In the eastern Himalaya (Bhutan), there are two distinct top-down-to-the-north segments of the South Tibetan detachment system. The outer segment is a diffuse ductile shear zone preserved as klippen in broad open synforms. New age constraints show that it was active until at least ca. 15.5 Ma and cooled by ca. 11.0 Ma, as constrained by sensitive high-resolution ion microprobe (SHRIMP) U-Pb geochronology of magmatic zircon and 40Ar/39Ar thermochronology of muscovite in ductilely deformed leucogranite sills. The inner segment is a ductile shear zone active at least until ca. 11.0 Ma (constrained by SHRIMP U-Pb geochronology of magmatic zircon) and overprinted by more recent brittle faulting. These age constraints indicate that ductile deformation continued on the South Tibetan detachment more recently in the eastern Himalaya than in central and western parts of the orogen. These improved constraints on timing of South Tibetan detachment segments allow for a more detailed reconstruction of continental collision in the eastern Himalaya in which the outer South Tibetan detachment segment was abandoned in the mid-Miocene and passively transported southward in the hanging wall of the Main Himalayan thrust (the basal detachment of the orogen), while top-to-the-north ductile to brittle shearing continued on the inner South Tibetan detachment segment. Hinterland stepping of the South Tibetan detachment to maintain an orogenic critical taper (frictional wedge model) is a possible mechanism for this tectonic reorganization of the South Tibetan detachment during the Miocene. However, our data combined with published geochronologic data for the eastern Himalaya demonstrate that foreland translation and exhumation of a midcrustal dome (viscous wedge model) is the more tenable mechanism.
Exposed midcrustal rocks in the Himalayan orogen are bounded by opposing-sense crustal-scale shear/fault systems, the Main Central thrust below and the normal-sense South Tibetan detachment system above (Fig. 1). The discovery that the timing of displacement along these two shear systems was coeval for at least part of the Miocene (Burchfiel et al., 1992; Hodges et al., 1992; see also review in Godin et al., 2006) has led to several geodynamic and kinematic models for continental collision in which midcrustal rocks were exhumed between coeval bounding shear zones (e.g., Grujic et al., 1996; Beaumont et al., 2001; Robinson et al., 2006; Webb et al., 2007). These models mark a major shift in the geological explanation for the tectonic evolution of large hot orogens such as the Himalaya.
In the eastern Himalaya, two geographically and structurally distinct segments of the South Tibetan detachment system, an outer, ductile shear zone preserved within broad synformal klippen and an inner ductile-brittle shear zone, have been identified (Edwards et al., 1996; Grujic et al., 2002; Hollister and Grujic, 2006) (Fig. 2). This South Tibetan detachment system geometry may be unique to the eastern Himalaya. While other klippen are found in the central (Kathmandu, Jajarkot, and Dadeldura) (Fig. 1A) and western (Simla) Himalaya, there is little agreement about whether the rocks preserved within are soled by South Tibetan detachment system top-to-the-north shearing, whether (in some cases) they are windows rather than klippen, or whether Tethyan strata are preserved (e.g., Upreti, 1999; Johnson et al., 2001; Robinson et al., 2001, 2003; Bollinger et al., 2006; Webb et al., 2007). In this paper, we constrain the timing of ductile shearing of the outer and inner segments of the South Tibetan detachment system in Bhutan using U-Pb and 40Ar/39Ar geochronological data. By clearly resolving the diachroneity of these two segments, we are able to propose mechanisms for their formation and preservation, and the consequent implications for geodynamic and kinematic models of the Himalayan orogen.
The Himalayan orogen is regarded as the epitome of a continent-continent collisional orogen. Following initiation of the collision in the early Eocene (Molnar, 1984; Rowley, 1998; de Sigoyer et al., 2000; Najman et al., 2001; Searle, 2001; Leech et al., 2005), the Eocene to late Oligocene was a period (Eohimalayan) of crustal thickening and Barrovian metamorphism in the midcrust. Evidence for crustal thickening is recorded by fold-and-thrust deformation in the upper crust (e.g., Ratschbacher et al., 1994; Aikman et al., 2008; Kellett and Godin, 2009), as well as kyanite-grade metamorphism in the middle crust (e.g., Simpson et al., 2000; Godin et al., 2001). This was followed in the early to mid-Miocene by the (Neohimalayan) exhumation of material from midcrustal depths to the southern Himalayan topographic front (Vannay and Hodges, 1996; Hodges et al., 1996). Exhumation was facilitated by vigorous surface erosion along the southern front of the orogen cooperating with active thrusting below and normal fault geometry tectonic denudation above the metamorphic rocks (see review in Yin, 2006).
The Himalayas are characterized by a series of southward-propagating, south-verging thrusts rooting into a common basal detachment thrust, the Main Himalayan thrust (e.g., Nelson et al., 1996). These thrusts bound the major lithotectonic units of the Himalaya (Fig. 1; see review in Le Fort, 1975). The northernmost lithotectonic assemblage is the Tethyan sedimentary sequence, bounded to the north by the Yarlung-Tsangpo suture, which divides Indian from Asian crust. Structurally beneath the Tethyan sedimentary sequence, there is the Greater Himalayan sequence of high-grade metamorphic rocks and Miocene leucogranites (Fig. 1). The boundary between the Tethyan sedimentary sequence and the Greater Himalayan sequence is the South Tibetan detachment system, a system of top-down-to-the-north normal-sense ductile shear zones and brittle faults (e.g., Burchfiel et al., 1992; Carosi et al., 1998; Searle et al., 2003). The dominantly sedimentary Lesser Himalayan sequence is separated from the Greater Himalayan sequence by the Main Central thrust. Farther south, the Neogene Siwalik molasse is separated from the Lesser Himalayan sequence by the Main Boundary thrust. The Main Frontal thrust marks the deformation front of the Himalayan orogen, and it consists of a blind active thrust below the Siwalik molasse against Quaternary alluvial sediments (Lavé and Avouac, 2000). Within the eastern Himalaya, a distinctive out-of-sequence thrust, the Kakhtang thrust (Fig. 1B), duplicates the thickness of the Greater Himalayan sequence (Davidson et al., 1997; Grujic et al., 2002).
SOUTH TIBETAN DETACHMENT SYSTEM
The South Tibetan detachment system is a laterally continuous network of normal-sense, mainly low-angle detachments, some of which were dominantly active in the Miocene and some of which are still active (Burg and Chen, 1984; Burchfiel and Royden, 1985; Burchfiel et al., 1992; Brown and Nazarchuk, 1993; Hurtado et al., 2001; Wiesmayr et al., 2002; Meyer et al., 2006). In the hanging wall of the detachment system, there are Cambrian to Eocene sedimentary rocks of the Tethyan sedimentary sequence (Garzanti, 1999). In the footwall, there are migmatite, amphibolite- to granulite-grade metamorphic rocks, and Miocene leucogranites of the Greater Himalayan sequence. Locally, greenschist- to amphibolite-facies metasedimentary rocks crop out within the South Tibetan detachment system. Regionally, these rocks are termed the Chekha Group (Bhutan), the Everest Series (eastern Nepal), the North Col Formation (eastern Nepal), the Annapurna Yellow Formation (central Nepal), and the Haimantas Group (NW India) and are probably latest Proterozoic to Ordovician in age (Frank et al., 1973; Gansser, 1983; Colchen et al., 1986; Burchfiel et al., 1992; Lombardo et al., 1993; Bhargava, 1995; Carosi et al., 1999; Searle et al., 2003; Myrow et al., 2009).
South Tibetan Detachment System in Bhutan
In Bhutan, there are two segments of the South Tibetan detachment system. The more internal, northern segment is referred to here as the inner South Tibetan detachment and the more external, southern segment is referred to as the outer South Tibetan detachment. In northwestern Bhutan, the inner South Tibetan detachment occurs as two closely spaced normal-sense faults, which dip shallowly (10°–30°) to the north, covered by Quaternary deposits (Figs. 2D and 3C) (Burchfiel et al., 1992). The northern brittle structure separates Upper Paleozoic shale, siltstone, and limestone of the Tethyan sedimentary sequence from metasedimentary rocks, while the southern ductile structure separates metasedimentary rocks from leucogranite, augen gneiss, and schist of the Greater Himalayan sequence. Variably deformed leucogranite sills and dikes increase upward in abundance, reaching nearly 100% mylonitic leucogranite in outcrops at the top of the Greater Himalayan sequence (Burchfiel et al., 1992; Edwards et al., 1996; Chakungal, 2006; our observations). Younger, brittle normal faulting along the Dzong Chu fault occurs north of, and truncates, the ductile inner South Tibetan detachment (Fig. 2) (Edwards et al., 1996, 1999). Normal faulting has continued in northern Bhutan to recent times, documented by faulted Quaternary moraines, although the most recent faulting is apparently caused by E-W rather than N-S extension (Meyer et al., 2006).
While metasedimentary rocks are only locally present at the inner South Tibetan detachment, thick sequences of the metasedimentary Chekha Group are preserved farther south within the outer South Tibetan detachment (Figs. 2 and 3). The outer South Tibetan detachment, with intensely folded gneisses of the Greater Himalayan sequence in its footwall and the Chekha Group in its hanging wall, is a diffuse shear zone exhibiting an upward decrease in metamorphic grade. Conjugated top-to-the-south and top-to-the-northwest shear bands and strong subhorizontal crenulation cleavage in the Chekha Group indicate a zone of general shear with components of both vertical shortening and top-to-the-north simple shear (Grujic et al., 2002; Carosi et al., 2006), similar to that described for the well-studied South Tibetan detachment in the Everest region (e.g., Law et al., 2004; Jessup et al., 2006). The shear zone is preserved within broad, noncylindrical synclines in central and southern Bhutan (Fig. 1A; Edwards et al., 1996; Grujic et al., 2002). Some of these klippen also preserve Tethyan sedimentary sequence rocks in their cores. The contact between the Chekha Group and the Tethyan sedimentary sequence marks a seemingly abrupt change in structural style and metamorphic grade, perhaps indicating a fault.
An interpretation of compiled published data from the Himalaya suggests that South Tibetan detachment structures may young toward the north (Godin et al., 2006). Edwards et al. (1996) suggested that the geometry of the South Tibetan detachment system in Bhutan is a result of later, brittle faulting cutting an earlier ductile South Tibetan detachment, implying synchroneity for both ductile segments of the South Tibetan detachment (inner and outer South Tibetan detachment), with younger brittle faulting cutting into the ductile shear zone in the north (inner South Tibetan detachment). We focus this study on constraining the timing of ductile components of the inner and outer South Tibetan detachment structures in areas where they are well accessible and separated enough to avoid overlapping of geological information.
The inner South Tibetan detachment in the eastern Himalaya has been shown, by Th-Pb and U-Pb dating of monazite in mylonitized granite bodies in the footwall, to have been active at least until ca. 12.5–12.0 Ma in northern Bhutan at Kula Kangri and at Wagye La (Edwards and Harrison, 1997; Wu et al., 1998, respectively). These ages can also be interpreted as constraining the onset of brittle deformation since the mylonitized granites are cut by brittle South Tibetan detachment faulting. In the adjacent area of Sikkim (India), zircon and monazite in deformed leucogranites in some component of the South Tibetan detachment have been dated at ca. 17 and 15–14 Ma, indicating that South Tibetan detachment–related ductile deformation may have occurred there until ca. 14 Ma (Fig. 2; Catlos et al., 2004). Only slightly younger muscovite 40Ar/39Ar cooling ages in granites in these areas of ca. 11–10 Ma (Maluski et al., 1988) indicate rapid cooling of the footwall and constrain the minimum timing of ductile deformation on the inner South Tibetan detachment. It has been recently suggested that ductile shearing occurred on the outer South Tibetan detachment between 24–22 Ma (Chambers, 2008) and 22–17 Ma (Grujic et al., 2002).
This study includes three transects across segments of the South Tibetan detachment system. The Lingshi and Ura transects cross the outer South Tibetan detachment where the Chekha Group is exposed in broad synforms in NW and central Bhutan (Figs. 2B and 2C). The Masang Kang transect in NW Bhutan crosses the inner South Tibetan detachment where it separates granulite-grade gneisses and leucogranites of the Greater Himalayan sequence from Tethyan sedimentary sequence rocks (Fig. 2D). The characteristic metamorphic assemblages observed in each transect are discussed briefly next.
Lingshi (Outer South Tibetan Detachment)
On the SW side of the Lingshi area, garnet-sillimanite gneiss of the Greater Himalayan sequence is separated from the Chekha Group by the Chung La leucogranite (Fig. 2B), which is a large sill intruded along the contact and tapering toward the east. A slightly deformed component of the Chung La granite has been dated by monazite U-Pb to ca. 23–22 Ma (R. Parrish, 2008, personal commun.). To the east of the Lingshi area, the Greater Himalayan sequence is composed of sheared garnet-bearing metapelite, quartzite, pelitic gneiss, and augen gneiss intruded by leucogranite dikes and sills. Near the contact with the Chekha Group, gneisses are migmatized, and leucosomes contain garnet, muscovite, and sillimanite. Above the Greater Himalayan sequence, there are sheared, intercalated, calc-silicate, garnet, and staurolite-bearing metapelitic schist, phyllite, and rare quartzite of the Chekha Group, which rapidly decrease in metamorphic grade from amphibolite facies at the base to greenschist facies farther up-section. Above the calc-silicate and schist, there is a thick sequence of monotonous gray marble. Leucogranite intrusions are common at the base of the Chekha Group and disappear ~2.5 km structurally above the contact. The Chekha Group appears to have a structural thickness of 4–5 km in the Lingshi area (Fig. 3A). The stratigraphic thickness is unconstrained because the metasedimentary rocks are isoclinally folded and sedimentary bedding has been transposed (Figs. 4A–4D). The contact between the Chekha Group and the overlying Tethyan sedimentary sequence is characterized by a sharp boundary from marble to immature sandstone in the east and from metapelitic schist to fine-grained graphitic slate, sandstone, and limestone in the west, accompanied by a decrease in metamorphic grade and transition from transposed bedding to preserved right-way-up stratigraphy (see Gansser  for further structural and lithologic details).
Ura (Outer South Tibetan Detachment)
In the Ura area, the Greater Himalayan sequence is characterized by sheared migmatitic garnet-bearing augen gneiss. Leucogranite dikes and sills are common within the augen gneiss. Some leucogranites contain magmatic cordierite, partially replaced by subsolidus sillimanite and muscovite, followed by andalusite. The overlying Chekha Group is composed of sheared amphibolite-facies garnet-bearing metapelitic schist and biotite-bearing quartzite. Amphibolite bands are present locally. To the north, two thick white marble bands are present. Garnet-staurolite schist becomes abundant (the Naspe Formation of Bhargava, 1995) toward the footwall of the Kakhtang thrust (Figs. 1B and 2C; Gansser, 1983; Davidson et al., 1997; Grujic et al., 2002). Leucogranite dikes and sills are pervasive throughout the Chekha Group. Tethyan sedimentary sequence rocks are not preserved within the Ura klippe.
Masang Kang (Inner South Tibetan Detachment)
The Greater Himalayan sequence in the Masang Kang area forms an east-west–trending, open, upright antiformsynform pair (Fig. 2C). Lithologies within the Greater Himalayan sequence include migmatite, augen gneiss, orthogneiss, calc-silicate, and quartzite with granulite-facies mafic and ultramafic lenses (Chakungal, 2006). Leucogranite dikes and sills intrude the Greater Himalayan sequence, and a large, variably sheared leucogranite body is exposed to the northeast. The leucogranite intrusions are cut by brittle faulting along the inner South Tibetan detachment and do not penetrate into the hanging wall Tethyan sedimentary sequence rocks.
Miocene leucogranite and granite dikes and sills are found throughout the Greater Himalayan sequence in Bhutan, and the largest bodies occur at the very top of the sequence and broadly correspond to the highest metamorphic grades (e.g., Wagye La in NW Bhutan; see Fig. 2B) (Gansser, 1983; Lombardo et al., 1993). Variably deformed leucogranite and granite dikes and sills are also found within the lower structural levels of the Chekha Group.
Dikes and sills emplaced within the Chekha Group at the outer South Tibetan detachment are generally leucogranitic (to granitic), coarse-grained to bimodal, and are characterized by abundant tourmaline and apatite, and minor garnet and retrograde subsolidus sillimanite. Tartan twinning in K-feldspar suggests crystallization at >450–550 °C (e.g., Brown and Parsons, 1989), whereas the presence of sillimanite suggests crystallization at ≥550 °C (using the aluminosilicate stability field of Pattison, 1992). The leucogranites are hosted by marble, calc-silicate, and schist and were therefore clearly injected (i.e., they did not form in situ). Leucogranite bodies emplaced within the top of the Greater Himalayan sequence at the inner South Tibetan detachment are characterized by magmatic andalusite and cordierite, and retrograde sillimanite. We suggest that the andalusite crystallized slightly above the solidus (it occurs interstitially and as marginal inclusions in magmatic phenocrysts) and, therefore, probably at a pressure ≤2.8 kbar (see details in Appendix A of the supplementary data1).
Samples for this study were collected from deformed leucogranite dikes and sills (Fig. 4). In the Lingshi and Ura areas, the sampled leucogranites were emplaced within Chekha Group rocks (Figs. 2B, 2C, 3A, and 3B). Sample DBH 003 was collected from an ~5 cm dike at the base of the Chekha Group, 100 m from the Chung La granite (Fig. 2B). The dike crosscuts the host calc-silicate, which has been isoclinally folded, and it has a pervasive axial planar cleavage (the dominant foliation in the area) that dips 40°–50° toward the ENE. The dike is folded into open folds with axial plane dipping 24° toward the SSW (Fig. 4A). The foliation is folded in a similar manner, although with a higher amplitude and tighter folds, indicating that the intrusion postdates the main foliation but is pre- to synkinematic with the later folding. Samples DBH 027, DBH 031, and DBH 036 were sampled from leucogranite sills progressively structurally up-section from the base of the Chekha Group on the eastern side of the Lingshi transect (Figs. 2B and 3A). The sills are all weakly folded and boudinaged, at the outcrop scale, within the host Chekha Group phyllitic schist (DBH 027) and calc-silicate (DBH 031, DBH 036) dipping ~10°–20° toward the W-SW (Fig. 3C). DBH 080 was sampled from a boudin of leucogranite sill emplaced in crenulated metapelitic schist near the western base of the Chekha Group in the Ura transect (Fig. 2C). The sill and main foliation of the schist both dip ~10° toward the ESE. The leucogranites in the Chekha Group show a weakly developed, coarse schistosity mainly defined by subparallel mica ± grain shape fabric of quartz grains, and no lineation. The style of deformation exhibited by the leucogranite bodies (boudinage, folding, or both) is a function of their emplacement orientation with respect to the outer South Tibetan detachment.
Sample BH 225 from the Masang Kang area was collected from a deformed leucogranite emplaced in and crosscutting migmatitic gneisses of the Greater Himalayan sequence ~1 km below the Greater Himalayan sequence–Tethyan sedimentary sequence contact (Fig. 3C). Below this, at the lower/southern deformation front of the inner South Tibetan detachment, top-to-the-NW shear bands appear, strain increases progressively northward into a mylonitic top-to-the-NW foliation, and S-C fabric is accompanied by conjugate shear bands, sillimanite mineral lineation, and a consistent stretching lineation. Microstructural evidence (described next) indicates that BH 225 has been deformed by the inner South Tibetan detachment.
The leucogranite samples collected from the hanging wall of the outer South Tibetan detachment do not have obvious chill margins (Fig. 4E). Microstructures include undulose extinction, subgrain rotation, and grain boundary migration (GBM) recrystallization in quartz grains (Figs. 4E and 4F). Subgrain rotation recrystallization dominates, although grain boundary migration recrystallization is locally significant. Deformation in feldspar includes undulose extinction, deformed twins, perthite and flame perthite, and microfractures (Figs. 4F and 4G). Muscovite is coarse, and generally displays undulose extinction, local recrystallization, and rare kinks (Fig. 4E). The main planar fabric of the leucogranites is defined by the alignment of muscovite. These microstructures suggest that ductile deformation of the leucogranites occurred at ~400–500 °C (Stipp et al., 2002; Passchier and Trouw, 2005).
Sample BH 225, collected from the footwall of the inner South Tibetan detachment, is weakly deformed. In thin section, both plagioclase (Pl) and K-feldspar (Kfs) show evidence of crystal-plastic deformation, including myrmekite, recrystallized grains (Kfs), undulose extinction (Pl), and deformation twins (Pl). Quartz grains display dynamic recrystallization dominantly by grain boundary migration (Fig. 4H) and locally by subgrain rotation. Combined, these microstructures suggest that South Tibetan detachment–related ductile deformation for this leucogranite occurred at ~500–600 °C (Stipp et al., 2002; Passchier and Trouw, 2005). In addition, sillimanite (Fig. 4H) is present as slickenfibers in top-to-the-north shear bands and also replaces andalusite. Muscovite replaces cordierite, andalusite, and feldspar, and is locally recrystallized.
SHRIMP ZIRCON U-Pb GEOCHRONOLOGY
All six studied leucogranites show evidence of solid-state ductile deformation, but they also crosscut earlier, more developed host rock fabrics; they must have been emplaced late synkinematically and crystallized synkinematically with top-to-the-north shearing both in the outer and inner South Tibetan detachment in Bhutan, consistent with previous observations (Edwards et al., 1996; Edwards and Harrison, 1997; Davidson et al., 1997; Wu et al., 1998; Grujic et al., 2002; Daniel et al., 2003; Chambers, 2008).
Zircons were analyzed for U-Pb and trace-element composition using the SHRIMP-RG (sensitive high-resolution ion microprobe–reverse geometry) co-operated by the U.S. Geological Survey and Stanford University. Sample locations are shown in Figures 1 and 2. Approximately 20 spots from 12 to 15 zircon grains were analyzed for each sample, except sample BH 225, for which ~70 spots were analyzed from ~40 zircon grains. Pits were ~30 μm in diameter and had a depth of ~5 μm. Mineral separation (standard and SelFrag for BH 225), analytical procedures, and detailed zircon characterization are outlined in Appendix B (supplemental material [see footnote 1]). In general, the outer South Tibetan detachment zircons exhibit dark-under-cathodoluminescence (CL) magmatic growth rims, spongy, inclusion-rich interiors, and rare inherited cores, while the inner South Tibetan detachment zircons exhibit magmatic growth rims and gray-under-CL interiors and commonly contain inherited cores.
Trace-Element Geochemistry of Zircon
The reverse-geometry arrangement of the SHRIMP-RG at Stanford University allows for the collection of trace-element geochemistry data coincident with U-Pb analyses (Mazdab and Wooden, 2006). For full data tables, see Appendix C (see footnote 1).
Zircon rims from leucogranites along the outer South Tibetan detachment have a consistent trace-element geochemistry characterized by a small positive Ce anomaly and a large negative Eu anomaly (Fig. 5A). Eu anomalies for zircon rims range from Eu/Eu* = 0.17 to as low as Eu/Eu* = 0.001; most analyses range from 0.05 to 0.001 (Fig. 5D). Heavy rare earth elements (HREEs) are somewhat to very depleted in zircon rims. In particular, sample DBH 080 yielded an average Yb/Gd ratio of ~2 (Fig. 5E). Zircon rim Th/U values are uniformly low (<0.1, Fig. 5E).
Such low Eu/Eu* values are likely an indication of crystallization in the presence of feldspar, which will take up all available Eu (Rubatto, 2002). This suggests zircon crystallization simultaneous with or after feldspar. Depletion in HREEs indicates preferred fractionation of HREEs by another mineral phase, probably garnet (Rubatto, 2002), which is present in samples DBH 027 and DBH 080. Assuming HREEs are largely stored in garnet, partial melting of the source rocks in which garnet is left in the residual phase could also cause HREE depletion in the leucogranite, in which case greater HREE depletion would indicate a smaller percentage of partial melt at the source. The consistency of the REE patterns of zircon rims compared to cores (Fig. 5B) indicates that mixing of age domains during microsampling was not an issue.
Zircon rims and gray-under-CL interiors from sample BH 225 that yield Miocene U-Pb ages (many gray interiors, as will be discussed later, yield early Paleozoic ages) are plotted in Figure 5C. The two growth phases are indistinguishable on the basis of their REE patterns, but they can be distinguished from the zircon rims from the outer South Tibetan detachment leucogranites by their less pronounced negative Eu anomaly, steeper HREE pattern, and lower Hf values (Figs. 5D and 5E). These zircons likely crystallized from a different protolith than those first described, and perhaps one that experienced a greater percentage of partial melting, consistent with higher country rock metamorphic temperatures in the footwall of the inner South Tibetan detachment (Chakungal, 2006).
U-Pb Geochronology Results
Results are presented in Table 1 and plotted in Figure 6,602. For all samples from the hanging wall of the outer South Tibetan detachment (Figs. 6A–6D), U-Pb ages of zircon rims do not define a single crystallization age but rather produce a spread of ages along concordia. This spread is discussed later herein. In general, 206Pb/238U age ranges to the nearest 0.5 Ma (at the 2σ confidence level) are reported through the text, unless otherwise indicated. Zircon rim data that are >1% discordant are shown in Table 1 and Figure 6,602, but they were not used for U-Pb age interpretation. For the Lingshi area, apparent 206Pb/238U ages range from ca. 24.5 to 16.5 Ma, 29.5 to 19.0 Ma, and 24.5 to 16.5 Ma (for DBH 027, 031, and 036, respectively). In the Ura area, sample DBH 080 zircon rims yielded 206Pb/238U ages of ca. 20.5–15.5 Ma. BH 225 from the footwall of the inner South Tibetan detachment displayed a similar spread of ages along concordia (Fig. 6E). Dark rim apparent 206Pb/238U ages range from ca. 15.5 to 11.0 Ma, while the gray interior zircon apparent ages range from ca. 29.0 to 11.5 Ma. Within individual grains, dark rims are younger than their gray interiors. Cores and interior zircon from BH 225 yielded a significant age population at ca. 500 Ma, as well as many Proterozoic ages, and a few younger discordant Paleozoic ages (Fig. 6F). Cores from DBH 003 from the Lingshi area were discordant but indicate Proterozoic ages (Fig. 6G).
Crystallization of the zircon rims is correlated with the presence of melt. Solid-state ductile deformation of the leucogranites must have therefore outlasted the youngest zircon rim crystallization ages. Thus, we conclude that ductile deformation continued after ca. 15.5 Ma for the outer South Tibetan detachment (based on the youngest zircon ages from DBH 080, DBH 027, DBH 031, and DBH 036) and after ca. 11.0 Ma for the inner South Tibetan detachment (based on the youngest zircon ages from BH 225).
With the exception of sample DBH 080, most of the Miocene-aged zircon phases in this study exhibit a high U content >2500 ppm (Table 1). Zircon with U content >2500 ppm has been reported as causing a systematic bias toward older apparent U-Pb ages in SHRIMP analyses independent of the zircon age (Williams and Hergt, 2000). We do observe a broad positive correlation between U content and age in our samples (Fig. 6H). This may provide an explanation for the observed spread of ages along concordia. For a systematic analytical bias, it should be possible to regress the age data back to a common hinge point at 2500 ppm, giving the “true” age of the zircon rims (e.g., Larson, 2009). However, if there is such a bias in apparent zircon age, it is not consistent within our samples. Furthermore, DBH 080, for which the U content of the zircon rims is entirely below 2500 ppm, still displays a spread along concordia in apparent age from ca. 20.5 to 15.5 Ma, though this is a smaller age range than that observed in other samples. Thus, while we agree that there is likely some bias in U-Pb age due to high U content, we argue that it is not consistent and thus cannot be quantified or corrected for in this data set. Since we use the youngest zircon ages for each sample (all of which have U contents <2500 ppm) to constrain the timing of ductile motion, this potential bias does not affect our conclusions.
The crystallization temperature of zircon was estimated using Ti-in-zircon thermometry (Watson et al., 2006). Spot locations were sampled adjacent to the spots for U-Pb analyses of the same zircons and have the same labels; analytical procedures can be found in Appendix B (see footnote 1). Data collected from zircon rims are shown in Table 2, and the results are plotted in Figure 7 against coincident SHRIMP U-Pb spot ages using the revised calibration of Ferry and Watson (2007). The calculated Ti-in-zircon temperatures are uncorrected for pressure.
Since quartz is present in abundance in the leucogranites, aSiO2 can be considered to be 1.0. For most igneous rocks, aTiO2 is ≥0.5 (Watson and Harrison, 2005). The studied leucogranites do not contain any Ti-bearing phases such as rutile, an indication that aTiO2 < 1.0. Our data are calculated for aTiO2 = 0.5 (Fig. 7) and probably represent maximum temperatures. For comparison, calculations of Ti-in-zircon temperatures for aTiO2 = 1.0 (which represent minimum temperatures) are also listed in Table 2, and they are in general ~30–40 °C lower.
In the Ura area, apparent crystallization temperatures range from ~620 to 500 °C for zircons with apparent ages ca. 20.5 to 15.5 Ma, with no obvious temperature-time (T-t) trend (Fig. 7A). In the Lingshi area, zircons have slightly higher apparent crystallization temperatures of between ~750 and 550 °C during ca. 25.0–16.5 Ma (DBH 036, Fig. 7B) and ~750 and 550 °C during ca. 24.5–16.0 Ma (DBH 027, Fig. 7C). Neither sample displays a T-t trend during zircon crystallization. In contrast, apparent crystallization temperatures for BH 225 from the inner South Tibetan detachment range from ~820 to 670 °C during ca. 15.5–11.0 Ma (Fig. 7D). The observed spread in apparent temperature at any given time (100–150 °C) may be due to local variations in Ti concentration during zircon crystallization, or the sampling of Ti-rich and Ti-poor subdomains in various analyses (Fu et al., 2008).
For all samples, there is an inverse correlation between Hf values and Ti-in-zircon crystallization temperatures (Fig. 7E). This has been observed in general for zircon (Fu et al., 2008) and can be taken as a reflection of the increasing degree of fractionation from least fractionated (BH 225) to most fractionated (DBH 080) (Claiborne et al., 2006). Thus, in general, all the outer South Tibetan detachment samples are more fractionated than the inner South Tibetan detachment sample.
Some apparent crystallization temperatures calculated for zircon rims from the outer South Tibetan detachment samples DBH 036, DBH 027, and especially DBH 080 are lower than even a minimum estimate for the solidus of the leucogranites of ~600–650 °C, assuming a pressure of ~4–6 kbar at the time of crystallization (Fig. 7; Davidson et al., 1997). However, the leucogranites contain abundant apatite and tourmaline, which suggest that the magmas contained high concentrations of B, P, H2O, and possibly other fluxes. At emplacement pressures >~2.7 kbar, high concentrations of fluxing elements may lower the solidus of graniticpegmatitic melts to as low as ~400–350 °C (Sirbescu and Nabelek, 2003). The Ti-in-zircon temperatures, tartan twinning in K-feldspar, and presence of retrograde subsolidus sillimanite indicate that crystallization of the studied leucogranites took place at temperatures >550 °C, but possibly at or below an average solidus of ~650 °C. The Ti-in-zircon thermometer was calibrated for pressure (P) = 10 kbar, and crystallization at lower pressures may cause temperatures to be underestimated by up to 50 °C (Ferry and Watson, 2007). Although there may be other, non-temperature-related factors influencing Ti concentrations (see Fu et al., 2008), the apparent Ti-in-zircon temperatures are consistent with our geochemical observations that the leucogranites crystallized at low temperature.
MUSCOVITE 40Ar/39Ar THERMOCHRONOLOGY
Muscovite from four leucogranite samples (DBH 003 and DBH 027 from the Lingshi area and DBH 067 and DBH 080 from the Ura area; Figs. 2 and 3) were selected for step-heating 40Ar/39Ar thermochronology to determine cooling ages of the leucogranite bodies in the hanging wall of the outer South Tibetan detachment. Muscovite makes up between 10% and 25% of the samples, and it is anhedral and coarse grained (up to 500 μm in length), with little recrystallization. Muscovite grains in samples DBH 003, DBH 067, and DBH 080 show a weak preferred orientation. Muscovite in sample DBH 027 has a preferred orientation that, along with deformed quartz grains, defines a foliation in the granite subparallel to the main foliation in the host rock.
The 40Ar/39Ar thermochronology of bulk coarse-grained (~200 μm average diameter) muscovite separates was carried out using a Heine-based Ta double-vacuum furnace at Dalhousie University, Halifax. Sample preparation and experimental procedures are outlined in Appendix D (see footnote 1). All plateau ages are reported at the 2σ confidence level.
All samples yield well-constrained plateau ages (Table 3; Fig. 8). Of the two samples analyzed from the Lingshi transect, sample DBH 003 yields a plateau age of 13.2 ± 0.2 Ma (93.3% of Ar released), while sample DBH 027 yields a plateau age of 11.5 ± 0.2 Ma (97.3% of Ar released). From the Ura transect, sample DBH 067 has a plateau age of 11.7 ± 0.2 Ma (96.8% of Ar released), and sample DBH 080 has a plateau age of 11.1 ± 0.2 Ma (86.0% of Ar released).
Recent advances in the understanding of Ar diffusion and distribution in muscovite warrant a short discussion of the concept of Ar closure temperature. The closure temperature of muscovite to Ar diffusion is nominally reported as ~350 °C (Hames and Bowring, 1994). It has been observed that high metamorphic temperatures of 500–600 °C have left Ar undisturbed in muscovite (see review in Villa, 2004), which is inconsistent with the assumption that Ar will always diffuse out of muscovite at temperatures greater than ~350 °C. This observation has been noted particularly in metamorphic rocks bearing multiple phases and generations of muscovite (e.g., Di Vincenzo et al., 2004). However, the preponderance of muscovite 40Ar/39Ar data for igneous rocks, both within and outside the Himalayan orogen, are consistent with muscovite 40Ar/39Ar ages < igneous crystallization ages recorded by minerals such as zircon, arguing for post-igneous-crystallization commencement of Ar retention in muscovite. Here, although minor dynamic recrystallization produced fine-grained muscovite, only coarse, igneous grains were selected for 40Ar/39Ar analysis. The granites have a relatively simple geologic history of crystallization from a melt, followed by subsolidus deformation during exhumation, and the well-defined plateaus produced during step-heating suggest that Ar systematics have not been disturbed in these grains since initial cooling. Ti-in-zircon data (Fig. 7) show that zircon was crystallizing from melt at fairly low temperatures of ≤600 °C, near the leucogranite solidus. It is possible, considering the aforementioned studies, that Ar did not diffuse out of the muscovite below 500–600 °C, and thus that the 40Ar/39Ar plateaus represent crystallization and not cooling ages of the muscovite, but we argue that in this case, such high closure temperatures are inconsistent with our Ti-in-zircon thermometry and U-Pb zircon ages.
Recent experiments on Ar diffusion in muscovite indicate that in cases where temperature is likely the main factor influencing Ar diffusion, ~425 °C may be a more appropriate approximation of Ar closure temperature than 350 °C (for pressures of ~5 kbar, muscovite grain diameters of ~200 μm, and a cooling rate of ~60 °C/Ma; determined from average slopes in Fig. 7; Harrison et al., 2009). Since our samples have experienced minor deformation, equations of volume diffusion may not be a strictly accurate way to estimate closure temperature (deformation may have kept Ar pathways open below the diffusion temperature, and thus 425 °C may be an over- rather than underestimate of Ar retention temperature).
A closure temperature of ~425 °C is slightly higher than the estimated temperature of the brittle-ductile transition zone of ~350 °C for quartzofeldspathic rocks (e.g., Sibson, 1977; Passchier and Trouw, 2005), and it coincides with the estimated temperature of ductile deformation of 400–500 °C for outer South Tibetan detachment granites (see previous discussion). Thus, the 40Ar/39Ar apparent cooling ages can be interpreted as closely preceding cessation of ductile deformation along the outer South Tibetan detachment. Since the inner South Tibetan detachment experienced ductile deformation until at least after 11.0 Ma, as constrained by U-Pb geochronological data (this study; Edwards and Harrison, 1997; Wu et al., 1998), followed by brittle deformation, then the outer South Tibetan detachment ceased operating while the inner South Tibetan detachment was still active as a ductile to ductile-brittle shear zone. A biotite 40Ar/39Ar cooling age from a leucogranite near Wagye La beneath the inner South Tibetan detachment of ca. 10.2 Ma and muscovite and biotite cooling ages from the Kula Kangri granite of ca. 11.4–10.7 Ma (Chakungal, 2006; Maluski et al., 1988, respectively) suggest that cooling of the inner South Tibetan detachment must have been rapid.
Timing of the South Tibetan Detachment System in the Eastern Himalaya
A recent review of published geochronological data suggests that, orogenwide, the timing of displacement of interpreted upper and lower components of the South Tibetan detachment system can be approximated as ca. 18 and ca. 22 Ma, respectively (Carosi et al., 1998; Godin et al., 2006). These upper and lower components of the South Tibetan detachment (e.g., the Qomolangma and Lhotse detachments, respectively, of the Everest region [Searle et al., 2003; Sakai et al., 2005] and the Phu and Chame detachments of the Annapurna region [Searle and Godin, 2003]) are not considered analogous to the outer South Tibetan detachment and inner South Tibetan detachment described here. In Nepal, the two detachments are structurally within 1–2 km of each other (e.g., Searle and Godin, 2003; Searle et al., 2003); they may be two components of the same shear zone acting simultaneously or progressively at different structural levels and rheological boundaries. In Bhutan, this might correlate to Greater Himalayan sequence–Chekha and Chekha–Tethyan sedimentary sequence bounding structures in the Lingshi klippe, both of which are part of the outer South Tibetan detachment (we suggest that the Chekha–Tethyan sedimentary sequence boundary may be a fault, although, thus far, it has been poorly defined in the field). However, there is a distinct time and geographic separation between the defined inner and outer detachments in Bhutan. Thus, we prefer the terms inner South Tibetan detachment and outer South Tibetan detachment to distinguish them from previously published upper and lower South Tibetan detachment terminology.
Structural, geochronological, and thermo-chronological data, previously published and ours (Table 4), indicate that ductile motion along the outer South Tibetan detachment began at ca. 24 Ma, continued until at least 16.0 (Lingshi) to 15.5 (Ura) Ma, but ceased by ca. 11.0 Ma, while ductile motion along the inner South Tibetan detachment remained active through ca. 11.0 Ma. Brittle faulting at the inner South Tibetan detachment followed (e.g., Wiesmayr et al., 2002; Meyer et al., 2006), while there is no field or geophysical evidence for any recent deformation associated with the outer South Tibetan detachment. These data suggest that the outer South Tibetan detachment was abandoned in the mid-Miocene. As discussed further later herein, the Kakhtang thrust, an out-of-sequence thrust within the Greater Himalayan sequence, separates the inner and outer detachments, at least in central and eastern Bhutan. Its age has been constrained to younger than 15 Ma by U-Pb dating of monazite in deformed leucogranites (Daniel et al., 2003), and it is postulated to be concurrent with the inner South Tibetan detachment (Grujic et al., 2002). Map relationships in central Bhutan indicate that the Kakhtang thrust cuts and therefore postdates the outer South Tibetan detachment (Figs. 1B and 2C).
Implications for Tectonic Models of the Himalayan Orogen
The two main questions arising from this study that have implications for the formation of the Himalayan orogen are: (1) What caused the outer South Tibetan detachment to be abandoned? (2) How did that abandonment affect the evolution of the orogen?
We suggest two possible mechanisms for abandonment of the outer South Tibetan detachment. First, the South Tibetan detachment may have stepped to the hinterland because the outer South Tibetan detachment segment of the shear zone became unviable. It has been documented that normal-sense detachment systems can progressively step structurally up-section as older, deeper shear surfaces become rotated toward the horizontal (progressively flattening during footwall exhumation) and, in cases, folded (e.g., Mancktelow and Pavlis, 1994), or they can even step down-section (e.g., Brichau et al., 2007). The outer South Tibetan detachment has been folded in a low-amplitude dome and basin geometry and may have been rotated toward the south (Wiesmayr et al., 2002). If both occurred during shearing, the changed orientation of the outer South Tibetan detachment surface may have prevented further slip and caused a new shear plane (inner South Tibetan detachment) to initiate. A second possibility is that the outer South Tibetan detachment became separated from the South Tibetan detachment due to doming within the Greater Himalayan sequence at depth. Exhumation of the dome would have eroded through a segment of the South Tibetan detachment, resulting in a laterally continuous “window” of younger and deeper Greater Himalayan sequence material separating the outer South Tibetan detachment from the active South Tibetan detachment system.
To examine these two possibilities, two general end-member mechanisms for orogenic wedge development widely applied to describe the Himalayan orogen may be considered: critical taper–frictional wedge theory (e.g., Webb et al., 2007; Kohn, 2008; Robinson, 2008) and channel flow–viscous wedge theory (e.g., Grujic et al., 1996; Beaumont et al., 2004; Carosi et al., 2006). Critical taper theory dictates that deformation in a frictional orogenic wedge occurs to maintain a critical taper angle between the hinterland-dipping basal detachment and the topographic surface of the orogen. That angle can decrease either by increased erosion rates or foreland propagation of thrusting or both, and it can increase by out-of-sequence thrusting or decreased erosion rates (Dahlen, 1990). It can be argued that hinterland stepping of a normal-sense fault in an orogenic wedge would also reduce the critical taper angle by widening the orogen. During 12–10 Ma, it appears that the outer South Tibetan detachment was abandoned synchronously with propagation of thrusting toward the foreland from the Main Central thrust to the Main Boundary thrust (at least in the western Himalaya; the timing of the Main Boundary thrust has not been well constrained for the eastern Himalaya; Meigs et al., 1995; Stüwe and Foster, 2001; Daniel et al., 2003). Intensified erosion also began at ca. 12–10 Ma as indicated by sediment accumulation rates in the Siwaliks (Najman, 2006, and references therein), which may be coincident with the start of the Indian monsoon (Dettman et al., 2003), although the date of the latter has been highly disputed, and it may have been already established by ca. 24 Ma (e.g., Clift et al., 2008). Thus, both orogenwide changes in thrusting and erosion rates and an arguably local hinterland stepping of normal-sense displacement all should have resulted in a widening of the orogenic wedge and narrowing of the critical taper angle. The expected tectonic response to these changes would be out-of-sequence thrusting to reestablish the critical taper angle. If the Kakhtang thrust arose in response, it was immediate, short-lived, and only locally developed. There is no evidence of other later, out-of-sequence thrusting in Bhutan. Furthermore, since both foreland propagation and intensified erosion occurred at the orogen scale, out-of-sequence thrusting within the Greater Himalayan sequence might be expected to be observable at the orogen scale.
Alternatively, channel-flow models propose that in large hot orogens, the gravitational potential between the thickened hinterland and the foreland results in an outward lateral flow of weakened midcrustal material in a channel bounded by more rigid crustal layers (e.g., Beaumont et al., 2004, 2006). Focused erosion at the deformation front of the orogen allows for the extrusion of midcrustal rocks to the surface along coeval yet opposite-sense shear zones; in the Himalayan orogen, these shear zones may be analogous to the Main Central thrust and the South Tibetan detachment (Beaumont et al., 2001). These models have shown that cool, underthrusted crust can form a crustal ramp on the basal detachment that will propagate into the orogen. Crustal ramps do occur in the Himalaya; a crustal ramp is interpreted beneath southern Tibet in the INDEPTH profile (Hauck et al., 1998), and similar, more external (thus younger) ramps may exist in the central Himalaya (Avouac, 2003). Other, more internal ramps may have been consumed. In channel-flow models, these ramps are found to destabilize the middle crust, resulting in doming of a weak, partially molten crustal layer; this process has been suggested as a mechanism for the formation of the North Himalayan gneiss domes (Beaumont et al., 2004). In channel-flow models, the tectonic fate of such a midcrustal dome depends on the strength of the upper crust and erosion efficiency. For example, the dome can be translated toward the exhumation front and juxtaposed on top of a previously exposed channel, potentially (but not necessarily) above an apparent out-of-sequence thrust (i.e., figure 12e of Beaumont et al., 2004; Jamieson et al., 2006). We suggest that exhumation of such a dome could have caused erosion through the South Tibetan detachment, cutting off the outer segment (outer South Tibetan detachment) from the enduring South Tibetan detachment (inner South Tibetan detachment).
We suggest that the now-exhumed outer South Tibetan detachment was first active as a near-horizontal top-to-the-north detachment deforming lower Chekha Group and upper Greater Himalayan sequence rocks, accompanied by anatexis at a depth of 10–30 km, from ca. 25 to 16 Ma (Fig. 9A; Grujic et al., 2002; Harris et al., 2004; Chambers, 2008). Partially melted Greater Himalayan sequence in the hinterland flowed southward by way of a pressure gradient between hinterland and foreland (Grujic et al., 1996). Melt pooled at a rheological boundary at the top of the Greater Himalayan sequence, resulting in the formation of plutons such as the Chung La and Kula Kangri granites (Scaillet and Searle, 2006). From ca. 16 to 12 Ma, the outer South Tibetan detachment began to exhume and cool (Figs. 6,602 and 8B) due to erosion at the surface, perhaps enhanced by underthrusting of the Lesser Himalayan sequence beneath the Greater Himalayan sequence, propagating a ramp on the Main Himalayan thrust toward the north. As the ramp advanced into the orogen, it caused the weak midcrust to destabilize, and a dome formed and was flattened and extruded to the south. Crystallized leucogranite dikes and sills in the hanging wall of the outer South Tibetan detachment underwent varying amounts of subsolidus ductile deformation before cooling at ca. 11 Ma (Fig. 9B). Coeval with rapid cooling of the outer South Tibetan detachment from ca. 12 to 10 Ma, ductile motion ceased on the Main Central thrust, and thrusting on the Main Boundary thrust began (Meigs et al., 1995; Stüwe and Foster, 2001; Daniel et al., 2003). The dome of Greater Himalayan sequence material was translated toward the foreland in the hanging wall of the Kakhtang thrust, cutting off the outer South Tibetan detachment from the inner South Tibetan detachment (Fig. 9C) and burying part of the outer South Tibetan detachment beneath the Kakhtang thrust. Note that an out-of-sequence thrust is not required in our interpretation, but doming and the corresponding bending of metamorphic isograds are. From that point, the outer South Tibetan detachment (and Chekha Group in its hanging wall) were passively carried southward in the hanging wall of the Main Himalayan thrust, and displacement occurring along the South Tibetan detachment at depth in the hinterland continued to be accommodated along the inner South Tibetan detachment. Subsequent to 11 Ma, continued displacement along the inner South Tibetan detachment, coeval with the Main Boundary thrust, facilitated rapid cooling and exhumation of footwall Greater Himalayan sequence rocks, and progressive overprinting of the ductile inner South Tibetan detachment by younger, brittle normal-sense faulting (Fig. 9D) (Edwards et al., 1996). By ca. 8–6 Ma, Greater Himalayan sequence rocks had cooled through the apatite fission-track closure temperature of ~110 °C (Grujic et al., 2006). However, the period of fast cooling and exhumation probably ended around the Miocene-Pliocene transition (Fig. 7; Grujic et al., 2006), allowing for the preservation of the outer South Tibetan detachment (appearing as klippen of Chekha Group and Tethyan sedimentary sequence rocks), which may have been lost to erosion elsewhere in the Himalaya. Our and published thermochronological data support a break in the cooling slope at around the end of the Miocene (Fig. 7).
Cooling through the Ar closure temperature of muscovite of metasedimentary rocks rimming the Kangmar dome, which is located to the north of NW Bhutan, occurred from 15 to 10 Ma (Maluski et al., 1988; Lee et al., 2000). This dome is thought to be an exhumed northern portion of the South Tibetan detachment that was brought to the surface by thrusting of that dome along a ramp in the Greater Himalayan sequence from 11.5 to 10.0 Ma (Lee et al., 2000). This deformation coincides with thrusting along the Main Boundary thrust, ductile shearing on the inner South Tibetan detachment, and possibly with thrusting along the Kakhtang thrust, as well as cessation of the outer South Tibetan detachment and the Main Central thrust, indicating that there was a major shift in deformation partitioning within the eastern Himalaya at 12–10 Ma.
The outer South Tibetan detachment appears to be unique to the eastern Himalaya, either because it did not form elsewhere, or it was not preserved. Erosion levels of the Greater Himalayan sequence in Bhutan are shallower than in the central Himalaya, most likely due to decreased erosion rates caused by a decline in mean annual precipitation caused by uplift of the Shillong Plateau from ca. 6 Ma (Grujic et al., 2006; Biswas et al., 2007). If outer segments of the South Tibetan detachment were similarly abandoned elsewhere in the Himalayan orogen, perhaps this difference in erosion can explain why they are not observed today.
Irrespective of the mechanism, once the outer South Tibetan detachment was abandoned in the mid-Miocene, it became coupled to the hanging wall of the Main Himalayan thrust and transported southward away from the active inner South Tibetan detachment. Thus, the outer South Tibetan detachment is now located in a more external position in the Himalayan orogen than when it was an active structure in the Miocene: its leading edge today does not mark the edge of the South Tibetan detachment in the Miocene.
The South Tibetan detachment system in Bhutan is composed of two distinct structures: the fossil outer South Tibetan detachment preserved in the cores of synforms in the Greater Himalayan sequence, and the younger inner South Tibetan detachment along the border between the Himalaya and Tibet.
Weakly deformed leucogranite sills emplaced in the hanging wall of the outer South Tibetan detachment are dated by zircon U-Pb SHRIMP geochronology at ≥15.5 Ma. Low zircon crystallization temperatures of ~500–800 °C as determined by Ti-in-zircon thermometry suggest subsolidus crystallization, but zircon textures and trace-element geochemistry point to magmatic crystallization in a low-temperature, highly fractionated melt. These rocks cooled through ~425 °C by ca. 11.0 Ma, as determined by muscovite 40Ar/39Ar thermochronology. Zircon in weakly deformed leucogranite emplaced in the footwall of the inner South Tibetan detachment crystallized at or prior to 11.0 Ma. Crystallization temperatures for magmatic zircon are ~700–800 °C, and they show more rapid cooling than that observed for leucogranites in the outer South Tibetan detachment.
These age constraints for the outer and inner South Tibetan detachment indicate that inner South Tibetan detachment ductile to brittle faulting coincided with both cessation of motion on the outer South Tibetan detachment and local out-of-sequence thrusting within the Greater Himalayan sequence. Furthermore, ductile shearing on both South Tibetan detachment segments appears to have been active more recently than ductile motion along the South Tibetan detachment system in the central and western parts of the orogen. The outer South Tibetan detachment was abandoned in the mid-Miocene and subsequently transported southward in the hanging wall of the Main Himalayan thrust to its present locations. While hinterland stepping of the South Tibetan detachment (frictional wedge model) is a possible mechanism for outer South Tibetan detachment abandonment, our data combined with published geochronologic constraints for the eastern Himalaya support foreland translation and exhumation of midcrustal domes (viscous wedge model) as the tectonic mechanism for outer South Tibetan detachment abandonment and subsequent southward translation.
We would like to gratefully acknowledge field support from Lhawang Dorji and our Bhutan field crew, analytical assistance from J. Wooden, F. Mazdab, P. Reynolds, and K. Taylor, and technical assistance and advice from D. Wheeler, A. Grist, and Y. Kettanah. Thanks are due to I. Coutand, C. Warren, and Y. Fedortchouk for discussions and to R. Carosi and an anonymous reviewer for providing comprehensive reviews that improved the manuscript. This work was supported by a Natural Sciences and Engineering Research Council of Canada (NSERC) grant to D. Grujic, and an NSERC Canada Graduate Scholarship–Doctoral scholarship and Geological Society of America research grant to D. Kellett.