The Pamir Mountains, located between the Tarim and Tajik basins to the east and west, respectively, stand as an orographic highland at the northwestern end of the Tibetan Plateau, shaped as a northward convex orogen (Figure 1) resulting from the northward indentation of the Pamir salient over the Tajik basin following the India-Asia collision (e.g., Burtman and Molnar, 1993; Sobel and Dumitru, 1997; Arrowsmith and Strecker, 1999; Coutand et al., 2002; Zubovich et al., 2010; Chapman et al., 2017; Chen et al., 2018; Li et al., 2020). Consequent shortening during the Cenozoic has driven lateral extrusion through extension and exhumation as well as tectonic denudation (e.g., Cao et al., 2013a; Stübner et al., 2013; Thiede et al., 2013; Rutte et al., 2017a, Rutte et al., 2017b; Worthington et al., 2020). Such large-scale deformation is characterized by compressional and transpressional structures along the western (Darvaz fault), northern (Main Pamir Thrust: MPT), and eastern (Kashgar-Yecheng transfer system: KYTS) Pamir, and metamorphic crystalline basement domes bounded by extensional faults emplaced in the interior (Figure 1). In the Pamir interior, large areas of metamorphic domes exhumed at the surface have experienced high-grade metamorphism since the Oligo-Miocene as well as continued exhumation, facilitated by extension until the late Miocene (e.g., Robinson et al., 2004; Robinson et al., 2007; Robinson et al., 2010; Cao et al., 2013a; Cao et al., 2013b; Stübner et al., 2013; Cai et al., 2017; Rutte et al., 2017a, Rutte et al., 2017b; Worthington et al., 2020).

In the eastern Pamir, the Kongur Shan and Muztaghata domes mostly consist of Triassic granite, gneiss and schist within their core and some slivers of Paleozoic meta-sedimentary rocks in the north and east, which are bounded by the KES (Figures 1, 2). The KES includes the right-lateral/normal Muji fault to the NW, the Kalagile and Kuke normal faults and Ghez dextral fault to the east, as well as the Kongur Shan detachment fault (KSF), Taheman and Tashkurgan normal faults to the west, and the Shen-ti normal fault to the south (Figure 1). In addition, grade of metamorphism, magnitude of E-W extension, and exhumation rates in the rocks exposed in their footwalls (e.g., Robinson et al., 2004; Robinson et al., 2007; Robinson et al., 2010; Thiede et al., 2013) vary along the different fault segments.

The Muztaghata dome is a south-vergent, northwest-to-west-plunging antiform, bounded by the steeply east-dipping Kuke normal fault to the east, south-dipping Shen-ti normal fault to the south and north-dipping North Muskol normal fault to the north (Figure 1; Robinson et al., 2007). This dome is interpreted as the eastward continuation of the west-plunging Shatput dome in the central Pamir (e.g., Robinson et al., 2007; Sobel et al., 2011; Rutte et al., 2017a, Rutte et al., 2017b). Th-Pb and U-Pb dating on monazites and zircons reveal peak metamorphism of the Muztaghata dome in the late Oligo-Miocene (Robinson et al., 2007; Cai et al., 2017). Multiple thermochronometers with different closure systems, including mica and biotite ⁴⁰Ar/³⁹Ar and apatite and zircon fission track, record rapid exhumation of the antiform at ∼12–8 Ma (Robinson et al., 2007; Sobel et al., 2011; Cao et al., 2013a; Cai et al., 2017), interpreted to result from doming accommodated by shearing along the Kuke, Shen-ti and North Muskol faults (Robinson et al., 2007; Sobel et al., 2011; Rutte et al., 2017a; Cai et al., 2017). Dome exhumation has continued until the Pliocene, possibly related to EW-oriented initial extension along the KSF and Taheman normal fault (Cao et al., 2013a; Thiede et al., 2013). Low-temperature thermochronology data input into one-dimension thermal modeling yields decreasing footwall exhumation rates from ∼1 to ∼0.4–0.5 ± 0.1 mm/yr from north to south of the Muztaghata dome since the Pliocene (Thiede et al., 2013). Average basin-wide erosion rates during the last glacial cycle, derived from ¹⁰Be cosmogenic nuclides, range from 0.6 to 1.4 mm/yr on the western flank of the dome (Seong et al., 2009a).

To the north, the Kongur Shan is defined as a NW-trending, NNW-plunging dome (Figure 1; Robinson et al., 2004). It is structurally continuous with the Muztaghata and Kingata Tagh massifs to the south and north, respectively, and is limited by the east-dipping Ghez dextral fault to the east, the low-angle west-dipping KSF to the west, and the north-dipping Kalagile normal fault to the north (Figure 1; Cao et al., 2013a; Cao et al., 2013b; Robinson et al., 2004, Robinson et al., 2007). The dome has experienced peak metamorphism at ∼9.3 Ma dated by Th-Pb monazite (Robinson et al., 2004). Dome exhumation initiated in the late Miocene and lasted until present due to extension along the KSF (Arnaud et al., 1993; Brunel et al., 1994; Robinson et al., 2004; Robinson et al., 2007; Robinson et al., 2010; Cao et al., 2013a, 2013b; Thiede et al., 2013). One-dimension thermal models with mica and biotite ⁴⁰Ar/³⁹Ar, apatite fission track and zircon (U-Th)/He data yield footwall exhumation rates of >2.5 ± 1.0 and ∼2 ± 0.5 mm/yr just north of the Kongur Shan massif, as well as in the NW Kingata Tagh massif, respectively (Thiede et al., 2013). Unlike the crosscutting relationship between the KSF and the Muztaghata dome, the Kongur Shan dome formation is interpreted to mostly be synchronous with EW-extension along the KSF taking place since the late Miocene (Arnaud et al., 1993; Brunel et al., 1994; Robinson et al., 2004; Cao et al., 2013a; Thiede et al., 2013). In summary, thermochronology ages suggest that the Kongur Shan and Muztaghata domes have undergone focused exhumation since the late Miocene (Cao et al., 2013a; 2013b), with maximum rates near the domes, decreasing northwards and southwards (Thiede et al., 2013). Despite a significant component of rock exhumation triggered by extension along the KES, prominent topographic relief in the core of the domes carved by glaciers and rivers indicates a significant contribution of glacial/fluvial erosion to domal exhumation, which remains poorly constrained (Cao et al., 2013b; Thiede et al., 2013; Schoenbohm et al., 2014).

The KES represents the most prominent tectonic structure in the eastern Pamir. As part of this extensional system, the ∼250 km-long KSF is the most striking low-angle normal fault (Robinson et al., 2004; Robinson et al., 2007). It bounds the western flank of the Kongur Shan and Muztaghata domes and controls the development of an intermontane basin named the Muji-Tashkurgan basin in its hanging wall (Figures 1, 2). The onset timing and magnitude of extension on the KSF is well constrained compared to other regional fault segments. At the northern end of the KSF, rapid cooling at ∼8–7 Ma in the footwall is interpreted to date the fault onset with an inferred ∼30 km of total E-W extension (Robinson et al., 2004). In the Kongur Shan massif, two-dimension thermo-kinematic modeling results based on mica and biotite ⁴⁰Ar/³⁹Ar ages suggest that the activity of the KSF started at ∼7 Ma and continued until recently at a constant exhumation rate of ∼4 mm/yr, hence a minimum E-W extension of ∼34 km (Robinson et al., 2004; Robinson et al., 2010). By contrast, the southern portion of the KSF may have initiated later, at ∼5–6 Ma, as recorded by overlapping apatite fission track and zircon U-Th/He cooling ages in the footwall (Cao et al., 2013a; Thiede et al., 2013), synchronous with initiation of normal faulting along the Tashkurgan fault that was constrained by accelerated exhumation in the footwall at 6–5 Ma recorded by zircon and apatite U-Th/He data (Chen and Chen, 2020). Together with extensive 5–6 Ma peak ages from detrital zircon fission track from modern glacial and river sediments at the base of the domes, this argues for roughly synchronous onset of E-W extension along the entire length of the KSF at ∼6–5 Ma (Cao et al., 2013a; Thiede et al., 2013). Accordingly, onset of extension along the southern segment of the KSF at ∼5–6 Ma would postdate the Muztaghata doming at 12–6 Ma (Robinson et al., 2007; Sobel et al., 2011; Cao et al., 2013a), suggesting that it would then have been mostly accommodated by extension along the Shen-ti and Kuke normal faults farther south and east. The magnitude of late Miocene E-W extension is estimated to be ∼20 km in the Muztaghata dome region, and to decrease dramatically to ∼3 km along the Tashkurgan fault to the south, with no clear offset along the Tahman fault (Robinson et al., 2007).

The KES is an active fault system that triggered large earthquakes, such as, the 2016 Aketao Ms 6.6 earthquake (e.g., Li et al., 2019) along the Muji fault, and the 1895 Ms 7–7.3 Tashkorgan earthquake along the Tahman and Tashkurgan faults (Li et al., 2011; USGS) (Figure 1). In general, late Quaternary E-W extension rates similarly decrease from north to south. While the dextral slip rate along the EW-striking Muji fault was determined as 4.5–9 mm/yr over ∼10 ka (Chevalier et al., 2011; Li et al., 2019), no rate estimate yet exists on active faults bounding the Kongur Shan and Muztaghata domes. Farther south, E-W extension rates are ∼1–2 mm/yr just south of the Muztaghata dome (Chevalier et al., 2015). This southward rate decrease is consistent with late Miocene footwall exhumation rates, which confirms faster extension along the KES in the north than in the south (Chevalier et al., 2011).

The present-day climatic patterns in the eastern Pamir region are dominated by the mid-latitude Westerlies that come from the Mediterranean, Black and Caspian seas (e.g., Aizen and Aizen, 1997). The Kongur Shan and Muztaghata domes are mostly semi-arid, with annual average precipitation of 70, 128 and 300 mm recorded at Tashkurgan, Bulunkol and Muztaghata stations (Figure 1), respectively (1956–2005, Duan et al., 2007; Seong et al., 2009a), which feeds glaciers and rivers.

The domal structures are extensively glaciated with >60% occupied by glaciers or glacial sediments. A variety of glacial landforms are identified in the valleys and the Muji-Tashkurgan basin, including U-shaped valleys, horn, arête, cirque, moraines, glacio-fluvial fans, debris flow and alluvial fans. Recent glacial morphology investigations suggested pronounced asymmetry in glacial area, debris cover, and headwall relief: less extensive glaciers, less debris cover and gentler headwall relief on the western (downwind) side of the range compared to the eastern side, interpreted to reflect the complex interplay between lithology, tectonics and climate (Schoenbohm et al., 2014). Glacial geology and ¹⁰Be cosmogenic dating of moraine boulders revealed that glaciers advanced westward from the Kongur Shan and Muztaghata massifs at least 12 times during the last two glacial cycles since ∼300 ka (Seong et al., 2009b; Owen et al., 2012), including 10 minor glacial advances since the Last Glacial Maximum (LGM ∼20 ka), hence suggesting a correlation with Northern Hemisphere climate oscillations (Seong et al., 2009b). The extent of glaciation varies along the range with limited glacier extent in the Muji-Tashkurgan basin since at least 43 ka, in contrast to extensive glaciation near the high domes since the LGM (Owen et al., 2012; Schoenbohm et al., 2014; Seong et al., 2009a; Seong et al., 2009b).

The gneiss domes are mainly incised by two tributary rivers, the Ghez and Tashkurgan Rivers, which are both fed by glacier meltwater and ultimately flow into the Tarim basin to the east (Figure 1). The Ghez River, which collects water from the Muji and Kengxuwar Rivers, cut across the Kingata Tagh, Kongur Shan and Muztaghata massifs, yielding a remarkable relief of ∼3,600 m (Ghez valley), which creates a prominent knickpoint where the river meets the KSF (Seong et al., 2009a). Some of the former glacier valleys may be reshaped by glacier meltwaters along the drainage divide, evident by river terraces located at the valley outlets (Schoenbohm et al., 2014). Farther south, the Tashkurgan River cuts across the southern Muztaghata dome, creating a topographic relief of >2,000 m along the river course.

The Iat index is an integrated parameter that measures relative active tectonics taking into account multiple geomorphic indices: hypsometric integral (HI), stream length-gradient index (SL), drainage basin shape (Bs), drainage basin asymmetry (Af-50), ratio of valley-floor width to valley height (Vf), and mountain front sinuosity (Smf), partially based on the determination of the basin’s mean elevation and slope (El Hamdouni et al., 2008; Pérez-Peña et al., 2009; Alipoor et al., 2011). The multiple geomorphic indices were extracted from the 30-m resolution DEM. The geomorphic index values listed above can be divided into three classes from high to low tectonic activity. The combination of these indices, called Iat index, averages the different classes of geomorphic indices, and values can be divided into four classes which provide quantitative information of relative tectonic activity from strong to weak. In this study, we apply a revised Iat index that does not consider the mountain front sinuosity (Smf) because mountain fronts of catchments along the KSF are relatively straight at larger-scale, due to uplift along the normal faults.

Average erosion rates of a drainage basin can be expressed as a function of stream gradient and drainage area that includes the temporal river channel change in elevation (Dietrich et al., 1994; Hack, 1973; Flint, 1974; Howard and Kerby, 1983; Whipple and Tucker, 1999; Snyder et al., 2000). In the eastern Pamir region, most of the selected drainage basins show typical features of river valleys because of their concave and convex hypsometric curves (Strahler, 1952; Figure 3). This is reflected by the presence of fluvial terraces along the Muji fault and KSF (Schoenbohm et al., 2014). Few of them have S-shaped hypsometric curves in glaciated drainage basins, that was interpreted as resulting from progressive glacial modification on initial fluvial landscape (Brocklehurst and Whipple, 2004; Figure 3). Therefore, it is plausible to use the stream power model to simulate average basin-wide erosion rates.

The stream power model can be expressed as below: d z / d t = U - E = U - K A m S n (7) where z is elevation of the river profile; U is the rock uplift rate; A is the upstream drainage area; S is the local channel gradient; K is the erodibility coefficient related to tectonics, lithology, and climate; and m and n are positive constants reflecting basin hydrology, geometry, and erosion processes (Dietrich et al., 1994; Whipple and Tucker, 1999). At steady state, river channel erosion rate is equal to rock uplift rate (E = U), and channel erosion does not change through time (dz/dt = 0), Eq. 7 can be deduced as equilibrium channel gradient (S _e ):  Se=(U/K ) 1/n A (-m/n) (8) Where, in the case of uniform U, K, m and n, Ks=(U/K)^1/n and θ = -m/n (0.4<θ < 0.6) are channel steepness indices and intrinsic channel concavity indices, respectively (Flint, 1974; Snyder et al., 2000; Kirby and Whipple, 2001; Brocklehurst and Whipple, 2002; Wobus et al., 2006; Brocklehurst and Whipple, 2007; Kirby and Whipple, 2012). In various geologic settings, the power-law relation in Eq. 8 can be defined as: S ( dz/dx ) = ( E/K ) 1/n A -m/n   (9) where S is the local channel gradient. At steady state, river channel erosion rate is equal to rock uplift rate (E = U). Using slope-area analysis (Kirby and Whipple, 2001; Wobus et al., 2006; Kirby and Whipple, 2012), a linear regression of the logarithm slope-area relationship can be obtained, so that the slope and intercept indicate θ and log(Ks).

Alternatively, the stream power parameters can be derived from an integral approach for steady state rivers (Perron and Royden, 2013). The transformed profile ‘χ plots’ can be described as: z ( x ) =z ( x b )  + ( E K A 0 m ) 1/n χ,   χ = ∫ x b x ( A 0 A ( x ′ ) ) m n dx ′ (10) where x _b is the elevation of the river outlet that is usually set to zero; A ₀ is the reference drainage area that is set to 1 km². The channel steepness (Ksn) is determined directly from the slope of χ plots.

Geomorphic index values are listed in Supplementary Table S1. For comparison, the various indices sensitive to tectonic activity were arbitrarily divided into three classes, with Classes one and three indicating high and low activity, respectively. The limit between classes was determined using the Natural Breaks (Jenks) method in ArcGis 10.2, which generally agrees with changes in the range of values for the various indices.

The long-term exhumation of the domes derived from low-temperature thermochronology data (Table 1) were interpolated to the average basin-wide exhumation/erosion rates during the Quaternary by stream power model. Average basin-wide erosion rates were calculated based on topographic parameters (n = 1, θ _ref = 0.45, K _king = 7.31 × 10^–6 m^0.1/a; K _kongur = 1.86 × 10^–5 m^0.1/a; K _muztagh = 5.77 × 10^–6 m^0.1/a) for the region northwest of the Kingata Tagh, Kongur Shan, and Muztaghata massifs, respectively, by solving Eq. 5 (Figure 6F; Supplementary Table S1). In general, the modeled erosion rates peak in the Kongur Shan area and decrease northwestward and southward. The average basin-wide erosion rates in the Kongur Shan area are ∼3.84 ± 1.0 mm/yr, consistent with footwall exhumation rates of >2.5 ± 1.0 and ∼4 ± 1.0 mm/yr derived from thermochronology data (Robinson et al., 2010; Thiede et al., 2013). Northwest of Kingata Tagh, basin erosion rates are <0.5–2 mm/yr and are lowest at the northwestern end, with not much difference between basins on the eastern and western sides. In the Muztaghata region, erosion rates range from 0.1 to 1 mm/yr, with little difference between both sides of the range. The average erosion rate is 2.0 ± 0.5 mm/yr for all basins on the eastern side, slightly higher than that (1.5 ± 0.4 mm/yr) on the western side.

Previous studies applied a combination of two indices (e.g., Vf and Smf) to determine the relative degree of tectonic activity (Bull and Mcfadden, 1977; Silva et al., 2003) and recurrence intervals of surface ruptures related to earthquakes along mountain fronts (Silva et al., 2003). More geomorphic indices (e.g., SL, Bs, Af-50, and HI) were later introduced to evaluate regional tectonics (Ramírez-Herrera, 1998; El Hamdouni et al., 2008; Pérez-Peña et al., 2009; Alipoor et al., 2011; Bhat et al., 2020). These parameters were averaged to generate an integrated index of relative active tectonics (Iat) (El Hamdouni et al., 2008); however, the validity of various morphologic indices has not yet been quantitatively evaluated despite the fact that inconsistent indices have been artificially excluded in the calculation (El Hamdouni et al., 2008).

In this study, we attempt to assess the sensitivity of various geomorphic indices to active tectonics that were used to generate the Iat index, by examining correlations between the values of Iat and SL, Bs, Af-50, Vf and HI, respectively (Figure 7). Our best-fit solution shows negative correlation between Iat values and SL and Bs, and positive correlation between Iat and Vf values, with much higher correlation coefficients than with Af-50 and HI values (Figures 7A–D). This suggests that SL, Bs, and Vf indices are more sensitive to tectonic activity than Af-50 and HI in the eastern Pamir domes. Accordingly, we only use the average classes of SL, Bs, and Vf to produce a new Iat index (Supplementary Table S1), which is consistent with the presence of active faults on the western side of the divide (Figure 6E). As such, we suggest that the validity of multiple geomorphic indices should be evaluated by correlation test and active fault mapping before being integrated as a reliable index of relative active tectonics.

Field investigation on the late Quaternary vertical and horizontal slip rates of the KSF and Muji fault have been undertaken (Chevalier et al., 2011, Chevalier et al., 2015; Li et al., 2019), but along-strike variations in tectonic activity as well as footwall exhumation along the KES have not been fully resolved yet. Whether the southern end of the KSF presents intense recent activity remains debated (Robinson et al., 2007; Yuan et al., 2013). In Figure 6E, Iat values suggest relatively strong normal faulting along the KSF on the western side of the divide compared to the Ghez and Kalagile faults to the east, consistent with eastward crustal tilting as revealed by low temperature thermochronology ages across the domal structures along the Ghez and Tashkurgan rivers (Sobel et al., 2011; Cao et al., 2013a; Thiede et al., 2013).

Interestingly, relatively intense tectonic activity concentrates on three segments of the western KES: the right-lateral Muji fault and the KSF bounding the Kingata Tagh—Kongur Shan, and the western and southern Muztaghata domes (Figure 6E), as supported by field observations. For example, numerous clear right-lateral (along segments oriented ∼ EW to WNW-ESE) and vertical offsets of Quaternary moraines and alluvial fans exist along the Muji fault and KSF, with Holocene horizontal rates of >4.5–9 mm/yr on the former (Chevalier et al., 2011; Li et al., 2019) and of ∼1 mm/yr on the latter at the base of the Muztaghata massif (Chevalier et al., 2015). Likewise, clear vertical offsets and triangular facets all along the KSF bounding the Kongur Shan massif reflect strong tectonic activity (Arnaud et al., 1993; Robinson et al., 2004; Cao et al., 2013a; Chevalier et al., 2015), with late Miocene exhumation of domal rocks at rates of >2–4 mm/yr (Cao et al., 2013a, Cao et al., 2013b; Thiede et al., 2013). Despite the fact that both extension rates and magnitude significantly decrease to the south (Robinson et al., 2007; Chevalier et al., 2011; Chevalier et al., 2015; Li et al., 2019), the KSF still shows evidence of clear tectonic activity immediately south of the Muztaghata massif, reflected by Quaternary vertical offsets of alluvial fans with throw rates of ∼1.7 mm/yr (Chevalier et al., 2015). In addition, two large landslides along the western Muztaghata dome have been interpreted to have been triggered by large earthquakes (Yuan et al., 2013) although the influence of glaciers or bedrock fractures cannot be excluded (Seong et al., 2009a). Therefore, we consider that the tectonic activity along the southern end of the KSF exists since its initiation at ∼5–6 Ma (Cao et al., 2013a; Thiede et al., 2013). Tectonic activity along the KES may shift drainage basins from mature to immature states, currently reflected by convex and S-shaped hypsometric curves on the KSF-controlled (western) side, in contrast to concave and S-shaped curves on the eastern side, indicative of greater mass loss due to glacial excavation, as suggested by the presence of larger glaciers (Schoenbohm et al., 2014).

Unlike the relative intensity of tectonic activity along the KES (Figure 6E), modeled average basin-wide erosion rates clearly differ: highest in the Kongur Shan dome area, and gradually decreasing to the north and south (Figure 6F). This trend in along-strike variations in the magnitude of basin-wide erosion rates broadly coincides with late Miocene exhumation rates in the footwall of the KSF, determined by low-temperature thermochronology (Thiede et al., 2013; Figure 8). While no clear correlation exists between Iat values and erosion rates for all catchments (Figure 7F), spatial coincidence of high erosion rates and strong active tectonics (low Iat) around the Kingata Tagh—Kongur Shan massifs indicates complex feedback between glacial/fluvial erosion and tectonics, as discussed in the next section. Interestingly, average erosion rates are also comparable to millennial-scale basin-wide erosion rates (Figure 1, Figure 6F). For the drainage basins W69, W70, and W71 located on the western flank of the Muztaghata dome for example, modeled average basin-wide erosion rates of 0.6 ± 0.1 to 1.3 ± 0.3 mm/yr (Supplementary Table S1) overlap well with average basin-wide erosion rates of ∼0.6–1.4 mm/yr in the Holocene determined by ¹⁰Be cosmogenic dating (Seong et al., 2009a). In addition, our results constrain for the first time, erosion rates <1.5 mm/yr in the footwall of the Muji fault, i.e., almost half that of the exhumation/erosion rates of 2–4 mm/yr in the footwall of the KSF near the Kongur Shan massif (Thiede et al., 2013). This trend most likely reflects the smaller vertical component along the mainly strike-slip Muji fault (Chevalier et al., 2011; Chevalier et al., 2015; Li et al., 2019), in contrast to the KSF to the south, even though less extensive glacial coverage has been suggested to influence such trend (Schoenbohm et al., 2014).

Detrital zircon fission-track data from modern glacial and river sediments reveal focused Quaternary exhumation of the domes (Cao et al., 2013b), but mechanisms for this episode of exhumation/erosion remains uncertain. In the Kongur Shan massif area, accelerated exhumation rates since 2 Ma were interpreted to be due to rapid, recent cooling (Arnaud et al., 1993). However, two-dimension thermal models suggested that this apparent increase in cooling rates may be explained by advection of isotherms resulting from rapid dome exhumation from ∼7 Ma to present at a constant rate of 4.2–6.5 mm/yr (Robinson et al., 2010). In the southern Muztaghata dome, a few apatite fission track ages of <3 Ma were suggested to record exhumation related to continued doming (Sobel et al., 2011). Alternatively, focused denudation of the dome system may be explained by positive feedback of extension along the KES in response to unloading of glacial/fluvial erosion (Cao et al., 2013b), since significant Quaternary exhumation is inevitably accomplished by glacial/fluvial erosion, as shown by the prominent topographic relief in the Kingata Tagh, Kongur Shan, and Muztaghata domes (Seong et al., 2009a; Thiede et al., 2013; Schoenbohm et al., 2014). More importantly, along-strike changes in average basin-wide erosion rates and footwall exhumation rates may be explained by other mechanisms including factors such as lithology, tectonics, and/or glacial/fluvial erosion, having the most control on the exhumation/erosion of the dome system during the Quaternary.

First, lithology is not a major factor controlling along-strike changes in Quaternary exhumation/erosion rates of the domes, although rock resistivity clearly influences surface erosion processes (Sklar and Dietrich, 2001; Korup and Schlunegger, 2009; Schoenbohm et al., 2014). In fact, the maximum exhumation/erosion rates coincide with the location of strong Triassic granite and gneiss in the Kongur Shan area, while weaker Paleozoic meta-sedimentary rocks correlate with reduced exhumation/erosion rates in the Kingata Tagh massif area (Korup and Schlunegger, 2009; Figure 6F, Figure 7). However, the fact that the Muztaghata dome has similar rock resistivity as that of the Kongur Shan dome, but yields much lower exhumation/erosion rates, suggests limited lithologic influence on the change in magnitude of exhumation/erosion rates along the range.

Second, E-W extension along the KSF likely plays a dominant role in controlling dome exhumation/erosion at long-term timescales. Thermochronology data combined with estimated magnitudes of extension yield long-term extension rates along the KSF of: 1) 3.8–4.3 mm/yr (∼30 km since ∼8–7 Ma) in the northern Kingata Tagh (Robinson et al., 2004); 2) >4.9 mm/yr (>34 km since ∼7 Ma) in the Kongur Shan area (Robinson et al., 2004; Robinson et al., 2010); 3) 3.3–4.0 mm/yr (∼20 km since 6–5 Ma) in the Muztaghata area (Robinson et al., 2007; Cao et al., 2013a); and 4) <0.5 mm/yr (<3 km since 6–5 Ma) south of the Muztaghata massif (Robinson et al., 2007; Cao et al., 2013a). Similar late Quaternary dextral rates of >4.5–9 mm/yr have been determined NW of Kingata Tagh along the Muji fault (Chevalier et al., 2011; Li et al., 2019), and ∼EW extension rates of 1.3–2 mm/yr immediately south of Muztaghata (throw rates of 1.1–1.7 mm/yr with KSF dip angle of 35–45°; Chevalier et al., 2015). Along-strike variations in both long-term exhumation rates and average basin-wide erosion rates in the footwall of the KSF positively correlate with extension rates along the fault since the late Miocene (Figure 8), suggesting that extensional tectonics impacts exhumation/erosion patterns along the domes at long-term timescales.

By contrast, negative correlation between exhumation/erosion rates and Iat values around the Kingata Tagh and south Muztaghata massifs (Figure 8) seemingly supports the idea that tectonic activity along the KSF is not the main driver for the Quaternary dome exhumation/erosion. This is supported by an absence of clear correlation between Iat values and modeled and measured erosion rates during the late Quaternary (Figure 8). In particular, the higher-relief regions (e.g., Kongur Shan and Muztaghata) with maximum glacier debris cover (Schoenbohm et al., 2014) correlate with higher exhumation/erosion rates (Cao et al., 2013a; Robinson et al., 2007), which highlights the importance of glaciers in shaping alpine landscapes. Indeed, glacial processes such as deepening and widening of the glacial valleys as well as headwater erosion, remove large amounts of bedrock which hence enhance valley relief (MacGregor et al., 2000; Brocklehurst and Whipple, 2002; Brocklehurst and Whipple, 2006; Brocklehurst and Whipple, 2007; Shuster et al., 2011; Schoenbohm et al., 2014). This is attested by relatively lower HI values and S-shaped hypsometric curves around the summits of the Kongur Shan and Muztaghata massifs, but with steeper headwall relief (Figures 3, 5). Additionally, topographic swath profiles show greater mean basin relief and slopes on the eastern side compared to the western side of the range (Figures 5D,E). This pronounced asymmetry is consistent with contrasting glacial landforms, with larger glacier surfaces, more debris cover, and steeper headwalls on the eastern (downwind) side of the range (Schoenbohm et al., 2014).

In actively deforming orogens, the glacial buzz-saw hypothesis is believed to exert a major control on the development of alpine topography, which predicts maximum glacial erosion around and above the mean long-term equilibrium line altitude (Aizen et al., 2001; Berger et al., 2008; Brozovic et al., 1997; Egholm et al., 2009; Spotila et al., 2004; Sternai et al., 2011). Where glacial buzz-saw is efficient, glacial morphology correlates with exhumation/erosion rates so that zones of high exhumation rates correlate with those of high erosion rates (Hallet et al., 1996), where taller glacier headwalls correspond to more resistant rocks (Brocklehurst and Whipple, 2002, Brocklehurst and Whipple, 2007; Naylor and Gabet, 2007; Scherler et al., 2011). In the eastern Pamir region, we observe that high average basin-wide erosion rates indeed correlate with high exhumation rates around the Kongur Shan and Muztaghata massifs, where taller headwalls and greater basin relief and slopes are more abundant than anywhere else along the range (Figures 5, 7; Schoenbohm et al., 2014; this study). Triassic igneous and metamorphic rocks that make up these two domes are indeed more resistant than the meta-sedimentary rocks to the north. The coupling of higher exhumation rates, higher erosion rates, stronger rock resistivity, and steeper glacial landforms supports the glacial buzz-saw mechanism in modulating the recent landscape of the eastern Pamir domes. Nevertheless, the summits of the Kongur Shan and Muztaghata domes rising much higher than the regional equilibrium line altitude (Seong et al., 2009a) imply that their formation may escape the effect of the glacial buzz-saw (Schoenbohm et al., 2014) and may be protected by glaciers (Thomson et al., 2010). However, the feedback between exhumation of deep-seated metamorphic rocks in response to surface glacial erosion through change of strain field remains unclear and merits further work.