1. Introduction
The Siwalik peripheral foreland basin of the Himalaya was formed during Neogene time as a result of India–Eurasia continent–continent collision, driven downward by flexure of the Indian plate (Dewey & Bird, Reference Dewey and Bird1970; Lyon-Caen & Molnar, Reference Lyon-Caen and Molnar1985). It longitudinally extended from the Potwar Plateau of Pakistan in the west to Arunachal Pradesh of India in the east. A thick (c. 7000 m) fluvial sequence (the Siwalik Group) of mudstone, sandstone and conglomerate was deposited in this basin from Early Miocene time until Pleistocene time (Tandon, Reference Tandon, Tandon, Pant and Casshyap1991), when it broke into two unequal parts along the Himalayan Frontal Thrust (HFT) (Valdiya, Reference Valdiya1998). The northern narrower part uplifted to form the Sub-Himalayan ranges (more commonly called the Siwalik Ranges) and the southern wider part became the Indo-Gangetic foreland basin (Fig. 1) (Valdiya, Reference Valdiya1998).
Figure 1. Generalized map showing thrust-bounded major lithotectonic terranes of the Himalaya. The Siwalik foreland basin sediments are exposed in the Sub-Himalaya. ITSZ – Indus Tsangpo Suture Zone; STD – South Tibetan detachment; MCT – Main Central Thrust; MBT – Main Boundary Thrust; HFT – Himalayan Frontal Thrust; IGB – Indo-Gangetic foreland basin; R. – River. Stars mark the locations of studied Lower Siwalik sectors. 1 – Chinji (southern Potwar Plateau); 2 – Khaur (northern Potwar Plateau); 3 – Jammu; 4 – south-central Uttarakhand; 5 – southeastern Uttarakhand (present study area); 6 – Surai Khola.
The Siwalik Group is subdivided into Lower, Middle and Upper Siwalik subgroups of Early to early Late Miocene, early Late Miocene to latest Late Miocene, and latest Late Miocene to Pleistocene ages, respectively (Pilgrim, Reference Pilgrim1910; Tandon, Reference Tandon, Tandon, Pant and Casshyap1991). Its palaeontology, sedimentology, stratigraphy and structural geology have been determined along many sectors (for a synthesis, see Tandon, Reference Tandon, Tandon, Pant and Casshyap1991; Burbank, Beck & Mulder, Reference Burbank, Beck, Mulder, Yin and Harrison1996; Sinha et al. Reference Sinha, Kumar, Sinha, Tandon and Gibling2007), but investigations are yet to be initiated along several other sectors. Moreover, most of the sedimentological investigations have been focused on the depositional environment and provenance of the sediments (e.g. Parkash, Bajpai & Saxena, Reference Parkash, Bajpai and Saxena1974; Tandon, Reference Tandon1976; Critelli & Ingersoll, Reference Critelli and Ingersoll1994; Garzanti, Critelli & Ingersoll, Reference Garzanti, Critelli and Ingersoll1996; Zaleha, Reference Zaleha1997; Nakayama & Ulak, Reference Nakayama and Ulak1999; Brozovic & Burbank, Reference Brozovic and Burbank2000; Friend et al. Reference Friend, Raza, Geehan and Sheikh2001; Sharma, Sharma & Singh, Reference Sharma, Sharma and Singh2001; Kumar et al. Reference Kumar, Ghosh, Mazari and Sangode2003; Huyghe et al. Reference Huyghe, Mugnier, Gajurel and Delcaillau2005; Najman, Reference Najman2006; Shukla, Bora & Singh, Reference Shukla, Bora and Singh2009). On the other hand, the channel morphology and hydrology of the rivers of the Siwalik basin have been determined only along a few sectors in northern Pakistan (Locations 1 and 2 in Fig. 1) (Willis, Reference Willis1993; Zaleha, Reference Zaleha1997), northern India (Location 4 in Fig. 1) (Shukla, Bora & Singh, Reference Shukla, Bora and Singh2009; Khan & Tewari, Reference Khan and Tewari2011) and western Nepal (Location 6 in Fig. 1) (Nakayama & Ulak, Reference Nakayama and Ulak1999; Ulak, Reference Ulak2005). These efforts, however, are insufficient to bring out a regional picture of the Siwalik fluvial systems and landscape. Detailed sedimentological investigations, therefore, are needed to be carried out along many other sectors to have a clear regional picture of the Siwalik fluvial systems including their channel patterns, morphologies, flow patterns and hydrology.
Bearing this in mind, we conducted the present study to obtain detailed information on channel pattern, channel morphology, flow pattern, hydrology and geomorphic positioning of a Lower Siwalik river system in the southeastern Uttarakhand state of India (Location 5 in Fig. 1), which fed the basin during 12.5 to 11.0 Ma (Kotlia et al. Reference Kotlia, Phartiyal, Kosaka and Bora2008). No sedimentological investigations have so far been conducted in this part of the Siwalik, mainly because of the dense forest cover and lack of easy access as well as scarcity of continuously exposed sections. The present study, thus, provides the first detailed sedimentological record of this part of the Lower Siwalik foreland basin.
We have determined the channel patterns through detailed lithofacies analysis and flow pattern through statistical analysis of the palaeocurrent data, and computed morphological and hydrological parameters of the channels in terms of depth, width, width/depth ratio, sinuosity, meander wavelength, meander bend width, radius of meander curvature and slope, along with mean annual discharge, mean bankfull discharge and silt-clay load percentage using empirical equations developed by various workers and tested on many ancient and modern rivers (Leopold & Maddock, Reference Leopold and Maddock1953; Leopold & Wolman, Reference Leopold and Wolman1960; Schumm, Reference Schumm1963, Reference Schumm, Rigby and Hamblin1972; Allen, Reference Allen1968, Reference Allen1970; Cotter, Reference Cotter1971; Miall, Reference Miall1976; Ethridge & Schumm, Reference Ethridge, Schumm and Miall1978; Gardner, Reference Gardner1983; Willis, Reference Willis1993; Tewari, Reference Tewari2005; Gibling, Reference Gibling2006; Shukla, Bora & Singh, Reference Shukla, Bora and Singh2009; Khan & Tewari, Reference Khan and Tewari2011; Tewari, Hota & Maejima, Reference Tewari, Hota and Maejima2012). Palaeocurrent directions and thicknesses of 168 cross-beds provided the basic data for computing these morphological and hydrological parameters. Given the analogy of climate, location, landscape and sedimentary processes between the Indo-Gangetic and Siwalik foreland basins (see Jain & Sinha, Reference Jain and Sinha2003 and references therein), the geomorphic positions of these streams within the foreland basin have been determined by comparing their sediment characters, palaeocurrent patterns, morphologies and hydrological parameters with those of the modern rivers of the Ganga (Gangetic) foreland basin. Furthermore, we have also made an attempt to provide a glimpse of the regional fluvial landscape of the Siwalik foreland basin during Middle Miocene time by comparing the morphological, flow and hydrological parameters of these streams with those of the Potwar Plateau (Locations 1 and 2 in Fig. 1) (Willis, Reference Willis1993; Zaleha, Reference Zaleha1997), Jammu (Location 3 in Fig. 1) (Sharma, Sharma & Singh, Reference Sharma, Sharma and Singh2001) south-central Uttarakhand (Location 4 in Fig. 1) (Shukla, Bora & Singh, Reference Shukla, Bora and Singh2009) and western Nepal (Location 6 in Fig. 1) (Nakayama & Ulak, Reference Nakayama and Ulak1999; Ulak, Reference Ulak2005) (Fig. 1).
The present data on channel pattern, morphology, flow pattern, hydrology and geomorphic positioning of the streams of this part of the Lower Siwalik foreland basin will contribute towards the understanding of the regional fluvial landscape of the Siwalik foreland basin during Middle Miocene time. The data will also be useful in spatial and temporal comparison of similar fluvial parameters on a regional as well as global scale. Furthermore, the study provides a model for studying ancient fluvial deposits through an integrated approach of lithofacies, palaeocurrent, palaeomorphological and palaeohydrological analyses.
2. Geological setting
In the study area, the Siwalik Range rises abruptly against the vast Ganga Plain. The Main Boundary Thrust (MBT) and HFT define its northern and southern structural boundaries, respectively (Fig. 2). The HFT thrusts the Siwalik rocks over alluvium of the Indo-Gangetic foreland basin (Karunakaran & Ranga Rao, Reference Karunakaran and Ranga Rao1979), whereas the Lesser Himalayan sequence is thrust over the Siwalik sequence along the MBT (Auden, Reference Auden1934).
Figure 2. Geological map of the study area. Bedding plane dips have been added to the map of Goswami & Yhokha (Reference Goswami and Yhokha2010). MBT – Main Boundary Thrust; HFT – Himalayan Frontal Thrust; BF – Bastia Fault; TF – Tanakpur Fault; KF – Kalaunia Fault. ‘A’ marks the location of the Kirora Nala section, and ‘B’ and ‘C’ mark the locations of the Tanakpur–Champawat motor road sections.
Only the Lower and Middle Siwalik subgroups are exposed in the area (Goswami & Yhokha, Reference Goswami and Yhokha2010). The whole strata are folded into a large syncline. The Middle Siwalik is in the core, whereas the Lower Siwalik is exposed in the 25° to 80° NW-dipping southern limb and 28° to 70° SW- to SE-dipping northern limb of the syncline (Fig. 2). The Kalaunia Fault (KF), Tanakpur Fault (TF), Bastia Fault (BF) and an unnamed thrust traverse the area (Karunakaran & Ranga Rao, Reference Karunakaran and Ranga Rao1979; Goswami & Yhokha, Reference Goswami and Yhokha2010) (Fig. 2). The latter two of these trend parallel to the Himalayan strike, whereas the former two trend transverse to it and are extensions of the basement structures of the adjoining Ganga foreland basin (Goswami, Reference Goswami2012). The Lower Siwalik Subgroup, comprising purple, dark to pale brown compact mudstones interbedded with fine-grained sandstones and siltstones, was deposited during c. 12.5 to 11 Ma (Kotlia et al. Reference Kotlia, Phartiyal, Kosaka and Bora2008). In many places the sequence is dominated by mudstone and siltstone, but in some other places it is dominated by sandstones. The overlying Middle Siwalik Subgroup, comprising salt-and-pepper grey, medium- to coarse-grained sandstones and subordinate brownish grey to reddish mudstones, was deposited during c. 11.0 to 4 Ma (Kotlia et al. Reference Kotlia, Phartiyal, Kosaka and Bora2008).
Geomorphologically, the Siwalik Hills in the area rise up to ~1200 m asl (metres above mean sea level). The antecedent Sarda River and its tributaries drain the terrane. The Sarda River originates in the perennially snow-clad high Himalayan ranges, and Siwalik streams join it either in the Siwalik terrane itself or in the adjoining Ganga Plain through deeply cut ‘V’-shaped valleys. The Siwalik Range in the area has a conspicuous sigmoid shape formed owing to bending of strata caused by the continuous northward pressing by a basement spur of the Ganga–Siwalik foreland basin (Goswami, Reference Goswami2012). This spur is one of the several digitations of the Delhi–Hardwar Ridge in the basement of the foreland basin (Raiverman, Kunte & Mukherjea, Reference Raiverman, Kunte and Mukherjea1983). Interestingly, the Sarda River also has a similar sigmoid course in the area owing to bends along its channel from the SW to WNW, again to the SW and finally to the S (Fig. 2).
3. Results
3.a. Lithofacies associations
Twelve stratigraphic sections of the Lower Siwalik, ranging in thickness from 36 to 280 m, along the Kirora Nala (also called Hathi Khor) stream and Tanakpur–Champawat motor road have been studied and measured for the lithofacies analysis (Fig. 2). A total of eight lithofacies have been identified in the area, which show marked lateral and vertical variations in characters. Specific sequences of two or more of these lithofacies constitute three distinct lithofacies associations, which have been named as ‘Facies Association A’, ‘Facies Association B’ and ‘Facies Association C’. The typical sedimentological logs of these three lithofacies associations, compiled from all the measured sections, are shown in Figure 3. Each one of these associations represents a specific depositional setting, albeit one or more of the constituent lithofacies of an association is generally missing in many sections. A brief description and interpretation of all the three lithofacies associations is given below.
Figure 3. Sedimentological logs showing the (a) sandstone-dominated ‘Facies Association A’, (b) sandstone–mudstone-dominated ‘Facies Association B’, and (c) mudstone-dominated ‘Facies Association C’ identified in the Lower Siwalik sequence of the study area.
3.a.1. Facies Association A
3.a.1.a. Description
Exposed only along the Kirora Nala section, this association begins with 30 cm to 70 cm thick units of the pebbly sandstone lithofacies. Having an erosional lower contact, this lithofacies comprises coarse- to very coarse-grained, poorly sorted pebbly sandstone, which is sandy conglomerate in some places (Fig. 4a). The granule- to pebble-sized clasts of the lithofacies are generally rounded. The larger clasts, however, are dominantly intrabasinal calcrete and mud. Sometimes parallel laminations and trough cross-beds are faintly preserved in the lithofacies. This lithofacies grades upwards into the cross-bedded sandstone lithofacies, which comprises coarse- to fine-grained, fining-upward, generally multistoreyed units having solitary or cosets of up to 82 cm thick trough, planar and low-angle cross-beds (Fig. 4b). The cross-beds generally amalgamate and indicate SSE- and WSW-directed prominent palaeocurrent directions, but in different locations. The individual units of the lithofacies are up to 3 m thick. Other sedimentary structures of the lithofacies include scour-and-fill, reactivation surfaces, ripple bedding and parallel laminations (Fig. 4c). The lithofacies grades upwards into a ripple-laminated sandy siltstone lithofacies, which comprises 1.8 to 2.5 m thick, often mottled, sandy siltstone units having centimetre- to decimetre-thick, discontinuous mudstone capping. The sedimentary structures of the lithofacies include small-scale trough cross-laminations, lenticular bedding, climbing-ripple cross-laminations, convolute laminations, load structures, parallel laminations and desiccation cracks (Figs 4d, 5a). Sometimes, a bioturbated mudstone lithofacies is developed on top of this lithofacies. Having a sharp lower contact, it comprises up to 1.5 m thick, grey to greyish green or purple, massive, bioturbated mudstone units, often alternating with siltstone units having similar characters. The sequence of lithofacies renders an overall fining-upwards character to this facies association.
Figure 4. Sedimentary structures of the ‘Facies Association A’. (a) Lower part of the pebbly sandstone lithofacies showing alignment of clasts along parallel laminations. Length of the hammer is 26 cm. (b) Trough cross-beds in the cross-bedded sandstone lithofacies. Length of the pen in 13 cm. (c) Parallel laminations in the cross-bedded sandstone lithofacies. Length of the hammer is 26 cm. (d) Climbing-ripple cross-laminations in the ripple-laminated sandy siltstone lithofacies. Length of the pen is 13.5 cm.
3.a.1.b. Interpretation
The erosional lower boundary, fining-upwards character, predominance of sand bodies, characters of constituent lithofacies and amalgamated large-scale cross-beds together suggest that this facies association represents the deposits of a sand-dominated braided river system (cf. Rust, Reference Rust and Miall1978; Walker & Cant, Reference Walker, Cant and Walker1984; Smith et al. Reference Smith, Ashworth, Best, Woodwords and Simpson2006). The basal pebbly sandstone lithofacies represents channel lag deposits emplaced by strong traction currents (Postma, Reference Postma, Colella and David1990). The overlying cross-bedded lithofacies was deposited in the lower bar environment owing to migration of sinuous and straight crested, 3D and 2D dunes and low-amplitude bedforms forming trough, planar and low-angle cross-beds, respectively (Coleman, Reference Coleman1969; Miall, Reference Miall and Miall1978). The ripple-laminated sandy siltstone lithofacies of the association was deposited in a subaerially exposed upper bar environment under decelerating flow conditions, but with some rapid fluctuations in the water level and sediment supply (Coleman, Reference Coleman1969; Reineck & Singh, Reference Reineck and Singh1980; Allen, Reference Allen1982; Bristow, Reference Bristow, Best and Bristow1993). The bioturbated mudstone lithofacies was deposited mainly through suspension fallout in channels or floodplains (Miall, Reference Miall and Miall1978; Bridge, Reference Bridge1984, Reference Bridge1993).
3.a.2. Facies Association B
3.a.2.a. Description
This facies association begins with 1.2 cm to 2.4 m thick, fining-upwards units of cross-bedded sandstone lithofacies comprising coarse- to fine-grained, often multistoreyed units having solitary or cosets of up to 72 cm thick, trough and planar cross-beds (Fig. 5b). About 42% of the cross-beds indicate a prominent S30°W to W-directed palaeocurrent, whereas 25% of them indicate a S to S60°E-directed palaeocurrent. Other sedimentary structures of the lithofacies include scour-and-fill, reactivation surfaces, ripple bedding and parallel laminations. Sharply overlying this lithofacies is either the bioturbated mudstone lithofacies or the fine sandstone–mudstone alternation lithofacies. The bioturbated mudstone lithofacies comprises 1.3 to 2 m thick units of grey, green, purple and chocolate brown, massive, mottled and bioturbated mudstone units, generally with a silt-dominated lower part. The units in places are finely laminated (Fig. 5c). Ferruginous and calcrete nodules are common. The units also contain plant fossils, rootlets and trace fossils (Fig. 5d). The fine sandstone–mudstone alternation lithofacies comprises 80 cm to 1.3 m thick prismatic or wavy, fine sandstone to silty sandstone units alternating with 2 to 3.5 m thick mudstone units. The sandy units have up to 6 cm thick trough and planar cross-laminations, ripple bedding, parallel laminations and ripple marks (Fig. 6a). The mudstone units are nodular and mottled. The lithofacies is bioturbated, contains ubiquitous plant fossils, coal chips/chunks and rootlets. Embedded within this lithofacies is the lensoid to prismatic sandstone lithofacies, which comprises medium- to fine-grained sandstone units having a maximum thickness of up to 2 m and a significantly variable width from a few metres to tens of metres. The lithofacies has an erosional lower contact. It contains parallel laminations, climbing-ripple cross-laminations and sometimes small-scale cross-laminations and ripple marks (Fig. 6b). The smaller units, however, are generally massive.
Figure 5. (a) Convolute lamination in the ripple-laminated sandy siltstone lithofacies of ‘Facies Association A’. Diameter of the coin is 2.4 cm. (b) Planar cross-bed in the cross-bedded sandstone lithofacies of ‘Facies Association B’. Length of the hammer is 26 cm. (c) Bioturbated, massive to parallel laminated buff coloured mudstone with thin, light coloured silty bands of the bioturbated mudstone lithofacies of ‘Facies Association B’. Diameter of the coin is 2.4 cm (d) Burrow tubes (indicated by arrows) in the bioturbated mudstone lithofacies of ‘Facies Association B’. Length of the hammer is 26 cm.
Figure 6. (a) Trough cross-lamination in the sandstone of the fine sandstone–mudstone alternation lithofacies of ‘Facies Association B’. Length of the pen lid is 5 cm. (b) Parallel laminations in the lensoid to prismatic sandstone lithofacies of ‘Facies Association B’. Length of the pen 14 cm. (c) The mottled siltstone–mudstone heterolith lithofacies of the ‘Facies Association C’. Length of the hammer is 26 cm. (d) Lensoid units of the fine sandstone lithofacies of the ‘Facies Association C’. Bedded calcrete of the mottled siltstone–mudstone heterolith lithofacies is also marked by arrows in the lower part of the photograph. Length of the hammer is 26 cm.
3.a.2.b. Interpretation
A fining-upwards character with a dominance of sandstone in the lower part and mudstone in the upper part, along with the characters of the constituent lithofacies suggest the deposition of this lithofacies in a meandering river system (cf. Walker & Cant, Reference Walker, Cant and Walker1984). The multistoreyed units of the cross-bedded sandstone lithofacies were deposited through lateral accretion of migrating 3D and 2D dunes in point bars (Miall, Reference Miall and Miall1978; Tewari & Gaur, Reference Tewari and Gaur1991; Bridge, Reference Bridge1993; Shukla & Singh, Reference Shukla and Singh2004). The overlying bioturbated mudstone lithofacies was deposited through suspension fallout in channels or flood plains and subsequently modified by pedogenic processes (Miall, Reference Miall and Miall1978; Bridge, Reference Bridge1984, Reference Bridge1993). The fine sandstone–mudstone alternation lithofacies was deposited by overbank flows during floods on natural levees, which supported much plant growth and is where pedogenesis is common (Reineck & Singh, Reference Reineck and Singh1980; Shukla & Singh, Reference Shukla and Singh2004). The lensoid to prismatic sandstone lithofacies within this lithofacies was deposited by traction currents in crevasses and splays (Reineck & Singh, Reference Reineck and Singh1980).
3.a.3. Facies Association C
3.a.3.a. Description
This facies association comprises mottled siltstone–mudstone heterolith lithofacies and fine sandstone lithofacies. However, the latter is less common and occurs as lenses within the former. Several cycles of this association are generally stacked together to form 20 to 36 m thick multistoreyed complexes. Covering vast areas, this facies association has sharp contacts with the other two facies associations, but it is seldom associated with their channel sand bodies. In many places, this association shows evidence of soil formation, but this has not caused any homogenization and the original lithofacies boundaries are still prominent. The mottled siltstone–mudstone heterolith lithofacies comprises red, maroon and reddish brown, nodular, mottled, variegated and bioturbated mudstone and siltstone alternations. The clay and silt proportions are highly variable laterally as well as vertically. However, in a few places the mottling is not so pervasive (Fig. 6c). The siltstone sometimes is finely laminated. The lithofacies is characterized by the presence of bedded calcrete (up to 35 cm thick), concentrated calcrete nodule horizons, disseminated calcrete and ferruginous nodules, coal stringers and chunks (Fig. 6d). The lensoid units of the fine sandstone lithofacies have a maximum thickness of 1.5 m (Fig. 6d). The units are mostly solitary, have an erosional to sharp base and become bioturbated siltstone upwards, which sometimes contains calcrete spherules. Sedimentary structures of the lithofacies include poorly preserved small-scale planar and trough cross-laminations and climbing-ripple cross-laminations.
3.a.3.b. Interpretation
The prevalence of thick muddy units, extensive calcretization, ferruginization and oxidation of sediments, and dissociation from any major fluvial channel sand bodies indicate that this facies association represents deposition in the upland interfluve regions (also known as Doab), which form vast raised areas further away from major river channels (Singh et al. Reference Singh, Srivastava, Sharma, Sharma, Singh, Rajagopalan and Shukla1999; Sharma, Sharma & Singh, Reference Sharma, Sharma and Singh2001). The modern sediment fill of the extensive upland interfluve areas of the Ganga basin show almost similar facies associations (cf. Kumar et al. Reference Kumar, Singh, Singh and Singh1995; Singh et al. Reference Singh, Srivastava, Sharma, Sharma, Singh, Rajagopalan and Shukla1999). The mottled siltstone–mudstone heterolith lithofacies was deposited on higher sloping surfaces, lower flat surfaces and in lakes/ponds of the upland interfluve (Kumar et al. Reference Kumar, Singh, Singh and Singh1995; Singh et al. Reference Singh, Srivastava, Sharma, Sharma, Singh, Rajagopalan and Shukla1999), whereas the fine sandstone lithofacies was deposited in creeks and small channels that commonly develop in upland interfluves and drain water only after rains (Singh et al. Reference Singh, Srivastava, Sharma, Sharma, Singh, Rajagopalan and Shukla1999). Similar facies associations of the Lower Siwaliks of the south-central Uttarakhand and Jammu regions have also been interpreted to represent the deposits of the upland interfluve regions (Singh et al. Reference Singh, Srivastava, Sharma, Sharma, Singh, Rajagopalan and Shukla1999; Sharma, Sharma & Singh, Reference Sharma, Sharma and Singh2001; Shukla, Bora & Singh, Reference Shukla, Bora and Singh2009).
3.b. Palaeocurrent analysis
As mentioned in Section 1, the present palaeocurrent analysis is based on measurements of 168 trough, planar and low-angle cross-beds as they are more reliable palaeocurrent indicators (Dott, Reference Dott1973). Cross-laminations that have a set thickness of < 5 cm are not included in these 168 readings. Separate analyses have been done for the braided and meandering streams. The cross-beds of braided and meandering streams have been sorted on the basis of detailed facies analysis. Out of the total 168 cross-beds, 130 belong to the bars of the braided stream and the remaining 38 belong to the bars of the meandering stream.
Given that the bedding plane dips in the area are invariably > 25°, the effects of tectonic tilt on palaeocurrent data have been removed following the method of Potter & Pettijohn (Reference Potter and Pettijohn1963). The palaeocurrent directions were grouped in 30° interval classes and rose diagrams were prepared. Data dispersions for the braided and meandering streams have been determined by calculating the vector mean (
$\bar \theta$
v), vector magnitude (L) and circular standard deviations (S) as proposed by Davis (Reference Davis2002) (Table 1). The uncertainty of the vector mean is tested at the 95% confidence interval following Batschelet (Reference Batschelet1981) (Table 1). The chi-square (χ2) test has also been applied to ascertain the uniformity of palaeocurrent distribution (Table 1).
Table 1. Formulae used for computing the morphological, hydrological and palaeocurrent parameters of the Lower Siwalik streams
Where, H – cross-bed set thickness; c – compaction of sand during burial, which varies from 1.1 to 1.3 (Ethridge & Schumm, Reference Ethridge, Schumm and Miall1978), in present calculations a value of 1.3 is taken following Gardner (Reference Gardner1983); c’ – variation in depth between straight and sinuous reaches, which varies from 0.585 to 1.0 (Ethridge & Schumm, Reference Ethridge, Schumm and Miall1978), in the present calculations a value of 1 is taken following Ethridge & Schumm (Reference Ethridge, Schumm and Miall1978); Dbf – bankfull channel depth; Wbf – bankfull channel width; F – width/depth ratio; g – force due to gravity; n – number of observations; π – a constant whose value is 3.14; k – concentration value corresponding to r value, which can be found in any statistics book; Oj – observed number of data in class j; Ej – expected number of data in class j (j varies from 1 to 12 for a 30° class interval).
The palaeocurrent data of the braided stream show a bimodal distribution with 47% of the total measurements indicating a southeasterly direction and 35% indicating a northwesterly direction (Fig. 7a). The remaining measurements indicate palaeocurrent directions in other azimuth classes but with an insignificant population of < 5% in each class, and, thus, are excluded from the rose diagram to enhance its clarity. The vector mean of all measurements is N204°, but with a smaller vector magnitude of 29%, a high standard deviation of 68° and a high 95% confidence interval value of ±24°. The low vector strength, and high values of the standard deviation and 95% confidence interval suggest sediment dispersal by multi-directional currents in multiple channels of the braided system. However, the distinct bimodal dispersion of sediments is considered to indicate a bend in the river channel (Fig. 8). Such bends in foreland basin river channels could be topographically or tectonically controlled, e.g. along several river courses of the Ganga basin (Raiverman, Kunte & Mukherjea, Reference Raiverman, Kunte and Mukherjea1983; Jain & Sinha, Reference Jain and Sinha2005; Goswami, Reference Goswami2012). The dispersions for both these two modal classes have also been determined separately. The southeasterly directed modes have a N150° vector mean with a fairly high vector magnitude of 97%, and the northwesterly directed modes have a N298° vector mean, again with a fairly high vector magnitude of 91%. These two modal classes have standard deviations of 11° and 23°, respectively and 95% confidence interval values of ±3° and ±7°, respectively. The lower values of the standard deviation and 95% confidence interval for these two modal classes indicate that the vector means of these two modal classes are statistically significant. Moreover, the fairly high calculated chi-square (χ2) probability value of 209 (much higher than the critical value of 31.26 at 99.9% confidence level) reaffirms that the palaeocurrent data shows two significant preferential directions.
Figure 7. Palaeocurrent roses of the (a) braided and (b) meandering streams of the study area. Bold arrows depict the vector means of the whole data, whereas dashed arrows depict the vector means of the two distinct modal classes: southeastern and northwestern for the braided stream and southeastern and southwestern for the meandering stream.
Figure 8. Schematic diagram depicting conceptual fluvial landscape of the Lower Siwalik foreland basin. The streams of the study area were positioned in medial-distal megafan (MF), interfan (IF) and upland interfluve (UI) setting. MBT – Main Boundary Thrust.
The meandering river shows a radial outward regional palaeocurrent pattern, swinging between ESE and WSW, but dominantly towards the southeast and southwest (Fig. 7b). The vector mean of all the measurements is N210° with a low vector magnitude of 57%, high standard deviation of 53° and high 95% confidence interval value of ±21°. The low vector strength, and high values of the standard deviation and 95% confidence interval suggest sediment dispersal by multi-directional currents, which might be related to avulsive, shifting channels on a megafan surface (cf. Leier, DeCelles & Pelletier, Reference Leier, Decelles and Pelletier2005). Nevertheless, the channels had prominent southeasterly and southwesterly flow directions, having a N139° vector mean with 89.00 vector magnitude and a N245° vector mean with 89.00 vector magnitude, respectively. The standard deviations of 26° and 25°, respectively, and 95% confidence interval values of ±14° and ±11°, respectively, indicate that that the calculated vector means for these two classes are more or less statistically significant. Moreover, the calculated chi-square (χ2) probability value of 35.36 (> critical value of 31.26 at the 99.9% confidence level) also shows that the palaeocurrent had significant preferential directions.
3.c. Channel morphology and hydrology
The empirical equations used to compute various morphological and hydrological parameters of the Siwalik streams are given in Table 1. The basic input for these calculations is the cross-bed thickness. It directly gives mean and bankfull channel depths (Dc and Dbf) (Allen, Reference Allen1968; Ethridge & Schumm, Reference Ethridge, Schumm and Miall1978; Gardner, Reference Gardner1983). This is a relatively robust estimate of channel depths for ancient deposits (Bridge & Tye, Reference Bridge and Tye2000). Moreover, in the presence of good rock exposure, such as the present study area, this method yields results encouragingly consistent with more direct estimators such as point-bar thickness (Leclair & Bridge, Reference Leclair and Bridge2001). The estimated channel depth is then used as an input to compute channel width (Wbf) (Allen, Reference Allen1968, Reference Allen1970). These two parameters are then inputs for all other morphological and hydrological parameters (Table 1).
The parameters have been computed separately for the braided and meandering streams. Again, only the 168, >5 cm thick, cross-beds have been taken for these calculations. The cross-bed thicknesses for the braided and meandering streams in the area range from 7 to 82 cm and 6 to 72 cm, respectively. However, about 50% of the former population is 16 to 30 cm thick, and about 55% of the latter population is 11 to 30 cm thick.
The average values of the calculated parameters are given in Table 2. It is clear from the palaeocurrent analysis that the overall surface gradient in the area was south-southwesterly. However, the low-sinuosity meandering stream had a gentler channel slope than the braided stream. Both the braided and meandering streams had a tranquil and low-flow regime at Froude numbers of 0.08 and 0.07, respectively, which account for the extensive development of cross-beds in the sandy units. The mean silt-clay percentage of <5%, along with a channel width/depth ratio of >40 suggest a bed-load nature for both the streams (cf. Schumm, Reference Schumm1968). Both the streams had variable hydrological parameters. The bankfull depth and width of the braided streams varied from 0.5 to 4.15 m and 31 to 305 m, respectively, whereas those of the meandering streams varied from 0.46 to 3.71 m and 18 to 180 m, respectively. The mean annual discharge of the braided and meandering streams varied from 2.30 to 515 m3s−1 and 1.02 to 264 m3 s−1, respectively, while bankfull discharge varied from 227 to 7045 m3 s−1 and 124 to 4070 m3 s−1, respectively.
Table 2. Average values of estimated morphological and hydrological parameters of the Lower Siwalik streams of the study area
4. Discussion
The vast Lower Siwalik foreland basin was fed in different sectors by braided, meandering and anastomosing streams of different dimensions (Willis, Reference Willis1993; Zaleha, Reference Zaleha1997; Friend et al. Reference Friend, Raza, Geehan and Sheikh2001; Sharma, Sharma & Singh, Reference Sharma, Sharma and Singh2001; Shukla, Bora & Singh, Reference Shukla, Bora and Singh2009). In the Khaur area of the northern Potwar Plateau, the fluvial network comprised single channel meandering and braided streams (Zaleha, Reference Zaleha1997), while the Chinji area of the southern Potwar Plateau had a main, braided channel belt and a large number of smaller channel belts, wherein the braided channel was constrained by terraces (upland) but the smaller channels migrated in response to avulsive floods (Friend et al. Reference Friend, Raza, Geehan and Sheikh2001). In the Jammu area, the main channel was shallow and meandering (Singh et al. Reference Singh, Srivastava, Sharma, Sharma, Singh, Rajagopalan and Shukla1999; Sharma, Sharma & Singh, Reference Sharma, Sharma and Singh2001). In south-central Uttarakhand, there was a well-developed network of braided, meandering and anastomosing rivers (Shukla, Bora & Singh, Reference Shukla, Bora and Singh2009), and just 55 km eastward, in the present area, the river network consisted of shallow, braided and low-sinuosity meandering streams. Further east, in western Nepal, it was a network of meandering streams (Nakayama & Ulak, Reference Nakayama and Ulak1999; Huyghe et al. Reference Huyghe, Mugnier, Gajurel and Delcaillau2005).
In the Khaur and Chinji areas of the Potwar Plateau and south-central Uttarakhand the Lower Siwalik rivers were located in the distal megafan areas (Zaleha, Reference Zaleha1997; Friend et al. Reference Friend, Raza, Geehan and Sheikh2001; Shukla, Bora & Singh, Reference Shukla, Bora and Singh2009). The radial outward drainage pattern of shifting, smaller, meandering streams along with dominantly southeastward-flowing main braided channels having very gentle channel slopes (< 21 cm km−1 or ~0.01°) and dominant sand size of fine to medium suggest that the Lower Siwalik rivers of the present area were also located in the medial-distal megafan areas (Fig. 8). The braided stream was the main feeder channel of the megafan, like that of the Chinji area. The NW to SSE bend along its course indicates strong autogenic forcing on sediment dispersal. The bend might have been controlled by some active topographic high in the basement, possibly the Tanakpur–Kasganj basement spur (Raiverman, Kunte & Mukherjea, Reference Raiverman, Kunte and Mukherjea1983) that is also known to control the present-day sigmoid course of the Sarda River as well as the similar shape of the mountain front in the area (Goswami, Reference Goswami2012). Tandon & Kumar (Reference Tandon and Kumar1984) have also demonstrated that such active basement highs caused drainage reversal during Upper Siwalik times. Moreover, similar bends along modern river courses of the Ganga foreland basin have also been considered to be controlled by active basement structures (e.g. Raiverman, Kunte & Mukherjea, Reference Raiverman, Kunte and Mukherjea1983; Jain & Sinha, Reference Jain and Sinha2005; Goswami, Reference Goswami2012).
A comparison between the morphological and hydrological parameters of the Lower Siwalik channels of northern Pakistan (Willis, Reference Willis1993; Zaleha, Reference Zaleha1997), western Nepal (Ulak, Reference Ulak2005), south-central Uttarakhand (Shukla, Bora & Singh, Reference Shukla, Bora and Singh2009) and the present study area is given in Table 3. The data is broadly comparable. The channels were < 400 m wide and < 5 m deep and generally had low sinuosity, except for the meandering stream of south-central Uttarakhand. The bankfull discharge varied spatially, being least in the Khaur area of the northern Potwar Plateau. The braided streams of the present area and south-central Uttarakhand had more or less the same dimensions and similar discharge. However, the meandering streams of the present area were of lower sinuosity, wider and shallower than in the neighbouring south-central Uttarakhand. These spatial variations in dimensions, shape and hydrological characters reflect mainly the spatial variations in the sediment and water budget (cf. Schumm, Reference Schumm1968), which are primarily controlled by the geographic location and dimensions of the catchment area.
Table 3. Computed morphological and hydrological parameters of the Lower Siwalik streams of the study area and other sectors
*Calculated using the data of Shukla, Bora & Singh (Reference Shukla, Bora and Singh2009); ** Calculated using the data of Willis (Reference Willis1993); ***calculated using the data of Zaleha (Reference Zaleha1997).
A comparison between the morphological, hydrological and flow characters of the Lower Siwalik rivers of the study area and the present-day, fine- to medium-grained sand-carrying rivers of the Ganga foreland basin (Table 4) reveals that the Lower Siwalik braided stream of the area is broadly comparable with the large modern rivers like the Kosi and Yamuna (width/depth ratios of 70 to 78) despite the greater channel width of the Kosi River. On the other hand, the Lower Siwalik meandering river of the area is comparable with the interfan rivers like the Bagmati, Burhi Gandak and Kamla–Balan (width/depth ratios of 12 to 44). However, it is interesting to note that the Kosi megafan in northern Bihar is presently fed by the mountain-fed, braided Kosi River and many plains-fed, meandering rivers (cf. Sinha & Friend, Reference Sinha and Friend1994), which exhibit an overall radial outward palaeocurrent pattern (Chakraborty et al. Reference Chakraborty, Kar, Ghosh and Basu2010). The meandering streams of the megafan are 6 to 150 m wide and have sinuosities of 1.2 to 2.5 (Chakraborty et al. Reference Chakraborty, Kar, Ghosh and Basu2010). Such a comparison is quite tenable owing to the similarity of the location, landscape and sedimentary processes between the Siwalik and Ganga foreland basins (Burbank & Beck, Reference Burbank and Beck1991; Willis, Reference Willis1993; Zaleha, Reference Zaleha1997) and comparability between the Middle Miocene and present-day tropical to sub-tropical and monsoonal climatic conditions of the Indian subcontinent (Raymo & Ruddiman, Reference Raymo and Ruddiman1992; Nakayama & Ulak, Reference Nakayama and Ulak1999). It is therefore inferred that the Lower Siwalik fluvial network of the area was somewhat like that of the present-day medial-distal megafan-interfan setting of the Ganga basin (Fig. 8). The main channel was braided and larger in dimensions with higher water discharge, whereas the low-sinuosity, avulsive meandering channel might have originated on the megafan analogous to those of the Kosi megafan (Fig. 8). Nevertheless, the presence of interfan rivers is also very likely. In addition, there were smaller and shallower interfluve (Doab) streams, which mainly redistributed the older alluvium (Fig. 8) (cf. Singh et al. Reference Singh, Srivastava, Sharma, Sharma, Singh, Rajagopalan and Shukla1999). Interestingly, the Lower Siwalik streams of the south-central Uttarakhand were also located in the medial-distal megafan-interfan setting (Shukla, Bora & Singh, Reference Shukla, Bora and Singh2009).
Table 4. Morphological and hydrological parameters of the modern rivers of the Ganga foreland basin
5. Conclusions
The present study has provided detailed information about a hitherto unstudied Middle Miocene river system of the Lower Siwalik foreland basin, which will be very useful in reconstructing the regional-scale landscape and understanding the fluvial systems of the Lower Siwalik foreland basin. The salient features of this river system can be enumerated as follows:
(i) The river system consisted of a network of shallow, bed-load braided and meandering rivers along with a few small seasonal streams.
(ii) The larger streams were located in the medial-distal megafan-interfan areas whereas the smaller seasonal streams were redistributing the upland interfluve sediments.
(iii) The braided stream had a bend induced by an active basement spur, similar to one identified by Tandon & Kumar (Reference Tandon and Kumar1984) in the basement of the Upper Siwalik foreland basin of the Chandigarh area.
(iv) The morphological and hydrological parameters of these rivers are comparable with the Lower Siwalik rivers of northern Pakistan, south-central Uttarakhand and western Nepal as well as some modern rivers of the Ganga basin.
(v) A comparison of sediment size, channel pattern, flow pattern, channel morphology and hydrology reveals that the medial-distal part of the Kosi megafan and adjoining interfan area provides a comparable modern analogue for the Lower Siwalik sedimentation in the study area.
Acknowledgements
We are grateful to our revered teacher Prof. Charu C. Pant, Dean, Faculty of Science of our university, for encouraging us to undertake this study, several fruitful discussions and constructive suggestions. We thank Prof. Santosh Kumar, the Head of our Department, for providing working facilities and financial assistance under SAP (CAS) of the UGC, New Delhi in our Department. Tanuja is thankful to UGC, New Delhi for awarding an RFMS research fellowship. We gratefully acknowledge the encouraging and constructive comments by learned reviewers.
