Uranium-lead (U-Pb) zircon geochronology, whole-rock geochemistry, and petrographic observations indicate that specimens from a suite of variably deformed granite and orthogneiss from the Okhaldungha region of east-central Nepal share a common origin. Microtextural characterization and quartz crystallographic fabric preferred orientation analyses of these same specimens outline a strain gradient that marks the location of a shear zone boundary. The location of this boundary, at the base of the orthogneiss, coincides with one of the interpreted locations of the Main Central thrust, though it cannot be uniquely identified as such. This study provides the first steps toward empirical constraints on the location and geometry of thrust structures in the region, helping to clarify the complex local kinematic framework. These methods not only help in assessing orogenic models of the Himalaya, but may also be applied to investigating other orogenic systems where the potential location of shear structures is contested.


Understanding the kinematic architecture of an orogenic belt is fundamental to understanding the structural processes accommodating convergence. Published literature contains many examples of contradicting interpretations of the location, nature, or even existence of various orogen-scale structures (e.g., Martin et al., 2005; Kruse and Williams, 2007; Johnston, 2008; Searle et al., 2008; Zanoni et al., 2014; Hildebrand, 2015; Martin, 2016). Contention surrounding the basic structural framework of the Himalaya has led to several different kinematic interpretations (e.g., Pearson and DeCelles, 2005; Searle et al., 2008; He et al., 2015) that have had a significant influence on our understanding of the evolution of the India-Asia collision as well as the general processes thought to dominate in collisional orogens (e.g., Searle and Szulc, 2005; Kohn, 2008). Primary amongst these contradictions in the Himalaya is the interpretation of the location and nature of the Main Central thrust (e.g., Martin, 2016). This orogen-scale, north-dipping thrust-sense shear structure was initially reported by Heim and Gansser (1939, p. 78) in the Kumaon Himalaya of India, bordering the western margin of Nepal, as a fault that places “crystalline rocks” (orthogneiss) upon “metamorphic limestone”. Gansser (1964) further described the Main Central thrust in the same region as separating “crystalline” rocks, dominantly biotite + alkali feldspar gneiss, in the hanging wall from commonly biotite to locally garnet- or even kyanite-grade phyllite or schist, sericitic quartzite, limestone or marble, and amphibolite. As noted by Searle et al. (2008), however, definition of the Main Central thrust based solely on the lithologies it separates has limited utility, as thrust faults cut up-stratigraphic section both in the direction of transport and potentially orthogonal to it. Various studies (as summarized by Searle et al., 2008) have mapped the trace of the Main Central thrust on the basis of lithologic contrasts (Bordet, 1961; Le Fort, 1975; Daniel et al., 2003), detrital zircon ages or bulk-rock Nd isotopes (Parrish and Hodges, 1996; Ahmad et al., 2000; Robinson et al., 2001; Martin et al., 2005; Richards et al., 2005; Mottram et al. 2014), or the location of anomalously young monazite ages (Harrison et al., 1997; Catlos et al., 2001, 2002). None of these methods, however, can uniquely identify the location of a fault or shear zone boundary between rock packages with variable lithologies (Searle et al., 2008), and as such the detailed kinematics of the Main Central thrust with respect to the architecture of the Himalaya remains enigmatic.

By its definition, a shear zone boundary corresponds to a strain gradient from relatively low-strain rocks outside the shear zone to progressively more highly strained rocks traced toward the center of the shear zone. Mapping the distribution of shear zone boundaries within an orogen provides a means by which to uniquely constrain areas of strain localization and thereby delineate the structural framework of the orogen. Attempts have been made to investigate the strain recorded across the exhumed former mid-crust of the Himalaya (e.g., Law et al., 2004, 2011, 2013; Goscombe et al., 2006; Larson and Godin, 2009; Long and McQuarrie, 2010; Long et al., 2011, 2016), but a general lack of strain markers has limited many studies to qualitative assessments (e.g., Martin et al., 2005). This lack of quantitative strain data in the frontal Himalaya has led to conflicting interpretations of both fault location(s) and the local and regional structural and kinematic framework.

A key region in this debate is east-central Nepal (Fig. 1) where previous workers have come to conflicting interpretations, inferring that: (1) mapped units are separated by a series of smaller structures rather than a regional-scale Main Central thrust (Ishida, 1969); (2) the Main Central thrust is located between mapped paragneisses and phyllites (Schelling, 1992); or (3) the Main Central thrust is located at the base of a distinctive mylonitic orthogneiss unit (Melung-Salleri augen orthogneiss) (Goscombe et al., 2006; Jessup et al., 2006; Larson, 2012; Larson et al., 2013; From and Larson, 2014; From et al., 2014). To attempt to distinguish between the various published structural interpretations of the region, and to assess the potential occurrence of a large-scale shear zone boundary near the base of the Melung-Salleri orthogneiss, a suite of specimens was collected across the Okhaldungha region of east-central Nepal (Fig. 1). These specimens were dated using U-Pb geochronology of zircon and characterized geochemically. Once it was determined that these rocks share a common protolith, they were examined microstructurally and subjected to quartz c-axis analysis to identify any relative changes in recorded strain and thereby assess the potential for the base of the orthogneiss to coincide with a shear zone boundary.


The Himalaya are commonly described as a series of orogen-parallel lithotectonic units separated by fault structures (e.g., Godin et al., 2006). From north to south these faults include the South Tibetan detachment, the Main Central thrust, the Main Boundary thrust, and the Main Frontal thrust (Fig. 1; Khanal and Robinson, 2013). Of relevance to this study are the rocks separated by the Main Central thrust, the Greater Himalayan sequence (hanging wall) and the Lesser Himalayan sequence (footwall). As introduced above, the nature and location of the Main Central thrust is a contentious issue with significant bearing on the structural framework of the orogen (Searle et al., 2008; Martin, 2016). Variation in the mapped location of the thrust can affect the interpreted affinity of rocks on either side and, in some orogenic models, their interpreted geologic history (e.g., Beaumont et al., 2001; Searle et al., 2006).

Investigations of the present study area, the Okhaldungha region of east-central Nepal, have mapped the Main Central thrust at different locations, resulting in significant variation in which rocks are ascribed to the Greater or Lesser Himalaya (Schelling, 1992; Larson et al., 2013; From and Larson, 2014). As indicated above, one interpretation is that the location of the Main Central thrust coincides with the base of the regionally extensive Melung-Salleri orthogneiss (Fig. 1; Goscombe et al., 2006; Jessup et al., 2006; Larson, 2012; Larson et al., 2013; From and Larson, 2014; From et al., 2014). The orthogneiss has been described as a biotite + muscovite gneiss characterized by augen structure (Ishida, 1969), a mylonitic augen gneiss of granitic composition (Schelling, 1992), and a strongly deformed, locally augeniferous quartz + alkali feldspar + plagioclase + muscovite + biotite orthogneiss (Larson et al., 2016). It is equivalent to a series of similar orthogneiss bodies that crop out across the Himalaya (Le Fort and Rai, 1999) that range in age from ca. 1.75 to 1.90 Ga (Kohn et al., 2010). These rocks have been variously interpreted to reflect felsic extrusive igneous rocks (Le Fort, 1975), metasomatized or tectonized basement (Arita, 1983), continental arc–related intrusive bodies (Kohn et al., 2010), or rift-related plutonic rocks (Sakai et al., 2013).

Recent field mapping in the study area identified a weakly deformed granite body within low-grade, phyllitic metasedimentary rocks just below the basal contact of the Melung-Salleri orthogneiss (Fig. 1). The granite intruding the metasedimentary rocks has the same basic mineral assemblage as the structurally higher orthogneiss, and in the field, it was suspected that the granite could represent the protolith to the orthogneiss. Four specimens were collected during field mapping along a roughly north-south transect across the orthogneiss body and the subjacent weakly deformed granite. The transect is subparallel to the overall regional tectonic transport direction. Three specimens (PK51, PK39, and PK57) were collected from within the orthogneiss, while the fourth (PK45), collected farthest to the south at the structurally lowest position, was collected from the weakly deformed granite body (Fig. 1). This suite of specimens, collected from rocks of potentially the same protolith across the base of the orthogneiss, allows investigation of the existence of a strain gradient across an interpreted shear zone boundary.


To assess genetic relationships between the rocks collected, all specimens were analyzed by geochronologic and geochemical methods. Full details of analytical methods employed are provided in the Supplemental Material1. Furthermore, all specimens were subject to both optical microscope–based microtectonic investigation and quartz c-axis fabric determination to compare their relative recorded strain history.



One- to three-kilogram (1–3 kg) representative rock specimens were crushed and separated using standard procedures. Zircon grains (n > 100) were then mounted in epoxy, imaged using cathodoluminescence (CL), and then analyzed for U-Pb age by laser ablation–multicollector–inductively coupled plasma mass spectrometry (LA-MC-ICPMS) at the University of California, Santa Barbara. Zircon CL images from orthogneiss specimens commonly show patchy zoning consistent with metamictization and mechanical fracturing (Fig. 2). Crystal domains associated with dark CL are typically metamict (i.e., amorphous) and have high U concentrations, whereas bright CL grains and cores are typically preserve oscillatory or patchy zoning. All ranges of CL intensity were targeted to capture the potential variability of ages contained in each specimen as well as inherited components (see Supplemental Table S12 for the full data set).

The majority of analyses in each orthogneiss specimen yield well-defined single-population (i.e., mean square of weighted deviates ∼1) discordia arrays with unanchored upper intercept dates of 1796 ± 6 Ma, 1799 ± 6 Ma, 1795 ± 6 Ma, and 1796 ± 6 Ma for specimens PK39, PK45, PK51, and PK57, respectively (2σ errors; Fig. 2). The lower intercepts of the discordia arrays are ca. 0 Ma (with large uncertainty) and are interpreted to reflect time-integrated radiogenic Pb loss and thus have no geologic significance. Analyses from bright CL cores typically do not fall on the discordia arrays but indicate inheritance from a variety of age sources ranging from ca. 2.0 to 3.6 Ga (not plotted in Fig. 2; see Table S1 [footnote 2]).


The whole-rock geochemistry of the specimens is presented in Table 1. The SiO2 content ranges from 74.9 to 77.1 wt%, while the total alkali content of the specimens ranges between 5.5 and 6.7 wt%, placing them in the rhyolite field of the International Union of Geological Sciences classification diagram (Fig. 3A). The specimens are peraluminous (Fig. 3B) and plot across the ferroan-magnesian boundary of the Fe* diagram [FeOtot/(FeOtot + MgO) versus SiO2] of Frost et al. (2001) with specimens PK57 and PK45 showing a ferroan association and PK51 and PK39 showing a magnesian association (Fig. 3C). Finally, when plotted on the modified alkali index (MALI) diagram of Frost et al. (2001) the specimens plot within the calcic and calc-alkalic fields (Fig. 3D). The peraluminous nature of the specimens and their distribution on the Fe* diagram and MALI plots are similar to those of peraluminous leucogranite bodies (Figs. 3B, 3C, 3D). It should be noted, however, that for such highly fractionated rocks, the Fe* diagram and MALI plot distributions also broadly overlap with those of arc-related sources and rocks from anorogenic environments.


In order to determine the strain recorded in the specimens collected and assess the potential for preservation of a strain gradient, the microstructures and crystal fabrics of our specimens were investigated using both a standard optical microscope and a Russell-Head Systems G60 automated crystal fabric analyzer. Quartz fabrics produced by similar instruments have been shown to be indistinguishable from those derived from electron back-scattered diffraction (Peternell et al., 2010) and X-ray goniometry (Wilson et al., 2007). Thin sections of each specimen were cut parallel to the macroscopic lineation and perpendicular to foliation (where present). Where observed, lineations and foliations are typically defined by aligned biotite and/or muscovite laths and quartz grain-shape fabrics (Figs. 4A, 4B).

Mineral Textures

The specimens display a range of microtextures that are roughly associated with sampling position on our north-south transect. Those collected farther north (PK57 and PK39) have a well-developed mylonitic foliation (Figs. 4A, 4B). The shear sense recorded by sigma-type feldspar porphyroclasts (specimen PK57; Fig. 4E) and asymmetric, partially recrystallized quartz porphyroclasts (specimen PK39; Fig. 4B) is consistent with a dominant top-to-the-south shear. Quartz grains within specimen PK57 are commonly pinned between mica-rich layers (Fig. 4E) and locally contain internal dislocation walls (Fig. 4E). These features are indicative of dynamic recrystallization involving grain boundary migration and sub-grain formation. Moreover, the generally straight quartz grain boundaries and common triple junctions indicate either significant static recrystallization post-deformation through grain boundary area reduction (GBAR) after penetrative deformation, or that recovery mechanisms were more efficient than dynamic recrystallization; the latter could suggest a decreasing strain rate. Specimen PK39 contains significantly more quartz than PK57, most notably in the form of large relict porphyroclasts (Fig. 4B). These relict grains are characterized internally by chessboard extinction patterns while the margins show evidence for new grain formation by extensive sub-grain rotation (Fig. 4F). These dynamically recrystallized grains are themselves characterized by straight boundaries and triple junctions (Fig. 4F) indicating significant GBAR.

In contrast to the well-developed tectonic foliation preserved in specimens PK57 and PK39, specimen PK51 is weakly foliated (Fig. 4C). The feldspar grains in the specimen form porphyroclasts or augen, are commonly perthitic, and are not sericitized. The latter observation contrasts with the modification of feldspars in specimen PK57 (Fig. 4E). Microstructural evidence for shear strain is recorded as localized extensional shear bands and asymmetric strain shadows (Figs. 4C, 4G), both of which indicate top-to-the south shear. Quartz within specimen PK51 occurs as both relict grain cores and recrystallized equigranular grains; as in specimen PK39, PK51 exhibits relict grain cores and chessboard extinction (Fig. 4H). The straight-sided new grains that typically meet in triple junctions (Fig. 4G) are interpreted to reflect sub-grain rotation dynamic recrystallization processes, with the quartz having since undergone significant GBAR.

Specimen PK45 is unfoliated (Fig. 4D) but contains local evidence of penetrative deformation including asymmetric strain shadows around a feldspar porphyroclast consistent with top-to-the-south shear (Fig. 4I). Quartz in the specimen occurs as equigranular, straight-sided recrystallized grains surrounding relict grain cores (Fig. 4J). As in both specimens PK39 and PK51, the relict grain cores display chessboard extinction (Fig. 4J). The smaller quartz grains surrounding the cores are interpreted to reflect sub-grain rotation dynamic recrystallization processes, while their polygonal texture is interpreted to reflect subsequent GBAR. In the rare occurrences where relict grains interact with each other, grain boundaries are characterized by bulging recrystallization textures (Fig. 4K).

The quartz chessboard extinction patterns in specimens PK39, PK51, and PK45 reflect orthogonal [c] and basal <a> slip (Mainprice et al., 1986), commonly associated with deformation temperatures >600 °C (Lister and Dornsiepen, 1982; Mainprice et al., 1986). Because chessboard extinction is only observed in relict grain cores within both mylontic and little-deformed specimens, it is here interpreted to reflect near-solidus deformation at the time of emplacement of the protolith to the orthogneiss (e.g., Blumenfeld et al., 1986; Büttner, 1999; Dawaï et al., 2013). In contrast, the substantial sub-grain recrystallization surrounding the relict grains is interpreted to reflect dynamic recrystallization during sub-solidus, penetrative shearing and strain accumulation.

Quartz Crystallographic Fabrics

An automated fabric analyzer was used to scan thin sections of each specimen to determine the unique orientation of quartz c-axes at an optical resolution of ∼5 µm. Results of these scans are represented as axial distribution or Achsenverteilungsanalyse (AVA) diagrams in the Supplemental Material (Figure S1 [footnote 1]). The quartz c-axis fabrics presented here were built by manually selecting individual quartz grains from the AVA diagram to avoid any potential misrecognition of other phases as quartz or the inclusion of poor-quality data through automated analyses. Moreover, relict grain cores (described above) were avoided, and only recrystallized grains were selected for building fabric patterns. The resulting fabrics provide information on the sense of shear during deformation (Bouchez and Pêcher, 1976; Lister, 1977; Lister and Williams, 1979; Lister and Hobbs, 1980) and deformation temperatures during fabric formation, assuming that: (1) temperature was the primary control on the critical resolved shear stresses for the operative glide systems; (2) no hydrolytic weakening occurred; and (3) strain rates did not vary significantly while the fabric was developing (Kruhl, 1996; Law et al., 2004; Morgan and Law, 2004; Law, 2014; Faleiros et al., 2016).

Quartz within specimen PK45 occurs mainly in disconnected pockets (Fig. 4D). Quartz c-axis orientations from 1219 recrystallized grains yield a generally disorganized fabric, with a hint of a type II crossed girdle, but the specimen is dominated by random c-axis orientations (Fig. 5A). If the weak crossed-girdle fabric is assumed to accurately reflect deformation conditions, the opening angle of ∼65° indicates a deformation temperature of 497 ± 50 °C (Faleiros et al., 2016). Quartz within specimen PK51 forms more continuous layers than that in PK45 (Fig. 4C) reflecting its foliated nature. A weakly organized fabric that outlines a type II crossed girdle topology is defined by 1339 c-axes from recrystallized grains (Fig. 5B). The opening angle of ∼65° between the fabric arms again indicates a deformation temperature of 497 ± 50 °C (Faleiros et al., 2016). Specimen PK39 contains both discrete quartz layers that are continuous on the scale of the thin section and a number of large, partially recrystallized quartz porphyroclasts (Fig. 4B). Recrystallized grains (n = 1415) from the discrete layers and surrounding porphyroclast cores yield a roughly symmetric type II crossed girdle (Fig. 5C). The opening angle of the fabric arms is measured at ∼62°, indicating a deformation temperature of 476 ± 50 °C (Faleiros et al., 2016). Finally, quartz within specimen PK57 defines the macroscopic foliation. The c-axis fabric (n = 1089) recorded in the quartz grains outlines an asymmetric type I crossed-girdle pattern indicating a dominant top-to-the-south shear (Fig. 5D). The fabric opening angle of 71° corresponds to a deformation temperature of 538 ± 50 °C (Faleiros et al., 2016).


Geochronologic and major element geochemical analyses indicate that the four specimens collected along our north-south transect have the same protolith magma crystallization age at ca. 1.8 Ga and have similar geochemical compositions consistent with derivation from a single crustal-dominated protolith. These similarities confirm previous geologic interpretations that place specimens PK51, PK39, and PK57 within a single unit (Schelling, 1992) and indicate that specimen PK45 is also originally from this same lithotectonic unit. This enables a direct and semiquantitative comparison of relative strains across this structurally important region of the Himalaya.

Microstructural characterization outlines significant, spatially correlated differences between specimens. Those collected farther north at structurally higher levels (PK39 and PK57) have more strongly developed tectonic foliations and more complete recrystallization of quartz (Fig. 4). This is reflected in the quartz c-axis fabrics measured, with more strongly developed fabrics in specimens collected to the north (Figs. 1 and 5). The change in c-axis fabrics can be quantified using a variety of statistical methods to assess the relative strength of their development (e.g., Woodcock, 1977; Lisle, 1985; Vollmer, 1990; Skemer et al., 2005). Of these, the approach of Vollmer (1990), which uses a triangular graphical distinction between point (P), girdle (G), and random (R) distributions calculated from the normalized eigenvalues of the fabric, satisfies the needs of this study by allowing an assessment of the degree of randomness of a fabric that is independent of the topology of the fabric (point or girdle). This can be expressed as a measure of the cylindricity index of the fabric, B, which ranges from 0 at R = 1, or a perfectly random fabric, to 1 for a perfectly ordered fabric (Vollmer, 1990).

Like the more qualitative assessments of strain already discussed, the triangular fabric plot (Fig. 6) demonstrates a shift toward more ordered fabrics traced toward the north and structurally higher positions (Fig. 6). This corresponds with B values that increase from a low in specimen PK45 of 0.1360, through 0.2113 in PK51, to 0.3628 and 0.5641 in PK39 and PK51 respectively (Table S2 in the Supplemental Material [see footnote 1]) and is consistent with increasing intensity, I (Lisle, 1985), values (Fig. 6). Because the specimens are of very similar lithology, with perhaps the exception of PK39 being more quartz rich, the strength of the c-axis fabrics recorded in the quartz grains may be viewed as a proxy for the total finite strain recorded in the rock. While it is recognized that deformation within quartz may be biased toward the later stages of deformation (e.g., Law et al., 2004), the relative difference in calculated randomness of the c-axis fabrics is consistent with qualitative strain determinations discussed above and is therefore interpreted to reflect significant finite strain differences. Moreover, the deformation temperatures indicated by fabric opening angles in the specimens are indistinguishable, consistent with penetrative shearing preserved in all specimens occurring during the same deformation event.

The documented spatial changes in recorded strain outline a strain gradient across the study area that coincides with the mapped lower boundary of the Melung-Salleri orthogneiss (Fig. 1). This is consistent with, but does not prove, the placement of the Main Central thrust at this approximate structural level by previously published studies of east-central Nepal (e.g., Goscombe et al., 2006; Jessup et al., 2006; Searle et al., 2008). The existence of such a gradient allows a comparison to be made with previous map interpretations and suggested fault distributions in the region (Fig. 7). While this comparison cannot be used to assess the relative merits of studies that map a structure at the base of the orthogneiss (Figs. 7A, 7B, 7D), it can be used to exclude those that do not (Fig. 7C). Moreover, documentation of a shear zone boundary provides an important first step in determining the detailed structural framework of the region, and the results of our study demonstrate the potential utility in applying semiquantitative microstructural and crystal fabric analyses methods to identify the location of inferred structures, such as thrust faults, in the Himalaya as well as other orogenic systems.

Constraining fault distribution and kinematics across the Himalaya is becoming increasingly important, given the variety of different tectonic models that have been proposed for its evolution. These range from models that evoke the stacking (e.g., McQuarrie et al., 2008; Robinson, 2008) or accretion (e.g., Webb, 2013; He et al., 2015) of discrete thrust slices, to those that argue for broad pervasive ductile strain (e.g., Searle and Rex, 1989; Searle et al., 2006). Applying a similar approach to that demonstrated in this study across a transect that includes a greater total structural thickness of the orogen may help to more accurately determine the spatial distribution of the structures that make up the kinematic framework, and thereby provide a means to test the proposed models. If multiple thrust-related structures are identified, that could lend support to models with numerous discrete faults, whereas the opposite would be consistent with models that imply pervasive, non-localized strain. This is notwithstanding potentially complicating factors such as changing rheology and strain localization as rocks move through an orogen (e.g., Larson and Cottle, 2014; Cottle et al., 2015; Parsons et al., 2016a, 2016b) that would need to be accounted for. Moreover, this type of analysis may be applied to other orogens in which the occurrence and/or nature of structures, such as the Monashee décollement in the Canadian Cordillera for example (Kruse and Williams, 2007; Gervais and Brown, 2011; Zanoni et al., 2014), is also disputed.


Geochronologic and geochemical data demonstrate that specimens collected at different structural and spatial positions within east-central Nepal were formed during the same magmatic episode. The microstructures and quartz c-axis fabrics recorded in these rocks outline a strain gradient that corresponds to a shear zone boundary at the base of the Melung-Salleri orthogneiss in the study region. This strain gradient coincides with thrust faults mapped by previous studies at this structural position, including the lower boundary of the Main Central thrust shear zone.

This study was supported by Natural Sciences and Engineering Research Council of Canada Discovery and Canadian Foundation for Innovation grants to K. Larson and U.S. National Science Foundation grant 1119380 to J. Cottle. H. Cavallin is thanked for assistance with mineral percentage determinations. Early versions of this manuscript benefited from discussion with A. Martin. Constructive reviews by R. Law and S. Long and editorial handling by M. Williams greatly improved the focus and clarity of this work.

1Supplemental Material. Full details of geochronologic and geochemical analytical methods employed, quartz AVA/c-axis data quality figures, and calculated geometric parameters of c-axis fabrics. Please visit http://doi.org/10.1130/GES01373.S1 or the full-text article on www.gsapubs.org to view the Supplemental Material.
2 Supplemental Table S1. LA-MC-ICPMS U-Th/Pb zircon isotopic data. Please visit http://doi.org/10.1130/GES01373.S2 or the full-text article on www.gsapubs.org to view Supplemental Table S1.

Supplementary data