New geochronological constraints on the thermal and exhumation history of the Lesser and Higher Himalayan Crystalline Units in the Kullu-Kinnaur area of Himachal Pradesh (India).

New geochronological, petrological and structural data from the Beas-Sutlej area of Himachal Pradesh (India) are used to reconstruct the tectonothermal and exhumation history of this part of the Himalayan orogen. Sm-Nd garnet ages at 40.5 ± 1.3 Ma obtained on a pegmatoid from the inverse metamorphic High Himalayan Crystalline (HHC) in the Malana-Parbati area probably mark local melting during initial decompression. Ongoing exhumation in ductilely deformed leuco-gneiss is constrained by Sm-Nd garnet ages at 29 ± 1 Ma and white mica Rb-Sr ages around 24-20 Ma, while Bt Rb-Sr ages indicate a drop of regional metamorphic temperatures below 300 °C between 15 and 12 Ma. The deep Sutlej gorge exposes medium-grade paragneisses and Proterozoic orthogneisses of the Lesser Himalayan Crystalline (LHC), overthrust by the HHC along the Main Central Thrust (MCT). Mica cooling ages in the HHC are in the range of 14-11 Ma. Above the extruded wedge of the HHC, the Leo Pargil leucogranite and associated dykes intrude the Haimanta Unit (HU) below the weakly metamorphic Palaeo-Mesozoic sediments of the Tethyan Himalayas (TH). The Leo Pargil leucogranite yielded a mean Sm-Nd garnet age of 19 ± 1 Ma and Rb-Sr muscovite and biotite cooling ages between 16.4 and 11.6 Ma. Marked young extrusion of LHC units resulted in differentiated exhumation/cooling of more frontal parts of the orogen. Very young ductile deformation of LHC gneisses near Wangtu is constrained by late-kinematic pegmatite intrusions crosscutting the main mylonitic foliation. Sm-Nd garnet and Rb-Sr muscovite ages of these pegmatites range between 7.9 ± 0.9 and 5.5 ± 0.1 Ma. Published apatite FT ages down to 0.6 Ma also document accelerated diachronous sub-recent exhumation of different parts of the orogen. Together with geochronological data from the literature, the new results demonstrate that the HHC and the HU were deformed by shortening and crustal thickening during the Eohimalayan phase (Late Eocene-Oligocene), followed by a strong thermal overprint and intrusions of granitoids during the Neohimalayan Phase (Early to Middle Miocene). The LHC experienced amphibolite facies metamorphic conditions in the Late Miocene prior to extrusion between the HHC and the very low-grade Lesser Himalayan sediments. In conjunction with climate changes, young tectonic activity in this central part of the Himalayan orogen may have strongly influenced fluvial incision and erosion, and therefore, contributed to the accelerated uplift, as indicated by extensive accumulation of Late Miocene to Early Pleistocene fluviatile-lacustrine sediments in the Zanda basin, the Transhimalayan headwaters of the Sutlej, in Western Tibet.

Introduction

Separated by the Main Central Thrust (MCT), the Higher and Lesser Himalayan Sequences are two key units in the Himalaya, with distinct tectono-metamorphic histories (Gansser, 1964, Fuchs, 1982, Le Fort et al., 1987). Using provenance studies based on U–Pb ages and Lu–Hf isotope data of detrital zircons and whole rock Nd isotopic compositions, some authors argue that these sequences can be clearly distinguished (Parrish and Hodges, 1996, Whittington et al., 1999, Ahmad et al., 2000, Robinson et al., 2001, Richards et al., 2005) and the Lesser Himalayan Sequence is Palaeoproterozoic to Mesoproterozoic (c. 2500–1000 Ma) in age, while the younger Higher Himalayan Sequence was deposited in Neoproterozoic to Cambrian times (c. 800–500 Ma). In the course of intracontinental collision between India and Asia, both sequences were remobilized/restructured as part of the northern Indian crust, and incorporated into the Himalayan orogen in several tectonic units, which experienced different subduction, metamorphic and exhumation histories.

Palaeomagnetic and biostratigraphic evidence constrain the India–Asia collision in the NW part of the Himalayan orogen at between c. 55 and 50 Ma (Patriat and Achache, 1984, Rowley, 1996, Matte et al., 1997, DeCelles et al., 2004; see, however, Klootwijk et al., 1992). Since detection of (U)HP-metamorphic rocks documenting deep subduction of Indian crust in the course of this continental collision, geochronological data from coesite-bearing assemblages have set additional time constraints for early Himalayan orogenesis. Sm–Nd, Lu–Hf, (Tonarini et al., 1993, De Sigoyer et al., 2000) and U–Pb mineral ages (Kaneko et al., 2003, Parrish et al., 2006, Leech et al., 2005) from the NW Higher Himalayan Sequence in the Kaghan and Tso Morari (Moriri) UHP assemblages are in the range of c. 55–46 Ma; these data set an upper time limit for peak pressure of metamorphism during continental subduction to mantle depths (Liou et al., 2004). On the other hand, age data constraining the subsequent exhumation-related (post-high-P) high-T stages in the metamorphosed Indian crust are still scarce (Vance and Harris, 1999, Prince et al., 2000). Cooling ages from the literature suggest that the post-peak-P evolution was differentiated, both across and along strike of the Himalayan chain. Some authors explained differences in the P–T ratio/post-P peak thermal evolution, and cooling along the orogenic belt by oblique/diachronic continental collision (Guillot et al., 1999, Lombardo and Rolfo, 2000). The orogenic profiles (SSW–NNE) in the Simla–Kinnaur-Leo Pargil and Kullu–Lahaul/Spiti-Tso Moriri transects (Fig. 1) show that the northern (i.e., hinterland) parts were exhumed to relatively shallow crustal depths early, by c. 40 Ma (Schlup et al., 2003), while regional cooling in the southern nappe pile of the Higher Himalayan Sequence seems to have been effective only since Miocene time, i.e., from c. 25–20 Ma onwards (Jain et al., 2000, Vannay et al., 2004). These and related observations led to an evolutionary model that discerns distinct tectonic phases, i.e. the Eo-Himalayan (Eocene–Oligocene) vs. Neo-Himalayan (Miocene) phases (Wiesmayr and Grasemann, 2002, Vannay et al., 2004, Schlup et al., 2011, and references cited therein). Unfortunately, few age results for high-T metamorphic minerals such as garnet are available. Garnet is useful for constraining the metamorphic peak and, thus, serves as a key-mineral for a reliable reconstruction of post-peak-P, high-T exhumation-related deformation and initial cooling stages of metamorphic belts. Geochronologically robust data for post-high-P Himalayan garnet have been published only from two far-apart localities, i.e., the Suru crystalline in Ladakh (Vance and Harris, 1999) and the Garhwal Himal (Prince et al., 2000).

Fig. 1

(a) Geologic–tectonic sketch map and generalised profiles of the NW Himalaya, based on: Fuchs, 1982, Thakur and Rawat, 1992, Frank et al., 1995, Steck et al., 1998, Dèzes, 1999, Vannay and Grasemann, 2001, Hager, 2003, Murphy et al., 2002, Kempf et al., 2009. The profiles of the Kullu (upper transect) and Sutlej transect (lower transect) follow Frank et al. (1977), and Wiesmayr and Grasemann (2002); observations from Chambers et al. (2009) and Kempf et al. (2009) have also been included. The inset sketch (upper right) is from Vannay and Grasemann (2001). (b) Geographic sketch of the area of investigation, showing main sampling sites (sample locations 1–5); additionally, locations of outcrops/samples as illustrated in Fig. 4a and b and Fig. 5a–c are shown.

This paper presents new geochronological data for garnet, white mica and biotite, together with petrological, geochemical and structural results from the Beas–Parbati and the Sutlej valleys in Himachal Pradesh, India, focusing on: (i) placing new constraints on the thermal-metamorphic and igneous evolution and, (ii) refining the late deformational history by evaluating the cooling age pattern with respect to the overall exhumation/cooling and erosion history of this part of the Himalayan orogen.

Regional geological setting

In the Kullu–Kinnaur region of Himachal Pradesh, NW Himalaya, the MCT tectonically juxtaposes rocks of the Lesser Himalayan Sequence, exposed in the Larji–Kullu–Rampur Window (LKRW), against rocks of the Higher Himalayan Sequence (Fig. 1). In the LKRW, one of the largest tectonic windows in the NW Himalaya extending along strike more than 100 km from NW to SE, the Lesser Himalayan Sequence consists of at least two tectono-metamorphic units: (i) The very low- to medium-grade metamorphic Proterozoic meta-quartzites, shales, dolomites and meta-igneous rocks (Frank et al., 1973, Frank et al., 1977, Miller et al., 2000, and references cited therein) of the Lesser Himalayan Sedimentary Succession and (ii) the amphibolite-facies ortho- and paragneisses of the Lesser Himalayan Crystalline Unit (LHC), i.e. the Jeori–Wangtu gneisses. The metasediments are separated from the overlying LHC by the Munsiari Thrust (Vannay and Grasemann, 1998) (MT in Fig. 1). The Higher Himalayan Sequence consists of three major tectonometamorphic units: (a) The amphibolite facies-grade to migmatitic metasediments of the High Himalayan Crystalline Unit (HHC) form an extruding wedge above the Main Central Thrust and below the South Tibetan Detachment System (STDZ) (Grujic et al., 1996, Vannay and Grasemann, 2001, Godin et al., 2006, Harris, 2007). The Northern Himalayan Crystalline (NHC) is considered to be a part of the Higher Himalayan unit, which was partly exhumed shortly after the collision between India and Asia (e.g., UHP of the Tso Morari area, de Sigoyer et al., 2000), partly late, by Miocene E–W extension and dome formation (e.g. Leo Pargil, Thiede et al., 2006; and Gurla Mandhata, Murphy et al., 2002) (Fig. 1). (b) Above the HHC wedge the medium- to low-grade Haimanta Unit (HU) represents the Proterozoic part of (c) metasediments of the Tethyan Himalayas (TH). Mainly Ordovician (and rarely older) granitoids within the HHC and the HU suggest a common palaeogeographic origin of these two units (Richards et al., 2005).

The two major river systems exposing the tectono-metamorphic sequences in this part of Himachal Pradesh are the Beas–Parbati system originating in the southern part of the central Himalayan chain, and the Sutlej, the headwaters of which are within the Indus-Tsangpo suture zone (IYSZ) in Lake Rakshas (Lagna Tso), near Mount Kailash, in western Tibet (Fig. 1). Apart from the Indus and the Tsangpo–Brahmaputra, the Sutlej is the only large Transhimalayan river that cuts through the entire mountain range, exposing some of the most spectacular tectono-metamorphic features of this young orogen.

Over the past c. 20 years, the cooling and exhumation history of the wider study area in the NW Himalaya has been addressed by various working groups, e.g., Searle, 1996, Jain et al., 2000, Vannay and Grasemann, 1998, Vannay and Grasemann, 2001, Wiesmayr and Grasemann, 2002, Schlup et al., 2003, Schlup et al., 2011, Vannay et al., 2004, Thiede et al., 2005, Thiede et al., 2006, Caddick et al., 2007, Chambers et al., 2008, Chambers et al., 2009. Geochronological data from these, and some older investigations, are integrated in the present study to better define the regional exhumation/cooling history.

Methodology and analytical techniques

Whole rock and mineral chemistry

Whole rock major and trace elements were determined on powder pellets by X-ray fluorescence using standard methods. The composition of mineral phases was determined with the JEOL superprobe JXA 8100/8200 at the University of Innsbruck, using wavelength dispersive analytical modes with 15 kV and 20 nA sample current on a benitoite standard. For detection and identification of micro-inclusions (<5 μ) in garnet backscattered electron imaging (BSE) and energy dispersive X-ray spectrometry (EDS) were used. Trace element and REE mineral compositions were determined by laser-ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) at the CNR-IGG, Section of Pavia, Italy with a laser probe consisting of a Q-switched Nd:YAG laser, model Quantel (Brilliant), whose fundamental emission in the near-IR region (1064 nm) is converted to 266 or 213 nm by harmonic generators. The ablated material was analysed by a double focussing sector-field analyser Element I (ThermoFinnigan MAT), in which the standard field regulator power stage of the magnet and the ICP torch were upgraded to those of the Element II model. Helium was used as carrier gas and mixed with Ar downstream of the ablation cell. NIST SRM 612 was used as external standard and 29Si as internal standard. Precision and accuracy were assessed from repeated analyses of the BCR-2 (g) reference standard and resulted usually better than 7% (1σ) and 10% (at ppm level), respectively. Mineral abbreviations are after Siivola and Schmid (2007).

Rb–Sr and Sm–Nd isotope analysis

The Sm–Nd and Rb–Sr analytical work was performed at the Laboratory of Geochronology, Center for Earth Sciences, University of Vienna. Carefully handpicked, optically inclusion-free pure garnet separates used for Sm–Nd analysis weighed between 47 and 97 mg. Before decomposition, or leaching, the fractions were washed for 30 min in warm (70 °C) 2.5 N HCl. Leaching procedures applied followed Anczkiewicz and Thirlwall (2003) using concentrated H2SO4. Pure mica concentrates (≫99%) were obtained by repeated grinding in an agate mill using alcohol, drying and sieving, and finally, magnetic purification. Chemical sample digestion, isotope dilution (ID), and element separation for Sm–Nd and Rb–Sr analysis follow those given in Thöni and Miller (2010). Total procedural blanks were <1 ng for Rb and Sr, and <50 pg for Nd and Sm. Sm, Nd, and Sr were run as metals from a Re double filament, using a Finnigan® MAT262 (for ID) and a ThermoFinnigan® Triton TI TIMS (for IC), while Rb was evaporated using a Ta filament. A 143Nd/144Nd ratio of 0.511847 ± 0.000005 (2σ; n = 55) and a 87Sr/86Sr ratio of 0.710251 ± 0.000010 (2σ; n = 54) were determined for the La Jolla (Nd) and the NBS987 (Sr) international standards, respectively, during the two periods of investigation. Within-run mass fractionation for Nd and Sr isotope compositions (IC) was corrected for relative to 146Nd/144Nd = 0.7219, and 86Sr/88Sr = 0.1194, respectively. Uncertainties on the 143Nd/144Nd and the 87Sr/86Sr isotope ratios are quoted as 2σm. For the 147Sm/144Nd and the 87Rb/86Sr ratios, a mean error of ±1% is applied (representing maximum errors), including blank contribution, uncertainties on spike composition, and machine drift. The regression calculation is based on these uncertainties and the isochron calculations follow Ludwig (2003). Age calculations are based on decay constants of 6.54 × 10−12 a−1 for 147Sm (Lugmair and Marti, 1978) and 1.393 × 10−11 a−1 for 87Rb (Nebel et al., 2011; all age data from the literature, based mostly on the 87Rb decay constant of Steiger and Jäger, 1977, have been recalculated), respectively; age errors are given at the 2σ level. For Nd, a continuous depletion of the upper mantle is assumed throughout geological time, and the following Depleted Mantle (DM) model parameters were used: 147Sm/144Nd = 0.222, 143Nd/144Nd = 0.513114 (Michard et al., 1985).

Results

Field evidence: sampling sites, structures, metamorphic zoning

Geographic locations of samples analysed in the present study (see Table 1, Table 2, Table 3, Table 4, Table 5; Fig. 1a) are numbered 1–5 in Fig. 1b. In addition, Fig. 1b shows the localisation of outcrops/samples (labelled F4, F5), as illustrated in Figs. 4a, b and 5a–c. UTM 43/44 coordinates are given for some samples, and the geographic terms for the sample locations from Fig. 1a and b are used in the text and Table 1, Table 4, Table 5 (see Fig. 1), as follows.

Table 1

Characteristics of analysed samples from the Kullu–Parbati and Sutlej areas, Himachal Pradesh, Indian Himalayas.

Sample	Unit	Location	Rock-type	Mineral pargenesis	Grt core	Grt rim
Kullu–Parbati area
T93	HHC	Malana	Ky–Grt schist	Qtz, Pl, Grt, Ky, Bt, Ms, Ap
T96	HHC	Malana–Chandarkhani Dhar	Leucocratic layer	Qtz, Pl, Kfs, Grt, Ms, Bt, Ap,Tur, Zrn	Alm68Prp8Grs4Sps20	Alm53Prp6Grs20Sps21
T278	HHC	Manikaran	Augen gneiss	Qtz, Pl, Grt, Ms, Rt/Ilm, Ap, Zrn	Alm73Prp15Grs9Sps3	Alm73Prp16Grs8Sps3
T394	HHC	Malana–Chandarkhani Dhar	Pegmatite	Qtz, Kfs, Pl, Grt, Ms, Bt, Ap, Tur, Zrn	Alm84Prp5Grs5Sps6	Alm82Prp4Grs4Sps10
T410	HHC	Manikaran	Leucocratic gneiss	Qtz, Kfs, Pl, Grt, Ms, Ap, Brl, Zrn, Py	Alm39Prp2Grs2Sps57	Alm41Prp3Grs4Sps52
Sutlej valley
09T13KIN	HHC	Peo-Morang	Ky-Grt schist	Qtz, Pl, Grt, Ky, Bt, Ms, Ap, Mnz, Zrn	Alm72Prp16Grs6Sps6	Alm67Prp11Grs5Sps17
HB25/00	HHC	Leo Pargil	Leucogranite
HB28/00	HHC	Leo Pargil	Leucogranite	Qtz, Kfs, Pl, Grt, Ms, Bt, Tur, Ap, Mnz, Zrn, Xtm	Alm57Prp1Grs1Sps41	Alm72Prp2Grs1Sps25
CH01-36	HHC	Leo Pargil	Leucocratic dyke
CH01-37	HHC	Leo Pargil	Leucocratic dyke
CH01-8	LHC	Wangtu	Pegmatite
CH01-9	LHC	Wangtu	Pegmatite	Qtz, Kfs, Pl, Grt, Ms, Ap, Mnz, Zrn, Xtm, U	Alm57Prp1Grs2Sps40	Alm62Prp1Grs2Sps35

Table 2

Representative trace element (ppm) data of garnet from Leo Pargil leucogranite HB28/00 and of garnet and muscovite from Wangtu pegmatite CH01-9.

	HB28/00	HB28/00	HB28/00	HB28/00	HB28/00	CH01-9	CH01-9	CH01-9	CH01-9	CH01-9	CH01-9	CH01-9
	Grt 1	Grt 2	Grt 3	Grt 5	Grt 6	Grt 1	Grt 2	Grt 4	Grt 5	Grt 6	Ms A	Ms E
Li	126	105	144	181	127	69.7	73.4	87.3	72.4	130	236	222
Sc	33.25	20.52	22.13	47.95	27,.43	5.72	4.37	3.99	3.12	4.04	10.3	7.7
Ti	156.96	110	165	231	143	223	178	161	224	208	875	1029
V	0.20	0.32	0.26	0.52	0.21	<0.01	0.02	0.01	0.04	<0.01	0.31	0.26
Cr	0.50	0.21	0.49	0.83	0.71	<0.19	<0.19	0.34	1.08	<0.16	92	24
Co	1.69	1.74	1.78	1.57	1.85	0.8	0.87	0.91	0.83	0.71	2.18	0.68
Rb	0.19	0.14	0.15	0.16	0.08	<0.04	0.16	0.13	0.18	0.36	1644	1783
Sr	0.28	0.26	0.34	0.36	0.23	0.30	0.33	0.35	0.29	0.64	1.29	0.37
Y	778	756	1113	1248	1089	776	1170	1406	771	1986	0.11	0.07
Zr	9.18	6.58	9.02	10.69	8.87	13.6	8.88	9.08	10.98	10.1	2.73	1.79
Nb	0.11	0.38	0.11	0.25	0.07	3.39	0.62	0.39	3.03	1.50	302	344
Cs	0.02	<0.02	<0.01	<0.01	0.02	<0.02	<0.02	<0.02	0.03	<0.01	46.6	49.8
Ba	<0.01	0.81	<0.01	0.01	<0.01	0.03	0.04	<0.01	0.01	<0.01	3.48	5.48
La	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	0.12	0.02
Ce	0.002	<0.001	0.007	0.005	0.002	0.03	0.01	0.01	0.02	0.01	0.08	0.02
Pr	0.004	0.001	0.001	0.004	0.002	0.03	0.01	0.02	0.01	0.02	0.01	0.01
Nd	0.163	0.07	0.116	0.146	0.11	0.58	0.35	0.27	0.41	0.38	0.03	0.02
Sm	1.19	0.84	1.37	1.46	1.17	5.47	3.58	3.45	4.58	4.43	0.01	0.005
Eu	0.062	0.048	0.061	0.058	0.066	0.02	0.02	0.01	0.01	0.03	<0.004	0.006
Gd	12.63	9.34	13.32	13.79	13.38	28.87	25.77	21.82	28.64	26.13	0.05	0.06
Tb	8.36	6.66	9.23	10.34	9.48	14.71	14.63	14.72	14.27	16.98	0.01	0.04
Dy	97.22	88.76	119.6	138.1	120.5	111.9	135.7	153.7	110.7	192.9	0.02	0.02
Ho	24.81	27.81	35.7	43.76	33.76	18.14	28.84	35.59	19.42	53.65	<0.004	0.002
Er	78.56	100.8	128.8	169.4	109.8	48.28	95.94	129.5	57.84	227.0	<0.007	0.008
Tm	12.59	18.0	22.84	32.53	17.74	10.16	23.32	34.69	13.18	70.88	0.01	0.005
Yb	87.58	133.8	167.3	255.4	120.1	91.48	240.7	398.3	129.7	911.4	0.13	0.087
Lu	11.21	19.09	22.26	36.89	15.08	11.02	33.97	58.24	17.09	155.4	0.04	0.025
Hf	0.32	0.29	0.36	0.47	0.36	0.82	0.48	0.52	0.84	0.64	0.57	0.42
Ta	0.40	2.19	0.35	0.83	0.25	5.54	1.66	0.99	9.91	3.21	35.2	46.8
Pb	0.02	<0.004	0.01	0.01	0.01	0.01	0.01	3.20	<0.004	<0.003	7.1	4.88
Th	0	0	0	0.12	0.04	0.01	0	0.09	0	0	0.11	0.03
U	0.06	0.07	0.05	0.13	0.04	0.27	0.07	0.07	0.21	0.16	0.09	0.08

Table 3

Major (wt%) and trace element (ppm) concentrations of granitic and pegmatitic rocks, Spiti–Sutlej.

Sample	HB25/00 HHC Leo Pargil	HB28/00 HHC Leo Pargil	CH01-36 HHC Leo Pargil	CH01-37 HHC Leo Pargil	CH01-8 LHC Wangtu	CH01-9 LHC Wangtu
SiO2	74.26	74.38	74.08	75.39	75.53	71.89
TiO2	0.07	0.04	0.06	0.01	0.03	0.02
Al2O3	14.54	14.87	15.79	15.29	13.49	16.27
Fe2O3	0.98	1.01	1.24	0.65	0.58	0.71
MnO	0.04	0.08	0.02	0.02	0.02	0.15
MgO	0.18	0.12	0.27	0.06	0.09	0.07
CaO	0.88	0.63	1.11	0.48	0.54	0.53
Na2O	4.25	4.22	3.34	5.26	3.43	4.24
K2O	3.80	4.24	2.73	2.12	4.41	5.98
P2O5	0.10	0.11	0.12	0.07	0.05	0.06
LOI	0.66	0.50	1.32	0.82	0.56	0.39
Total	99.76	100.20	100.08	100.17	98.73	100.31
F	250	100	<50	<50	550	150
Cl	220	<20	120	<20	<20	<20
Sc	3	3	<1	<1	<1	<1
V	21	7	21	8	9	<2
Cr	229	54	124	105	82	30
Co	<1	<1	<1	<1	<1	<1
Ni	7	5	7	5	5	4
Cu	2	1	<1	<1	<1	3
Zn	37	41	14	33	9	9
Ga	17	18	20	21	24	23
Rb	241	300	174	224	379	518
Sr	214	78	93	9	17	5
Y	11	14	5	5	6	21
Zr	31	30	27	25	7	26
Nb	8	11	32	16	32	23
Sn	14	13	51	49	19	7
Cs	40	36	24	10	<5	<5
Ba	383	154	249	13	39	7
La	10	9	<5	<5	<5	<5
Ce	<10	<10	<10	<10	<10	<10
Pr	<3	<3	<3	<3	<3	<3
Nd	9	<5	7	<5	6	<5
Sm	<5	<5	<5	<5	<5	<5
U	8	8	20	9	8	8
Pb	79	82	49	36	50	42
Th	4	4	5	2	2	7
ASI	1.22	1.23	1.67	1.34	1.24	1.17
Norm. Crn	2.09	2.46	5.67	3.66	2.25	1.99

< = Below detection limit.

Table 4

Rb–Sr analytical data and wr–mineral age results from the Malana–Parbati and the Sutlej valleys, Indian Himalayas.

Sample	Unit/location	Rb, ppm	Sr, ppm	87Rb/86Sr	87Sr/86Sr	±2σm	Mineral–wr age (Ma, ±2σ)
(A) Kullu–Parbati (locs. 1, 2)
T93 Bt	HHC Ky–Grt schist, Malana	486.7	1.87	772.0	0.966026	0.000011	15.35 ± 0.15
wr		151.7	39.0	11.37	0.803358	0.000004
T96 Bt	HHC, Malana–Chandarkhani Dhar	934.0	0.867	3379	1.568739	0.000460	17.08 ± 0.16
Bt, replicate		958.9	0.858	3512	1.582476	0.000088	16.72 ± 0.16
wr		203.8	109.8	5.402	0.765908	0.000004
T278 Bt	HHC augen gneiss, Manikaran	544.9	6.40	249.1	0.814742	0.000009	12.68 ± 0.12
white mica		284.2	76.61	10.80	0.773618	0.000007	20.46 ± 0.26
wr		66.8	111.0	1.752	0.771038	0.000004
MANG5 Bt	HHC, Ky–Grt gneiss, Manikaran	655.2	2.49	774.1	0.897227	0.000011	13.29 ± 0.13
wr		180.5	133.1	3.942	0.754604	0.000004
T410 white mica	HHC leucocratic geiss, Parbati	772.3	8.37	271.6	0.875083	0.000008	23.86 ± 0.25
wr		247.8	90.5	7.985	0.787439	0.000003
(B) Sutlej (locs. 3, 4, 5)
CH01-8 wr	Pegmatite, LHC near Wangtu	380.3	17.66	63.72	0.937917	0.000104
8C Ms		992.1	1.004	2988	1.163258	0.000239	5.53 ± 0.06
CH01-9 wr	Pegmatite, LHC near Wangtu	527.9	5.01	321.1	1.245588	0.000110
wr, replicate		511.7	4.89	318.9	1.245798	0.000016
9A Ms		1683.8	0.339	17124	2.668633	0.000060	6.08 ± 0.39
9E Ms/2		1628.1	0.720	6968	1.364671	0.000050	1.28 ± 0.1
9E Ms/2, replicate		1614.0	0.475	10544	1.446359	0.000030	1.41 ± 0.1
CH01-11 wr-V (vein)	Extension gash in amphibolite,	156.5	15.9	28.5	0.728231	0.000010	11.21 ± 0.1
Bt	HHC Karcham-Sangla	759.3	6.82	324.3	0.774582	0.000010
CH01-11 wr-NG	Wall rock of extension gash	601.9	54.4	32.11	0.732578	0.000010	10.3 ± 0.1
Bt		759.3	6.82	324.3	0.774582	0.000010
09T13KIN Bt	Grt–Ky schist, HHC Peo-Morang	456.1	3.180	419.9	0.826827	0.000007	14.56 ± 0.14
wr		192.8	210.5	2.659	0.742200	0.000004
HB 25/00 wr	Granitic gneiss HHC,	232.8	226.9	2.974	0.726094	0.000060
B2 Ms	Spiti–Sutlej confluence	675.5	15.0	130.8	0.752806	0.000010	15.0 ± 0.21
A Ms		689.7	17.6	113.8	0.751413	0.000010	16.41 ± 0.21
Bt		1449.3	2.35	1851	1.093787	0.000020	14.27 ± 0.11
HB28/00 wr	Grt-bearing leucogranite	301.5	81.7	10.70	0.731559	0.000010
Bt	(less deformed)	1996.6	3.55	1672	1.000610	0.000010	11.62 ± 0.1
B2 Ms	Spiti–Sutlej confluence	872.5	10.6	240.2	0.779601	0.000010	15.0 ± 0.21
CH01-36 wr	S-parallel folded dyke in Leo Pargyal	174.4	97.2	5.219	0.753742	0.000050
-36A Ms	Granite, HHC, c. 5 km NNW Nako	605.4	20.9	84.56	0.773881	0.000146	18.25 ± 0.21
CH01-37 wr	Dyke intruded in axial plane of	223.4	8.32	78.13	0.767958	0.000092
-37 B Ms/2	A fold; HHC, c. 5 km NNW Nako	1165.8	0.80	4698	1.876077	0.000040	17.23 ± 0.82
-37 F Ms/2		2265.1	0.555	18644	5.849877	0.000140	19.67 ± 2.0

Table 5

Sm–Nd analytical data and age results from the Malana–Parbati and the Sutlej valleys, Himachal Pradesh (Indian Himalayas).

Sample	Sm, ppm	Nd, ppm	147Sm/144Nd	143Nd/144Nd	±2σm	Mineral–wr age (Ma, ±2σ)	ε (t)	TDMNd (Ga)
(A) Kullu–Parbati area
MAN G5 wr	8.500	43.59	0.1179	0.511923	0.000002			1.74
T278 wr	8.472	42.87	0.1195	0.512065	0.000002			1.56
T93 wr	6.506	34.95	0.1125	0.511660	0.000002			2.02
T394 wr	0.051	0.171	0.1803	0.512039	0.000005
Grt (1)	0.093	0.019	2.8949	0.512744	0.000045	39.7±2.5	−11.6
Grt (2)	0.108	0.027	2.3898	0.512628	0.000020	40.7±1.5	−11.6
T96 wr1 (Bt schist + leucocratic l.)	3.491	18.08	0.1167	0.511708	0.000004			2.03
wr2, leucocratic layers	0.235	0.478	0.2974	0.511803	0.000004
Grt R (from wr2)	0.489	0.204	1.4511	0.513009	0.000043	(160 ± 6)	−18.3
T410 wr	1.562	5.134	0.1839	0.511953	0.000004
Grt R (bulk separate)	1.084	0.585	1.1196	0.512128	0.000004	29.5 ± 1.0	−13.3
Grt 1MF	1.081	0.611	1.0697	0.512125	0.000006	29.6 ± 1.3	−13.3
Grt 2MF	1.069	0.647	0.9977	0.512102	0.000007	28.0 ± 1.5	−13.3
(B) Sutlej valley
09T13KIN wr	6.588	34.25	0.1163	0.511889	0.000002			1.76
HB28/00 wr 1	1.434	4.674	0.1854	0.511895	0.000008
wr2	1.280	4.125	0.1875	0.511901	0.000004
Grt R1 (H2SO4, fine)	1.557	0.677	1.3877	0.512045	0.000007	18.5 ± 1.0	−14.3
Grt R2 (H2SO4, fine)	1.597	0.760	1.2695	0.512035	0.000005	19.1 ± 0.9	−14.3
Grt R3 (H2SO4, fine)	2.079	2.417	0.5199	0.511942	0.000004	19.3 ± 2.5	−14.3
Grt (H2SO4, not ground)	2.306	3.336	0.4179	0.511950	0.000004	(32.6 ± 3.2)
Grt (not leached)	2.455	4.165	0.3563	0.511951	0.000008
CH01-9 wr 1	1.297	3.512	0.2233	0.511283	0.000010
wr 2	1.089	3.002	0.2193	0.511284	0.000002
wr 3	1.117	2.960	0.2281	0.511284	0.000003
Grt (not leached)	5.328	7.309	0.4406	0.511304	0.000006	(14.3 ± 4.4)	−26.4
Grt/1 (R, H2SO4)	4.081	2.476	0.9961	0.511318	0.000004	6.69 ± 0.81	−26.4
Grt/2 (R, H2SO4)	4.291	2.606	0.9951	0.511324	0.000004	7.92 ± 0.89	−26.4

Note: age results set in parenthesis are considered as geochronologically meaningless (see text).

Fig. 4

(a) Pegmatitic dyke (A) in mylonitic Wangtu orthogneiss (B), transected by a brittle to ductile right-lateral fault (C). Lower hemisphere equal area projection shows sub-vertical pegmatite (solid line) cross-cutting E-W trending foliation planes (dashed lines). Arrows indicate orientation of slickenlines; sampling locality of CH01-9, central Sutlej valley. View is towards N. Lon/Lat coordinates: 78.0351913/31.56244934. (b) Pegmatitic dyke in mylonitic Wangtu orthogneiss close to sample location described in (a). Note deflection of foliation planes towards the contact indicating slight rotation of the dyke after emplacement; sampling locality of CH01-8, central Sutlej valley. View is towards W. Lon/Lat coordinates: 78.03620517/31.56458974.

Fig. 5

(a) Example of synkinematic mineral growth: extreme syn-metamorphic “rotation of garnet” in mica schist (plane polarised light, sample T366; 1340 m a.s.l., NW of Kullu town). Length of picture is 5 mm. (b) Late crystallization of idiomorphic staurolite in mica schist, partly discordantly overgrowing inclusion trails of opaque minerals. Near Khoksar/Rohtang Pass. Length of picture is 6.7 mm. (c) Xenoliths (“schollen”) of paragneiss in unfoliated, Early Palaeozoic meta-granite, HHC Kullu–Malana (Deo Tibba, western flank) area. Note that different scales are used for parts a–c of the Figure.

Sample location 1: Chandarkhani Dhar/Malana valley (samples T96, T394, T93).

Sample location 2: Manikaran/Bajhoni Dhar, Parbati valley (T278, T419, MANG5).

Sample location 3: Spiti-Sutlej junction (HB25-00, HB28-00), Spiti valley (CH01-36, CH01-37), Leo Pargil site.

Sample location 4: Sutlej valley near Wangtu (CH01-8, CH01-9).

Sample location 5: Sutlej gorge near Recong Peo (09T13KIN; 5a) and Baspa valley, SE Karcham (CH01-11; 5b).

In the following, petrographic and (micro-)structural features of the main sampling sites analysed in this study are summarised.

Malana–Manikaran (Parbati valley) (sample localities 1, 2)

East of Beas river in Kullu area, in the Malana–Parbati valleys, an inverse metamorphic sequence is exposed as the basal part of the HHC slab, concordantly overlying the thin, variegated and strongly deformed “Bajaura nappe” that forms a distinct element between the low-grade metamorphic Lesser Himalayan Berinag quartzites of the LKRW in the footwall and the HHC in the hanging wall. (Frank et al., 1973, Frank et al., 1977, Frank et al., 1995, Thöni, 1977) (Fig. 1). The lower c. 3–5 km of the HHC profile is composed of biotite ± garnet-bearing schist and paragneiss, with a few intercalations of augen-gneiss and leucocratic gneiss, and topped by an extensive body of Early Palaeozoic granites/granitic gneisses. Metamorphic grade within the HHC increases upwards from biotite–garnet to staurolite–kyanite-grade, reaching sillimanite grade in the middle Parbati valley. The main foliation dips towards NE, a few amphibolite-facies grade shear sense indicators suggest top-to-the SSW–SW thrusting. Garnet growth is predominantly synkinematic with respect to overall ductile deformation, though post-kinematic idiomorphic rims and late retrogression (chloritization) of both garnet and staurolite has been observed locally and regionally (Fig. 5a and b). There is no evidence for late-stage faulting disrupting the inverse metamorphic sequence.

Spiti–Sutlej junction, Leo Pargil area (sample locality 3)

Above the South Tibetan Detachment (here locally called Sangla Detachment) W of the village of Morang (Fig. 1b), the base of the HU, which in parts reaches amphibolite facies conditions, is intruded by the Ordovician Kinner Kailash granite (Marquer et al., 2000, Tripathi et al., 2012). Between the Kinner Kailash granite and the confluence of the Spiti river with the Sutlej the HU forms a regional scale syncline. Metamorphic grade in the HU decreases rapidly up-section from kyanite–staurolite–garnet schists at the base to lower greenschist facies in the centre of the syncline, where sedimentary structures in the HU are preserved (Chambers et al., 2009). East of the Spiti river, north of the confluence with the Sutlej, the Leo Pargil dome (Fig. 1) exposes high grade metamorphic rocks intruded by chaotically-oriented discordant leucogranite veins. The dome forms a marked morphologic rise, extending in a SW–NE direction (i.e., perpendicular to the strike of the orogen) up to the main suture zone (IYSZ) in the Indus valley; it is flanked by both ductile and brittle detachments, but no detachment has yet been traced south across the Sutlej valley. Unmetamorphosed Tethyan sedimentary rocks lie in the hanging wall of the detachment that exhumed the Leo Pargil dome (Thiede et al., 2006). In sections to the west they clearly unconformably overlie the HU (Wiesmayr and Grasemann, 2002).

The dominant fabric in the HU is a penetrative crenulation cleavage marked by muscovite and biotite and mesoscopic folds with NW–SE striking fold axes, which overprint an earlier fabric oriented parallel to the partly preserved bedding, indicating NE–SW crustal shortening. At higher metamorphic grade, garnet, kyanite and staurolite crystallized during or at a late stage of folding (Chambers et al., 2009). Pegmatitic dykes associated with the intrusion of the Leo Pargil pluton clearly postdate the main fabric but are affected by ptygmatic folding or pinch-and-swell stretching depending on their spatial orientation.

Wangtu–Sangla (LHC-MCT-HHC connection) (sample localities 4, 5)

Within the southeastern LKRW, the LHC is emplaced along the Munsiari Thrust over the low-grade metasedimentary rocks of the Lesser Himalayan Sedimentary Succession (MT in Fig. 1). The LHC is composed of the amphibolite-grade Jutogh or Jeori metasediments and the 1.85 Ga Wangtu orthogneiss (Vannay et al., 2004, Richards et al., 2005). Although there is a gradual transition between these two units some authors suggest a tectonic contact (Jain et al., 2000). The LHC records an inverted metamorphic field gradient from garnet to sillimanite, which has been interpreted to be the result of top-to-the SW thrusting kinematics associated with general shear extrusion (Vannay et al., 2004). Whereas the main foliation generally dips towards NE, the stretching lineation gradually changes its orientation from W-plunging in the centre of the LHC to NE plunging at the top and the bottom of the wedge.

The contact between the LHC and the HHC is marked by a thrust, considered as the equivalent of the MCT (Fig. 1). Consistent amphibolite facies shear sense indicators at the structurally lower part of the HHC record top-to-the SW thrusting. The contact is overprinted by a zone of lower-greenschist to cataclastic deformation, which is localised with normal-sense offset in graphitic schists (Karcham Normal Fault). Within the HHC, the metamorphic field gradient increases from garnet grade at the base to sillimanite grade and migmatites at the top. The shear sense associated with a generally E-dipping foliation is also marked by a NE plunging stretching lineation that changes from top-to-the SW at the base to top-to-the NE on the top of the HHC (Vannay et al., 2004).

Sample description and mineral chemistry

Petrographic features, whole rock and mineral compositions are summarised in Table 1, Table 2, Table 3, geographical locations are also given in Table 1. Two samples, leucogranite HB28/00 (locality 3) and pegmatite CH01-9 (locality 4), are described in more detail below.

Leucogranite HB28/00 (UTM 276241E, 3519880N) (Leo Pargil)

This sample contains K-feldspar, plagioclase, quartz, muscovite, biotite, garnet, tourmaline and accessory apatite, monazite, zircon and xenotime. In addition, a few scattered grains of bismuth were observed. The K-feldspar is anhedral with an average composition of Or90Ab9.9An0.1. The analysed plagioclase is Na-rich and contains less than 1.8 wt% CaO + K2O. Both feldspars are slightly altered. Muscovite is Fe-rich and Ti-poor, with approximately 0.20 total Fe + Mg + Ti atoms per 11 oxygen anhydrous formula. In comparison with ideal muscovite, it is poor in Al, with 2.7 atoms per formula unit, and contains excess Si (3.1 apfu). Biotite is much less abundant than muscovite. In composition, it has high Fe/Mg ratios and high Al2O3 contents, plotting in the field of peraluminous magmatic biotite (Abdel-Rahman, 1994). Tourmaline is Mg-rich schorl.

Garnets are subhedral, measure up to 1100 μm in diameter and contain abundant inclusions of quartz in addition to muscovite and apatite. BSE images and EDS revealed the presence of micro-inclusions of zircon, apatite and xenotime. Analysed garnets are almandine–spessartine rich, with minor pyrope and grossular contents (Alm56–70Spess40–25Gross0.5–2Py1.4–1.8) and show a growth zoning with decreasing MnO and increasing FeO from core towards the rim (Fig. 2a).

Fig. 2

(a) Mn X-ray elemental map of garnet in Leo Pargil leucogranite HB28/00, illustrating decreasing MnO from core to rim. Darker shades of grey indicate higher element concentrations. The garnet contains inclusions of quartz (black). (b) BSE image of garnet in Wangtu pegmatite CH01-9. The garnet is characterised by a high Fe–Mn core domain and micro-inclusions of zircon. (c) BSE image of garnet aggregate in sample T410, characterised by partly euhedral rims and inclusions of Qtz, Pl, Ms, beryl and micro-inclusions of apatite (see BSE image below), Fe-sulphide and uraninite. The garnet is Mn-rich and slightly zoned with increasing Ca, Fe and Mg and decreasing Mn from interior parts towards rims. Interior compositions are in the range Alm39.0–40.3Prp1.6–2.6Grs0.7–2.8Sps54.2–57.9 and rim compositions are Alm40.5–42.6Prp3.2–3.8Grs1.4–5.5Sps50.2–52.0. (d) X-ray elemental map of garnet in sample T96 illustrating an inhomogeneous distribution of Ca. Lighter shades of grey indicate higher Ca concentrations. Inclusions are apatite (white), quartz, biotite and muscovite (black).

Analysed garnet grains (Table 2) have low or very low contents of LREE, Rb, Cs, Ba, Sr, V, Cr, Co, Ta, Th, Zr, Hf, with moderate concentrations of Sc (22–48 ppm), Li (105–181 ppm) and Ti (110–247 ppm). Y contents are high at 452–1327 ppm and coupled with high HREE contents (YbN = 205–1426). Sm/Nd ratios range between 7.8 and 12. Chondrite-normalised REE patterns (Fig. 3a) show strong LREE/HREE fractionation (YbN/NdN = 267–6867) and negative Eu anomalies (Eu/Eu∗ = 0.04–0.05).

Fig. 3

Chondrite-normalised (Boynton, 1984) REE patterns of garnet in (a) Leo Pargil leucogranite HB28/00 and (b) Wangtu pegmatite CH01-9. Note the pronounced negative Eu-anomaly in all analysed grains and the core (black squares and circles) vs. rim (open squares and circles) zonation.

Pegmatite CH01-9 (UTM 218977E 3496352 N) (LHC Wangtu)

The photograph of the sampled outcrop in Fig. 4a shows pegmatitic dykes that crosscut the main foliation of the Wangtu orthogneiss, which is itself transected by late brittle-ductile faults. The pegmatite consists of K-feldspar, plagioclase, quartz, muscovite, garnet and accessory apatite, monazite, zircon, xenotime and uraninite. K-feldspar has a composition of Or93–94Ab6–7 and is commonly altered. Plagioclase is albite (Ab92–93An6–7Or1) and frequently slightly altered. Garnet is euhedral with well-developed crystal faces and varies in size from 30 to 500 μm (Fig. 2b). The analysed garnet grains are essentially almandine–spessartine solid solutions in which the spessartine component ranges from 32 to 40 mol%. The largest garnet analysed is 450 μm in diameter with a core and rim composition of Alm57.3Spess40.2Gross1.8Py0.6 and Alm62.0Spess35.4Gross1.8Py0.8, respectively. Compositional maps always show a typical growth zoning with increasing XFe and decreasing XMn from core to rim. Some garnet grains contain two-phase fluid inclusions and backscattered electron (BSE) images reveal that micro-inclusions (<5 μm) of zircon, xenotime ± uraninite are common (Fig. 2b).

LA-ICP-MS analyses (Table 2) show that analysed garnets are rich in Y (506–1986 ppm) and HREE, but contain extremely low abundances of Cr (<1.1 ppm), Rb (<0.37 ppm), Sr (0.25–0.77 ppm), V (<0.04 ppm), Th (<0.09 ppm), Ba (<0.03 ppm), La (<0.005 ppm), Ce (<0.02 ppm), Pr (<0.03 ppm) and Eu (<0.08 ppm). Li and Ti contents are moderate, ranging from 54 to 139 ppm and from 122 to 277 ppm, respectively. Sc contents are 3–6 ppm, Zr is low at 7–14 ppm, and coupled with 0.4–0.8 ppm Hf. Chondrite-normalised rare earth element (REE) patterns (Fig. 3b) display pronounced enrichments of HREE over LREE (YbN/NdN = 267–6867) and distinct negative Eu anomalies of 0.003 to 0.01. HREE contents in garnet vary widely, with YbN ranging from 203 to 5028. Sm/Nd ratios are high (9.4–16.3).

Both mica fractions (A and E) have high FeO (3.35 ± 0.36 and 3.03 ± 0.34 wt%, respectively), minor MgO (0.41 ± 0.04 and 0.38 ± 0.04), TiO2 (0.15 ± 0.09 and 0.20 ± 0.11) and very low MnO contents (≈0.04 wt%). CaO is below detection limit, Na2O contents are low and range between 0.41 and 0.57 wt%. In comparison with ideal muscovite, both fractions are poor in Al, with 2.6 atoms per formula unit, and contain excess Si (3.2 apfu). Muscovite crystals are rich in Li, Rb, Cs, Ti, Nb and Ta, but have extremely low concentrations of V, Y, Th, U and REE, and minor Sc, Cr, Co, Sr, Zr, Ba, Hf and Pb (Table 2).

Whole-rock chemistry

The chemical composition of two leucogranites and four pegmatite samples from the Sutlej section is presented in Table 3. These rocks are leucocratic, with FeO + MgO + TiO2 ranging from 0.64 to 1.44 wt%. They are silica-rich (71.9–75.5 wt% SiO2) and strongly peraluminous, with ASI values [molar Al2O3/(CaO + Na2O + K2O) ratios] of 1.2–1.7, resulting in 1.9–5.7 wt% CIPW-normative corundum. Relative to most granites, they have high concentrations of Rb, ranging from 174 to 528 ppm. With 10–40 ppm Cs and 13–51 ppm Sn, samples HB25/00, HB28/00, CH01-36 and CH01-37 are notably enriched in these elements relative to upper crustal abundances. Conversely, Ti, V, Sc, Co, Ni, Cu, Sr, Y, Zr, Ba and light rare earth element (LREE) contents are low or very low relative to average granites (Table 3). The initial, i.e. emplacement-related, 87Sr/86Sr ratios based on mineral isochrons are highly variable, ranging from 0.726 to 1.21, and ε(t)Nd values for leucogranite HB28/00 and pegmatite CH01-9 are −14.3 and −26.4, respectively.

Geochronology

Rb–Sr data

Rb–Sr analyses have been performed on whole rock–biotite pairs (four samples from the Kullu–Parbati areas, five samples from the Sutlej transect) and whole rock–white mica pairs (two samples from the Parbati area, six samples from the Sutlej valley) to constrain the exhumation and cooling path. Rb–Sr analytical data and age results are listed in Table 4, the age results are plotted in T–t diagrams (Fig. 8).

Fig. 8

Temperature–time (T–t) plots for the present study and previously published data. (a) Kullu–Rohtang and Malana–Parbati area (sample locations 1 and 2). Data from: Frank et al., 1977, Mehta, 1977, Jain et al., 2000, Schlup et al., 2011, and this study; (b) HHC in the Sutlej section (sample locations 3 and 5). Data from: Jain et al., 2000, Vannay et al., 2004, Thiede et al., 2005, Chambers et al., 2008, and this study; (c) Lesser Himalayan Larji–Kullu–Rampur Window series (sample location 4). Data from: Jain et al., 2000, Vannay et al., 2004, Thiede et al., 2005, Caddick et al., 2007, Chambers et al., 2009, and this study.

Kullu–Malana–Parbati section

Biotite–whole rock ages

Two garnet–kyanite bearing samples (T93, T96) from the Malana area (between Beas and Parbati valleys) gave biotite–whole rock ages at 15.35 ± 0.15 and 16.7 ± 0.2 Ma, while augen gneiss T278 and garnet–kyanite gneiss MANG5 from near Manikaran (Fig. 1b; Bajhoni Dhar) have ages of 12.68 ± 0.12 and 13.29 ± 0.13 Ma (Table 4).

White mica–whole rock ages

Two white mica–whole rock ages from the Bajhoni Dhar profile near Manikaran/Parbati valley range at 20.5 ± 0.3 Ma (augen gneiss T278) and 23.9 ± 0.3 Ma (leucocratic gneiss T410), respectively (Table 4).

Sutlej section

Pegmatites cross-cutting the LHC Wangtu gneiss complex

Two pegmatites from the upper part of the Larji–Kullu–Rampur Window close to the MCT (CH01-8 and CH01-9) that cross-cut the main foliation of the Wangtu gneiss (see Fig. 4) gave white mica–whole rock ages of 5.5 ± 0.1 and 6.1 ± 0.4 Ma (0.16–0.3 mm grain size), respectively, at variable, but very high Rb/Sr ratios (Fig. 7e). A more strongly magnetic white mica fraction of pegmatite CH01-9, however, yielded a very young, but reproducible age at 1.3 ± 0.1 and 1.4 ± 0.1 Ma (CH01-9E Ms/2 and CH01-9E Ms/2 replicate in Table 4).

Fig. 7

Sm–Nd and Rb–Sr isochron plots for whole rocks, garnet and mica fractions: (a) pegmatite T394 (Malana); (b) leucocratic gneiss T410 (Manikaran); (c) leucogranite HB28/00 (Leo Pargil); (d) LHC pegmatite CH01-9 (Wangtu). Note the drastically different Nd initial isotope ratios (εNd) for the HHC-Tethyan Himalaya leucogranite source and the LHC pegmatite source in (c and d). (e) Muscovite–wr Rb–Sr plot for Wangtu pegmatites CH01-8 and -9. Symbol sizes partly exceed analytical errors. See text for discussion.

Biotite–whole rock ages, HHC

A HHC garnet–kyanite bearing schist (09T13KIN) from near Recong Peo gave a biotite–whole rock age of 14.56 ± 0.14 Ma. Biotite separated from an extension gash of the basal HHC (MCT) series between Karcham and Sangla yielded 10.3 ± 0.1 Ma for the vein material, and 11.2 ± 0.1 Ma for the biotite separated from the wall rock (amphibolite; CH01-11, Table 4).

HHC Leo Pargil leucogranite, and dykes

Two variously deformed leucogranites sampled at the Spiti–Sutlej confluence (HB25/00 and HB28/00) gave biotite–whole rock ages of 14.3 ± 0.1 and 11.6 ± 0.1 Ma, respectively, while three white mica fractions (muscovite) separated from these two samples resulted in whole rock–mineral ages of between 16.4 ± 0.2 and 15.0 ± 0.2 Ma (Table 4).

White mica (muscovite) from two dykes (CH01-36, CH01-37) of the Leo Pargil intrusive suite was also analysed. Interestingly, these rocks gave somewhat scattered and older ages at 17.2 ± 0.8, 18.3 ± 0.2, and 19.7 ± 2.0 Ma, at strongly varying Rb/Sr ratios and Sr concentrations (Table 4).

Sm–Nd data

The Sm–Nd analytical results listed in Table 5 include both metasedimentary and (meta-)igneous lithologies of variable metamorphic grade. Ten whole rock powders and twelve garnet fractions handpicked from five different (meta-)igneous samples (leucogranites, granite gneisses, pegmatites) were analysed.

Pegmatite sample T394 (W Chandarkhani Dhar; 3250 m a.s.l.; UTM Z43S: 711900E/3552600N)

Garnet grains are chemically homogeneous and virtually free of REE-rich inclusions, as also indicated by the very low Nd concentrations of 19 and 27 ppb, respectively, for the two handpicked fractions (Table 5). Calculated with the whole rock data point, two analysed garnet fractions yielded ages of 39.7 ± 2.5 (fraction 1) and 40.7 ± 1.5 Ma (fraction 2), at an initial εNd of −11.6 (Table 5). The corresponding three-point isochron result is 40.5 ± 1.3 Ma (Ndi = 0.511991 ± 6; MSWD = 0.45) (Fig. 7a).

Leucocratic layers in Bt-schist T96 (Chandarkhani Dhar, 3850 m a.s.l.; UTM Z43S: 712900E, 3553600N)

Garnet separated from cm-sized leucocratic (plagioclase-rich) layers in an isoclinally folded Bt-schist show distinct Ca–Fe zoning, with low Ca in the somewhat irregular core parts (Fig. 2d). A handpicked bulk garnet fraction resulted in a surprisingly old “age” of 160 ± 6 Ma.

Leucocratic gneiss T410, Manikaran (Bajhoni Dhar profile, 3420 m a.s.l.; UTM Z43S: 725200E/3547000N)

One garnet fraction handpicked from the bulk separate (0.15–0.42 mm sieve fraction) of this strongly foliated gneiss was leached with hot H2SO4 (Anczkiewicz and Thirlwall, 2003). The garnet is Mn-rich and slightly zoned, with increasing Ca from core to rim (Fig. 2c). Therefore, two additional garnet separates were handpicked from narrowly defined magnetic splits of the same bulk crushate to resolve eventual grain-internal age-zonation, but a leaching experiment was not applied. The resulting individual garnet–whole rock two-point isochron ages are identical within analytical errors for all three garnet separates, ranging between 29.6 ± 1.3 and 28.0 ± 1.5 Ma (Table 5), at similar 147Sm/144Nd ratios of between 1 and 1.12. The mean age is 28.71 ± 0.86 Ma (n = 4; MSWD = 1.8), at an initial Nd isotopic composition of Ndi = 0.511919 ± 5 (εNdt = −13.3) (Fig. 7b).

Leo Pargil leucogranite, near Spiti–Sutlej confluence (sample HB28/00) (UTM 276241E, 3519880N)

Two whole rock splits and five garnet fractions of sample HB28/00 were analysed (Table 5). As shown by the microprobe results and in thin section, the garnets are characterised by growth zoning (Fig. 2a) and the presence of variable amounts of REE-rich micro-inclusions. As expected, the Sm–Nd analytical results reveal a clear-cut impact of LREE-rich micro-inclusions (probably incompletely reset old monazite) on Nd concentration and Sm/Nd, with the effect of strongly increasing the individual apparent “age“ of the garnet (Table 5; data points labelled “Grt’s, n.l.” in Fig. 7c). Therefore, leaching experiments were applied to three carefully handpicked separates, which were thoroughly ground in an agate mortar after handpicking. The individual garnet–whole rock regression ages for the three leached fractions (Grt R1, Grt R2, and Grt R3 H2SO4 fine; Table 5) are identical within errors, ranging at 18.5 ± 1.0, 19.1 ± 0.9 Ma, and 19.3 ± 2.5 Ma, respectively, at variable Sm/Nd ratios. The mean garnet–whole rock regression age resulting from these three leached garnet fractions and the two whole rock splits (n = 5) of 18.82 ± 0.73 Ma corresponds to a best-fit isochron (MSWD = 0.9), with an initial Nd isotopic composition of Ndi = 0.511877 ± 4 (εNdt = −14.3) (Fig. 7c).

Garnet from pegmatite CH01-9, LHC near Wangtu, central Sutlej valley (UTM 218977E, 3496352 N)

The garnets are chemically homogeneous and idiomorphic, but may contain REE-rich micro-inclusions (apatite, xenotime). Therefore, two out of the three handpicked garnet separates were leached with hot H2SO4, whereas one separate was analysed without applying the leaching experiment. The effect of leaching is similar to that shown for the Leo Pargil leucogranite garnets. The apparent age for the unleached, Nd-rich, low-Sm/Nd fraction (garnet–whole rock: “14.3 ± 4.4 Ma”; Table 5) is obviously distorted, due to the effect of un-equilibrated inclusions in the garnet; therefore, this fraction was not included in the final regression calculation. The two H2SO4-leached separates (Grt/1 R and Grt/2 R, Table 5) yielded within-error identical, and very young garnet–whole rock ages of 7.92 ± 0.89 and 6.69 ± 0.81 Ma, respectively (Fig. 7d). The strongly negative initial Nd isotopic composition (εNdt = −26.4) indicates a considerably older source for the pegmatitic melt compared to most High Himalayan rocks. The primary magmatic age of the Wangtu gneiss in which the pegmatite intrudes is Palaeoproterozoic (c. 1.8–1.9 Ga; Kwatra et al., 1986, Richards et al., 2005). The mean garnet Sm–Nd age for two leached garnet fractions and three whole rock splits (Table 5) of sample CH01-9 is 7.24 ± 0.64 Ma (MSWD = 1.6; Ndi = 0.511273 ± 2; εNd = −26.4), and 7.24 ± 0.65 Ma (MSWD = 2.5; Ndi = 0.511273 ± 2; εNd = −26.4), for n = 4 (i.e., if wr 1, with the larger uncertainty, is excluded from the regression calculation) (Fig. 7d).

Discussion: timing of tectono-thermal events

Hints for pre-Himalayan relics

Conclusive evidence for pre-Himalayan tectono-metamorphic events in the HHC is scarce. The Cambrian/Ordovician unconformity at the base of the sediment succession in Spiti and the extensive magmatic activity “around 500 Ma” prove the existence of an “Early Palaeozoic event” in these rocks (Miller et al., 2001). This “Early Palaeozoic magmatic event” has been documented on a regional scale from the Himalaya (e.g., Debon et al., 1986, Girard and Bussy, 1999, with data compilation). Near Deo Tibba in the upper Kullu, xenoliths of paragneiss (pre-existing rocks incorporated during magma emplacement) were observed locally within undeformed parts of these early Palaeozoic granites (Fig. 5c). Marquer et al. (2000) presented evidence for pre-Himalayan high-grade metamorphism in the area surrounding the Lower Palaeozoic Kinner Kailash granite (central Sutlej valley). In addition, Argles et al. (1999) reported a single Sm–Nd date indicating the presence of pre-Himalayan garnet in a HHC leucogneiss from Garhwal (Alaknanda valley), while Catlos et al. (2002) described Cambro-Ordovician monazite preserved in garnet from the HHC of the Dudh Kosi–Everest transect in Nepal.

Our Sm–Nd results add new evidence for the existence of pre-Himalayan garnet in the HHC. Sample T96 (Table 5) was collected from the lower part of the HHC, some 1.5 km above the MCT, NW of Malana (Fig. 1). The unexpectedly old garnet–whole rock “age” of 160 ± 6 Ma from leucocratic layers in Bt-schist (T96) is probably due to Nd isotopic memory, caused by incompletely reset, pre-Himalayan relictic domains in garnet; the “age” could result from mixing of young, Himalayan garnet with pre-Himalayan, relict garnet (core) domains (Fig. 2d).

The strongly negative εNdt value of −18.3 also suggests involvement of fairly old source material that escaped complete Himalayan-age Nd isotopic re-homogenisation. It could even indicate involvement of sources related to Lesser, rather than higher Himalayan isotopic signatures. Some published data sets (Whittington et al., 1999, Robinson et al., 2001) indicate a clear contrast in εNd isotopic signatures between Higher Himalayan and Lesser Himalayan metasedimentary protolith sources, as the latter have much more negative values due to their higher mean age compared with Higher Himalayan source areas. The U–Pb zircon, Hf and Nd isotopic data of Richards et al. (2005) clearly demonstrate the existence of these two distinct geological terranes, LHC and HHC, in the Sutlej section, although the HHC “represents a mixture of several sources” in their provenance model (Richards et al., 2005).

Timing of the metamorphic peak and Himalayan “inverted metamorphism”

Inverted metamorphic sections are typical of the HHC along almost the entire Himalayan orogenic belt, and both the Beas and Sutlej valleys are classical examples of this intriguing feature (Frank et al., 1973, Vannay and Grasemann, 2001, Caddick et al., 2007). Jain and Manickavasagam (1993) and Manickavasagam et al. (1999) proposed “ductile shear displacements along ubiquitous, closely spaced S–C planes” within a major shear zone as the main mechanism for producing inverted metamorphic sequences. Vannay and Grasemann (1998) suggested diachronic equilibration at different temperatures, but rather constant pressure, to explain such “inverted metamorphic sections“. Their interpretation was based on field evidence, structural and petrological data in the Sutlej valley, where metamorphism increases upwards from biotite-garnet to sillimanite-K-feldspar grade (Vannay and Grasemann, 1998, Vannay and Grasemann, 2001). In order to explain the apparent inversion of the palaeo-isotherms but constant or decreasing peak metamorphic pressures recorded in the LHC and HHC slabs, a general shear kinematic model was used to explain extrusion and differential rotation of material lines during non-coaxial deformation (Grasemann et al., 1999, Vannay and Grasemann, 2001). Numerous other models like telescoping (Searle and Rex, 1989, Law et al., 2011), tectonic extrusion (Burchfiel et al., 1992), channel flow (Beaumont et al., 2001), underplating (Bollinger et al., 2006) have been proposed to explain the inverted metamorphism within the HHC. Although mechanically different, all these models have in common that two shear zones with opposite vorticity (Grujic et al., 1996, Grasemann et al., 1999) operated simultaneously, at least during the Early Miocene (Hodges et al., 1993).

Knowledge of the early evolution phases of burial and prograde metamorphism, i.e., for the time span of continental subduction and collision between c. 55 and 45 Ma in the HHC (e.g., De Sigoyer et al., 2000, Guillot et al., 2003, Leech et al., 2005), is limited. It is typically restricted to the HP-UHP metamorphic NHC nappes, in the vicinity of the main subduction zone. The oldest age in the present study from the HHC related to Himalayan orogeny is the garnet Sm–Nd age of pegmatite sample T394 (Chandarkhani Dhar). Based on composition and element zoning of the garnet, the mean isochron age of 40.5 ± 1.3 Ma (including 3 data points; Fig. 7a) suggests garnet crystallization during local decompression melt emplacement, arguably at an early stage of exhumation of the subducted Indian crust from depth. Interestingly, the result is similar to some Sm–Nd garnet ages reported from the Garhwal Himalaya (range: 40–35 Ma, and younger; Foster et al., 2000, Prince et al., 2000), and to Sm–Nd garnet and U–Th–Pb monazite ages from corresponding HHC metamorphic units of the Nanga Parbat–Haramosh Massif in the Pakistan Himalaya (range: 46–37 Ma; Foster et al., 2000) and in the Nepal Himalaya (45.2 ± 2.1 Ma; Catlos et al., 2002). Initiation of prograde metamorphism between c. 45 and 40 Ma was also inferred from fine-grained white mica Ar–Ar dating in the very low- to low-grade metamorphic Tethyan series of the nearby area of Spiti (Wiesmayr and Grasemann, 2002).

Available mineral ages from the inverted metamorphic sections in the Malana–Parbati valleys (e.g., Manikaran, Bajhoni Dhar profile) are too few to reliably calculate the difference in timing of isotopic closure across the profile. For the hanging wall part of the sampled section (c. 2.800–3.500 m above sea level), just below the massive slab of early Palaeozoic granites, peak metamorphic temperatures of 650 °C, or higher, are assumed (presence of beryl). Taking into account the mean age difference for garnet (Sm–Nd), white mica (Rb–Sr), and biotite (Rb–Sr), a mean cooling rate of c. 22 °C/Ma may be calculated for the time window between 28.7 and 12.7 Ma. The “mean” age of 28.71 ± 0.86 Ma (Fig. 7b) for garnet from leucocratic gneiss T410 is interpreted to trace garnet crystallization close to the metamorphic peak, where major synmetamorphic ductile deformation partly outlasted the final stages of garnet crystallization, in the course of exhumation of the HHC crystalline slab in late Oligocene time. This 29±1 Ma garnet age is similar to Sm–Nd garnet ages from the Suru area in Ladakh (33–26 Ma range; Vance and Harris, 1999). Considering the area of investigation, monazite ages that constrain garnet growth along the prograde P–T path to between >34 Ma and c. 28 Ma were reported from the base of the Haimanta Unit (HU) further to the southeast, in the upper Sutlej Valley (Chambers et al., 2009), (Fig. 1). By applying the closure temperature model to the geochronological data (mineral pairs) in Table 4, Table 5, the resulting exhumation rates for the Manikaran section are 1.05–0.82 mm/a for the T–t range of c. 650–500 (±50) °C/29–21 Ma (Grt–Ms), and 0.83–0.52 mm/a for the T–t range of c. 500–300 (±50) °C/21–12.5 Ma (Ms–Bt closure), and assuming a “normal” (30 °C/km) geothermal gradient. Since the dated higher retentive minerals (garnet, muscovite) are from a ductilely deformed, feldspar-rich assemblage, they document pervasive internal deformation of the extruding crystalline slab, at the time, when the early MCT was arguably active, and favouring a ductile extrusion model for the HHC (sensu Mancktelow, 1995, Grujic et al., 1996).

Exhumation-related deformation, post-peak melt emplacement, and constraints on cooling history from mica and FT ages

Fig. 8a–c summarises the age data from the present study and from the literature in a temperature vs. time (T–t) plot for three domains: the HHC of the Kullu–Rohtang and Malana–Parbati area in the W (a), the HHC in the central Sutlej section (b), and the LHC in its footwall, in the area around Wangtu–Karcham (c).

The following closure temperatures (Tc) have been used: Apatite FT: 130 ± 20 °C (Jain et al., 2000); zircon FT: 240 °C (Jain et al., 2000); biotite Rb–Sr: 300 ± 50 °C (Jäger et al., 1967); white mica Ar–Ar (K–Ar): 380 ± 50 °C (Purdy and Jäger, 1976); white mica Rb–Sr: 500 ± 50 °C (Jäger et al., 1967); garnet Sm–Nd: ⩾650 °C (Ganguly et al., 1998, Van Orman et al., 2002, Tirone et al., 2005); Mnz U–Pb: ⩾750 °C (Foster et al., 2002).

Assessment of leucogranite formation and the timing of pegmatitic melt emplacement in the Leo Pargil site (Kinnaur/Spiti–Guge border region)

Using the Y + Nb vs. Rb discrimination diagram of Pearce et al. (1984; Fig. 6), most data points of the analysed leucogranites and pegmatites (HB28/00 and CH01-9) plot in the field of syn-collisional granites. Their peraluminous nature, high Rb/Ba ratios (>0.63), high 87Sr/86Sr initial ratios, the presence of garnet, muscovite and monazite are all diagnostic of S-type granites. The crystallization of magmatic tourmaline in sample HB28/00 is also consistent with a metasedimentary source (Bénard et al., 1985).

Fig. 6

Rb vs. (Y+Nb) discrimination diagram for granites (Pearce et al., 1984) showing that Leo Pargil (LP) leucogranites, associated dykes and Wangtu pegmatites plot in the field of syn-collisional granites generated from partial melting of (meta-) sedimentary source rocks (syn-COLG). The grey field represents other HHC leucogranites (see text for references). VAG = volcanic arc granite; ORG = ocean-ridge granite; WPB = within-plate granite.

Low concentrations of TiO2 and Ba argue against extensive melting of biotite in the source, but the relatively high Rb concentrations suggest that generation of these granitic rocks involved incongruent dehydration melting of muscovite. In addition, the composition of the analysed samples compares well with published analyses for Himalayan leucogranites (e.g. Scaillet et al., 1990, Inger and Harris, 1993, Guillot and Le Fort, 1995, Searle et al., 1997) and melts produced by fluid-absent melting of muscovite-bearing starting materials (e.g. Patiño Douce and Harris, 1998, Acosta-Vigil et al., 2003). Fluid-absent melting of a pelitic source is also suggested by the observed high Rb/Sr ratios coupled with Sr/Ba ratios in the range of 0.4–0.7 since vapour-present conditions during melting of pelitic protoliths would result in enrichment of both Rb and Sr in the melt but depletion of Ba (Harris and Inger, 1992). High 87Sr/86Sr and low to very low 143Nd/144Nd initial ratios (Table 4, Table 5) indicate that these granitic and pegmatitic rocks crystallized from melts that were produced through partial fusion of compositionally heterogeneous crustal materials (cf. Ferrara et al., 1991). The εNd values of the leucogranites (−14, see Table 5) are, however, within the range of “typical” HHC metasedimentary sources. Hence, in contrast to the Wangtu pegmatite (εNd = −26) there is no argument for involvement of much older, LHC-type protolith material in these melts. Temperature estimates based on zircon solubility (Watson and Harrison, 1983) are in the range 662–677 °C for the Leo Pargil leucogranites and 580–652 °C for the Wangtu pegmatites.

Both, the Leo Pargil leucogranite and the Wangtu pegmatite garnets (see below) show the strong enrichment of HREE over LREE and pronounced negative Eu anomalies (Fig. 3) that are characteristic features of spessartine-rich garnets crystallized from silicic melts (e.g. Irving and Frey, 1978, Harris et al., 1992, Bea et al., 1994, Villaseca et al., 2003). In contrast, negative Eu anomalies are not apparent in metamorphic garnet formed by subsolidus reactions (Harris et al., 1992). Thus, REE patterns, low Ca- and high Mn-contents clearly argue against a restite origin of garnet, indicating a primary magmatic origin. Therefore, garnet can be used to constrain the magmatic crystallization age of these rocks.

The c. 19 ± 1 Ma Sm–Nd age of the Leo Pargil garnet (sample HB28/00; Fig. 7c) is interpreted as documenting the final steps of emplacement/the onset of regional cooling after intrusion of the leucogranitic melts. This age is similar to U–Th–Pb age results of other HHC/Tethyan leucogranites from areas adjoining further northwest and southeast (e.g., Zanskar, Kumaon, etc.; see Searle, 1996). Taking into account the mica closure ages, the cooling rate of the Leo Pargil is slower than that of many other Himalayan granites, e.g., in the Zanskar area (Gumburanjun leucogranite; Ferrara et al., 1991, Dèzes, 1999). It should be noted, however, that the dated garnet-bearing sample relates to the dyke swarms at the rim, not to the central part of the Leo Pargil intrusive complex. The entire intrusion history could well extend over a period of time exceeding 2 Myr as suggested by the fact that some of the dykes are deformed, while others are not, and by the somewhat scattered Rb–Sr muscovite ages (Table 4). This interpretation is in accordance with scattered zircon U–Pb and monazite U–Th–Pb data from the Leo Pargil dome (between c. 27 and 16 Ma; Leech and Sas, 2006, Languille et al., 2011).

Rock chemistry and garnet REE composition, among others, characterise the Leo Pargil leucogranite as crustal melt. The spatial and chronological relationship of extensional deformation and rapid exhumation/decompression for anatectic melt generation and emplacement has been addressed for other leucogranite occurrences in the NW Himalaya (e.g., Dèzes, 1999). Our results agree well with the crystallization ages of most Himalayan leucogranites that were emplaced within the HHC over an along strike distance of >1000 km in the time span 24–16 Ma (Deniel et al., 1987, Guillot and Le Fort, 1995, Noble and Searle, 1995, Searle et al., 1997, Harrison et al., 1997, Harrison et al., 1999, Tripathi et al., 2012), but can be as young as 13–12 Ma (data compilation in Searle, 2007).

Late Miocene pegmatite melts vs. Proterozoic protoliths in the Lesser Himalayan Wangtu gneiss, central Sutlej valley

For the LHC (Wangtu) pegmatite CH01-9, the 2σ-errors of the individual garnet–whole rock–whole rock Sm–Nd regressions overlap, at ages of 6.69 ± 0.81 Ma and 7.92 ± 0.89 Ma (Table 5; Fig. 7d-e). Inclusion of four data points (two garnet and two whole rock fractions) in a single regression yields a date that statistically does not quite fulfil the criteria of an isochron sensu stricto, but is close to it (MSWD = 2.5, for n = 4). The resulting “mean garnet age” is 7.24 ± 0.65 Ma (Ndi = 0.511273 ± 2; εNd = −26.4). The strongly negative initial Nd isotopic composition indicates a very high crustal residence time for the protolith of the melt. The somewhat limited fit of the data points (MSWD = 2.5; or, alternatively, MSWD = 1.6 for 5 data points, i.e., if wr 1 with the higher uncertainty is included in the regression calculation), i.e., the difference in age for the two individual fractions, could indicate either a time span >1 Myr for garnet crystallization (i.e., grain internal age zonation; in this case, the fairly precise “mean” age would result from statistical underestimation of the time span of mineral growth), or alternatively, reflect the effect of some remaining LREE-inclusions in the leached garnets. The latter mechanism is suggested from comparing the LA-ICP-MS REE-concentration data (Table 2) vs. TIMS results on handpicked grains in Table 5. Therefore, the Sm–Nd age result in Fig. 7d must be interpreted with caution. However, given the similar Sm and Nd concentrations and, especially, the identical Sm/Nd ratios for the handpicked, leached garnet fractions in Table 5, any remaining LREE-rich micro-inclusions are expected to be rather homogeneously distributed. Most importantly, given the huge difference in time between pegmatite melt generation and emplacement (c. 9–6 Ma, see Table 4, Table 5) and the age of the country rock (c. 1.8–2.0 Ga; Kwatra et al., 1986, Miller et al., 2000, with references; Richards et al., 2005) as a possible source of the melt, a significant contribution from old un-equilibrated micro-inclusions (such as xenotime and apatite) that could falsify the age of the garnet, can obviously be excluded. Therefore, the two individual garnet–whole rock Sm–Nd ages of 6.69 ± 0.81 Ma and 7.92 ± 0.89 Ma (Table 5, Fig. 7d) should define the time window for pegmatite melt emplacement and garnet crystallization of pegmatite CH01-9. Including the error bars, the whole time window derived from the Sm–Nd isotopes ranges between 8.8 and 5.9 Ma, being well in line with the Rb–Sr (cooling) ages of 6.1 ± 0.4 and 5.5 ± 0.1 Ma obtained on muscovite of CH01-9 and another nearby pegmatite (CH01-8), respectively (Table 4; Figs. 4b and 7e). These data have important implications on the regional thermal and deformational history, since (i) the pegmatites were intruded while mylonitic ductile deformation in the Wangtu gneiss was still effective (Hager, 2003, p. 59), and (ii) the pegmatites themselves were transected by s-parallel ductile to brittle shears (Fig. 4). The Rb–Sr muscovite ages indicate that the rock pile cooled through the 500 ± 50 °C blocking temperature between c. 6.6 and 5.2 Ma B.P.

Correlation of geochronology and tectonic processes: cooling and exhumation rates

Together with the studies of Frank et al., 1977, Jain et al., 2000, Vannay et al., 2004, Thiede et al., 2005, Thiede et al., 2006, Caddick et al., 2007, Chambers et al., 2008, Chambers et al., 2009, and Schlup et al. (2011), the results from the present study (Fig. 8a–c) show that white mica K–Ar (Ar–Ar) and biotite Rb–Sr cooling ages from the HHC between the upper Kullu valley–Rohtang area and the Malana–Parbati valleys are in the range c. 17–11 Ma (Table 4, Fig. 8a). Data from the Sutlej valley generally fit this trend or are slightly younger (14–10 Ma), including the cooling data from the Leo Pargil leucogranite (14–11 Ma for biotite; Table 4; see also Thiede et al., 2006). Biotite from the Karcham Normal Fault, close to the MCT, gave an age of 10.3 ± 0.1 Ma, while undeformed Bt from a late extension gash in amphibolite yielded a Rb–Sr age of 11.2 ± 0.1 Ma (Table 4), possibly indicating the waning stages of greenschist facies metamorphic conditions in these rocks, if a closure temperature of 300 ± 50 °C is applied.

Chambers et al. (2009) published in situ Ar–Ar data from Late Proterozoic schists of the HU from near Morang in the upper Sutlej valley; three fine-grained white mica samples yielded mean ages between 14.23 ± 0.73 and 10.4 ± 2.3 Ma. If compared to results of Vannay et al. (2004) from both the Tethys Himalayan Sequences (19.3–17.1 Ma) and the HHC (17.7–15.3 Ma), these somewhat younger ages can be explained by the much finer grain sizes used for analysis (Chambers et al., 2009).

In contrast to the Tethyan Himalaya and HHC, the mineral age spectrum in the footwall of the MCT, the LHC, is considerably younger (Fig. 8c). Caddick et al. (2007) and Chambers et al. (2008) published U–Pb data for monazite and uraninite between 11.1 ± 2 Ma and 6.4 ± 0.5 Ma from Lesser Himalayan Jutogh schist and leucogranite in the central Sutlej valley (near Sarahan; Fig. 1); the data were interpreted as crystallization ages, correlating with prograde to peak (amphibolite-facies grade) metamorphic conditions and reflecting “footwall heating due to rapid overthrusting of hot rock” (i.e., the HHC, onto the LHC). In the Wangtu area, which represents the strongly deformed hanging wall part of the LHC just below the HHC, the white mica Ar–Ar ages (Vannay et al., 2004, Thiede et al., 2005) are constrained – towards the high T side – by the new Rb–Sr and Sm–Nd data of the pegmatite samples described above (Fig. 7d-e). These pegmatites crosscut the ductile deformation structures of the Wangtu gneiss and are, in turn, transected by younger, localised brittle fault zones (see Hager, 2003, and Fig. 4a). White mica Ar–Ar ages from this area range between 6.7 and 4.3 Ma (see Thiede et al., 2005, for data compilation). A few zircon FT ages (Jain et al., 2000) cover the range c. 5–2 Ma, while FT ages for apatite plot in the time window c. 2–0.5 Ma (Fig. 8c). A single apatite FT age (1.9 ± 0.9 Ma) from the western part of the LKRW (near Sainj/Larji) fits into this youngest age group (Schlup et al., 2011).

When comparing the cooling paths from the different areas shown in Fig. 8, the following mean cooling rates can be calculated: (a) Manali–Rohtang area: c. 10 ± 3 °C/Ma for the time window <40 Ma, and Malana–Parbati area: 20 ± 3 °C/Ma for the time c. 30–10 Ma (upper and lower curves in Fig. 8a); (b) HHC Sutlej valley: 25–30 °C/Ma for the time <20 Ma (Fig. 8b); (c) LHC central Sutlej valley (Wangtu–Sangla): 50–100 °C/Ma for the time window al. (2007) and Chambers et al. (2008) argue for a c. 11–9 Ma metamorphic peak in the LHC Jutogh Series of central Sutlej section, based on U–Pb data on monazite and uraninite. For the footwall part of the HHC at Wangtu, muscovite documents cooling below 500 ± 50 °C in the time window 6.1–5.5 Ma. Clearly, the latter area shows the fastest exhumation during very young geologic times, resulting in differential uplift in this part of the Himalayan orogen, as postulated by Vannay et al. (2004) and Thiede et al. (2005). Notable Miocene to Pleistocene exhumation was also documented by apatite and zircon FT data for the Dhauladar–Chamba–Gianbul section, to the northwest of Kullu (Deeken et al., 2011), and for the Kumaon Himalaya to the east of Sutlej (Patel et al., 2011).

Regional considerations: sedimentological and tectonic processes in the Sutlej traverse and uplift of the Himalaya

The Sutlej is one of the few “Transhimalayan” rivers that cut through all tectonic units of the Himalaya, thus providing an excellent possibility to study the interaction of tectonothermal and erosional processes (Vannay et al., 2004, Thiede et al., 2005). Its headwater area, east of Mount Leo Pargil, is characterised by the formation of the most extensive inter-montane basin fill of the Himalayan orogen, the so-called Zanda (or Zhada) basin, in Western Tibet (Fig. 1). The basin formed at a time, when the STD as the principal “roof fault” (Searle, 2007) had demonstrably become inactive (e.g., Leloup et al., 2010).

The Zanda basin fill consists of a 800–900 m thick fluviatile–lacustrine sedimentary succession (gravels, sands, marlstones), which formed by a huge delta and lake system. Its present day extent measures more than 200 km in NW–SE direction, from near Mt. Leo Pargil in eastern Spiti to Lagna Tso (Lake Rakshas), the source of the Sutlej, near Mt. Gurla Mandhata (Fig. 1). The Zanda sediments have recently been studied by different working groups (Meng et al., 2008, Wang et al., 2008, Kempf et al., 2009, Saylor et al., 2009, Zhu et al., 2010). Their age range is given as c. 9/7–1 Ma (Kempf et al., 2009), or c. 6–0.8 Ma (Meng et al., 2008).

Considering the growth history of Tibet (Gloaguen and Ratschbacher, 2011), some recent publications argue that the plateau sensu stricto, i.e. to the north of the Indus-Tsangpo (Yarlung) suture zone (IYSZ, Fig. 1), had reached an elevation of 4 km by c. 35 Ma (end of the Eocene), and remained at this height since then (e.g., Hetzel et al., 2011). The Zanda basin is located to the south of the main suture zone (IYSZ, Fig. 1), within the sedimentary belt of the Tethyan Himalaya. Some authors relate its formation with the activity on the Karakoram fault bordering the Aylari Range to the NE of the basin (Valli et al., 2007, Wang et al., 2008), but a clear-cut tectonic model for formation of the basin does not exist. E–W extension-related faulting is generally active in southern Tibet (and the central Himalaya), producing N–S striking graben structures since Late Miocene to Pliocene time (e.g., Blisniuk et al., 2001, Ratschbacher et al., 2011). Hintersberger et al. (2010) relate E–W extension in the NW-Himalaya to “propagation of extension driven by the collapse of the Tibetan Plateau.” Specifically, Late Cenozoic faulting and updoming is constrained both for the Leo Pargil site (Thiede et al., 2006) and the Gurla Mandhata “detachment system” (Murphy et al., 2002) which border the Zanda basin sediments at their western and eastern termination (Fig. 1). These higher-grade metamorphic areas show a differentiated cooling history: the fast exhumation of these “domes” relative to near-by low-grade crustal sections is explained by orogen-parallel normal-faulting in an overall contractional orogenic setting (Murphy et al., 2002, Thiede et al., 2006). In the Gurla Mandhata detachment system (Fig. 1), Murphy et al. (2002) claim garnet crystallization for the time window c. 17(±1)–11(±1) Ma, based on Th–Pb ages of monazite inclusions in garnet, white mica and biotite Ar/Ar ages down to c. 7 Ma for the footwall of the detachment, and Th–Pb ages for magmatic monazite from Gurla Mandhata leucogranites as young as 7 Ma. For the Leo Pargil “dome”, Thiede et al. (2006) documented a three-stage cooling history since c. 16/14 Ma (Ar/Ar and FT data), with final pronounced accelerated cooling/exhumation rates at c. 4–3 Ma. These data may help to explain the sedimentary evolution in the Zanda basin situated between these two metamorphic “domes” (Meng et al., 2008, Kempf et al., 2009). It may be argued that tectonic processes essentially controlled formation and development of the basin, especially in the initial stages, and that control on formation of the Zanda basin was exerted by uplift of both the Gurla Mandhata and Leo Pargil domes flanking the basin to the east and west, respectively, since c. 9 Ma (Murphy et al., 2002, Thiede et al., 2006).

A crucial point in the reconstruction of the uplift history in the Sutlej section is the interpretation of paleo-elevations/paleo-temperatures. Based on magnetostratigraphic, palaeontological and sedimentological evidence from the Zanda basin sedimentary sequence, Meng et al. (2008) argued that climatic conditions in this area changed from warm and wet (at >5 Ma) to cool at 2.7 Ma, reflecting a steady, but variable uplift of the southern Tibetan plateau from 1000 to c. 3000 m above sea level. These climatic changes are explained by the successive blocking of the warm and wet Indian summer monsoon, due to the uprise of the Himalayan mountains from c. 1000 m (at 5.3 Ma) to >3500 m asl at c. 2.7 Ma, when a still faster uplift of the southern Tibetan plateau began (Meng et al., 2008). Repeated climatic changes and fast uprise of the Himalaya, leading to rapid young tectonic uplift of southern Tibet, were also deduced from the 7–1.7 Ma old fluvio-lacustrine sediment succession of the Gyirong Basin, east of Gurla Mandhata (Wang et al., 1996, Hong et al., 2010), and the main uplift stage for the northwestern Tibetan plateau is regarded as being post-Miocene (younger than c. 5 Ma) (Zheng et al., 2000). In contrast, carbon and oxygen isotope studies from the Zanda basin sediments led to different conclusions (Saylor et al., 2009). Though the data support a 2-phase, both pre- and post-mid-Miocene uplift of the southern Tibetan Plateau, the authors argue that the Sutlej headwater area was at high altitude (“at least as high as today”) already by 9 Ma B.P., when deposition of the Zanda sediments began (Saylor et al., 2009; see also DeCelles et al., 2011). This latter view agrees with other studies supporting an early (Eocene to Oligocene) high elevation of the greater Tibetan Plateau (Currie et al., 2005, Van der Beek et al., 2009, Hetzel et al., 2011).

Deposition of the Zanda sediments correlates with a tectonically very active period in the Himalayan region. Other inter-montane sedimentary basins, such as the Kathmandu basin and the Kashmir basin, also developed during these late evolutionary stages as tectonic depressions on the margins of the Greater Himalayan orogenic belt (Burbank and Johnson, 1982, Fujii and Sakai, 2002). Ratschbacher et al. (2011) emphasise that “the currently active neotectonic deformation” in Tibet started c. 5 Ma ago. An intensified phase of young (post-Miocene) tectonic activity related to the MCT, both regionally (cf. Catlos et al., 2002) as well as for more restricted areas (i.e. the Himachal Himalayas), was recently suggested, based on seismic reflection data and neodymium isotope results in the Indus delta, the present estuary of the Sutlej. This has been explained by a tectonically driven reorganization of the major river systems in the central northwestern Himalayas of Himachal Pradesh, including the Sutlej, Beas and Ravi, after 5 Ma (Clift and Blusztajn, 2005). Even if a definite interpretation of the analytical data concerning the paleoelevation/paleoclimate history in SW-Tibet is not yet possible, and the faunal and depositional characteristics of the sediments of the Zanda and Gyirong basins do not clearly document the climatic evolution over the post-10 Ma time interval, it should be stressed that, for several millions of years, the situation at the southern coast of the Zanda lake must have been such that the draining river (the Sutlej) “… was never able to deeply incise into the bedrock of the outflow region” (Kempf et al., 2009). This would imply a differentiated uprise and erosion history of this part of the Himalayan orogenic belt since Late Miocene/Pliocene, and possibly lasting into Pleistocene time (Zhu et al., 2010), just as deduced from the cooling history of deeper buried, metamorphic rocks from more interior parts of the orogen, which experienced – and are still actively driven by – fast exhumation in very young and recent geologic time.

Conclusions

The petrological and structural results and the new geochronological data of this study corroborate literature data; they also add important evidence in favour of a differentiated thermal and exhumation/cooling history in the Kullu–Sutlej area of Himachal Pradesh.

Sm–Nd data from garnet bearing pegmatoids signal local decompression melting in the HHC of Malana/Kullu at c. 40 ± 2 Ma, indicating regional pressure release during initial exhumation of crystalline from depth.

Ongoing orogenic compression at amphibolite-facies grade led to synkinematic growth of metamorphic minerals over major parts of the HHC. Slightly zoned garnet from muscovite-garnet gneiss of the basal parts of the inverse metamorphic HHC slab near Manikaran (Parbati valley) crystallized at 29 ± 1 Ma, close to the metamorphic peak (c. 650 °C) and probably at a time, when the MCT as a distinct master fault began to evolve at the base of the HHC. This data may also set a time marker for regional cooling in this area.

Garnet Sm–Nd and mica (Ms, Bt) Rb–Sr ages from the Leo Pargil leucogranite and related dykes in the Kinnaur-Spiti-Guge border region constrain the final crystallization and cooling history (c. 650–500–300 °C) of this extended crustal re-melting event at between 20 and 14/12 Ma.

The cooling history in the southern/central main units of the HHC in Himachal is constrained by Rb–Sr muscovite ages of between 30 and 20 Ma (Rohtang, Parbati area), Ar–Ar white mica ages of 19–14 Ma (central Sutlej section), and biotite Rb–Sr ages of 17–13 Ma, with a tendency of younger ages to cluster towards the MCT. Zrn and FT ages (literature data) are mostly between 15 and 5 Ma, while Ap FT ages are generally younger than 5 Ma.

Pegmatite veins cross-cutting the Palaeoproterozoic Wangtu gneiss (central Sutlej valley) which has been ductilely deformed during the later stages of Neo-Himalayan tectono-thermal overprinting, are themselves crosscut by ductile-brittle faults. Sm–Nd (garnet) and Rb–Sr (muscovite) ages between 7.9 ± 0.9 and 5.5 ± 0.3 Ma from these pegmatites set a lower time limit for ductile deformation at c. 6 Ma B.P. Cooling from c. 650 °C to surface temperature in these rocks was accomplished within some 7–6 Myr, resulting in very fast cooling rates of 70–100 °C/Myr. These values are in line with FT ages for Zrn and Ap (previously published data) from the wider area of the LHC around Wangtu, which are in the range of 4–2 Ma (Zrn) and 2–0.6 Ma (Ap).