INTRODUCTION
Sediment aggradation and incision along Himalayan rivers are influenced by precipitation changes and tectonics (Goodbred, Reference Goodbred2003; Bookhagen et al., Reference Bookhagen, Thiede and Strecker2005; Srivastava and Misra, Reference Srivastava and Misra2008; Srivastava et al., Reference Srivastava, Tripathi, Islam and Jaiswal2008; Ray and Srivastava, Reference Ray and Srivastava2010; Scherler et al., Reference Scherler, Bookhagen, Wulf, Preusser and Strecker2015). Understanding of such riverine processes together with their spatial and temporal changes is central to the understanding of landscape evolution in the Himalaya.
At the valley scale, river aggradation and incision of valley fill sediments have been linked to climate (Pratt-Sitaula et al., Reference Pratt-Sitaula, Burbank, Heimsath and Ojha2004; Ray and Srivastava, Reference Ray and Srivastava2010; Scherler et al., Reference Scherler, Bookhagen, Wulf, Preusser and Strecker2015). Incision into the bedrock along with the formation of strath terraces has been linked to rock uplift by tectonic movements (Pazzaglia et al., Reference Pazzaglia, Gardner and Merritts1998; Hancock and Anderson, Reference Hancock and Anderson2002; Dortch et al., Reference Dortch, Owen, Dietsch, Caffee and Bovard2011a; Srivastava et al., Reference Srivastava, Ray, Phartiyal and Sharma2013). Studies along rivers draining the southern front of the Himalaya—namely, the Sutlej, Yamuna, Bhagirathi, Alaknanda, and Gandaki, which receive waters from the southwest (SW) Indian summer monsoon (ISM)—have provided an understanding of aggradation and incision cycles and their relationship to climate and tectonics (Lavé and Avouac, Reference Lavé and Avouac2000; Thiede et al., Reference Thiede, Bookhagen, Arrowsmith, Sobel and Strecker2004; Bookhagen et al., Reference Bookhagen, Thiede and Strecker2005; Juyal et al., Reference Juyal, Sundriyal, Rana, Chaudhary and Singhvi2010; Ray and Srivastava, Reference Ray and Srivastava2010; Devrani and Singh, Reference Devrani and Singh2014; Morell et al., Reference Morell, Sandiford, Rajendran, Rajendran, Alimanovic, Fink and Sanwal2015; Sharma et al., Reference Sharma, Bartarya and Marh2016a). However, little attention has been given to the rivers that drain the more arid regions of the northwest (NW) Himalaya, which besides the summer monsoon also receive waters from the midlatitude westerlies. These rivers, which include the Indus and its tributaries, are influenced by the tectonic movements associated with the Indus-Tsangpo Suture Zone (ITSZ), the western syntaxes, and the Karakoram Fault (Burbank et al., Reference Burbank, Leland, Fielding, Anderson, Brozovic, Reid and Duncan1996; Leland et al., Reference Leland, Reid, Burbank, Finkel and Cafee1998; Jamieson et al., Reference Jamieson, Sinclair, Kirstein and Purves2004; Pant et al., Reference Pant, Phadtare, Chamyal and Juyal2005; Phartiyal et al., Reference Phartiyal, Sharma, Upadhyay and Sinha2005; Dortch et al., Reference Dortch, Owen, Dietsch, Caffee and Bovard2011a; Blöthe et al., Reference Blöthe, Munack, Korup, Fülling, Garzanti, Resentini and Kubik2014; Sinclair et al., Reference Sinclair, Mudd, Dingle, Hobley, Robinson and Walcott2017). The Indus River originates in Tibet, flows through Ladakh, and crosses the western syntaxes and forms a megafan on the plains of Pakistan before reaching the Arabian Sea. Sediments deposited by the river provide useful information on the climate and tectonic regimes through time.
Studies in the upper reaches of the Indus River system have so far dealt with the following: (1) chronology of past glaciations (Derbyshire and Owen, Reference Derbyshire and Owen1997; Owen and Benn, Reference Owen and Benn2005; Owen et al., Reference Owen, Caffee, Bovard, Finkel and Sharma2006a; Dortch et al., Reference Dortch, Owen and Caffee2013; Sharma et al., Reference Sharma, Chand, Bisht, Shukla, Bartarya, Sundriyal and Juyal2016b; Orr et al., Reference Orr, Owen, Murari, Saha and Caffee2017); (2) catchment-scale erosion rates (Dortch et al., Reference Dortch, Owen, Schoenbohm and Caffee2011b; Clift and Giosan, Reference Clift and Giosan2014; Munack et al., Reference Munack, Korup, Resentini, Limonta, Garzanti, Blöthe, Scherler, Wittmann and Kubik2014); (3) neotectonic deformation (Sinclair et al., Reference Sinclair, Mudd, Dingle, Hobley, Robinson and Walcott2017); (4) phases of aggradation and incision along segments of the river and their relationship to climate (Dortch et al., Reference Dortch, Owen, Dietsch, Caffee and Bovard2011a; Sant et al., Reference Sant, Wadhawan, Ganjoo, Basavaiah, Sukumaran and Bhattacharya2011a, Reference Sant, Wadhawan, Ganjoo, Basavaiah, Sukumaran and Bhattacharya2011b; Blöthe et al., Reference Blöthe, Munack, Korup, Fülling, Garzanti, Resentini and Kubik2014); (5) paleolake records (Phartiyal and Sharma, Reference Phartiyal and Sharma2009; Sangode et al., Reference Sangode, Phadtare, Meshram, Rawat and Suresh2011, Reference Sangode, Rawat, Meshram, Phadtare and Suresh2013; Sant et al., Reference Sant, Wadhawan, Ganjoo, Basavaiah, Sukumaran and Bhattacharya2011b); and (6) paraglaciation (Owen et al., Reference Owen, Bailey, Rhodes, Mitchell and Coxon1997, Reference Owen, Gualtieri, Finkel, Caffee, Benn and Sharma2001; Pant et al., Reference Pant, Phadtare, Chamyal and Juyal2005; Saha et al., Reference Saha, Sharma, Murari, Owen and Caffee2016; Sharma et al., Reference Sharma, Chand, Bisht, Shukla, Bartarya, Sundriyal and Juyal2016b). These studies, however, do not provide a synoptic view of the evolution of the upper Indus River and its relation to climate and tectonics.
In the upper reaches of the Indus River, two terrace types exist: (1) valley fill river terraces that rise from the riverbed and preserve sedimentary records of valley aggradation followed by incision; and (2) strath terraces characterized by a fluvially incised bedrock overlain by a thin alluvial sediment cover. Chronology of both the types of terraces enables computation of climate-induced aggradation and bedrock incision rates (Bull, Reference Bull1990; Pazzaglia et al., Reference Pazzaglia, Gardner and Merritts1998; Hancock and Anderson, Reference Hancock and Anderson2002; Starkel, Reference Starkel2003; Ray and Srivastava, Reference Ray and Srivastava2010).
Geodetic surveys between Leh and the Karakoram during 1998 and 2008 show statistically insignificant horizontal velocities (0.4±1 mm/a), and this makes it difficult to quantify internal deformation within the ITSZ (Jade et al., Reference Jade, Rao, Vijayan, Gaur, Bhatt, Kumar, Jaganathan, Ananda and Kumar2011). Contrastingly, a recent study on the tectonic geomorphology of the river around Leh indicates a slip of 21 cm/ka during the past ~45 ka. This slip is on the Stok Thrust between Indus Molasse and the Ladakh Batholith (Sinclair et al., Reference Sinclair, Mudd, Dingle, Hobley, Robinson and Walcott2017).
The present study builds on the work of Sinclair et al. (Reference Sinclair, Mudd, Dingle, Hobley, Robinson and Walcott2017) and provides an optically stimulated luminescence (OSL) chronology for valley fill and strath terraces at 18 locations between the villages of Nyoma (34°37.96′N, 76°27.40′E) and Dah (33°9.37′N, 78°44.12′E). OSL ages combined with geomorphic and sedimentologic data help define the timing and duration of aggradation and bedrock incision and enable the development of a model for the formation of river and strath terraces.
CLIMATE AND GEOLOGY
The study area is located in the rain shadow of the Greater Himalaya and is a cold desert of the Köppen’s BWk type (Fig. 1). Summers (May to September) are short with mean temperatures of 12.4°C, and winters (October to March) have mean temperatures of −1.4°C (Demske et al., Reference Demske, Tarasov, Wünnemann and Riedel2009). Annual precipitation in Leh is ~100 mm and occurs mostly as snowfall (Lang and Barros, Reference Lang and Barros2004).The region experiences two climatic systems, the SW ISM and midlatitude westerlies. The ISM advects moisture from the Arabian Sea and the Bay of Bengal, brings rainfall to the Himalaya during the summer, and provides ∼50% of the hydrologic budget for the Indus River (Bookhagen et al., Reference Bookhagen, Thiede and Strecker2005; Anders et al., Reference Anders, Roe, Hallet, Montgomery, Finnegan and Putkonen2006; Hobley et al., Reference Hobley, Sinclair and Mudd2012). The midlatitude westerlies provide snowfall and contribute the remaining ~50% to the hydrologic budget (Bookhagen and Burbank, Reference Bookhagen and Burbank2010).
Figure 1 (color online) (A) Geology of the study area, Indus River, and study locations. (B) Geologic cross section along X-Y showing deformation in molasse and presence of north-verging Zanskar Thrust, Choksti Thrust, and Upshi-Bazgo Thrust (after Searle et al., Reference Searle, Pickering and Cooper1990).
The river crosses the two principal geologic units of the ITSZ: (1) the Indus Molasse comprising basal shallow marine Zanskar platform sediments (the Tar group) followed by the terrestrial sediments of the Indus group (Thakur and Misra, Reference Thakur and Misra1984; Sinclair and Jaffey, Reference Sinclair and Jaffey2001; Wu et al., Reference Wu, Clift and Yang2007; Henderson et al., Reference Henderson, Najman, Parrish, BouDagher-Fadel, Barford, Garzanti and Andò2010, Reference Henderson, Najman, Parrish, Mark and Foster2011; Singh et al., Reference Singh, Sahni, Jain, Upadhyay, Parcha, Parmar and Agarwal2015); and (2) the granitic Ladakh Batholith (Fig. 1).
The rocks of the Tar and Indus groups form part of the Zanskar ranges and are separated by a high-angle, north-verging back thrust, known as the Choksti Thrust (Thakur, Reference Thakur1983; Searle, Reference Searle1986; Searle et al., Reference Searle, Pickering and Cooper1990; Sinclair and Jaffey, Reference Sinclair and Jaffey2001). In addition, the Indus rocks are internally deformed by a back thrust called the Stok Thrust. The entire package of Indus Molasse is thrusted over the Ladakh Batholith along the north-verging Upshi-Bazgo Thrust (Fig. 1B; Brookfield and Andrews-Speed, Reference Brookfield and Andrews-Speed1984; Searle et al., Reference Searle, Pickering and Cooper1990; Sinclair et al., Reference Sinclair, Mudd, Dingle, Hobley, Robinson and Walcott2017).
METHODOLOGY
A Shuttle Radar Topography Mission (SRTM) digital elevation model (DEM) with a resolution of 3 arc seconds, Survey of India topographic sheets (1:50,000), and field survey were used: (1) to delineate river patterns; (2) to identify terrace types; (3) to construct cross-valley and longitudinal river profile; and (4) for stream length (SL) gradient index analysis.
Morphostratigraphy
Morphostratigraphy of the sections used terrace elevation above the contemporary river level. The active floodplain was named T-0, and the terraces at successively higher elevation were named T-1 (youngest) and T-2 (older). The height of terraces above the contemporary river level (arl) was measured using a total station with a vertical resolution of 2–3 mm. The survey was carried out during August, and therefore, elevations are with respect to the flood stage. Bedrock incision rate based on strath terraces was determined by dividing the height of the terrace by its age. An uncertainty of ±5% was assumed owing to lateral geomorphic variability of the strath surfaces. Errors in the age and strath height were propagated in the computation of incision rate.
The terrace sediments were described using gravel sizes, roundness, sorting, imbrication, lithology, and matrix percentage. Using these, various lithofacies were identified and vertical graphic sedimentary logs were prepared. Further, genetically related sediment units were classified as lithofacies associations (e.g., facies formed as result of channel processes were considered as genetically related). Clast composition and textural data were determined using a 1×1 m grid that was placed on the exposed face of the sections using Howard (Reference Howard1993).
Chronology
The chronology was based on OSL. Use of radiocarbon dating in the study area has been challenging because of the paucity of datable material and contamination by older carbon supplied from the limestone terrain (Juyal et al., Reference Juyal, Pant, Basavaiah, Bhushan, Jain, Saini, Yadava and Singhvi2009; Phartiyal et al., Reference Phartiyal, Sharma, Srivastava and Ray2009).
OSL dating is based on the zeroing of the stored luminescence in sediment grains by daylight exposure during erosion, transport, and deposition. Reaccumulation of luminescence is initiated on burial because of irradiation arising from the decay of naturally occurring radionuclides. The luminescence accumulation continues unabated until excavation, and this luminescence carries the information on the age. Conversion of luminescence intensity to radiation dose units is done using OSL measurements and the annual radiation dose. The dating protocol thus includes the following: (1) estimation of the equivalent laboratory beta dose (D E ), which is the laboratory beta dose that induces the luminescence intensity equal to that which the sediment has received naturally; and (2) computation of annual dose rate (D T ) determined by measuring the elemental concentration of natural U, Th, and K radionuclides and cosmic dose. Samples for OSL dating were collected using stainless-steel tubes and processed using the standard laboratory protocols (Aitken, Reference Aitken1998). Single aliquot regeneration (SAR) method of Murray and Wintle (Reference Murray and Wintle2000) was used to determine D E values. Double SAR method was used in seven samples where the feldspar contamination could not be removed by chemical treatment (see Supplementary Data 1 for SAR and double SAR protocols; Jain et al., Reference Jain, Murray, Bøtter-Jensen and Wintle2005).
A critical issue in the reliability of OSL ages is the completeness of zeroing of luminescence prior to burial. This depends on the amount of daylight exposure that was available to mineral grains in the sediment. Turbulence and turbidity of the transporting medium attenuates the intensity and modifies the spectrum daylight. Such conditions lead to heterogeneous zeroing at the grain level and to a wide distribution of D E values. Based on the existing dating methods and protocols, a mean age model was used for samples with normally distributed D E values (Galbraith et al., Reference Galbraith, Roberts, Laslett, Yoshida and Olley1999).
The samples were considered well bleached if (1) the luminescence intensity and D E values for all the measured aliquots were poorly correlated (see Colls et al., Reference Colls, Stokes, Blum and Straffin2001) and (2) the ratio of standard deviation and mean of D E values (the coefficient of variation S n ) was greater than 0.1. Use of criterion 1 indicated that all 33 samples were well bleached, and the use of criterion 2 suggested that 21 samples were well bleached (Clarke, Reference Clarke1996; Srivastava et al., Reference Srivastava, Brook, Marais, Morthekai and Singhvi2006; see Supplementary Data 1 for details). Thus, overall, the mean age model was considered to be appropriate. The DT was computed by measuring the concentration of U, Th, and K in samples using X-ray fluorescence. Water content was assumed to be 10±5% by weight, and the cosmic dose was determined using Prescott and Hutton (Reference Prescott and Hutton1994). More details on the methodology and dating protocols are presented in Supplementary Data 1.
River parameters
The SL gradient index and steepness index (K s ) helped delineate tectonic perturbations in the river catchment (Kirby et al., Reference Kirby, Whipple, Tang and Chen2003; Morell et al., Reference Morell, Sandiford, Rajendran, Rajendran, Alimanovic, Fink and Sanwal2015). SL index was computed using the equation of Hack (Reference Hack1973):
$${\rm SL}{\,\equals\,}{{H_{1} {\minus}H_{2} } \over {{\rm ln}\,L_{2} {\minus}\ln L_{1} }},$$
where H 1 and H 2 were the elevation from mean sea level for the points at distances L 1 and L 2 from the source. The Ks index provides the spatially averaged uplift rate (Whipple and Tucker, Reference Whipple and Tucker1999), but it ignores aspects such as: (1) nonlinearity in incision (Whipple et al., Reference Whipple, Hancock and Anderson2000); (2) channel morphology (Lavé and Avouac, Reference Lavé and Avouac2000, Reference Lavé and Avouac2001); and (3) riverbed morphology (Hancock and Anderson, Reference Hancock and Anderson2002; Whipple and Tucker, Reference Whipple and Tucker2002). However, given that the SL index is related to flow resistance and erodibility contrast at a particular reach of the channel, it depends on lithological changes and tectonics. We used it as a first-order tool to elucidate possible tectonic controls. To compute the SL index, longitudinal river profile was prepared by changing the projection of SRTM DEM from WGS 84 to Lambert conformal conic (to transform the units from degrees to meters). The Environmental System Research Institute’s ArcGIS 9.3 software was used for corrections and data processing. Hydrologic tools of ArcGIS and standard protocols to generate a stream from DEM were utilized. Small channels with pixel values less than the trunk channel were deleted, and the fragments were merged to form a single linear stream. The length of the trunk channel was then divided into small segments, each separated by the SRTM DEM pixel size of 90 m. These line features were then converted into points and were added with field x, y (latitude, longitude). The z values (raster values) were extracted from the DEM in the attribute table using the “extract values to the points” tool. The attribute table was exported in text format and used to compute the longitudinal river profile and SL gradient index.
K s , the ratio of channel gradient and drainage basin area, was also utilized to identify and quantify the knick points, and we used the data of Munack et al. (Reference Munack, Korup, Resentini, Limonta, Garzanti, Blöthe, Scherler, Wittmann and Kubik2014).
RESULTS
Geomorphic setting and longitudinal river profile of the Indus River
Based on channel gradient, knick points, valley width, and channel pattern, the river was divided into four segments: segment 1, Nyoma to Mahe; segment 2, Mahe to Upshi; segment 3, Upshi to Spituk; and segment 4, Spituk to Dah (Fig. 2A). Table 1 collates the geomorphic characteristics of each segment. Valley widths in these segments are 4000 m (segment 1), 200 m (segment 2), 6000 m (segment 3), and 100 m (segment 4). The channel gradient ranged from 0.75 to 7.5 m/km (Fig. 2B).
Figure 2 Longitudinal river profile (A) of Indus River valley width (B) from Mahe upstream to Dah (after Munack et al., Reference Munack, Korup, Resentini, Limonta, Garzanti, Blöthe, Scherler, Wittmann and Kubik2014). (C) Stream length (SL) gradient index plotted along the Indus River. (D) K s plotted along the channel (after Munack et al., Reference Munack, Korup, Resentini, Limonta, Garzanti, Blöthe, Scherler, Wittmann and Kubik2014). Note the four segments of Indus River with distinct channel gradient and SL index. Segment 1 seems to be the southwest edge of the Tibetan plateau. (E) Google Earth image showing geomorphology along the Indus River as it flows in a narrow gorge in segment 2 (Mahe to Hymia). Note the study locations. Yellow stars locate various studied sections. The yellow rectangles are insets to Figures 3 and 4. m amsl, meters above mean sea level. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Table 1 Valley configuration, geomorphology, and channel gradient along the Indus River.
The SL index was the lowest (490) in segment 1 and highest in segment 4 (>19,500). Noteworthy observations include the following: (1) in segment 2, although the river traversed through the same lithology (the Indus Molasse), a sudden increase in the SL index (~4500) was seen; (2) the SL decreases in segment 3 where the river flows along the contact between the Ladakh Batholith and the Indus Molasse; and (3) the SL increases in segment 4 with deformed Indus Molasse. Within this segment, the river shows a sharp increase in SL index downstream from Skyurbuchan (34°26.01′N, 76°42.90′E), from where the river flows across the Ladakh Batholith (Fig. 2C). The K s index in segments 1 and 3 is low, whereas it is high in segments 2 and 4 (Munack et al., Reference Munack, Korup, Resentini, Limonta, Garzanti, Blöthe, Scherler, Wittmann and Kubik2014), which suggests a definitive role of tectonics.
Sedimentation and lithofacies
We prepared 28 graphic lithological logs, and based on sediment characteristics, seven lithofacies were identified. These were coded as per Miall (Reference Miall1996). Table 2 presents the characteristics of individual lithofacies. The gravel lithofacies had three subfacies: (1) subfacies A, where the constituent clasts were granitic derived from the Ladakh Batholith; (2) subfacies B comprised clasts from Indus Molasse; and (3) subfacies C comprised a mixed source and represented deposition in the trunk channel. Lithofacies were grouped into two lithofacies associations and were classified as (1) channel bound sequence and (2) fan bound sequence.
The channel bound sequences are 2.5 to 27 m thick and up to ~50 m wide, multistoried units. Laterally, these units exhibit lensoidal geometry and fining upward trends that internally consist of clast-supported, horizontally stratified gravels (Gh), massive gravels (Gcm), planar cross-bedded gravels (Gp), and planar cross-laminated sand (Sp) lithofacies. Generally, in a vertical sequence several such units separated by a lens of horizontally laminated sand facies (Sp) is common. At places, these units are overlain or underlain by matrix-supported graded gravel (lithofacies Gmg).
The fan bound sequences are 5 to 30 m thick and comprise matrix-supported graded gravels (Gmg), matrix-supported massive gravels (Gmm), horizontally laminated sand (Sh), and planar cross-bedded sand (Sp) lithofacies. Individual units had a fining upward trend, often tabular and laterally extending over several tens to hundreds of meters. In places, they are capped by horizontal or planar cross-bedded sandy facies. The constituent gravels are angular and poorly sorted and have a single provenance (either of Ladakh Batholith or Indus Molasse).
Morphostratigraphy and OSL chronology
Morphostratigraphically, the upper Indus River could be divided into two segments: (1) from Nyoma to Spituk, characterized by varying channel gradient, a level of fill terrace (T-1), and alluvial fans; and (2) Spituk to Dah, characterized by persistence of T-1 and one or two strath terraces. In the following terrace configuration, sedimentary profiles and OSL chronology of different sections are described.
The OSL measurements indicated that most samples had poor luminescence sensitivity, were well bleached, and showed typical quartz shine down and luminescence versus dose growth curves. Preheat and dose recovery tests yielded signal stability between 200°C and 240°C. As the recuperation ratio was lowest at ~220°C, a preheat of 220°C was used. Table 3 provides OSL ages with relevant data.
Segment: Nyoma to Spituk
In this segment, 10 sections (Mahe, Niornis, Kiari, Gaik, Tirido, Hymia, Upshi, Kharu, Stakna, and Spituk) were studied to understand the sedimentary profile of fill terrace (Fig. 2E). At Mahe, where the river cuts through the Indus Molasse, a 10.5-m-thick valley fill comprised the basal 5.5 m of channel bound sediment overlain by a 5-m-thick, poorly sorted matrix-supported debris flow unit with ~90% of the constituent gravels sourced from the Indus Molasse (Fig. 3A). Two samples from this section yielded OSL ages of 41±2 ka (LD-1047) and 14±2 ka (LD-1433; Fig. 3A).
Figure 3 (color online) The Google Earth image showing the studied sections Mahe-L and Mahe-R (A) and Niornis (B). Lower panel shows the morphostratigraphy and luminescence chronology of these sections. m amsl, meters above mean sea level; m arl, meters above river level.
The section at Niornis, ~50 km downstream from Mahe, had a ~70-m-thick valley fill comprised of 20 fining upward cycles of channel bound sequence. A sample at ~40 m above the base yielded an OSL age of 28±4 ka (LD-1048; Fig. 3B). Approximately 8 km farther downstream at Kiari, an 18-m-thick fill comprised the basal 5 m of alluvial fan bound debris flows mainly derived from Ladakh Batholith. These were overlain by a 13-m-thick unit of channel bound sequence of mixed lithology. A sample 12 m below the top yielded an OSL age of 26±4 ka (LD-1063). From Gaik to Hymia, for ~18 km (Fig. 4), the river flows into a narrow gorge across granites of Ladakh Batholith. At Gaik, a 42-m-thick fill sequence comprised the basal 24 m of channel bound sequence of mixed lithology followed by 11 m of alluvial fan bound sequence with source in Ladakh Batholith and a 7-m-thick unit of channel bound sequence (Fig. 4A). Two samples at 20 (basal channel bound unit) and 7 m below the surface (top of the fan sequence) yielded OSL ages of 13±1 ka (LD-1064) and 15±1 ka (LD-1065).
Figure 4 (color online) The Google Earth image showing the studied sections Gaik and Tirido (A) and Hymia (B). Lower panel shows the morphostratigraphy and luminescence chronology of these sections. m amsl, meters above mean sea level; m arl, meters above river level.
Approximately 9 km downstream at Tirido, a 24.5-m-thick section comprised two units of channel bound sequence of mixed lithology, and two samples at 16 and 0.5 m depth yielded OSL ages of 29±3 ka (LD-1066) and 26±2 ka (LD-1067), respectively (Fig. 4A). At Hymia, ~14 km farther downstream from Tirido (Fig. 4B), a 14-m-thick section comprising the basal 2.5 m channel bound sequence was overlain by a 2.5-m-thick unit of alluvial fan bound sequence and ~9 m of channel bound sequence. Samples from the base and another at a depth of 7 m have ages of 23±1 ka (LD-1069) and 24±1 ka (LD-1068).
Downstream from Hymia, the low gradient channel is flanked by (1) steeply dipping alluvial fans emerging from the Indus Molasse (Zanskar ranges; Drew, Reference Drew1873), (2) low gradient amphitheaters of deglaciated valleys draining the Ladakh Batholith, and (3) valley fill terraces (Fig. 5A). Toward the toe of alluvial fans, a flatter terrace, T-1, rests on the modern riverbank. A 38-m-thick section at Upshi (Upshi-R; Fig. 5B) comprised the basal ~30 m channel bound sequence of mixed lithology with an OSL age of 37±3 ka (LD-1070; Fig. 5B). An 8-m-thick alluvial fan bound unit sourced from the Ladakh Batholith overlay this. At Upshi-L, a 26-m-thick section (Fig. 5C) on the southern bank of the river comprised the basal 15-m-thick, mixed sourced channel bound sequence whose top was OSL dated to 30±2 ka (LD-1046; Fig. 5C).This was overlain by an 11-m-thick unit of molasse sourced alluvial fan deposit. The upper alluvial fan event was traceable to 12 km downstream at Kharu where it made an ~30-m-thick section. The top of this section had an OSL age of 33±2 ka (LD-1045; Fig. 5D).
Figure 5 (A) Geomorphology of segment 3 around Leh (Upshi to Spituk). Note the alluvial fans sourced from Indus Molasse and amphitheaters from Ladakh Batholith. The red line with triangles shows the traverse of Upshi-Bazgo Thrust. Valley cross section, morphostratigraphy, and lithologs of fill terraces of studied sections at Upshi-R (B), Upshi-L (C), Kharu (D), Stakna (E), and Spituk (F). m amsl, meters above mean sea level; m arl, meters above river level. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Similarly, at Stakna (Fig. 5E), valley fill terrace sections comprised three depositional units (Figs. 5E). The basal 4 m of this unit comprised alluvial fan bound sediments sourced from molasse rocks. This had an OSL age of 47±1 ka (LD-1015). This unit was overlain by a 3-m-thick channel bound unit made up of gravels of mixed lithology with an OSL age of 28±1 ka (LD-1016). A 5 m debris flow unit above this was OSL dated to 25±1 ka (Stakna, LD-1044; Fig. 5E).
At Stakna-1, a 12-m-thick section comprised the basal 1 m channel bound sequence overlain by an ~11-m-thick alluvial fan bound sequence sourced from the batholith. These had OSL ages of 29±1 ka (LD-1017) and 30±1 ka (Stakna-1, LD-1043; Fig. 5E).
The third section, ~5 km downstream of Stakna (Stakna-2), on the northern bank, represents a relict alluvial fan prograding from the Ladakh Batholith. This had a 49-m-thick sequence of channel bound, partially weathered, pale-yellow granitic gravels and comprised two depositional events. The basal event was dated to 46±4 ka (Stakna-2, LD-986; Fig. 5E), and the upper to 31±1 ka (LD-985).
At the western Leh at Spituk (34°07.13′N, 77°30.66′E), an ~45-m-thick section was exposed on the northern bank of the river valley (Fig. 5F). The lower ~10 m comprised channel bound sediments with intervening, ~1-m-thick lensoidal bodies of massive, brown coarse sand sourced from the Indus Molasse. This was overlain by 30-m-thick lacustrine sediments followed by ~5 m of eolian sand toward the top. A sample from the bottom gray sand unit yielded an OSL age of 52±4 ka (LD-1003; Fig. 5F), and that from the topmost eolian unit yielded an OSL age of 20±2 ka (Kumar et al., Reference Kumar, Srivastava and Meena2016). Similarly, a 36.5-m-thick section located on the left bank at Spituk had ~30 m of alluvial fan deposit at its base, also sourced from the Indus Molasse. The sequence comprised more than 20 units of matrix-supported angular clast lithofacies and was overlain by 2.8-m-thick lacustrine deposits.
Segment: Nimu to Dah
This ~120 km stretch of the Indus River traverses a gorge and exhibits strath and valley fill terraces (see Fig. 6A; Table 4). At Nimu, at the confluence of the Indus and Zanskar Rivers (34°09.81′N, 77°20.15′E), a valley fill terrace (T-1) and a strath terrace (T-2) are present (Fig. 6B). The fill terrace comprised an ~12-m-thick channel bound sequence and the strath terrace T-2 at 124±6.2 m arl (Fig. 6C). This was overlain by a 12.5-m-thick sequence of channel bound sediments. The strath terrace exhibited entrenched meander topography, and a sample from the youngest meander yielded an OSL age of 55±6 ka (LD-1221; Fig. 6C). Blöthe et al. (Reference Blöthe, Munack, Korup, Fülling, Garzanti, Resentini and Kubik2014) dated an older meander of this terrace to 206 ka using cosmogenic 10Be. The OSL age implies a bedrock incision rate of 2.2±0.4 mm/a.
Figure 6 (A) Google Earth image showing segment 4 of the Indus River from Nimu to Dah. The studied sections are marked by yellow stars. The contact between Indus Molasse and Ladakh Batholith is marked by a red line with triangles. The yellow rectangles are the insets for Figures 6B, 7, and 8. (B) Terraces between Nimu and Saspol. Note one level of fill terrace T-1 (yellow) and strath terraces T-2 (brown). The Indus River flowing in the narrow gorge carved into Indus Molasse. Geomorphology at Nimu and Saspol sections. Note the incised meander at Nimu section and sampling locations of Blöthe et al. (Reference Blöthe, Munack, Korup, Fülling, Garzanti, Resentini and Kubik2014) and this work. Valley cross section and morphostratigraphy and lithologs of fills at Nimu (C) and Saspol (D) indicating one level of filled and one level of strath terrace. m amsl, meters above mean sea level; m arl, meters above river level. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
About 29 km downstream from Nimu, at Saspol (34°14.84′N, 77°6.46′E) (Fig. 6D, Table 4), the river cuts through the Indus Molasse and two terraces, T-1 and T-2. T-1 comprised 19-m-thick channel bound sediments, and a sample at 10 m below the surface gave an OSL age of 17±2 ka (T-1, LD-1000; Fig. 6D). Strath terrace T-2 was at 81.2±4 m arl and was covered by 17 m of channel bound sediments. A sample at a depth of 11.5 m yielded an OSL age of 50±5 ka (T-2; Fig. 6D). Terrace T-2 can be traced downstream for ~1 km, yielding a bedrock incision rate of 1.6±0.2 mm/a.
At Nurla, ~10 km downstream of Saspol, a former meander of the Indus River is present on the north bank of the river (Fig. 7A). The strath terrace relating to this meander was at 148±7.4 m arl and had a 20-m-thick alluvial cover. A sample near the bottom gave an OSL age of 78±5 ka (LD-989; Fig. 7B). The estimated bedrock incision rate here was 1.9±0.2 mm/a. Approximately 20 km further downstream at Khalsi, the river continues to flow through the Indus Molasse and has three terraces: T-1, T-2’, and T-2 (Fig. 7C, Table 4). Terrace T-1 comprised a 31.2-m-thick channel bound sequence, and a sample near the top yielded an age of 18±1 ka (T-1, LD-998; Fig. 7C). Strath terrace T-2’, at an elevation of 50.2±2.5 m arl, had a 7-m-thick sediment and dated to 41±4 ka (T-2’, LD-996; Fig. 7C). The uppermost terrace, T-2, at 158.8±8 m arl had an ~20-m-thick cover of channel bound sequence, which was OSL dated to 52±5 ka (T-2, LD-997; Fig. 7C and D). The bedrock incision rate from T-2 was 3.0±0.4 mm/a and from T-2’ was 1.2±0.2 mm/a.
Figure 7 (color online) (A) Google Earth image showing the fill and strath terraces from Nurla to Khalsi. Note the presence of paleomeander of Indus and the corresponding strath surface at Nurla. Morphostratigraphy and chronology of Nurla (B) and Khalsi (C). (D) Field photograph of section at Khalsi showing one level of fill and two levels of strath terraces. m amsl, meters above mean sea level; m arl, meters above river level.
The stretch between Dumkhar and Skyurbachan Gompa has a fill terrace T-1 and a strath terrace T-2 (Fig. 8A). At Dumkhar (Fig. 8B), a valley fill terrace (T-1) comprised a 33.3-m-thick channel bound sediment followed by a strath terrace (T-2) at an elevation of 132.5±6.6 m (arl) overlain by 18-m-thick channel bound gravel (Fig. 8B). A sample from the fluvial gravel of terrace T-2 yielded an OSL age of 57±3 ka (LD-995) and provided a bedrock incision rate of 2.3±0.2 mm/a. This terrace configuration continues for another 6.5 km, and at Skyurbuchan Gompa terrace T-2 was at 142.5±7 m (arl) and overlain by 15.5 m of fluvial gravels (Fig. 8C, Table 4). The base of the overlying gravels yielded an OSL age of 56±7 ka (LD-988) and thereby provided an bedrock incision rate of 2.5±0.4 mm/a. Similarly, ~5 km downstream of Skyurbachan Gompa, a 25-m-thick, scree-covered fill terrace (T-1) and two strath terraces (T-2’ and T-2) at elevations of 40±2 m (arl) and 150±7.5 (arl) were present. The alluvial cover of T-2’ and T-2 were OSL dated to 47±7 (LD-993) and 83±7 ka (LD-987; Fig. 8D) and yielded bedrock incision rates of 0.85±0.2 and 1.8±0.2 mm/a, respectively.
Figure 8 (color online) (A) Google Earth map showing the strath terrace located on the hanging block of Upshi-Bazgo Thrust. Valley cross section, terrace configuration, and lithologs and chronology at Dumkhar, Skyurbuchan Gompa, and Skyurbuchan Downstream sections by elevation (B) and height (C). m amsl, meters above mean sea level; m arl, meters above river level.
Downstream from there, the river flows through the Ladakh Batholith and carves a deep and narrow gorge with only one level of fill terrace. A sediment fill of 17 m thickness at Biamah and 15 m at Dah, ~35 km farther downstream, was noted. A debris flow sample from the top of the sequence at Biamah yielded an OSL age of 7±1 ka (LD-991; Table 1). The strath height, ages of the alluvial cover, and estimated bedrock incision rates are collated in the Table 4.
DISCUSSION
Sedimentation pattern and paleoclimate
Sedimentation pattern and aggradation in a river valley depends on the relative proportions of water and sediment in the channel network, vegetation cover, and the system state with respect to its being supply or transport limited (Blum and Törnqvist, Reference Blum and Törnqvist2000; Srivastava et al., Reference Srivastava, Tripathi, Islam and Jaiswal2008). The sediment load of a channel depends on contemporary climate and relief, driven by tectonics. Most of the stretch from Mahe to Dah exhibited widespread valley fills in the form of terrace T-1 and alluvial fan sequences. The stratigraphy of terrace T-1 and alluvial covers of strath terraces suggest that channel bound and alluvial fan bound debris flows controlled the aggradation in the valley. Lithofacies composition (Gh, Gp) of channel bound deposits suggests that the traction current-mediated processes resulted in the aggradation of channel bars. Internally, these bars comprised large-scale planar two-dimensional bed forms suggestive of sustained flood conditions with broad hydrograph, characteristic of wetter conditions (Miall, Reference Miall1996; Suresh et al., Reference Suresh, Bagati, Kumar and Thakur2007; Shukla, Reference Shukla2009; Ray and Srivastava, Reference Ray and Srivastava2010). Fan bound deposits represented tributary or outwash alluvial fan aggradation via sheet type, nonchannelized floods originating from the glaciers or from flash floods (Srivastava et al., Reference Srivastava, Ray, Phartiyal and Sharma2013). Intercalations of these two types of sequences suggest deposition under subhumid-paraglacial climates where the trunk channel flooded regularly and the channel bars aggraded vertically and episodically because of high sediment supply from glacier-fed tributaries that brought sediments via flash floods (Miall, Reference Miall1996). Nearly 50% of the cumulative thickness of the sequences were channel bound gravels of subfacies C and comprised homogenized multiprovenance sediments. The remaining 50% was of a single provenance (subfacies A and B). This indicates that, although regular floods in the Indus River that fetch sediment from all over the catchment were responsible for aggradation, the sediment supplied by tributaries via episodic discharge was of equal importance.
The Spituk lake sequence on the alluvial fan emanating from molasses suggests that (1) this sequence postdated the alluvial fans, and (2) the paleolake formed by progradation of fans, both from the Ladakh Batholith and the Indus Molasse. Detailed sedimentary architecture of this sequence and the surrounding area presented by Sinclair et al. (Reference Sinclair, Mudd, Dingle, Hobley, Robinson and Walcott2017) indicates that neotectonically forced northeastward movement of the Indus River controlled the sedimentary architecture of the valley at Spituk, and the Stok Thrust played an active role.
The present work suggests that the sediments at Spituk were deposited between 52 and >20 ka. Sinclair et al. (Reference Sinclair, Mudd, Dingle, Hobley, Robinson and Walcott2017) OSL dated the upper part of the Spituk section to 27±3 ka, which falls within this age bracket. Previously, infrared-stimulated luminescence ages of 177–72 ka (Blöthe et al., Reference Blöthe, Munack, Korup, Fülling, Garzanti, Resentini and Kubik2014) and 14C–accelerator mass spectrometry ages of 10–1.5 ka (Phartiyal et al., Reference Phartiyal, Sharma and Kothari2013) were considered unreliable because of the possibility of incomplete bleaching of feldspar in lacustrine environments and contamination of carbon by meteroric waters.
Understanding the aggradation record in terms of climate requires analysis of the contemporary climates of the region. The records suggest the following: (1) five glacial advances in the Leh region since 430 ka (Owen et al., Reference Owen, Finkel, Barnard, Haizhou, Asahi, Caffee and Derbyshire2005; Ali and Juyal, Reference Ali and Juyal2013; Dortch et al., Reference Dortch, Owen and Caffee2013); (2) stronger monsoon conditions during 60–30 ka and weaker monsoon conditions during 25–18 ka as evidenced from the Guliya Ice Core, located ~400 km northeast (NE) of the study area (Fig. 9A; Thompson et al., Reference Thompson, Yao, Davis, Henderson, Mosley-Thompson, Lin, Beer, Synal, Cole-Dai and Bolzan1997); and (3) penetration of the SW monsoon to up to >75 km beyond the orographic barrier during 14–9 ka and 50–30 ka, as well as its role in assisting hill slope erosion, accentuating landslides, and enhancing the sediment budget of rivers (Bookhagen et al., Reference Bookhagen, Thiede and Strecker2005, Dortch et al., Reference Dortch, Owen, Haneberg, Caffee, Dietsch and Kamp2009; Srivastava et al., Reference Srivastava, Ray, Phartiyal and Sharma2013). This is evidenced in the lacustrine and glacial records from Ladakh and Tibet that indicate postglacial warming culminating at the Holocene climatic optimum (Prell and Kutzbach, Reference Prell and Kutzbach1987; Fang, Reference Fang1991; Shi et al., Reference Shi, Yu, Liu, Li and Yao2001; Owen et al., Reference Owen, Caffee, Bovard, Finkel and Sharma2006a, Reference Owen, Caffee, Finkel and Seong2008; Jung et al., Reference Jung, Kroon, Ganssen, Peeters and Ganeshram2009, Reference Jung, Kroon, Ganssen, Peeters and Ganeshram2010; Wünnemann et al., Reference Wünnemann, Demske, Tarasov, Kotlia, Reinhardt, Bloemendal and Diekmann2010; Dortch et al., Reference Dortch, Owen and Caffee2013; Owen and Dortch, Reference Owen and Dortch2014; Rawat et al., Reference Rawat, Gupta, Sangode, Srivastava and Nainwal2015a, Reference Rawat, Gupta, Srivastava, Sangode and Nainwal2015b; Fig. 9B).
Figure 9 (color online) Valley aggradation in the Indus River and paleoclimate record. (A) Climate record from the Guliya Ice Core (after Thompson et al., Reference Thompson, Yao, Davis, Henderson, Mosley-Thompson, Lin, Beer, Synal, Cole-Dai and Bolzan1997) and Arabian Sea NIOP905 87KYr Benthic Stable Isotope Data (Jung et al., Reference Jung, Kroon, Ganssen, Peeters and Ganeshram2010). (B) Paleoglacial record of Ladakh. (C) Age distribution of the channel bound and outwash fan sediment. (D) Frequency distribution of all the ages. (E) Major phases of aggradation shown by prominent river in northwest (NW), Nepal, and northeast (NE) Himalaya. Note that widespread aggradation of channel took place during the Marine Oxygen Isotope Stages (MIS) 1 and MIS 3, and sedimentation of alluvial fans in Ladakh took place during the deglaciation of MIS 3 only.
A comparison of OSL ages from fluvial fill (related to both T-1 and T-2), outwash fans, and debris flows with past climate records indicates that valley alluviation occurred in three pulses at 52, 28, and ~16 ka (Fig. 9C and D). Aggradation on fans occurred from ~47 to 29 ka, and three debris flow events occurred at ~27 ka. All these were suggested periods of stronger monsoon conditions. Records from the Zanskar River suggest valley aggradation between 50 and 20 ka (Blöthe et al., Reference Blöthe, Munack, Korup, Fülling, Garzanti, Resentini and Kubik2014). Thus, the phases of higher rainfall during 60–30 ka and 10–5 ka mobilized the sediment from the catchment into the channel and assisted aggradation. It is noteworthy that river aggradation chronologies of rivers across the Himalaya—namely, Spiti, Sutlej, and Alaknanda (NW Himalaya); Marsyandi (central Himalaya, Nepal); and Brahmaputra (NE Himalaya)—that derive moisture largely from the SW Indian monsoon also experienced valley aggradation during 60–30 ka and ~10–5 ka (Pratt-Sitaula et al., Reference Pratt-Sitaula, Burbank, Heimsath and Ojha2004; Srivastava et al., Reference Srivastava, Bhakuni, Luirei and Misra2009, Reference Srivastava, Ray, Phartiyal and Sharma2013; Juyal et al., Reference Juyal, Sundriyal, Rana, Chaudhary and Singhvi2010; Ray and Srivastava, Reference Ray and Srivastava2010; Sharma et al., Reference Sharma, Bartarya and Marh2016a).
In the NW Himalaya, two rainfall-induced phases of increased landslides are reported during 37–29 and 15–3 ka (Dortch et al., Reference Dortch, Owen, Haneberg, Caffee, Dietsch and Kamp2009). Rivers in northern Tibet in the Kunlun Mountains also built large outwash fans during ~30 ka (Owen et al., Reference Owen, Finkel, Haizhou and Barnard2006b). Therefore, we conclude that on a regional scale the rivers flowing in the drier part of the Himalaya aggraded during the phases of stronger monsoon conditions. Further, when the catchment had a higher vegetation cover that reduced the sediment to water ratio, incision occurred. This in general was true for all the rivers along the Himalayan arc (Ray and Srivastava, Reference Ray and Srivastava2010; Scherler et al., Reference Scherler, Bookhagen, Wulf, Preusser and Strecker2015; Sinclair et al., Reference Sinclair, Mudd, Dingle, Hobley, Robinson and Walcott2017).
Channel gradient indices, strath terrace formation, and incision rates
Ideally, the channel gradient indices exhibit higher values for regions with molasse (sandstone and shale) and lower values for regions of granitic batholith (Troiani et al., Reference Troiani, Galve, Piacentini, Seta and Guerrero2014). Our results, however, do not show this trend (Fig. 2). K s values show a similar trend (i.e., higher for regions with granite and lower for regions with molasse) (Munack et al., Reference Munack, Korup, Resentini, Limonta, Garzanti, Blöthe, Scherler, Wittmann and Kubik2014). These suggest that for the Indus River the channel steepness was controlled by tectonic uplifts and the bedrock erodibility played a subordinate role. Channel steepness for nonglaciated catchments bordering the Indus River and in the southeast (SE) Tibetan plateau front provides a similar inference (Kirby et al., Reference Kirby, Whipple, Tang and Chen2003; Sinclair et al., Reference Sinclair, Mudd, Dingle, Hobley, Robinson and Walcott2017).
In segment 2, a steep channel gradient and high SL index indicate the river’s adjustment to active tectonics and uplift in its headwater region. Similar observations accrue from SE Tibet, where longitudinal river profiles were steep and adjusted to differential uplift between Tibet and its surrounding (Kirby et al., Reference Kirby, Whipple, Tang and Chen2003). Varied evidence based on sedimentology and geomorphology has suggested recent uplifts in southern Tibet (Liu, Reference Liu1981; Li and Zhou, Reference Li and Zhou2001), and these lead to vertical incision. 10Be-based modern erosion rates above and below the knick point at Mahe indicate strong headward erosion and incision by the Indus (Munack et al., Reference Munack, Korup, Resentini, Limonta, Garzanti, Blöthe, Scherler, Wittmann and Kubik2014). This also suggests that the incision in segment 2 was partly driven by the elevation of the Tibetan plateau.
OSL chronology in segment 3 indicates that the alluvial fan aggradation and river valley filling occurred during 47–29 ka. Therefore, excess sediment delivery from the Zanskar ranges that were additionally pushing the river northward led to a lower SL index. The Stok, Upshi-Bazgo, and Choksti Thrusts located along and south of the Indus River, respectively, were active, and tectonic activity along these thrusts aided by the high relief of the Zanskar ranges led to excess sediment supply. 10Be-derived denudation rates of 69.8 mm/ka in Zanskar bound rivers are high as compared with 29 mm/ka of those draining the batholith suggesting that the Indus Molasse sequence is actively deforming and rising (Searle et al., Reference Searle, Pickering and Cooper1990; Sinclair and Jaffey, Reference Sinclair and Jaffey2001; Munack et al., Reference Munack, Korup, Resentini, Limonta, Garzanti, Blöthe, Scherler, Wittmann and Kubik2014; Sinclair et al., Reference Sinclair, Mudd, Dingle, Hobley, Robinson and Walcott2017). High sediment supply and aggradation in this segment helped lower the gradient and the SL index of the channel.
In segment 4, chronologies of the alluvial cover preserved over the strath terraces indicate incision rates of ~0.8–3.0 mm/a. The height and chronology of strath terraces indicated two levels of former riverbed profiles of the Indus (Fig. 10A and B). The upper profile is at an average elevation of 134±24 m arl and has an OSL age of 62±16 ka, and these imply an average incision rate of 2.2±0.9 mm/a. The lower profile, at 45±5 m arl with an OSL age of 44±8 ka (Fig. 8D), implies an erosion rate of 1.0±0.3 mm/a. Interpolation of these profiles upstream in the present river profile indicates that: (1) the lower profile truncates into the fill sequences preserved in segment 3 (Leh valley) as both bear similar ages; and (2) the upper profile is older, and sediment of equivalent age might be present in the subsurface in segment 3 and upstream. The sequences above the present riverbed, upstream from Nimu, do not have any sediment that is equal to, or older in age than 62±16 ka. This suggests that a divergence of both lower and upper profiles occurred because of a fall in the base level in the downstream region (Pazzaglia et al., Reference Pazzaglia, Gardner and Merritts1998; Crosby and Whipple, Reference Crosby and Whipple2006; Wobus et al., Reference Wobus, Whipple, Kirby, Snyder, Johnson, Spyropolou, Crosby and Sheehan2006). Bedrock incision rates deduced in this study and rock uplift rates in the region may not correlate. Investigations in Marsyandi River, Nepal, show that while the uplift continues, the alluvial fill protects the riverbed from incision, and later after the removal of the fill, the river incises at rates faster than the mean (Pratt-Sitaula et al., Reference Pratt-Sitaula, Burbank, Heimsath and Ojha2004). We consider this to be the same in the study area.
Figure 10 (color online) (A) Longitudinal river profile of the Indus River with strath terraces. Note the downstream divergent nature of the strath profile. (B) Longitudinal profile of the Indus showing reconstruction of two levels of paleoriverbeds and ages of the strath terraces. (C) Comparison of bedrock incision rates in the Nanga Parbat–Harmosh Massif region of the western syntaxes (after Leland et al., Reference Leland, Reid, Burbank, Finkel and Cafee1998) and the incision rates deduced in this study. Note the systematic downstream increase in the incision rates. m amsl, meters above mean sea level.
Bedrock incision rates and base level fall at Nanga Parbat–Harmosh Massif (NPHM)
In NPHM of the western syntaxes, two sets of strath terraces along the Indus are dated using terrestrial cosmogenic nuclides to ~7 ka and between 67 and 27 ka (Burbank et al., Reference Burbank, Leland, Fielding, Anderson, Brozovic, Reid and Duncan1996; Leland et al., Reference Leland, Reid, Burbank, Finkel and Cafee1998) yielding bedrock incision rates of 9–12 mm/a during the Holocene and 1–6 mm/a during pre–last glacial maximum. Modern incision rates of the Indus in NPHM are ~12 mm/a, and these reduce to 3–6 mm/a ~50 km upstream at Skardu. Postglacial incision rates of Shigar, a tributary of Braldu River that flows into the Indus near Skardu, are 2–29 mm/a (Seong et al., Reference Seong, Owen, Bishop, Bush, Clendon, Copland, Finkel, Kamp and Shroder2008). Therefore, the present study and the published data suggest that the incision rates along the Indus River increase sharply from 0.8–3.0 to ~12 mm/a as it flows from Nimu to NPHM in the western syntaxes (Fig. 10C). 10Be-based bulk erosion rates along the Indus also show an increase in NW direction that reaches its maxima at the western syntaxes (Garzanti et al., Reference Garzanti, Vezzoli, Andó, Paparella and Clift2005; Van Der Beek et al., Reference Van Der Beek, Melle, Guillot, Pêcher, Reiners, Nicolescu and Latif2009; Ali and de Boer, Reference Ali and de Boer2010; Munack et al., Reference Munack, Korup, Resentini, Limonta, Garzanti, Blöthe, Scherler, Wittmann and Kubik2014). This suggests that erosion and downcutting by the Indus River were controlled by the base level fall at NPHM.
Increase in the channel slope and bedrock incision in the downstream reaches can produce a knick point that moves upstream and that in the long term can induce incision as far as 300 km upstream (Gardner, Reference Gardner1983). In the case of the Indus River, the presence of a wide basin at Skardu that lies between NPHM and Nimu can possibly hinder the upstream movement of the knick point. However, in the Skardu basin and NPHM, the channel gradient are 3.3 and 10.8 m/km, respectively. These are higher than that 1.8 m/km at Leh. These, therefore, indicate progressive steepening of the river valley in the downstream reaches. Structural and geomorphic analysis suggests that incision and aggradation in the Skardu basin itself was controlled by the far-field effect of rapid uplift in the NPHM zone and north–south compression along the Himalayan arc (Cronin, Reference Cronin1989). Thus, incision rates determined by earlier studies and those in the present study provide insight into the response of tectonics and the far-field effects of rapid uplift in the NPHM of the western syntaxes and formation of strath terraces in segment 4.
Distribution of strath terrace T-2 along the longitudinal profile shows initiation of the knick points in the most downstream part at ~83 ka (Skyurbuchan Downstream section; Fig. 8D) and that moved in the upstream, such that terrace at the same level yields younger ages of ~50 ka. The age of ~78 ka for the strath terrace, T-2, at Nurla section is an exception and still needs a proper explanation through structural mapping of the area.
Formation of strath terraces in segment 4, between Nimu and upstream of Dah, can be understood in two different scenarios as well. The first scenario considers that in segment 4 river flows through the Indus Molasse sequence, which is deformed and comprises several north-verging thrusts like the Choksti, Stok, and Upshi-Bazgo Thrusts. Chronological data on strath terraces indicate that these thrusts are neotectonically active and the river has responded to uplift along these thrusts. OSL ages of 42 ka for fluvial terraces and the geomorphic setting around Leh indicate neotectonic deformation of the Indus Molasse along the Stok Thrust and that a shortening of ~10 m controlled the landscape development (Sinclair et al., Reference Sinclair, Mudd, Dingle, Hobley, Robinson and Walcott2017). Further evidence of tectonic movement accrues from the following: (1) steepness index in the hanging wall of the Stok Thrust (Sinclair et al., Reference Sinclair, Mudd, Dingle, Hobley, Robinson and Walcott2017), the basin morphometry, and hypsometric indices of tributaries draining into the Indus and Shyok Rivers between Leh and Khalsi (zone B of Jamieson et al., Reference Jamieson, Sinclair, Kirstein and Purves2004); (2) increasing channel sinuosity with river downcutting in the form of entrenched meanders at Nimu (Gardner, Reference Gardner1975; Schumm et al., Reference Schumm, Mosley and Weaver1987; Rogers et al., Reference Rogers, Kárason and van der Hilst2002); and (3) records of seismites along the Indus River at Spituk, Saspol, and Lamayuru, as well as other surrounding areas that point toward past seismic activity in the region (Bagati et al., Reference Bagati, Mazari and Rajagopalan1996; Singh and Jain, Reference Singh and Jain2007; Phartiyal and Sharma, Reference Phartiyal and Sharma2009).
An alternative scenario is where fills overlying the strath terraces in the Nimu–Dah segment and those in the Leh valley are coeval and represent a single phase of valley alluviation in the Indus as suggested earlier (Blöthe et al. Reference Blöthe, Munack, Korup, Fülling, Garzanti, Resentini and Kubik2014). Excess sediment brought down by the Zanskar River potentially choked the gorge and dammed the Indus near Spituk, and a lake formed that existed for at least 100 ka (Blöthe et al. Reference Blöthe, Munack, Korup, Fülling, Garzanti, Resentini and Kubik2014). This suggests that the ~35 m Spituk lake sequence took >100 ka to aggrade; implying an average sedimentation rate of 0.35 mm/yr, which is not realistic. The varve and sedimentary rythmites in the Spituk sequence that are supposed to be annual sedimentary layers are less than a centimeter in thickness indicating a much higher sedimentation rate (3–4 mm/yr; Nag et al., Reference Nag, Phartiyal and Singh2016). The chronological data published from the paleolake sequences from across the Himalaya also seem to suggest much higher sedimentation rates (Fort, Reference Fort2000; Korup et al., Reference Korup, Strom and Weidinger2006; Phartiyal et al., Reference Phartiyal, Sharma, Srivastava and Ray2009; Anoop et al., Reference Anoop, Prasad, Krishnan, Naumann and Dulski2013; Srivastava et al., Reference Srivastava, Ray, Phartiyal and Sharma2013; Nag and Phartiya, 2015).
CONCLUSION
The upper Indus River that we studied exhibits a complex interaction among the channel aggradation, progradation of alluvial fans, landslides, and high sediment supply from tributaries that caused local damming and transient filling of the river. Several possible explanations for valley filling, lake formation, and strath chronology in the Indus River exist, but the more preferred explanation is the climate-driven valley aggradation and incision attributable to tectonics and the fall in NPHM base level.
The valley filling was controlled by bar aggradation associated with channel and alluvial fan progradation. Aggradation occurred in three pulses at ~52, ~28, and ~16 ka; aggradation on alluvial fans took place from ~47 to 29 ka. These were during the well-documented periods of strengthened SW monsoon conditions.
Strath terraces in the Nimu–Dah segment suggest two past levels of the Indus. The upper profile at an elevation of 134±24 m arl has an age of 62±16 ka yielding an average erosion rate of 2.2±0.9 mm/a. The lower profile at 45±5 m arl has an age of 44±8 ka and an average erosion rate of 1.0±0.3 mm/a.
Incision rates from NPHM suggest that the strath development in segment 4 was because of a far-field effect of uplift and rapid incision in the western syntaxes.
Active tectonic deformation of the Indus Molasse during the past ~50 ka led to the incision of the river and influenced the sedimentary architecture of fills.
Results of our study pose an important question as to what degree do tectonics influence landscape evolution along the upper Indus valley.
Table 3 Sample location, dosimetry, paleodose, dose rate, and ages of all the samples collected from the Indus River.
Table 2 Lithofacies, subfacies, description, and interpretation.
Table 4 Terrace configuration (e.g., fill or strath type), thickness of fill and strath above the river level (arl), chronology, and bedrock incision rates as studied along the Indus River.
ACKNOWLEDGMENTS
Professor Anil K. Gupta, the director of the Wadia Institute of Himalayan Geology, is thanked for his support. Comments and annotations provided by Drs. Jason Dortch and Craig Dietsch and two anonymous reviewers helped in improving the manuscript significantly. Prof. Lewis Owen helped immensely in improving the manuscript and all the figures. Drs. K. Morell, Oliver Korup, Jan Blöthe, and Rasmus Thiede are thanked for helpful discussions. The personnel at the state works department at Khalsi are thanked for their help with total station mapping. Measurement of the Nimu section was done during the fieldwork sponsored by Department of Science and Technology, New Delhi, via grant project no. SR/FTP/ES-41/2012. This forms part of AK’s PhD dissertation.
Supplementary materials
To view supplementary material for this article, please visit https://doi.org/qua.2017.19
