The timing of geologic events along the India-Asia suture in southern Tibet remains poorly understood because minimal denudation prevents widespread exposure of structurally deep rocks. In this study, we present geologic maps of two structurally deep domes, cored by mylonitic orthogneisses, across the India-Asia suture zone in southwestern Tibet. New U-Pb zircon ages and rock textures indicate that core orthogneisses are originally Gangdese arc rocks that experienced Late Eocene prograde metamorphism, probably during crustal thickening. Crosscutting leucogranite sills underwent northwest-southeast extension related to slip along a brittle-ductile shear zone here designated the Ayi Shan detachment. The timing of shear along detachment is bracketed by zircon U-Pb ages of 26–32 Ma for these pre- to syn(?)-extensional leucogranites, and by a 40Ar/39Ar muscovite age of 18.10 ± 0.05 Ma for a rhyolitic dike. This rhyolite dike crosscuts a widespread siliciclastic unit that was deposited across the detachment, which we correlate to the Kailas Formation. The Great Counter thrust defines the surface trace of the India-Asia suture zone; it cuts the Kailas Formation, and is in turn cut by the Karakoram fault. A new 40Ar/39Ar muscovite age of 10.17 ± 0.04 Ma for the Karakoram fault footwall is consistent with published thermochronologic data that indicate Late Miocene transtension in southwestern Tibet.
Intercontinental collision between the Indian subcontinent and central Asia resulted in widespread Cenozoic crustal thickening and surface uplift that extends nearly 2000 km from northern India to Kazakhstan, a region encompassing over 7,000,000 km2. This continent-scale event played an important role in the evolution of Asian river systems (Brookfield, 1998) and global climate change (Ruddiman and Kutzbach, 1989; Molnar et al., 1993; Quade et al., 1995), and served as the testing ground for a wide variety of geodynamic models of orogenic processes (Houseman and England, 1996; Royden, 1996; England and Molnar, 1997; Flesch et al., 2001; Beaumont et al., 2004; Bendick and Flesch, 2007). Because of the potential feedback between collision and other physical phenomena, it is critical to constrain the timing of deformational events.
The initial collision between India and Asia is estimated to have occurred during the Paleocene–Early Eocene (Besse et al., 1984; Gaetani and Garzanti; 1991; Rowley, 1996; Zhu et al., 2005), although some suggest it occurred earlier (latest Cretaceous–Paleocene) (Yin and Harrison, 2000; Ding et al., 2005) and even others have argued for a much later time (ca. 34 Ma) (Aitchison et al., 2007). The suture between India and Asia juxtaposes the Cretaceous–Tertiary Gangdese arc, which formed along the southern margin of Asia due to northward subduction of Tethyan oceanic lithosphere, against the Tethyan sedimentary sequence (TSS) which represents material deposited along the former passive margin of the Indian subcontinent. Our present understanding of the history of the Gangdese arc indicates that it did not experience collision-related deformation until at least 30 Ma (Yin et al., 1994; Harrison et al., 2000), some 25 Myr after the most widely accepted initiation estimates of intercontinental collision (Zhu et al., 2005). To the north of the Indus-Yalu suture zone (IYS) (also referred to as the Indus-Tsangpo suture and Yarlung-Tsangpo suture) crustal shortening and related basin development was ongoing since 40 Ma in northeastern Tibet in the Qilian shan and Fenghuo shan-Nangqian thrust belt (Zhang and Zheng, 1994; Yin and Harrison, 2000; Horton et al., 2004) and since 28 Ma in the Eastern Kunlun ranges (Yin et al., 2008). To the south of the suture, U-Pb ages of zircons from intrusive rocks that crosscut folded TSS strata indicate that significant crustal shortening within the Tethyan fold-thrust belt occurred prior to the mid-Eocene (Aikman et al., 2008). These observations imply that the Gangdese arc was capable of resisting deformation or transmitting collisional related stresses while crustal shortening to its north and south was ongoing. This is at odds with Middle to Late Eocene ages of plutonic rocks of the Gangdese arc (Honegger et al., 1982; Xu et al., 1985; Harrison et al., 2000; Wen et al., 2008), which suggest that the arc was hot and therefore weak during the early stages of the collision. Observations that cast further doubt on the inference that the Gangdese arc escaped early collision-related deformation are the increased sedimentation rates and development of a collisional foredeep on the adjacent passive margin of the Indian subcontinent at this time (Rowley, 1996; Ding et al., 2005; Zhu et al., 2005). These features imply the presence of a large crustal load to its north along the southern margin of Asia. England and Searle (1986) and Kapp et al. (2007) show that a retroarc thrust belt spanning southern Tibet developed north of the Gangdese arc between 105 and 53 Ma. Their studies indicate that the thrust belt likely roots into the Gangdese arc and therefore predict that it was significantly thickened immediately prior to the initial collision. This raises the possibility that the arc did not escape deformation, but rather transmitted the stress elsewhere due to its thickened state.
To better understand the geologic history of the Gangdese arc and the IYS, we have undertaken a field and geochronologic study of well-exposed orthogneisses exhumed from deep structural levels along the IYS in southwestern Tibet. Our results indicate that the protolith of these orthogneisses are Gangdese arc rocks that experienced Late Eocene prograde metamorphism, which we attribute to burial via crustal thickening. This event is followed by the development of a brittle-ductile extensional shear zone involving Gangdese arc rocks.
GEOLOGY OF THE AYI SHAN
The trace of the IYS trends subparallel to the crest of the Ayi Shan (shan means mountain range in Mandarin) in southwestern Tibet (Fig. 1). Cretaceous–Tertiary granites that represent the Gangdese arc are exposed along the eastern flank of the Ayi Shan whereas rocks of the TSS crop out along the western flank. Our geologic mapping (Fig. 1) shows that the geologic framework of the Ayi Shan consists of three primary structural features (Figs. 2 and 3). They are, from oldest to youngest, the (1) Ayi Shan detachment, (2) the Great Counter thrust, and (3) the Karakoram fault system.
Ayi Shan Detachment
Two northwest-southeast elongated, doubly plunging structural domes exist in the southern and northern portions of the Ayi Shan and are mantled by a brittle-ductile shear zone, which we refer to as the Ayi Shan detachment (Figs. 1–3). The Ayi Shan detachment juxtaposes porphyritic granite in its upper plate (hanging wall) against metamorphic rocks in its lower plate (footwall). Although the shear zones in the southern dome and northern dome are possibly different structures, we refer to them collectively as the Ayi Shan detachment based on their structural similarities (Figs. 4A and 4B). Valleys crossing the Ayi Shan expose ∼1500 m of lower plate (footwall) rocks (present-day thickness). Immediately below the detachment, the lower plate locally contains 10–20 m of chloritic, quartzofeldspathic breccia. Exposures of the chloritic breccia are common on the eastern side of the southern Ayi Shan and rare on the western side. Beneath the chloritic breccias, where present, the lower plate metamorphic rocks can be separated into three lithologically distinct units. Immediately below the chloritic breccia is a sequence of mylonitic biotite schist that reaches a thickness of 600 m. Primary and secondary foliations (S, C, and C′ foliations) within the schist are defined by biotite and recrystallized quartz, whereas the mineral stretching lineation is defined by smeared quartz grains and aligned clots of muscovite. Locally, discrete top-to-southeast shear zones exist within the top 100 m of the mylonitic schist. In the southern dome, high-angle normal faults cutting the upper plate locally sole into the Ayi Shan detachment, implying they are kinematically linked with the detachment (Fig. 4C). Below the schist is an ∼400-m-thick sequence of mylonitic orthogneiss (Fig. 4D). The foliation within the orthogneiss is defined by alternating hornblende + biotite rich layers and feldspar + quartz rich layers, whereas the mineral stretching lineation is defined by the aligned biotite clots as well as quartz and feldspar grains. The structurally deepest rocks observed in the northern dome are quartzofeldspathic migmatitic gneiss (Fig. 2), which is at least 500 m thick.
The characteristic mineral assemblage of the schist exposed in the footwall of the Ayi Shan detachment is biotite + plagioclase + muscovite + rutile + garnet, which corresponds to amphibolite facies metamorphic conditions. This mineral assemblage is present throughout the footwall from the highest to the lowest exposed level. Garnets do not preserve an internal fabric and therefore no evidence for rotation during growth. The abundant ductilely deformed feldspar porphyroclasts suggest that deformation was occurring at temperatures of 400–500 °C (Tullis and Yund, 1991), which likely corresponds to mid-crustal depths. Feldspar porphyroclasts range in size from 1 to 3 cm in the long dimension. The orientation of the mineral stretching lineation defined by biotite clots, quartz, and feldspar in the footwall rocks indicates the mean slip direction along the Ayi Shan detachment is S60°E.
The mylonitic foliation in the footwall rocks is folded at all scales. The most abundant folds are those with axes oriented subparallel to the stretching lineation, which we refer to as corrugations. First- and second-order corrugations are defined by the surface trace of the Ayi Shan detachment and also by the foliation within its lower plate in northern and southern domes. The corrugations trend toward the southeast (parallel to the Ayi Shan) (Fig. 2) and are relatively symmetric (Fig. 3). The age of these corrugations is not clear. If the corrugations are syn-kinematic with shearing along the Ayi Shan detachment, they likely formed in a constrictional strain field. This is consistent with the local presence of L-tectonites in the southern Ayi Shan dome. Some amount of folding must have occurred after slip along the Ayi Shan detachment since it is folded (Fig. 3).
Variably deformed leucogranite bodies make up ∼5%–10% of the footwall of the Ayi Shan detachment (Figs. 4E and 4F). Sills are more pervasive than dikes. Sills and dikes range between tens of centimeters up to 3 m thick. Generally, they display straight contacts with the country rock. Leucogranite dikes typically cut the mylonitic foliation and leucogranite sills at deep structural levels but are deformed at higher structural levels and transposed parallel to the mylonitic foliation. We interpret this relationship to indicate that emplacement of leucogranite bodies broadly occurred during shearing within the footwall of the Ayi Shan detachment.
Great Counter Thrust
The southwest-dipping Great Counter thrust (GCT) defines the surface trace of the IYS. It juxtaposes the TSS in its hanging wall against Cretaceous–Tertiary granite and Tertiary siliciclastic rocks in its footwall (Figs. 1 and 2). The TSS in the Ayi Shan region ranges in age from Ordovician to Cretaceous and consists of a wide variety of sedimentary and low-grade metasedimentary rocks (Cheng and Xu, 1987). The GCT strikes northwest-southeast, subparallel to the Ayi Shan. Shear sense indicators show that it accommodates top-to-northeast motion. Exposed in the immediate hanging wall of the GCT are fault-bounded lenses of highly fractured metabasalt, metagabbro, and peridotite, which locally contain lenses of podiform chromite that are elongated parallel to the trace of the GCT (Fig. 2B).
The footwall of the GCT typically consists of pebble, cobble, and boulder conglomerate. This rock sequence is correlated to the Kailas Formation based on its composition and structural position (Cheng and Xu, 1987; Liu, 1988; Murphy et al., 2000; Murphy et al., 2009). The Kailas Formation in the Ayi Shan lies unconformably on Cretaceous–Tertiary granite and the lower plate of the Ayi Shan detachment, indicating that it postdates movement along the Ayi Shan detachment (Fig. 2A).
Karakoram Fault System
The eastern margin of the Ayi Shan coincides with the trace of the Karakoram fault system (Figs. 1–3). Deformation along the Karakoram fault occurs within a narrow zone (2–20 km wide) consisting of right-lateral faults, normal faults, and right-lateral ductile shear zones that strike northwest. Fault-slip data show that the Karakoram fault accommodates right-lateral displacement with a minor normal dip-slip component (Ratschbacher et al., 1994; Murphy et al., 2000; Kapp et al., 2003). The Karakoram fault cuts, and therefore must be younger than, the GCT in the Ayi Shan. Thermochronologic studies along the Karakoram fault in the vicinity of the Ayi Shan show a rapid cooling event in its footwall (south side) at ca. 10 Ma, which is interpreted to have resulted from slip along the Karakoram fault at that time (Arnaud, 1992). However, Valli et al. (2007) suggest it was continuously active since the Early Miocene, based on U-Pb zircon ages and petrofabric analyses of rocks exposed on its south side (footwall). The Karakoram fault is presently active as it offsets Quaternary surficial deposits (Armijo et al., 1989; Chevelier et al., 2005; Brown et al., 2002; Murphy and Burgess, 2006; Sanchez et al., 2010).
Uranium-lead zircon geochronology of granitic and gneissic rocks collected from the southern and northern gneiss domes was undertaken to (1) test the possible genetic link between the rocks making up the core of the domes with Cretaceous–Tertiary granites exposed to the east within the Gangdese batholith, and (2) assess the age of metamorphism. Seven orthogneiss samples were analyzed from the southern dome (AY-1 –AY-7), three orthogneiss samples (AY-8–AY-10) were analyzed from the northern dome, and two mylonitized leucogranite sills (AY-11 and AY-12) were analyzed from the northern dome (Figs. 1, 5, and 6). Sample locations are shown on Figures 2A and 2B. We obtained cathodoluminescence images of most zircons after performing U-Pb analyses to evaluate zoning patterns and internal structure with respect to the location of the U-Pb analyses. U-Pb isotopic data is available in the Supplemental Table File1.
U-Pb geochronology of zircons was conducted by laser ablation–multicollector– inductively coupled plasma–mass spectrometry (LA-MC-ICP-MS) at the Arizona LaserChron Center. Zircons were extracted from the rock samples with a standard mineral separation process. Handpicked zircon grains were mounted in epoxy and then polished to provide flat exposure of the interiors of the grains. The analyses involved ablation of zircon with a New Wave/Lambda Physik DUV193 excimer laser (operating at a wavelength of 193 nm) using a spot diameter of 15–35 μm. The ablated material is carried with helium gas into the plasma source of a GV Instruments IsoProbe, which is equipped with a flight tube of sufficient width that U, Th, and Pb isotopes are measured simultaneously. All measurements are made in static mode, using Faraday detectors for 238U and 232Th, an ion-counting channel for 204Pb, and faraday collectors for 208–206Pb. Each analysis consists of one 12-s integration on peaks with the laser off (for backgrounds), 12 one-second integrations with the laser firing, and a 30 s delay to purge the previous sample and prepare for the next analysis. Common Pb corrections are accomplished by using the measured 204Pb and assuming an initial Pb composition from Stacey and Kramers (1975). Measurement of 204Pb is unaffected by the presence of 204Hg because backgrounds are measured on peaks (thereby subtracting any background 204Hg and 204Pb), and because very little Hg was present in the argon gas during this analytical session. In-run analysis of fragments of a large zircon crystal (generally every fifth measurement) with a known age of 564 ± 4 Ma (2σ error) is used to correct for instrumental fractionation. The ages reported on each sample include both random and systematic errors associated with uncertainties in common Pb composition, age of the standard, and decay constant, and all final uncertainties are reported at the 2σ level.
Samples AY-1 through AY-7 are orthogneiss samples collected from the footwall of Ayi Shan detachment in the southern Ayi Shan range. Cathodoluminescence (CL) imaging of zircon grains from AY-1 show mottled and irregular CL patterns (Fig. 7). A few grains are rounded. Omitting three outlier spots, individual 206Pb/238U ages range from 43.9 ± 0.9 Ma to 47.4 ± 0.9 Ma. U/Th ratios range from 0.4 to 7.5 (Table 1 in the Supplemental Table File [see footnote 1], Fig. 5). The weighted mean of all 206Pb/238U ages is 45.53 ± 0.53 Ma (n = 31). We interpret this age as magmatic age due to the low U/Th ratios for all the zircon grains (Vavra et al., 1999).
Zircon grains from AY-2 are euhedral to subhedral. The individual spot 206Pb/238U ages range from 57.9 ± 0.6 Ma to 67.0 ± 0.8 Ma. U/Th ratios range from 0.7 to 4.2 (Table 2 in the Supplemental Table File [see footnote 1]). CL imaging of zircon grains shows oscillatory zonation consistent with igneous crystallization (Fig. 7). The weighted mean of all the ages is 61.78 ± 0.92 Ma (n = 21) (Fig. 5).
Most of the zircon grains extracted from AY-3 are subhedral. Individual spot 206Pb/238U ages range from 62.5 ± 10.1 Ma to 81.9 ± 7.7 Ma. U/Th ratios range from 1.1 to 8.9 (Table 3 in the Supplemental Table File [see footnote 1]). CL imaging of AY-3 zircon grains does not show typical oscillatory zonation and distinct rim-core domains as expected for igneous zircons. The weighted mean age is 73.9 ± 1.4 Ma (n = 18).
Zircon grains from AY-4 and AY-5 are euhedral and CL imaging of zircon grains shows igneous cores (low U/Th) and metamorphic rims (high U/Th) (Fig. 7C) (Vavra et al., 1999; Rubatto, 2002). The measured 206Pb/238U ages of individual spots range from 33.7 ± 0.8 Ma to 484.8 ± 4.7 Ma and 37.3 ± 0.4 Ma to 540.0 ± 9.0 Ma for samples AY-4 and AY-5, respectively; U/Th ratios range from 1.2 to 458.4 and 0.9 to 252.8, respectively (Tables 4 and 5 in the Supplemental Table File [see footnote 1]). The U-Pb isotope data of each sample plotted on concordia diagrams (Fig. 5) define chords with upper and lower intercept ages. These are 504 ± 46 and 42 ± 31 Ma for sample AY-4 and 558 ± 27 and 43 ± 47 for sample AY-5, respectively. The weighted average age of laser spots that yielded U/Th ratios in excess of 20 is 38 ± 1.3 Ma (n = 12) for sample AY-4 and 37.43 ± 0.87 Ma for sample AY-5, and we interpret these to reflect the age of zircon overgrowth.
The zircon grains from AY-6 are dominantly euhedral to subhedral with a few that are rounded. CL imaging shows that most zircons display zoning patterns that are consistent with igneous derivation (Fig. 7D). The measured 206Pb/238U ages of individual spots range from 43.8 ± 1.5 Ma to 49.9 ± 0.8 Ma and U/Th ratios range from 0.7 to 2.7 (Table 6 in the Supplemental Table File [see footnote 1], Fig. 5). The weighted average age is 48.47 ± 0.58 (n = 30). Based on the U/Th ratio and zoning patterns in CL, we interpret the weighted average age to reflect the age of the igneous protolith.
Zircons from AY-7 are mostly subhedral with a few of them being rounded. CL imaging shows that some of the zircon grains have oscillatory zonations with distinct rim-core domains. Individual spot 206Pb/238U ages, range from 30.4 ± 2.4 Ma to 455.3 ± 16.6 Ma, and U/Th ratios range from 1.6 to 111.8 (Table 7 in the Supplemental Table File [see footnote 1]). The U-Pb isotope data of each sample spot plotted on concordia diagrams (Fig. 5) define chords with upper and lower intercept ages. These are 489 ± 28 Ma and 42 ± 18 Ma, respectively. The weighted average age of laser spots that yielded U/Th ratios in excess of 20 is 31.55 ± 0.69 Ma (n = 10) and we interpret these to reflect the age of zircon overgrowth.
AY-8 and AY-10 are orthogneiss samples taken from the footwall of the Ayi Shan detachment in the northern Ayi Shan. The zircon grains from both samples are dominantly euhedral to subhedral with a few that are rounded. CL imaging of AY-8 shows that most zircons display zoning patterns that are consistent with igneous derivation (Fig. 7E). The measured 206Pb/238U ages of individual laser spots range from 36.1 ± 1.9 Ma to 44.8 ± 3.0 Ma and U/Th ratios range from 0.5 to 3.2 (Table 8 in the Supplemental Table File [see footnote 1], Fig. 5). The weighted mean 206Pb/238U age is 38.74 ± 0.68 (n = 21). Based on the U/Th ratio and zoning patterns in CL, we interpret the weighted mean age to reflect the age of the igneous protolith.
Individual laser spot 206Pb/238U ages from AY-10 zircons range from 54.0 ± 4.2 Ma to 95.0 ± 8.6 Ma (n = 31). CL images of AY-10 zircon grains do not show oscillatory zonation and distinct rim-core domains (Fig. 7F). U/Th ratios range from 0.7 to 2.6 (Table 10 in the Supplemental Table File [see footnote 1]). The weighted mean 206Pb/238U age is 87.4 ± 1.7 Ma (n = 31) (Fig. 5).
Mylonitic Leocogranite Sills
Samples AY-11 and AY-12 are samples of mylonitized leucogranite from the footwall of the Ayi Shan detachment in the northern Ayi Shan. Metamorphic foliation cuts across the contact between leucogranite and country rock schist and gneiss. Individual laser spot 206Pb/238U ages for samples AY-11 and AY-12 range from 32.2 ± 3.0 Ma to 788.8 ± 14.8 Ma (n = 26) and 25.7 ± 1.4 Ma to 1782 ± 89 Ma (n = 19), respectively. U/Th ratios range from 0.7 to 62.3 and 3.8 to 386, respectively (Table 11 in the Supplemental Table File [see footnote 1]). The weighted average 206Pb/238U ages of laser spots that yielded U/Th ratios in excess of 20 are 34.8 ± 1.7 Ma (n = 7) for sample AY-11 and 26.9 ± 1.1 Ma for sample AY-12 (n = 10) (Fig. 5) and we interpret these to reflect the ages of zircon overgrowth.
40Ar/39Ar thermochronology was conducted on muscovite and biotite from rock samples from the southern Ayi Shan gneiss dome. Samples were irradiated for 7 h along with the standard Fish Canyon Tuff sanidine (FC-2) with an estimated age of 28.02 Ma (Renne et al., 1998) at the U.S. Geological Survey (USGS) TRIGA Reactor in Denver, Colorado. Age spectrum analyses were conducted by step-heating mineral separates within a double vacuum Mo resistance furnace at the New Mexico Geochronology Research Laboratory, New Mexico Institute of Mining and Technology. Samples were targeted that could bracket the timing of geologic events that postdate movement along the Ayi Shan detachment. One sample (AY-14) comes from a dike that cuts the Kailas Formation, and the other sample (AY-15) comes from the south side of the Karakoram fault in its footwall (Fig. 2B).
AY-14 is from a 25-cm-thick rhyolitic dike that cuts across sandstone and conglomerate beds in the Kailas Formation adjacent to the southern Ayi Shan gneiss dome. 40Ar/39Ar age spectra from muscovite are flat for the majority of gas released and yield an age of 18.10 ± 0.05 Ma (Fig. 8). AY-15 is from a granitoid in the hanging wall of the Ayi Shan detachment ∼6 km west of the Karakoram fault. 40Ar/39Ar age spectra from muscovite are flat over the majority of gas released and yield an age of 10.17 ± 0.04 Ma (Fig. 8). Because both age spectra are flat we interpret both to be recording rapid cooling of both rock samples.
Protolith of Lower Plate Orthogneisses
A key element in the architecture of the IYS in southwestern Tibet is the lower plate of the Ayi Shan detachment. These rocks lie beneath the surface trace of the suture and therefore provide information on the suture zone at depth. Lower plate samples AY-1, AY-2, and AY-3 from the southern dome yield U-Pb zircon ages with low U/Th values of 45.53 ± 0.53 Ma, 61.78 ± 0.92 Ma, and 76.1 ± 1.4 Ma, respectively. Similarly, lower plate samples AY-8 and AY-10 from the northern dome yield low U/Th ratios with 206Pb/238U ages of 38.74 ± 0.68 Ma and 87.4 ± 1.7 Ma, respectively. U-Pb zircon ages of upper plate granites collected from the southern Ayi Shan range from 47 to 50 Ma and have low U/Th ratios indicating that the zircons are recording crystallization ages (Wang et al., 2009). These ages, along with regional studies on the age of Gangdese arc magmatism (Honegger et al., 1982; Xu et al., 1985; Harrison et al., 2000; Wen et al., 2008) lie within the range of our geochronologic results suggesting that the protolith of the Ayi Shan lower plate orthogneisses is the Gangdese arc. This interpretation is supported by the observation that both the upper plate granite and the lower plate orthogneisses contain distinctive large orthoclase phenocrysts and porphyroclasts, respectively.
Age of Prograde Metamorphism
Orthogneiss samples AY-4, AY-5, and AY-7 contain zircons that yield old ages in their cores and younger ages along their rims. Laser ablation analyses define chords on U-Pb Concordia diagrams and suggest that each sample contains zircons that grew in two stages and/or suffered Pb loss. The older, mostly Paleozoic 206Pb/238U ages define a chord and have low U/Th ratios. The younger rims (Fig. 7) are generally Late Eocene–Early Oligocene and have high U/Th ratios, which we interpret to be consistent with the growth of young zircon around an older core. The weighted average 206Pb/238U ages of the laser spots that yield U/Th ratios > 20, which we use to discriminate between mixed spot ages and those primarily from the rim, are 38 ± 1.3 Ma for AY-4, 37.43 ± 0.87 Ma for AY-5, and 31.55 ± 0.69 Ma for AY-7. We interpret the age of zircon overgrowths in samples AY-4, AY-5, and AY-7 to have occurred during prograde metamorphism likely associated with burial and crustal shortening (Fig. 6).
Because structurally deep exposures of Gangdese arc and associated country rocks are rare, the regional extent of this metamorphic event is uncertain. However, sandstone petrology of Paleocene to Early Eocene clastic rocks in the Tethyan Himalaya of southern Tibet (86°42′E) indicate a transition from a continental block provenance, interpreted as Indian craton, to recycled orogen provenance, interpreted to be a Gangdese arc-trench system (Zhu et al., 2005). This implies that the Gangdese arc was a developing topographic high that was being eroded in the Eocene. We propose that this uplift reflects isostatic uplift due to crustal thickening and associated horizontal shortening.
The age of the Late Eocene–Early Oligocene metamorphic event that affected the Gangdese arc is coeval with Eohimalayan metamorphism (Hodges et al., 1988; Pêcher, 1989; Vannay and Hodges, 1996). Eocene–Oligocene crustal thickening within the Tethyan Himalaya (Godin et al., 1999; Godin et al., 2001; Lee et al., 2000; Aoya et al., 2005; Kellett and Godin, 2009; Aikman et al., 2008) is widely thought to have induced prograde metamorphism and anatectic melting within Greater Himalayan rocks (Godin et al., 2006; Aoya et al., 2005; Lee and Whitehouse, 2007; Larson et al., 2010). Our interpretation of Middle Eocene to Early Oligocene crustal thickening of the Gangdese arc implies that the orogenic wedge at this time extended from the Tethyan Himalaya in the south, northward across the India-Asia suture zone to the Gangdese batholith.
Early to Late Oligocene Extension
Shear sense indicators show that the orthogneisses have undergone southeast-northwest–directed stretching with dominantly top-to-southeast displacement of the upper plate with respect to the lower plate. These data indicate the orthogneisses originated from underneath the upper plate granite to the southeast and were subsequently incorporated into the structurally shallow portions of the suture zone by way of shearing along the Ayi Shan detachment. This interpretation predicts that the Gangdese arc to the east of the Ayi Shan (near Mount Kailas) was vertically thinned. U-Pb ages of zircon from stretched mylonitic leucogranite sills that intrude the orthogneisses indicate vertical thinning occurred in the Early to Late Oligocene. This result is in agreement with previously reported geochronologic data from the Ayi Shan othogneisses and mylonitic leucogranite sills that southeast shear was ongoing during the Late Oligocene (Lacassin et al., 2004; Valli et al., 2007, 2008). These previous studies investigated field relationships and rocks exposed along the eastern margin of the Ayi Shan range along the Karakoram fault zone (Fig. 1). Because these studies were limited to the eastern side of the Ayi Shan, this unfortunately led to poorly constrained interpretations of the timing of geologic events and fault system geometry (compare figure 3 in Valli et al.  to Figure 3 in this paper). Valli et al. (2007, 2008) interpret that the metamorphic rocks and the igneous rocks which intrude them are a product of long-lived, continuous strike-slip deformation along the Karakoram fault. Valli et al. (2007) show steeply dipping faults bounding the Ayi Shan metamorphic core in the northern and southern domes and interpret that the Karakoram fault and the Great Counter thrust functioned together as a regional-scale positive flower structure (transpressional deformation). Our study extends across the entire range and shows that the metamorphic rocks are not localized along the Karakoram fault zone, but rather are restricted to the lower plate of the Ayi Shan detachment and are cut by the Karakoram fault zone. Fault slip data from the Great Counter thrust show that it facilitates NE-SW shortening along the length of the Ayi Shan (Murphy et al., 2009), rather than oblique strike-slip displacement as suggested by Lacassin et al. (2004). Strike-slip shear sense indicators along the Great Counter thrust used to support the interpretation by Lacassin et al. (2004) are located in the Mount Kailas area (southeast of the Ayi Shan) where the thrust trace coincides with that of the Karakoram fault. Our geologic mapping presented in this study along with fault-slip data presented in Murphy et al. (2009) indicate that oblique shear sense indicators along the thrust are not characteristic features, but instead, are local features present in the Mount Kailas area probably due to local strike-slip reactivation.
Deformation Cycles in Southwest Tibet
Taking into account the field relationships exposed across the entire range and geochronologic results presented here, we envision a more complex geologic history than previously suggested (Lacassin et al., 2004; Valli et al., 2007, 2008). Figures 9 and 10 illustrate our interpretation of two cycles of shortening (vertical crustal thickening) and extension (vertical crustal thinning) in southwestern Tibet since the Late Eocene. We envision that deformation represents that occurring at middle to shallow crustal depths; this is represented by four stages (Fig. 9) described below.
Stage A (40–31 Ma)
Burial metamorphism related to crustal shortening of the Gangdese arc closely followed behind or possibly overlaps in time with waning arc magmatism (Fig. 9A)(Kapp et al., 2007). Although the timing in southwest Tibet is not known, we envision this metamorphic event coincided with development of the Tethyan fold-thrust belt as suggested by data in southern Tibet (Aikman et al., 2008). In this scenario, it is possible that the Tethyan fold-thrust belt roots into the Gangdese arc, implying that the Tethyan fold-thrust belt and Gangdese arc were part of the same thrust wedge during the Eocene. An implication of this interpretation is that the suture zone in the upper and middle crust was translated southwards with respect to the suture zone in the lower crust and mantle lithosphere.
Stage B (32–26 Ma)
Displacement along the Ayi Shan detachment resulted in attenuation of the Gangdese arc, facilitated exhumation of metamorphosed Gangdese arc rocks, and juxtaposed deep arc rocks against shallow arc rocks (Fig. 9B). The kinematics of extension is approximately parallel to the arcuate-shape Himalayan orogen. Since the Kailas Formation is depositional on the lower plate of the Ayi Shan detachment, its age provides an upper bound on the timing of slip. The 40Ar/39Ar muscovite age of the rhyolitic dike (AY-14) that cuts the Kailas Formation indicates that the Kailas Formation is older than ca. 18 Ma. Therefore slip on the detachment must predate 18 Ma. This age constraint is consistent with the U-Pb detrital zircon ages from the Kailas Formation, which are interpreted to record its deposition between 26 and 24 Ma (DeCelles et al., 2011). DeCelles et al. (2011) argue on the basis of provenance and paleoflow data, strata relationships, and lithofacies patterns, that the Kailas Formation was not deposited in an overall contractional setting. Rather, they interpret the bulk of the formation to have been deposited in a rift or transtensional strike-slip basin. Their interpretation is consistent with our results, which indicate that the Kailas Formation was deposited on vertically thinned Gangdese arc rocks. Moreover, the sandstone petrology of rocks at the base of the formation indicates that it was derived from deeply eroded Gangdese arc rocks (DeCelles et al., 2011). This is also consistent with our results that indicate exhumation of Gangdese rocks in the Late Oligocene.
Stage C (18–15 Ma)
A second phase of shortening is evident by the development of the north-directed GCT (Fig. 9C). The Kailas Formation is cut by the GCT in the Ayi Shan (Figs. 1 and 2), in the Gangdese shan near Mount Kailas (Yin et al., 1999), and farther east near Lopukangri (Murphy et al., 2009, 2010). Although the relationship between the rhyolitic dike (sample AY-14) and the Great Counter thrust is not known, we did not observe any igneous dikes that cut the Great Counter thrust and therefore the fault may postdate intrusion at ca. 18 Ma. This age is consistent with other studies that interpret the thrust to have been active during the Early to Middle Miocene (Ratschbacher et al., 1994; Yin et al., 1999). Slip along the GCT results in northward translation of the suture zone at shallow structural levels with respect to its position deeper in the crust (Fig. 9C) (Ratschbacher et al., 1994; Yin et al., 1999; Murphy and Yin, 2003).
Stage D (15 Ma to Holocene)
The Great Counter thrust is cut by the Karakoram fault as well as by a large-magnitude extensional shear zone in the Leo Parghil range and Gurla Mandhata region (Fig. 9D; Thiede et al., 2006; Murphy et al., 2000, 2009, 2010; Murphy and Copeland, 2005). A K-feldspar from a leucogranite sample collected along the Karakoram fault system in the Zhaxigang area (Fig. 1) yields an age spectrum that indicates rocks on its south side (footwall) were rapidly cooled ca. 10 Ma (Arnaud, 1992). The 40Ar/39Ar muscovite age from sample AY-15, 10.17 ± 0.04 Ma, is consistent with this result and suggests that the Karakoram fault in the vicinity of the Ayi Shan was active during the Late Miocene. The Leo Parghil range is bounded by the Leo Parghil shear zone, a top-to-WNW extensional shear zone. Zhang et al. (2000) showed that its footwall on the SE side of the range contains ductile deformed garnet-bearing schists. Leucogranite bodies in the footwall locally contain an extensional shear fabric and yield a K-Ar age of 16–15 Ma. On the westside of the Leo Parghil range 40Ar/39Ar white mica ages from rocks in the footwall of the Leo Parghil shear zone of 16–14 Ma indicate a phase of rapid cooling possibly due to slip along the shear zone (Thiede et al., 2006). Southeast of the Ayi Shan is the Gurla Mandhata metamorphic core complex (Fig. 2) (Murphy et al., 2000; Murphy and Copeland, 2005). It is bounded by the top-to-the-west Gurla Mandhata detachment system, which is interpreted to be kinematically linked to the Karakoram fault. U-Pb zircon ages from extensionally sheared leucogranite bodies indicate that extension initiated at ca. 15 Ma. The timing of extension in this region is broadly the same as that documented farther east in south-central Tibet (Lee et al., 2011), and therefore we interpret this to be a regional event, rather than a local event associated with transtensional deformation along the Karakoram fault.
Implications for the Development of the Himalayan Orogenic Wedge
A current view on the growth of orogenic wedges describes their evolution as a function of the balance among processes responsible for energy accumulation (e.g., crustal thickening) and energy dissipation (e.g., lateral spreading and hinterland extension) (Hodges et al., 1996; Hodges, 2000). These processes compete to maintain a self-similar wedge geometry described by the taper angle (e.g., Dahlen, 1990). The first-order structural elements in the central Himalaya have recently been integrated by Larson et al. (2010) into a conceptual model describing orogenic wedge development characterized by two cycles of accumulation (increase in taper angle) and dissipation (decrease in taper angle) of gravitational potential energy (Fig. 9). Key stages of this evolution are as follows: (1) Eocene–Oligocene crustal thickening leading to an increase in the taper of the orogenic wedge (increase in gravitational potential energy); (2) Early Miocene foreland-directed lateral spreading resulting in a decrease in the taper (decrease in gravitational potential energy); (3) Middle Miocene hinterland thickening leading to a renewed buildup of the taper (increase in gravitational potential energy); and (4) Late Miocene to present lateral spreading instigating a decrease in the taper (decrease in gravitational potential energy). The deformation cycles described above and illustrated in Figure 9 are consistent with the interpretation that the taper of the Himalayan orogenic wedge has increased and decreased twice since the Late Eocene. Within the context of a critical taper model, we interpret the geology of the Ayi Shan as recording deformation in the hinterland of the Himalayan orogenic wedge (Fig. 9). In this scenario, Late Eocene–Early Oligocene crustal thickening within the Gangdese arc rocks recorded in the Ayi Shan would have resulted in surface uplift, thus increasing the taper of the Himalayan orogenic wedge. Vertical thinning of upper and middle crustal rocks in this region followed soon after crustal thickening and therefore would have assisted in localized subsidence, thereby facilitating a decrease in the taper angle. This event temporally overlaps with movement along the Main Central thrust zone and South Tibet detachment, which are interpreted to facilitate foreland-directed spreading of Greater Himalayan rocks (e.g., Hodges, 2000). Both hinterland vertical spreading and foreland-directed spreading are processes that assist in decreasing the taper of the orogenic wedge and therefore dissipate the gradient of the gravitational potential energy. Initiation of the Great Counter thrust, broadly bracketed to have occurred between 18 and 15 Ma, is interpreted to represent a phase of renewed crustal thickening and therefore surface uplift and an increase in the taper angle. This event temporally correlates to the development of the North Himalayan antiform, an ∼700-km-long shortening structure in the Himalayan hinterland (Lee et al., 2000, 2004; Godin et al., 2006; Larson et al., 2010). This implies that this renewed phase of hinterland thickening is a regional event. Initiation of the large-magnitude extensional and transtensional shear zones (Leo Parghil shear zone, Gurla Mandhata detachment system, Karakoram fault system) at ca. 15 Ma indicates that the Himalayan hinterland transitioned to a phase of vertical thinning. In the Himalayan foreland, age estimates of thrusting indicate that the locus of horizontal shortening and vertical thickening propagated toward the south (toward the foreland) (DeCelles et al., 2001). Foreland propagation of the thrust belt and vertical thinning in its hinterland are both processes that assist in decreasing the taper angle of the orogenic wedge.
The kinematics of both phases of hinterland extension (32–26 Ma [18 Ma?] and 15 Ma to present) are approximately parallel to the arcuate-shaped Himalayan thrust belt. One hypothesis explains that arc-parallel extension in the Himalaya is a result of outward radial growth (spreading) since the Middle Miocene (e.g., Murphy et al., 2009). We propose that this is also a viable explanation for Oligocene arc-parallel extension in the Ayi Shan and highlights the influence of the shape of the orogenic wedge in controlling the kinematics of orogenic wedges.
Geologic mapping combined with geochronologic studies of rocks along the IYS in southwest Tibet reveal a geologic history we interpret to result from two cycles of shortening and extension. Our primary results are as follows.
(1) The U-Pb zircon ages and textural observations indicate that the protolith of orthogneisses exposed in the Ayi Shan is Gangdese arc rocks and associated early Paleozoic country rock.
(2) U-Pb ages and isotope systematics of zircon rims from the orthogneisses indicate that these rocks experienced a Late Eocene–Early Oligocene prograde metamorphic event which we attribute to burial via crustal thickening/shortening.
(3) Geologic mapping shows that the orthogneisses and overlying mylonitic schist are mantled by a top-to-the-southeast brittle-ductile detachment that we refer to as the Ayi Shan detachment. Shear along the detachment is coeval with intrusion of leucocratic dikes and sills into its lower plate. U-Pb zircon ages from these igneous bodies indicate that extension occurred during the Early to Late Oligocene.
(4) Initiation of the Great Counter thrust, broadly bracketed to have occurred between 18 and 15 Ma, is interpreted to represent a phase of renewed crustal thickening along the IYS. This regional shortening structure is cut by a system of strike-slip and extensional shear zones (Leo Parghil shear zone, Karakoram fault, Gurla Mandhata detachment system) that are estimated to have initiated between 16 and 15 Ma and are presently active.
(5) Our results show that the crust in the vicinity of the IYS experienced two cycles of vertical thickening followed by vertical crustal thinning. Within the context of critical taper theory, vertical thickening and thinning in the Himalayan hinterland assist in increasing and decreasing the taper angle of the orogenic wedge, respectively. Moreover, these hinterland events can be linked to deformation patterns recognized in the Himalayan foreland and together support the idea that the evolution of the orogen can be explained, at least qualitatively, by critical taper models.
We thank Peter DeCelles, Paul Kapp, and Alex Robinson for valuable discussions regarding the regional geologic implications of our results. We also thank Victor Valencia, Alex Pullen, and Scott Johnston for their assistance with U-Pb analyses. We also thank Matt Kohn and Kyle Larson for very helpful reviews of an earlier version of this paper. This study was supported by a grant from the National Science Foundation (grant EAR 0438826 to Murphy).