Field, structural, kinematic, and deformation temperature analyses were conducted on rocks from the Lhagoi Kangri gneiss dome (southern Tibet) in order to establish the geologic history of the dome, identify major phases of deformation within the dome, and to relate these phases of deformation to the tectonic evolution of the Himalayan middle crust. The Lhagoi Kangri dome, one of a series of gneiss-cored domes in southern Tibet, records stratigraphy and structural features similar to previously studied north Himalayan gneiss domes. Field mapping reveals a sequence of rocks that comprise a cover of unmetamorphosed to amphibolite-grade siliciclastic and minor carbonate rocks overlying a core predominantly composed of foliated and lineated orthogneiss intruded by relatively undeformed granite, which also intrudes the cover rocks both concordantly and discordantly. Field observations and microstructural analyses suggest that the contact between the core and cover rocks was originally a nonconformity, but we do not rule out the possibility of subsequent slip along the surface, as has been reported for correlative structures in other domes. Lhagoi Kangri rocks were pervasively deformed during at least two major tectonic phases. The earliest deformation event (D1) resulted in shortening and thickening of crust, the record of which is largely eliminated, particularly in lower structural levels, by transposition and recrystallization during the second phase of deformation (D2). Ductile deformation during D2 is characterized at higher structural levels by crenulation cleavage that tightens with depth, while at lower structural levels D2 manifests as a distributed shear zone that records some evidence of both plane strain and coaxial flattening, possibly indicating overall heterogeneous general shear. The shear zone is ∼3 km thick and contains rocks with mostly symmetrical top-to-north and top-to-south shear sense indicators with a dominant top-to-north component at lower structural levels. Microstructural analyses and quartz c-axis fabrics indicate a range of D2 shear zone temperatures from 200 to 300 °C at the upper boundary to ≥630 °C at the lowest structural levels sampled with minimal evidence of lower temperature overprinting. The interpreted temperatures define a wide range in thermal field gradients (18–90 °C/km) that suggest that temperature indicators were locked in at relatively late stages of D2. The structural framework and kinematic history of the Lhagoi Kangri dome are similar to previously studied north Himalayan gneiss domes as well as to transects through the South Tibetan detachment system, which supports previous interpretations of structural continuity between the north Himalayan gneiss domes and other middle crustal exposures in the Himalaya. The Lhagoi Kangri distributed shear zone, in particular, may represent a deeper ductile manifestation of the South Tibetan detachment system.


The North Himalayan antiform delineates a series of structural domes, the north Himalayan gneiss domes (NHGD), that trends approximately parallel to the orogenic front (Fig. 1). Granite and orthogneiss within the cores of the NHGD are often interpreted to correlate to rocks of the Greater Himalayan sequence (GHS), the Himalayan tectonostratigraphic unit representing exhumed middle to lower crust (e.g., Nelson et al., 1996; Zhang et al., 2004; Lee et al., 2006; Lee and Whitehouse, 2007; King et al., 2011). Separating the GHS from the overlying Tethyan sedimentary sequence (TSS) is the South Tibetan detachment system (STDS), a series of north-dipping, low-angle normal faults and top-to-north sense shear zones that can be traced nearly continuously along strike of the orogen (e.g., Burg et al., 1984; Burchfiel et al., 1992; Kellett and Grujic, 2012). In many NHGD, the contact between the core and the TSS is a top-to-north-sense shear zone that is interpreted as an exhumed, downdip projection of the STDS (e.g., Chen et al., 1990; Lee et al., 2000, 2006; Quigley et al., 2006; Larson et al., 2010; King et al., 2011). This interpretation is also corroborated by geophysical evidence at the approximate longitude of the Kangmar dome (Fig. 1; Nelson et al., 1996). Given their unique position within the Himalaya-Tibet orogenic system and their correlation to the GHS, the NHGD have been the subject of research questions with direct implications to Himalaya tectonics, i.e., (1) if there is structural continuity between the GHS and the NHGD and how is this manifested in the domes, (2) how the NHGD were formed and exhumed, and (3) the structural and stratigraphic relationships between core and carapace rocks in the gneiss domes. Our study represents the first comprehensive investigation of the sparsely documented Lhagoi Kangri dome (LKD). In this contribution, data from field and detailed structural analyses are used to establish a general geologic history for LKD rocks that in turn is used to revisit the above questions and to provide additional data to test previous models regarding the tectonic evolution of the Himalayan middle crust.

Previous Work

At least 15 domes have been identified along the North Himalayan antiform (e.g., Burg et al., 1984; Debon et al., 1986; Watts et al., 2005), and six of these domes have been studied in detail. Of the six previously studied domes, Changgo (Larson et al., 2010) Malashan (Aoya et al., 2005; Kawakami et al., 2007; Gao et al., 2013; Gao and Zeng, 2014), Yardoi (Aikman et al., 2008; Zeng et al., 2009, 2011, 2014), and Mabja-Sakya (Zhang et al., 2004; Lee et al., 2006; Lee and Whitehouse, 2007) contain granite bodies with Eocene to Miocene (35–10 Ma) U/Pb zircon ages, whereas Kangmar (Schärer et al., 1986; Lee et al., 2000) and Kampa (Quigley et al., 2008) contain only granitic gneiss with Neoproterozoic to Cambrian (560–500 Ma) U/Pb zircon ages. Granitic gneiss with Neoproterozoic to Cambrian age is also reported in Mabja-Sakya (Lee and Whitehouse, 2007) and Yardoi (e.g., Zeng et al., 2009). These data support two end-member origins for the gneissic cores: (1) Eocene–Miocene synorogenic plutons that intruded their host rock (e.g., Malashan and Changgo domes; Aoya et al., 2005; Larson et al., 2010), or (2) granites that crystallized in the Neoproterozoic–Cambrian and were subsequently metamorphosed and deformed during the Eocene–Miocene (e.g., Kangmar and Kampa; Lee et al., 2000; Quigley et al., 2008). The Mabja-Sakya dome is interpreted as a mixture of the two end members: Neoproterozoic–Cambrian orthogneiss that was migmatized and intruded by anatectic melts during the Oligocene–Miocene (e.g., King et al., 2011). In terms of the early tectonic framework of the Himalaya-Tibet system, these results indicate that the cores of the NHGD are the structurally highest levels of the GHS that underwent Cenozoic metamorphism and were deformed in a northern equivalent (e.g., Quigley et al., 2008) or a distributed mid-crustal portion of the STDS (Lee et al., 2000, 2004; Zhang et al., 2004; King et al., 2011).

Tectonic Models

Several mechanisms have been proposed for formation of the North Himalayan antiform and doming of NHGD rocks. Le Fort et al. (1987) attributed doming in the NHGD to buoyant rise of lower crustal melt, whereas Burg et al. (1984) suggested that the Kangmar dome was produced by formation of a thrust duplex. In contrast, Chen et al. (1990) preferred an interpretation where the Kangmar dome formed exclusively from extension, likening it to metamorphic core complexes in the North American Cordillera. Lee et al. (2006) interpreted Mabja-Sakya doming to be the result of out-of-sequence, south-vergent thrusting, and patterns in 40Ar/39Ar cooling ages from Kangmar dome are also suggested to arise from ramping on a blind thrust (Lee et al., 2000). Similarly, Larson et al. (2010) suggested that doming and exhumation of the Changgo culmination resulted from out-of-sequence thrusting and folding, while Yin et al. (1999) and King et al. (2011) concluded that doming of the NHGD could also have resulted from rise of melt-weakened GHS in response to regional horizontal compression. Thermomechanical modeling of the Himalayan orogeny produced NHGD-like features from a southward-flowing mid-crustal channel through either contemporaneous, localized upper crustal extension or by underthrusting triggered by impingement of a rigid body, such as the Indian lithosphere, into the channel (Beaumont et al., 2004). In these models the STDS represents the upper boundary of the channel. Recent studies on structural transects through the STDS south-southwest of LKD emphasize middle crustal strain patterns, particularly that the degree of apparent telescoping of thermal field gradients varies spatially, to support channel flow or wedge extrusion models (e.g., Cottle et al., 2011; Law et al., 2011). At the very least, the apparent STDS strain patterns are most compatible with Himalayan tectonic models in which the STDS and related structures accommodated appreciable amounts of displacement during Oligocene–Miocene time (e.g., Cottle et al., 2011; Law et al., 2011).


The LKD is a ∼20 × 40 km elliptical exposure with an east-west–trending long axis located 100 km north-northeast of Mount Everest (Fig. 1; Tibet Bureau of Geology and Mineral Resources, 1993; Watts et al., 2005). Burg et al. (1984) originally identified the LKD during regional reconnaissance mapping. Schärer et al. (1986) published zircon (501 Ma) and monazite (15.5 ± 0.8 Ma) 206Pb/238U ages from a single sample of Lhagoi Kangri leucogranite, but they did not elaborate on where the sample was collected or its geologic context. Rolfo et al. (2004) presented preliminary data regarding the lithology, structure, and metamorphic conditions recorded in LKD rocks that they interpreted to be similar to that of other NHGD. Watts et al. (2005) produced a geologic map of LKD using band ratios of thermal emission imagery from ASTER (Advanced Spaceborne Thermal Emission and Reflection Radiometer) that showed the presence of both gneiss and a two-mica granite body, the Gyaco La pluton (Debon et al., 1986), in the core of the dome. In addition, they identified a crescent-shaped body in map view as a muscovite-rich schist or quartzite layer that emphasizes the domal geometry.

Lhagoi Kangri Lithology

The rock types composing the LKD broadly correspond to the sequence defined for the adjacent Mabja-Sakya dome (Lee et al., 2004) and that of surrounding southern Tibet as originally mapped by the Tibet Bureau of Geology and Mineral Resources (1993). LKD age assignments follow the conventions of previous regional stratigraphic and field mapping studies (e.g., Tibet Bureau of Geology and Mineral Resources, 1993; Liu and Einsele, 1994). Despite the relative absence of vegetation, both exposure and accessibility throughout the LKD are poor. As a result, the number of mapped units is intentionally parsimonious (Figs. 2–4).

The structurally lowest formation in LKD, unit og, is a dominantly felsic orthogneiss complex composed of intervals with variable mode, texture, and local ptygmatic segregations of leucosome (Fig. 4). The orthogneiss forms rounded subcrops at the highest elevations in LKD (Fig. 5A), and its exposure is typically limited to felsenmeer, except along cliff lines where it is sheltered by overlying metasedimentary rocks. Most orthogneiss documented in this study is coarse grained and contains the assemblage quartz + plagioclase + biotite with large (1–5 cm) K-feldspar augen. In particular instances the orthogneiss also contains sillimanite and/or cordierite. At lower structural levels, the orthogneiss unit contains amphibolite, dikes of undeformed pegmatite, and fine-grained leucogranite. Within ∼200 m of the contact with the overlying schist and quartzite, the orthogneiss is fine grained and comprises garnet- or cordierite-bearing assemblages. Orthogneiss exposures typically exhibit a well-developed foliation defined by sheets of biotite ± white mica and a lineation of aligned quartz aggregates that is typically weakly developed if present it all. The contact between unit og and the overlying Paleozoic schist and quartzite (Ps) has high structural relief (∼1 km) along its trace. The nature of the contact is not clearly defined due to poor exposure. However, in the central portion of the dome where the most detailed sampling transects were carried out (Fig. 2), the orthogneiss is interlayered with quartzite, indicating an originally intrusive contact.

The lowest metasedimentary unit is a Paleozoic package (Ps) of micaceous and quartzose schist (Fig. 5B), quartzite, and paragneiss with minor calc-silicate (Fig. 5C). Throughout the LKD, unit Ps forms distinct reddish-brown to yellow cliff bands, although it also manifests as blockfields that are difficult to distinguish from those of the underlying orthogneiss. All unit Ps rock types exhibit well-developed foliation and poorly to strongly developed lineation. In schist and quartzite outcrops, the main foliation is defined by aligned micas and it contains a mineral lineation defined by elongate quartz and biotite as well as by elongate porphyroblasts like staurolite and kyanite. Foliation in the calc-silicate intervals manifests as green, beige, and purple layers (e.g., Fig. 5C) that presumably reflect compositional domains. Where present, lineations in calc-silicate are defined by elongate quartz and biotite aggregates. Quartzite mineralogy in unit Ps appears to vary systematically with depth. Structurally lower quartzite horizons contain biotite as the only mica phase, whereas white mica joins biotite at higher structural levels. Rocks in the lower one-third of unit Ps are intruded by centimeter- to meters-thick leucogranite and pegmatite bodies with variable contact geometry: some are concordant with the main foliation and are isoclinally folded and/or boudinaged, while other bodies intrude obliquely to the dominant fabric and are largely undeformed (e.g., Figs. 5B, 5D). Within 100 m of the contact with the subjacent orthogneiss, unit Ps contains amphibolite layers that are concordant with the main foliation and with compositional contrasts in the metasedimentary host. Garnet-andalusite schist and epidote-clinopyroxene calc-silicate are also present near the contact with the orthogneiss and are locally spatially related to the amphibolite layers; however, particularly on the south side of the dome, calc-silicate layers also delineate a transition between Ps schist and overlying Paleozoic marble (Pm). The upper contact of Ps is relatively sharp where exposed, and along its trace it juxtaposes both Pm and the Triassic siliciclastic unit (Ts) over reddish-brown biotite-bearing quartzose schist.

Unit Pm is a discontinuous, highly weathered quartz- and phlogopite-bearing marble with a maximum thickness of ∼350 m and thinning to zero on the north side of the dome (Figs. 2 and 3). This observation is consistent with previous regional mapping that shows Permian–Carboniferous limestone deposits pinching out north of the LKD (Liu and Einsele, 1994). Marble outcrops are scarce throughout the study area, but they form narrow yellow cliff bands where present. The marble contains a pervasive foliation defined by alternating yellow and white bands that represent relatively quartz-rich and quartz-poor intervals. Where present, lineations on primary foliation surfaces are defined by elongate quartz and biotite aggregates. The upper contact of unit Pm is not well exposed in most accessible areas of the LKD, but the marble is a distinct lithologic package in that it is the only calcite-rich rock type observed in the metasedimentary carapace, apart from vein fill in epidote-bearing intervals of unit Ps.

Within the study area defined, Triassic siliciclastic rocks (Ts) dominate the highest structural levels on all sides of the dome (Figs. 4 and 5A). Unit Ts stands out in contrast to the other metasedimentary rocks in that it typically forms spectacular ridgelines of ≥1 km of relief (e.g., Fig. 5E). The lower portion of Ts consists of pervasively foliated, dark gray chloritoid- and garnet-bearing graphitic phyllite, and the upper portion consists of gray interbedded carbonaceous argillite, siltstone, and fine-grained sandstone with minor quartz pebble conglomerate. On the north side of the dome, finer grained intervals in the lower portion are phyllitic and crenulated with small, randomly oriented laths (0.1–0.2 mm) of chloritoid that overprint the crenulations.


Based on field observations and structural analysis, two ubiquitous deformation phases are recorded in LKD rocks. Structures that define the first phase (D1) are not well documented in the field. Instead D1 is implied by pervasive cleavage development at higher structural levels that is progressively overprinted by a second pervasive deformation event (D2) over a few hundred meters down structural section (Fig. 3). D2 is characterized by a pervasive foliation and mineral stretching lineation at lower structural levels within the LKD (Figs. 3 and 6). In the lowermost schist, D1-related fabrics are only weakly preserved, and D1 fabric is indistinguishable in the core orthogneiss. In addition to pervasive D1 and D2 deformation, metasedimentary cover rocks preserve structures that are only observed locally (D3?), such as open folds with gently northeast-inclined axial planes that refold mesoscale F2 folds. Brittle deformation features are relatively rare in the LKD, but joint sets that strike approximately north-south and dip steeply east are present on both the south and north sides of the dome. Given the paucity of data relating to other deformation phases, this study focuses primarily on D2 deformation. The study area investigated comprises four structural domains that are defined by variation in average dip direction of S2 foliation (Figs. 6B–6E), and each domain encompasses one or two sampling transects (Fig. 2). The northern and southern domains are the most complete structural sections, from weakly metamorphosed unit Ts to the core orthogneiss. Consequently, the northern and southern transects are the only ones where unobscured D1 fabrics are documented. The central domain comprises ∼600 m of section above the core-cover contact plus ∼100 m below it, while the eastern domain represents only ∼500 m of the metasedimentary cover rocks, spanning the lowermost Triassic siliciclastic unit down to the uppermost Paleozoic schist.

D1 Deformation

Mesoscale Observations

Mesoscale D1 is ambiguous throughout much of the study area because it is generally weakly developed at higher structural levels and progressively overprinted by D2 at lower structural levels, factors that compound poor accessibility to parts of the dome. Within the areas investigated, D1 fabric manifests as a discrete cleavage (S1) at a low angle to compositional layering in unit Ts argillite. In the northern domain, S1 cleavage is more apparent at low structural levels where it forms well-developed pencil cleavage with intersection lineations (L2x1) that plunge gently to the north-northwest (Fig. 7A). In the southern domain D1 fabric is ambiguous except where it is crenulated by D2 (Fig. 7B).

Microscale Observations

Samples at the highest structural levels in the northern and southern domains preserve a weakly developed microscale fabric that, similarly to the mesoscale counterpart, is subparallel to compositional heterogeneity (Fig. 7C). S1 fabric is defined by sheets of chlorite, biotite, ilmenite, and quartz. Microscale fabrics are also most easily identified at lower structural levels where they are crenulated by S2 foliations in phyllitic intervals (Fig. 7D).

D2 Deformation

Mesoscale Observations

Regional D2 deformation is characterized by a penetrative foliation and mineral stretching lineation throughout most of the structural section sampled during this study. S2 progresses from crenulation cleavage at higher structural levels to a continuous, gently to moderately dipping schistosity that is concentric about the dome (Figs. 2, 3, and 6). In the northern domain S2 also defines axial planes that dip gently to moderately northward. L2 stretching lineations are defined by elongate porphyroblasts like chloritoid and staurolite in fine-grained metasedimentary rocks and by aligned quartz and biotite aggregates in coarse-grained metasedimentary intervals and in the core orthogneiss. In the northern and central domains, S2 dips northward and L2 lineations plunge approximately downdip (Figs. 6B, 6C). In the eastern domain S2 dips gently east while L2 mineral stretching lineations trend north-northeast, approximately parallel to the strike of S2 (Fig. 6D). The southern domain contains S2 foliation that generally dips southward instead of northward, and trends of L2 lineations are scattered (Fig. 6E). At moderate structural levels in the southern domain, finer grained intervals record well-developed F2 crenulations with average S2 axial planar attitude of 077/45°SE (e.g., Fig. 7B) that is not as prominent in the northern domain.

Microscale Observations

Microstructural analyses of LKD samples indicate variation in D2 fabric development with structural depth similar to the mesoscale variation. Starting in the Triassic schist and progressing downsection into unit Ps and the orthogneiss, S2 becomes increasingly pervasive, forming a continuous aligned-quartz and mica-chlorite cleavage in quartz-rich intervals and a domainal cleavage with submillimeter spacing defined by quartz microlithons and biotite ± white mica in pelitic intervals (Figs. 8A, 8B). Marble (unit Pm) from the northern domain also contains a pervasive S2 foliation defined by biotite aggregates. In unit Ps, F2 crenulations are commonly preserved within the S2 mica domains even at lower structural levels where S2 dominates the fabric (Fig. DR11). Some pelitic schist (unit Ps) samples from the central domain contain elongate, kinked porphyroblasts such as kyanite and plagioclase with tails that merge with well-developed S2 schistosity (e.g., Fig. 8C).

The core orthogneiss also exhibits a variety of D2 textures, from fine grained and weakly foliated to coarse grained with well-developed gneissic banding (Fig. DR1). S2 foliation in coarser grained intervals is defined by elongate biotite (±white mica) that forms continuous 0.5–1.0-mm-thick bands, while the finer grained fabric is defined by thin (≤0.5 mm) discontinuous biotite layers and isolated individual biotite grains with a preferred orientation. The orthogneiss leucosome also contains alkali feldspar porphyroclasts. In samples from both the northern and central domains, K-spar porphyroclasts have myrmekite lobes that are localized along grain boundaries parallel to the matrix foliation (Fig. 8D).

D2 Kinematic Indicators

Mesoscale features throughout the LKD record shear sense associated with D2 deformation. Centimeter-thick shear bands with decimeter-scale spacing that are subparallel to the regional S2 foliation are preserved in unit Ps on the north side of the dome and record mixed top-to-south-southwest and top-to-north-northeast shear sense (e.g., Fig. 9A). Shear bands are also preserved in the core orthogneiss, although the example documented (Fig. 9B) is not in situ. Rocks in the northern and central domains also contain boudin trains comprising veins of quartz ± calcite. At the resolution of the study only three examples of boudinage were documented, but these examples derive from various structural levels within the metasedimentary cover sequence, and all are accompanied by pervasive S2 schistosity (Figs. 5C, 9C, and 9D). Boudin attitudes were not measured, but in one instance where exposure and fracturing allowed observation on X/Z and approximate Y/Z planes, their geometry suggests that the boudins are three-dimensional chocolate tablets (Ghosh, 1988) that are subparallel to the local attitude of S2 (265/20°NE).

Microscale kinematic indicators such as strain shadows on porphyroblasts and C′-type shear bands, in both the metasedimentary rocks and the orthogneiss, variably record symmetrical and asymmetrical shear sense. Near the core-cover contact, particularly in the central domain, samples record dominantly top-to-north-northeast shear (Table 1; Figs. 3 and 6). In the northern and eastern domains, garnet and chloritoid porphyroblasts in unit Ts graphitic schist contain both symmetric (Φ type) and mixed top-to-north-northeast and top-to-south-southwest σ-type strain shadows in S2 matrix schistosity (Fig. 10A; Fig. DR1). Unit Ps in the northern and central domains contains garnet and staurolite porphyroblasts framed by σ-type strain shadows of quartz with chlorite and biotite that record mixed top-to-north-northeast (Fig. 10B) and top-to-south-southwest (Fig. DR1) shear sense. Garnet schist in the southern domain also contains porphyroblasts with asymmetric strain shadows on garnet with both top-to-south-southwest and top-to-north-northeast shear sense (Table 1), although data are too sparse to suggest an apparent relationship between shear sense and structural position as noted for the central domain.

The lowermost schist and orthogneiss in the northern, central, and southern domains contains microscale C′-type shear bands, mica fish, and, in the orthogneiss, asymmetric myrmekite. The C′-type shear bands in these units record both top-to-north and top-to-south shear sense, but are predominantly top-to-north indicators (e.g., Figs. 10C, 10D; Fig. DR1; Table 1). Mica fish in schist and paragneiss intervals also record mixed top-to-north and top-to-south shear sense where present (Fig. 10E; Fig. DR1). Some K-feldspar porphyroclasts in the LKD orthogneiss exhibit myrmekite overgrowths and recrystallized tails similar to features that Simpson and Wintsch (1989) identified as deformation induced and suggested were viable kinematic indicators. In LKD samples shear sense inferred from myrmekite overgrowth is consistent with that from porphyroclast tails and matrix shear fabric (e.g., Fig. 10F), suggesting their viability as kinematic indicators in the LKD as well. However, myrmekite could also have grown prior to D2 deformation, and it is beyond the scope of this study to determine myrmekite growth mechanisms.

D2 Quartz, Feldspar, and Calcite Microstructures

Background and methods. Quartz and feldspar microstructures (e.g., Tullis and Yund, 1987; Hirth and Tullis, 1992; Pryer, 1993; Stipp et al., 2002b) are documented for 28 samples, and calcite twin morphology (Ferrill et al., 2004) is documented for 4 samples (Table 1). The implications these microstructures have for conditions during D2 deformation in the LKD are addressed in the Discussion.

Quartz microstructures in LKD samples fall into two of three recrystallization types originally defined by Stipp et al. (2002b) for naturally deformed quartz in the Italian Alps: (1) subgrain rotation recrystallization (SGR) and (2) grain boundary migration recrystallization (GBM), which can be further subdivided into GBM I and GBM II on the basis of impurity-hindered versus non-impurity-hindered grain boundary migration. In that study temperatures associated with the quartz recrystallization mechanisms were constrained by thermobarometry of synkinematic mineral assemblages. The temperatures determined are 400–500 °C for SGR, >500 °C for GBM, and 630 ± 30 °C for chessboard extinction, which is a characteristic feature of GBM II. Quartz dynamic recrystallization microstructures are also controlled by other factors, such as strain rate and hydrolytic weakening (e.g., Kronenberg and Tullis, 1984; Post and Tullis, 1998; Stipp et al., 2006); Stipp et al. (2002a) demonstrated that recrystallization mechanisms in the Alps samples were also influenced by strain rate.

Feldspars exhibit dislocation creep mechanisms similar to that of quartz (e.g., Tullis and Yund, 1987, 1991), and plastic deformation usually occurs in naturally deformed samples at ∼450 °C and higher (e.g., Pryer, 1993). However, deformation mechanisms in feldspar are also influenced by strain rate and hydrolytic weakening (e.g., Tullis and Yund, 1991). Feldspar grains can also fracture or boudinage at higher temperatures if they are included in a rheologically weaker matrix such as quartz and mica (e.g., Fitz Gerald and Stünitz, 1993; Stünitz and Fitz Gerald, 1993).

In coarse-grained limestone and marble, calcite twinning is an important deformation mechanism at low temperatures (≤400 °C), and Ferrill et al. (2004) determined that there is a positive correlation between calcite twin width and deformation temperature such that thin twins (Type I, <1 µm) are dominant at temperatures <170 °C while thick twins (Types II–IV, >1 µm) are dominant at temperatures >200 °C. The geometry of the twin boundaries varies with increasing temperature, from straight boundaries formed at lower temperature (Type II), curved at moderate temperature (Type III), to patchy (indicative of partial dynamic recrystallization) at higher temperatures (Type IV, >250 °C). Complete dynamic recrystallization typically occurs at temperatures ≥300 °C (e.g., Weber et al., 2001; Ferrill et al., 2004). However, in calcite veins or localized fault zones, incipient dynamic recrystallization can begin at temperatures lower than 250 °C in response to high strain rates (Ferrill et al., 2004). Calcite twinning is also influenced by total strain, and so interpretations of calcite twin morphology must be made cautiously and with consideration of alternatives.

LKD calcite microstructures. Representative samples of phlogopite-bearing marble were collected from all three domains, and exhibit a range in calcite microstructures from Type I through Type IV twins (Table 1; Fig. 11; for additional photomicrographs, see Fig. DR1). A sample from the southern domain (LK-04) exhibits Type I and II calcite twins (Fig. 11A), whereas a correlative sample from the eastern domain (LK-65) exhibits Type III and IV twins (Fig. 11B), indicating at least partial dynamic recrystallization. From the northern domain, sample LK-17 (Table 1; Fig. DR1) contains Type I and II twins as well as thick twins with tapered boundaries (Type III). Patchy twin boundaries (Type IV) are not recorded in this sample, but clusters of small, unstrained grains surrounding large twinned grains indicate dynamic recrystallization in this sample as well (Fig. DR1).

LKD quartz and feldspar microstructures. The Guodong and Muddy Lake transects from the central domain (Figs. 2 and 6; Table 1) are representative of the range in quartz and feldspar recovery mechanisms associated with D2 deformation. The following observations are presented in order of highest to lowest structural level, summarized in Table 1, and shown in Figures 12A–12F. For additional examples of quartz and feldspar microstructures, see Figure DR1.

Sample LK11-74 is a mica-bearing quartzite that marks the top of the Guodong transect, sampled ∼600 m above the contact with the core orthogneiss. LK11-74 contains dynamically recrystallized quartz grains with amoeboidal grain boundaries characteristic of GBM I and subgrain formation indicating SGR (Fig. 12A). These quartz microstructures are dispersed throughout the sample rather than partitioned into particular domains (e.g., microscale shear zones), which suggests that both dynamic recrystallization mechanisms (SGR and GBM) were operative contemporaneously. Structurally lower quartzite and quartz schist samples from the transect exhibit dominant GBM I recrystallization (Table 1) with grain interiors that are free of subgrains (e.g., Fig. 12B). Orthogneiss at the base of the transect also exhibits quartz grain boundary migration microstructures and chessboard extinction (collectively, GBM II).

The adjacent Muddy Lake transect spans a range in structural levels similar to that of the Guodong transect and contains abundant quartzite that is compositionally similar to LK11-74 (i.e., impure quartzite with biotite ± white mica flakes). The Muddy Lake quartzite samples are characterized by quartz GBM microstructures at all structural levels covered by the transect (Table 1; e.g., Fig. 12C). At lower structural levels, starting ∼150–250 m above the orthogneiss contact, there is an apparent shift in recrystallization mechanism from impurity-hindered (GBM I) to non-hindered grain boundary migration (GBM II), and chessboard extinction is also more common at lower structural levels (Table 1; Fig. 12D; Fig. DR1).

Feldspar microstructures documented indicate an apparent correlation with structural depth. At the upper part of the Guodong transect, feldspar grains exhibit brittle microcracks and undulose extinction while quartz grains in the same sample exhibit GBM I microstructures (LK11-73; Fig. 12E; Table 1). Deformation lamellae and undulose extinction are not evident in quartz grains, suggesting that (1) the quartz and feldspar microstructures developed under similar conditions, but the quartz was rheologically weaker (e.g., Stünitz and Fitz Gerald, 1993), or (2) the feldspar microstructures were inherited from some previous deformation. In contrast, many orthogneiss samples from the Guodong and Muddy Lake transects contain feldspar grains with bulging grain boundaries and subgrain structures (e.g., Fig. 12F; Table 1; Fig. DR1). In these same orthogneiss samples, quartz grains generally record GBM II and chessboard extinction microstructures.

The southern transect encompasses a wider range in structural levels, and sampling resolution is much lower than in the Guodong and Muddy Lake transects, but the samples collected suggest a similar correlation between structural depth and quartz and feldspar microstructures. The structurally highest quartz schist sample (LK-05; Table 1) records quartz SGR microstructures, and at progressively lower structural levels there is an apparent shift to GBM I (sample LK-48) and then to GBM II (sample LK-52) microstructures.

D2 Quartz Crystallographic Preferred Orientation Patterns

Background and methods. Quartz crystallographic preferred orientation (CPO) patterns (Lister et al., 1978; Lister and Hobbs, 1980; Lister and Dornsiepen, 1982; Mainprice et al., 1986; Schmid and Casey, 1986; Law, 1990; Kruhl, 1998) were analyzed from 10 LKD samples (Table 1). The CPO patterns were collected by electron backscatter diffraction (EBSD) analysis using an FEI Quanta 400 scanning electron microscope equipped with a Nordlys 2 EBSD camera and housed at the University of California, Santa Barbara. Samples chosen for EBSD analysis mostly overlap with those chosen for quartz and feldspar microstructural analysis, and all samples are derived from structural levels that generally encompass the regime of GBM I recrystallization (Table 1; Fig. 13). Thin sections were polished using 0.25 µm diamond grit followed by 3 h on a vibratory polisher coated with 0.05 µm colloidal silica gel. The samples contain large grain sizes and weakly developed, relatively homogeneous fabric. For these reasons, EBSD analyses for each sample were carried out over the entire area of a standard rectangular thin section (27 × 46 mm) using a 100 µm step size, with the exceptions of samples LK11-57 and LK11-87b, which were carried out over approximately half of the total thin section area at a step size of 100 µm. Analyses were set up to only index quartz and were processed using HKL CHANNEL 5 EBSD software. Data processing involved filling unindexed pixels in which five neighbor pixels share a common orientation. Individual grains are defined on the basis of a 15° critical misorientation angle. Because grain sizes in the LKD samples are generally larger than the step size, CPO plots presented in Figures 13B–13D were plotted using a one point-per-grain (PPG) approach in which individual grains are given equal weight, although for the LKD samples there is little or no difference in CPO fabric between raw and PPG data (e.g., Fig. DR2).

Quartz CPO patterns, particularly [c]-axis orientations, reflect active crystallographic slip systems during deformation (e.g., Schmid and Casey, 1986) that are controlled predominantly by temperature, strain rate, strain path, and degree of hydrolytic weakening (Tullis et al., 1973; Lister, 1977; Lister and Hobbs, 1980; Lister and Dornsiepen, 1982; Tullis and Yund, 1989; Law, 1990). In practice there are several possible quartz crystallographic slip systems (e.g., Schmid and Casey, 1986), but only basal <a>, prism <a>, and prism [c] slip are relevant to the LKD samples. Basal and prism define the lattice planes of the crystal along which slip occurs, while <a> and [c] refer to crystallographic orientations that define the slip direction. In instances where orientations of [c] axes from a sample viewed parallel to the XZ plane of strain (i.e., perpendicular to foliation, parallel to lineation) form a crossed-girdle pattern, the opening angle, defined as the angle between the two girdles (e.g., Fig. DR2), has a positive correlation to temperature, although it can also be affected by strain rate, shear strain amount, and hydrolytic weakening (Tullis et al., 1973; Law, 1990; Kruhl, 1998; Law et al., 2011). Modeling (Lister, 1981) and experimentation (Gleason et al., 1993) suggest that changes in [c] fabric opening angle are related to a shift in dominant crystallographic slip system from basal <a> to prism [c] as a function of increasing temperature; however, it is important to emphasize that water and strain rate influence the activity of these systems.

Quartz CPO patterns. CPO patterns from the LKD are weakly to moderately developed in quartzite samples and weakly developed in two orthogneiss samples, one quartz schist sample, and one quartz vein sample (Fig. 13).

Three samples from the southern transect (Fig. 13B) bracket 500–700 m of structural section within unit Ps. The structurally lowest sample, LK-51, is from a foliation-parallel quartz vein that exhibits a relatively strong [c] axis maximum. The next lowest sample (LK-54) from the southern transect is a quartzite derived from ∼200 m structurally below LK-51. The [c] axis pattern for sample LK-54 is defined by a strong Y-axis (central) maximum and a weakly developed crossed-girdle fabric that indicates dominant prism <a> slip and a component of basal <a> slip. The [c] fabric opening angle is 72–73° (Fig. 13B), and asymmetry in the peripheral arms of the crossed girdle, as well as asymmetry in the strength of <a> axis fabric, indicate a top-to-north shear sense. The structurally lowest EBSD sample from the southern transect (LK-43) is quartz schist from ∼500 m below LK-54. The [c] axis plot for sample LK-43 resembles a CPO indicative of a combination of prism <a> and prism [c] slip (e.g., Lister and Dornsiepen, 1982; Bouchez et al., 1984; Morgan and Law, 2004), but the pattern is too weak to decipher.

One quartzite sample from the structurally highest level of the Guodong transect (LK11-74; Table 1; Fig. 13C) has a CPO with an elongate central [c] axis maximum, indicative of prism <a> slip, and weakly developed peripheral arms that resemble a Type I crossed-girdle pattern (e.g., Lister and Dornsiepen, 1982). The contour pattern is weak, but point data yield a fabric skeleton (Fig. DR2) with an opening angle of 57–60°.

Samples from the Muddy Lake transect characterize CPO variation with structural level over ∼300 m of lowermost unit Ps and another ∼100 m of the underlying orthogneiss (Fig. 13D; Table 1). Three of four quartzite samples from unit Ps (LK11-57, LK11-54, LK11-48) exhibit well-defined CPO patterns with a characteristic [c] axis maximum near the center of the plot. The stereogram for the fourth quartzite sample (LK11-52) appears to be aligned obliquely to the XZ plane, but the amount of rotation required to produce an interpretable CPO pattern is inconsistent with the orientation of the thin section cut relative to fabric in the hand specimen. The two structurally highest samples (LK11-57 and LK11-54) exhibit well-developed Type I crossed girdles that comprise slightly weaker intermediate and peripheral maxima and indicate a component of basal <a> slip in addition to prism <a> slip implied by the strong central maxima. In sample LK11-54 the [c] fabric peripheral arms and density of <a> axes record asymmetry that indicates top-to-north-northeast sense shear. Opening angles for samples LK11-57 and LK11-54 are 76–77° and 74–75°, respectively. Sample LK11-48, located proximal to the orthogneiss contact and ∼70 m below LK11-52, exhibits a CPO characterized by a well-developed central maximum but no peripheral arms, indicating dominant prism <a> slip. Samples LK11-87b and LK11-90 are biotite orthogneiss located 85 and 100 m, respectively, structurally below the contact with the overlying metasedimentary rocks (∼85–100 m below sample LK11-48). Compared to the quartzite, the orthogneiss samples have relatively poorly defined CPO patterns. While they do exhibit faint Type I crossed-girdle fabrics, the most clearly defined aspect of the orthogneiss fabric is a Y axis maximum that indicates dominant prism <a> slip in sample LK11-87b.

Porphyroblast Inclusion Trails

Porphyroblast inclusion trail geometry was documented in order to constrain relative timing of deformation and metamorphism. Sigmoidal and spiral inclusion microstructures are commonly used to interpret early stages in the strain paths of deformed rocks (e.g., Passchier et al., 1992; Williams, 1994; Johnson, 1999; Williams and Jiang, 1999; Trouw et al., 2008). In LKD samples that contain them, porphyroblast inclusion trails exhibit variable geometry with respect to matrix S2 fabric, constrain relative timing of porphyroblast growth and D2 fabric development, and in some instances record D2 kinematics.

Sigmoidal inclusion trails are preserved in chloritoid, garnet, and staurolite porphyroblasts from most pelitic rock samples with S2 matrix fabric (units Ts and Ps). The curved trails are defined by elongated individual grains and aggregates of smaller grains that align to form the trails (Figs. 14A–14E). In a representative unit Ts chloritoid porphyroblast (Fig. 14A), inclusion trails are continuous with matrix S2 fabric and the porphyroblasts are framed at top and bottom by microfolds. Garnet porphyroblasts in unit Ts contain inclusion trails that are commonly deflected at the grain boundary (e.g., Fig. 14B), but they are still continuous with matrix S2 fabric. Similarly, garnet porphyroblasts in pelitic samples of unit Ps have sigmoidal inclusion trails that are typically continuous, or at least curve into parallelism, with S2 matrix fabric (e.g., Figs. 14C–14E). Staurolite inclusion trails are rare and poorly defined when present. In a representative sample, staurolite porphyroblast inclusion trails are inclined to matrix fabric near the porphyroblast cores and are curved but continuous with matrix fabric near the rims (LK11-60; Fig. 14F). In contrast to the simpler inclined or sigmoidal trails recorded in most porphyroblasts, the structurally lowest metapelite sampled in the central domain (LK11-42; Table 1) contains snowball garnet characterized by spiral-shaped inclusion trails (Figs. 14G, 14H). The spiral shapes are defined not only by curvilinear trails, but also by arc-shaped clusters of inclusions (Fig. 14G).

In LKD samples, most porphyroblast inclusion trails are continuous with matrix foliation (e.g., Figs. 14A, 14B, 14D–14F), suggesting intertectonic (between D1 and D2) to syn-D2 porphyroblast growth (e.g., Trouw et al., 2008; Johnson, 1999; Lee et al., 2000, 2004). In the samples where sigmoidal inclusion trails are not continuous with S2 matrix foliation (e.g., LK-52; Fig. 14C), the trails still appear to curve into parallelism with matrix fabric and also indicate that their hosts are intertectonic to syn-D2. In the case of LK11-02a (Fig. 14A), the microfolds adjacent to the porphyroblasts are best interpreted as deflection planes (Passchier et al., 1992) that indicate porphyroblast rotation during D2 fabric development and, together with inclusion trail and strain shadow asymmetry, record top-to-310° shear sense. Likewise, the geometry of garnet-hosted inclusion trails in LK-67a and the continuity between inclusion trails and matrix foliation are indicative of early stages of synkinematic growth (e.g., Passchier et al., 1992; Trouw et al., 2008). These features imply top-to-335° shear sense, in agreement with asymmetric strain shadows on garnet from the same sample (Table 1). In other examples (Figs. 14C, 14D) distinction between syn-D2 rotation and crenulation overgrowth is not possible. The snowball garnet porphyroblasts hosted in LK11-42 are unequivocal evidence of synkinematic growth given that: (1) only two pervasive deformation fabrics (D1 and D2) are recorded in the LKD rocks, and neither are likely to have rotated through >180° (e.g., Jiang and Williams, 2004); and (2) the shear sense implied by rotation (top-to-015°) of the porphyroblasts is consistent with shear sense recorded by C′ shear bands in the matrix (Figs. 14G–14I; Table 1). These observations indicate that chloritoid, garnet, and staurolite grew post-D1 to syn-D2.


Interpretation of D2 Microstructures and Quartz CPO Patterns

Quartz and feldspar microstructures and quartz CPO fabrics show generally close agreement if interpreted as deformation temperature indicators and assuming average geologic strain rates (e.g., Kruhl, 1998; Stipp et al., 2002b) for D2 deformation. As noted here, microstructures and CPO fabrics are influenced by a number of other variables, particularly strain rate and water content, that are not constrained for LKD rocks. This should be reemphasized for calcite microstructures as well, with the added caveat that total strain may also influence calcite twin width and intensity (Ferrill et al., 2004). As described in previous sections and summarized in Table 1, quartz microstructures exhibit a correlation with structural depth characterized by a progression from SGR, through transitional SGR-GBM to GBM I, and finally GBM II at the lowest structural levels. Quartz [c] axis fabrics also exhibit a trend of increasing opening angle toward lower structural levels at least through the metasedimentary rocks if not the core orthogneiss (Figs. 12B–12D). Furthermore, the quartz [c] axis fabric topology can be broadly described as transitioning from crossed girdles to single [c] axis maxima with increasing structural depth in the Muddy Lake transect (Fig. 12D), a fabric transition that has been ascribed to progressively increasing temperatures or shifts in recrystallization mechanism (e.g., Gleason et al., 1993; Stipp et al., 2002b). These patterns in microstructures and CPO fabrics could be primarily controlled by (1) increasing deformation temperatures, (2) decreasing strain rate, or (3) increasing water content. The third scenario is possible given that rocks at lower structural levels probably reached higher metamorphic grade that might be accompanied by greater degrees of dehydration (e.g., Harlov, 2012). Chessboard extinction microstructures and the activation of prism [c] slip, which is assumed to be responsible for chessboard extinction (e.g., Stipp et al., 2002b), are particularly sensitive to the presence of water (e.g., Bouchez et al., 1984; Garbutt and Teyssier, 1991; Morgan and Law, 2004). In the lower structural levels of the LKD where chessboard extinction is common, quartz CPO fabrics do not exhibit prism [c] or transitional basal <a>–prism [c] patterns (i.e., Type II crossed girdles). This suggests that hydrolytic weakening is not the primary factor controlling microstructures and CPO fabrics in the LKD. Ultimately, none of these scenarios can be ruled out for the LKD, but independent temperature estimates can provide a test for the first scenario: deformation temperatures increasing with depth.

Rolfo et al. (2004) indicated that the LKD rocks from the base of the metasedimentary section reached granulite facies conditions, characterized by peak temperatures of 650–700 °C. In that study Rolfo et al. (2004) found that metamorphic mineral growth occurred concurrently with D2 deformation, consistent with the observations that chloritoid, garnet, and staurolite in the metasedimentary rocks grew pre-D2 to syn-D2, suggesting overlap between deformation and metamorphic temperatures. Kinked kyanite grains (Fig. 8C) and fibrolite-bearing C′ shear bands (Fig. 10C) provide additional petrologic evidence to suggest that there was some overlap between metamorphic and deformation temperatures in the LKD. Although precise metamorphic temperature constraints are beyond the scope of this study, we suggest that temperatures from the correlative metasedimentary rocks in the adjacent Mabja-Sakya dome are appropriate proxies. Calculated temperatures ranging from ∼570 ± 50 °C in the garnet zone to ∼700 ± 65 °C in the sillimanite zone (Lee et al., 2004) agree with the peak temperatures indicated by Rolfo et al. (2004) for the LKD. In addition, Langille et al. (2010) correlated quartz microstructures and [c] fabric opening angles to metamorphic temperatures in the Mabja-Sakya dome to constrain deformation temperatures of 450–700 °C. These considerations are a reasonable basis for interpreting temperatures from LKD microstructures and quartz opening angles. The proposed metamorphic temperatures require that higher temperature ranges for SGR (450–550 °C) and GBM (>550 °C), as suggested by Law (2014) for Himalayan samples, should be applied to the LKD samples (Table 1). These alternate temperature ranges derive from along the Main Central thrust (MCT; Fig. 1) at the base of the GHS, but are consistent with temperatures of combined SGR + GBM (490–530 °C) and exclusive GBM (>530 °C) interpreted for samples along the STDS in the Mount Everest region (Fig. 1; Law et al., 2011), assuming geologic strain rates of 10–12–10–15 s–1. Because the distribution of samples with chessboard extinction is consistent with the interpreted progression in quartz microstructures toward lower structural levels, as well as with assumed metamorphic temperatures, it is also considered a valid indicator of deformation temperatures of ∼630 ± 30 °C as interpreted by Stipp et al. (2002b). Previous studies reported temperatures as low as 550 °C (e.g., Garbutt and Teyssier, 1991) and as high as 750 °C (Morgan and Law, 2004) for prism [c] slip, depending on water activity. Although water activity is not constrained for LKD samples, the presence of chessboard extinction microstructures at only the lowest structural levels suggests that it was active at or near the peak temperatures (650–700 °C) estimated by Rolfo et al. (2004), that are compatible with the value estimated by Stipp et al. (2002b). Calcite twin microstructures are harder to reconcile with deformation temperatures, but note that the structural position of the samples that contain them is consistent with the pattern noted in quartz microstructures and CPOs.

Interpreted D2 deformation temperatures are summarized with respect to structural position in Figure 15. The southern domain (Fig. 15A) covers a much larger structural thickness (∼3 km) than the northern and central domains (Fig. 15B), but the latter are sampled at a much higher resolution. Two samples (LK11-04 and LK11-36) that did not yield interpretable D2 microstructure are included to constrain an approximate upper boundary for D2 deformation in the northern and central domains. The lower structural limit of D2 is the structurally deepest sample collected, so it should not be considered the true lower boundary of D2 deformation. However, the structural intervals over which deformation temperatures are constrained in this study yield comparable thermal gradient estimates, suggesting that they are representative of the entire D2 field gradient. The interpreted temperature indicators are active over a range, some of which have no defined upper limit (i.e., quartz GBM and chessboard extinction), so average LKD thermal field gradients cannot be calculated, and instead reasonable minimum and maximum values are defined here and shown in Figure 16. In the southern domain, a minimum thermal field gradient of 100 °C/km, over a present-day structural thickness of ∼3 km, is bracketed by the upper limit of sample LK-04 (300 °C) and the lower limit for sample LK-41, which from the lower limit of the chessboard extinction temperature (630 ± 30 °C) would be 600 °C. If sample LK-04 is rejected because calcite microstructures remain questionable temperature indicators for the LKD, then the minimum field gradient is instead bracketed by the upper temperature limit for sample LK-05 (550 °C; Table 1), which yields a much lower field gradient of 18 °C/km. For the combined northern and central domains, a minimum field gradient is bracketed by the upper limit of sample LK11-74 (530 °C; Table 1) and the lower limit of sample LK11-26 (600 °C), which results in a minimum thermal field gradient of ∼58 °C/km over 1.2 km of structural thickness. Calculation of maximum gradients is more problematic because many of the temperature indicators used do not have defined upper limits. However, a reasonable maximum D2 thermal field gradient can be calculated along the southern transect. If sample LK-04 is rejected because the calcite twin temperatures are suspect, the lower temperature limit for the transect is set by the minimum deformation temperature for sample LK-05 (450 °C; Table 1). The upper temperature limit is set by the upper limit for sample LK-54 (630 °C; Fig. 15; Table 1) because a maximum temperature cannot be assigned to sample LK-41. These sample constraints yield a thermal field gradient of 90 °C/km over ∼2 km. Collectively, the calculations that dismiss calcite microstructure thermometry indicate a range of 18–90 °C/km (Fig. 16A) for the D2 thermal field gradient in the LKD.

Evolution of the Lhagoi Kangri Dome

Structural, kinematic, and deformation temperature analyses in the LKD indicate that the rocks underwent a sequence of events that is generally consistent with the histories reported from six previously studied domes (Fig. 1; Lee et al., 2000, 2004; Aoya et al., 2006; Quigley et al., 2006; Larson et al., 2010; King et al., 2011; Gao et al., 2013).

In the LKD, D1 is not well characterized in terms of geometry, kinematics, or major associated structures, so interpreting the origins of the mesoscale and microscale fabric in the highest structural levels investigated (S1) is somewhat problematic. However, pervasive S1 schistosity occurs in the upper interval of unit Ts, which is stratigraphically and structurally correlative to rocks in the adjacent Mabja-Sakya dome that record thickening by folding and thrusting following initial collision between India and Asia (Lee et al., 2004). Porphyroblast microstructures and preliminary petrology indicate that a phase of metamorphism in the LKD, defined by chloritoid and garnet growth in unit Ts and by garnet, staurolite, and kyanite growth in unit Ps, postdates D1 deformation and could conceivably be a manifestation of regional Barrovian metamorphism (Burg et al., 1984) documented to occur near the end of crustal thickening in the Mabja-Sakya dome (e.g., Lee et al., 2004; Lee and Whitehouse, 2007; Smit et al., 2014). Timing of onset and the precise duration of LKD metamorphism are not established here, but microstructural evidence, particularly continuity between porphyroblast inclusion trails and matrix fabric, suggest that metamorphism and D2 were at least partially contemporaneous.

Mesoscale and microscale structures, a mix of symmetrical and asymmetrical (top-to-north-northeast) kinematic indicators, and penetrative rock fabrics suggesting plane strain all indicate that D2 deformation defines a distributed ductile shear zone spanning a present-day thickness of ≥3 km. The present-day thickness includes at least 650 m of core orthogneiss up to the lowermost Triassic schist (Figs. 3 and 17). The original dip of the shear zone is likely modified by doming, but relatively consistent north-northeast–trending L2 lineations in the northern, central, and eastern domains suggest that the average dip of S2 in the northern domain (25° north-northeast) is a reasonable approximation. The top of the D2-dominant interval is denoted by slaty S1 cleavage in unit Ts that is progressively overprinted by S2 crenulations down structural section. Within unit Ts but at even lower structural levels, crenulations are transposed and overprinted by penetrative S2 schistosity and microscale S-C and C′ shear bands (e.g., Figs. 8, 10, and 17). Pervasive L-S fabric, associated microscale C′-type shear bands, samples with symmetrical crossed-girdle quartz CPO patterns that generally indicate plane strain deformation (e.g., Lister and Hobbs, 1980; Law, 1990), and rare S2 foliation-parallel chocolate tablet boudins that locally indicate flattening (e.g., Ghosh, 1988) may suggest that D2 deformation involved a component of thinning across the shear zone. In particular, symmetrical top-to-north-northeast and top-to-south-southwest microscale shear sense indicators and generally symmetrical quartz CPO patterns could indicate a component of pure shear, as was previously suggested from similar observations in the Mabja-Sakya dome (Lee et al., 2004). However, in the case of the Mabja-Sakya dome, Langille et al. (2010) interpreted symmetrical shear sense indicators to reflect either local bulk pure shear or heterogeneous strain partitioning in the presence of graphite (e.g., Selverstone, 2005), which is abundant in unit Ts in the LKD. It must then be considered that L-S fabrics are not exclusive to pure or general shear, and some LKD rocks also preserve asymmetric CPO patterns (e.g., LK-54 and LK11-54; Figs. 12B, 12D) and asymmetric boudinage (e.g., Fig. 5C) that suggest simple shear deformation. With careful consideration for the relative time when deformation temperatures were locked in, calculated D2 thermal field gradients may be indicative of a bulk D2 deformation history. At face value, the range in D2 field gradient (18–90 °C/km) is compatible with expected continental geotherms (15–30 °C/km; e.g., England and Thompson, 1984), perhaps suggesting that there is no attenuation across the shear zone. In general, mesoscale and microscale observations and interpreted deformation temperatures documented in this study define D2 deformation as a ductile, layer-parallel (S2) shear zone penetrative through a current structural thickness of ≥3 km.

Structures or rock fabrics that are unambiguously related to doming were not observed, but the concentric trajectory of S2 schistosity about the dome acts as a proxy. The foliation pattern could be primary (syn-D2) or secondary (post-D2), and in determining possible mechanisms for doming, it must also be considered that L2 stretching lineations have a relatively consistent north-northeast trend in the northern, central, and eastern domains. This is particularly relevant to diapirism models, wherein the host rock is expected to have concentric trajectory about the intrusion, similar to that in the LKD, but stretching lineations related to diapirism are predicted to have a radial or tangential trajectory about the dome (Dixon, 1975; Bateman, 1984). Furthermore, in pure diapirism models kinematic indicators should record orthogneiss-up (relative to the metasedimentary cover) sense of shear, and LKD kinematic indicators record variable core-up and core-down shear sense, particularly in the northern and central domains (Figs. 3, 6, 9, 10, 14, and 15; Table 1). However, concurrent regional strain patterns or lateral crustal flow can produce concentric foliations and unidirectional stretching lineations (e.g., Brun and Pons, 1981; Whitney et al., 2004) like those documented in the LKD. With this in mind, if LKD doming occurred synchronously with D2 deformation, the observed structural fabrics reported here would permit a model in which both D2 deformation and diapirism contributed to dome formation. If doming postdated D2, then a diapirism contribution could be ruled out because post-D2 diapirism would be expected to overprint D2 fabric with radial and/or tangential lineations and a core-up shear sense, none of which were documented in the LKD.

Contractional (Burg et al., 1984) and extensional (Chen et al., 1990) end-member models for doming are not as easily evaluated for LKD data, although at the resolution of the field study, significant regional structures that could be implicated in these end-member doming models were not observed. In the case of end-member extension models such as that proposed by Chen et al. (1990) for the Kangmar dome and a variant proposed by Quigley et al. (2006) for the Kampa dome, the contact between orthogneiss and metasedimentary cover is an extensional detachment fault. Exposure of the core-cover contact in the LKD was limited within the areas we investigated, but in the few places that it is exposed, the contact appears to be intrusive, particularly in the central domain where unit Ps quartzite and og orthogneiss are interfingered. Models in which doming occurs by buckling (Larson et al., 2010) or uplift in response to plate flexure (King et al., 2011) also cannot be ruled out.

Comparison to Other NHGD

Deformation characterized by mesoscale and microscale structures that are similar to D2 in the LKD is ubiquitous among previously studied NHGD, although the details vary between each. Interpretation of D2 deformation in the LKD as a diffuse shear zone, as opposed to a localized fault or mylonite zone at the core-cover contact, is generally consistent with findings in Kangmar and Mabja-Sakya, where it is inferred that D2 deformation is a northward distributed mid-crustal manifestation of structurally higher, localized deformation on the STDS (Lee et al., 2000, 2004; Zhang et al., 2004; King et al., 2011). Of particular note is the observation that L2 stretching lineations on the northern flanks of both the LKD and Mabja-Sakya consistently trend north to north-northeast, indicating structural continuity, at least through D2 time, between these domes. Results for the LKD, Mabja-Sakya, and Kangmar differ from studies in Changgo (Larson et al., 2010), Kampa (Quigley et al., 2006), and Malashan (Kawakami et al., 2007), where deformation is marked by relatively localized shearing adjacent to the core-cover contact, but is also inferred to be a manifestation of the STDS (Changgo and Kampa) or a related fault (Malashan). The STDS in the Mount Everest region dips generally north-northeast and projects beneath the LKD and Mabja-Sakya (e.g., Lee et al., 2004); this is compatible with geophysical evidence at the longitude of the Kangmar dome (Nelson et al., 1996). The structural evolution interpreted for LKD supports previous investigations in Kangmar (Lee et al., 2000), Mabja-Sakya (Lee et al., 2004, 2006; King et al., 2011), and Changgo (Larson et al., 2010) that indicate structural continuity between the GHS and NHGD prior to exhumation. Our interpretations are particularly consistent with kinematic analyses from the Kangmar dome (Wagner et al., 2010) that indicate a shift from symmetric to asymmetric top-to-north shear sense at lower structural levels, although it is important to reiterate that kinematic indicators at discrete structural intervals within the respective D2 shear zones may not be explicitly related to bulk D2 strain (e.g., Langille et al., 2010).

Tectonic Implications

The following discussion is predicated upon the assumption that the D2 deformation in the LKD correlates to D2 events in the Mabja-Sakya and Kangmar domes, which in turn have been correlated to deformation along the STDS (e.g., Lee et al., 2000, 2004, 2006; Lee and Whitehouse, 2007). In order to place the LKD into the broader context of middle crustal evolution during the Himalayan orogeny, LKD thermal field gradients are compared to similar data sets from (1) the STDS in the Mount Everest region (Law et al., 2011), (2) the STDS north of Mount Everest (Ra Chu and Gondasampa transects; Jessup and Cottle, 2010), (3) the STDS northeast of Mount Everest (Dzakaa Chu transect; Cottle et al., 2007), (4) the Kangmar dome (Wagner et al., 2010), and (5) the Mabja-Sakya dome (Langille et al., 2010). Only temperatures interpreted from quartz and feldspar microstructures, or from quartz [c] axis fabric opening angles, are considered since these are the data sets used for the LKD. However, metamorphic temperatures are also available for the other listed locations, except Ra Chu and Gondasampa, and the spatial and temporal association between metamorphic field gradients and deformation temperature field gradients for these locations are also key to constraining their tectonic evolution (e.g., Law et al., 2011). Two fundamental structural styles are represented by the selected study areas. In the case of the LKD, Mabja-Sakya, Kangmar, and Dzakaa Chu, the shear zone is distributed and penetrative through a present-day thickness of ≥1 km (Fig. 16A). In contrast, Ra Chu, Gondasampa, and Rongbuk valley transects are characterized by a highly attenuated shear zone of GHS rocks structurally separated from overlying TSS by a discrete, layer-parallel, brittle-ductile detachment (Law et al., 2004, 2011; Jessup et al., 2006; Jessup and Cottle, 2010; Figs. 16B, 16C). Specific sources for stratigraphic position and structural thicknesses, as well as a brief explanation for how thermal gradients were derived from other sources, are given in Appendix 1.

The LKD deformation temperatures interpreted from microstructural analysis and quartz [c] axis opening angles suggest a thermal field gradient of 18–90 °C/km over a current structural thickness of ∼3 km (Figs. 15–17). The Dzakaa Chu transect records a field gradient of 310 °C/km over 1 km structural thickness (Cottle et al., 2011), while Mabja-Sakya has an estimated thermal gradient of 30 °C/km over ≥6.5 km, and Kangmar records deformation temperatures that define a composite gradient of 40 °C/km in the upper ∼3.5 km and 130 °C/km for the lower 2–2.5 km (Fig. 16A; cf. Law et al., 2011). It is interesting that a linear regression through the opening angle temperatures for the LKD (Fig. 16A) also suggests an average gradient of 132 °C/km at lower structural levels, although the R2 value (0.4) is low. Law et al. (2011) calculated deformation temperature gradients of 369, 385, and 420 °C/km for three transects in the Rongbuk valley over structural thicknesses of ≤550 m (Fig. 16B). Ra Chu and Gondasampa deformation temperature data indicate thermal gradients of ≥1800 °C/km over <100 m (Fig. 16C), an order of magnitude larger than any thermal gradient estimated from the other study areas. Such high estimated field gradients are probably due to the relative scale of the two transects, but they could also reflect variation in structural style with close proximity to the brittle detachment surface (e.g., Law et al., 2011). Calcite temperatures are included for the upper part of the Ra Chu transect, but eliminating those data would actually increase the estimated thermal field gradient.

Interpretations from differences in thermal field gradients between different locations should be made cautiously given that (1) temperature-depth profiles may be linear at one scale but nonlinear at another (e.g., the significantly larger gradients associated with smaller transects in Ra Chu and Gondasampa or the apparent steepening of the Kangmar field gradient with depth), and (2) if the STDS cuts downsection through the GHS it is expected that thermal field gradients would decrease even without the influence of telescoping along the STDS (e.g., Beaumont et al., 2004; Law et al., 2011). The effect in (2) may be evident when comparing thermal field gradients from Rongbuk valley, Dzakaa Chu, Lhagoi Kangri, and Mabja-Sakya. Note that between these transects thermal field gradients progressively decrease from ≥369 °C/km in Rongbuk valley to 30 °C/km in Mabja-Sakya. An explanation for this spatial pattern must also account for when in the strain history the deformation temperatures were locked in, both within individual settings and between study areas, as well as from what structural level of the crustal deformation zone these transects derive (e.g., Langille et al., 2010; Law et al., 2011).

In the Rongbuk valley, Ra Chu, and Gondasampa areas the deformation history is characterized by a progression in deformation temperatures (high to low) through time, interpreted to reflect changes during exhumation, telescoping of hotter isotherms, and strain localization (e.g., Law et al., 2004, 2011; Jessup et al., 2006; Jessup and Cottle, 2010). Similarly, Cottle et al. (2011) suggested that the deformation temperature gradient in Dzakaa Chu is best explained by progressive telescoping of isotherms from hotter to cooler during exhumation, although it was not accompanied by strain localization. The studies from the Kangmar and Mabja-Sakya domes provide more reasonable end-member possibilities for when LKD temperatures were set: (1) prior to or during early stages of D2 (Kangmar dome; Wagner et al., 2010), or (2) after or during late stages of vertical thinning (Mabja-Sakya dome; Lee et al., 2004). The low Mabja-Sakya dome thermal field gradient (30 °C/km) would be expected of late-stage D2 deformation at a time when Lee et al. (2004) suggested that regional isotherms relaxed toward a normal geothermal gradient, whereas the Kangmar dome has deformation temperatures that overlap with peak metamorphic temperatures (Wagner et al., 2010) and a higher field gradient that is best explained by deformation temperatures locked in during relatively early D2 deformation (Law et al., 2011). The LKD thermal field gradient (18–90 °C/km) is similar to that of the Mabja-Sakya dome and the upper part of the Kangmar dome, suggesting that the D2 temperatures in the LKD also record late-stage deformation. This is compatible with similarities in lithology and structural fabric between the LKD and Mabja-Sakya. However, it was postulated that the LKD opening angle temperatures might reflect a higher thermal field gradient (132 °C/km) at lower structural levels. Although this higher calculated LKD gradient could be an artifact of few data points or the smaller structural thickness over which it applies, it is consistent with the composite thermal field gradient interpreted for the Kangmar dome, which would indicate that deformation temperatures from lower structural levels were locked in during early stages of D2 and temperatures at higher structural levels were locked in during later stages of D2. This idea is at least compatible with porphyroblast-matrix relationships that suggest D2 deformation occurred contemporaneously with metamorphism.

Regardless of whether the LKD field gradient is linear in the range of 18–90 °C/km or steepens to ∼132 °C/km at lower structural levels, the absence of microstructural overprinting in the LKD samples has direct implications for the way in which D2 isotherms, and thus deformation, evolved along the LKD shear zone. Similar to the interpretations made by Cottle et al. (2007) for the STDS at Dzakaa Chu, the lack of an S2-parallel brittle detachment in the LKD suggests that D2 deformation records early, higher temperature stages of deformation and shear zone development that, in the Everest area, are partially overprinted by extreme telescoping along the brittle STDS (e.g., Law et al., 2004; Jessup et al., 2006). The relatively low gradients in deformation temperature across the LKD shear zone are consistent with this interpretation. If the thermal gradient steepens at lower structural levels in the LKD, then some telescoping must have occurred within the shear zone. Even if a linear thermal field gradient is applied, some telescoping is implied by the minimum gradient in the northern and central domains (58 °C/km) and the maximum gradient from the southern domain (90 °C/km). Under this paradigm, the absence of low-temperature overprinting is most consistent with a model in which originally higher temperature portions of the shear zone were passively translated to higher structural levels, which could be accomplished without substantial fabric overprinting by upward migration of the upper shear zone boundary through time (e.g., Williams et al., 2006), as proposed for Rongbuk valley and Dzakaa Chu (Cottle et al., 2011; Law et al., 2011). Another process that might be expected to result in an apparent progression from high-temperature to low-temperature microstructures within the shear zone is strain localization during exhumation, which is a characteristic of crustal-scale faults (Sibson, 1977) or, perhaps more appropriately for the Himalaya, extrusive channel flow (e.g., Grujic et al., 1996, 2002). However, these processes would be predicted to produce lower temperature fabrics that overprint higher temperature fabrics. One other possibility is simultaneous deformation across the entire shear zone. This scenario cannot be ruled out entirely, but it is unlikely given that if the shear zone had any original dip, as assumed here, displacement would result in exhumation in which lower temperature overprint of higher temperature fabrics would be predicted. Therefore, the favored model for D2 in the LKD is that of a zone of heterogeneous simple or general shear in which the upper boundary migrated to structurally lower levels through time.

Ultimately, the D2 deformation temperature field gradient preserved in the LKD is a complex function of both spatial and temporal variations in middle crustal ductile deformation during early Himalayan orogenesis. The data presented for the LKD shear zone do not distinguish between gravitational collapse-driven middle crustal deformation (e.g., Burchfiel and Royden, 1985) versus channel flow–type middle crustal deformation (Beaumont et al., 2004), but they do indicate that the LKD shear zone is a prominent structure that accommodated penetrative middle crustal ductile deformation. These findings are consistent with previous studies discussed herein that suggest that STDS-related structures played a significant role in transport and exhumation of the GHS.


Field observations and preliminary structural and petrographic analyses in the LKD suggest that the region underwent a multistage history of deformation and metamorphism during the Himalayan orogeny, consistent with events documented in other NHGD and the GHS. The first stage (D1), largely preserved only at higher structural levels in Triassic siliciclastic rocks, involved burial and local development of slaty cleavage. Heating of the rocks during burial triggered the onset of prograde metamorphism recorded in the lowermost unit Ts and Ps pelitic assemblages. This initial overthickening and heating sufficiently weakened the LKD rocks so that they deformed through a distributed zone of penetrative ductile shearing (D2) that was confined to the lowermost Triassic rocks down through the core orthogneiss, a present-day structural thickness of ∼3 km. A wide range of associated deformation temperatures with no evidence of lower temperature overprinting also suggests that D2 strain was never localized. A possibly telescoped deformation temperature gradient (58–90 °C/km at high structural levels, 132 °C/km at low structural levels) as well as local indications of flattening and plane strain suggest that the D2 shear zone was defined by heterogeneous general shear, although simple shear is not ruled out. These observations are compatible with previously documented regional strain patterns and associated tectonic models that indicate that the STDS and related middle crustal ductile shear zones accommodated displacement between the GHS and the TSS during southward migration of the GHS. The Lhagoi Kangri geologic history is similar to that of the adjacent Mabja-Sakya dome and Dzakaa Chu transect, but more LKD data are necessary to exhaustively test ideas regarding its tectonic development, specifically its metamorphic evolution, and to further define its place in the early tectonic framework of the Himalaya.

We thank Tinley for his invaluable assistance with logistics for two expeditions into Lhagoi Kangri. A. Willingham (graduated from University of California, Santa Barbara) provided assistance in the field. G. Seward at the University of California, Santa Barbara conducted the electron backscatter diffraction analysis and processing. E. Heider and M. Smith contributed to petrographic analyses of samples from the Southern and Muddy Lake transects, respectively, as undergraduates at the University of Tennessee. We also thank J. Langille and K. White for insightful discussions regarding mid-crustal processes in the Himalaya and C. Hughes for assistance in interpreting and manipulating quartz crystallographic preferred orientation plots. We thank A. Weil for editorial handling and feedback. A previous version of this manuscript was greatly improved by comments from two anonymous reviewers. We thank K. Larson, J. King, and M. Stipp for their thorough and thoughtful reviews. Funding for this project was provided by the National Geographic Society (NG-CRE-8490-08) and the National Science Foundation (grants EAR-0911561 and EAR-0911416) to Jessup and Cottle. We also thank the DigitalGlobe Foundation for their generous imagery grant, which included decimeter-scale satellite images that helped improve the accuracy of the Lhagoi Kangri geologic map.


Thermal gradients were calculated in previous investigations in the Rongbuk valley (Law et al., 2011, fig. 10 therein) and Dzakaa Chu transect (Cottle et al., 2011, fig. 8 therein). Thermal gradients for Ra Chu and Gondasampa were calculated here from data presented by Jessup and Cottle (2010, table 1 therein). Law et al. (2011) also calculated thermal gradients in the Kangmar and Mabja-Sakya domes. In the case of the Kangmar dome, Law et al. (2011) measured thermal gradients off of a schematic cross section that contained interpreted deformation temperature isotherms (Wagner et al., 2010, fig. 5c therein), and in the case of the Mabja-Sakya dome, they derived thermal gradients from a schematic structural section given in terms of distance below an interpreted chloritoid-in isograd (Langille et al., 2010, fig. 9 therein). In an attempt to produce a more consistent comparison between the LKD and the other two NHGD, thermal gradients were recalculated for the Kangmar and Mabja-Sakya domes using quantitative structural depth versus deformation temperature data (Fig. DR3) also derived from Wagner et al. (2010, table 1 and fig. 8 therein) and Langille et al. (2010, table 1 and fig. 9 therein). The results, which are reported in Figure 16, are not much different from those reported by Law et al. (2011), with one distinction. For the Kangmar dome, Law et al. (2011) identified a three-part thermal gradient in which the overall trend is characterized by an increase in thermal gradient with deepening structural position. In contrast, the Kangmar thermal gradient measured in this study from visual best fit in Figure DR3 only delineates a two-part thermal gradient. However, the relative increase of thermal gradient with lower structural level is also exhibited in our recalculated two-part thermal gradient.

1GSA Data Repository Item 2016194, additional photomicrographs, CPO plots with fabric skeletons, and unaltered deformation temperature data from previous studies, is available at www.geosociety.org/pubs/ft2016.htm, or on request from editing@geosociety.org.