Embed Size (px)
Citation preview
JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 96, NO. 83, PAGES 4065--1082. MARCH IO. I99I
Paleomagnetic Study of Mesozoic Continental Sediments Along the Northern
Tien Shan (China) and Heterogeneous Strain in Central Asia
YaT.T cTTpN,I JEAN.PASCAL COGNE,I,2 VnvceX'T COURTILLC)T,I JSAN-PHILIPPE AVOUAC,I PRUT TAPPO}.INER,1
GoNcer.JE weNcj MExrANc Bal5 HoNcz youj mrc u a AND cHlNsrctrrc wH 4 gRrc gurrgr.q,ur.3
A paleomagnetic snrdy of rocks from the northern foot of the Tien Shan and the southern
border of the Dzungar Basin, east of Urumqi (44.2"N, 86.008)' sparuring ages from middle
Jurassic to early Tertiary was carried out to constrain the tectonic evolution in central Asia
since Mesozoic time. Five middle Jurassic sites reveal a remagnetized direction close to the
presenr Earth field in geographic coordinates: D = 6.6" , I = 72.6" (o95 = 7.4o), Thirteen out of
l7 upper Jurassic and lower Cretaceous sites yield a characteristic direction (stratigraphic
coordinates) of D = 12.7o,1= 48.6o (495 = 5'5o)' Nine of 16 upper Cretaceous and lower
Tertiary sites provide a characteristic direction of D = 12.5",1= 51.3o (495 = 6.9o). The latter
two directions pass fold and reversal tests. The pole positions are close to each other and to the
Besse and Courtillor [1989, 1990] Eurasian apparent polar wander path, for ages ranging from
130 to 70 Ma. However, the diffcrence in paleolatitudes amounls to about 5.9" t 3.7o, which
could indicate significant continental shortening in the Altai Mountains and perhaps further
north, subsequenito India-Asia collision. The pole positions from the Dzu_ngar Basin are close
ro those found for rhe Tarim [Li et al., 1988a], leading to an insignificant paleolatitude
difference (3.0" t 6.9o), but showing a larger difference in declination (8.6' t 8.7'). These
paleomagnetic results are compatible with a model of heterogeneous deformation in the
western part of the collision zone between India and Siberia. A significant shortening in the
Altai, a slight counterclockwise rotation of the Dzungar block, the westward-increasing
shortening in rh" Ti"n Shan with attendant clockwise rotation of the Tarim block are all
consisteni with this model, in which Tibet, the Tien Shan anà the Altai undergo differential
suain along strike in a relay fashion, with the total India-Siberia convergence remaining
approximately constant.
IMTRoDUC,NON
Central Asia consiss of a mosaic of blocks which accreted
since the early Paleozoic. Based mostly on paleontology and
stratigraphy, and on the recognition of tectonic boundaries, a
number of blocks with different geological histories have now
been identified: Siberia, Tarim, Kazakhstan, North China,
South China, Indochina, and the Tibetan blocks. However, the
timing of individual collisions is still debated, sometimes
hotly so. For instance, Li et al. [1982] propose that the North
and South China blocks were accreted in the Triassic, whereas
Laveine et al. ll987l and Mattauer et al. U9851 propose'a
Paleozoic age.Paleomagnetism provides an independent source of data
bearing on this problem. Opdyl<e et aI. [19861 use such data !o
favor a post-Triassic age of North versus South China
collision. McElhinny et al. ll98ll and Li et a/. [1988b]believe that Tarim and North China behaved as independent
I hrtitut de Physique du Globe de Paris and Dépanernent des Sciencesde la ærrc, Universiré Paris Vtr.
2 CeEss. Universiré Rennes I3 C"nt." National de la Recherche Scientifique, Laboratoire de
Paléontologie des Vertébés, Univenité Paris VI.4 Xitr.li"ng Engineering Institute, People's Republic of China.5 Burruo of seismology of Xinjiang, People's Republic of China.
Copyright l99l by the American Geophysical Union.
Paper number 90J802699.0 | 48 -022't t9 | t90JB -0269 9$05.00
units in the Permian, whereas t, [1980] finds no evidence for asuture between the two.
A significant amount of paleomagnetic research has
recently been published or is underway in central and easternAsia. Data have been collected from Tibet [Achache et al',19841, South China [e.g., Kent et al., 1986: Enkin et al.,1991a, â), and farther north from North China le.g.Lin,19841 urd the Tarim [e.g., Li et al., 1988a, à]' This leaves theimportant Dzungar block relatively unexplored. Wiù a surfacearea of 130,000 kmz, this large basin has traditionally beenconsidered a part of the Kazakhstan, wedged between the Tarimand Siberia blocks (Figures 1 and 16). Its boundary with theformer lies along the Tien Shan range and with the latter alongthe Altai range, where ophiolite suites and ophiolite mélangeshave been found [tr et aL,1982;Zhang andWu,1985).
In recent years, research carried out in the Gansu corridorberween Tarim and North China, and in the Tarim and Dzungarbasins has focused primarily on Pgrmian and Triassicformations [Bai et al., 1987; Li, 1988; Li et al., 1989]' Yetthe configuration of the Eurasian mosaic has been profoundlyaltered by the India-Asia collision fMolmr and Tappornier,1975; Tapponnier and Molnar, 1979; Tapponnier et al',19861. The f irst results from the French-Chinesepaleomagnetic work in Tibet suggested some 2000 km ofshortening between Tibet and Siberia lAchache et al.,19841and total shortening could be even larger, based on revisionsof the apparent polar wander path of Eurasia lBesse andCourtillot, 1991] and on the age of the onset of collision
lJaeger et al.,1989).It is therefore important to reconstruct the precollisional
paleogeography of Asia and to determine in a quantitativefashion the amount of subsequent intracontinental
4065
CHEN ET AL: MESOZOIC CONTINENTAL SEDIMENTS
Fig. l. Topographic map of Cenrral Asia showing main basins (larim and Danngar) and ranges (fibet, Pamir, Tien Shan andAltai); contotrrs every 1000 m. The box near Urumqi corrcsponds to the studied area (Figurc 2) [aher Tapporuier and Molnar,1979; l.P. Avouac eI aJ., manuscript in prcparation, l99ll. The inset is a sketch map of the main tectonic unirs (see alsoFiguæ 16).
8 1 .
deformation. This requires collection of Mesozoic andCenozoic formations, which have apparently been lessstudie4 particularly in the Dzungar block. This paper reportsthe first results of a paleomagnetic sampling in the regionwest of Urumqi and north of the foot of the Tien ShanMountains (44.2'N, 86.0"E), along the southern edge of theDzungar basin (Figure 2).
Grolocynro SewuNc
The well-exposed rock sequences generally consist of soft,coarse-grained continentâl sediments. Some harder strata werefound to be better suited for paleomagnetic sampling. Drillcores (445) were taken from 38 sites in rocks dating from themiddle Jurassic to the early Tertiary. Dating the continental
r qu*.,t*y
EII Lower euatemaryE g"66
E siæFig. 2. Schcmaric geological map of sampling arca, including site locations (solid circles). Cross sections AA' and BB' are
shown in Figure 3.
E r*-
* El ;ulassiç
l-'l Ct"t**us @ Triassic and older
EEI Roadarduill"g" ffi F"ult
CHEN ET AL: MESOZOIC CONTINENTAL SEDIMENTS
TABLE 1. Site Data
WanS U9851, this formation is Bathonian to Callovian in
age. Ostracodes reported from this formation, such as
T-imi r ias ev i a c ate n ular i a M and e ls t atn and D arw inula imp ud ic a
Sharapova also suggest a Middle Jurassic age (J. F' Babinot'
personal communication, 1989). From this sequence we
coilected a total of 60 samples in five sites in deep-red suata
along the east side of the Totuenhe River (fable I and Figure
2).
Late Jurassic Qigou (J jq)-Ecnly Cretaceous
H utubi ( K 1 fi Forrnations
These iormations make up a total of more than 1000 m of
sediments. The rocks consist of beds of coarse to medium-
grained sandstones, some of them dark red, others grey-green'
+067
rocks of the Dzungar basin with some accuracy is difficult
because of the scarcity of fossils' Thus' we divided our
paleomagnetic sampling into three age sequences: middle
iurassic.-late Jurassic-early Cretaceous and late Cretaceous-
earlv Tertiary. The sites are distributed along a 120-km-long
streich west of Urumqi (Table I and Figure 2)'
Middle Jurassic Toluenhe Fornntion (J2)
This sequence of dark-red and green-grey medium-grained
,urràrrorr" i, .o." than 500 m thick' It is monoclinal with
n!* u"rti""t dips (even overturned, Figure 3).and contains
i"*lf pf*" tu"it * Ginkgoites sibericus' B.aieria cf' gracilis--Àd
nrroptyllum sp. ftut"ou of Peu^oleum. Geology of
finjiu"g, 'personal
éommunication' 1988)' According to
Fossils
t lt 21 3l 4r s
0 10 203041 92 0) 1' ) )
242 52 8293 03 1a )J J
r 2t 2t 1
1 2t 2
206260264257a / 1
8 68 68 87 81'l
Dark-Red Coarse Medium-Grained Sandstone (Jh-K& (JtKU)
Ginkgoites sibericusBaiera cf. gracilisPterophyllum sp.
RhinocyprisC y p r idea unicos t ata G al.C y p r idea t it a Lj ub imov aO rigoiliocypris cirrita Mad.
gastropod bivalve
l 61 6t 6l 6I
r 29
r 29
1 0
l 1t 21 3r 2r 3
5 45 65 69 89 75 0
7 97 5646 83 0J U
3 21 0 5r00
t 2t z
1 19
r 3r 2
1 0L 2t zl 4t 2
9'l ')
r 2
050 6070 809l 01 6t 7l 82 6273 43 53 63 73 8
284284) R )
274a À a
239l l 89 9
l l 38 9
1 3 02743082 8 12-17281
5 05 05 8437 5t )
7 57 61 À
5 66 62 62 81 82 6a ^
) 1
Pale Red Color Fine-Grained Sandstone (Kæ-El-22-(KuJD)
8 69 38 8
274269270247245241
8 99 0
t2ll l 69 9
2801 1 1
Abbreviarions are n = number of minicores collected from site; s, / = strike and dip (in degrees) of siæ beds (dips greater than
90o denote inverted u"os; tuit" is counterclockwise from downward dip). Formation names arc 121= Totuenhe Fm. (middle
Jurassic), J3o = eigou Fm. (late Jurassic), Klh = Hutubi Fm. (early Crelaceous), K2d = Donggou Fm' (late Cretaceous, and E1-
2z=ZniquaÀ Fm. (earlY TertiarY).
4068
The Qigou Formation has yielded ostracodes suggesring aJurassic age. The Hutubi Formation is reported to containEarly Cretaceous ostracodes such as Cypridea unicostataG aleeva, C. trita Lj ubimova, R hirccyp r is echinata Lj ubimovaal'd Origoiliocypris cirrita Mandelstam, The latter two speciessuggest a Barremian age (J. F. Babinot, personalcommunication, 1989).
Stratigraphically, the Hutubi Formation is parr of rheTugulu Group, which according to Chen [983] spans theentfue early Cretaceous. The Hutubi Formation is overlain bythe Shengjinkou Formation, itself overlain by the LianmuqinFormation. In Wuerhe district, in the norrhwestern parr of theDzungar Basin, the Lianmuqin Formation has yielded a fairlyvaried vertebrate fauna [Dong, 1973] containing, amongothers, the dinosaur Psittacosaurus. Psittacosaurus is known inseveral other localities in Asia, from Siberia and Mongolia tonorthern China and Thailand fBuffetaw et al., 19891, in rockswhich seem to correspond to the later part of the earlyCretaceous (Aptian/Albian). The older Hutubi Formation isthus probably pre-Aptian in age.
The strata outcrop on steeply dipping (>50o) limbs ofanticlines with roughly east-west fold axes (Figures. 2 and 3).Alogether, 211 cores were collected from 17 sites. Twelve ofthe sites are from lhe northern limbs of the folds (Table I andFigure 2).
Late Cretaceous Donggou (KZù - Early TertiaryZiniquan (E f Z) Formatiotts
The Donggou Formation has yielded ostracodes; its lateralequivalent, the Ai l ikehu Formarion, has also yieldedhadrosaurid dinosaurs indicating a late Cretaceous age. Theseformations are referred to the Maastrichtian by Chen [19831and to the Coniacian/Sanronian by Hao er a/. [1986]. In theZiniquan Formation, we have collected gastropods andbivalves of Tertiary aspect (A. Lauriat-Rage, personalcommunication, 1989). This formation has yielded fewfossils. It is referred to the PaleoceneÆarly Eocene by Li[ 1e84] .
This strongly folded sequence crops our farrhesr North ofthe Tien Shan Mountains. The dips of the srrata are generallygreater than 50o and are sometimes overturned. The fold axeshave orientations similar to those of folds to the south.Sampled rocks are fine-grained sandstones with a pale redcolor. A total of 174 cores were sampled from 16 sires, 8 ofwhich are on the northern limbs of the anticlines (Figure 3).
P.rLrotrl,ccNgnc REST:LTS
Minicores were collected with an electrically poweredportable drill with standard 2.S-cmdiameter drill bits and wereoriented with magnetic and Sun compasses. The averagemagnetic declination in the region of study is 4oriy'. In thelaboraory the samples were cut into 2.2-cmlong specimens.
Most magnetic measurements were carried out in themagnetically shielded room of the Paleomagnetic Laboratoryat the Institut de Physique du Clobe de Paris (IPCP), wirh a
CHEN ET AL: MESOZOIC CONTINENTAL SEDIMENTS
Fig. 3. Schemaric N-S cross-sections in eastem pan of sampling area (see Figure 2 for location).
I
?,
/\t ll ,
three-component cryogenic magnetometer. The remainder ofmeasurements were done at the Paleomagnetic Laboratory ofthe Université de Rennes (UR), using a single axis cryogenicmagnetometer and a Schonstedt spinner f lux g atemagnetometer. A total of 282 specimens were measured. The
- intensities of narural remanent magnetization (NRM) rangedfrom 1 to 200 mA/m, averaging 2 mA/m. Most specimenswere thermally demagnetized, in the laboratory-built fumace atthe IPGP and in a Schonstedt furnace at the UR, and severalspecimens were demagnetized by alternaring field (af).Thermal demagnetization was found, in general, to be moreefficient than af in sgparating magnetic components. Sampleswere usually demagnerized over 10 to 16 steps, sampleorientation in the furnace being inverted at each step to detectâny systematic magnerization resulting from an ambientmagnetic field in the furnaces. Demagnetization results wereplotted as orthogonal vector diagrams and as equal-areaprojections. After each temperature step, the bulksusceptibi l i ty of each specimen was measured. Curietemperature (CT) analyses in air and isothermal remanentmagnetization (RM) experiments were carried out on about 20specimens.
Paleomagnetic directions and planes were determined using
principal component analysis (PCA; Klrsclvint [1980]), andsite means using McFadden's method [McFadden andMcElhinny, 19881. Twenty-eight out of 38 sites providedconsistent results, which are described sequentially below.
Middle Jurassic Totuenhe Formation (J2s)Both thermal and af demagnetizarion methods were used,
but thermal demagnetization was much more effective for allfive sites (Figures 4a versus b). In general, vector diagramsfrom both types of demagnetization demonsEate the presenceof two magnetic components with closely aligned directions.The low temperature component (LTC) was cleaned by about20OoC. The Fisher average for this component, based on 20specimens from the five sites, is Dg=t.Jo, Ig=66.9o (k=54.9
and g-95=4.2o) in geographic coordinates, i.e., close ro rhedirection of the present Earth's field (PEF) (D=0", I=62').
In specimens from site 14, the high temperarure component(HTC) is unbiocked by about 575oC for rhe specimens of sire14 (Figure 5b); measurements of IRM and Tc on rhis site(Figures 4c ud 4d) suggest thar rtre remanence is carried bymagnetites. For the other sites, the HTC has an unblockingtemperature of 620"C suggesting that hematite is acontributing carrier (Figures 4a and 5a). Measurements of low-field susceptibility do nor indicare major changes of magnericmineralogy during thermal demagnetizarion (Figure 4e).Twenty samples from five sites yielded HTC directions ratherclose to the PEF in geographic coordinates, rhough with asiightly steeper inclination: Dg=6.6o, Ig=72.6o (t=107.9 andrl95=7.4o). This direction becomes very shallow and mostlyupward in stratigraphic coordinates: Ds= 353.9o, 1s=-10.1o(k=50.7 and og5=10.8o; see Figures 6a ud 6b and Table 2).
Because this formation is almost monocl inal. no
Mmax = 16.2 mA/m
Tcmperelurc ('C)
CHEN gT AL: MESOZOIC CONTINENTAL SEDIMENTS 4069
1 n
xoE
1 . 0
X(5E
60
=a
È Â ôc4a
=!
c
! z oI
t
I
I
4
Altcrnating field (mT)
b
1 00 200 300 400 500 600 700
T€mperrture (oq
d Pigure 4
Mrgnctizing l ield (mT)
c
Fig. 4. Results of magneric study for middle Jurassic (J2) samples. (a, à) Normaiized NRM intensity cuwes showing an
unblocking temp€raftre of about 650'C during thermal demagnetization in Figure 4a, and a rarher high coercivity during af
demagnetization in Figure 4à. Mmax is rhe maximum magnetization measured during demagnetization. (c) Acquisition of
IRM indicating predorninant magnerite for site 14. (d) Thermomagnelic curve in air showing a decrease of magneric
intensity ar about 580oC, also indicative of magnerite for the same site. The applied.magnetic field is about 0.5 T. (e)
Suscepribility curye demonstrating absence of magnetic mineral change during thermal demagnetization.
Mmax = 13.5 mA/m
Tempcreturc ('C)
4070 CHEN ET AL: MESOZOIC CONTINENTAL SEDIMËNTS
Geographic Coordinates Scale = 2 mA/m Scale = 5 mA/m
Fig. 6. Equal-arca projection of high tcmpcrature cornponent (HTC) directions of all middlc Jurassic (J2f specimens (a)beforc and (à) afær bcdding correction. Note that most dtections are steeper than present Eanh field @EF, shown by a star)in gcographic coordinatcs and very shallow in stratigraphic coordinates. Closed (open) symbols are for directions in lower(uppcr) hernispherc.
E U Pa o
Fig. 5. Orthogonal vector projecrion of representative thermal demagnetizarions from middle lurassic (J21) samplesdemonstraring (l) different unblocking temperatures for differcnt siæs; (2) sæeper dirccrions rhan presenr Earth field @EF)in (a) geographic coordinates and (à) horizontal dircctions in stratigraphic coordinates. Solid symbols are in horizpntaiplane, and open symbols in NS vertical plane. Numben adjacent to data points indicate rempentur€s (degrces Celsius).
1 1 -1368
14-1738'
+
+
+
+
*'ill.. T q
+
+
+
+
+
+
+
+
Geognphic Coordlnrtes
+ + + + + + + + + + + + + + + + +
Stntlgrephlc Cocdlnrtes
CHEN ET AL: MESOZOIC CON'NNENTAL SEDIMENTS
TABLE 2. Paleomagnetic Directioru for J6 Sites
Site "/ru Dg Ig Ds /r k c95
l I 414 16 .8 69 .3 3 5 9 . 0 - 4 1 . 5I L )
t4 .4
12 414 346 . 75 .4 348 .45
1 3 4 1 4 2 3 . 6 7 0 . 3 4 . 0
1 4 4 1 4 3 3 8 . 6 8 . 1 3 4 4 . r4
107 I
- 28.4 17.61 0 . 6- 36 .3 15 .5r 5 . 6- 6 6 . 1 1 1 . 41 0 . 1
L5 414 28 .8 74 .2 354.3 ? .8 3 ! .8 16 .5Me"n nlN Dg Ig Ds Is & ç195
7 . 41 0 . 8
55
6 . 6 7 2 . 6 1 0 7 . 9353.9 - 50 .7
1 0 . 1
The n/N is number of entries in the statistics/number of
demagnetized samples;
D, I, k,s95 is declination, inclination and Fisher [1953]stat ist ics of si te-site data. Subscripts g and s stand for
geographic and stratigraphic coordinates, respectively.
significant fold test could be performed' although the best
esiimate (k) of the precision parameter decreases by a factor of
2 upon tilt correction (this is not significant at the 5Vo leve|
for N=5 sites). Also, it is very unlikely that these beds have
acquired their magnetization before folding, since this should
imply a subequatorial paleolatitude of the area by the middle
Jurassic. On the other hand, if the characteristic component
corresponds to a remagnetization, it is clear from Figure 6 and
Table 2 that its direction is distinct from that of the PEF' but
with an inclination intermediate between PEF (1=62') andbedding plane dip (85"N). Two possible causes could accountfor this observation. Either (1) there is a discreteremagnetization event before the end of folding of the units.In this case, about l07o of partial unfolding would restore HTCdirections parallel to PEF; or (2) there is a deviating effect dueto a strong anisotropy of magnetic suscePtibility (AMS). Afew measurements of the AMS of these samples were made on aDigico apparatus le.g., Col l inson, 19831 at RennesUniversiry (e.g., see Cogné ll988al for the procedure)' Theyshow that the AMS ellipsoid is oblate with minimumsusceptibility normal to bedding and that the ratios ofmaximum to minimum susceptibility are rather high,averaging about I .15. A deviat ing effect of AMS onmagnetization aquisition is thus conceivable, as has alreadybeen documented in other examples [Cogné, 1988à]. \flith thedata presently available, it is nor yet possible to concludewhich explanation is correct. However, a recentremagnetization event leading to a PEF direction and deviatedby the anisotropy is probably a good hypothesis. Asmentioned above, the rocks making up this formation arerather coarse and soft, so that the remagnetization waspossibly caused by chemical/fluid reactions.
Late Jurassic Qigou (J jq) - Êaily CrctaceousHutubi (K1) Formations
It is clear that thermal demagnetization is much moreefficient than af for most specimens (Figures 7a and 7b).Specimens from this formation exhibit two components ofmagnetization. The low temperature component (LTC)unblocks by 200" or 300"C. The mean direction of this
component is D8=2.8o, Ig=62.0o (k= 23.6' 6195=2.1o and
n=200) in geographic coordinates, and is not distinguishablefrom the PEF direction. The t value is 10.7 times larger ingeographic than in strat igraphic coordinates, clearly
indicating a recent overprint. Because the Curie balanceexperiments were carried out in air, the breakdown of the lowtemperature part of the thermomagnetic curve in Figure 7d is
partly due to weight loss. However, the weight loss wasmeasured after the experiments and appeared to besuperimposed on a decay of magnetization due either to
reaching the Curie point or to the breakdown of a magneticphase. This suggests the presence of hydroxides such asgoethites, and/or sulfides such as pyrrhotites.
As far as the high temperature component (HTC) isconcerned, we can divide demagnetization chatacteristics intothree types of behavior. The first type shows magnetizationswhich cannot be separated from the LTC, close to the PEF(Figure 8a). The second type reveals two NRM components; a
LTC with unblocking temperature < 300oC and a HTCcharacteristic component with both normal and reversedpolari t ies (Figures 8b and 8c). In the third type,demagnetization could not be completed, and the endpointsshow only a tendency toward reversed polarity (Figure 8d).Comparison of nearby samples with Wpe 2 behavior suggests
that the IITC is the same in both tyPes 2 and 3. Hence there is
oniy one characteristic component in these samples. The
intersection of great circles, given by tJlre McFadden and
McElhinny [988] combined line and plane data analysis
method, was used to determine the HTC of type 3 samples,
which could then be used in the computation of site mean
directions.The FITCs have unblocking temPeratures ranging from 300o
to slightly below 600'C (Figure 7a), and IRM acquisjtion
curveJ shôw thatg}Vo saturation is reached in fîelds of 150 mT
for most specimens (Figure 7c). This evidence indicates that
the main magnetic carrier is magnetite' However, in some
samples, the high temPerature Part of thermomagnetic curves
reveals Curie poins above 600oC (Figure 7d). This suggests
that hematite is sometimes present (e.g., Figures 8â to 8d) but
does not carry a recoverable comPonent of magnetization. No
change of bulk susceptibility is found during heating (Figure
7 e) .
Mmax = 22.2mNm
4072
From the 17 sites of this formation. four sites werediscarded (sites 3, 22, 23 and 24) because of totalremagnetization in the PEF direction, or erratic behavior athigh demagnetization temperafures (Figure 8a). Results fromthe remaining 13 sites are listed in Table 3. The method of
x(!E
CHEN ET AL: MESOZOIC CONTINENTAL SEDIMENTS
200 300 400Mrgnctlzin3 fldd (n'I)
e
Fig. 7. Results of magnetic study for upper lurassicJowcr Cretaceous (J3o-K1) samples. (a, â) Normalized demagnerizationintensity cuwes (as in figurcs 4a and4b, but with unblocking temperâture at about 580'C and lower coercivity); (c)Acquisition of IRM showing that prominent magnetic carrier is magnetite. (d) Thermcnagnetic curve in air. Noæ that Curiepoints of goethite and magnetite arc rather clearly discemed at about l00o and 580oC, togethcr wirh some hematite withCurie point above 600oC. The applicd magneric field is about 0.6T. (e) Susceptibility curve for a rcpresentative specimenshowing rhe lack of magnetic mineral change during thermal demagnerizarion.
1 . 0
x
E
t
q
È
v)I0
ê s 0ila
tâ
600
McFadden and Lowes [1981] was used to check whether thereversed and normal polarities belong to one direction group.The statistic p=(R 1+R --R2 l(R ++R -))12(N-R+-R -) (where R, R*and R- are the lengths of the vector sums for the total numberof samples (N), normal-polarity samples only (N1) and
Mmax = 8.60 mÉr'm
3u
a
!a
I
30
20
1 0
200 400Tcrnpcreturc ("C)
.r073
Gcograph ic Coordlnater
Strâtlgraphlc Coordlnstes
CHEN ET AL: MESOZOIC CONTINENTAL SEDIMENTS
Scalc - 2mA/n
E U Po o
Fig. 8. Orthogonal veclor Projections- of thermal demagnedzation
diJolavins, differcnt types of magnetic behavior from upper furassic.-
ioi"i'Ci"ï""."s (J3o:k1) samplàs' (a) PEF onlv; (â' c) wèll-separaæd
comDonents and dual'polarities; (d) rendency toward reversed polarity-'
;;;';iil ;;;- direitions (not sirown) above 575oC; (e) an example
from reiected siæ 22 showing that magnetic intensity decreases rapidly
"il."i1*p"-*rc and HTCIs difficuit to isolate, although a revened-
poi"f,y cômponent is discemable. Symbols and conventions as in
Figure 5.
b
N No o
Scdc - 5 mA/m Stratlgraphic Cærdinates
Scale = I mA/rn
4074 CHEN ET AL: MESOZOIC CONTINENTAL SEDIMENTS
TABLE 3. Paleomagnetic Directions for JuKl sites
0 102041 92 02 r252 8293 0J I
3 23 3
7lg6 l t l7 l r08/88/98/87ls8t96 l t l8 /1 l7 t 88/8
7 7 . 67 6 . 473 .3-42.0- 3 6 . 6-34.420 . I-22.679 .36 6 . I7 7 . 76 6 . I6 8 . 9
1 . 3 l 5 . l9 . 949.95 r . 11 0 8 . 85 0 . 836.7r8.29 7 . 9l 1 8 . 6t < r
57.534 .8
22.48 . 67 . 85 . 38 . 61 0 . 1l 3 . 39 . 35 . 11 8 . 78 . 0
1 7 1 . 833.723.23 1 0 . 1305.7313.2t96.7208.5343.63 4 8 . 15 0 . 83 6 .9
2 r . 51 9 . 51 0 . 31 8 1 . 8196.7r 8 0 . r202.0204.r3 5 5 . 38 . 024.32 3 . 37 . 4
42.93 0 . 5- 5 7 . r-59 .6-59.3-42.0-47.9Â a 1
5 r . 653.743.842 .0J .
Mean n lN Ds Ig Ds /"I
cl95 kstkg f(Rs) f(Rs) f(Rc)
Normal 18
5 5 . 8
t4.3
69.2
7 6 . riz.tI 1 . 9
iso. t
48.6
45.5'-53.6
4 . 257.63 7 . 762.3n a6 1 . 9
7.30 0 . 5 25 . 5 13.7 0.01
Reversed 5 àos. r - :o.s9 . 17 . 169.39 . 8
Paleolatitude = 26.6o t 4.8o. Pole: latinrde = 72.3o,longitude = 227.3o, A95 = 4.B.The nlN, D, I, h a95 are same as in Table2' ThefiRg)
Td-l(nrl are computed asflR) = (Rs+Rru - R72\(R5+Rù)/2 (N-Rlu-Rs), where Rr, xs and rRly are the lengrh of the
vector suns of the site mean directions of all sites (R7), sites from the southem limbs (R5) and from the northern [r"U, 1ntr),with l/ = NS sVN;,f(Rc) is the critical values zt the 95vo probability level; if flR)>/(Rc), thî hypothesis of a common true meandirection may be rejecæd [McFadden and Jorcs, l98ll.
reversed-polariry samples (N_), with il=N++rV_, respectively)gives 0.028 which is much smaller than rhe critical values arrhe 95Vo probabiliry level, 0.28 (R_ = 4.935, R+ = 6.917. R -11.82, N=13). This indicates ùar rhe rwo poladry groups arenot significantly different from antipodal.
Only sites 25 and 28 are on the south limb of the fold.However, the stratigraphic correction produces an increase of& from 4.2 to 57.6. The fold test is significanr ar the 99goconfiderrce level using both the McElhinny [1964] and the{cFadden and. Jones [1981] fold rests (Figures 9à nd 9b,Table 3). The mean paleomagnetic dircction for this formarionis thus Ds=12.7o, Is-48.6o (k= 57.6 and o,95=J.5o, i1stratigraphic coordinates, Table 3).
Late Cretaceous Do.nggou (KZû - EoceneZiniquan ( E t -Z) F orrnaions
About half of the samples showed only the pEF or unstableor random directions. The rest (45 out of 102 samples),corresponding to 9 out of 16 sites, yielded two recognizablecomponents. Analysis of IRM and thermomagnetic curves, aswell as thermal demagnetizations indicaæ that-LTCs and IITCscorrespond to distinct types of magnetic carriers. For LTCs,the unblocking temperature ranges from l00o to 300oC(Figures lla ro llc). The low temperarure part of thethermomagnetic curves (Figure l0e) suggests, as in theprevious. case, the presence of goethites and/or pyrrhotites.The statistics of this componeni shows a pEF direction asabove. For the HTCs, the unblocking temperarures andmagnetic saturations reveal two carriers. One, with a Curietemperature of about 650oC (Figure lOe), an unblockingtemp€rature of 650oC (Figure lOa) and high coercivity (FigurelOc) is probably hematite. The other, which reachei 90% ofmagnetic saluration at 150 mT (Fig. lOd) and has an
unblocking temperarwe of about 580.C (Figures l0ô and lla)is probably magnerire.
The behavior of the HTCs is similar to that of the LateJ,urallilEarly Cretaceous formations. The sites showing onlythe PEF direction were eliminated from further analysis. Sitj-mean directions were computed using the combined line andgfqne a1t1 analysis method of McFadden and McElhinny[1988]. No major change of bulk susceptibility was founâduring heating (Figure l0/). Sire mean and formation-meandirections are listed in Table 4.
Two of nine sites showed reversed polarities. Comparing!h9_tw_o groups of polariries, rhe reversal test staristiè p is0.17 (R+= 6.900, N+=7, R-= 1.980, N_= 2, R=8.860, N= 9).This is smaller than the crirical value at the 95Ço confidencelevel (0.534), which meens rhat rhe normal-polarity andreversed-polarity groups are not significantly different fromantipodal.
Figures l2a md 12à show siæ means wittr o95 confidencelimits on equal-area projecrions before and after stratigraphiccorrections. The sites display a large range of structuralattitudes. After structural correction, t increases ltom 2.7 n56.9. The fold test is positive at the 99go confïdence level forboth the McElhirmy and rhe McFadden and Jones fold tests(Table 4). The formarion mean direction is D.rl2.5o. /.e51.3o(ê.56.9 and 495=6.9o, in stratigraphic coordinares, Table 4).
Dscusstox .cND Coxeusrcx
A first sampling rrip !o the Northwestern foot of the TienShan, West of Urumqi (44.2"N, 86.0.E), along the sourhernedge of the Dzungar basin, has provided paleomagnetic resultsfrom three age groups. Because age determinations on thesecontinental sediments are unfortunately rather coarse, the
CHEN ET AL: MESOZOrc CONNNEIqIAL SEDIMENTS
Str.tlgraphlc Cmrdlnrtes
1075
Gcogrephlc Coordlnrtcs
Fig. 9. Equal-area projecuon of HTC site mean direcdons from upper Jurassic-lower.cretâceous (J:o-Ktt) sites' (a) before
and (â) after bcdding corrcction, wirh circles of 957o confiden"' à"*on'toting posidve fold æst' SyinUots and convenlions
as in Figurc 6.
three age Sroups cannor be resolved more finely than middle
i**tii, lite J^urassic-early Cretaceous, and laæ Cretaceous-
early TertiarY.-lûiaOt" Jurassic sites appear to have been completely
,"t*g""tir"a in a direction "lot" to the PEF' after tectonic
deforiration. It seems Particularly difficult to ottain primary
oni"o-"stt"tic directions from middle Jurassic formations'f#ôd;;
as has been found in South China by Enkin et al'
ilS-trbf and M. Steiner et al' (personal communication'
\*iej.'ciirillot ard Eesse [1986] als-o sgss3.sfed that lower
i*r"ti" resule from both North and South China quoted by
Lin et al. U9851 might be later remagnetizations'- fne- t"tà Jruassic--early Cretaceous and the late Cretaceous-
early Tertiary age groups yield characteristic- paleomagnetic
à""i,r"* *fii"f .Ë likely primary with dual polarities and
l9i, potitit" fold tests. gien in thlse formations' about 507o
of the samples were remagnetized in the PEF.or showed random
àl*ir*.'rrt" Ju-Kl andKu-Tl poles, based on 13 and 9 siæs
ôt,h t4 and 45 specimens) respécdvely, are listed in Tables 3
and 4 and are shown in Figure 13'It is clear that the t'"o-pol"t (triangle and diamond symbols
i" -n;;;
i:l which lie o'ltttitt ttt" jàint intersection of their
a 9 5 confidence intervals, ate not statistically
diJtiirguishable; indeed, their angular distance is 2'3o* 8'0o'g;"u:t" the magnetizations yielding these two poles predate
tectonic deformation, because they have recorded dual
p"ioltl"t and are carried by distinct minerals' they are most
iit"iy a be primary' and these two poles can b-e. assigned ages
ào"f ,o thË ages-of the formations from which they were
obtained.--en impficit assumption in the following discussion is that
the recovlered paleomagnetic directions -are parallel to the
;;;;" axiat aipote Geo) pa*neH directi,on' Indeed' we
irust address the question of whether these directions have
suffered from compaction-induced inclination shallowing
before the paleomagnetic inclinations can be interpreted in
terms of paleolatitudes. This problem has. recently been
analvzed for natural sediments fArason atd Levi' 1990a1'
i"â"'pltirca sediments lLevi and Banerjee, 19901' synthetic
sediàents fDeamer and Kodama, 19901 or by numerical
modeling -lArason
and Levi, l990bl' It appears that
"à*f".ti6" àf magnetite-bearing clay-rich deep'sea sediments
may produce a significant inclination shallowing, as high as
iS;, "sp"cintly in tire range 450-60o of initial inclination'
Howeuer, this phenomenon' which is na! systematic in deep-
sea sediments, is believed to be insignificanç if not absent' in
ouartz-rich sandstones lDeamei and Kodama' 1990:
3;;;;ï;; et at., 19891. Because of the nature of the
sediments in our study (coarse-grain continental sandstones)'
we believe that paleomagnetic inclinations parallel that of the
GAD field within the range of experimental error'- ih" D"uttgar Basin was aPParently attached to the
sunounAing biocks (Kazakhstan, Siberia and Tarim) by the
"ô"i i*Ëtic, i.e., 'the
time for which we have ou oldest
oaieomasnetic constraints, with no major tectonic
âirof"."al*o mking place since. We nexl comPare our results
with those available for other blocks'
Dz ungaila ver s us S iberia--pi*-"gnedc results from Siberia have been obtained by
Soviet authàrs. However, as has been discussed by Wesplul .et"i. tfSAOt and Besse and Courtillor t19911, the data reliability
i, aifnc"it to evaluate. On the other hand, we cân follow the
urruÀption of Besse and Courtillot that for the periods of
i"a"t* here, Siberia was rigidly attached lo-Eurasia' We can
il;;;f"t; deduce the paleô-pôsitions of Siberia from the
"pp-""l p"f- wander p"rft (epWp) of Europe' where data from
{di"" dai" and Nortir America have been included' based on
;;;;;; t;"onstructions lBesse and Courtillat' 1988'
isgil.iltit APWP is also shown in Figure 13' going back to
200 Ma at 10-Ma intervals.-- fii"-a"o of Figrue 13 are plotted in another way in Figure
+ + + + + + +
+
+ + f + + + + + + +
+
+
Mmax = 1.75 mA"/m
TarçanonfC)
4076 CHEN ÊT ÀL: MESOZOIC CONTINENTAL SEDIMENTS
0 100 200 300 400 500 600 700 0 200 400 600Tcrnpcnhæ (oC) Tctlpcnlùn(oC)
e f
Fig. 10. Rcsults of magnetic study for upper Crcteceous-lower Tertiary (KZA-E1-2ù samples. Normalized denragnerizationintcnsity curves prcscnting clear unblocking tcmpcraures at about (a) 650" and (â) 550"C ; IRM acquisitior ctrves for rwotypcr of sernples, with (c) high cocrcivity minerrls (gocthite and hcmatiæ) rnd (d) a low cocrcivity mineral (magneriæ); (a)thcrmcnegnctic curvc in air, dcmonsrnring thc existcncc of gocrhiæ (about 100"C), hematitc (about 650"C) and a fractionof magnetitc (580'C), the applicd mrgnaic ficld ic about 0.6T; (/) susceptibility curve showing stablc susceptibility duringthcrrnel dcmrgnctization.
E
.I
I
aI
I
I
aa
a
a
a:Ia{a!
!
a!I
aIé
!
Àaq
Mmax = 45.1 mAlm
2
I
200 400Tc|lpcnlùn (oC)
CHEN ET AL: MESOZOIC CoNTINENTAL SEDIMENTS 4077
17-2A18
18-2198
Shrtigrrphlc Coordlnrter Scale = l0 mA/m
Stratigrephic Coordlnrtec
14, where the paleolatitudes and declinations obtained for theDanngar Basin are compared with those deduced from the Besseand Courtillot APWP for a reference site at 44.20N, 86.0oE. Itis clear from Figure 14â that at both periods the declinationsare virtually identical to those predicted for Eurasia. There is aslightly larger, more systematic difference in terms ofpaleolatitudes (Figure l4c). This difference is smallest, andapproximately constant, if the recorded ages of themagnetizations are older than 70 Ma and younger than 130Ma. Although the ages based on paleontology are allowed tobe some l0 Ma younger and older than these values,respectively, this would imply increased paleolatitudedifferences for which there is no geological evidence:significant convergence and a suturing event in the upperJurassic (?) on one hand, and divergence and possibly basin
Scalc = l0 mA.lrn
Scale=.5 mA/rn
opening in the Eocene, i.e., at the time of onset of the India-Asia collision. We assume that samples with ages older than130 Ma and younger than 70 Ma rep,resent only a small part ofour collection and have therefore little influence on the means.The J/K and IVT poles are therefore assigned ages of 130-ll0Ma and 90-70 M4 respectively.
We see in Figure 13 that the IÇT pole of the Damgar blockis concordant with 70-90 Ma poles of the Besse and CourtillotAPWP, with a large joint intersection of confidence intenralsand an angular distance of 6.4ot6.6o. In the 90-70 Ma window,the paleolatitude and declination differences between Eurasiaand Dzungaria af,e not significantly different from zero (5.5ot6.6o, and 3.4o t7.2", respectively, Figures 14a and 14â). TheJ/K pole of the Dzungar block is also close to the 130-110 Mapoles of the Besse and Cowtillot APWP (Figure 13), although
c
Fig. ll. Orthogonal vcctor projcction of thcrmrl dcmagnctization from uppcr Crctaccous-lower Tertiary (KU-E1-Zù
samplcr, dirplaying differcnt typcs of magnctic bchavior. (a,à) Well-scparated dircctiqrs with dual polarity; (c)represcntetivi of some spccimenr ihowing otrly a tendency toward r final dircction. Symbols and convcntions as in Figure5 .
4078 CHEN ET AL: MESOZOIC CONTINENTAL SEDIMENTS
TABLE 4. Paleomagnetic Directions for Ku/Tl Sites
Site nlN Dg DsIg /s k og5
051.0l 6r 7l 82 63 43 53 6
st97 t 96167le8/8) t 1
41104t9, t <
1 8 8 . 7 l . 342.2 85 .3
126.5 49 .1320.8 -29 .1106.2 46.2
2 . 3 - 5 . 628.9 28.62 r . 9 1 7 . 0l r . 4 2 2 . 2
195.6 -5 r .24 . 4 3 7 . 38 . 4 4 9 . 3
t78.2 -64.719.3 37 .05 . 2 5 8 . 3
2 7 . 2 5 8 . 517.8 48 .2r2.9 54.6
19.9 r7 .622.8 r2.9
1 9 1 . 0 4 . 95 1 . 3 8 . 52 8 . 6 I 0 . 5
17 .7 22 .59 . 9 2 6 . 9
Mean nlN Dg IsDsIg cl95 ks lk s f(R s) ÂRs) f(Rc)
All 4t .4 41.9
43.236.6
- - 3 1 . 12 4 6 . 1
r 2 . 5 5 1 . 3
13 .3 49 .3
r s8.o -sà.2
39.46 . 9 2 t . l
49.37 . 9
0 . 1 9 0 . 9 42 . 7
56.9J . J
s 9 . 9
3 . 9 1
Normal 7
Reversed 2
Paleolat inrde =32.0o t 6.4o, Pole: lat i tude =74.3o,longin:de =223.1o, A95 = 6.4o, lægend as inTable 3.
the intersection of confidence intervals is smaller and theangular distance is statistically distinct from zero, at 6.2o t5.1o. In the 130-110 Ma window, rather constant andconsistent paleolatitude and declination differences of5.9"t5.2o and 2.5ot5.8o, respectively, are found (Figure l4).
Dzungaria versus TarimWe next compare our Dzungar results from the northern
piedmont of the Tien Shan with those of Li et al. [988a]
which were obtained on the southem piedmonr of the range, atthe northem edge of the Tarim craton (Figures 1 and 16). Wenote that the upper Cretaceous pole position listed by Li et al.unfortunately contains a misprint, which is carried rhroughtheir paper. Correct values are listed in Table 5 and displayedin Figures 13 and 15. The two poles of Li et al. [1988a] arewithin the joint intersection of their 957o confidenceintervals, which are rather large, and therefore not. statisricallydistinct at this probability level. If we compare rhe Dzungar
+ + + + + + + + + + + + + + + + +
Gcogrephic Cærdlnrtes Stratgrephic Coadlnates
Fig. 12. Equal-aree projection of HTC site mean dircctions from upper Cretaceous-lower Terriary (KZa-Et-Zù sites, (a)bcfore and (à) afær bedding correction, with circles of 957o confidence demonstrating positive fold test. Symbols andconventions as in Figurc. 6.
CHEN ET AL: MESOZOIC CONTINENTAL SEDIMENTS
trP,oN oP,"E 495, deg. LP,oN op'"8
4079
TABLE 5. Paleomagnetic poles of Tarim, Dzungar and Eruasia
64.6
72.3
7 5 . 5
208.4
1 1 1 1
201.6
6 . 4
1 . 8
9 . 6
4 . 8
t . 9
66.3
7 4 .3
7 6 . 4
,r t , , o*
2 2 3 . 1
1 9 9 . r
8 . 7TarimlLi et a1.,19881 (recalculated)
Dzungar(this paper)
EurasiaIBesse and Courtillot, l99ll
* The pole longitude, 0p, of the upper Cretaceous Kuche section is misprinted in the paper oî Li et a/. [1988] as 2I4" both in
their Table 2 and in the-text, but their global average is correct'
and Tarim poles for the JK and IÇT periods respectively, wefind that the Dzungar poles lie barely outside the largeuncetainry intervals of the Tarim poles of conesponding age.The angular distances between the pairs of poles are 10.3o t10.7' fôr J/K and 8.0o 110.80 for K/I (Figure l3). Thistranslates into a Dzungaria versus Tarim paleolatitudedifference of 1.9o t 10.?o and a declination difference of ll.5of ll.0o at J/K time, and corresponding values of 6.7o + 10'8oand 5.2o + 11.0o at I?T time (Figure 15).
T ectonic I nterP r e tationThe first-order result evident in Figures 13 to 15 is the
overall compatibility of the Siberian, Dzungar and Tarimpaleomagnetic poles at J/K and IVT times. This is consistentwith the proposal that both Tarim and Dzungaria were attachedto Kazakhstan and Siberia (and the rest of Eurasia to the northof Cenozoic ranges) prior to upper Jurassic time. The new
I
fr
O
-200 - 5 0
6t
9 0
70
q ô
3 0
t 0
o.tt
ql
O6lÈ
Fig. 13. Apparcnt polar wander path of Eurasia efter Bessc ard
Civtillot tiistl, rna*ed wirh opcn squares connected by a line (ime.
inærvel bctwcen succcrsive poles is l0 Ma). Othcr poles are solid
squarc = J[/1( fiom TailmlLi ct al.,l988al, solid circle =Ku from Tarim
lii ct a1.,1988a1. triangle = Ju/I(l from Dzungar Basin (this study)'àiamond = Ku/Tl from Danngar Basin (this study) and asterisks =
Danngar and Tarirn Basin sampling arcas.
- 2 0 0 - I 5 0 - r 0 0 - 5 0 0
Age (My)
Fig. 14. (a,â) Prcdicted paleolatitudc and paleodeclination of sampling
area calculated frqn Eurasian AP1VP of Eassc ard Courtillot [1991]. (c)
Predicted paleolarirudc from /niag atd lrving [19821 APWP' Squares
are rcsults from this snrdy for Ju-Kl and Ku'Tl inærvals including enor
bars on ages and mean dircctions (stippled areas). Differcnces betwecn
results frqn this study and rcference cuwes are given.
I
h Hg hh '
çl
4080
- r 0 0Age (My)
Fig. 15. Detail of (a) Figures l4a and 14à, including results from theTarim (circles; zfter Li et aJ. [988o]) and the Dzungar (squares; thisstudy) basins.
poles indirectly lend support to the hairpin loop in theEurasian APWP at Jurassic and Cretaceous times [Besse atdCourt i l lot, 1988, 1991; see also Enkin et al. , l99la, b).These poles would have been interpreted to indicate largelatirudinal displacements, had the Irving and Iming Q9821APWP been used. This is clear for instance in Figure 14c,where the larger discrepancies and uncertainties in thepaleolatinrdes are readily apparent. Note that Li et al. [988a,p. 219), when discussing Courtillot and. Besse's [1986]speculations about differences between the East China andSiberia Cretaceous poles, failed to quo[e one of theirsuggestions, namely lhat the Cretaceous poles of Siberiamight be in error. This suggestion has been confirmedsubsequently by Besse ard Courtillot [988, l99l].
Ar a finer level of analysis, we note the followingobservations. The small declination and paleolatitudedifferences between Dzungaria and Siberia, and betweenDzungaria and the Tarim on the other hand, appear to besystematic. These differences can be interpreted in terms ofdeformation after ttre deposition age. If we average the IÇT andJ/I( values and correspondingly reduce the uncertainties, weobtain a mean paleolatitude difference of 5.9o x 3.7o(equivalent to 650 + 410 km of NS shortening) and adeclination difference (or CCW (counterclockwise) rotation) of2.6" t 4.5o of Dzungaria with respect to Siberia. In the case ofDzungaria with respect to the Tarim, we find a meanpaleolatitude difference of 3.0o t 6.9" (330 t 760 km) and adeclination difference (CW (clockwise) rotation of Tarim withrespect o Dzungaria) of 8.6o t 8.7o. The large uncertainties onthe Tarim poles of course adversely affect thesedeterminations. The values which are significantly differentfrom zero at the 95Vo confidence level are the shorteningbetween Dzungaria and Siberia, i.e., in the Altai and SayanTuva ranges, and (at the edge of significance) the rotarionbetween Dzungaria and the Tarim, i.e., in the Tien Shan.
CHEN ET AL: MESOZOIC CONTINENTAL SEDIMENTS
5 0
These values can be undersnod in terms of a simple modelof heterogeneous intracontinental deformation resulting fromthe India-Asia collision (Figure 16 and J. P. Avouac et al.,manuscript in preparation, 1991). It is of interest to examinethis deformation in relation to the NNE penetration of Indiainto Eurasia since the Eocene (Figure 16). AlthoughDzungaria" Kazakhstan and Siberia appear to be parts of rhesame Eurasian kernel. assembled before the end of theMesozoic, Dzungaria and Siberia are separated by a major zoneof Cenozoic shortening, the Altai and Sayan Tuva ranges. Thiszone is approximately 700 km wide and 2000 m high on theaveragei consistent with a S0-km-thick crust, based onisostatic compensation. If this crust had an original thicknessof about 35 krn, the present thickness implies some 300 km ofshortening. This value is likely to be an underesrimare, sincewe have not taker into account additional components ofshortening related to (l) the fact that the continental crust inthis part of Asia might have been initially thinner, (2) theamount of material removed by erosion, or (3) displaced bylateral extrusion. A total value of about 400 km is possibleand compatible with the paleomagnetic estimate of 650 t 410km. The Dzungar Basin itself can be approximated by alosange or domino-shaped block, bounded by the overlappingAltai and Tien Shan ranges to the Norrh and Southrespectively, and by NW to NNW striking right lareral faulrs(Figure 1). NNE directed shortening should lead to CCW
Fig. 16. Schematic topographic map showing major blocks andmountain belts in central Asia, with cross section AB from Tarim toSiberia. Thin great circles crudely ourline shape of Tien Shan (J.P.Avouac et al., manuscripr in preparation, l9l) and dashed grcat circleindicaæs CW rotation of northem edge of Tarim. Arrows show thepalcodeclinations of the different blocks. SIB = Siberia, KAZ =Kazakhstan, DZUN = Dzungar, TAR = Tarim, TIB = Tibet, IND = India,STS = southem foot of Tien Shan, NTS = nonhem foot of Tien Shan, A= southcm foot of Altai, S = nonhem foot of Sayan. Numben on croJJsection AB are in kilometers (topography from Simkin et al., |9891(Mercator projection)).
l 0
o
g
(l
d
r 0
2 0
0
6t
q,
Flg|,. ta
CHEN E'f AL: MESOZOIC CONTINEI'I-IAL SEDIMENTS 108 I
rotation of the basin le.g., Cobbold and Davy, 19881' Despiteits large uncertainty of 4.5o, the mean rotation which we find(2.6') is in the conect sense and would conespond to a smallarnount of additional shortening (20 km).
The Dzungar Basin and the Tarim are separated by the TienShan range. This range is very narrow east of Hami, becomeswider to the west and reaches a maximum width of about 400km north of Kashgar (i.e., some 2000 km to the west),suggesting that the amount of NS convergence increaseswestward along the range. The wedge shape of the zone of highrelief of the range, as defined by its 2000 m altirude contour,may be approximated by two great circles intersecting at apivot point P some 300 + 200 km to ttre East of Urumqi(Figure 16). The mean elevation (about 3000 m) at thelongitude of Kashgar implies a crustal thickness of 60 km andsome 280 km of shortening with the same simple calculationused above for the Altai. This corresponds to a CW rotation ofTarim with respec! to Dzungaria of some 9o about P (the
distance between P and Kashgar is about 1800 km, hence 280km/1800 km = 0.16 rad = 9o, with a combined uncertainry ofabout 4o). The shortening between our sites in Dzungaria andthose of Li et al. [l988a] in the Tarim is inferred to be of theorder of 150-175 km (Figures 15 and 16). Despite largeuncertainties which we have stressed above, the orders ofmagnitude are compatible with the paleomagnetic estimates(9o t 4o to be compared with 8.6o t 8.7o, and 150-175 km t507o with 330 f 760 km).
Paleomagnetic constraints now available indirectly forSiberia fBesse and Courtillot, 19881 and directly fromDzunguia (this paper) and the Tarim [Lr et al., 1988a1 are thusconsistent with the simple model of heterogeneousintracontinental deformation due to the India-Asia collisionsummarized in Figure 16. The rather large shortening in theAltai range, slight CCW rotation of Dzungaria with respect toSiberia range, increasing shortening in the Tien Shan fromEast to West and =10o CW rotation of is southern piedmontwith respect to its northern piedmont, observed or compariblewith the paleomagnetic data, are predicæd by this model. Notethat, as often seems to be the case in paleomagnetism, meanvalues appear to make more sense than their uncertaintiesmight suggest, leading to the inference that theseuncertainties might be overestimated.
The tectonic model embodies rhe idea that the total amountof shortening between India and Siberia, which isapproximately constant in the EW or ESE-WNW direction
between 70o and llOoE longitude, is spent in a nonuniformway along the suike of the "overlapping" Tibetan Plateau,Tien Shan and Altai ranges, with recent strain diminishingtoward the west in the Altai and Tibet and toward the east in the
Tien Shan. The Indian indenter is closest ro the Tarim at thelongitude of Kashgar. The width of Tibet is minimum there and
the northwest tip of India and the Pamirs directly abut the
deforming northem boundary of the Tarim, whereas morédeformation is absorbed within Tibet to the east. This impliesCrff rotations of the Tarim and CCW rotalion of Dzungaria,which are trapped between the overlapping convergencezones. This means that the Tarim poles of Li et al. [l988al andthose given here for Dzungaria cannot be considered to berigidly fixed to either ttre Siberia block or the North Chinablock. Further work is required to reduce the uncertainties inthe Tarim data, to extend the geographical data base for theTarim and Dzungar blocks and to confirm the preliminaryresults and interpretation outlined here. Direct Siberian dataare also badly needed. Field triPs to Siberia, northernDzungaria and southern Tarim are therefore plarured. Untilmore data become available, the robust conclusion from thissftdy and that of li et al. $988a1 is the first-order agreementof Dzungar and Tarim paleomagnetic poles with the Besse atdCourtillot [1988, 1991] APWP of Eurasia, and the rough
agreement of second-order differences with an independently
derived tectonic model of heterogeneous deformation in
overlapping zones of convergence.
Acknowledgment.r. R. Butler, N. Opdyke and a third anonymousreviewer made valuable comments on the manuscript. We are alsothankful for the suggestions made by R. Enkin. This research has beensupported by the Institur National des Sciences de I'Univers andXinjiang Engineering lnstitute. This is IPGP contribution 1147.
REFERENCES
Achache. J., V. Counillot, and Y. X. Zhou, Paleogeographic andtectonic evolution of South Tibet since middle Cretaceous time:New paleomagnetic data and synthesis, J. Geophys. Res., 89,l 0 3 l l - 1 0 3 3 9 , 1 9 8 4 .
Arason. P., and S. Levi, Compaction and inclination shallowing indeep-sea sediments from rhe Pacific Ocean, J. Geophys. Res.,95,4501-4510, 1990a.
Arason, P., and S. Levi, Models of inclination shallowing duringsediment compaction, J. Geophys. Res., 95, 44814500, 1990ô.
Bai, Y. H., G. L Cheng, Q. G. Sun. Y. H. Sun, U Y. H., Y. J. Dcng andD. J. Sun, Late Paleozoic polar wander path for the Tarim platformand its tectonic significance, Tectonophysics, /J9, 145-153,1 9 8 7 .
Besse, J., and V. Courtillot, Paleogeographic maps of the continentsbordering rhe lndian Ocean since the Early furassic, l. Geophys.Res . , 93 , 11 ,791 -11 ,80E , 1988 .
Besse, J., and V. Counillot, Revised and synthetic apparant polarwander palhs of the African, Eurasian, Nonh American and IndianPlates. and true polar wander since 200 Ma, J. Geophys. Res., inpress, 1991.
Buffetaut, E., N. Sartayarak, and V. Suteethom, A psittacosauriddinosaur from rhe Cretaceous of Thailand and its implications forthe palaeogeographical history of Asia, Terra Nova,l,370-373,I 989 .
Chen, P. J,, A survey of the non-marine Cretaceous in China,Cretaceous Res., 4, 123-143, 1983.
Cobbold, P. R. and P. Davy, Indentation tectonics in nature andexperimen!, 2, Central Asia, Bull. Geol. Inst. Univ. Uppsala, 14,1 4 3 - 1 6 2 , 1 9 8 8 .
Cogné, J. P., Strain, magnetic fabric, and paleomagnetism of thedeformed redbeds of the Pont-Réan formation, Brittany, France, J.Geophys. Res. , 93, 13,673-13,687, 19884.
Cogné, J. P.. Strain-induced AMS in the granite of Ramanville and itseffects upon TRM acquisition, G eop hys. 1., 92, 445-453, 1988b.
Collinson, D. W ., Methods in Rock Magnelism and Paleonagnetism,503 pp., Chapman and Hall-London, 1983.
Courtillot, V., and J. Besse, Mesozoic and Cenozoic evolution of theNonh and South China blocks, lfalzrc, 320, 86-87 , 1986.
Deamer, G. A., and K. P. Kodama, Compaction-induced inclinadonshallowing in synthetic and natural clay-rich sediments, "/.Geophys. Res., 95, 4511-4530, 1990.
Dong, Z. M., Dinosaurs from Wuerho, Acod. Sin. Mem. Vert.Palaeontol. atd Palaeoanthropol., (ia Chinese), I I,l-7, lyl3 .
Enkin, R. J., and V. Counillot, L Xing, L 7Àang, L Zhuang, and l.Zhang, The stationary Cretaceous paleomagnetic pole of Sichuan(South China Block), Tectonics, in press, l99lâ.
Enkin, R. J., Y. Chen, V. Courtillot, I. Besse, L S. Xing, Z. H.7hang,Z. H. 7Àuang and J. X. Zhang, A Creaceous pole from South China,and the Mesozoic hairpin tum of rhe Eurasian APWP, ,/. Geophys.Res., in press, l99la.
Fisher, R. A., Dispenion on a sphere, Proc. R. Soc. London Ser.4.,217,295-305, t953.
Hao, Y. C., D. Y. Su, J. X. Yu, P. X. U, Y. G. Li, N. W. Wang, H. Qt, S.Z. Guan, H. G. Hu, X. Li, W. B. Yang, L. S. Ye, Z X. Shou, and Q.B. Zhang, The Cretaceous System of China, 301 pp., GeologicalPublishing House, Beijing, 1986.
Irving, E., and G. A. Irving, Apparent polar wander pathsCarboniferous through Cenozoic and rhe assembly of Gondwana,Geophys. Ssrv.,5, l4l-188, 1982.
Jaeger, J. I., V. Courtillot, and P. Tapponnier, Paleontological view ofthe ages of the Deccan Traps, the Cretaceous/Tertiary boundary, andthe India-Asia collision, Geology, 17,316-319, 1989.
Kent, D., G. Xu, K. Huang, W. Zvnç and N. Opdyke, Paleomagnetic
4082 CHEN ET AL: MESOZOIC CONTNENTAL SEDIMENTS
results of upper Crelaceous rocks from south Cltina, Earth Planet.Sci. Leu.,79, 179-1E4, 1986.
Kirschvink, J. L, The least-squares line and plane and the analysis ofpaleomagnetic data, Geophys. l. R. astron. 9oc.,62, 699-718,1980.
l:vcinc, J. P., Y. Lemoigne. X. Li, X. rilu, S. Zhang, X, Zhao. W.Zhao, W. Zhu, and I. Zhu, Paleogeography of China inCarboniferous time at the light of paleobotanical data, incomparison with \Vcsæm Europc Carboniferous assemblages, C.R. Acad. Sci. Paris,34,391-394, 1987.
l-evi S., and S. Banerjee, On rhe origin of inclinarion shallowing inrcdeposited sediments, J. Geophys. Res., 9 5, 43834390, 1990.
Li, C. Y., A preliminary study of plate tectonics of China, (inChinese). BulI. Chin. Acad. Geol. Sci.,2(1), 1-19, 1980.
I4 C. Y., Q. Wang, X. Y. Liu, and Y. Q. Tang, Explonatory Notes tothe Tccanic Map of Asia, 49 pp., The Canographic PublishingHouse, Beijing, 1982.
Lr" Y. P., R Sharps, M. McWilliams, A. Nur, Y .4. U, Q. Li, and W.Zrang, Paleomagnetic rcsults from Late Palcozoic dikes from thenorthwestcm Dzungar Block, nonhwestem China, Earth Planet.Sci. ku.,94, 123-130, 1989.
U. Y. P., Z. K. Zlrang, M. McWilliams, R. Shaçs, Y. l. Zhu, Y. A U,Q. Li, and A. Cox, Mesozoic paleomagnetic results of the Tarimcnton: Tertiary relativc motion between China and Siberia?,Gcophys. Res. Leu., 15,217-220, 1988r�.
Li, Y. P., M. McWilliams, A. Cox, R. Sharys, Y. A. Li, Z. I. Gao,Z. K.Zrang, and Y. 1.Z-Àai, Late Permian paleomagnetic pole fromdikcs of thc Tarim craton, China, Gcology,16,275-278,1988b.
Li, Y. T., The Terriary System of China, 362 pp., GeologicalPublishing House, Beijing, 1984.
Lin, J. L, Thc apparent polar wander paths for the north and southChina blocks, Ph.D. thesis, 248 pp., Univ. of California at SanuBaôara. 19E4.
Lin, J. L., M, Fuller, and W. Zhang, Prcliminary Phanerozoic polarwander paths for the North and Souù China blocks, Nature,3l3,444-449, 1985.
Mattauer, M., P. H. Matte, J. Malavieille, P. Tapponnier, H. Maluski,X. Z. Qin, L Y. Lun, and T. Y. Qin, Tectonics of the Qinling Belt:Build up and evolution of eastem Asia, Nature,3I7, 496-500,1985.
McElhinny, M. W., Starisrical signilicance of the fold test inpalcomagnetism, Geophys. J. R. Astron. Soc.,8,338-340, 1964.
McBhinny, M. W., B. J. J. Embleton, X. H. Ma, and L K. Zhang,Fragmentation of Asia in the Permian, Nature,293, 212-216,1 9 8 1 .
McFadden, P. L, and D. L. Jones, The fold test in paleomagnetism,Geophys, J. R. Astron.loc.,67,53-58, 1981.
McFedden, P. L., and F. J. Lowes, The discrimination of meandirections drawn from Fisher distributions, Geophys. L R. astron.loc.,67, l9-33, 1981.
McFadden, P. L., and M. W. McElhinny, The combined analysis of
remagnet izat ion c i rc les and di rect observar ions inpaleomagnetism , Earth Planet. Sci. Leu., 87, 16l-172, 19E8.
Molnar, P., and P. Tapponnier, Cenozoic tectonics of Asia: Effects of acontinental collision, Science, I89, 419-426, 1975.
Opdyke, N. D., K. Huang, G. Xu, W. Y. Zhang, and D. V. Kenr,Paleomagnetic results from the Jurassic of rhe Yangrze plarform, -/.Geophys. Res., 9 l, 9553-9568, 1986.
Simkin, T., R. I. Tilling, J. N. Taggan, W. J. Jones and, H. Spall, This .dynamic planet-Vy'orld map of volcanoes, eanhquakes and plaretectonics, U. S. Geological Survey, Washington D. C., 1989.
Stamatakos, J.A., K.P. Kodama, L. Vittorio, and T. L. Pavlis,Paleomagnetism of Crctaceous and Paleocene sedimentary rocksacross the Castle Mountain Fault, south central Alaska, in DeepStructure and Past Kinematics of Accrered Terranes, Geophys.Monogr. Ser., vol.50, edited by J. W. Hillhouse, pp. l5L-171 ,AGU, Washington, D.C., 1989.
Tapponnier, P., and P. Molnar, Active faulting and Cenozoic tectonicsof the Tien Shan. Mongolia, and Baykal regions, l. Geophys Res.,84,3425-3459, 1979.
Tapponnier, P., G. Peltzer and R. Armijo, On rhe mechanics of thecollision between India and Asia, Geol. Soc. Spec. Publ.,19, ll5-1 5 7 , 1 9 8 6 .
Vr'ang, S., The Jurassic System of China, 350 pp., GeologicalPublishing House, Beijing, 1985.
Westphal, M., M. L. Bazhenov, l. P. Lauer, D. M. Pechersky, and J. C.Sibuet, Paleomagnetic implications on the evolurion of the Terhysbelt from the Atlantic Ocean to the Pamirs since the Triassic,T ectono p hys ics, 1 2 3, 37 -82, 1986.
Zhang, L. C., and N. Y. Wu, The geotectonic evolution of Tien Shan,Xinj iang Geol . ,3(3) l -14, 1985.
ZijdeweLd, J. D. A., A. C. demagnetization of rocks: Analysis ofresults, in Methods in Paleonagnerrsrn, edited by D. W. Collison,K. M. Creer, and S. K. Runcom, pp. 254-286, Elsevier, New York,1967 .
--TÏ]lî6o"c, Y. Chen, J.-P. Cogné, V. Courtillot, and P.
Tapponniet, Insritut de Physique du Globc de Paris, Géornagnétisme et
Paléomagnétisme (CNRS: U A 7 Z9), 4, place Jus sieu, 7 5252 P e;isCedex 05 France.
M. Bai, Burcau of seismology of Xinjiang, People's Republic ofChina.
E. Buffetaut, CNRS laboratoire de Paléontologie des VenébÉs,Univeniré Paris W, France.
M. U. G. tffang, C. Wei, and H. You, Xinjiang Engineering
Institute, People's Republic of China.
(Received April 24, 1990;rcvised November 12, 1990;
acceped December 13, 1990.)
