Morphometric properties of the trans-Himalayan river catchments: Clues towards a relative chronology of orogen-wide drainage integration Parthasarathi Ghosh, Sayan Sinha, Arindam Misra PII: DOI: Reference:
S0169-555X(14)00539-X doi: 10.1016/j.geomorph.2014.10.035 GEOMOR 4970
To appear in:
Geomorphology
Received date: Revised date: Accepted date:
7 April 2014 14 October 2014 23 October 2014
Please cite this article as: Ghosh, Parthasarathi, Sinha, Sayan, Misra, Arindam, Morphometric properties of the trans-Himalayan river catchments: Clues towards a relative chronology of orogen-wide drainage integration, Geomorphology (2014), doi: 10.1016/j.geomorph.2014.10.035
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Morphometric properties of the trans-Himalayan river
Parthasarathi Ghosh*, Sayan Sinha, Arindam Misra
SC R
IP
wide drainage integration
T
catchments: clues towards a relative chronology of orogen-
MA
* corresponding author:
[email protected]
NU
Geological Studies Unit, Indian Statistical Institute, 203 B. T. Road, Kolkata 700108, India
D
Abstract
TE
The geomorphological evolution of the Himalayan mountain belt both in terms of crustal deformation and concomitant erosion by surface processes has been suggested to have a profound
CE P
influence on a number of earth system processes and has been extensively researched through a number of different techniques. The huge catchments of the trans-Himalayan rivers are the product
AC
of long-term fluvial erosion of the landscape. This work attempts to understand their evolution through a study of drainage network, morphology, and internal organization of the smaller watersheds nested within each catchment.
Using morphometric techniques applied to an orogen-wide digital elevation data grid, we characterized the drainage network structure and catchment of all the 18 trans-Himalayan rivers situated between exists of the Indus and Brahmaputra rivers and constructed rectangular approximations of the catchment geometries. With the help of catchment dimensions measured transverse and parallel to the strike of the orogen, and by analysing the dimension and spatial dispositions of the rectangular approximations, we demonstrate that the trans-Himalayan catchment shapes cannot be explained only as a product of the headward enlargement of drainage 1
ACCEPTED MANUSCRIPT networks on a topographic slope, or orogenic taper. Within individual catchments we identified the existence of orogen-transverse drainage components (watersheds) that are organized in a
T
systematic manner with respect to the first-order physiographic features of the Himalayas that were
IP
formed at different periods of geological time. They each show distinct morphometric characteristics
SC R
that are indicative of differences in processes and / or time scale involved in their formation. The hypsometric properties of the watersheds occupying the upper part of the catchments suggest that they are the remnants of pre-orogenic drainage that became confined to the leeward side of the
NU
Himalayas before the advent of monsoon circulation. The shape and organization of the transverse watersheds occurring in the middle of the catchments resembles a series of small drainage basins
MA
formed on the precursor topography of the modern Himalayas. The lower parts of the catchments were shaped instead by drainage diversions induced by deformations related to the frontal thrust.
D
We show how the shape of the catchments represents an integration of processes such as headward
TE
drainage enlargement, capture of pre-existing drainage, and diversion of drainage in response to
CE P
crustal deformation at successive stages of Himalayan mountain growth. We further show that there is a systematic change in the morphological characters and organization of the watersheds, nested in the catchments, from the middle towards the flanks of the Himalayas indicating the variations in
AC
relative influence of different drainage evolution processes and the orogen-scale heterogeneity in tectonic style.
Keywords: River catchment, Drainage network, Morphometry, Himalayas, Outlet spacing, Drainage evolution.
2
ACCEPTED MANUSCRIPT 1. Introduction Surface processes, and predominantly the action of rivers, have carved out large valley systems on
T
the Himalayan topography. A huge population of rivers, very large to small, drains the southern
IP
slope of the Himalayas and flows out into the foreland alluvial plains after crossing the southern
SC R
mountain front (Burrard and Hyden, 1907; see also Fig. 1 in Brookfield, 1998). The smaller streams cross only the lower part of the southern slope, whereas the larger rivers traverse the entire
NU
southern slope as well as small regions lying to the north of the topographic divide (Fig. 1). These rivers are termed trans-Himalayan in this work. In contrast, the two syntaxial rivers, the Indus and
MA
Brahmaputra originate in the Tibetan Himalaya. They flow parallel to the orogen behind the Himalayas for hundreds of kilometres towards the west and east, respectively (Fig. 1). Their
D
tributaries drain the region lying between the northern slope of the orogen and the southern edge
TE
of the Tibetan Plateau (Fig. 2). These two rivers turn south at the western and eastern syntaxial
CE P
bends to cut across the mountain belt and exit in the foreland plains.
The catchments of the modern trans-Himalayan rivers are very large (Table 1, Fig. 1). It would require continuous erosion for a few tens of millions years to produce erosional features of such
AC
dimensions (Shroder and Bishop, 2004). Capturing insights into the history of these catchments could, therefore, provide scope for better understanding the erosional history of this orogen on a time scale of 107 years and, in turn, for better constraining links between orogen-scale denudation and global changes. However, relying on clues contained among the complex geomorphic features of landscapes such as those of the trans-Himalayan river catchments is possible only on condition that their mode of formation can be clearly understood. Apart from a few earlier studies on the evolution of the Himalayan drainage systems (e.g., Gupta, 1997; Brookfield, 1998; Friend et al., 1999) a comprehensive understanding of the evolution of these huge geomorphic features is not available.
3
ACCEPTED MANUSCRIPT Global surveys of the drainage networks in linear orogens and footwall uplands (Hovius, 1996, Talling et al., 1997; Purdie and Brook, 2006; Walcott and Summerfield, 2009) indicate that the
T
drainage outlets of the range-scale catchments are regularly spaced along the mountain front
IP
irrespective of scale, slope, lithology, climate or tectonic setting. It has been also noted that the
SC R
outlet spacing (s) maintains a consistent relationship with the width of the orogen (w). For many orogens, the spacing ratio (w/s) takes a value of about 2 (Hovius, 1996; Walcott and Summerfield, 2009). This consistency among spacing ratios has been attributed to an expression of Hack’s Law
NU
(Hovius, 1996), which states that the length of the trunk stream maintains a power law relationship with the drainage area (Hack, 1957) during the growth of the drainage network. Despite being more
MA
or less regularly spaced, unlike other orogens the spacing ratio of outlets of the trans-Himalayan drainages has been found to be much less, about 1.7 (Hovius, 1996; Walcott and Summerfield,
D
2009). This would indicate that the catchment-evolution process, in this case, is different. Large,
TE
tectonically active and widening orogens like the Himalayas have a complex history of physiographic
CE P
evolution due to variations in tectono-climatic forcing over a millennial time scale. It is expected that, in such cases, additional factors and mechanisms will affect drainage development (Friend et al., 1999; Castelltort and Simpson, 2006; Gupta, 1997). It has been argued that drainage
AC
reorganization induced by currently ongoing tectonic deformation at the frontal part of this active orogen produces an unstable arrangement of the basins (Walcott and Summerfield, 2009), which gives rise to a computed spacing ratio smaller than 2. However, the mechanics of catchment evolution that might explain the present-day shape and organization of the trans-Himalayan basins remains to be fully investigated. It is possible that known mechanisms such as enlargement of the basins into the hinterland (Densmore et al., 2005; Perron et al., 2008), accretion of the foreland (Castelltort and Simpson, 2006), and annexation (capture) of antecedent river catchments have interacted at different stages to give rise to the modern-day configuration. However, identification of the signatures of these different classes of mechanism is problematic due to the fact that the highly dynamic morphotectonic setting has kept changing the relative positions of the divide and of 4
ACCEPTED MANUSCRIPT the range front. The long time interval over which these processes interact, and the low preservation potential of the morphological signatures in a high-denudation setting, together tend to obscure the
T
evidence (Walcott and Summerfield, 2009).
IP
In contrast to earlier methods, we demonstrate here the value of breaking down the larger
SC R
catchments into their smaller catchment components (hereafter termed watersheds in order to emphasize the distinction of scale), by showing that various sub-regions nested within the larger
NU
trans-Himalayan catchments preserve, through their morphometric signatures, a memory of evolutionary history, which has therefore only partly been erased by drainage integration and deep
MA
denudation. The shape and spatial arrangement of watersheds that are nested within the larger trans-Himalayan catchments thus provide clues about the relative chronology of catchment growth.
D
By establishing a link between distinct hypsometric watershed properties and watershed position
TE
within the orogenic belt, we show that the watersheds belonging to different sub-regions within a catchment can be correlated with different stages of Himalayan landscape evolution which are
CE P
consistent with independently established geological events. Based on a conceptual, orogen-wide catchment evolution model, we finally discuss how these distinct watershed populations could have
today.
AC
interacted during the Cenozoic in order to give rise to the complex catchment geometries observed
This study also extends earlier observations on the morphometric properties of the Himalayan drainage basins (Hovius, 1996; Walcott and Summerfield, 2009) by including all the trans-Himalayan catchments that occur between the two syntaxial rivers of the orogen (i.e., Indus in the west and Brahmaputra in the east). With this extended data set, as well as a few new metrics, we also highlight variations in morphometric attributes from the middle towards the flanks of the Himalayan orogen. The pattern highlights the variations in relative influence of different drainage evolution processes and the orogen-scale heterogeneity in tectonic style.
5
ACCEPTED MANUSCRIPT
2. Geomorphological setting and geological history of the Himalayas
IP
T
2.1 Topography and drainage
SC R
The Himalayas form a high barrier between the plateau of Tibet to the north and the alluvial plains of the Indian subcontinent to the south. More than 100 peaks between the Nanga Parbat (8126 m) in Kashmir and the Namcha Barwa (7756 m) in eastern Tibet rise to elevations exceeding 7000
NU
metres. The mountain belt is ~2500 km-long and 200–400 km-wide and extends between the two
MA
syntaxial bends (Hazara in the west and Namcha Barwa in the east) where the strike of the mountain ranges turns abruptly south. The Himalayan orogen is delimited by the Indus–Tsangpo suture in the north, the left-lateral Chaman strike-slip fault in the west, the right-lateral Sagaing strike-slip fault in
TE
D
the east, and the Main Frontal Thrust (MFT) in the south (Yin, 2006).
CE P
The mountain belt consists of four parallel mountain ranges of varying widths, each with distinct physiographic features and a distinctive geological history. These are designated, from south to north, as the Sub-Himalaya; the Lesser Himalaya; the Great Himalaya; and the Tethys, or Tibetan,
AC
Himalaya. Farther north lies the Trans-Himalaya, in Tibet.
The Sub-Himalaya includes the Siwalik Range in the south and flat-floored structural valleys to its north. The main Siwalik Range displays a steeper southern slope facing the Indian plains and a gentler northern slope that descends northward to flat-floored basins called duns. Except for small gaps in the east, the Siwalik Range extends the entire length of the Himalayas. It presents a maximum width of 100 km. The southern boundary of the Siwalik Range occurs at an elevation of approximately 275 m and rises an additional 750 metres to the north. To the north of the SubHimalaya occurs a massive mountainous tract, the Lesser Himalaya. In this 80 km-wide range, mountains rise to 4500 m and valleys occur at elevations of ~900 m. Neighbouring summits share similar elevations, creating the appearance of an intensely dissected plateau. The backbone of the 6
ACCEPTED MANUSCRIPT entire Himalayan mountain system is the Great Himalaya, rising into the zone of perpetual snow. The range reaches its maximum height in Nepal. Among its peaks, 10 of them exceed 8000 metres.
T
The range trends northwest–southeast from Jammu and Kashmir to Sikkim, India. East of Sikkim it
IP
strikes east–west for another 420 km through Bhutan and the eastern part of Arunachal Pradesh,
SC R
and finally bends northeast, terminating at Namcha Barwa. There is no distinct boundary between the Great Himalaya and the mountain ranges, plateaus and basins lying to its north, which are generally grouped under the names of Tethys or Tibetan Himalaya. It is widest in Kashmir and
NU
Himachal Pradesh, forming the Spiti Basin and the Zaskar Range.
MA
The crest of the Great Himalaya defines the main topographic divide between the southern and northern flanks of the orogen (Fig. 2). Peaks higher than 5000 m line up along this topographic divide
D
(Fig. 1). This line is very distinct in the central and eastern Himalayas but becomes less so west of the
TE
Sutlej catchment, where it appears to branch out into sub-parallel ridges separated by strike valleys. The southern slope of the orogen receives large monsoon precipitation from June to September. The
CE P
topographic divide along the crest of the Great Himalaya prevents the moisture-laden monsoon from reaching the northern slopes, which are consequently arid.
AC
The main topographic divide and the main drainage divide are nonetheless not coincident in the Himalayas (Figs. 1, 2). For the region lying in between the two syntaxial rivers (i.e., Indus and Brahmaputra) the drainage divide, a composite of the upper extremities of the trans-Himalayan rivers, is situated about 50 km north of the topographic divide, except for a small part of the eastern Himalayas (Walcott and Summerfield, 2009). In contrast, the northern extremities of the Indus and Brahmaputra catchments are situated a few hundred kilometres north of the topographic divide.
2.2 Tectonic setting and physiographic evolution of the mountain belt
The India–Eurasia continental collision began around 45–55 m.y. ago (late Yepresian–Lutatian) with closure of the Neotethys Ocean (Rowley, 1996) leading to crustal deformation and formation of the 7
ACCEPTED MANUSCRIPT Himalayan mountain belt. Around 27 to 22 million years ago, the Himalayas existed as a positive topographic feature (Rowley, 1996; Einsele et al., 1996). Deep denudation (DeCelles et al., 1998;
T
Szulck et al., 2006) generated large masses of detritus shed by the rising orogen, and these have
IP
been preserved in the continental Himalayan foreland basins and in the Indus and Bengal fans. The
SC R
large physical, chemical and biological perturbations induced by the erosion of this mountain belt and the temporal variations in erosion rate might have caused a number of global changes including climate (e.g., inception of ice age and monsoon), river and sea-water chemistry, and even could have
NU
influenced crustal processes like thrust propagation and isostatic adjustments (Molnar and England, 1990; Burbank, 1992; Milliman and Syvitski, 1992; Raymo and Ruddiman, 1992; Derry and France-
MA
Lanord, 1996; Burbank and Anderson, 2001: pp. 221–223; Galy and France-Lanord, 2001; Zeitler et al., 2001; Wobus et al., 2003; Anders et al., 2006; Bookhagen and Burbank, 2006; Clift et al., 2008;
D
Singh et al., 2008; Whipple, 2009). The history of erosion of this mountain belt at various temporal
TE
and spatial scales continues to be studied using a number of techniques that include reconstruction
CE P
of deformed strata (Einsele et al., 1996), metamorphic petrology, thermochronology and other dating techniques (Yin, 2006), and numerical modelling (Kirby and Whipple, 2001; Lavé and Avouac,
AC
2001; Finlayson et al., 2002).
Generally speaking, ramping-up of crustal material along a number of northward-dipping, orogenparallel thrust zones sums up the principal mechanism behind the rise of the Himalayan mountain belt. The forces related to underthrusting of the Indian plate under the Tibetan plate have led to the formation of three major thrust systems, i.e. Main Central Thrust (MCT) in the north, the Main Boundary Thrust (MBT) in the middle, and Main Frontal Thrust (MFT) in the south (Yin, 2006). These thrust zones were formed one after the other due to progressive collision between the two plates. The deformational processes related to the MCT were most active between 22 and 16 Ma (DeCelles et al., 1998; Najman, 2006; Szulc et al., 2006). The main tectonic movements along the MBT system occurred around 10 Ma (Yin, 2006; Clift et al. 2008) and that for the MFT system around 1 Ma. Due
8
ACCEPTED MANUSCRIPT to continuing convergence, the deformational front has continued shifting south, thus emplacing the younger thrust systems to the south side of the older ones. The crustal material is uplifted near the
T
thrust ramps and is also tectonically transported towards the foreland over the thrust soles. The
IP
positions of the three Himalayan ranges (i.e., the Great, the Lesser and the Siwalik Range) are
SC R
roughly coincident with the locations where these three major thrust systems crop out. During the successive stages of frontal accretion, ongoing reactivation of internal structures narrowed the range and the interplay between internal thickening and frontal propagation determined whether the
NU
width of the mountain range would grow or shrink. However, the net outcome in the case of the Himalayas has been a widening of the orogen through addition of younger ranges on the Indo-
D
MA
Gangetic foreland side over the last ~20 million years.
TE
3. Methods for capturing the morphometry of catchments and watersheds
CE P
In this study, the structure of trans-Himalayan drainage networks as well as the geometry of the catchments and their constituent watersheds were obtained by using standard tools for hydrological
AC
analysis available in the Hydrology Toolbox of the ArcGIS software (version 10), and by applying them to a digital elevation model (DEM) of the Himalayan region. The SRTM DEM, available from http://srtm.csi.cgiar.org/, was used. The chosen area encompasses the catchments of all the rivers draining the southern slope of the Himalayas with exits through the mountain front located between those of the Indus and Brahmaputra. Each cell of the DEM raster represents a rectangular area ~88 m in length and width. The absolute vertical resolution of the SRTM data for the Himalayan region is ± 6.2 m and the relative vertical error is ± 8.7 m (Rodriguez et al., 2006). All the length and area measurements reported here correspond to geometries projected using UTM (zone 45 North) system and the WGS-84 datum.
9
ACCEPTED MANUSCRIPT The location of the mountain front (Fig. 1) was determined using the following method. It was assumed that a point lying within the mountainous region was more likely to have a higher slope
T
than a point lying in the plain given that the former is a deformed bedrock terrain and the later is
IP
predominantly a depositional surface. Therefore, to begin with a raster map of slope values was
SC R
prepared using the DEM. The cells of that raster were classified into two classes (i.e., slope < 5°: plain, and slope ≥ 5°: mountain). The boundary between these two types of cells, in theory, locates the position of the mountain front in the classified map. In this case that boundary was found to
NU
closely match the position of the mountain front shown in physical and topographic maps. However, closer scrutiny of the classified map reveals the presence of some noise induced, for example, by
MA
incised river channels within the alluvial plain and flat valley floors within the mountains. The noise was removed by repeatedly applying majority filter on the classified raster in order to obtain a raster
D
map displaying two regions clearly separated from each other. The line separating these two regions
TE
(Fig. 1) is defined here as the mountain front and was used to create a mask excluding the plains.
CE P
The updated DEM was used for subsequent analyses.
The hydrological tools were used at first to extract the structure of the stream network from the
AC
DEM (Fig. 3a). A cell in the DEM raster was considered to belong to drainage network only if it received flow from at least 3000 upstream cells (minimum upstream buffer area of 5 x 5 km). The networks of the trans-Himalayan rivers were isolated from the smaller networks and retained for further analyses (Fig. 1). The stream order was determined using Strahler’s method (Fig. 3a; Strahler, 1957). The points where rivers intersect the mountain front (drainage outlet; Fig. 3a) were used to determine the perimeter of the upland drainage basins using the hydrological tool of ArcGIS.
The drainage networks of the trans-Himalayan rivers reveal three distinct components. Near the mountain front, the network is characterized by long segments of tributary streams flowing parallel to the local strike of the mountain front and meeting the trunk stream at closely-spaced points (Fig. 3a). Upstream, the character of the network changes. This part is characterized by sub-networks of 10
ACCEPTED MANUSCRIPT tributary systems each featuring longer stream segments that are orthogonal to the mountain front. Further upstream, the network comprises a few, very long tributaries flowing sub-parallel to the
T
topographic divide and meeting to form the main trans-Himalayan river. Using the downstream
IP
endpoints (pour points) of each recognizable sub-network (here termed watershed) within a
SC R
catchment and watershed tool of ArcGIS software, we determined the watershed boundaries for each watershed (Fig. 4a). For example, the points downstream of which strike-parallel tributaries are significantly longer than the transverse stream segments were chosen as pour points (Figs. 3a, 4a).
NU
These define a population of Middle Transverse Watersheds (MTWs). The point where the main stream first enters a MTW serves as the pour point for the population of Upper Longitudinal
MA
Watersheds (ULWs), whereas the outlets of the catchments along the mountain front define the pour points of the Lower Longitudinal Watersheds (LLWs). The percent hypsometric curve (see
D
Strahler, 1952) for each watershed (Fig. 5) is constructed using the software tool developed by
TE
Perez-Peña et al. (2009). Subsequent to that, we measured a set of morphometric parameters
CE P
related to the two-dimensional shape of the trans-Himalayan catchments as well as those for the watersheds nested within the individual catchments.
AC
Hovius (1996) measured the distance between the two adjacent drainage outlets at the mountain front in order to determine the drainage “spacing”. However, when determining the lateral extent of a catchment a metric needs to be adopted that can relate to a single catchment and not to two neighbouring ones. For that purpose, earlier studies considered the average of the distances measured between the outlets of the neighbours lying to its left and right (Hovius, 1996; Walcott and Summerfield, 2009). It has been noted that for most trans-Himalayan catchments the outlet is shifted laterally (often only locally) from the axis of the main catchment due to large-scale drainage diversions near the frontal part of the orogen (Gupta, 1997). Thus, in the Himalayas, the position of the outlet is often not truly in line with the main axis of the huge trans-Himalayan catchment occurring upstream. Moreover, the location of the outlet is also affected by subjectivity in defining
11
ACCEPTED MANUSCRIPT the position of mountain front. Also, due to the sinuous nature of the front, in most cases the distances to the neighbouring outlets (lying to the left and right) cannot the measured along a single
T
straight line. We emphasize that the main objective here is to measure the orogen-parallel width of
IP
the catchment in between two transverse drainage divides. In the Himalayas, the divide shared by
SC R
two adjacent catchments bifurcates at their downstream end to give place to interstitial drainage basins (see Walcott and Summerfield, 2009; Figs. 3b, 3c). It follows that the width of the catchments, when measured between the downstream ends of two lateral catchment divides, provides a more
NU
straightforward, tangible metric (Ld), which possibly reflects the first harmonic of the topography
MA
more closely (Fig. 3c).
In order to determine the orogen-transverse width and also to characterize the overall shape of the
D
catchments, we constructed smallest rectangles for each catchment. The feature of these rectangles
TE
is that they completely enclose the catchment of interest and have one of the sides parallel to the local strike of the mountain front (Fig. 3c). The length of the orogen-transverse side of the rectangle
CE P
is taken as the orogen-transverse catchment length (Lt), whereas the length of the range-frontparallel side (Lp) defines the maximum orogen-parallel width (Fig. 3c). These three metrics, i.e., Ld,
AC
Lt, and Lp, together characterize the overall catchment shape (e.g., aspect ratio, triangular vs. rectangular). Similar methods were adopted to determine the corresponding metrics (i.e., WLd, WLt and WLp) for the watersheds nested within the individual catchments.
4. Results: catchments and drainage networks of the trans-Himalayan rivers The hydrological analysis revealed a number of catchments of varying shapes and dimensions that could be grouped into three classes according to their size and position within the landscape fabric
12
ACCEPTED MANUSCRIPT of the Himalayas — i.e., the mountain front, the composite drainage divide, and the topographic divide (Figs. 1 and 2).
T
The catchments of the two syntaxial rivers, Indus and Brahmaputra, are the largest. Both are
IP
situated to the north of the topographic divide and extend parallel to the orogen (Figs. 1, 3c). The
SC R
extreme ends of these catchments curve southwards near the two syntaxials bends of the mountain belt. Together, these two catchments occupy the entire northern part of the orogen from west to
NU
east and enclose the catchments of the other (the trans-Himalayan) rivers on three sides. The drainage network is dominated by long, orogen-parallel stream segments and short, transverse
MA
segments near the syntaxial bends (Fig. 3a). At their outlets on the southern mountain front, both the rivers flow as 7th-order streams (Table 1). Again, in both cases the trunk stream attains the
D
highest order before entering the southern slope of the orogen. The points at which the streams
TE
attain the highest order are situated behind both the topographic divide as well as the composite
CE P
drainage divide of the trans-Himalayan rivers (sensu. Walcott and Summerfield, 2009; Fig. 3a).
The catchments of other trans-Himalayan rivers (Fig. 3c) are much smaller than those of the syntaxial rivers. These catchments were referred to as “range-scale basins” by Walcott and
AC
Summerfield (2009), and their drainage networks were referred to as “High-order” and “Intermediate-order” rivers by Friend et al. (1999). The northern extremities of all these basins form a composite drainage divide that is oriented parallel to the topographic divide but situated about 150 to 50 km north of it in the central part of the orogen (Figs. 2, 3c). Only in the area drained by the river Jhelum of the western Himalayas and the group of rivers Tista, Torsa, and Raidak of the eastern Himalayas, does the upper drainage divide roughly coincides with the topographic divide (see also Walcott and Summerfield, 2009).
The variation in orogen-transverse (Lt) and orogen-parallel (Ld) dimensions of the 18 transHimalayan catchments is shown in Fig. 6a. We note that for these 18 catchments the mean value of
13
ACCEPTED MANUSCRIPT Lt is 182 km (s.d. 61) and that of Ld is 139 km (s.d. 66). However, at places their orogen-parallel dimension is larger than the transverse dimension, whereas at other places it is just the reverse (Fig.
T
6a). Figure 6b shows the along-strike variation in the ratio of Lp and Ld. It is apparent that all the
IP
catchments lying to the east of the Sutlej are roughly rectangular in shape (Lp/Ld values range
SC R
between 1 and 2) and the catchments occurring west of the Jamuna are very prominently triangular (Figs. 3c, 6b).
NU
In the central part of the Himalayas the catchments (i.e., Ghagra, also called Girwa, Gandak, and Kosi) are much larger than those of the flanking regions. They are rectangular, with their longer side
MA
parallel to the orogen (Fig. 3c). The Ld values are noticeably higher than the corresponding Lt values (Fig. 6a). The upper boundaries of these catchments are fairly straight and situated north of the
D
topographic divide. Among them, the Kosi and Ghagra have well-developed wing-like lateral
TE
projections in their upper part, which makes the Lp/Ld values slightly higher than that for the Gandak (Fig. 6b). To the east of these large basins, a group of four smaller rectangular catchments,
CE P
i.e. the Tista, Torsa, Raidak and Sankosh, contrasts with the earlier ones in having their longer side transverse to the orogen (Figs. 3c). The northern edges of these catchments are roughly coincident
AC
with the topographic divide. The group of centrally located, rectangular catchments (i.e., Ghagra to Sankosh) is flanked on either side by nearly square catchments, i.e., where Ld ≈ Lt: Jamuna, Ganga, Sarada in the west, and Manas and Bhareli in the east. Among them, the Ganga, Sarda, and Manas have their headwaters situated to the north of the topographic divide (Fig. 3a). The catchments occurring near the western and eastern extremities of the orogen have more complex shapes. All the catchments situated in the westernmost part (i.e., Jhelum, Chenab, Ravi and Sutlej) are “T”shaped and Lp/Ld values are greater than 2 (Table 1, Figs. 3c, 6b). They have a prominent, elongated, orogen-parallel upper component that is connected with the mountain front through a narrow orogen-transverse valley. In contrast, the Beas in the west and the Subansiri in the east are shaped like parallelograms, with their lower edge displaced towards the lateral ends of the orogen.
14
ACCEPTED MANUSCRIPT The orogen-transverse dimension (Lt) of the 18 trans-Himalayan catchments have been found to be correlated with their area (Fig. 6c) and the relationship can be expressed as Lt = 6.16 A0.34. Hack’ law
T
(Hack, 1957) describes the relationship between the length of the main stream (Lv) and the basin
IP
area (A), as Lv = cAb, where the coefficient (c) can vary between 1.4 and 3.0 and the exponent (b)
SC R
between 0.55 and 0.6 (Hovius, 1996). If we consider that the transverse dimension of the transHimalayan catchments approximates the length of the trunk stream, then both the coefficient and the exponent fall clearly outside the respective ranges. This deviation strongly suggests that the
NU
evolutionary process of the trans-Himalayan catchments deviates significantly from the drainage
MA
basins studied by Hack.
At drainage outlets the rivers Sutlej, Kosi and Subansiri flow as 6th order streams (Table 1). In the
D
cases of both the Sutlej and Kosi, the highest-order stream rises north of the topographic divide (Fig.
TE
3a). The highest-order stream of the Subansiri catchment, however, originates only at the downstream end of the network, i.e. within the Sub-Himalaya and very close to the mountain front.
CE P
There exists nonetheless one 5th-order stream in that network which rises north of the topographic divide. The data reveal ten catchments where the highest stream order is 5 (Table 1). For all the “T”-
AC
shaped catchments displaying wing-like projections in their upper part (i.e., Jhelum, Chenab and Ghagra), the highest-order streams rise north of the topographic divide. Among the remaining catchments, the highest-order stream rises at the downstream end of the catchments, except in the case of the Sankosh. There are five other basins where the highest stream order is 4, and the sources occur at locations immediately south the topographic divide (Table 1; Fig. 3a).
Innumerable rounder and much smaller watersheds occur within triangular interstices outlined by the lateral edges of two adjacent trans-Himalayan basins and the southern mountain front (Fig. 3b). These basins are the “interstitial basins” described by Walcott and Summerfield (2009) or “Loworder, Lower Himalayan and Sub-Himalayan rivers” of Friend et al. (1999). Their upper extremities are located midway between the mountain front and the topographic divide. The headwaters rise 15
ACCEPTED MANUSCRIPT within the Lesser Himalaya or the Siwalik Ranges, and these small rivers drain only the lower part of the southern slope of the orogen. The network patterns are predominantly dendritic. The maximum
T
stream order is 2. The following discussion ignores these small interstitial basins and their drainage
SC R
IP
networks.
NU
4.1 Drainage network patterns
The stream network of the trans-Himalayan catchments varies in character from the mountain front
MA
in the south to the upper drainage divide. Restricting analysis to 3rd and higher order streams, we focus on three outstanding drainage features that constitute individual components of most trans-
TE
longitudinal component (Fig. 3a).
D
Himalayan rivers: a lower longitudinal component, a middle transverse component, and an upper
CE P
The lower longitudinal component adjacent to the mountain front is characterized by a pinnate pattern. The longest stream segments are parallel to the local strike of the mountain front. The lower longitudinal drainage component is best developed for the three large centrally located basins
AC
(Ghagra, Gandak and Kosi) as well as for the Indus, Sutlej and Jamuna. The lower longitudinal drainage component is poorly defined for the Jhelum, Chenab, Ravi, Tista, Torsa, Raidak, Sankosh, Subansiri and Brahmaputra. In places, the trunk streams are strongly deflected parallel to the orogen immediately north of the mountain front. Large displacements of the trunk streams were noted in the drainage networks of the Sutlej, Ghagra and Bhareli (Fig. 3a).
The middle transverse component is dominated by a parallel drainage pattern, with the longest stream segments predominantly orthogonal to the strike of the topographic divide. This component is best developed in the Indus, Jamuna, Ganga, Ghagra, Gandak, Kosi, Tista, Manas, Subansiri and Brahmaputra. This pattern is weakly developed in the “T”-shaped basins of the western Himalayas
16
ACCEPTED MANUSCRIPT (Jhelum to Sutlej) and in transversely elongated basins of the eastern Himalayas (Torsa to Sankosh). These networks show dominance of a trellis pattern. The networks of the Sarda in the central
T
Himalayas, and Bhareli in the eastern Himalayas are, however, dominated by a dendritic pattern.
IP
The upper longitudinal drainage component is also dominated by a trellis pattern and contains very
SC R
long stream segments running parallel to the topographic divide. The upper longitudinal component is best developed for the two syntaxial rivers and for some of the trans-Himalayan rivers such as the
NU
Jhelum, Chenab, Sutlej, Ghagra, Kosi, and Subansiri.
MA
4.2 Morphometric properties and organization of the nested watersheds
Figure 4a shows the populations of watersheds corresponding to the Upper Longitudinal Watershed
D
(ULW), Middle Transverse Watershed (MTW) and Lower Longitudinal Watershed (LLW) groups,
TE
respectively.
CE P
ULWs are roughly rectangular and extend parallel to the orogen. These watersheds, where present (Table 1), are situated to the north of the topographic divide. To their north they are bound by the composite drainage divide of the trans-Himalayan basins and to their south by the topographic
AC
divide. All the “T”-shaped catchments of the western Himalayas (i.e., Jhelum, Chenab, Ravi, and Sutlej) contain well-developed ULWs. The same is true for the large, rectangular, central Himalayan catchments that exhibit wing-like projections in their upper parts (i.e., Ghagra and Kosi). The upper longitudinal watershed is poorly developed in the Gandak catchment. A small longitudinal watershed occurs in between the two MTWs: Gandak_A and Gandak_H. The ULWs are poorly developed for all the catchments lying between the Tista and the Subansiri. Though small ULWs are present in the Sankosh and Manas catchments, they are completely absent in others.
Usually, the ULWs occur in between the Indus–Tsangpo Suture Zone (ITSZ) in the north and the South Tibetan Detachment (STD) in the south, and occupy the region that is dominantly made up of
17
ACCEPTED MANUSCRIPT rocks belonging to the Tethyan Himalayan Succession (sensu. Yin, 2006). However, in the case of the Jhelum, Chenab, Beas, and Gandak catchments, the ULWs occur between the surface trace of the
T
MCT in the south and that of the STD in the north and occupy the region dominantly composed of
IP
rocks belonging to the Greater Himalayan Crystallines (Yin, 2006). The upper watershed of the
dominated by the Gangdese Batholith (Yin, 2006).
SC R
Subansiri catchment extends to the north of the ITSZ; to the southern edge of the Lhasa Block
NU
Within the individual trans-Himalayan catchments, the MTWs are situated between LLWs of the downstream and ULWs of the upstream parts of each catchment. If we consider all the MTWs of all
MA
the catchments, we note that they are arranged in an orogen-parallel linear array occupying the upper part of the southern slope of the orogen (Fig. 4a). Individual watersheds are triangular to
D
rectangular in shape and they are elongated transverse to the orogen. Their upper divide coincides
TE
with the main topographic divide and their downstream ends (“pour points”) are situated close to the boundary between the Siwalik and the Lesser Himalayan ranges. A few MTWs, however, extend
CE P
upstream to the northern topographic divide (e.g., Gandak_A, Gandak_H, Kosi_B, Manas_D and Manas_E). It appears that these extended watersheds follow the trend of prominent transverse
AC
structures. For example, Gandak_A follows the Kali Gandaki gorge of Nepal to the prominent transverse rift (Longge rift; see Yin, 2006). The same applies to the Gandak_H and Kosi_B MTWs. In general the transverse watersheds occupy the region lying in between the surface traces of the South Tibetan Detachment in the north and the MBT in the south. This region is made up of rock assemblages belonging to the Greater Himalayan Crystallines as well as those of the Lesser Himalayan Sequence (Yin, 2006).
The LLWs are roughly rectangular and extend parallel to the strike of the local mountain front. Welldeveloped LLWs are present in the catchments of the Indus and Jhelum, in all the catchments from the Beas to the Gandak as well in those from the Bhareli to the Brahmaputra. They are poorly
18
ACCEPTED MANUSCRIPT developed in the Kosi and Manas catchments. LLWs are almost absent from the eastern Himalayan
IP
T
catchments of the Torsa, Raidak and Sankosh rivers.
SC R
4.3 Hypsometric properties of the watersheds
The percentage hypsometric curves (Strahler, 1952) for the ULWs, MTWs and LLWs are shown in
NU
Figure 5. It is evident that the hypsometric properties of the ULWs are distinctly different from both the MTWs and LLWs. The curves for the ULWs are dominantly convex upward, particularly in the
MA
higher altitudes. The overall shape of these curves indicates that a high proportion of land area resides in the flattish upland and sloping midland. A small portion of the area resides in the
D
lowlands. Therefore, the landscape of the ULWs resembles plateaus with sharp edges, dissected by
TE
deep and narrow canyons. In contrast, the hypsometric curves for the MTWs and LLWs are characterized by gently sloping middle and lower parts along with a distinctly concave upper part.
CE P
The shape of these curves suggests that most of the area is concentrated in narrow midlands and wide lowlands, and a very small portion of the area occurs in the uplands. Therefore, it might be
AC
suggested that both the MTWs and the LLWs represent topographies dominated by wide valley floors, flattish midlands and very jagged uplands. These characteristics are better expressed among the LLWs. The distinctive hypsometric properties of the ULWs highlight the fact that the topography of the southern slope of the orogen is much more deeply dissected than that of the northern slope. It can be inferred that the processes controlling the form of the watersheds on the southern slope of the orogen (MTWs and LLWs) contrast sharply with those of the watersheds situated in the lee of the topographic divide. Rapid and intense bedrock erosion by a dense network of rivers supplied by intense monsoon precipitation on the southern slope plays a key role in shaping the MTWs and LLWs, whereas the dominance of gravitational slope processes under much lower precipitation,
19
ACCEPTED MANUSCRIPT coupled with higher retention of valley alluvium in the lee of the topographic divide, has produced
IP
T
watersheds (ULWs) that are significantly different in shape.
SC R
4.4 The MTWs and the relationship with their orogen-transverse length and area
Each of the three larger trans-Himalayan catchments (i.e. the Ghagra, Gandak and Kosi) individually
NU
hosts 4 to 6 MTWs (Fig. 7a; Table 1). Each of the catchments lying to the west and the east of the central region hosts 1 to 2 MTWs, except for the two sub-equant catchments (Manas and Bhareli)
MA
situated near the eastern extremity of the Himalayas, which each contain four narrow, closely spaced MTWs. If we consider the MTWs of all 18 catchments situated between the Indus and
D
Brahmaputra, their sizes are roughly uniform and their outlets are regularly spaced (Figs. 4a, 7a).
TE
Average WLd is 60 km (s.d. 32). WLt (average 106 km, s.d. 46) takes a larger value than the corresponding WLd and shows a similar pattern of variation (Fig. 7a) along the Himalayas from west
CE P
to east. The average WLt:WLd ratio is 2.1 and the average aspect ratio of the rectangular approximations is ~1.5. The orogen-parallel dimension of the rectangular approximations of the
AC
watersheds (WLp) shows values very similar to WLd (average: 74 km, s.d.: 34). Except for one MTW in the Ganga catchment, the MTWs are elongated and strike across the orogen (Fig. 7a). However, the MTWs belonging to the catchments of the Gandak across to the Sankosh are more rectangular compared to the MTWs occurring to the west and east of this region (Fig. 7b). The power function relating WLt to watershed area (A) takes the form: WLt =1.39 A0.52 (Fig. 7c). Considering that the orogen-transverse dimension of the transverse watersheds approximates the length of the corresponding main river valleys, we find that both the coefficient and the exponent of this relationship are similar to those of Hack’s law (Hovius, 1996). This relationship indicates that the erosional shape produced by growing drainage within the MTWs is still preserved to a certain extent.
20
ACCEPTED MANUSCRIPT 5. Discussion 5.1. Evolution of the trans-Himalayan river catchments
IP
T
Given a sloping surface with parallel upper and lower edges, it has been argued that the development of a drainage network begins at the lower edge (base level) as a number of small,
SC R
closely spaced rills and gullies (Horton, 1945). All of these tend to grow in an upstream direction and enlarge with time (Densmore et al., 2005; Perron et al., 2008). However, only a few can evolve into a
NU
network that is large enough to reach the upper edge of the surface (≈ to the topographic divide for two-sided orogens). The “successful” streams do so at the cost of their neighbours (Talling et al.,
MA
1997). Results from numerical simulations indicate that a particular combination of advective and diffusive processes promotes development of a branching stream network while others do not
TE
D
(Perron et al., 2008).
For all the catchments that extend all the way from the lower to the upper edges of the
CE P
experimental slope, the lateral divides separating the neighbours represent the lines along which the rates of widening of adjacent basins are balanced. Therefore, where the lateral divides are parallel to
AC
each other and extend all the way to the upper edge of the topographic slope it can be assumed that the neighbouring catchments are growing in a synchronous manner and that their growth rates (both in longitudinal and transverse directions) are similar. In other words, if the basin outlines are approximated by rectangles, the neighbouring rectangles will display very few overlaps. It should be noted that a similar configuration has been observed in the passive margin setting of South Africa (Walcott and Summerfield, 2008, 2009). The rectangular approximations of the basins studied by Walcott and Summerfield (2008), if constructed, will have very little overlap with the adjacent ones.
Based on the study of a number of drainage networks, Hack (1957) showed that the length of the longest path in a stream network was correlated with the catchment area and that the length could be expressed as a power function of area. The exponent (~0.5) and the coefficient (~ 1.5) of Hack’s
21
ACCEPTED MANUSCRIPT equation suggest that the catchment length is one and half times the side of a square with the same area as the catchment. In other words, it implies an aspect ratio of 2.25 for a rectangular catchment.
T
Earlier studies have noted that in a number of cases (except the Himalayas) the orogen-transverse
IP
and orogen-parallel dimensions are correlated and the ratio takes a value of 2 (Hovius, 1996; Talling
SC R
et al., 1997; Purdie and Brook, 2006; Walcott and Summerfield, 2009). Therefore, it is expected that the rectangular approximations of the drainage basins will have an aspect ratio of about 2.
NU
With the help of our extended data on 18 trans-Himalayan catchments, we document instead the following evidence: (i) the rectangular approximations of adjacent catchments show very large
MA
overlaps (Fig. 3c); (ii) the orogen-parallel dimension of the catchments is not uniform (Fig. 6a) along the Himalayas from west to east; (iii) the parallel and transverse dimensions are not correlated with
D
each other (Fig. 6a); (iv) the ratio between the two dimensions is much less than 2 (mean Lt/Ld is
TE
1.5; Table 1). In the case of the three central Himalayan catchments, the ratios are less than 1 (Table 1). In addition, we note that (v) the relationship between catchment length and catchment area
CE P
deviates significantly from Hack’s equation (Fig. 6c).
It is evident from the above observations that the shape of the trans-Himalayan catchments cannot
AC
be explained as the product of normal drainage evolution on a topographic slope as might be inferred by scaling up from a theoretical unit slope to a critical taper the size of the Himalaya. It is possible that mechanisms such as expansion of the transverse catchments into the hinterland (Densmore et al., 2005), orogenic accretion into the foreland (Castelltort and Simpson, 2006), and capture of antecedent watersheds have interacted at different stages of evolution of the Himalayan orogen to give rise to the modern-day catchment pattern.
It is generally difficult to identify the signatures of these different types of processes due to the low preservation potential of landforms in high-energy landscapes. In this study, we have nonetheless highlighted that the trans-Himalayan catchments can be subdivided into three watershed classes,
22
ACCEPTED MANUSCRIPT each with distinctive morphologies and links with the first-order topographic features of the orogen. The corresponding watersheds classes have been termed ULW, MTW and LLW. The morphologies
T
and organization of these three types of watersheds help us to identify the imprints of different
IP
processes and their mutual interactions during the growth of the Himalayas.
SC R
The reason why the rectangular approximations of the adjacent catchment basins overlap can be attributed to the presence of well-developed ULWs that makes Lp much higher than Ld. We propose
NU
that the catchments with a “T”-shaped geometry (e.g., Sutlej) are better explained by considering the existence of an older longitudinal drainage (represented by the ULWs) on the leeward side which
MA
was later captured by headward growth of transverse drainage on the southern slope, eventually breaching the topographic divide (Fig. 8). A similar mechanism has also been suggested by Brookfield
D
(1998) and Cina et al. (2009). In such catchments, the higher order streams rise from the northern
TE
slope of the orogen (Fig. 3a), which also indicates capture of a well-developed, pre-existing drainage on the northern side by a younger, lower-order transverse stream (Fig. 8). Following that event, the
CE P
entire upper longitudinal drainage was diverted southward.
The high rate of erosion by fluvial processes induced by intense monsoon precipitation on the
AC
southern slope probably helped the transverse streams to cross the topographic divide. The “S”shaped, or convex-up, hypsometric curves for the UTWs are distinctly different from the concave-up curves of the MTWs and LLWs occupying the middle and lower parts of the catchments, respectively. This suggests that the middle and lower parts of the trans-Himalayan catchments are in a state of advanced landscape dissection, in spite of the fact that the topography of the southern slope is more strongly influenced by the deformational processes given that the deformational front kept shifting southwards over time. The southern slope experiences very heavy annual precipitation from the southwest monsoon (Bookhagen and Burbank, 2006; Anders et al., 2006). The annual precipitation drops drastically to the north of the topographic divide. The rainshadow region is dominated by diffusive hillslope processes, whereas advective processes contribute significantly to shaping the 23
ACCEPTED MANUSCRIPT landscape of the windward side. The hypsometric properties of the watersheds on the southern slope thus more directly reflect erosion by surface processes rather than crustal deformation. One
T
may thus speculate that the catchments comprising ULWs became confined to the leeward side of
IP
the topography before monsoon circulation began to operate across South Asia, thus documenting
SC R
the relative antiquity of these catchments.
A mechanism similar to that for the Sutlej also explains the shape of the Kosi catchment (Fig. 4a),
NU
where, among a number of MTWs, the drainage of the Kosi_E connects with the drainage of the Kosi_A (ULW). Other MTWs do not extend beyond the topographic divide. In this case, however, the
MA
shape of the catchment is not “T”-shaped as the drainage diversions at the downstream end had combined the flows of the laterally disposed siblings. It is difficult to confirm why only one or two
D
among the MTWs within a single catchment breaches the topographic divide and connects to the
TE
ULWs. It may, however, be noted that those unique drainage systems flow through lower altitudes and have a lower overall slope than the drainage of adjacent MTWs. It is thus possible that the
CE P
successful ones trace the trends of the orogen-transverse rifts that formed around 8-9 m.y. ago (Yin and Harrison, 2000; Yin, 2006). The presence of these rifts and other transverse tectonic structures
(Fig. 6b).
AC
might also be related to the more triangular geometry of the catchments lying to the west of Jamuna
The middle part of the trans-Himalayan catchments comprises an array of nested MTWs. The large catchments of the central Himalayas individually contain 4 to 6 MTWs. These watersheds have reasonably straight lateral drainage boundaries that are transverse to the orogen. Their rectangular approximations do not overlap significantly (Fig. 4b). Moreover, the orogen-parallel dimension of these watersheds is uniform and shows a variation pattern similar to that of their transverse dimensions (Fig. 7a). The average value of the ratio between the transverse and parallel dimensions of the watersheds is greater than 2 (2.1). The power-law relationship between the transverse length and the area has an exponent of 0.52 and a coefficient of about 1.4 (Fig. 7c), similar to those of 24
ACCEPTED MANUSCRIPT Hack’s relationship. Morphologically, therefore, MTWs resemble a series of small drainage basins at their early stage of growth. We propose that the MTWs are a legacy of of the drainage that basins
T
formed on the precursor topographic “taper” of the modern Himalayan mountain belt. They may be
IP
equated to the transverse rivers shown in Gupta (1997, Fig. 4 therein). The mean transverse (WLt)
SC R
and parallel (WLd) dimensions of ~74 km and ~57 km, respectively, suggest that the MTW rivers, at that time, were draining a ~100 km-long topographic slope.
NU
The Siwalik Range was formed due to the up-thrusting of the foreland sedimentary deposits along the MFT in a later stage of orogeny. This rising mountain range laterally diverted the pre-existing
MA
rivers draining watersheds to meet the flows of the neighbouring watersheds (see also Gupta, 1997). The structural depressions (represented by the LLWs) lying behind the range are characterized by
D
long range-front-parallel tributaries meeting the trunk stream of the trans-Himalayan drainage (Fig.
TE
3a).
CE P
The number of MTWs whose flows have been joined varies from catchment to catchment (Fig. 7a). In the central Himalayan region, each individual catchment contains 4 to 6 MTWs (Fig. 7a), hence their elongated shapes parallel to the strike of the orogen (Figs. 3c, 6a). Given that the diversion of 3
AC
to 4 streams is required to produce the observed pattern, and considering further that the individual MTWs are about ~60 km wide, the catchment outlets need to be separated by uplift zones at least 180 to 240 km long. The exposure belt of the Siwalik rocks are either absent or very narrow near the outlets of the Tista to the Sankosh. Each of these contains one (at best two) MTW. Here, the transverse structures aligned along the divide between the catchments of the Tista and Torsa possibly interfered with the formation of closely spaced, coherent thrust-related deformational structures in the frontal zone. Deformations near the outlets of the Manas, Bhareli and Subansiri (eastern Himalayas) are associated with flow diversion of 2 to 3 MTWs. It follows that the zones of deformation in that area were relatively shorter (~120–180 km) than in the central Himalayas. Finally, each of the catchments lying to the west of the Ghagra contains only few MTWs. The 25
ACCEPTED MANUSCRIPT outlines of many of these MTWs are very irregular (Fig. 4a). Possibly, the deformations related to the transverse structures (e.g., Pulan–Gurla Mandhata extensional system: Yin, 2006) close to the Delhi
T
Ridge played a role in perturbing the movements along the frontal thrust in the Sutlej region,
IP
causing less extensive diversions compared to that of the Kosi and thereby hindering the formation
SC R
of large LLWs. Had the downstream drainage diversion been as intense, it is likely that the group of catchments comprising the Beas, Sutlej, Jamuna, Ganga and Sarda (all with MTWs similar to
NU
Sutlej_B) would have been combined into a single catchment with a shape similar to that of the Kosi.
MA
5.2. A relative chronology of trans-Himalayan drainage basin evolution
D
We propose that the trans-Himalayan drainage basins comprises a number of geomorphologically
TE
distinct components associated with different physio-tectonic and climatic events that prevailed at different time intervals. The Himalayan mountain belt is an active, compressional, orogen that is
CE P
widening progressively due to uplift of newer parts of the foreland lying to its south. Both the tectonic and physiographic fronts have shifted towards the foreland with time. Therefore, the model
AC
of headward enlargement of the gullies starting from the mountain front cannot be directly applied to explain the evolution of trans-Himalayan basins that have a much more complex outline and host multi-component drainage networks. Therefore, the evolutionary model needs to take into account the effects of widening of the orogen, enlargement of the basins due to headward erosion as well as the capture of earlier drainage, if any.
The evolution took place in three major stages (Fig. 8):
Stage 0 (before 20 Ma). The collision between the Indian and Tibetan continents has begun but major movements along the MCT system are yet to happen. The Himalayan mountains are yet to form. A set of roughly east–west, sub-parallel river systems are present on the gently sloping
26
ACCEPTED MANUSCRIPT southern edge of the Tibetan land. These rivers possibly originated from the central Tibetan highlands and terminated in the seas lying to the west and the east. A similar configuration has also
T
been suggested by Wadia (1953), Brookfield (1998), Sinclair and Jaffy (2001), Yin (2006), Cina et al.
IP
(2009), and Robinson et al. (2014).
SC R
Stage 1 (between 20 Ma and 10 Ma). The Tethyan accretionary wedge expands and grows into a mountain range. Major thrusting events take place along the MCT system and large deformational
NU
structures develop due to thrust propagation. As a result of the tectonic deformations related to the MCT and STD systems, the precursor of the Greater Himalayan Range forms (Fig. 8) and begins to
MA
function as the topographic divide separating a distinct southern slope from the northern one. The river systems of Stage 0 are now confined to a set of structurally-guided parallel valleys situated to
D
the north of the new topographic divide. A similar scenario has also been suggested by Brookfield
TE
(1998), Hallet and Molnar (2001) and Sinclair and Jaffey (2001). A number of regularly-spaced channel networks form along the southern slope of the mountain. They progressively grew towards
CE P
the hinterland and evolved in a manner similar to that suggested by Densmore et al. (2005) and Perron et al. (2008). Some of them might have widened laterally by competitive capture. However,
AC
the new transverse drainages remained disconnected from the pre-existing, east–west-flowing rivers on the northern side. The rivers affiliated to this evolutionary stage have mainly derived sediments from the Greater Himalayan Crystaline Complex, or GHC (sensu. Yin, 2006).
Stage 2 (10 Ma to ~1 Ma). The major tectonic movements take place along the MBT system. Deformational structures produced by thrust propagation develop to the south of the existing topographic divide. The Lesser Himalayan Range forms and the orogen becomes wider as the mountain front shifts to the southern edge of the newly formed range. The flows of some of the transverse catchments that formed during Stage 1 are diverted by the rising ranges, giving rise to larger catchments. A set of closely-spaced transverse drainage systems also developed along the southern slope of the newly formed range (see also Friend et al., 1999). Monsoon circulation begins 27
ACCEPTED MANUSCRIPT to operate and the southern face begins to receive large amounts of precipitation. As a consequence of enhanced fluvial activity, the southern face starts to erode rapidly. The rate of headward growth
T
of the existing transverse drainages increases. Some of them breach the topographic divide and
IP
begin to capture the river systems to the north. The cross-structure (transverse rifts) that formed ca.
SC R
8–9 Ma might have aided the process of divide breaching. The complex interaction among the different co-existing drainages is probably represented by the organization of the MTWs and the ULWs that we observe today. It is possible that, during this stage, the southern slope of the
NU
Himalayas was about 100 km wide, in contrast to the ~150 km observed today (Fig. 2). The transverse rivers flowing along the southern slope during Stage 2 eroded rocks belonging to the GHC
MA
and the Lesser Himalayan Sequence, or LHS (sensu. Yin, 2006) and supplied the detritus to the foreland sedimentary basin(s) situated to the south of the mountain front at that time. Those
TE
continental margin.
D
sediments are now preserved within the Siwalik sequences and in the submarine fans at the
CE P
Stage 3 (late Quaternary). Major tectonic movements occur along the MFT system and an array of thrust-related deformational structures forms to the south of the Lesser Himalayan Range, giving
AC
rise to the Siwalik Range. The southern edge of the orogen at this stage roughly defines the modern mountain front. The pre-existing transverse rivers of Stage 2 are diverted parallel to the orogen and begin to collectively contribute to the trunk streams (cf. Friend et al., 1999). The diversion combined a number of adjacent catchments into unusually wide trans-Himalayan catchments (Fig. 3c). At the same time, new drainages begin to evolve on the slopes of the newly formed ridges of the Siwalik Range. They represent some of the interstitial basins occurring between the outlets of the transHimalayan rivers (Figs. 8 and 3b).
28
ACCEPTED MANUSCRIPT 6. Conclusions In this work we studied the morphological characteristics of all 18 trans-Himalayan river catchments
T
situated between the exits of the two syntaxial rivers, the Indus and the Brahmaputra. The aim was
IP
to capture clues about the relative chronology of drainage integration in response to tectonic uplift.
SC R
Our observations demonstrate that the shape of the trans-Himalayan catchments cannot be explained only as the product of normal drainage evolution on a topographic slope due to headward
NU
enlargement of drainage networks. In this case, additional mechanisms such as the headward capture of pre-existing longitudinal drainage systems situated on the northern side of the main
MA
topographic divide of the Himalayas, as well as drainage reorganization near the southern mountain front, were equally important in determining the morphology of these catchments.
D
Within individual catchments we identified the existence of distinct orogen-transverse drainage
TE
components that are arranged in a systematic manner with respect to first-order physiographic
CE P
features of the Himalayas formed at different periods of geological time. The morphometric attributes of three classes of watershed, i.e. sub-catchments nested within the 18 larger catchments, were used to define these drainage components. They each show distinct characteristics that are
AC
indicative of different stages or processes that were involved in shaping the composite transHimalayan catchments. The shape of the transverse watersheds occurring in the middle of the catchments resembles a series of small drainage basins formed on the precursor topography of the modern Himalayan mountain belt, which can be conceptually simplified as a wedge. The corresponding south-facing slope at that time was possibly ~100 km wide. The strike-parallel watersheds occurring at the upstream end of the catchment (ULWs) and situated on the northern side of the main topographic divide display distinctly different hypsometric properties compared to the watersheds occupying the middle and lower part of the catchments. The effects of intense fluvial dissection of the landscape induced by monsoon precipitation in the middle and lower watersheds are absent in the upper watersheds. The lower part of the catchments is characterized by flow 29
ACCEPTED MANUSCRIPT diversion of a number of pre-existing transverse drainages, indicating the influence of growing deformational structures along the mountain front. However, the number of diverted streams per
T
catchment varies from the central part of the Himalayas to its western and eastern sides, indicating
IP
a lateral variation in the nature of deformation relating to the frontal thrust system as well as to the
SC R
influence of the transverse structures.
Based on the observed relationships between the different kinds of drainage components and the
NU
first-order physiographic features of the Himalayas, we propose that a set of transverse drainages developed on the southern slope after the formation of the main topographic divide of the
MA
Himalayas (Greater Himalayan Range), about 20 million years ago. The rates of headward growth of some among these transverse drainages were greater than others. The more “successful” streams
D
breached the topographic divide and captured parts of pre-existing strike-parallel drainages of the
TE
northern slope. The initiation of monsoon circulation and the formation of cross-structures (Neogene rifts) around 10 to 7 million years ago possibly promoted the headward erosion of some of
CE P
the transverse drainages. The formation of newer ranges to the south of the topographic divide, i.e., the Lesser Himalaya and the Siwalik Range, diverted the downstream parts of the laterally adjacent
AC
south-flowing streams. The upland basins of these diverted drainages combined to form larger ones. In summary, in its initial phase the orogen-transverse dimension of the drainage basins increased due to headward erosion of the streams. In the later phase, although headward enlargement continued unabated, the process of capturing upstream, strike-parallel drainage components and the annexation of the uplifted foreland area at the downstream end gained increasing influence on the orogen-parallel dimension of the basins. The shape of the trans-Himalayan catchments and their spatial links with the physiographic and tectonic features can thus help to understand the chronology of Himalayan evolution over the last 20 million years.
30
ACCEPTED MANUSCRIPT Acknowledgments The authors remain grateful to the Geological Studies Unit of the Indian Statistical Institute, Kolkata,
T
India for providing the necessary computational facility for this study. We also remain thankful to
IP
Ms. I. Sengupta and Mr. K. Hatui for their help in data preparation and analysis. The revised version
SC R
of the manuscript benefited immensely from the suggestions of Hugh Sinclair, Yanni Gunnell and
AC
CE P
TE
D
MA
NU
Adrian Harvey.
31
ACCEPTED MANUSCRIPT References Anders, A.M., Gerard H. Roe, G.H., Hallet, B., Montgomery, D.R., Finnegan N.J., Putkonen, J., 2006.
T
Spatial patterns of precipitation and topography in the Himalaya. Geological Society of America,
IP
Special Paper 398, 39–52.
SC R
Bookhagen, B., Burbank, D.W., 2006. Topography, relief, and TRMM-derived rainfall variations along the Himalayas: Geophysical Research Letters 33, L08405.
NU
Brookfiled, M.E., 1998. The evolution of the great river systems of southern Asia during the Cenozoic
MA
India-Asia collision: rivers draining southwards. Geomorphology 22, 285–312. Burbank, D.W., 1992. Causes of recent Himalayan uplift deduced from deposited patterns in the
D
Ganges basin. Nature 357, 680–683.
TE
Burbank, D.W., Anderson, R.S., 2001. Tectonic Geomorphology, Blackwell Science, Oxford, 274 pp.
CE P
Burrard, S.G., Hayden, H.H., 1907. Geography and Geology of Himalayan Mountains and Tibet, Part 3: The Rivers of Himalaya and Tibet. Government of India Press, Calcutta, pp. 119-230.
AC
Castelltort, S., Simpson, G., 2006. River spacing and drainage network growth in widening mountain ranges. Basin Research 18, 267–276. Cina, S.E., Yin, A., Grove, M., Dubey, C.S., Shukla, D.P., Lovera, O.M., Kelty, T.K., Gehrels, G.E., Foster, D.A., 2009. Gangdese arc detritus within the eastern Himalayan Neogene foreland basin: Implications for the Neogene evolution of the Yalu–Brahmaputra River system. Earth and Planetary Science Letters 285, 150–162. Clift, P.D., Hodges, K.V., Heslop, D., Hannigan, R., Long, H.V., Calves, G., 2008. Correlation of Himalayan exhumation rates and Asian monsoon intensity. Nature Geoscience 1, 875–880.
32
ACCEPTED MANUSCRIPT DeCelles, P.G., Gehrels, G.E., Quade, J., Ojha, T.P., Kapp, P.A., Upreti, B.N., 1998. Neogene foreland basin deposits, erosional unroofing, and the kinematic history of the Himalayan fold-thrust belt,
T
western Nepal. Geological Society of America Bulletin 110, 2–21.
IP
Densmore, A.L., Dawers, N.H., Gupta, S., Guidon, R., 2005. What sets topographic relief in
SC R
extensional footwalls? Geology 33, 453–456.
Derry, L.A., France-Lanord, C., 1996. Neogene Himalayan weathering history and river 87Sr/86Sr:
NU
impact on the marine Sr record. Earth and Planetary Science Letters 142, 59–74. Einsele, G., Ratschbacher, L., Wetzel, A., 1996. The Himalaya-Bengal fan denudation-accumulation
MA
system during the past 20 Ma. Journal of Geology 104, 163–184.
D
Finlayson, D.P., Montgomery, Hallet, B., 2002. Spatial coincidence of rapid inferred erosion with
TE
young metamorphic massifs in the Himalayas. Geology 30, 219–233.
CE P
Friend, P.F., Jones, N.E., Vincent, S.J., 1999. Drainage evolution in active mountain belt: extrapolation backwards from present-day Himalayan river patterns. In: Smith, N.D., Rogers, J. (Eds.), Fluvial Sedimentology VI. Special Publications of International Association of Sedimentologists 28, 305–313.
AC
Galy, A., France-Lanord, C., 2001. Higher erosion rates in the Himalaya: Geochemical constraints on riverine fluxes. Geology 29, 23–26. Gupta, S., 1997. Himalayan drainage patterns and origin of fluvial megafans in the Ganges foreland basin. Geology 25, 11–14. Hack, J.T., 1957. Studies of longitudinal stream profiles in Virginia and Maryland. U.S. Geological Survey, Special Paper 294-B, pp. 45–97. Hallet, B. Molnar, P., 2001. Distorted drainage basins as markers of crustal strain east of the Himalayas. Journal of Geophysical Research 106, 13,697–13,709.
33
ACCEPTED MANUSCRIPT Horton, R.E., 1945. Erosional development of streams and their drainage basins; hydrophysical approach to quantitative morphology. Geological Society of America Bulletin 56, 275–370.
IP
T
Hovius, N., 1996. Regular spacing of drainage outlets from linear mountain. Basin Research 8, 29–44.
Geology 29, 415–418.
SC R
Kirby, E., Whipple, K.X., 2001. Quantifying differential rock-uplift rates via stream profile analysis.
Lave, J., Avouac, J.P., 2001. Fluvial incision and tectonic uplift across the Himalayas of central Nepal.
NU
Journal of Geophysical Research 106, 26,561–26,591.
MA
Milliman, J.D., Syvitski, J.P.M., 1992. Geomorphic/tectonic control of sediment discharge to the ocean: The importance of small mountainous rivers. Journal of Geology 100, 525–544.
TE
chicken or egg. Nature 346, 29–34.
D
Molnar, P., England, P., 1990. Late Cenozoic uplift of mountain ranges and global climate change:
CE P
Najman, Y., 2006. The detrital record of orogenesis: a review of approaches and techniques used in the Himalayan sedimentary basins. Earth-Science Reviews 74, 1–72.
AC
Perez-Peña, J.V., Azañon, J.M., Azor, A., 2009. CalHypso: An ArcGIS extension to calculate hypsometric curves and their statistical moments. Applications to drainage basin analysis in SE Spain. Computers and Geoscience 35, 1214–1223. Perron, J.T., Dietrich, W.E., Kirchner, J.W., 2008. Controls on the spacing of first-order valleys. Journal of Geophysical Research—Earth Surface 113, F04016. Purdie, H., Brook, M., 2006. Drainage spacing regularity on a fault-block: a case study from the Ruahine Range. New Zealand Geographer 62, 97–104. Raymo, M.E., Ruddiman, W.F., 1992. Tectonic forcing of late Cenozoic climate. Nature 359, 117–122.
34
ACCEPTED MANUSCRIPT Robinson, R.A.J., Brezina, C.A., Parrish, R.R., Horstwood, M.S.A., Oo, N.W., Bird, M.I., Thein, M., Walters, A.S., Oliver, G.J.H., Zaw, K., 2014. Large rivers and orogens: The evolution of the Yarlung
T
Tsangpo–Irrawaddy system and the eastern Himalayan syntaxis. Gondwana Research 26, 112-121.
SC R
Photogrammetric Engineering & Remote Sensing 72, 249–260.
IP
Rodriguez, E., Morris, C.S., Beltz, J.E., 2006. A global assessment of SRTM performance.
Rowley, D.B., 1996. Age of initiation of collision between India and Asia: a review of stratigraphic
NU
data. Earth Planetary Science Letters 145, 1–13.
Shroder, J.F., Bishop, M.P., 2004. Mountain geomorphic systems. In, Bishop, M.P., Shroder, J.F.Jr.
MA
(Eds.), Geographic Information Science and Mountain Geomorphology. Praxis Publishing Limited,
D
Chichester, pp. 33–73.
TE
Sinclair, H.D., Jaffy, N., 2001. Sedimentology of the Indus Group, Ladakh, northern India: implications
151–162.
CE P
for the timing of initiation of the palaeo-Indus River. Journal of the Geological Society of London 158,
Singh, S.K., Rai, S.K., Krishnaswami, S., 2008. Sr and Nd isotopes in river sediments from the Ganga
AC
Basin: sediment provenance and spatial variability in physical erosion. Journal of Geophysical Research 113, 1–18.
Strahler, A.N., 1952. Hypsometric (Area-Altitude) Analysis of Erosional Topography. Geological Society of America Bulletin 63, 1117–1142. Strahler, A.N., 1957. Quantitative analysis of watershed geomorphology. Transactions, American Geophysical Union 38, 913–920. Szulc, A.G., Najman, Y., Sinclair, H.D., Pringle, M., Bickle, M., Chapman, H., Garzanti, E., Andò, S., Huyghe, P., Mugnier, J-L., Ojha, T., DeCelles P., 2006. Tectonic evolution of the Himalayas
35
ACCEPTED MANUSCRIPT constrained by detrital 40Ar–39Ar, Sm-Nd and petrographic data from the Siwalik foreland basin succession, SW Nepal. Basin Research 18, 375–391.
IP
outlets from linear fault blocks. Basin Research 9, 275–302.
T
Talling, P.J., Stewart, M.D., Stark, C.P., Gupta, S., Vincent, S.J., 1997. Regular spacing of drainage
SC R
Wadia, D.N., 1953. Geology of India (third edition). MacMillan, London, 553 pp. Walcott, R.C., Summerfield, M.A., 2008. Scale dependence of hypsometric integrals: an analysis of
NU
south-east African basins. Geomorphology 96, 174–186.
MA
Walcott, R.C., Summerfield, M.A., 2009. Universality of variability in basin outlet spacing: implications for the two-dimensional form of drainage basins. Basin Research 21, 147–155.
TE
Geoscience 2, 97–104.
D
Whipple, K.X., 2009. The influence of climate on the tectonic evolution of mountain belts. Nature
CE P
Wobus, C.W., Hodges, K.V., Whipple, K.X., 2003. Has focused denudation sustained active thrusting at the Himalayan topographic front? Geology 31, 861–864.
AC
Yin, A., 2006. Cenozoic tectonic evolution of the Himalayan orogen as constrained by along-strike variation of structural geometry, exhumation history, and foreland sedimentation. Earth-Science Reviews 76, 1–131.
Yin, A., Harrison, T.M., 2000. Geologic evolution of the Himalayan–Tibetan orogen. Annual Review of Earth and Planetary Sciences 28, 211–280. Zeitler, P.K., Meltzer, A.S., Koons, P.O., Craw, D., Hallet, B., Chamberlain, C.P., Kidd, W.S.F., Park, S. K., Seeber, L., Bishop, M., Shroder, J., 2001. Erosion, Himalayan geodynamics, and the geomorphology of metamorphism. GSA Today 11, 4–9.
36
ACCEPTED MANUSCRIPT Figure Captions
T
Figure 1. Map showing the first-order physiographic features of the Himalayan mountain belt. The
IP
area drained by the two syntaxial rivers, Indus and Brahmaputra, as well as that by the trans-
SC R
Himalayan rivers are marked.
Figure 2. Generalized physiographic profile of the Himalayan mountain belt along a traverse
NU
extending from the mountain front in the south to the edge of the Tibetan plateau in the north. The average elevation of the main topographic divide, the composite drainage divide of the trans-
MA
Himalayan rivers and the southern edge of the Tibetan plateau are plotted against distance from the mountain front. Vertical bars represent the standard deviations. The elevations were measured at
D
the outlets of the trans-Himalayan rivers for the mountain front; at the peaks, higher than 5000
TE
meters, for the topographic divide; along the trunk streams for the upper reaches of trans-
CE P
Himalayan rivers; along the composite drainage divide, along the trunk streams of the Indus and Brahmaputra and along the northern catchment boundaries of the two syntaxial rivers. The distances are averages of measurements made along a number of orogen transverse lines in the
AC
western and eastern Himalayas.
Figure 3. A: Map showing the structure of drainage network within the catchment of eighteen transHimalayan and two syntaxial rivers. The catchment outlets are marked by filled circles and the origin of the highest order stream in each network is marked. Note the well-developed longitudinal drainage components in the Indus, Brahmaputra, Sutlej, Ghagra and Kosi rivers and that the highestorder streams originate on the northern slope of the Himalayas. Also note prominent distortions in the trunk-stream near the mountain front in the networks of Sutlej, Jamuna, Ganga Ghagra, etc. B: small, sub-equant interstitial basins occurring between the Kosi and Manas catchments. C: rectangular approximations for the trans-Himalayan catchments. The catchments are shaded differently for clarity. The method of measuring Ld, Lt and Lp for one of the catchments is shown. 37
ACCEPTED MANUSCRIPT
Figure 4. Watersheds nested within the catchments of the trans-Himalayan and syntaxial rivers. A:
T
map showing the organization of the Upper Longitudinal, Middle Transverse and Lower Longitudinal
IP
watersheds within individual syntaxial and trans-Himalayan catchments. The pour points of the
SC R
watersheds are shown by filled (white) circles. Note that the MTW pour points are regularly arranged along an orogen-parallel line passing midway between the mountain front and the topographic divide. Also note that the upper extremities of the MTWs are roughly coincident with
NU
the topographic divide, except for a few that extend to the northern side of the divide. B:
MA
rectangular approximations for the Middle Transverse Watersheds (MTW). Figure 5. Watershed percentage hypsometric curves. A: Upper Longitudinal Watersheds. B: Middle
D
Transverse Watersheds. C: Lower Longitudinal Watersheds.
TE
Figure 6. West-to-east variation in morphological parameters and the length vs. area relationship of
CE P
the 18 trans-Himalayan catchments. A: Variation in orogen-transverse (Lt) and orogen-parallel (Ld) dimensions. B: Variation in the ratio between orogen-parallel dimension of the rectangular approximation (Lp) and Ld. C: Relationship between the orogen-transverse length (Lt) and area (A)
AC
among the population of trans-Himalayan catchments. Figure 7. West-to-east variation in morphological parameters and the length vs. area relationship of the MTWs. A: Variation in orogen-transverse (WLt) and orogen-parallel (WLd) dimensions. B: Variation in the ratio between orogen-parallel dimension of the rectangular approximation (WLp) and WLd. C: Relationship between the orogen-transverse length (WLt) and area among the population of MTWs. Figure 8. Schematic representation of different stages of evolution of the trans-Himalayan catchments in map view (left panel) and along a hypothetical orogen-transverse profile (right panel). Not to scale. S0: catchment width of the transverse streams that developed during the initial stage.
38
ACCEPTED MANUSCRIPT S1: width of the upland catchments after drainage reorganization at their downstream end. Note that S1 is two to four times larger than S0. Changes in the orogen-transverse width of the evolving
T
catchments are indicated by a family of “W”s. Subscripts u and d differentiate the upstream and
IP
downstream components of the width variable, and the numerals after “d” define the development
SC R
stage. The vertical, broken lines, in the right-hand panel, indicate the positions of crest-line of orogen-parallel ranges, whereas, the thick broken lines (left panel) mark the outlines of the
AC
CE P
TE
D
MA
NU
catchments in Stage 3.
39
MA
NU
SC R
IP
T
ACCEPTED MANUSCRIPT
AC
CE P
TE
D
Figure 1
40
NU
SC R
IP
T
ACCEPTED MANUSCRIPT
AC
CE P
TE
D
MA
Figure 2
41
MA
NU
SC R
IP
T
ACCEPTED MANUSCRIPT
AC
CE P
TE
D
Figure 3a
42
NU
SC R
IP
T
ACCEPTED MANUSCRIPT
AC
CE P
TE
D
MA
Figure 3b
43
MA
NU
SC R
IP
T
ACCEPTED MANUSCRIPT
AC
CE P
TE
D
Figure 3c
44
MA
NU
SC R
IP
T
ACCEPTED MANUSCRIPT
AC
CE P
TE
D
Figure 4a
45
MA
NU
SC R
IP
T
ACCEPTED MANUSCRIPT
AC
CE P
TE
D
Figure 4b
46
MA
NU
SC R
IP
T
ACCEPTED MANUSCRIPT
AC
CE P
TE
D
Figure 5a
47
MA
NU
SC R
IP
T
ACCEPTED MANUSCRIPT
AC
CE P
TE
D
Figure 5b
48
MA
NU
SC R
IP
T
ACCEPTED MANUSCRIPT
AC
CE P
TE
D
Figure 5c
49
TE
D
MA
NU
SC R
IP
T
ACCEPTED MANUSCRIPT
AC
CE P
Figure 6a
50
CE P AC
Figure 6b
TE
D
MA
NU
SC R
IP
T
ACCEPTED MANUSCRIPT
51
NU
SC R
IP
T
ACCEPTED MANUSCRIPT
AC
CE P
TE
D
MA
Figure 6c
52
AC
Figure 7a
CE P
TE
D
MA
NU
SC R
IP
T
ACCEPTED MANUSCRIPT
53
AC
Figure 7b
CE P
TE
D
MA
NU
SC R
IP
T
ACCEPTED MANUSCRIPT
54
NU
SC R
IP
T
ACCEPTED MANUSCRIPT
AC
CE P
TE
D
MA
Figure 7c
55
TE
D
MA
NU
SC R
IP
T
ACCEPTED MANUSCRIPT
AC
CE P
Figure 8
56
ACCEPTED MANUSCRIPT Table 1. Basic morphometric properties of the trans-Himalayan rivers. See text for abbreviations.
1
Jhelum
2
Chenab
3
Ravi
4
Beas
5
Sutlej
6
Jamuna
Syntaxial
568
-
TransHim. TransHim. TransHim. TransHim. TransHim. TransHim. TransHim. TransHim. TransHim. TransHim. TransHim. TransHim. TransHim. TransHim. TransHim. TransHim. TransHim. TransHim. Syntaxial
305
97
3.14
Rectangularlongitudinal T-shaped
184
129
1.43
95
66
151
75
306
143
7
Ganga
8
Sarda
9 10
Ghagra (Girwa) Gandak
11
Kosi
12
Tista
13
Torsa
14
Raidak
15
Sankosh
16
Manas
17
Bhareli
18
Subansiri
19
Bramhaputra
132
CE P
AC
No. of MTWs 2
5*
Very large Large
T-shaped
5*
Large
1
1.44
T-shaped
4
Large
1
2.01
Parallelogram
5+
Small
1
2.13
T-shaped
6*
Large
1
4
Small
2
5
Small
2
5
Small
2
5*
Large
4
5
Small
5
6*
Large
6
4
Small
2
4
Absent
1
4
Absent
1
5
Small
1
5
Small
4
-
SC R
Indus
ULW
NU
0
Lt/Ld Overall Shape
Highest stream order 7*
T
Ld (km)
IP
Lt (km)
MA
Type
2
114
1.16
150
1.28
D
Name
159
149
1.07
208
237
0.88
200
263
0.76
241
263
0.92
139
92
1.51
118
78
1.53
131
35
3.74
169
72
2.35
197
162
1.22
Rectangulartransverse Rectangulartransverse Rectangulartransverse Rectangularlongitudinal Rectangularlongitudinal Rectangularlongitudinal Rectangulartransverse Rectangulartransverse Rectangulartransverse Rectangulartransverse Equant
117
132
0.89
Equant
5
Absent
4
246
156
1.58
Parallelogram
6+, 5*
Large
2
373
-
-
7*
Very large
2
191
TE
Sr. No.
Rectangularlongitudinal + * Rising north of the topographic divide. Very locally developed.
57
ACCEPTED MANUSCRIPT Highlights We document some of the morphometric characters of these drainage basins, stream networks and organization of watersheds of the trans-Himalayan drainage basins.
The shapes of these basins are different and more complex than what could have been produced by head ward enlargement of branching drainage networks.
Individual basins comprise a number of geomorphologically distinct components disposed systematically with respect to the first order physiographic features of the Himalayas.
The shape of the basins embodies the effects of interplay between head ward erosion, capture and diversion and their interaction with the major mountain building processes for the last 20 million years.
AC
CE P
TE
D
MA
NU
SC R
IP
T
58