A Late Miocene/Pliocene Origin of the Inverted Metamorphism of the Central Himalaya
T.M. Harrison F.J. Ryerson P. Le Fort A. Yin O.M. Lovera
This paper was prepared for submittal to the Geological Society of Amerika Abstracts with Progress Conference Denver, CO October 28-31,1996
‘llds documentwas prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government northeUniversity ofCalifornia noranyoftheir employees, makesany warrsrq, express orimplied, orassumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringeprivately owned righta. Referenceherein to any specificoxnrnercial producL process, or service by trade name, trademark, manufacturer, or otherwise, does not neceaaarily constitute or imply its endorsement, recommendation, or favoring by the United Statea Government or tie University of California. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United Statea Government or the Universityof California,and shallnot be used for advertisingCRprcduct endorsement purposes.
A Late Miocene/Pliocene
of the inverted metamorphism of the Central
Himalaya T. Mark
P. Le Fort’,
An Yin and Oscar M. Lovera
Department of Earth a Space Sciences md lGpp, University of California, Los Angeles, CA 90024, USA *Lawrence Livemore Nation~ ~kmtoryt Liverrnore, CA 94550, USA ‘Institut Dolomieu. C.N. R.S., Grenoble, 38301, France
The spatial recognized identify
in the Himalaya
ago’, has inspired
the best known
of this remarkably in Himalayan
in the footwall
such as the need for exceptional
range, and transcends
at ca. 5
such as why the MCT (and not the more recently
of the MCT fault occurred
marks the break in slope of the present day mountain
Using a new
we have discovered
It has been widely
activation of a broad
at that time.
the MCT zone which juxtaposed
to have been active during the Early Miocene.
in situ Th-Pb dating of monazite
is that found immediately
that the pattern
Despite its role in absorbing a significant fraction of Indo-Asian convergence, fundamental issues regarding the evolution of the MCT, such as the timing of its initiation and cessation, remain poorly known. As all the major geologic elements of the Himalaya (i.e., Ieucogranites, South Tibet Detachment System (STDS), and inverted metamorphism; Fig. 1) have been related to the evolution of tie [email protected]
its SUP~story is key to
understanding the nature of this unique erogenic belt. AIthough the question of a genetic link betwt%n Early Miocene anatexis in the MCT hanging wall and slip on the MCT and STDS is being debated”,
there has been no
disagreement that Himalayan inverted metamorphism is temporally related to Early Miocene activity on the MCI’. The S-directedthrust faults of the Himalaya, principally the MCT and the Main Boundary l%ms~ appear to sole in a common decollement at depth%’[email protected]
1). In general, the Mm places high-grade gneisscs of the TIMtan
Slab (= Greater Himalayan Crystalline) and the Main BoundW ~st
atop schists of the Midlands Formations (= Lesser Himalayan Formations),
those schists against unrnetarnorphosed Neogene molasse (Fig. 1).
Studies of the MCT h~ging wall indicate that deformation and anatexis were Occurnng at 22tl M% while cooling ages in the upper potion of tie Tibetan Slab suggest that ductile deformation had ceased there by -18 Mas’1*i3.The amount of slip ~ong tie Mm is inferred to be about 150-300 km, based on flexural modeling of gravity &ta and balanced cross-sections ‘619. Juxtaposition of the T:betan Slab atop the Midlands along much of the Himalaya is asswiated with a zone of inverted metamorphism that has long intrigued geologists (Fig. 2). Reposed causes of the apparent increase in metamorphic grade with higher structural position (i.e. shallower depth) include; the heating of the colder lower plate by emplacement of a hot nappe, either due to thermal relaxation alone or aided by strain heating*O=u, the folding of pm-existing isogradsa, imbricate thrusting,
mantle delaminationti, and the ductile shearing of an existing zone of
right-way-up metarnorphismm. Within the central Himalaya the Tibetan Slab increases in thickness from -3 km in the Kali Gandaki section to >10 km in the Burhi Gandaki section @lg. 1). ‘his correlates with an eastward increase in thickness and peak grade of the inverted metamorphic section, reaching kyanite grade at Darondi Kbola and comndum grade at Ankhu Khola (Hg. 1). The metamorphism within the Midlands rig. 2) typically increases in grade over a N-S distance of 10-20 km from chlorite through biotite, garne~ and kyanite, reaching sillimanite grade in the hanging wa112.Ttte region approximately bounded by the garnet isograd and the MCT is usually characterized by a highly shard
4-8 km thick zone of distributed deformation ([email protected]
2) commonly referred to * the ‘MCT
zone’*32. Kinematic indicators within this broad zone document that the shear sense is uniformly topto-the-south. Sheath folds, asymmetric Wnage,
mylonitic gneiascs wi~ S-C f~c%
abundam We examined b
acceamy mineralogy of pelitea from the Midlands of centrtd Nepal, which vary b
grade in the structurally lowemrmat samples to kyanite grade directly adjacent to the Mff
(13g. 2), with the goal of
establishing the timing of inverted metamorphism. Monazke [CeFOJ is unstable in pelitic rocks during diageneia and does not reappear until a temperature of -5(KPC is reaehedw%. During the period of instability, the constituents of monazite are carried either by sdhmite or light rare earth element (LREE) 0xidcs3*M. RoMM not greatly exceed @O”d,
diffbaive loss of radiogenic Pb (I%*) produced by rit situ decay of ‘XIIand U in monazite
is not significant. SEM studies of our rocks indicatethatatchlorite and biotite grades, allanitc is the principat host of LREE. Monazite overgrowths ~gin to appear on allanite hosts C1O* to ti garnet isograd and -1050 pm-sized neoformed monads Four montite
~ present within the garnet grade rocks. grains identified in a thin section of sample AP332 (Fig. 2) wereanalyzedusing the ‘Phi’h
ion microprobe method.
technique is well-suited to determining in situ crystallization ages of Tertiary
monazites ss it permits textural relationships between monazite and mineral host to be retained, and the spatial selectivity permits inherited monazite cores to be identified and avoided. Th concentrations are typically very high in monazite (-3-8%) which, coupled with low levels of common Pb, results in relatively high levels of ‘I%* in young samples. The short time scale for secular equilibrium to be achieved among the intermediate daughters of ‘h
precludes monaxite from containing excess 2~b*.
The high ‘Pb”
abundances and high ionization efficiency
of Pb from monazite under 02” bombardments permit Th-Pb age determinations to be made in a -10 pm spot on a CU.5 Ma specimen in -20 min. Ages are determined by direct reference to a standard, in this case 554 monazite which yields a ‘Pt#h of N%’.
isotope dilution age of 45i2 Ma”. The standard calibration has a typical reproducibility
Performing the ion microprobe measurements in polished section rather than as separated grains in an
epexy mount has presented no technical difficulties. The regions within the polished thin section containing monaz.ite were isolated by sawing and then mounted with several grains of 554 monazite in a 1” diameter epoxy mount and Au coated for analysis (see ref. 8 for additional details of analytical methods). Based on previous geochronological studies of the Tibetan Slab’”s, our expectation was that crystallization ages might range from -20 to >30 Ma. Surprisingly, four different monazite grains yield %@%
agea between 4.S
and 8.0 Ma (Table 1). Of potentially greatest signifbnce
is a monazite grain included within a mm-sized garnet in
AP332 @lg. 3). Garnet-biotite-plagioclase-muscovite-qu-
thermobarometry (see ref. 38) for this sample yields a
P-T of 68 kbar and 55NO”C,
typical of conditions found within the MCT zone in this region~. Since chlorite
would have become unstable had the pteasure exti
suggests that the pressure estimate reflects the peak value. l%e two Ilt-Pb measurements on this grain yield an average age of 5.tktO.1 Ma. Since the garnet and monazite isograds are essentially coincident in temperature (i.e., -525°C), we conclude that these data are pfinruf-ie
evidencethat garnet growth in the presently exposed rocks was
occurring at a depth of-25 km at -5-6 Ma. Preliminary measurements of monazite in garnet from sampk DH-73-
96, Darondi ICbcda(Rg. I), yield ag= between 9-11 Ma Wtich. thowghsomewhat older, are consistent with the pattern of inved Several otir a N-S section ~ugh
in the =nw~ ~~aya
fines of evide~e ~SO suggestthatthelvICTwas active during the Pliocene. ‘ArFAr the D~on~ ~d BLuhi @Iu
-4 Ma witin ~d @=ent
to the Ma
v~ley’s drop from -16 Ma in the upper MCX hanging wail to
zone before climbing to pre-Tertiary ages (Fig. 4d). A U-T1 oxide
precipitated in m ~bitic alterationzonewithin the Tibetan Slab yielded U-Pb ages of 4.8-5.2 MalJ. Although these young age-swere previously ascri&d, respectively, to isotopic resetting and crystallization
due to hot fluids channeled
along the MCT at 5 Ma’3, the -7 kbar pressure recorded by AP332 rules out this interpretation. This is because both the upper portion of the Thetan Slab and the lower grade portion of the Midlands are known from thermochronomeUy 13’14 to have been relatively COICL and therefore relatively close (<15 km) to the Earth’s surface, at 5 Ma when AP332 was at a &pth of -25 km. Although wedge extrusion (e.g., ref. 4) of the MCI’ zone during the Pliocene can explain the pattern of mineral ages (Fig. 4d), this mechanism would require the MCI’ to have beut reactivated as a normal fault and field evidence for a topto-the-south relationship across the MCI’ zone is overwhelming~.
The hypothesis we favor is tha~ following termination of a phase of slip during the Early
Miocene, the M(2T was reactivated during the Late Miocene. Stip was initially resticted to the Ma
resulted in the burial metamorphism of the upper Midlands (as recorded by sample AP332). Uplift induced advection of heat within the hanging wall may have driven fluid circulation and produced the localized greenschiat alteration13. In a second phase of deformation, strain was accommodated across the broad MCT shear zone resulting in juxtaposition of two right-way-up metamorphic sequences (see ref. 28). We have consolidated our obaenations in the framework of a numerical model in order to cunatrain a varietyof possible Late MiOc.end%cctte
deformation histories that are consistent with our thermobarochronologic result. We
assume that the MCI’ was the footwall ramp in a fault-bend thrust geometryw and use a tinite-difference solution to the diffusion-advcctioa equation to teat models involving both hanging wall deformation alone and combined hanging wall and footwall deformation. An example of the latter model which closely matchea our isotopk constraints is shown in Fig. 4. In this model, the MCT fault is activated at 8 Ma (foilowing 12 my. of inactivity) with a tad disphwexncnt rate of 22 mdyr
4a). At 6 MA =tiviw s~fts from ~
boundary of t.lMMCT zone (I%g.4b). Although the MCI’ zone is without question a broad shear zone (yellow in
Fig. 4~b), for simplicib wc Mve m~el~ effect is that garnet_
it ss a single fault at the b=
of the shear zone. In either case, the net
to the h~ging Wd which continues to be transported Up the
ramp. Slip terminaks at 4 Ma. Together with the kinetic parameters for Ar diffusion in biotite’”, the calculated thermal histories co~esponding to mineral ~ting localities prdict a biotite age distribution (solid curve) that C1O=1Y match= ~ emPiric~ ~sultsi’ (~lg. ‘@. The model alSOpredicts that the presently exposed upper Midlands would have reached the garnet isograd at -6 Ma at a P-T of 7 kbar and 550”C, in excellent agreement with the monazite-ingarnet result for AP332 (square in Fig. 4c). Therrnobarometry41 of rocks collected in a roughly horizontal N-S traverse through the Tkan
that pressures as high as -7 kbar were achieved adjacent to the MCI’, whereas peak pressures at the top of the section were only -34 kbar. When pressure is plotted against inferred structural height above the MCT, these results yield a ‘normal’ Iithostatic gradient of 0.27 kbsdkm” implying that these rocks were once sub-vertical and subsequently rotated -30°. Our numerical model predicts that hanging wall rocks from the range of depths comsponding to these pressures would today be exposed at the surface together as a result of the rotation associated with their sequential transport up the ramp (see Fig. 4b). Although the results of our thermal modeling are not unique, jwticularly with respect to the timing of activation of the MCI’ zone, they demonstrate both the physical plausibility of our hypothesis and that a wide vuiety of observations can be rwonciled within its framework. If Late Miocene activation of the MCT zone and the subsequent development of inverted metamorphism were a widespread occurrence across the Hirnalay& two long--g
easily resolved. Although the surface expression of the presently active Main Frontal ‘Ihrust and recently active Main Boundary Thrust are exposed at elevations of <1 krm it is the MCT zone that marks the most dramatic topographic break in S1OP within the Himalaya’n’.
If the MCT zone had been inactive for the past -20 my., why
would this feature persist? Although a shallow (- 15°) ramp on the decdlement”
at -20 km depth maybe a
contributory factor in elevating the MCI’ zone, it would not produce a sharp break in slope over such a short distance unless the overlying rocks were unusually flexurally weak. More likely, this feature simply reflects the fact that significant displacement along the MCT zone was occurring aa recently as 4 Ma. The one common element among the numerous attempts to account for the thermal energy required by Tibetan Slab artatexis is the requirement for some extraordinary process to have aided melting. A variety of special mechanisms have been proposed including
high (100 MPa) shear swess ~ong the [email protected]
mantle delaminationti, very long (>25 my.) or short (S1 my.)
thermal incuhtion periodsux, high concentrations (mmy 100’s”of ppm)of B, F, and Li’s, or rapid (>5 mdyr) decompmssiond. Although tk juxtaposition of rocks undergoing partial melting with rhe relatively cold rocks of the adjacent MCX zone Undergoing endothermic reactions’” appears enigmatic, our hypothesis provides a simple expl~ation.
The (presumably higher grade) footwall rocks that were juxtaposed against the thrust flat of the MCT
during anatexis have since been displaced northward and replaced by a tectonically telescoped section of lower grade rocks. As a consequence, any attempt to explain the present juxtaposition of Tibetan Slab and Midlands Formations without recognizing the diachroneity of deformation is ordained to require extraordinary circumstances.
1. Oldham, R.D., Rec. Geol. WV. India 16, 193-198 (1883). 2. Colchen, M., Le F- ~:, a ~&hcs. A. A~purna-Manaslu-Ganesh Hind notice de la cane geologiquie au IZ200.00S [email protected]
~l~on: F~nch-Engllsh. p*s: Centre Nationaldela RechercheScientifique ( 1986). 3, LeFort, P.. Cuney, M- ~niel) C.. F~ce-~ord. Tectonophysics 134, 39-57 (1987).
C.. Sheppard, S.M.F., Upreti, B.N., A Vidal, P.,
4. BurchfieL B.C., Chen! z., Hod8es. K.V.. Liu Y., Royden, 269, 41 p. (1992). 5. Yin, A. J-
C., C%XU, J. Ceoi. Sot. Am Spec. pap.
M. 98, 14,245-14,256 (1993).
6. Harris, N. a Massey, J. Tectonics 13, 1537-1546 (1994), 7. Le ForL P. in The Tectonics of Asia (eds. A. Yin & T.M. Harrison), Cambridge University Press, 95-109 (1996). 8. Harrison, T.M., McKeegan, K.D., & LeFort, P. Earth Planet. Sci. Left. 133, 271-282 (1995), 9. Gansser, A. The Geology of the Himcdayar, Wiley InterScience, New York, 289p, (1964). 10. Le Fort, P. AnL J. Sci. 275A, 1-44 (1975). 11. Burbank, D.W., Beck, R.A. & Mulder, T. in The Tectonics of Asia (eds. A. Yin & T.M. Harrison), Cambridge University Press, 149-188 (1996). 12. Zhao, W. et al., Nature 366, 557-559 (1994). 13. Copeland, P. , Harrison, T.M., Hodges, K.V., Mar6ujol, P., LeFort, P., & P&her, A. Jour. Geophys. Res. %, 8475-8500 (1991). 14. Hubbard, M.S. & Harrison, T.M. Tectonics 8,865-88 (1989). 15. Parrish, R.R. & Hodges, K.V. Geol. Sot. AnL Abs. w. Progs. 2S, A174 (1993). 16. Lyon-C~n,
H. & Molnar, P. Tectonics4,513-538
17. Schelling, D. Tectonics, 11,925-943 18. Sckl]ing, 19. Srivastava 20. Aria
K. Tectonics 10, 851-862 (1991).
P. & Mitr4 G. Tectonics 13, 89-109 (1994).
K. Tectotwphysics 95,43-60
21. Graham, C.M. & England, P.C. Earth PlaneL Sci. Mt. 31, 142-152 (1976). 22. England,
P. & Thompson, A.B. J. Petrol. 2S, 894-928 (1984).
Molnar, P.& England, P. J. Geophys. Res. 95,4833-4856
England, P., Le [email protected]
P., Molnar, P., & Pi?eher, A. Jour. Geophys. Res. 97,2107-2128
25. England, P. & Molnar, P. Tectonics 12, 145-157 (1993). 26. Searle, M.P. & Rex, D.C. J. Metamorpk
Geol. 7, 127-134 (1989).
27, Brunei, M., & KienxsL J.R. Ciam J. Eiuth Sci. 23, 1117-1137 (1986). 28. Hubbard, M.S. J. GeoL 104, 1996. 29. Hubbard, M.S. Ph.D. disaermtion, Mass. Inst. Tech., 1988. 30. Bouchez, J.L. & P&her, A. Tecmophysics
31. MaeFarlane, A.M. Jour. Metamorph GeoL 13,595-612 (1995). 32. P&her, A. J, Metantorph Petal. 7,31-41 (1989). 33. Overstseet W.C. U.S. Geof. Sum. Projl Paper 530, 327p.
34. Smith, H.A. & Barreiro, B. Contrib. Mineral Petal. 105,602-615
3S. Hanison, T. M.. figsbury,
J.A. & Miller, C.D. 2nd Gofdschmidt Cortf. Abs, 52(1990).
J.A.?Miller.C.F..Wooden,J.L. & Harrison, Smith.H.A.~ ciletti}B. EOS75, 692 (1994).
36. K.ingsbWs 37.
38. SpeaT, F.S. Me~~hic Mineralogic~
T.M. Chem Geof. 110, 147-168 (1993).
799 p. ( 1993).
39. SUppe, J. AnL J. Sci. 283, 684-721 (1983).
40. Grove, M. & Htison.
T.M. An Min. 81, 940-951 (1996).
41. Hodges, K.V., Lc ForL P. & P&her, A. Geology 16, 707-710(1988). 42. Nakaw T. in Tectonics of the Western ffimafaya (eds. L.L. Malinconico & R.J. Lillie), Spec. Pap. Geol. Sot. Am. 232,243-264 (1989). 43. Masek, J.G,, Isacks, B.L., Gubbels, T.L. & Fielding, E.J. Jour. Geophys. Res. 99, 13,941-13,956 (1994). .-4bstmctsof the 1lth Hima&yan-Karakorutn-Tibet Workshop, Flagstaff, AZ April 29-May 1, 1996, 171-172(1996).
44. Yeats, R.S.
Guillot, S, & Le [email protected]
P. L.ithos 35, 221-234 (1995).
Bird, P. J. Geophys. Res. 83, 4975-4987(1978).
M. Tatsumoto, personal communication,
Acknowledgments We thank Kevin McKecgan and Chris Coath for their assistance with the ionmicqxob ~urements Dan Farber for the SEM characterization
by grantsfrom the NSF. We
particularity wish to thank the directors and stafT of the W.M. Keck Foundation f- their generous suppwt.
Support Lawrence Depar~ent
for this research was provided
by The Geophysics
and worked performed
under the auspices of the U.S. under Contract W-7405-Eng+.
Table 1. T%-Pb ion microprobe monazit.e ages grain
‘pb~ spot ‘UPb/D’Th a~e (%)
5.7io-7 24 8.&tO.4 84 81 5.5ti.2 21 5. 7i-o. 1 87 2 6.0M3.9 34 51 4.6i0.8 28 61 5.0+0.3 77 2 5.5Ml.2 80 3 :alculated assuming common ‘wPb/wPb = 39 11
map of the Himalaya and Southern Tibt (after ref. 3) showing tie relationship betw~n
the Tibetan Slab md Midlands Formations including reference isograds of the inverted metamorphic section. The Thetan Slabis juxtaposed
against the Tethyan metasediments by the STDS. The section marked A-A’
is shown in Fig. 2. KG = Kali Gandaki, BG = Burhi Gandaki, D = Darondi Khola, AK = Ankhu Khola. Fig. 2 Generalized
Cross section through the central Himalaya (after ref. 2) illustrating the pattern of inveti
metamorphism beneath the Main Central Thrust (MCI’), MBT = Main Boundary ThrusL STDS = Southern Tibetan Detachment System. The projected location of sample AP332 is shown. [email protected]
3 Reflected light photornicrograph of sample AP332 showing garnet (gt) in a biotite (hi) and quartz (qt) matrix.
Therrnobarometry yields a P-T of 7*1 kbars and 53M0°C.
Two Th-Pb analyses of the monazite
inclusion yield an average age of 5.6MI.1 Ma indicating garnet formation cccurred at the time at a depth of -25 km. The discolored region near the monazite is where the Au coating has km removed by sputtering. Fig. 4 Thermal modeling was undertaken assuming a fault-bend fold geometry, with top and bottom boundaries maintained at constant temperatures of 25°C and 1025°C, respectively, (resulting in an initial thermal gradient of 25°CAcm)and reflecting side boundaries. Denudation equals surface uplif%although topography was simulated in some runs. In this model, both the hanging wall and footwall move at 11 mndyr in opposite directions. (a) Slip along the Mm begins at 8 Ma. Hanging wall rocks are fmt transported along the flat and then up the ramp. NL25, the sample closest to the MCI’ and thus at the highest peak pressure (-7 kbar), is the fmt to reach the MCT (see refs. 13,41). It is followed sequentially up the ramp by U129 (-6 kbar) and U752 (-4 kbar) (see refs. 13, 41). Between 8 and 6 Ma aampk AP332 experiences burialheatingdueto overthmstingby the‘13betanSlab.(b) At 6 M% the MCT zone (modeled as a single fault at the base of the shear zone) is activated accreting h
Slab, onto the hanging wall which continues to be transported up the ramp. Slip terminates at 4 ML (c) Temperature histories predicted by the thermal model for Tibetan Slab samples NL25, U129, U752, MCT zone sample AP332, and sample AP524 fkom beneath the MCI’ zone. The model predicts that the exposed MCI’ zone would have reached the garnet isograd at -6 Ma at a P-T of 7 kbarand550% in agreement with AP332 result (box; Table 1). (d) Biotite ages from a N-S traverse across the MCI’ zone and into the Tbetan
Slab(~f. 13). m
hrmal historiescalculated by numericalmodel predict biotite ages (black curve) that
closely match the observed distribution.
Technical Information Department • Lawrence Livermore National Laboratory University of California • Livermore, California 94551