leucograniticplutons of middle to late Cenozoic age are dis-tributed along the highest part of the Himalaya inthe northern part of the Greater Himalayan zone(LeFort, 1981, 1986; Ferrara et al., 1983; Schäreret al., 1986; Deniel et al., 1987; France-Lanordand LeFort, 1988). Because these plutons, andthe gneisses they intrude, have yielded extremelyhigh 87Sr/86Sr ratios (Deniel et al., 1987; France-Lanord and LeFort, 1988), they are believed to bean important source of radiogenic Sr driving theCenozoic rise in seawater 87Sr/86Sr (Edmond,1992). The southern boundary of the GreaterHimalayan zone is the Main Central thrust sys-tem, which is a several-kilometer-thick zone ofnumerous thrust faults and intense shear strain(Brunel and Kienast, 1986; Macfarlane et al.,1992; Hodges et al., 1996).

Several large, generally synformal thrust sheetscomposed of garnet- to biotite-grade schist andorthogneiss in their lower parts and low-grademetasedimentary to unmetamorphosed sedimen-tary rocks in their upper parts are located on top ofthe Lesser Himalayan zone rocks to the south ofthe Greater Himalayan zone (Figs. 1 and 2).These are known as the crystalline, or LesserHimalayan, nappes or thrust sheets (e.g., Gansser,1964; Valdiya, 1980). Late Cambrian–EarlyOrdovician (Schärer and Allègre, 1983; Einfalt etal., 1993) plutons of two-mica, cordierite-bearinggranite are present in several of these thrust sheets(e.g., LeFort et al., 1986; Einfalt et al., 1993). Inwestern Nepal and adjacent Kumaon, India, theDadeldhura-Almora thrust sheet occupies a largepart of the source terrane directly north of the sec-tions of the Siwalik Group that we studied (Figs. 1and 2; Valdiya, 1980). The Dadeldhura thrustsheet consists of an ~10-km-thick sequence ofPrecambrian phyllite, quartzite, slate, sericiticquartzite, garnet-mica schist, quartz-feldspar-mica schist, chloritic schist, augen gneiss, biotitegneiss, and local metavolcanic units (Fig. 3;Bashyal, 1986; Einfalt et al., 1993; Upreti,1996a). Comparison with the Kathmandu crys-talline thrust sheet in east-central Nepal suggeststhat several kilometers of metasedimentary rockhave been eroded from the upper part of theDadeldhura sheet. A large Cambrian–Ordoviciangranitic pluton, the Dadeldhura granite, crops outin the southern and central parts of the thrust sheet(Fig. 1; Einfalt et al., 1993). The Dadeldhurathrust sheet as a whole is folded into a northwest-southeast–trending synform (Figs. 1 and 2). In

KINEMATIC HISTORY OF THE HIMALAYAN FOLD-THRUST BELT

Geological Society of America Bulletin, January 1998 3

Nepal and northern India, most workers have in-terpreted the crystalline thrust sheets as the south-ern continuations of the Greater Himalayan zonerocks (e.g., Gansser, 1964; Stöcklin, 1980;Valdiya, 1980; Schelling, 1992). Upreti (1996a)argued that in western Nepal, the Dadeldhura

thrust sheet is structurally distinct from theGreater Himalayan rocks in the Main Centralthrust hanging wall, displaying a lower grade ofmetamorphism (garnet and lower) and an associ-ation with the Cambrian–Ordovician granites.However, the distinction between the rocks of the

Greater Himalayan zone and the Dadeldhurathrust sheet is not straightforward on the basis ofthese criteria, because granitic orthogneisses inthe Greater Himalayan zone have produced LateCambrian Rb/Sr ages (Ferrara et al., 1983; LeFortet al., 1986) and U-Pb data that are consistent with

DECELLES ET AL.

4 Geological Society of America Bulletin, January 1998

Dhangadhi

DA

B

KARNALI R.

KARNALI R. Tansen

DB

Surkhet

SWK

SKButwal

MAH

AKA

LIR

.

29°N

30°N

30°N

RT

DT

MCT

X'

X

MCT

MFT

KK

N

MBT

81°E

81°E

MK

BK

GALYANG - LAKARPATA FMS. (PALEOZOIC)

DAMGAD AND MELMURA FMS.(LATE PRECAMBRIAN-EARLY PALEOZOIC)

KUSHMA & RANIMATA FMS. (PRECAMBRIAN)

BURHI GANDAKI & SALYANI GAD GNEISSES,KALIKHOT FM. (PRECAMBRIAN)

HIMAL GROUP GNEISSES & PARAGNEISSES(PRECAMBRIAN)

MELPANI, BHAINSKATI, & DUMRI FMS.(CRETACEOUS-EARLY MIOCENE)

SIWALIK GROUP (MIOCENE - PLIOCENE)

TERTIARY LEUCOGRANITE

Greater Himalayan Zone

Lesser Himalayan Zone

Subhimalayan Zone

GRANITE (CAMBRIAN-ORDOVICIAN)

Crystalline Thrust Sheets

100 km

PHYLLITES AND CARBONATES(CAMBRIAN-EOCENE)

Tibetan Himalayan Zone

Tulsipur

INDIA

N

BANGLADESH

TIBETNEPAL

Figure 1. Generalized geologic map of western Nepal modified from Amatya and Jnawali (1994). The locations of Macheli Khola (MK), KhutiaKhola (KK), Babai Khola (BK), Swat Khola (SWK), Surai Khola (SK), and Dumri Bridge (DB) sections are shown. Major thrust faults are shownas lines with barbs on hanging-wall side (dashed where not continuously exposed), and are abbreviated as follows: MFT—Main Frontal thrust;MBT—Main Boundary thrust; RT—Ramgarh thrust; DT—Dadeldhura thrust; MCT—Main Central thrust. Star indicates location of samplefrom Dadeldhura granite that yielded a U-Pb age of 492 ± 6 Ma. Villages: DA—Dadeldhura; B—Baitadi. Line X–X′ indicates location of cross sec-tion shown in Figure 2.

an ~500 Ma age in the Annapurna Range of north-central Nepal (Hodges et al., 1996). It is thereforeplausible that the Dadeldhura thrust sheet is thesouthward, structurally shallower, continuation ofthe hanging wall of the Main Central thrust. Fromthe standpoint of source terrane composition,however, the Dadeldhura thrust sheet, with itsthick metasedimentary stratigraphic section andlower grade of metamorphism, is different fromthe MCT sheet (Fig. 3).

To the south of the Greater Himalayan zone,and surrounding the crystalline thrust sheets, is theLesser Himalayan zone, which consists of Prot-erozoic(?) to Devonian phyllite, quartzite, lime-stone, and dolostone (Valdiya, 1980; Stöcklin,1980; Bashyal, 1986; Upreti, 1990, 1996b; Fig. 3).The southern boundary of the Lesser Himalayanzone is the Main Boundary thrust, and its internal

structure to the north of the crystalline thrust sheetsis characterized by a large antiformal duplex (incentral Nepal; Schelling, 1992) or hinterland-dipping duplex (in western Nepal and northernIndia; Srivastava and Mitra, 1994; Fig. 2). The de-velopment of this duplex probably was responsiblefor folding of the overlying crystalline thrustsheets (Dhital and Kizaki, 1987; Schelling, 1992;Srivastava and Mitra, 1994). South of the Dadel-dhura thrust sheet, the Lesser Himalayan rockscrop out in a narrow belt between the MainBoundary thrust and the northward-dippingRamgarh thrust below the base of the Dadeldhurathrust sheet (Fig. 2; Valdiya, 1980; Shrestha et al.,1987; Srivastava and Mitra, 1994). Active seis-micity in the Himalayan thrust belt is concentratedat depths of 5–20 km along a major ramp in thebasal Himalayan thrust (Ni and Barazangi, 1984;

Zhao et al., 1993; Pandey et al., 1995). This rampis located below the duplex in Lesser Himalayanrocks (Fig. 2; Pandey et al., 1995). The maximumrates of surface uplift and horizontal convergencebetween India and southern Tibet are also locatedabove this zone of active seismicity, which projectsvertically upward to the surface trace of the MainCentral thrust (Bilham et al., 1997). Whereas someof the uplift may be a response to Main Centralthrust displacement, the microseismic data suggestthat displacement may also be occurring along theramp in the basal thrust (Pandey et al., 1995). Thiswould imply that the duplex is still growing. Thekinematics implied by the cross section (Fig. 2)transfer displacements on individual thrusts withinthe duplex updip and southward into the frontalparts of the Ramgarh, Main Boundary thrust, andMain Frontal thrust systems.

KINEMATIC HISTORY OF THE HIMALAYAN FOLD-THRUST BELT

Geological Society of America Bulletin, January 1998 5

TIBETAN (TETHYS)

B.

DEFORMED LENGTH = 115 kmRESTORED LENGTH = 343 kmTOTAL SHORTENING = 228 km (66%)

20 km

MAIN FRONTAL DADELDHURA-RAMGARH

MAIN CENTRAL

SIWALIK GROUP

MAIN BOUNDARYSWAT & DUMRI FMS.

LAKARPATA GROUP

GALYANG FM.

RANIMATA FM.

KUSHMA FM.

MELMURA & DAMGAD FMS.

SALYANI GAD GNEISS

DADELDHURA GRANITE

KALIKHOT FM.

VAIKRITA (HIMAL) GROUP

MFT

MB

T RT

MC

T

DT

STD

S

-10

-5

0 km

5

-10

0 km

-5

5

-15

SOUTH NORTHXX'

A.DADELDHURA

DT

MB

T

MCT

DADELDHURATHRUST SHEET

LESSER HIMALAYAGREATERHIMALAYA

TIBETANHIMALAYA

SUBHIMALAYA

20 k

m

Figure 2. Regional structural cross section X–X′ of the Himalayan fold-thrust belt in western Nepal (see Fig. 1 for location). MFT—MainFrontal thrust; MBT—Main Boundary thrust; RT—Ramgarh thrust; DT—Dadeldhura thrust; MCT—Main Central thrust; STDS—SouthTibetan detachment system. Eroded rocks in the lower part of the Dadeldhura thrust sheet are shown above present Lesser Himalayan topogra-phy in order to emphasize the minimum amount of sediment derived from the Dadeldhura sheet that was available to Miocene–Pliocene deposi-tional systems of the Indo-Gangetic foreland basin system. Note the difference in scale between deformed- and restored-state cross sections.

6G

eological Society of A

merica B

ulletin,January 1998

2

0

4

6

8

10

12

14

16

18

0

TWO-MICATOURMALINE-BEARINGLEUCO-GRANITES;MIOCENE

MICASFELDSPARS

RO

LWA

LIN

G-K

HU

MB

UP

AR

AG

NE

ISS

ES

:B

IOT

ITE

PA

RA

GN

EIS

S; C

ALC

-S

ILIC

AT

E G

NE

ISS

; MA

RB

LE;

GR

AN

ITIC

GN

EIS

S &

AU

GE

NG

NE

ISS

; PR

EC

AM

BR

IAN

RO

LWA

LIN

G-K

HU

MB

U-K

AN

GC

HE

NJU

NG

AM

IGM

AT

ITE

S:

GR

AN

ITIC

, MIG

MA

TIC

OR

TH

OG

NE

ISS

WIT

HIN

TE

RC

ALA

TE

D P

AR

AG

NE

ISS

AN

D L

OC

AL

AM

PH

IBO

LIT

ES

; PR

EC

AM

BR

IAN

JUN

BE

SI G

NE

ISS

: FO

LIA

TE

D G

AR

NE

T-B

IOT

ITE

SC

HIS

T;

FE

LDS

PA

TH

IC G

NE

ISS

;C

ALC

-SIL

ICA

TE

SC

HIS

T;Q

UA

RT

ZIT

E;A

MP

HIB

OLI

TE

GR

AN

ITIC

AU

GE

N G

NE

ISS

;P

RE

CA

MB

RIA

N

BIOTITEFELDSPARSSILLIMANITE

BIOTITEMUSCOVITEFELDSPARSSILLIMANITE

KY

AN

ITE

SIL

LIM

AN

ITE

FE

LDS

PA

RS

MIC

AS

, GA

RN

ET

QUARTZBIOTITEMUSCOVITEFELDSPARS

MUSCOVITEBIOTITECHLORITESTAUROLITEFELDSPARSKYANITEGARNET

MCT SYSTEM: GREATERHIMALAYAN ZONE

DT SYSTEM: CRYSTALLINETHRUST SHEET

MBT SYSTEM: LESSERHIMALAYAN ZONE

MFT SYSTEM:SUBHIMALAYAN ZONE

QUARTZCHLORITEMICAPHYLLITEVOLCANICS

PHYLLITE

SIW

ALI

K G

RO

UP

: MU

DS

TO

NE

,S

AN

DS

TO

NE

, CO

NG

LOM

ER

AT

E;

MIO

-PLI

OC

EN

E

LITHICSANDSTONE,RECYCLEDMETASED.CLASTS

ROCKTYPES

KEY GRAINTYPES

ROCKTYPES

KEY GRAINTYPES

ROCKTYPES

KEY GRAINTYPES

ROCKTYPES

KEY GRAINTYPES

kmkm

BURHIGANDAKIGNEISS

KALIKHOTFM. SCHIST

SALLYANIGAD GNEISS

MELMURAFM. PHYLLITE

DAMGAD FM.QUARTZITE

DA

DE

LDH

UR

A G

RO

UP

(P

RE

CA

MB

RIA

N)

DADELDHURAGRANITE:TOURMALINEGRANITE;492 Ma

QUARTZCARBONATEMINORPHYLLITE

PH

ULC

HA

UK

I

GR

OU

P(P

ALE

OZ

OIC

)

ULLERI FM.AUGENGNEISS

KUSHMA FM.QUARTZITE &PHYLLITE

RANIMATA FM.PHYLLITE &QUARTZITE;BASICINTRUSIVES

GALYANG FM.PHYLLITE,QUARTZITE

PHYLLITEQUARTZ

SANGRAM FM.PHYLLITE,QUARTZITE

SYANGJA FM.QUARTZITE

LAKARPATAFM.CARBONATESPHYLLITES

MELPANI FM.QUARTZITE

SURKHET GR.SS, LS, SHALE

QUARTZCARBONATEPHYLLITECHLORITE

QUARTZFELDSPARMICA

2

4

6

8

10

MID

LAN

D G

RO

UP

(UP

PE

R P

RE

CA

MB

-UP

PE

R P

ALE

OZ

)T

ER

T.

CRET.

TOP ERODED

TOP ERODEDOR CUT BYSTDS

TOP ERODED ORCUT BY THRUST

TOP ERODED ORCUT BY THRUST

Figure 3. The lithological composition of each of the major thrust sheet systems in thewestern Nepal Himalaya, with key mineral and grain type assemblages and ages. Com-piled from Stöcklin (1980), Valdiya (1980), Shrestha et al. (1987), Schelling (1992), andSrivastava and Mitra (1994). Because Greater Himalayan rocks in western Nepal are

poorly documented, the section compiled by Schelling (1992) in eastern Nepal is de-picted for the Main Central thrust system; lithostratigraphic names used in this sectionare not strictly applicable to Himal Group (Greater Himalayan zone) rocks in westernNepal. Abbreviations as in Figure 2.

The Main Frontal thrust is generally mapped atthe frontal topographic break of the Himalaya,where folded and thrusted Siwalik Group rocksoverlie Quaternary gravels (e.g., Nakata, 1989;Schelling, 1992; Mugnier et al., 1993). Reflectionseismic data and low-relief hills in the Quaternaryalluvium south of the topographic front of therange indicate that the locus of active shorteningalong the front of the thrust belt is in the subsur-face beneath the northern Indo-Gangetic plain(Mugnier et al., 1993). Between the Main Bound-ary thrust and the Main Frontal thrust are two tofour thrust faults that carry north-dipping, foldedpanels of the Siwalik Group, suggesting that a re-gional decollement exists in the lower Siwalikrocks at a depth of ~5–6 km (Fig. 2; Dhital andKizaki, 1987; Schelling et al., 1991; Schelling,1992; Srivastava and Mitra, 1994; Mugnier et al.,1993; Acharyya, 1994).

Estimates of Shortening and Timing ofThrust Faults

Quantitative estimates of thrust displacementsand crustal shortening in western Nepal have notbeen made. Any cross section through theHimalaya is likely to underestimate shorteningbecause of the general absence of hanging-wallcutoffs and the pervasive penetrative deformationand large volume losses incurred by the meta-morphic rocks north of the Main Boundarythrust. Nevertheless, minimum estimates are use-ful in developing palinspastic restorations of theorogenic wedge and the Indo-Gangetic forelandbasin system, and in assessing provenance data.The cross section in Figure 2 exhibits ~228 km ofhorizontal shortening in Lesser Himalayan andSubhimalayan rocks, exclusive of Dadeldhurathrust and Main Central thrust displacements.The resulting 66% shortening is essentially iden-tical to the 65%–70% shortening determined bySrivastava and Mitra (1994) in the Subhimalayaand Lesser Himalaya of northern India, ~100 kmwest of our study area. Displacements on thrustswithin the duplex were fed into the Ramgarhthrust, which is interpreted as the roof thrust ofthe duplex. Because the Dadeldhura thrust sheetis above the Ramgarh thrust, it also would haveundergone significant southward displacementand folding into a large antiform-synform pairduring growth of the duplex. Simultaneously, asection of rock ~15 km thick was eroded from thehanging wall of the Dadeldhura thrust betweenits present northern trace and the trace of theMain Central thrust (Fig. 2). We have not at-tempted to restore the Main Central and Dadel-dhura thrust systems, but Schelling (1992) esti-mated ~175–210 km of shortening on the MainCentral thrust in eastern Nepal, and Srivastavaand Mitra (1994) estimated ~193–260 km on the

combined Almora (equivalent to the Dadeldhura)and Main Central thrust systems in northern In-dia. Combining our estimates for the Sub-himalaya and Lesser Himalaya with the estimatesby Srivastava and Mitra (1994) for the Almorathrust and Main Central thrust in Kumaon, andthe estimates by Ratschbacher et al. (1994) forthe Tibetan Himalaya, yields a total of ~556–623km of shortening for the Himalayan fold-thrustbelt in western Nepal. This value is comparableto the total estimated by Srivastava and Mitra(1994), but is considerably greater than Schelling’sestimate in eastern Nepal. The difference be-tween the estimates in India and western Nepalcompared with eastern Nepal arises mainly fromthe large amounts of shortening required to buildthe Lesser Himalayan duplex in the western crosssections and the absence of comparable internalshortening in the duplex in Schelling’s (1992)cross section.

The ages of major displacement events onHimalayan thrusts are well known only for theMain Central thrust system in north-centralNepal. Thermochronologic studies of the MainCentral thrust document a major cooling event at~22–20 Ma (Hubbard and Harrison, 1989;Copeland et al., 1991; Harrison et al., 1992;Macfarlane et al., 1992; Macfarlane, 1993;Coleman, 1996). Hodges et al. (1996) demon-strated that faults in the Main Central thrust sys-tem experienced several displacement events be-ginning ~22.5 Ma in the Annapurna Range ofnorth-central Nepal, and presented evidence forsignificant out-of-sequence displacements onthrusts within the Greater Himalayan zone.Macfarlane (1993) presented evidence for lateMiocene–Pliocene reactivation of the Main Cen-tral thrust. Recent Th-Pb dating by Harrison et al.(1997) of synkinematic monazite inclusionswithin garnet crystals proximal to the Main Cen-tral thrust suggests that a major displacementevent occurred on the Main Central thrust systemas recently as ~5.5 Ma, although some of the up-lift necessary to convey these rocks to the surfacecould have been facilitated by passive transportabove the Lesser Himalayan duplex. Srivastavaand Mitra (1994) also reported crosscutting rela-tionships that indicate that the Main Centralthrust may have had a major episode of out-of-sequence (break-back) displacement. The SouthTibetan detachment system was active between~21 and 16 Ma, approximately synchronous withMain Central thrust emplacement (Burchfiel etal., 1992; Hodges et al., 1996; Coleman, 1996).

In western Nepal, along the north limb of theDadeldhura syncline, the southward-dippingDadeldhura thrust cuts the Dumri Formation, ofprobable early Miocene age. This means that theDadeldhura thrust must have had a major phaseof displacement (at least several tens of kilome-

ters, equal to the bed length of the preserved syn-formal part of the thrust) after deposition of theDumri Formation. The Main Boundary thrust inNepal must postdate the emplacement of thecrystalline thrust sheets because the latter werefolded by growth of the Lesser Himalayan du-plex, which fed slip southward into the MainBoundary thrust. The Main Boundary thrust cutsrocks as young as the Pliocene upper SiwalikGroup and carries rocks as old as early middleMiocene; thus its age is between ~15 and ~2 Ma.Meigs et al. (1995) reported provenance and agedata from the Siwalik Group in northern Indiathat suggest that the Main Boundary thrust wasactive by ~9 Ma and possibly as early as ~11 Ma.We present data in this paper that support an ~11Ma age of initiation of the Lesser Himalayan du-plex. However, the Main Boundary thrust is onlythe most recently active of the thrusts that rootbeneath the duplex, and our data indicate that itwas not active until Pliocene time. The MainFrontal thrust and associated hanging-wall imbri-cates that cut the Siwalik Group must, at least inpart, postdate the Pliocene upper member, whichis the youngest unit in the hanging walls of thesethrusts. The ~45° of progressive rotation of bed-ding in the upper Siwalik Group in the SuraiKhola section suggests the presence of a progres-sive unconformity related to displacement on oneof the intra-Siwalik thrusts during Pliocene time.That Quaternary deposits are deformed by theMain Frontal thrust system shows that the thrustsystem is still active (Nakata, 1989).

TERTIARY FORELAND BASIN DEPOSITS

Sedimentology of the Dumri Formation

The terms Dumri and Suntar Formations areused synonymously in western Nepal for a thickunit of fluvial sandstone and red mudstone thatconstitutes the youngest stratigraphic unit in thehanging wall of the Main Boundary thrust sys-tem. Although a detailed analysis of the DumriFormation is beyond the scope of this paper, webriefly discuss the sedimentology of the unit inorder to provide some context for Dumri petro-graphic data and hypotheses for Neogene ero-sional unroofing of the Himalayan thrust belt.Our reconnaissance of the Dumri Formation wasrestricted to exposures in five areas: (1) along theDadeldhura road, where the folded Dadeldhurathrust juxtaposes Precambrian(?) schist againstthe Dumri; (2) along Patu Khola, northwest ofTulsipur; (3) in Swat Khola, northwest of Surkhet(Birendranagar); (4) in road cuts north ofSurkhet; and (5) at the type section of the Dumrialong the Tansen-Butwal road on either side ofthe Dumri bridge (Fig. 1).

The Dumri Formation in the Surkhet area is at

KINEMATIC HISTORY OF THE HIMALAYAN FOLD-THRUST BELT

Geological Society of America Bulletin, January 1998 7

least 1300 m thick, but only ~750 m of the unit areexposed in the Tansen area (Sakai, 1989). TheDumri consists of gray and greenish sandstoneand interbedded red mudrocks. The proportion ofsandstone relative to mudstone decreases up-section. The sandstone bodies are typically 10–40m thick and are composed of medium- to fine-grained, trough cross-stratified, and horizontallylaminated sandstone. Bedding is lenticular andbasal surfaces of sandstone bodies are erosional.Fossil logs and rip-up clast conglomerates arepresent in the lower parts of some sandstone units,and upward-fining sequences are common. Weinterpret these sandstone bodies as the deposits oflarge fluvial channels. Less-abundant, finer-grained, rippled, tabular sandstone beds are inter-preted as crevasse-splay deposits. The associatedred mudrocks are generally structureless and mot-tled, and contain rooted zones, plant fragments,Mn/Fe-rich mottles, and occasional carbonatenodules; these deposits probably represent paleo-sols. The Dumri Formation is therefore inter-preted as fluvial channel, crevasse splay, and over-bank deposits (Sakai, 1989). Sakai’s (1989)sparse paleocurrent data from unspecified typesof cross-stratification in the Dumri indicate gener-ally southward paleoflow. Our paleocurrent mea-surements (~500 measurements) from troughcross strata in the Tansen and Surkhet areas showa strong west-southwestward vector mean.

The Dumri Formation has yet to be reliablydated. In western Nepal and northern India, theunit overlies well-dated middle to late Eocenenummulitic limestone and black shale of theBhainskati Formation (also referred to as theSwat or Subathu Formation). At the DumriBridge locality, the contact between the Dumriand the underlying Bhainskati is marked by a3–4-m-thick, extremely mature paleosol—probably an oxisol—that has conspicuous Fe-rich mottles. We interpret this as a deeplyweathered soil profile at a major unconformitybetween the Dumri and Bhainskati Formations.Regional lithostratigraphic correlations suggestan early Miocene age for the Dumri Formation(Sakai, 1983, 1989; Kayastha, 1992), althoughGautam (1989) inferred a late Eocene age on thebasis of paleomagnetic inclination data. Innorthern India, the Eocene Subathu Formationis overlain by the poorly dated Dagshai Forma-tion and the early–middle Miocene Kasauli For-mation (Sahni, 1953; Najman et al., 1994). InPakistan, similar fluvial red beds in the MureeFormation are considered to be of earlyMiocene age (Burbank et al., 1996). Regardlessof the exact regional correlation, it appears thata major unconformity, spanning perhaps mostof the Oligocene Epoch, exists between well-dated Eocene rocks and poorly dated lowerMiocene rocks throughout the northern part of

the Himalayan foreland basin (Fig. 4). Theyounger age limit of the Dumri Formation isalso poorly known, mainly because the top ofthe formation is everywhere either faulted oreroded. The oldest Siwalik Group rocks are ofmiddle Miocene age (~14 Ma; Quade et al.,1995) and are lithologically similar to upperDumri Formation lithofacies. These combinedobservations, therefore, suggest that the Dumriis of early to early-middle Miocene age.

Sedimentology of the Siwalik Group

Stratigraphy and Age. Data reported in thispaper are derived mainly from a new ~3.4-km-

thick measured section at Khutia Khola in farwestern Nepal, and from sections at Macheli,Babai, and Surai Kholas (Figs. 1 and 4). TheSiwalik Group of far western Nepal can be di-vided into informal lower, middle, and uppermembers as defined by Quade et al. (1995). Al-though this scheme is useful for broad regionalcomparisons, it is merely lithostratigraphic andcannot be used for detailed chronostratigraphy. Inthe Khutia Khola section, the measured thick-nesses of the lower and middle Siwalik membersare, respectively, 862 m and 2468 m. The base ofthe Khutia section is not exposed, and the top ofthe section is truncated by a thrust fault thatplaces the lower member on top of the upper

DECELLES ET AL.

8 Geological Society of America Bulletin, January 1998

km

14

13

12

7

10

Ma

8

9

1110.8 Ma

0

1

2

3

4

5

6

FLUVIAL CONGLOMERATE

FLUVIAL SANDSTONE

FLUVIAL MUDSTONE

DU

MR

I FM

.S

IWA

LIK

GR

OU

P (

LOW

ER

MIO

CE

NE

-QU

AT

ER

NA

RY

)

BHAINSKATI(SWAT) FM.

LOW

ER

MID

DLE

UP

PE

R

(LO

WE

RM

IOC

EN

E ?

)

(EOCENE) NUMMULITIC LIMESTONE

MARINE BLACK SHALE

CANDE & KENT (1995)

Figure 4. Simplified stratigraphic section of the Tertiary rocks of western Nepal, based onmeasured sections at Dumri Bridge and Khutia and Swat Kholas. The contact between theDumri Formation and Siwalik Group is unknown because it is either eroded or removed byfaulting at all exposures we have studied. The magnetostratigraphy of the Siwalik Group isbased on unpublished data of T. P. Ojha and D. Richards. The magnetic polarity time scale ofCande and Kent (1995) is shown for reference, and the position of lower-middle Siwalik Groupboundary is highlighted at ~10.8 Ma.

member. The upper member in the area is at least1000 m thick.

The age range of the Siwalik Group in Nepalis known to be between ~14 Ma and ~1 Ma onthe basis of vertebrate fossils (West et al., 1978,1991; Corvinus, 1994) and magnetostratigraphicstudies (Tokuoka et al., 1986; Harrison et al.,1993; Appel and Rössler, 1994; Quade et al.,1995). Studies to develop the precise ages ofboundaries between lower, middle, and upperSiwalik members are in progress, but work thusfar indicates that the age of the lower-middleboundary in the Khutia Khola section is ~10.8Ma (Fig. 4), whereas the same boundary in theSurai Khola section is ~8 Ma (T. P. Ojha andD. Richards, unpub. data). At Bakiya Khola ineastern Nepal, the lower-middle boundary is ~9Ma and the middle-upper boundary is <4.5 Ma(Quade et al., 1995).

Lithofacies Descriptions and Interpreta-tions. The Siwalik Group is characterized by al-ternating siltstone and gray sandstone units.Siwalik lithofacies are divisible into five assem-blages (Table 1) that characterize fluvial depositsand have been documented in the Siwalik Groupof Pakistan, India, and central Nepal (e.g., Willis,1993a, 1993b; Hisatomi and Tanaka, 1994;Tanaka, 1994; Burbank et al., 1996). Although adetailed treatment of the sedimentology of theSiwalik Group is beyond the scope of this paper,we provide general descriptions and interpreta-tions in order to assess major changes in the na-ture of Siwalik fluvial systems through time.

The five major lithofacies assemblages docu-mented in the Khutia Khola section consist ofsandy fluvial channel, gravelly fluvial channel,crevasse splay, paludal flood plain, and paleosoldeposits (Table 1). Channel sandstones (Fig. 5A)and imbricated conglomerate bodies (Fig. 5B) al-

ternate with crevasse splay (Fig. 5C) and paleosoldeposits (Fig. 5D), and the paludal facies typi-cally are at the bases of channel bodies. On thebasis of sandstone body thicknesses, channeldepths ranged from a few meters to ~12 m duringdeposition of the lower Siwalik units and up toseveral tens of meters during deposition of themiddle Siwalik units. The limited lateral expo-sure in our sections prevents estimation of bank-full discharges, but these channel deposits areprobably comparable with those documented byWillis (1993a, 1993b) in the Siwalik rocks ofnorthern Pakistan, which indicate channel-beltdischarges in the range of 103–104 m3/s.

The paleosols are mostly red and orange in thelower Siwalik Group and gray and yellow in themiddle and upper members. Lower Siwalikpaleosols are rich in pedogenic carbonate nod-ules, similar to lower Siwalik paleosols in centralNepal (Quade et al., 1995). The dark gray paleo-sols of the upper and middle Siwalik units repre-sent pedogenesis in poorly drained conditions,analogous to Histosols in the modern soil nomen-clature. Paleosols formed in better drained condi-tions are typified by the bright yellow paleosolsof the upper middle and upper Siwalik units, andby darker yellow to red paleosols in the lowerSiwalik units.

The paludal deposits, although relatively mi-nor in volume, are distinctive because they arewell laminated and contain abundant fossilleaves, occasional bone fragments, coal, and lig-nite seams. That they occur almost exclusivelydirectly beneath fluvial channel deposits suggeststhat the paludal facies may have formed in poorlydrained, topographically low areas that ulti-mately were the targets of channel avulsions.

The gravelly channel deposits of lithofacies as-semblage 5 (Table 1) are characterized by sedi-

mentary structures and textures typical of grav-elly braided river deposits (e.g., Smith, 1974;Rust, 1978; Miall, 1996; Fig. 5B). These featuresare also common in the modern gravel-bed riversand streams that drain the Himalaya in westernNepal. Some of these modern rivers (e.g., theKarnali) exhibit anastomosing, rather than classi-cal braided, channel-belt morphologies. Duringthe dry season, when flows are many meters be-low bankfull stage, these channels have internallybraided networks of anabranches. It is plausiblethat some Siwalik rivers had similar anastomos-ing morphologies.

Paleocurrent Directions. Limbs of troughcross-strata were measured in 50 of the channeldeposits throughout the Khutia Khola section andin 6 channel deposits in the Surai Khola section.Each measurement station consists of a singlechannel sandstone from which 15–25 individualtrough limbs were measured, yielding an averagetrough-axis orientation for each station (accordingto method 1 of DeCelles et al., 1983). Whereverpossible, trough axes were measured directly onthe outcrop; however, such exposures are ex-tremely rare in the sections we have studied. In theconglomeratic upper member, we measured 10imbricated clasts per station in the Surai, Khutia,Macheli, and Babai Khola sections.

Average trough axes for the Khutia Khola sec-tion are consistently within the azimuth range of120°–200° (Fig. 6, A and B); the overall averageis 160°. Only one channel deposit produced anazimuth outside of this range. The average paleo-current azimuths of the lower and middle mem-bers are indistinguishable. At Surai Khola, themean paleocurrent direction is 208° (Fig. 6C).The data from the upper member exhibit morescatter; the mean azimuth is ~200° (Fig. 6D).These data demonstrate that the rivers that de-

KINEMATIC HISTORY OF THE HIMALAYAN FOLD-THRUST BELT

Geological Society of America Bulletin, January 1998 9

TABLE 1. SUMMARY OF SIWALIK GROUP LITHOFACIES

Lithofacies Description Stratigraphic Interpretationassemblage occurrence

1 2–42 m thick, lenticular bodies of medium- to coarse-grained sandstone; Present mainly in lower and middle Fluvial channel deposits; size oferosional basal surfaces; upward fining; trough and planar cross- Siwalik members; thicknesses of channels increases fromstratification; parallel lamination; ripples; epsilon cross-stratification; units and degree of amalgamation average 8–10 m deep in lowerdewatering structures. increase upsection. Siwalik Group to more than 20

m in middle Siwalik Group.2 Several-meter-thick packages of fine- to very fine-grained sandstone Present throughout lower and middle Crevasse splay deposits.

beds, 10–50 cm thick, interbedded with thin beds of siltstone; beds are Siwalik Group, but most abundanttabular, rippled, horizontally laminated, and bioturbated (Scoyenia in lower member.ichnofacies).

3 Variegated, mottled, red, yellow, gray, and orange siltstone; no primary Red and orange colors are prevalent Paleosols; reddish units are morebedding; root traces, carbonate nodules, Fe nodules, burrows; basal in lower Siwalik Group; gray and oxidized and less organic richcontacts gradational with underlying lithofacies; thickness ranges from yellow more abundant in middle than gray units.20 to 200 cm. and upper Siwalik Group; overall,

this lithofacies most abundant in lower member.

4 5–150-cm-thick beds of laminated, organic-rich siltstone; fossil leaves Most abundant in lower and middle Paludal deposits in overbank and bones, coal and lignite are common; characteristically occur on Siwalik Group. areas.top of units of lithofacies assemblage 3 and below units of assemblage 1.

5 Pebble- to cobble-conglomerate; clast-supported; beds 50–150 cm Characteristic of upper Siwalik Group. Large gravelly braided-riverthick; imbricated; trough and planar cross-stratification and crude deposits.horizontal stratification; well-organized, often upward-fining.

posited the Siwalik Group in western Nepalflowed consistently southward, essentially identi-cal to the modern fluvial drainage pattern in thenorthern Indo-Gangetic flood plain. We havefound no evidence for axial (i.e., parallel to themountain front) or northward drainage in any partof the Siwalik Group of western Nepal. Hisatomiand Tanaka (1994) found no evidence for axialdrainage in the Siwalik Group of central Nepal.

Stratigraphic Trends in Lithofacies and Flu-vial Characteristics. As in other areas of theHimalayan foreland basin system (Tokuoka et al.,1986; Harrison et al., 1993; Quade et al., 1995;Burbank et al., 1996), the Siwalik Group of west-ern Nepal exhibits an overall upward-coarseningtrend (Fig. 4). The lower member is characterizedby alternating red and yellow paleosols and rela-tively thin (~10 m) and laterally restricted channeldeposits, and associated crevasse splay deposits.The abundance of oxidized, calcareous paleosols

in the lower member suggests that the middleMiocene flood plain was relatively stable and welldrained, but subject to occasional flooding. Theprevalence of crevasse splay deposits implies thatlower Siwalik channels had well-developed lev-ees, which in turn suggests that they were anasto-mosing and/or meandering. In contrast, the mid-dle member is dominated by dark coloredHistosols and laterally extensive, thick (com-monly >20 m) channel deposits. At bankfullstage, middle Siwalik Group rivers must have haddischarges comparable to those of large modernrivers draining the thrust belt, about five timesgreater than those of lower member rivers.Crevasse splays are not well developed, implyingthat the middle Siwalik channels lacked promi-nent levees and were more laterally mobile thanlower Siwalik channels. The upward transitionfrom middle to upper Siwalik members is gradualand probably takes place at different stratigraphic

levels in different areas along the Himalayanthrust belt (Tokuoka et al., 1986; Sah et al., 1994).

The evidence discussed herein suggests thatthe middle and upper Siwalik rivers were similarin morphology to the modern transverse tribu-taries of the Ganges, wherein gravel deposition isrestricted to within ~20 km of the topographicfront of the mountain belt. The modern Indo-Gangetic flood plain north of the Ganges River isdominated by two types of rivers that flow gener-ally south-southeastward, transverse to the trendof the mountain front: (1) large rivers (dischargesof ~104 m3/s) that have catchments extendinginto the interior of the mountain belt and(2) smaller rivers (discharges of ~102–103 m3/s)that have catchments limited to the frontal part ofthe mountain belt. The former produce fluvialmegafans with areas of 104–105 km2, but topo-graphic relief of only ~20 m (Wells and Dorr,1987; Gohain and Parkash, 1990; Mohindra et

DECELLES ET AL.

10 Geological Society of America Bulletin, January 1998

Figure 5. The Siwalik Group lithofacies. (A) Stacked channel sandstone bodies (lithofacies assemblage 1) in the middle Siwalik Group. Noteperson for scale. (B) Imbricated conglomerate of lithofacies assemblage 5 in the upper Siwalik Group. (C) Tabular beds of lithofacies assemblage2. Beds are 20–50 cm thick. (D) Paleosol deposits in lower Siwalik member, showing carbonate nodules and large tubular burrows. Scale in B andD is 20 cm long.

al., 1992; Sinha and Friend, 1994; Gupta, 1997).The rivers on these fluvial megafans are highlymobile and have variable morphologies, includ-ing anastomosing, braided, and meandering.Highly sinuous meandering rivers flow parallelto braided or anastomosing rivers over large dis-tances. Shallow ponds and oxbows are locallyabundant. The flood plain receives large amountsof precipitation (1000–1500 mm/yr) and majorfloods during the monsoon. During the dry sea-son, however, the water level in channels of ma-jor transverse rivers drops more than 10 m. Themiddle and upper Siwalik Group fluvial depositsmost likely accumulated under a similar mon-soonal climatic regime (Quade et al., 1995).

Independent evidence from fossil leaves andcarbon and oxygen isotopic studies of paleosolcarbonate and organic material support the con-tention that the rivers that deposited the middleand upper Siwalik Group were influenced bymonsoonal discharge. In most sections, fossilleaves indicative of tropical rainforest vegetationare abundant in the lower member, but disappearwithin a few hundred meters above the base ofthe middle member (Quade et al., 1995). Paleo-botanical studies indicate the presence of wide-spread, broad-leaved tropical evergreen forestsduring lower Siwalik deposition, which were re-placed by dry subtropical plants and grasses dur-

ing upper Siwalik deposition (e.g.,Awasthi et al.,1994). Carbon isotopic data from Siwalik Grouppaleosol carbonate nodules also indicate a bio-mass transition from C3 to C4 plants at ~7–8 Ma,which Quade et al. (1995) interpreted as a shiftfrom semitropical evergreen forests to mon-soonal grasslands. However, the carbon isotopictransition postdates the lithostratigraphic transi-tion from lower to middle Siwalik Group by 1–3m.y.; thus it does not appear to coincide with anymajor sedimentological changes in this part ofthe foreland basin.

Because the paleocurrent data indicate thatflow directions were generally southwardthroughout Siwalik Group deposition, the up-section changes in sedimentology from the lowerto the middle members cannot be explained sim-ply as a result of southward progradation of thealluvial system in front of the southward-propagating Himalayan thrust front. Had thisbeen the case, the lower member rivers shouldhave been larger than the upper member rivers,because the former would have been fartherdownstream in the overall depositional mosaic ofthe foreland. The paleocurrent data also excludethe possibility that the transition into the middlemember took place in response to a change fromtransverse to axial-trunk drainage. Neither can anincrease in subsidence rate explain the lithofacies

transition, because although subsidence rate mayinfluence channel density, it cannot influence flu-vial discharge. The most plausible explanation ofthe lithofacies transition from lower to middle Si-walik Group is an increase in channel bankfulldischarge, at least on a seasonal basis (Hisatomiand Tanaka, 1994). Because the paleobotanicaldata indicate that overall effective moisture de-creased from the lower to middle Siwalik Group,channel sizes must have increased in response toconcentration of discharge in a rainy season, sim-ilar to that of today in the Indo-Gangetic forelandbasin system. The sedimentological evidence canbe explained by gradual intensification of sea-sonality beginning ~11 Ma, eventually sufficientto trigger C4 grass expansion on the Indo-Gangetic flood plain by ~8 Ma.

To summarize, the Siwalik Group in westernNepal records a change from small channelsand well-drained flood plains to much largerchannels and more poorly drained flood plainsfrom middle to late Miocene time. Rivers had avariety of morphologies, but highly sinuousrivers seem to have dominated during earlyMiocene time, whereas anastomosing andbraided rivers prevailed during middle and lateMiocene time. The changes in fluvial style arebest explained as a response to increased rainyseason discharge.

KINEMATIC HISTORY OF THE HIMALAYAN FOLD-THRUST BELT

Geological Society of America Bulletin, January 1998 11

A. KHUTIA KHOLAMIDDLE SIWALIKSTROUGH LIMBSn=62133 STATIONS

LOWER SIWALIKSTROUGH LIMBSn=37620 STATIONS

UPPER SIWALIKSIMBRICATIONSn=808 STATIONS

B. KHUTIA KHOLA

C. SURAI KHOLALOWER & MIDDLESIWALIKSTROUGH LIMBSn=1246 STATIONS

D. SURAI, KHUTIA, BABAI,& MACHELI KHOLAS

Figure 6. Rose diagrams with vector means (arrows) and 95% confidence intervals (arc lines) summarizing paleocurrent data from the Khutia,Surai, Babai, and Macheli Khola sections. See Figure 1 for locations of sections. Exact stratigraphic locations of measurement stations are avail-able from DeCelles.

Provenance

Methods.We cut samples of medium-grained,fluvial channel sandstones from the Dumri Bridgeand Khutia Khola sections into standard petro-graphic thin sections, stained them for Ca-plagio-clase and K-feldspar, and point-counted (mini-mum of 450 grains per slide) according to theGazzi-Dickinson method (e.g., Ingersoll et al.,1984). Traditional parameters were also docu-mented in order to distinguish certain coarse-grained rock fragments (Table 2). Samples ofsand from modern rivers draining specific litho-tectonic zones of the thrust belt were point-counted in order to provide a basis for interpreta-tion of the Tertiary sandstones in terms of knownsource terranes. In addition, nine samples of Si-walik Group sandstone were crushed, sieved, andpanned for dense minerals in the field, and threesamples of modern river sand were panned fordense minerals. Zircons were separated from thepanned concentrates using heavy liquids, a mag-netic separator, and disposable sieve screens.Uranium-lead ages of zircons >145 φin sieve sizewere determined by isotope dilution–thermal ion-ization mass spectrometry. The larger grains wereseparated into populations on the basis of theircolor, shape, and clarity, and representatives ofeach population were then selected for analysis.All of the zircons analyzed were abraded to abouttwo-thirds of their original diameter prior to dis-solution, and were analyzed as individual grainsfollowing the methods of Gehrels et al. (1991).

Petrographic Results. Dumri Formation sand-stones have average modes1 of QmFLt = 68, 2,30; QtFL = 78, 2, 20; and QmPK = 98, 2, 0 (Figs.7 and 8). Lower Siwalik Group sandstones havemodes of QmFLt = 54, 10, 36; QtFL = 72, 10, 18;and QmPK = 84, 13, 3. Middle and upper SiwalikGroup sandstones have modes of QmFLt = 53,17, 30; QtFL = 68, 17, 15; and QmPK = 76, 11,13. In ternary QFL space, framework composi-tions shift from the Q-L binary in Dumri andlower Siwalik sandstones toward more feld-spathic compositions in middle and upper Siwaliksandstones. This trend away from relativelyquartz-rich (during Dumri time), to plagioclase-rich (during lower Siwalik time), to K-feldspar-rich (during middle and upper Siwalik time) com-positions is most clearly seen in the mono-mineralic populations (QmPK) (Figs. 7B and8A). Samples of the Dumri Formation containedvirtually no K-feldspar, but significant amounts ofplagioclase (Fig. 8A). Similar results were ob-tained in studies of the Siwalik Group by

Hisatomi (1990) and Critelli and Ingersoll (1994).Lithic fragments are represented in all samples

by phyllite, quartz-mica schist (Fig. 9A),quartzite, sericitic quartzite (Fig. 9B), limestone,dolostone, marble (Fig. 9C), and minor amountsof volcanic grains (Fig. 9D). Foliated polycrys-talline quartz grains are common in some sam-ples. Stratigraphic trends in lithic grain contentare also evident: phyllite grains decrease and car-bonate grains generally increase upsection (Fig.8, B and C). Both muscovite and biotite are abun-dant in most middle and upper Siwalik Groupsandstone samples. Accessory grains includegarnet, kyanite, zircon, epidote, staurolite, silli-manite, pyroxene, hornblende, chlorite, andcordierite. The higher grade minerals (kyaniteand sillimanite) first appear in the upper part ofthe lower Siwalik Group (Fig. 8E). Cements aremainly quartz (overgrowths and pore-fillinggranular aggregates) in the Dumri Formation andcalcite, kaolinite, quartz, and local anhydrite inthe Siwalik Group.

Conglomerates in the upper Siwalik memberare dominated by quartzite and contain subordi-nate amounts of micritic limestone, marble, and

recycled sandstone clasts from the middle Siwa-lik member (Figs. 5B and 10). Notably lacking inthe upper Siwalik conglomerates in westernNepal are clasts of medium- to high-grade meta-morphic and igneous rocks.

Modal petrographic data from grain mounts ofmodern river sands provide comparative infor-mation about sands derived from known lithotec-tonic zones (Fig. 11; see Appendix for sample lo-cations). Samples from the Sun Kosi River,which drains the Greater Himalayan zone in thehanging wall of the Main Central thrust, are en-riched in quartz, plagioclase, micas, and lithicfragments of garnet-mica schist and sillimaniteschist; accessory minerals include cordierite,sillimanite, garnet, staurolite, and kyanite (Fig.11). The Sun Kosi samples are remarkably defi-cient in K-feldspar (Fig. 11, B and C). In contrast,sands from Rao and Lamikheti Kholas (see Ap-pendix), which are derived entirely from rocks ofthe Dadeldhura thrust sheet, contain abundant K-feldspar, schist, phyllite, and coarse-crystalline,quartzo-feldpathic (granitic) lithic fragments. Ac-cessory minerals are dominated by amphibolesand garnets. Sands from the Rapti River in cen-

DECELLES ET AL.

12 Geological Society of America Bulletin, January 1998

TABLE 2. PETROGRAPHIC PARAMETERS

Qm Monocrystalline quartzQp Polycrystalline quartzQpt Foliated polycrystalline quartzQms Monocrystalline quartz in sandstone/quartzite lithic grainC ChertS SiltstoneQt Total quartzose grains (= Qm + Qp + Qpt + Qms + C + S)

K Potassium feldspar (including perthite, myrmekite, microcline)P Plagioclase feldspar (including Na and Ca varieties)F Total feldspar grains (= K + P)

Lvm Mafic volcanic grainsLvf Felsic volcanic grainsLvv Vitric volcanic grainsLvx Microlitic volcanic grainsLvl Lathwork volcanic grainsLv Total volcanic lithic grains (= Lvm + Lvf + Lvv + Lvx + Lvl)

Lsh MudstoneLph PhylliteLsm Schist (mica schist)Lsc Schist (calc schist)Lma Marble (foliated, coarse-grained)Lls LimestoneLd DolostoneLm Total metasedimentary lithic grains (= Lsh + Lph + Lsm + Lsc + Lma + Lls + Ld)

Dense minerals, typically monocrystalline:BiotiteChloriteCordieriteEpidoteGarnetHornblendeKyaniteMuscovitePhosphatePyroxeneSillimaniteStauroliteTourmalineZircon

1GSA Data Repository item 9805, tables for re-calculated modal point-count data and for U-Pb iso-tope data, is available on request from Documents Sec-retary, GSA, P.O. Box 9140, Boulder, CO 80301.E-mail: editing@geosociety.org.

tral Nepal, which cuts through the Mahabharatthrust sheet (another of the crystalline thrustsheets containing large Cambrian–Ordoviciangranite bodies, similar to the Dadeldhura sheet),are similar to those from the Dadeldhura thrustsheet and include abundant K-feldspar, schist,phyllite, and medium-grade metamorphic miner-als (especially garnet). Samples collected at thetopographic front of the range from large riversthat drain the entire thrust belt (Mahakali andKarnali Rivers) contain a mixture of grain typesthat can be assigned to each of the major lithotec-tonic zones (Fig. 11): plagioclase, schist, abun-dant mica, and high-grade metamorphic mineralsfrom the Greater Himalayan zone; K-feldspar,phyllite, chlorite schist, and foliated polycrys-talline quartz from the crystalline thrust sheets(the Almora and Dadeldhura sheets); and carbon-ate, quartzite, and phyllite grains from the LesserHimalayan zone.

Isotopic Results.Of the 130 detrital zircongrains from Siwalik Group sandstones and mod-ern rivers that were analyzed (Table DR2, seefootnote 1), 113 grains yielded concordant toslightly discordant ages. The 207Pb*/206Pb* ages

are used as interpreted crystallization ages forthese grains, whereas the more highly discordantgrains are excluded from further discussion. Zir-cons from the Narayani, Karnali, and MahakaliRivers exhibit age clusters in the 750–1000 Ma,1700–1900 Ma, and 2200–2600 Ma ranges(Figs. 12 and 13). The Karnali River yielded onegrain in the 480–500 Ma range, and a few ages of1900–2400 Ma were obtained from zircons fromall three rivers. All of the Siwalik sandstone sam-ples contained grains of ~800–1000 Ma and~2500 Ma (Figs. 12 and 13). The lower and mid-dle Siwalik samples contained grains of 460–530Ma, but only one grain in this age range wasfound in the upper member sample. Conversely,the middle Siwalik sample lacked grains in the1800–2000 Ma age range, whereas lower and up-per Siwalik samples contained several grains ofthis age range.

The detrital zircon ages are consistent withwhat little is known about the ages of zircons inmetamorphic and plutonic sources in the GreaterHimalayan and Lesser Himalayan zones. Fewgeochronological data are available from Hima-layan terranes for comparison with the zircon

ages reported herein. Eight individual zircongrains from a sample of the Dadeldhura granitecollected along the Dadeldhura road (Fig. 1) wereanalyzed in this study. Four grains yield a concor-dant age of 492 ± 6 Ma, and the other four grainshave significant inherited components of ~1687Ma (n= 2, MSWD [mean square of weighted de-viates] = 0.4) and ~2465 Ma (n= 2, MSWD = 13)(Fig. 14). Parrish and Hodges (1996) reportedsingle-grain U-Pb ages from Lesser Himalayanrocks and the Main Central thrust zone in centralNepal. Lesser Himalayan ages range between~1867 and ~2657 Ma, and clusters are in theranges ~1867–1884 Ma and ~1921–1962 Ma; afew early Proterozoic and Archean ages were alsoreported (Fig. 13). Ages associated with the MainCentral thrust zone are scattered between ~967and ~1710 Ma; a cluster of three slightly discor-dant analyses is in the ~967–973 Ma range. TheMain Central thrust zone zircons are entirelyyounger than the zircons from Lesser Himalayanrocks. In our study, zircons from the modernrivers yielded age clusters that nearly match theages reported by Parrish and Hodges (1996) andreflect derivation from both the Greater andLesser Himalayan zones: the ~750–1000 Ma agesindicate sources in Greater Himalayan rocks,whereas the dominant ~1700–1900 Ma and~2500 Ma ages and subordinate ~1900–2400 Maages reflect sources in Lesser Himalayan rocks.All of the Siwalik sandstone samples from theKhutia Khola section contained zircons havingages consistent with sources in Greater Hima-layan (~800–1000 Ma) and Lesser Himalayan(~2500 Ma) rocks (Fig. 13). The sands and sand-stones sampled in this study provide a distributionof detrital zircons from a much larger region thanthe area covered by the Parrish and Hodges(1996) study. The close match between the detri-tal ages and the in situ ages suggests that the rangeof ages reported by Parrish and Hodges (1996)may be generally representative of the Himalayanfold-thrust belt in Nepal, with the importantexception of Cambrian–Ordovician zircons thatare present in the sand and sandstone samples.The ~460–530 Ma grains in the Siwalik Groupsandstones suggest sources in the Cambrian–Ordovician Dadeldhura granite in the hangingwall of the Dadeldhura thrust, and possibly thesparsely dated Cambrian–Ordovician (Ferrara etal., 1983; LeFort et al., 1986; Hodges et al., 1996)orthogneisses in the Greater Himalayan zone. Thenumber of older (>1700 Ma) grains increasessteadily from the lower through upper Siwaliksandstones and into the modern rivers. None ofthe zircon grains analyzed in this study has aCenozoic age, but this may be attributable to theinheritance problems that characterize the Ceno-zoic leucogranites of the Greater Himalayan zone(e.g., Hodges et al., 1996).

KINEMATIC HISTORY OF THE HIMALAYAN FOLD-THRUST BELT

Geological Society of America Bulletin, January 1998 13

Qm

F Lt

Qm

F Lt

Qt

F L

Qm

P K

Qt

F L

60% Qm

Qm

KP

LOWER SIWALIKGROUP

MIDDLE & UPPERSIWALIK GR.

DUMRI FM.

A. B.

Figure 7. Ternary diagrams. (A) All point-count data. (B) Means and standard deviations.Each point represents a single modal analysis of 450 grains. Qm—monocrystalline quartz; Qt—total quartzose grains; F—total feldspar; Lt—total lithic grains; L—lithic grains exclusive ofquartzose lithics; P—plagioclase feldspar; K—potassium feldspar.

Interpretation and Kinematic History

Figure 15 illustrates a hypothetical, struc-turally controlled unroofing sequence that couldaccount for the petrographic and zircon prove-nance indicators in the Dumri Formation andSiwalik Group. Thrusting and uplift in theTibetan (Tethyan) Himalayan zone took placeduring Eocene to early Miocene time (Searle,1991; Ratschbacher et al., 1994). The high-grademetamorphic rocks of the Main Central thrustsheet were being shortened and uplifted in theductile regime by ~22–20 Ma (e.g., Hubbard andHarrison, 1989; Harrison et al., 1992; Hodges etal., 1996). Initial displacement along strands ofthe South Tibetan detachment system was un-derway by ~19–18 Ma (Burchfiel et al., 1992;Hodges et al., 1996). Erosion of sedimentary andlow-grade metasedimentary cover rocks in theTibetan and Greater Himalayan zones probablyproduced the quartzose sedimentary and meta-sedimentary lithic detritus of the Dumri Forma-tion (Fig. 15A). The small amounts of plagio-clase may imply that crystalline basement rockshad been exposed in the Main Central thrusthanging wall by Dumri time, perhaps in re-

sponse to tectonic unroofing by the South Ti-betan detachment system. This is supported byNd isotopic data that indicate that since ~17 Ma,sediment deposited on the Bengal fan has beenderived predominantly from the Greater Hima-layan zone (France-Lanord et al., 1993).

The initial phase of displacement on theDadeldhura thrust occurred after deposition ofthe Dumri Formation, because the fault cuts theDumri along its northern trace (Figs. 1 and 2).This phase of thrusting must have involved atleast 60 km of southward displacement (the ap-proximate bed length of the Dadeldhura sheetthat is above the Dumri Formation; Fig. 2). Fur-ther evidence of major early to middle Miocenedisplacement of the crystalline thrust sheets inNepal comes from the Copeland (1996) study of40Ar/39Ar cooling ages of rocks in the synformalKathmandu thrust sheet, where muscovite cool-ing ages span ~22–13 Ma. Thus, lower SiwalikGroup sandstones should record the early un-roofing history of the crystalline thrust sheets aswell as continued erosion of the Main Centralthrust hanging wall, yielding plagioclase andhigh-grade metamorphic minerals such as kya-nite and sillimanite from the Greater Himalayan

zone, plus metasedimentary lithic grains from theupper part of the Dadeldhura thrust sheet. Abun-dant plagioclase and accessory kyanite and silli-manite seem to typify sands derived from GreaterHimalayan zone rocks in Nepal (Fig. 11). The U-Pb zircon dates are consistent with sources inGreater Himalayan high-grade metamorphicrocks and Lesser Himalayan metasedimentaryrocks, probably in the upper parts of the Dadel-dhura thrust sheet. The Cambrian zircon agescould reflect derivation from the granites of theDadeldhura thrust sheet and/or the Greater Him-alayan orthogneisses.

The increase in coarse-grained, unaltered K-feldspar in sandstones of the upper part of thelower Siwalik Group is an indication of wide-spread erosion of granitic rocks beginning at ~11Ma. K-feldspar is sparse in modern sands derivedfrom the Greater Himalayan zone in Nepal, butabundant in sands derived from the Dadeldhurathrust sheet. Most of this K-feldspar is derivedfrom the Dadeldhura granite and the graniticgneisses that surround it. We therefore tentativelyinterpret the influx of K-feldspar in the upper partof the lower Siwalik Group to be an indication ofwidespread, deep erosion of the Dadeldhura

DECELLES ET AL.

14 Geological Society of America Bulletin, January 1998

0 m

2000

3000

4000

0 0.2 0.4 0.6 0.8

MID

DLE

SIW

ALI

K

LOW

ER

SIW

ALI

KD

UM

RI F

M.

K/(K + P)

0.2 0.4 0.6 0.8 1.0

PHYLLITE

0.2 0.4 0.6

CARBONATE

0.714

0.716

0.718

0.720

0.722

0.724

0.726

0.728

0.730

0.732

A. B. C. D.

1000

~10.8 Ma

Ky SiE.

Sr/ Sr87 86

Figure 8. Plots showing the stratigraphic distributions. (A) K-feldspar and plagioclase. (B) Phyllite grains. (C) Carbonate lithic grains, in-cluding dolostone, limestone, and marble. (D) The 87Sr/86Sr composition of paleosol carbonate (diamonds) and carbonate fraction of associateddetrital silt (stars) from Quade et al. (1997). (E) Occurrence of kyanite (Ky) and sillimanite (Si). Phyllite and carbonate contents shown in B andC are normalized to total volcanic, carbonate, and phyllitic lithic grains.

thrust sheet. The zircon ages provide supportfor a significant source of K-feldspar in theCambrian–Ordovician granites, and the presenceof accessory high-grade metamorphic mineralgrains (kyanite and sillimanite) and abundant800–1000 Ma zircons attests to continued supply

of sediment from the Greater Himalayan zone.The increased amounts of older (>2.0 Ga) zir-cons also suggest erosion of Lesser Himalayanrocks. Together, the data imply that a major dis-placement event began on the system of thrustfaults that built the duplex beneath the Dadel-

dhura thrust sheet at about the time of depositionof the upper part of the lower Siwalik Group (~11Ma). Horses of the Galyang and Lakarpata For-mations were imbricated upon each other andfault slip was fed updip and southward into a roofthrust (the Ramgarh thrust) near the base of theDadeldhura thrust sheet (Fig. 15C). As the duplexgained structural relief, the overlying Dadeldhurathrust sheet was folded into its present synformalshape, and much of the thrust sheet was eroded.The timing of this displacement event is synchro-nous with the initiation of Main Boundary thrustdisplacement in northern India as interpreted byMeigs et al. (1995). However, in the model pre-sented here, the Main Boundary thrust (in thestrict sense) was the last thrust carrying LesserHimalayan rocks to be emplaced, and thus itsmain phase of displacement may have beensomewhat later.

KINEMATIC HISTORY OF THE HIMALAYAN FOLD-THRUST BELT

Geological Society of America Bulletin, January 1998 15

Figure 9. Photomicrographs of Siwalik Group sandstones, all under crossed polarizers. (A) Quartz-mica schist (QM), monocrystalline quartz(Qm), and foliated polycrystalline quartz (Qpt). (B) An orthoquartzite grain in the lower Siwalik Group. (C) A metacarbonate grain, probablya fragment of marble derived from the Lesser Himalayan zone, in the middle Siwalik Group. (D) Volcanic lithic (Lv) and plagioclase feldspar(P) grains, in the lower Siwalik Group. Widths of A, B, and D are ~1 mm; width of C is ~0.5 mm.

CLAST COUNTS

0% 50% 100%

SK1

SK2

KK1

KK2 CARBONATE

CHERT

QUARTZITE

SS/SILT

Figure 10. Chart showing composition of upper Siwalik Group conglomerates, based on clastcounts of ~100 each at Khutia Khola (KK1, KK2) and Surai Khola (SK1, SK2).

Upper Siwalik Group conglomerates were de-rived from nearby high-relief sources of low-grade metasedimentary and sedimentary rocks inthe Main Boundary thrust sheet, probably duringits main phase of emplacement. Clasts of lithicsandstone, undoubtedly derived from the middleSiwalik Group, constitute a significant propor-tion of some clast counts. These were probablyderived from the hanging walls of intra-Siwalikthrusts and/or from now completely eroded Si-walik outcrops on the hanging wall of the MainBoundary thrust. By the time of deposition of theupper Siwalik Group, the thrust front had ad-vanced into the preserved outcrop belt of forelandbasin deposits and, as a result, middle Siwaliksandstones were recycled (Fig. 15D). The persis-tence of Cambrian–Ordovician, ~800–1100 Ma,and >1700 Ma zircons in the upper SiwalikGroup, as well as K-feldspar and high-grademetamorphic minerals, indicates continued sup-ply of sand-sized material from the Greater Him-alayan zone and the Dadeldhura thrust sheet.Some of the older (Precambrian) zircons alsocould have been recycled from Dumri and/orSiwalik sandstones.

Earthquake seismic evidence indicates that thelarge ramp beneath the Lesser Himalayan duplexis still active (Pandey et al., 1995), and this dis-placement may be accommodated in the frontalthrust belt by displacement in the Main Frontalthrust system (Nakata, 1989). Recent global po-sitioning satellite (GPS) studies suggest that themajority of convergence measurable at the sur-face is occurring between the Greater Himalayanzone and the northern part of the Lesser Hima-layan zone (Bilham et al., 1997). The overall rateof crustal shortening accommodated within theLesser Himalaya and Subhimalaya since ~11 Mais ~21 mm/yr, which is comparable to modernrates of underthrusting of the northern Indianplate beneath the Himalaya calculated by Bilhamet al. (1997).

DISCUSSION

Axial Paleodrainage

The data presented here have direct bearing onthe timing of development of a large axial (i.e.,eastward-flowing) river system during Miocene–Pliocene time in the Indo-Gangetic forelandbasin. It has been inferred that an axial river sys-tem developed during deposition of the middle Si-walik Group in response to rapid subsidence inthe proximal foreland basin during emplacementof the Main Boundary thrust sheet (Burbank et al.,1996), and that the present transverse drainagepattern in the proximal foreland is a relatively re-cent (post-Pliocene) development, due to ero-sional unloading and isostatic rebound of the fore-

DECELLES ET AL.

16 Geological Society of America Bulletin, January 1998

Qm

F Lt

SK3

SR

RR2

RR1

RK

MKKR2

KR1SK5

LK

Qm

P K

RR1

SR

SK3

RR2

RK

MKKR2

KR1SK5

LK

GHS, SUN KOSI RIVERDT/MT SHEETS

ENTIRE THRUST BELTLHS, SETI AND RAPTI RIVERS

A.

B.

C.

D.

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

RR

1S

R

SK

3

RR

2

RK

MK

KR

2K

R1

SK

5

LK

K/(K + P)

CARBONATEGRAINS

Figure 11. Ternary diagrams showing the modal framework compositions of modern sandsfrom the Sun Kosi (SK3, SK5), Karnali (KR1, KR2), Mahakali (MK), Seti (SR), Rao (RK),Rapti (RR1, RR2), and Lamykheti (LK) Rivers of Nepal. See Appendix for locations of samplesites. In A and B, solid symbols are for sands in rivers draining the entire thrust belt and onlythe Greater Himalayan zone, and open symbols are for sands in rivers draining the crystallinethrust sheets and the Lesser Himalayan zone. Shaded columns in C and D are for rivers drain-ing the Lesser Himalayan zone and crystalline thrust sheets.

land basin (Burbank, 1992). The paleocurrent andlithofacies data indicate that the transversedrainage pattern in the Indo-Gangetic forelandbasin has existed since at least middle Miocenetime. Whereas the data presented in this papersupport the interpretation that the middle SiwalikGroup was deposited during the growth of theLesser Himalayan duplex (of which the MainBoundary thrust is an imbricate), Siwalik sand-stones that crop out in the frontal thrust belt inwestern Nepal contain no evidence for axialdrainage. If the deposits of such a river system arepreserved at the latitude of our study, they must bein older rocks. An obvious candidate in the appro-priate stratigraphic position is the Dumri Forma-tion. Channel sandstones in the lower Dumri For-mation are generally much thicker than those inthe lower Siwalik Group, suggesting that therivers that deposited the early Dumri Formation

were larger. However, paleocurrent data from theDumri Formation indicate predominantly west-southwestward paleodrainage (Sakai, 1983; au-thors’ unpub. data), as would be expected for anaxial system flowing into the Indus portion of theforeland basin. Until more data are available onthe sedimentology, age, and regional paleo-drainage patterns for the Dumri, this interpreta-tion will remain speculative. The deposits of anaxial system correlative with the Siwalik trans-verse systems are probably in the subsurface be-neath the Indo-Gangetic plain.

Implications for Geochemical Models ofHimalayan Unroofing

The data presented herein provide new con-straints on geochemical models of Himalayanunroofing. A major increase in 87Sr/86Sr of paleo-

sol carbonate nodules in Siwalik flood-plain de-posits occurred at ~9–10 Ma (Quade et al., 1997;Fig. 8D). The shift in paleosol carbonate 87Sr/86Sris mirrored by data from detrital carbonate silt inpaleosol parent material (Fig. 8D). In paleosolcarbonate from other sections of the SiwalikGroup in Nepal and in resedimented pedogenicclays in the Bengal fan (Derry and France-Lanord, 1996), the 87Sr/86Sr ratio rises abruptly at~9–7 Ma and then falls to values found in mod-ern Himalayan rivers after ~3 Ma (Quade et al.,1997). A similar fall in the 87Sr/86Sr ratio in theKhutia Khola section has not been documentedbecause the upper part of the section has been re-moved by faulting.

Can the trends in the 87Sr/86Sr ratio in the Si-walik Group and Bengal fan be coupled to trendsin petrographic data and kinematic-erosionalevents in the thrust belt? In the Khutia Khola sec-tion, it is tempting to correlate several short-termpeaks in the 87Sr/86Sr ratio of paleosol carbonatewith peaks in the carbonate and K-feldspar graincontents of associated sandstones (Fig. 8). That acorrelation with carbonate grain content shouldexist is suggested by the general coherence of thelimited isotopic data from detrital carbonate silt(only 10 measurements) and paleosol carbonate(Fig. 8D). In addition,87Sr/86Sr ratios of watersfrom springs in the Dadeldhura granite outcropbelt are >0.730 (Quade, unpub. data) and whole-rock initial Sr ratios from the granite itself aregreater than 0.726 (Einfalt et al., 1993). However,the major rise in 87Sr/86Sr since ~9 Ma postdatesthe onset of increases in detrital carbonate and K-feldspar grains in Siwalik sandstones by ~1–2m.y. (Fig. 8, A, C, and D). The modern sand pet-rographic data strongly suggest that the mainsources of K-feldspar for Siwalik sandstoneswere granitic rocks in the crystalline thrustsheets, and the carbonate grains were clearly de-rived from metamorphosed dolostone and lime-stone in the Lesser Himalayan zone. Low oxygenisotopic values (δ18O [Peedee belemnite] =–15‰ to –9‰) and low carbon isotopic values(–6‰ to –2‰) from the paleosol detrital carbon-ate also suggest a metamorphosed carbonatesource terrane (Quade et al., 1997). Thus, the keyto understanding the 87Sr/86Sr record is the bulk-rock kinematic history of the Himalayan thrustbelt and the evolving distribution of rock typesand sources of radiogenic Sr.

Consideration of the structural geometry andkinematic history of the Himalayan thrust belt inwestern Nepal (Figs. 2 and 15) provides a plausi-ble explanation for the observed Neogene trendsin the 87Sr/86Sr ratio of flood-plain and Bengalfan sediments. Deep erosion of the Dadeldhurathrust sheet and the other crystalline thrust sheetsmust have commenced with growth of the LesserHimalayan duplex, beginning ~11 Ma (Fig.

KINEMATIC HISTORY OF THE HIMALAYAN FOLD-THRUST BELT

Geological Society of America Bulletin, January 1998 17

Figure 12. U-Pb concordia diagrams for detrital zircons from: (a–c) three modern river sandsin western (Mahakali) and central (Karnali and Narayani) Nepal, and (d–e) lower, middle, andupper Siwalik Group sandstones at Khutia Khola. Each square represents a single detrital zir-con grain; the solid squares are considered to be concordant enough to represent crystallizationages and are included in summary histograms in Figure 13.

15C). This would have delivered abundant K-feldspar grains to the foreland basin, and carbon-ate grains would have become increasingly avail-able as the crystalline sheets were erosionallybreached and the underlying Lesser Himalayanduplex was exposed to weathering (Fig. 15, Cand D). For ~5–6 m.y., the tandem high 87Sr/86Srsources in the Cambrian–Ordovician granites ofthe crystalline thrust sheets and the highly solu-ble metamorphosed carbonate rocks of the LesserHimalayan duplex would have supplied a large

flux of radiogenic Sr. As more of the crystallinethrust sheets were eroded, however, their contri-butions to the flux of radiogenic Sr would havediminished, and the overall ratio preserved in theSiwalik Group and Bengal fan would have de-creased. If correct, this scenario would imply thatthe regional rate of erosion of the 10–15-km-thick Dadeldhura thrust sheet must have been~2–3 mm/yr. In the northern Indian Ocean, deep-sea sediments record an abrupt increase in massaccumulation rate during the period ~11–6 Ma

(Rea, 1992), which suggests a linkage with rapiderosion of the crystalline thrust sheets and thegrowing Lesser Himalayan duplex.

The increase of 87Sr/86Sr preserved in Neo-gene foreland basin deposits and the Bengal fanmay have been caused by the development of theLesser Himalayan duplex, erosion of the crys-talline thrust sheets, and incision into the under-lying Lesser Himalayan rocks. However, the ef-fect of this major kinematic-erosional event onthe Neogene marine 87Sr/86Sr record (Hodell andWoodruff, 1994) is not clear. The marine recordexhibits a nearly inverse correlation with massaccumulation rate in the northern Indian Ocean(Rea, 1992), and with the 87Sr/86Sr records of theBengal fan (Derry and France-Lanord, 1996) andSiwalik Group (Quade et al., 1997). Changes inSr flux as well as 87Sr/86Sr ratio must have playedimportant roles in controlling the seawater ratio.It is also quite likely that Greater Himalayanrocks provided radiogenic Sr throughout theNeogene, because the petrographic data suggestthat these rocks were exposed by early Miocenetime during deposition of the Dumri Formation.The 87Sr/86Sr ratio of seawater has been risingrapidly since ~40 Ma, so it is probable that multi-ple sources of radiogenic Sr, each coming intoplay at a time dictated by the kinematic history ofthe thrust belt, contributed to the long-term trendin Sr composition.

CONCLUSIONS

(1) The three-fold, informal lithostratigraphicsubdivision of the Siwalik Group used in otherparts of the Indo-Gangetic foreland basin isapplicable to the Siwalik Group in westernNepal, the lower member consisting of >850 mof fluvial channel sandstones alternating with ox-idized calcic paleosols, the middle member con-sisting of >2400 m of very thick (>20 m) channelsandstones and drab-colored Histosols, and theupper member comprising at least 1000 m ofmainly gravelly braided river deposits. The in-crease in fluvial channel size from the lower tomiddle Siwalik Group probably occurred in re-sponse to increased seasonal discharge.

(2) Paleocurrent data indicate that the overalldrainage pattern in the northern part of the Indo-Gangetic foreland basin has been similar to thepresent-day drainage pattern since middleMiocene time.

(3) Provenance data from the Dumri Forma-tion and the lower-middle Siwalik Groupdemonstrate an overall upsection enrichment infeldspar, carbonate, and high-grade metamor-phic minerals at the expense of quartzose grainsand low-grade metasedimentary and sedimen-tary lithic grains. K-feldspar grains increaseabruptly in the upper part of the lower Siwalik

DECELLES ET AL.

18 Geological Society of America Bulletin, January 1998

Figure 13. Summary histograms of U-Pb zircon ages from Siwalik Group sandstone samples,sand samples from modern rivers, and for comparison, ages reported by Parrish and Hodges(1996) from the Main Central thrust (MCT) zone and Lesser Himalayan source terranes in theLangtang area of north-central Nepal. Star is age of Dadeldhura granite reported in this paper.

Figure 14. U-Pb concordia diagram for eight individual zircon grains from the Dadeldhuragranite. Four grains are concordant at an age of 492 ± 6 Ma, and the other four have inheritedcomponents of about 1687 Ma and 2465 Ma when regressed through 492 Ma.

Group, suggesting that granitic rocks becamewidely exposed at ~11 Ma. Modal point countsof modern river sands derived from knownsource terranes suggest that the main sources ofK-feldspar were the Cambrian–Ordovician gran-ite and associated orthogneiss of the Dadeldhurathrust sheet. Provenance data from the upper Si-walik Group indicate local low-grade metasedi-mentary source terranes in the Lesser Himalayanrocks of the Main Boundary thrust sheet, in ad-dition to continued influx from the Dadeldhurasheet and the Greater Himalayan zone.

(4) Combined provenance data and crosscut-ting relationships outline the history of majorthrust displacements in western Nepal. The MainCentral thrust was active during early Miocenetime and probably was the major source of theDumri Formation. The Dadeldhura thrust cuts theDumri Formation and probably was emplaced~15–11 Ma, mainly during deposition of thelower Siwalik Group. Beginning ~11 Ma, thetrailing portion of the Dadeldhura sheet wasfolded into a regional antiform by the growth of alarge duplex in underlying Lesser Himalayanrocks. Slip on faults within the duplex was fedinto a roof thrust near the base of the Dadeldhurasheet (the Ramgarh thrust) and eventually (by~4–5 Ma) into the Main Boundary thrust. Con-tinued duplex growth is demonstrated by concen-trated microseismicity, and slip is transferredsouthward into the Main Frontal thrust system.On the basis of minimum estimates of shorteningin Lesser Himalayan and Subhimalayan rocks,the overall rate of crustal shortening since ~11Ma is ~21 mm/yr, which is comparable to mod-ern rates of underthrusting of the northern Indianplate beneath the Himalaya.

(5) Comparison of U-Pb ages of detrital zir-cons from the Siwalik Group and modern riverswith available zircon dates from Himalayansource terranes indicates that this method may bea powerful provenance tool in Himalayan stud-ies. Cambrian–Ordovician (460–530 Ma) detritalzircons were probably derived from the Dadel-dhura granite (which yielded a U-Pb zircon ageof 492 ± 6 Ma) in the hanging wall of the Dadel-dhura thrust, and perhaps from poorly datedGreater Himalayan orthogneisses. There are800–1000 Ma zircons, probably derived from theGreater Himalayan zone, present throughout theSiwalik Group, and these indicate that the rocksin the hanging wall of the Main Central thrustsystem have been a major source of sedimentsince early middle Miocene time. Zircons olderthan ~1.8 Ga were probably derived from LesserHimalayan rocks and perhaps the upper part ofthe Dadeldhura sheet. None of the 113 detritalzircons dated in this study has an age consistentwith derivation from the Cenozoic leucogranitesof the higher Himalaya.

KINEMATIC HISTORY OF THE HIMALAYAN FOLD-THRUST BELT

Geological Society of America Bulletin, January 1998 19

DT SHEET

DT

LOWER SIWALIK GR. (~15 -11 Ma)

B.

PLAG. + QTZ. + METAMORPHICS

GHZ

DUMRI FM. (EARLY MIOCENE)

A.

TIBETAN THRUSTS

STDSMCT

QTZ. + PLAG. + PHYLLITE

GHZ

LHS METASEDS.

GHZ

MFT

UPPER SIWALIK GR. (<5 Ma)

D.

RECYCLED SIWALIK GR.+ LHS METASEDS

RT-DT

GHS METAMORPHICS

GHZ

MIDDLE SIWALIK GR. (~11-5 Ma)

C.

K-SPAR + METASEDS.

MCT

LHZ

LHZMBT

GHS METAMORPHICS

K-SPAR + METASEDS.

DADELDHURA GRANITE

LESSER HIMALAYAN DUPLEX

LESSER HIMALAYAN DUPLEX

THZ

THZ

THZ

THZ

Figure 15. Schematic cross sections showing a hypothetical early Miocene–Pliocene unroof-ing history for the Himalayan thrust belt in western Nepal. Abbreviations: THZ—Tibetan Him-alayan zone; GHZ—Greater Himalayan zone; LHZ—Lesser Himalayan zone; STDS—SouthTibetan detachment system; DT—Dadeldhura thrust; RT—Ramgarh thrust; other abbrevia-tions as used in text. (A) Deposition of the Dumri Formation: Sediment is derived mainly fromthe THZ and the cover rocks of the GHZ during emplacement of Main Central thrust. Minoramounts of plagioclase could have been derived from GHZ gneisses exposed by initial displace-ment on the South Tibetan detachment system. (B) Lower Siwalik Group: Sediment is derivedmainly from the GHZ and cover section of the active Dadeldhura thrust sheet. (C) MiddleSiwalik Group: Growth of large duplex in LHZ folds and passive uplifts trailing part of DTsheet (large vertical arrow). K-feldspar-rich sediment is derived from metamorphic and graniticrocks of the DT sheet, including the Dadeldhura granite, and from continued erosion of GHZrocks. Carbonate grains increase as LHZ duplex is erosionally breached. (D) Upper SiwalikGroup: Main Boundary thrust is emplaced as a southernmost imbricate of the LHZ duplex, andslip is transferred to thrusts in the Main Frontal thrust system. Conglomerate provenance isdominated by local sources in the frontal part of the Main Boundary thrust sheet, but sand-sizedsediment is continually supplied from the DT sheet, the LHZ duplex, and the GHZ.

(6) Combined petrographic and U-Pb iso-topic provenance data indicate that Cambrian–Ordovician granites in the hanging wall of theDadeldhura thrust system (and similar granites incentral and eastern Nepal) and metamorphosedLesser Himalayan carbonate rocks may haveplayed a heretofore unrecognized role in control-ling the 87Sr/86Sr composition of middle–lateMiocene synorogenic sediments in the Indo-Gangetic foreland basin, the Bengal fan, andglobal seawater. However, multiple potentialsources of radiogenic Sr exist in the Himalaya;the isotopic characteristics of these potentialsource terranes must be determined if we are tomake progress in understanding the effects ofHimalayan weathering on seawater Sr composi-tion. Because the exposure of Himalayan sourceterranes was ultimately controlled by the kine-matic tempo of the thrust belt, a much more de-tailed understanding of the latter is required in or-der to assess the 87Sr/86Sr contributions of thevarious terranes.

ACKNOWLEDGMENTS

This research was supported by National Sci-ence Foundation grant EAR-9418207 and a grantfrom the University of Arizona Foundation. Weare grateful to David Richards and Robert Butlerfor assistance with paleomagnetic sample collec-tion and analysis, and to Gautam Mitra, PeterCopeland, Roger Bilham, and Mark Harrison foruseful information about the Himalayan fold-thrust belt. Kip Hodges, Doug Burbank, and GarySmith reviewed the manuscript and providedmany helpful suggestions for improvement.

APPENDIX. LOCATIONS OF SAND SAMPLESFROM MODERN RIVERS.

SK3—Sun Kosi River; lat 27°40′08.7′′N, long85°43′49.7′′E. Source terrane includes mainly GreaterHimalayan zone, plus some Lesser Himalayan zone.

SK5—Sun Kosi River; ~50 m downstream fromFriendship Bridge, at town of Kodari on the border be-tween Tibet and Nepal. Source terrane includes onlyGreater Himalayan zone.

KR1—Karnali River; ~1 km upstream from Karnalibridge at mouth of Karnali Gorge. Source terrane in-cludes entire Himalayan thrust belt.

KR2—Karnali River; ~2 km downstream from Kar-nali bridge at mouth of Karnali Gorge. Source terraneincludes entire Himalayan thrust belt.

MK—Mahakali River; ~0.5 km north of Mahen-dranagar Bridge where the river exits the Himalaya.Source terrane includes entire Himalayan thrust belt.

RR1—Rapti River; lat 27°28′37.3′′N, long85°02′32.0′′E. Source terrane includes Mahabharatthrust sheet and Lesser Himalayan rocks in hangingwall of Main Boundary thrust.

RR2—Rapti River; lat 27°30′28.4′′N, long85°03′10.2′′E, ~50 m upstream from yellow bridgeat Bhainsedhoban. Source terrane includes only theMahabharat thrust sheet.

RK—Rao Khola; lat 29°18′17.9′′N, long 80°40′

08.3′′E. Source terrane includes only the Dadel-dhura thrust sheet.

LK—Lamikheti Khola; lat 29°18′16.4′′N, long80°46′20.7′′E. Source terrane includes only theDadeldhura thrust sheet.

SR—Seti River (of western Nepal); ~1 km upstreamfrom the confluence with Rao Khola. Source terrane in-cludes Lesser Himalayan and Greater Himalayanzones, plus crystalline thrust sheets northeast ofDadeldhura thrust sheet.

REFERENCES CITED

Acharyya, 1994, The Cenozoic foreland basin and tectonics ofthe eastern Sub-Himalaya: Problems and prospects: Him-alayan Geology, v. 15, p. 3–21.

Amatya, K. M., and Jnawali, B. M., 1994, Geological map ofNepal: Kathmandu, Nepal, Department of Mines andGeology, scale 1:1 000 000.

Appel, E., and Rössler, W., 1994, Magnetic polarity stratigra-phy of the Neogene Surai Khola section (Siwaliks, SWNepal): Himalayan Geology, v. 15, p. 63–68.

Awasthi, N., Sarkar, S., and Prasad, H., 1994, Vegetation andpalaeoenvironment of Siwalik succession in Surai Kholaarea, Nepal: Himalayan Geology, v. 15, p. 107–119.

Bashyal, R. P., 1986, Geology of the Lesser Himalaya, farwestern Nepal: Nancy, France, Centre National de laRecherche Scientifique, Science de la Terre Memoire 47,p. 31–42.

Bilham, R., Larson, K., and Freymuller, J., 1997, Indo-Asianconvergence rates in the Nepal Himalaya: Nature, v. 386,p. 61–64.

Brunel, M., and Kienast, J.-R., 1986, Etude pétro-structuraledes chevauchements ductile himalayens sur la transver-sale de l’Everest-Makalu (Népal oriental): Canadian Jour-nal of Earth Sciences, v. 23, p. 1117–1137.

Burbank, D. W., Beck, R. A., and Mulder, T., 1996, The Hima-layan foreland basin,inYin,A., and Harrison, T. M., eds.,The tectonics of Asia: London, Cambridge UniversityPress, p. 205–226.

Burbank, D. W., 1992, Causes of recent Himalayan upliftdeduced from deposited patterns in the Ganges basin:Nature, v. 357, p. 680–683.

Burchfiel, B. C., Zhiliang, C., Hodges, K. V., Yuping, L.,Royden, L. H., Changrong, D., and Jiene, X., 1992, Thesouth Tibetan detachment system, Himalayan orogen:Extension contemporaneous with and parallel to short-ening in a collisional mountain belt: Geological Societyof America Special Paper 269, 51 p.

Burg, J. P., and Chen, G. M., 1984, Tectonics and structuralzonation of southern Tibet: Nature, v. 311, p. 219–223.

Cande, S. C., and Kent, D. V., 1995,A new geomagnetic polar-ity time scale for the Late Cretaceous and Cenozoic: Jour-nal of Geophysical Research, v. 100, p. 6093–6095.

Coleman, M. E., 1996, Orogen-parallel and orogen-perpendic-ular extension in the central Nepalese Himalayas: Geolog-ical Society of America Bulletin, v. 108, p. 1594–1607.

Copeland, P., LeFort, P., Ray, S. M., and Upreti, B. N., 1996,Cooling history of the Kathmandu crystalline nappe:40Ar/39Ar results: Flagstaff, Arizona, 11th Himalaya-Karakoram-Tibet workshop, abstracts, p. 39.

Copeland, P., Harrison, T. M., Hodges, K. V., Maruejol, P.,LeFort, P., and Pêcher, A., 1991, An early Pliocene ther-mal disturbance of the Main Central thrust, central Nepal:Implications for Himalayan tectonics: Journal of Geo-physical Research, v. 96, p. 8475–8500.

Corvinus, G., 1994, The Surai Khola and Rato Khola fossilif-erous sequences in the Siwalik Group, Nepal: HimalayanGeology, v. 15, p. 49–61.

Critelli, S., and Ingersoll, R. V., 1994, Sandstone petrology andprovenance of the Siwalik Group (northwestern Pakistanand western-southeastern Nepal): Journal of SedimentaryResearch, v. A64, p. 815–823.

DeCelles, P. G., Langford, R. P., and Schwartz, R. K., 1983,Two new methods of paleocurrent determination fromtrough cross-stratification: Journal of Sedimentary Petrol-ogy, v. 53, p. 629–642.

Deniel, C.,Vidal, P., Fernandez,A., LeFort, P., and Peucat, J. J.,1987, Isotopic study of the Manaslu granite (Himalaya,Nepal): Inferences on the age and source of Himalayan

leucogranites: Contributions to Mineralogy and Petrol-ogy, v. 96, p. 78–92.

Derry, L. A., and France-Lanord, C., 1996, Neogene Hima-layan weathering history and river: Impact on the marineSr record: Earth and Planetary Science Letters, v. 142,p. 59–74.

Dhital, M. R., and Kizaki, K., 1987, Structural aspect of thenorthern Dang, Lesser Himalaya: University of Ryukyus,Bulletin of the College of Science, no. 45, p. 159–182.

Edmond, J. M., 1992. Himalayan tectonics, weatheringprocesses, and the strontium isotope record in marinelimestones: Nature, v. 258, p. 1594–1597.

Einfalt, H. C., Hoehndorf, A., and Kaphle, K. P., 1993, Radio-metric age determination of the Dadeldhura granite,Lesser Himalaya, Far Western Nepal: SchweizerischeMineralogische und Petrographische Mitteilungen, v. 73,p. 97–106.

Ferrara, G., Lombardo, B., and Tonarini, S., 1983, Rb/Srgeochronology of granites and gneisses from the MountEverest region, Nepal Himalaya: Geologisch Rundschau,v. 72, p. 119–136.

France-Lanord, C., and LeFort, P., 1988, Crustal melting andgranite genesis during the Himalayan collision orogene-sis: Royal Society of Edinburgh Transactions, v. 99,p. 183–195.

France-Lanord, C., Derry, L., and Michard,A., 1993, Evolutionof the Himalaya since Miocene time: Isotopic and sedi-mentologic evidence from the Bengal Fan,in Treloar, P. J.,and Searle, M., eds., Himalayan Tectonics: p. 605–623.

Gansser, A., 1964, Geology of the Himalayas: London, Inter-science, 289 p.

Gautam, P., 1989, Magnetic properties of some late Paleozoicto Tertiary sedimentary rocks of Tansen area, LesserHimalaya, Nepal: Hokkaido University, Journal of theFaculty of Science, ser. 4, v. 23, p. 467–487.

Gehrels, G. E., McClelland, W. C., Samson, S. D., and Patchett,P. J., 1991, U-Pb geochronology of detrital zircons from acontinental margin assemblage in the northern CoastMountains, southeastern Alaska: Canadian Journal ofEarth Sciences, v. 28, p. 1285–1300.

Gohain, K., and Parkash, B., 1990, Morphology of the Kosimegafan,in Rachocki, A. H., and Church, M., eds., Allu-vial fans: A field approach: Chichester, United Kingdom,John Wiley and Sons, p. 151–178.

Gupta, S., 1997, Himalayan drainage patterns and the origin offluvial megafans in the Ganges foreland basin: Geology,v. 25, p. 11–14.

Harrison, T. M., Copeland, P., Kidd, W. S. F., and Yin,A., 1992,Raising Tibet: Science, v. 255, p. 1663–1670.

Harrison, T. M., Copeland, P., Hall, S. A., Quade, J., Burner, S.,Ojha, T. P., and Kidd, W. S. F., 1993, Isotopic preservationof Himalayan/Tibetan uplift, denudation, and climatichistories in two molasse deposits: Journal of Geology,v. 100, p. 157–173.

Harrison, T. M., Ryerson, F. J., Le Fort, P., Yin, A., Lovera,O. M., and Catlos, E. J., 1997, A late Miocene–Plioceneorigin for the central Himalayan inverted metamorphism:Earth and Planetary Science Letters, v. 146, p. E1–E7.

Hisatomi, K., 1990, The sandstone petrography of the Churia(Siwalik) Group in the Arung Khola-Binai Khola area,west central Nepal: Wakayama University of Natural Sci-ence, Bulletin of the Faculty of Education, v. 39, p. 5–29.

Hisatomi, K., and Tanaka, S., 1994, Climatic and environmen-tal changes at 9 and 7.5 Ma in the Churia (Siwalik)Group, west central Nepal: Himalayan Geology, v. 15,p. 161–180.

Hodell, D. A., and Woodruff, F., 1994, Variations in the stron-tium isotopic ratio of seawater during the Miocene: Strati-graphic and geochemical implications: Paleoceanogra-phy, v. 9, p. 405–426.

Hodell, D. A., Mueller, P. A., McKenzie, J. A., and Mead, G. A.,1989, Strontium isotope stratigraphy and geochemistry ofthe late Neogene ocean: Earth and Planetary Science Let-ters, v. 92, p. 165–178.

Hodges, K. V., and Silverberg, D. S., 1988, Thermal evolutionof the Greater Himalaya, Garhwal, India: Tectonics, v. 7,p. 583–600.

Hodges, K. V., Parrish, R. R., and Searle, M. P., 1996, Tectonicevolution of the central Annapurna Range, NepaleseHimalayas: Tectonics, v. 15, p. 1264–1291.

Hubbard, M. S., and Harrison, T. M., 1989,40Ar/39Ar con-straints on deformation and metamorphism in the Main

DECELLES ET AL.

20 Geological Society of America Bulletin, January 1998

Central thrust zone and Tibetan Slab, eastern NepalHimalaya: Tectonics, v. 8, p. 865–880.

Ingersoll, R. V., Bullard, T. F., Ford, R. L., Grimm, J. P., Pickle,J. D., and Sares, S. W., 1984, The effect of grain size ondetrital modes: A test of the Gazzi-Dickinson point-counting method: Journal of Sedimentary Petrology,v. 54, p. 103–116.

Kayastha, N. B., 1992, Stratigraphy of the lower Tertiary rocksof Nepal Himalaya: Journal of Nepal Geological Society,v. 8, p. 21–30.

LeFort, P., 1981, Manaslu leucogranite:A collision signature ofthe Himalaya: A model for its genesis and emplacement:Journal of Geophysical Research, v. 86, p. 10545–10568.

LeFort, P., 1986, Metamorphism and magmatism during theHimalayan collision,in Coward, M. P., and Ries, A. C.,eds., Collision tectonics: London, Geological SocietySpecial Publication 19, p. 159–172.

LeFort, P., Debon, F., Pêcher, A., Sonet, J., and Vidal, P., 1986,The 500 Ma magmatic event in Alpine Southern Asia: Athermal episode at Gondwana scale: Nancy, France, Cen-tre National de la Recherche Scientifique, Science de laTerre Memoir 47, p. 191–209.

Macfarlane, A. M., 1993, Chronology of tectonic events in thecrystalline core of the Himalaya, Langtang National Park,Central Nepal: Tectonics, v. 12, p. 1004–1025.

Macfarlane, A. M., Hodges, K. V., and Lux, D., 1992, A struc-tural analysis of the main Central thrust zone, LangtangNational Park, central Nepal Himalaya: Geological Soci-ety of America Bulletin, v. 104, p. 1389–1402.

Meigs, A. J., Burbank, D. W., and Beck, R. A., 1995,Middle–late Miocene (>10 Ma) formation of the MainBoundary thrust in the western Himalaya: Geology, v. 23,p. 423–426.

Miall, A. D., 1996, The geology of fluvial deposits: Berlin,Springer-Verlag, 582 p.

Mohindra, R., Parkash, B., and Prasad, J., 1992, Historicalgeomorphology and pedology of the Gandak megafan,middle Gangetic plains, India: Earth Surface Processesand Landforms, v. 17, p. 643–662.

Mugnier, J. L., Mascle, G., and Faucher, T., 1993, Structure ofthe Siwaliks of western Nepal: An intracontinental ac-cretionary prism: International Geology Reviews, v. 35,p. 32–47.

Najman,Y. M. R., Enkin, R. J., Johnson, M. R. W., Robertson,A. H. F., and Baker, J., 1994, Palaeomagnetic dating ofthe earliest continental Himalayan foredeep sediments:Implications for Himalayan evolution: Earth and Plane-tary Science Letters, v. 128, p. 713–718.

Nakata, T., 1989, Active faults of the Himalaya of India andNepal,in Malinconico, L. L., Jr., and Lillie, R. J., eds.,Tectonics of the western Himalayas: Geological Societyof America Special Paper 232, p. 243–264.

Ni, J., and Barazangi, M., 1984, Seismotectonics of the Him-alayan collision zone: Geometry of the underthrustingIndian plate beneath the Himalaya: Journal of Geophys-ical Research, v. 89, p. 1147–1163.

Pandey, M. R., Tandukar, R. P., Avouac, J. P., Lave, J., andMassot, J. P., 1995, Interseismic strain accumulation onthe Himalayan crustal ramp (Nepal): Geophysical Re-search Letters, v. 22, p. 751–754.

Parrish, R. R., and Hodges, K. V., 1996, Isotopic constraints onthe age and provenance of the Lesser and Greater Hima-layan sequences, Nepalese Himalaya: Geological Societyof America Bulletin, v. 108, p. 904–911.

Pêcher, A., 1989, The metamorphism in the central Himalaya:Journal of Metamorphic Geology, v. 7, p. 31–41.

Quade, J., 1993, Major shifts in the 87Sr/86Sr ratios of large pa-leorivers draining the Himalayas of central Nepal over thepast 10 Ma: Geological Society of American Abstractswith Programs, v. 25, no. 6, p. A175.

Quade, J., Cerling, T. E., and Bowman, J. R., 1989, Develop-ment of the Asian Monsoon revealed by marked ecologicshift during the latest Miocene in northern Pakistan: Na-ture, v. 342, p. 163–166.

Quade, J., Cater, J. M. L., Ojha, T. P., Adam, J., and Harrison,T. M., 1995, Late Miocene environmental change inNepal and the northern Indian subcontinent: Stable iso-topic evidence from paleosols: Geological Society ofAmerica Bulletin, v. 107, p. 1381–1397.

Quade, J., Roe, L., DeCelles, P. G., and Ojha, T. P., 1997, Thelate Neogene 87Sr/86Sr record of lowland Himalyanrivers: Science, v. 276, p. 1828–1831.

Ratschbacher, L., Frisch, W., and Guanghua, L., 1994, Distrib-uted deformation in southern and western Tibet duringand after the India-Asia collision: Journal of GeophysicalResearch, v. 99, p. 19917–19945.

Raymo, M. E., and Ruddiman, W. F., 1992, Tectonic forcing oflate Cenozoic climate: Nature, v. 359, p. 117–122.

Rea, D. K., 1992, Delivery of Himalayan sediment to the north-ern Indian Ocean and its relation to global climate, sealevel, uplift, and seawater strontium,in The Indian Ocean:A synthesis of results from the Ocean Drilling Program:American Geophysical Union Geophysical Monograph70, p. 387–402.

Richter, F. M., Rowley, D. B., and DePaolo, D. J., 1992, Sr evo-lution of seawater: The role of tectonics: Earth and Plan-etary Science Letters, v. 109, p. 11–23.

Ruddiman, W. F., and Raymo, M. E., 1988, Northern Hemi-sphere climate regimes during the past 3 Ma: Possible tec-tonic connections: Royal Society of London Philosophi-cal Transactions, ser. B, v. 318, p. 411–430.

Rust, B. R., 1978, Depositional models for braided alluvium,inMiall, A. D., ed., Fluvial sedimentology: Canadian Soci-ety of Petroleum Geology Memoir 5, p. 605–625.

Sah, R. B., Ulak, P. D., Gajurel, A. P., and Rimal, L. N., 1994,Lithostratigraphy of Siwalik sediments of Amlekhganj-Hetauda area, sub-Himalaya of Nepal: Himalayan Geol-ogy, v. 15, p. 37–48.

Sahni, B., 1953,Angiosperm leaf impressions from the Kasaulibeds, N. W. Himalaya: The Palaeobotanist, v. 2, p. 83–87.

Sakai, H., 1983, Geology of the Tansen Group of the LesserHimalaya in Nepal: Kyushu University, Memoirs of theFaculty of Science, ser. D, v. 15, p. 27–74.

Sakai, H., 1989, Rifting of Gondwanaland and uplifting of theHimalayas recorded in Mesozoic and Tertiary fluvial sed-iments in the Nepal Himalayas,in Taira, A., and Masuda,F., eds., Sedimentary facies in the active plate margin:Tokyo, Terra Scientific Publishing Company, p. 723–732.

Schärer, U., and Allègre, C. J., 1983, The Palung granite (Him-alaya): High resolution U-Pb systematics in zircon andmonazite: Earth and Planetary Science Letters, v. 63,p. 423–432.

Schärer, U., Xu, R. H., and Allègre, C. J., 1986, U-(Th)-Pb sys-tematics and ages of Himalayan leucogranites, south Ti-bet: Earth and Planetary Science Letters, v. 77, p. 35–48.

Schelling, D., 1992, The tectonostratigraphy and structure ofthe eastern Nepal Himalaya: Tectonics, v. 11, p. 925–943.

Schelling, D., Cater, J., Seago, R., and Ojha, T. P., 1991, A bal-

anced cross-section across the central Nepal SiwalikHills: Hetauda to Amlekhganj: Hokkaido University,Journal of the Faculty of Science, ser. 4, v. 23, p. 1–9.

Searle, M. P., 1986, Structural evolution and sequence of thrust-ing in the High Himalayan, Tibetan Tethys and Indus su-ture zones of Zanskar and Ladakh, western Himalaya:Journal of Structural Geology, v. 8, p. 923–936.

Searle, M. P., 1991, Geology and tectonics of the KarakoramMountains: Chichester, United Kingdom, John Wiley andSons, 358 p.

Shrestha, S. B., et al. 1987, Geological map of far westernNepal: Kathmandu, Nepal, Royal Nepali Department ofMines and Geology, Scale 1:250 000.

Sinha, R., and Friend, P. F., 1994, River systems and their sedi-ment flux, Indo-Gangetic plains, northern Bihar, India:Sedimentology, v. 41, p. 825–845.

Smith, N. D., 1974, Sedimentology and bar formation in theupper Kicking Horse River: A braided meltwater stream:Journal of Geology, v. 82, p. 205–223.

Srivastava, P., and Mitra, G., 1994, Thrust geometries and deepstructure of the outer and lesser Himalaya, Kumaon andGarhwal (India): Implications for evolution of the Hima-layan fold-and-thrust belt: Tectonics, v. 13, p. 89–109.

Stöcklin, J., 1980, Geology of Nepal and its regional frame:Geological Society of London Journal, v. 137, p. 1–34.

Tanaka, S., 1994, Palaeosols in the Churia (Siwalik) Group, westcentral Nepal: Himalayan Geology, v. 15, p. 181–204.

Tokuoka, T., Takayasu, K.,Yoshida, M., and Hisatomi, K., 1986,The Churia (Siwalik) Group of the Arung Khola area, westcentral Nepal: Shimane, Japan, Shimane University,Memoirs of the Faculty of Science, v. 20, p. 135–210.

Upreti, B. N., 1990, An outline of the geology of far westernNepal: Journal of Himalayan Geology, v. 1, p. 93–102.

Upreti, B. N., 1996a, The Lesser Himalayan nappes of Nepal:Problems of their origin: Flagstaff, Arizona, Himalayan-Karakoram-Tibet Workshop, 11th, p. 165.

Upreti, B. N., 1996b, Stratigraphy of the western Nepal LesserHimalaya: A synthesis: Journal of Nepal Geological So-ciety, v. 13, p. 11–28.

Valdiya, K. S., 1980, Geology of the Kumaon Lesser Himalaya:Wadia Institute of Himalayan Geology, 291 p.

Wells, N. A., and Dorr, J. A., 1987, Shifting of the Kosi River,northern India: Geology, v. 15, p. 204–207.

West, R. M., Lukacs, J. R., Munthe, J., and Hussain, T., 1978,Vertebrate fauna from Neogene Siwalik Group, Dang val-ley, western Nepal: Journal of Paleontology, v. 52,p. 1015–1022.

West, R. M., Hutchison, J. H., and Munthe, J., 1991, Miocenevertebrates from the Siwalik Group, western Nepal: Jour-nal of Vertebrate Paleontology, v. 11, p. 108–129.

Willis, 1993a, Ancient river systems in the Himalayan fore-deep, Chinji village area, northern Pakistan: SedimentaryGeology, v. 88, p. 1–76.

Willis, 1993b, Evolution of Miocene fluvial systems in theHimalayan foredeep through a two kilometer-thick suc-cession in northern Pakistan: Sedimentary Geology, v. 88,p. 77–121.

Zhao,W.,Nelson,K. D.,and Project IN DEPTH team,1993,Deepseismic reflection evidence for continental underthrustingbeneath southern Tibet: Nature, v. 366, p. 557–559.

MANUSCRIPTRECEIVED BY THESOCIETY JUNE OCTOBER17, 1996REVISEDMANUSCRIPTRECEIVEDAPRIL 30, 1997MANUSCRIPTACCEPTEDMAY 15, 1997

KINEMATIC HISTORY OF THE HIMALAYAN FOLD-THRUST BELT

Geological Society of America Bulletin, January 1998 21

Printed in U.S.A.