Abstract

In the central Himalaya, past researchers have identified a distinctive transition from the physiographic Lower Himalayan ranges in the south to the Higher Himalayan ranges in the north. Local relief and hillslope gradient, as well as erosion and surface uplift rates, increase abruptly across this transition to the north. In the eastern Himalaya, the same physiographic transition exists, but it is less dramatic. We describe here a previously undocumented steep, north-dipping, brittle structure that is roughly coincident with this physiographic transition in eastern Bhutan—the Lhuentse fault. Low-temperature (U-Th)/He apatite data suggest that the Lhuentse fault has been active since the Pliocene, and (U-Th)/He dates on offset hydrothermal hematite deposits from within the fault zone demonstrate a component of Quaternary slip. Although we identified no definitive evidence of fault kinematics based on field or petrographic analysis of the fault rocks, the disrupted pattern of (U-Th)/He apatite dates suggests normal-sense displacement, contrary to what was expected given previous studies of an analogous transition in the central Himalaya. We regard the existence and activity of the Lhuentse fault as evidence of (1) recent evolution in the tectonic regime of the eastern Himalaya from one of near-exclusive north-south shortening to one in which both transcurrent and normal faulting are increasingly important in the region north of the Himalayan deformation front, or (2) an active duplex south of the physiographic transition in the middle latitudes of Bhutan.

INTRODUCTION

The transition from the Indo-Gangetic Plains northward to the Tibetan Plateau is a dramatic, 5-km-high escarpment (Fig. 1). The Himalayan landscape in between rises thousands of meters across a set of steps, or physiographic transitions (Fig. 1), that have been correlated to regions of focused uplift. The northernmost physiographic transition, physiographic transition 1 (PT1), marks the dramatic change from the rugged and high peaks of the Himalayan range to the lower relief of the Tibetan Plateau. The most dramatic of these transitions, PT2, generally separates the physiographic Lower Himalaya to the south from the Higher Himalaya to the north. The southernmost, PT3, denotes the meeting of the edge of the Himalayan range and the Indo-Gangetic Plains. See Hodges et al. (2001) and Hodges and Adams (2013) for a more thorough discussion.

In central Nepal, PT2 corresponds spatially with a discontinuity of vertical rock and surface uplift rates that are higher to the north than to the south (Jackson and Bilham, 1994). Various geological, geochemical, geodetic, and geophysical data sets have been employed to interpret this discontinuity as a manifestation of one of three possible tectonic processes: (1) a simple ramp in the north-dipping sole thrust at the base of the Himalayan orogenic wedge (Fig. 2A; e.g., Lyon-Caen and Molnar, 1983, 1985; Cattin et al., 2001; Lavé and Avouac, 2001); (2) an actively growing thrust duplex (Fig. 2B; e.g., Robinson et al., 2003; Bollinger et al., 2004, 2006; Pearson and DeCelles, 2005); or (3) a Quaternary, out-of-sequence, south-vergent thrust fault (Fig. 2C; Wobus et al., 2003, 2005, 2006a; Hodges et al., 2004).

PT2 has been studied in less detail in other sectors of the Himalaya. For example, it is possible to identify a distinctive PT2 feature in Bhutan (Duncan et al., 2003), even though the topographic profile is quite different than in central Nepal. Like Duncan et al. (2003), we recognize that the front of the range in Bhutan exhibits ruggedness similar to the High Himalaya in central Nepal (Fig. 1B). This observation gives the sense that Bhutan is “missing” the typical Lower Himalayan topography. However, Grujic et al. (2006) have proposed that a zone of low relief in the middle latitudes of Bhutan represents Pliocene to Quaternary uplifted surfaces of Lower Himalayan topography. These perched low-relief landscapes abut a zone of high relief to the north (Fig. 1B). This juxtaposition of high- and low-relief topographies creates a physiographic transition analogous to that of central Nepal.

Because PT2 in central Nepal is so closely related—in a spatial sense—to buried or emergent deformational structures that have played important roles in the evolution of the Himalaya, it is reasonable to posit that observation of this landform in other parts of the orogen could serve as an indicator of the existence of similarly important structures, and could help focus future structural studies in less well-characterized sectors. However, such a hypothesis needs to be tested in each location where PT2 is observed. Variability in the consistency or form of PT2 along strike could suggest a change in tectonic mechanism, or magnitude and duration of tectonic forcing. The study of PT2 along the strike of the range could illuminate architectural or temporal variations in the development of the orogen.

Here, we describe the topographic and geologic character of PT2 in central and eastern Bhutan and evaluate whether or not it marks a zone of active deformation in this sector of the orogen as it does in central Nepal. Bedrock mapping in Bhutan has revealed the existence of a major north-dipping, contractional fault in the vicinity of PT2, the Kakhtang thrust (Figs. 3 and 4), but previous workers have ascribed a Miocene age to all activity on the structure (Gansser, 1983; Swapp and Hollister, 1991; Davidson et al., 1997; Grujic et al., 2002; Kellett et al., 2009). We present evidence here showing that PT2 is also coincident with the trace of a steep, north-dipping structure with activity extending into the Quaternary, referred to here as the Lhuentse fault. Although there is no direct structural evidence indicative of the kinematics of the fault, a (U-Th)/He thermochronometric data set for samples collected across the structure is interpreted to provide evidence of a component of normal-sense displacement, with rocks to the south having been exhumed more rapidly than rocks of the higher-relief terrains to the north during the past 2–3 m.y. This unanticipated result is interpreted as evidence of a Pliocene to Quaternary transition in the kinematic evolution of the Himalayan hinterland at this latitude.

CHARACTERISTICS OF PT2 IN THE BHUTAN HIMALAYA

Morphometric Observations

Although Duncan et al. (2003) first established the general position of PT2 in Bhutan, our search for field evidence of structural disruption along this transition required a more precise location effort. Following the approach of Wobus et al. (2006a), we conducted a topographic analysis of the region using an Advanced Spaceborne Thermal Emission Reflection Radiometer (ASTER) digital elevation model (DEM) with ∼30 m/pixel resolution.

We focused our geomorphic analyses on three metrics: (1) local relief, (2) hillslope gradient, and (3) channel steepness. Our local relief map (Fig. 4A) was calculated by finding the difference between the highest and lowest elevations within a 5-km-diameter circular moving window. The slope map (Fig. 4B) was calculated across an ∼90 m square and then smoothed using a 1-km-diameter circular moving window. We used the normalized channel steepness index (ksn) to quantify patterns of river gradients that are scale independent, allowing us to compare river gradients across Bhutan (Fig. 4C) regardless of drainage basin size or location (e.g., Wobus et al., 2006b). We calculated ksn over 10 km lengths of channel using a reference concavity index of 0.45 (e.g., Wobus et al., 2006b). The patterns of channel steepness revealed from this analysis can help constrain the ways in which landscapes are responding to external forcing factors (e.g., uplift, rock strength, and climate; e.g., Wobus et al., 2006b).

Our analysis shows that PT2 within Bhutan is completely contained within the mapped extent of the Greater Himalayan sequence of rocks, which corresponds to the metamorphic hinterland of the Bhutan Himalaya (Fig. 3; Long et al., 2011a). PT2 can be readily observed as an abrupt, but discontinuous increase in hillslope gradient, local relief, channel steepness, and mean elevation to the north. In most cases, the mountains directly to the north of PT2 climb rapidly to elevations in excess of 5000 m. Here, the regional mean elevation and local relief appear to be limited by glacial erosion (Brozović et al., 1997; Meigs and Sauber, 2000) rather than fluvial valley incision. Glacial erosion in these high northern regions is evident in visible satellite imagery and previously published maps (Gansser, 1983; Komori, 2008) in the form of cirques, tarns, moraines, and active glaciers. Such evidence for extensive glacial activity limits the utility of attempts to infer differential exhumation from simple morphometric comparisons of landscapes north and south of PT2 (Brocklehurst and Whipple, 2007).

In comparison to PT2 in central Nepal (Wobus et al., 2003, 2005, 2006a), the north-south gradient across PT2 in Bhutan in any morphometric parameter is subdued and less abrupt, and it is segmented and more difficult to trace continuously along its trend. It is not possible to follow the trace of PT2 across the deep canyons that surround these perched low-relief landscapes. Curiously, the two well-constrained segments in central Bhutan are located near the northern edge of the perched low-relief surfaces (Grujic et al., 2006; Fig. 4). This may imply that these low-relief surfaces and PT2 have a shared history.

Spatial Relationship of PT2 with Previously Mapped or Inferred Structures

Of all the major structures previously mapped in Bhutan (Gansser, 1983; McQuarrie et al., 2008; Grujic et al., 2011), PT2 most closely follows the Kakhtang thrust. However, there is a considerable uncertainty regarding the trace (Fig. 3), throw, and timing of the Kakhtang thrust. Gansser’s (1983) initial observations of the Kakhtang thrust lacked a clear description of the structure or its trace. In 1991, Swapp and Hollister observed a previously undocumented inversion in metamorphic grade that they argued was created by displacement on a thrust fault, and roughly coincident with the shear zone proposed by Gansser (1983). No exposures of the Kakhtang thrust have been described in the literature, but it is often generally mapped near the position of a change in abundance of leucogranitic anatexites or the second sillimanite isograd within the Greater Himalayan sequence (e.g., Swapp and Hollister, 1991).

The rough coincidence of PT2 and the Kakhtang thrust is particularly good in eastern Bhutan (Fig. 4D). However, the trace of PT2 and most previously mapped traces of the Kakhtang thrust in central Bhutan (e.g., Long et al., 2011a; Grujic et al., 2011) differ substantially. In areas where the mapped position of the Kakhtang thrust is not coincident with PT2, there is no geomorphic signature of differential uplift across the thrust. This may not be surprising, as previous interpretations of U-Pb dates from luecogranites deformed by the Kakhtang thrust suggest an exclusively Miocene deformation history (Grujic et al., 2002; Kellett et al., 2009). Combined, these observations suggest that PT2 is not simply a physiographic manifestation of hanging-wall uplift of the Kakhtang thrust (i.e., this does not seem to support the hypothesis in Fig. 2C), but it may have a more complex relationship with the tectonic architecture of Bhutan.

While some research groups have interpreted the transition in uplift rate across PT2 in central Nepal to be a consequence of hanging-wall deformation associated with a deep-seated ramp in the Himalayan sole thrust (Fig. 2A; e.g., Lyon-Caen and Molnar, 1983; Jackson and Bilham, 1994; Lavé and Avouac, 2001), a similar interpretation of the phenomenon in Bhutan is more difficult. There is considerable debate regarding the existence and along-strike extent of a ramp in the Himalayan sole thrust beneath Bhutan. Hauck et al. (1998) proposed a ramp in the Himalayan sole thrust in western Bhutan, but their preferred position of that ramp is far north of PT2 and could not explain the inferred differential uplift across the transition. Recent balanced sections by McQuarrie et al. (2008) and Long et al. (2011b, 2012) have provided evidence for a prominent Miocene ramp within transects of eastern Bhutan. However, the positions of these ramps and duplexes at depth are to the south of PT2 and cannot be related to uplift north of PT2 relative to the perched low-relief landscapes. The presence of a ramp in central or western Bhutan is not as clear from balanced sections (Long et al., 2011b, 2012). Conversely, inversions of thermochronologic data by Robert et al. (2011) suggest that the Himalayan sole thrust is a listric structure throughout Bhutan and that there is no evidence for a significant ramp in this part of the orogen.

FIELD OBSERVATIONS OF PT2—LHUENTSE FAULT

As is the case throughout much of the central and eastern Himalaya, the area of PT2 in Bhutan is heavily vegetated, and rocks are highly weathered. Outcrops are generally poor and restricted to river gorges or sparse road cuts. In order to better evaluate the causal mechanism for PT2 in Bhutan, we studied one transect along the Kuri Chu in detail where PT2 could be projected on the basis of our morphometric analysis (Fig. 5A) and where the prospects for bedrock exposures seemed promising. At this projected point in the Kuri Chu valley, located ∼4 km south of the mapped trace of the Kakhtang thrust (Fig. 5A; Long et al., 2011b), we found a distinctive zone of brittle deformation in Greater Himalayan sequence rocks that we refer to here as the Lhuentse fault.

Most of our observations regarding the structural character of the Lhuentse fault are based on a continuous bedrock outcrop, parallel to the Kuri Chu drainage, which has a structural thickness of ∼20 m. The fault zone is expressed as a series of closely spaced, subparallel faults with a mean orientation of 063°/80° (Fig. 5B). Each fault is marked by several tens of centimeters of cataclasite, including coarsely fragmented bedrock, fault gouge, and (typically) syn- to postkinematic hydrothermal mineralization (Figs. 5C and 5D). On closer inspection, the faults are marked by anastomosing zones of finer-grained fragmented rock and gouge surrounding larger (∼1–2 m) phacoidal blocks of less-deformed rock. In outcrop, the crush zones are more easily eroded, leaving the more-resistant phacoids as prominent knobs. Despite a concerted effort, we were unable to identify any convincing slickenlines or asymmetric features on any of the structures in the Kuri Chu outcrop that would permit kinematic interpretation.

A thin section from the fault zone shows variously comminuted quartz grains (Figs. 6A and 6B). This style of quartz deformation is consistent with grain-size reduction principally by cataclasis at temperatures likely <300 °C and relatively high strain rates (Passchier and Trouw, 1996). We saw no indication of crystallographic grain-shape preferred orientation in the quartz, and we found no useful kinematic indicators in thin section.

Petrographic analysis of the hydrothermal precipitates revealed that most consisted of an aggregate of quartz and an opaque mineral. Raman spectroscopy revealed that the opaque mineral is hematite (α-Fe2O3; see GSA Data Repository1). Based on the absence of hydrothermal mineralization in the rocks outside the fault zone, we infer that the fault zone itself served as the principal conduit for mineralizing fluids. Both outcrops and the thin section display clear evidence of multiple generations of hydrothermal mineralization. In section, fine-grained precipitates appear as speckled and rounded clusters. They are surrounded by younger, coarse-grained quartz + hematite masses. These quartz + hematite masses are also crosscut by younger microfaults (Figs. 6C and 6D), which unfortunately exhibited no preferred orientation or consistent offsets of preexisting hydrothermal veins that might indicate slip direction. We infer that hydrothermal quartz + hematite precipitation commenced after the earliest stages of deformation and continued at least episodically during slip, but it had largely ceased prior to the latest stages of deformational activity. Although we have no direct constraints on the crystallization temperatures of the syndeformational hydrothermal minerals in the Kuri Chi cataclasites, we note that fluid inclusion studies of hydrothermal quartz deposits elsewhere along the Himalayan front suggest temperatures ca. 300 °C (Derry et al., 2009).

ALONG-STRIKE EXTENT OF THE LHUENTSE FAULT

Attempts in the field to follow the Lhuentse fault along strike to the west and east out of the Kuri Chu valley were unsuccessful due to poor exposure. However, high-resolution (∼0.5 m/pixel) WorldView satellite imagery indicates that the fault may be the southernmost manifestation of a broader domain of high-angle faulting up to 13 km in width (Fig. 5A). The images reveal spaced lineaments with an average strike (070°) similar to the Lhuentse fault (Fig. 5E). These lineaments were also observed in remote-sensing data sets by Das (2004), who interpreted them as potential strike-slip faults. However, we found no evidence of lateral offsets of drainages on the Lhuentse or any other ENE-trending lineaments in the course of our remote-sensing study.

(U-Th)/He THERMOCHRONOLOGY—INDICATIONS OF NORMAL FAULT OFFSET

By analogy with PT2 in central Nepal (Wobus et al., 2003, 2005, 2006a), we anticipated that PT2 in Bhutan might also mark a discontinuity in bedrock cooling ages as determined by noble gas thermochronometry. To explore this possibility, we collected bedrock samples (paragneisses and orthogneisses and fine-grained schists) for (U-Th)/He thermochronology along an ∼13 km transect that crosses the Lhuentse fault in the Kuri Chu drainage (Fig. 5A). Six samples were found to have datable apatite, but only three of those contained datable zircon. Using facilities at Arizona State University’s Noble Gas Geochronology and Geochemistry Laboratories (NG3L), we measured 21 (U-Th)/He apatite (AHe) dates and 14 zircon (ZHe) dates on single, handpicked crystals (Table 1; Fig. 7A). Analytical procedures closely followed those described in Schildgen et al. (2009) and van Soest et al. (2011).

For each sample, zircon and apatite single-crystal dates clustered reasonably well, but—as is frequently the case in (U-Th)/He thermochronometry—the dispersion of dates for a particular mineral in a particular sample was greater than would be predicted by analytical imprecision alone. We elected to interpret the AHe or ZHe date for a specific sample as the analytical error-weighted mean for all dated crystals, where the number of crystals (n) ranged from three to five, depending on the sample. The reported error for the mean dates (Table 1) represents both the propagated analytical errors as well as an error magnification factor based on the mean square of weighted deviates (MSWD) and Student’s t test for n – 1 degrees of freedom (Wendt and Carl, 1991; Cooper et al., 2011). These mean dates and their calculated errors (reported at the 95% confidence level) form the basis for interpretations of the low-temperature cooling history presented in the following paragraphs.

Compared to the central Nepal region where PT2 has been studied in detail, the Kuri Chu AHe dates are older, suggesting less aggressive erosional exhumation since the late Pliocene in Bhutan, particularly north of PT2. For example, in the Annapurna range of Nepal, AHe dates north of PT2 are uniformly younger than 1 Ma (Blythe et al., 2007; Nadin and Martin, 2012). In the Kuri Chu transect, AHe dates for samples collected north of the Lhuentse fault trace ranged from 2.78 ± 0.41 Ma to 3.6 ± 1.3 Ma. Two samples collected south of the trace yielded younger dates of 2.41 ± 0.47 Ma and 2.42 ± 0.35 Ma. Zircon dates for three samples collected north of the fault ranged from 4.23 ± 0.27 Ma to 4.793 ± 0.085 Ma. Although we found no datable zircons in any samples collected south of the fault, Long et al. (2012) reported a ZHe date of 4.05 ± 0.07 Ma for a sample collected ∼7.5 km south of the trace (Fig. 7A).

In order to estimate variations in exhumation rate across the transect from the thermochronometric data, we pursued one-dimensional (1-D) thermal modeling (AGE2EDOT; Brandon et al., 1998) of the AHe data that involved simultaneous solution of steady-state advection and diffusion equations and assumed uniform constant exhumation rates for each sample and reasonable values for material properties. Our assumptions regarding material properties were similar to those of Whipp et al. (2009): thermal conductivity = 2.75 W/m K; radiogenic heat production = 0.8 mW/m3; specific heat = 1000 J/kg K; rock density = 2700 kg/m3; a surface temperature of 10 °C; and a constant temperature boundary condition of 490 °C at 30 km. In addition, we employed the experimental data of Farley (2000) in order to model the diffusive loss of helium in apatite during exhumation. For simplistic 1-D models, a thermochronometric date—taken here as the error-weighted mean AHe date for each sample—uniquely specifies a time-averaged exhumation rate (ɛ). Our justification for not employing more complex two- or three-dimensional models for such calculations is based on the work of Whipp et al. (2007), who concluded that the thermal field in active orogens such as the Himalaya is predominantly controlled by the vertical (i.e., 1-D) advection and conduction.

Time-averaged exhumation rates modeled in this way from the Kuri Chu apatites range from 0.58 ± 0.14 to 0.93 ± 0.15 km/m.y. (Fig. 7B). When erosion rates are in this range, assumptions regarding material properties have significant influence on 1-D modeling results, and conductive heat transfer becomes more significant relative to advective heat transfer. However, the quoted 95% confidence interval for these calculated exhumation rates reflects only propagated analytical uncertainties in the thermochronologic data. We feel that this is reasonable because it is unlikely that material properties vary substantially along the ∼13 km length of the transect. Nevertheless, the exhumation rate values themselves should be viewed only as rough estimates. A much more robust result is the apparent difference in rates north and south of the trace of the Lhuentse fault. Samples to the north can be reasonably interpreted as indicating an average erosion rate (ɛn) of 0.686 ± 0.049 km/m.y., whereas the two samples to the south indicate an average erosion rate (ɛs) that is statistically distinctive at 0.922 ± 0.096 km/m.y.

A distinction between the cooling histories of samples north and south of the fault is less obvious from the distribution of AHe dates in Figure 7A than from the distribution of modeled exhumation rates in Figure 7B. This reflects the fact that the samples were collected over a range of elevations, from 1254 m (BT0970) to 2360 m (BT0962). In order to better illustrate the distinctions, we used ɛn and ɛs to normalize the AHe cooling dates to the mean elevation of the transect (1740 m). The elevation-corrected dates are shown in Figure 7C (see GSA Data Repository for method and discussion [see footnote 1]). North of the fault, the normalization process results in a much smaller scatter in corrected AHe dates (AHen), with a mean of 2.67 ± 0.16 Ma. South of the fault, corrected AHe dates (AHes) are younger, with a mean of 1.92 ± 0.28 Ma.

We conclude from this exercise that the apatite (U-Th)/He thermochronologic data from the Kuri Chu transect indicate preferential uplift of lithologies exposed south of the Lhuentse fault trace relative to those north of the trace over the past ∼2 m.y. Given the observed dip direction of the fault, we interpret the data as indicative of normal-sense offset on the structure. Using the mean exhumation rates from each side of the fault to calculate the elevation for a cooling date, we find an offset of ∼500 m in the AHe isochrones across the Lhuentse fault. Whether this reflects pure normal slip or possibly normal-oblique slip is presently unknown.

CONSTRAINTS ON THE AGE OF THE LHUENTSE FAULT FROM HEMATITE (U-Th)/He GEOCHRONOLOGY

Although the disruption of AHe cooling age patterns across the Lhuentse fault indicates slip subsequent to 2 Ma, an additional constraint is provided by (U-Th)/He dating of syndeformational hydrothermal hematite in the Lhuentse fault cataclasites. Although the usefulness of hematite as a (U-Th)/He chronometer has been noted elsewhere (e.g., Kula and Baldwin, 2012), our study may be the first published application of the technique to dating fault activity. Since procedures for such work at NG3L have not been published previously, they are detailed briefly here. We began by breaking a large, centimeter-sized piece of ore-rich cataclasite sample into millimeter-sized fragments. In order to eliminate the need for alpha ejection corrections of measured 4He concentrations, interior fragments were enclosed in a pneumatic chamber and self-abraded for 3 d to eliminate the outer several tens of microns of material from each fragment. Four well-rounded, polished fragments were selected and loaded into Nb tubes for analysis. Our approach to analyzing the hematite samples followed closely the protocols described by van Soest et al. (2011) for zircon analysis, although we found that high concentrations of H2O and CO2 liberated from the hydrothermal samples required our use of a cryotrap and longer gas purification times (∼6 min) to purify the extracted gases in preparation for helium measurement on an Australian Scientific Instruments Alphachron system at NG3L. The samples were then dissolved in a mixture of HF, HNO3, and HCl in Parr digestion vessels prior to parent element measurements on a ThermoElectron X-Series inductively coupled plasma–source mass spectrometer.

The four specular hematite fragments yielded (U-Th)/He dates ranging from 70 ± 8 ka to 130 ± 15 ka (Table 2; dates reported at the 95% confidence level), with a variation larger than would be anticipated based on analytical imprecision alone. One plausible interpretation of these results is that the variation reflects variable postcrystallization loss of 4He. We consider this unlikely, however, based on presently available experimental constraints on the diffusivity of 4He in hematite (Kula and Baldwin, 2012). Using the kinetic parameters from that paper and assuming rapid cooling of the fault zone subsequent to hydrothermal precipitation (at rates of greater than 1000 °C per m.y., in all likelihood), we calculated bulk 4He closure temperatures of around 300 °C for all samples, for an effective diffusion dimension of 500 mm. Inasmuch as this is comparable to the maximum temperatures consistent with the quartz deformation textures observed in the cataclasites, we suggest that the measured dates are crystallization ages rather than cooling dates. We interpret the variability of dates to reflect polyphase precipitation of hematite in an evolving brittle shear zone. While the data do not constrain the duration of deformation, we infer that the Lhuentse fault was experiencing brittle deformation during the Pleistocene. Beyond the significance of these data for understanding Lhuentse fault activity, they generally affirm the utility of hematite (U-Th)/He chronology for dating the activity of faults that serve as conduits for Fe-rich hydrothermal fluids.

EVOLUTIONARY GEODYNAMICS OF PT2 IN BHUTAN

As is the case in the central Nepal Himalaya, PT2 in Bhutan may indicate comparatively rapid uplift of the Higher Himalaya to the north in Late Tertiary time. However, we found no evidence in the course of our study that this has been the case over the past 2 m.y. Thermochronometrically constrained palinspastic reconstructions of the Himalayan thrust belt in Bhutan imply that both the Kakhtang thrust and several more deep-seated structures of the Lesser Himalayan duplexes were active in this sector of the orogen between 15 and 9.5 Ma (Long et al., 2012). Although PT2 in Bhutan may be a relict Miocene physiographic feature, it seems likely that either the formation or the preservation of PT2 may have been associated with the creation of the perched low-relief landscapes that lie to the south.

Several lines of evidence suggest a late Miocene–Holocene change in the tectonics of Bhutan concomitant with the development of the Shillong Plateau to the south and partitioning of a significant component of India-Eurasia convergence to structures south of the Himalayan thrust belt of Bhutan (e.g., Clark and Bilham, 2008). We infer that this north-to-south broadening of the region of convergence south of PT2 had an important effect on deformational patterns in our study area by Pliocene time, such that a greater amount of rock uplift (and consequent exhumation) occurred south of PT2. We regard the Lhuentse fault as, at least in part, a structure that initiated in response to differential uplift on either side of PT2.

We postulate that the underlying cause of local uplift south of PT2 was increasing structural relief at depth due to the activity of a blind duplex (Fig. 8). Long et al. (2011b) have suggested the existence of such a structure south of the trace of the Kakhtang thrust near the location of our cross section A-Aʹ based on palinspastic reconstruction of the Himalayan thrust belt in Bhutan. We propose that the Lhuentse fault developed as a very steep normal fault on the back side of the duplex, dipping toward the hinterland, to accommodate duplex growth and its southward migration as the orogenic wedge grew toward the foreland (Platt, 1986; Roure et al., 1991).

Whether or not the existence of a recently active fault at PT2 is only a local phenomenon or of regional significance in the eastern Himalayan region is unclear. Seismic studies in Bhutan reveal a complex modern deformation field that includes components of both north-south shortening and transcurrent slip (predominantly right-lateral) on structures that strike northwest-southeast at various angles relative to the overall east-west trend of the Bhutanese Himalaya (Fig. 9; De and Kayal, 2003; Rajendran et al., 2004; Andronicos et al., 2007; Drukpa et al., 2006; Velasco et al., 2007; Hazarika et al., 2010). We know of no evidence indicating that transcurrent deformation played an important role in the largely Miocene construction of the primary tectonic architecture of Bhutan (Gansser, 1983; Grujic et al., 1996; McQuarrie et al., 2008; Long et al., 2011a), which suggests that that seismic evidence for transcurrent deformation in Bhutan may be a recent phenomenon. It is plausible that the Lhuentse fault has accommodated both transcurrent and normal slip during its history, although field or remote-sensing evidence for the latter is presently lacking. If it occurred, transcurrent slip along a fault of its orientation would have enabled progressive eastward extrusion of the southeastern Tibetan Plateau and easternmost Himalaya in a manner similar to that inferred by Antolín et al. (2012).

A change in the distribution and/or mode of strain in the eastern Himalaya associated with a transfer of slip to the Shillong Plateau (e.g., Clark and Bilham, 2008) or a focus of uplift created by a reactivated duplex needs to be accounted for within models of mountain building and landscape evolution, as well as geohazard potential maps. This hypothesis is counter to models of fold-and-thrust belts that suggest the most active loci of deformation are near the foreland of the range.

CONCLUSIONS

Morphometric analysis of central and eastern Bhutan has led to the definition of a sharp transition between topographically subdued landscapes to the south and rugged landscapes of the Higher Himalaya to the north, analogous to, but discontinuous and more subdued than, the PT2 feature studied more extensively in central Nepal (e.g., Hodges et al., 2001; Wobus et al., 2006b). Field studies in the Kuri Chu drainage of central Bhutan demonstrated the existence of a previously unrecognized brittle deformational structure at PT2: the ENE-WSW–striking, steeply north-dipping Lhuentse fault. Remote-sensing analysis of the region suggests that the Lhuentse fault may be one structure of a family of high-angle, ENE-WSW–striking faults in the hinterland of the Bhutan Himalaya. Although exposures of the Lhuentse fault zone are good, we were unable to determine the structure’s kinematics on the basis of outcrop or thin section analyses of the associated cataclasites. Regional patterns of seismicity are consistent with structures oriented similarly to the Lhuentse fault having right-lateral transcurrent slip components, but we found no direct structural or geomorphic evidence of such kinematics. However, the results of zircon and (especially) apatite (U-Th)/He thermochronometry across the fault are consistent with ∼500 m of normal-sense displacement over the past ca. 2 m.y. Dating of syndeformational specular hematite precipitates in the fault zone by the (U-Th)/He method yielded results indicative of Quaternary deformation on the structure at least as recently as a few tens of thousands of years. We postulate that the Lhuentse fault, while having only minor displacement, is related to a larger-scale reorganization of the tectonic regime of the eastern Himalaya caused by southward expansion of the Himalayan orogenic wedge. Increasing complexity of the deformation field in what had been the hinterland of the Himalayan thrust belt in recent times should influence our assessments of seismic hazards in the region.

This work was supported by a National Science Foundation Tectonics Program grant EAR-0708714 to Hodges and a joint Tectonics and Geomorphology and Landuse Dynamics Programs grant EAR-1049888 to Whipple. We thank Emmanuel Soignard of the LeRoy Eyring Center for Solid State Science at Arizona State University for his help with the Raman spectrometer. This work was greatly strengthened by the assistance of and conversations with Frances Cooper (School of Earth Sciences, University of Bristol) and Arjun Heimsath (School of Earth and Space Sciences, Arizona State University), both in the field and laboratory. Field work would not have been possible without the support of our friends and colleagues in Bhutan: Peldon Tshering (National Environment Commission), Ugyen Wanda (Department of Geology and Mines), Karma Choden, and Ugyen Rinzen (Yangphel Adventure Travel). We thank Arlo Weil, Nadine McQuarrie, and two anonymous reviewers for their helpful comments on an earlier version of this manuscript.

1GSA Data Repository Item 2013228, additional thermochronometry sample information, ore identification via Raman spectroscopy, and elevation normalization of thermochronometry data, is available at www.geosociety.org/pubs/ft2013.htm, or on request from editing@geosociety.org, Documents Secretary, GSA, P.O. Box 9140, Boulder, CO 80301-9140, USA.