Reprinted from

CATENA ELSEVIER

Catena 26 (1996) 85-98

Multi-channel pattems of bedrock rivers: An example from the central Narmada basin, India Vishwas S. Kale ",Victor R. Baker

Sheila Mishra

a Department of Geography, Uniuersity of Pune, Pune 411007, India Department of Geosciences, University of Arizona, Tucson, AZ 85721, USA Department of Archaeology, Deccan College, Pune 411006, India

Received 25 January 1995; accepted 21 August 1995

Abstract

Catena 26 (1996) 85-98

Multi-channel pattems of bedrock rivers: An example from the central N m a d a basin, India Vishwas S. Kale ", Victor R. Baker

b,

Sheila Mishra "

a Department of Geography, Uniuersiry of Pune, Pune 411007, India Departmenr of Geosciences, University of Arizona, Tucson. AZ 85721. USA Department of Archaeology. Deccan College. Pune 41 1006, lndia

An anomalous multiple channel pattern in bedrock is observed on a predominantly downcutting reach of the Narmada River. The multi-channel reach (800-2750 m in width and 8500 m in length) is bounded by major faults, and is underlain by granite and gneiss bedrock. Geomorphological investigations reveal differences among the upstream, middle and downstream sub-reaches of the multi-channel study area. Whereas the upstream sub-reach is dominated by deep flows and fine sediments, the lower sub-reach is characterized by a steep gradient and rapids. The middle sub-reach is the widest, and is marked by thickly forested islands and boulder berms. The characteristics of the three different sub-reaches suggest control by the interactions of lithology, flood processes and tectonics. Estimations of Hack's (1973) stream-gradient index values indicate considerable variations for the SL values along the length of Narmada River. The highest value of gradient index (SL = 797) is associated with the multi-channel reach, implying lithologic or tectonic control. Given the dimensions of the reach and its channels, it appears that the present hydrological regime is inadequate to produce the feature. We hypothesize that the multi-channel pattern development in bedrock was initiated by block or domal uplift. Enhanced gradients and extreme floods permitted the system to exploit linear weaknesses in the bedrock, leading to the development of anabranches and establishment of multiple channels in bedrock. Abrupt changes in the channel planform and morphology at the study site indicate that the river is adjusting its channel geometry (width, depth, gradient and plan configuration) to a new equilibrium channel morphology through the action of the extreme floods characteristic of this fluvial environment.

1. Introduction River channel planforms in alluvial and bedrock channels have generally been considered to be fundamentally different in character (Howard, 1980; Schumm, 1985; 0341-8162/96/$15.00 O 1996 Elsevier Science B.V. All rights resewed SSDI 0341 - 8 162(95)OOO35-6

V.S. Kale et al. /Catena 26 11996) 85-98

Ashley et al., 1988). However, the great preponderance of channel morphological studies have been conducted in alluvial channels. Single and multiple channel systems have been identified to characterize the river morphology and channel geometry of alluvial rivers (Schumm, 1985). In alluvial channels, two possible multiple channel patterns are usually recognized: braided and anastomosed (Leopold and Wolman, 1957; Fairbridge, 1968). In recent years, a clear distinction has been made between the two (Rust, 1978; Miller, 1991a). Ouchi (1985) used the term reticulate pattern to represent a widespread multiple channel network that has an angular cross-channel development. Recently Nanson and Knighton (1996) proposed a general classification of anabranching alluvial patterns, defined as systems of multiple channels separated by islands. Several other varieties of alluvial anabranching rivers are recognized as well. The term "anastomosis" has been used to designate multiple interconnected channelways whether in alluvium or bedrock. For example, Garner (1974) follows Bretz (1923) defining an anastomosing channel system as "... an erosionally developed network of channels in which the insular flow obstructions represent relict topographic highs and often consist of bedrock." In the Channeled Scabland Bretz 1923, Bretz, 1928 recognized anastomosing patterns as the result of cataclysmic flood water invading relatively small preflood valleys. Spilling of water across divides generates the anastomosing pattern as a kind of channel "overfitness" (Baker, 1978), using the terminology of Dury (1964). Relatively few studies have considered the development of river channel planforms in bedrock. Among these are works by Kirkby (1972), Shepherd and Schumm (19741, Braun (1983), Kelsey (1988), Ashley et al. (1988), Harden (1990), Seidl and Dietrich (1992), Wohl, 1992, 1993, Wohl et al. (1994), and Lorenc et al. (1994). The emphasis has been on single channels and particularly the development of meanders. As in alluvial rivers, channel pattern and associated hydraulic geometry in bedrock probably represent an equilibrium between hydrologic regime and the geologic environment (Begin and Schumm, 1984), reflecting the river's adjustment in energy expenditure (Chang, 1979). Studies of the development of meanders and multiple channel patterns in bedrock are therefore, necessary for understanding their control by bedrock characteristics as well as by the hydraulic regime. The major objectives of this paper are to report and explain the morphologic character of the multi-channel reach of the Narmada River in central India.

2. Geological and geomorphological setting The multi-cllnnel reach is located at the upstream end of the incised and structurally controlled section of the Punasa Gorge, in the central Narmada Basin (Fig. 1). The Narmada River occupies the Son-Narmada-Tapi (SONATA) lineament zone, a megatectonic feature (Kale, 1989; Brahman, 1990; Ravi Shankar, 1991). The zone consists of several longitudinal fault-bound blocks that have an episodic history of vertical and lateral movements. Strong neotectonism and moderate seismicity characterize the SONATA zone (Ravi Shankar, 1991). Fig. 2 provides an outline geological map of the central Narmada Basin. Three major faults and a number of cross faults traverse the area. The Narmada River and its main southern tributaries, the Ajnal and the

V.S. Kale et al. / Catena 26 (1996) 85-98

Fig. 1. Location map.

Machak, show strong alignments along the regional geofracture system. The alternation of structurally controlled bedrock gorges with alluvial reaches probably reflects differential movement between blocks bounded by major faults.

3. Reach characteristics Channel banks upstream and immediately downstream of the study reach are typically composed of thick Pleistocene alluvium. Just upstream of Joga Kalan (Fig. l), the river, which is aligned along a major fault, turns abruptly southward and, after cutting through a ridge of chert breccia, widens and splits into multiple channels over basement granitic gneiss and gneissic granite. Further downstream chert breccia is exposed in the channel. The maximum width of the study reach is about 2750 m, and the length is about 8500 m. Further downstream a knick point is located at the head of a 20-m deep bedrock gorge. The average gradient of this reach is 0.0012, and at rapids it exceeds 0.03. The mean sinuosity of the study reach does not exceed 1.2. However, upstream of Handia, the alluvial channel of Narmada is highly sinuous ( P = 1.3 to nearly 1.5). The gauging station located at Handia, about 25 km upstream of the study reach, recorded a peak discharge of 46,000 m3 s-' on August 19, 1984. Local inquiries reveal that while the cross-channels located on the islands are active during monsoon floods, the interchannel

V.S. Kale et al. / Catena 26 (1996) 85-98

V.S. Kale et al. /Catena 26 (1996) 85-98 ..

DECCAN TRAPS ( f l o w s & dykes)

[KATKUT

SANDSTONE FORMATION

0

MANDHATA

-

GROUP

~p

Chertbreccia

KISHANGAO GROUP

Dolorntte BARKESAR

,

---.

I

GROUP

I

I

Fig. 2. Geological map of central Narmada River Basin (after Vivek Kale, 1989).

islands are totally submerged only during very large floods. Such large floods most recently occurred in 1961, 1968, 1970 and 1984. A geomorphological map of the reach was prepared from aerial photographs (1 : 60,000), topographical maps (1 : 50,000) and Indian Remote Sensing Satellite imagery (1 : 50,000) (Fig. 3). Observations of the morphology, sedimentology, and distribution of primary and auxiliary channels were made during geomorphological investigations in the field. These studies reveal differences among the upstream, middle and downstream sub-reaches of the multi-channel study area (Table 1). Whereas the upstream sub-reach is dominated by fine sediments and deep flows, the lower sub-reach is characterized by a steep gradient and rapids. The middle sub-reach is the widest, and is marked by thickly forested islands or boulder berms (Fig. 4). The islands are composed of flood sediments or boulders with bedrock cores. Primary channels and, cross-channelclinking the primary ones, appear to be controlled by the regional lineaments at the large scale, and joints and intrusive dikes at the small scale.

4. Formation of multi-channel patterns in bedrock Given the deeply incised character of the Narmada River along most of its course (Rajaguru et al., 19951, the multi-channel reach described here is an anomaly. Given the

Fig. 3. Indian Remote Sensing satellite image (FCC) of the multi-channel study reach (February 4, 1991).

dimensions of the reach and its channels, it might seem that the present hydrological regime is inadequate to produce the feature. Similar reasoning led Garner (1966, 1967) to hypothesize that the anastomosing bedrock channels of the Rio Caron, Venezuela, were the result of wet, tropical rivers invading landscapes previously formed under arid

V.S. Kale el al. / Catena 26 (19%) 85-98

V.S. Kale et al. / Catena 26 (1 996) 85-98

Table 1 Geornomholoeical characteristics of sub-reaches Sub-reach

Number of primary channels

Maximum width (m)

Upper

5

2350

Deep channel; quiet flows; a high proportion of fine sediments Large vegetated inter-channel islands; boulder berms; incised primary and auxiliary channels

Middle Lower

Features

4

1900

Numerous rapids; high gradients; sandy bars with megaripples; abraded rock surfaces and boulder berms

I

\@

.

BHENSWARA

climatic morphogenesis. Garner (1967) describes such patterns as "rivers in the making". However, for the Narmada multi-channel patterns are not observed along other reaches of the river; some sort of local control seems responsible. The anomalous development of multiple channels in bedrock might be attributed to one or more of the three processes: (1) differential resistance to erosion, (2) flood processes, and (3) tectonic movements.

4.1. Lithological and structural controls The multi-channel reach of the Narmada River is bounded by two major faults (Fig. 2). The lithological constituents include gneissic granite and granitic gneiss that belong to Barkesar Group (Kale, 1989). These basement rocks are characterised by joints and lineaments, and are intruded by numerous dikes (10 cm to > 8 m in width). The abrupt change in channel pattern with an equally abrupt change in the rock type at the study site might suggest some sort of lithologic control. However, we do not have detailed data on rock mass strength with which to test this hypothesis. Erosion of bedrock channels requires both weathering and detachment. Field observations at the study site indicate that the basement rock is characterized by joints on which plucking erosion can operate. Fig. 4 clearly reveals that the alignments of the primary and cross channels are, to some extent, controlled by major lineaments. It appears that the system exploited and tapped linear weakness zones in the underlying granite and gneissic rocks leading to the establishment of multiple channels in bedrock. Geomorphological studi$>by Kale and Shingade (1987) indicate that multiple bedrock channels can form by coalescence of grooves and potholes along joints in basalt bedrock. The development of inner channels, beginning with longitudinal lineations and grooves is also indicated by Shepherd and Schumm's (1974) flume experiments. Our observations in the middle Krishna Basin of India have revealed an association between multi-thread and wide channels, and granitic rocks characterized by joints. Thus, litho-structural control may be important for the reach as a whole, but this factor alone does not explain the geomorphological and sedimentological differences between the three sub-reaches.

GEOMORPHIC MAP SCABLAND ........._...,... ............ COARSE SAND ...........

G R A S S E S / FINES

DENSE FOREST

DmT MAJOR FAULTS

--

-

LINEAMENTS

Fig. 4. Sketch map of the multi-channel reach shovmg ~najorgeomorphological features.

V.S. Kale er al. / Catena 26 (1996)85-98

V.S. Kale et al. / Catena 26 (1996185-98

4.2. Flood processes

The Narmada River experiences occasional high-magnitude floods. During such large floods the incised channel is overtopped at several places. Several high magnitude floods in this century (1961, 1968, 1970, 1973 and 1984) have filled the incised channel of the Narmada River. A catastrophic flood event in such a flood-prone river is likely to give rise to spectacular erosional and depositional features, including scabland, potholes, boulder berms and inner channels (Baker, 1978; Baker and Pickup, 1987; Wohl, 1992, 1993; Lorenc et a]., 1994). Evidence of high magnitude floods is present throughout the anomalous reach of the Narmada (Kale et al., 1994). Boulder berms ranging in length from 5 to > 15 m are commonly observed in the interchannel areas. Flood debris and flood scars on trees occur up to 13 m above the low water level. Similarly, scabland-like features, flute marks, polished and abraded rock surfaces, and occurrence of megaripples formed of sand (spacing = 1.5 to 3 m), suggest the effect of large floods. Although gradient is not directly determined by hydraulic regime in bedrock channels, the rate of erosion depends upon flow of water and sediment over the bed (Howard, 1980). Abrasional surfaces and flute marks demonstrate that erosion of bedrock is taking place under the present hydraulic regime. This is also supported by the presence of large boulders of local origin. The occurrence of polished and abraded rock surfaces and potholes suggest high velocities and Froude numbers close to 1 for a sustained length of time (Kale et al., 1994). Given the high flood velocities and Froude numbers at this reach, one would expect considerable erosion and sediment transportation capabilities. Flood power of the river in these floods can be estimated from relevant parameters of the study reach, including the mean gradient of 0.0012. Local slopes at the rapids reach at least 0.03, and it is these values that are most important for bedrock erosion and initiating boulder transport. The discharge of 46,000 m3 s-' (peak on record at Handia) and the varying channel width (800 to 2750 m) can be used to yield a rough estimate of the stream power per unit area w according to the formula (Baker and Costa, 1987):

where y is the specific weight of water (9800 N m-3), Q is discharge (m3s-'1, S is slope and w is the water-surface width. The estimated values range between 200 w a t t ~ m -for ~ the widest sub-reach (slope = 0.0012 and width = 2750 m) and > 7000 wattsm-2 for the lower sub-reach (for slope = 0.03 and width = 1900 m). Similarly, at the rapids the shear stress (slope-depth product) can exceed 3000 N m-2. These values indicate the l&h competence of the river during extreme floods, particularly in the lower sub-reach. The values are comparable to those noted on other rivers displaying spectacular bedrock erosion (Baker and Pickup, 1987; Rathburn, 1993; O'Connor, 1993). Baker and Pickup (1987) have described the occurrence of a series of anastomosing channels and scour pools in upper Katherine Gorge, Australia. Because of morphologic similarity to flood erosional forms in the Channel Scabland of Washington, they have attributed these features to extreme flood flows. Longitudinal grooves and shallow linear depressions parallel to flows, in Piccaninny Creek in northwestern Australia were

inferred by Wohl (1993) to result from longitudinal vortices and turbulent vortices during high-magnitude flood flows. While cataclysmic flood erosion seems clearly to be occurring on the lower sub-reach, the upper sub-reach remains problematic because of its association with fine sediments and a low proportion of rocky channels and scabland areas. Thus, flood processes alone do not seem sufficient to explain the multi-channel pattern. 4.3. Tectonic processes Fluvial anomalies, such as local changes in channel patterns and local widening or narrowing of channels are possible indicators of active tectonics r reg or^ and Schumm, 1987). Channel patterns, which constitute a fourth degree of freedom in various equilibrium schemes of river behaviour (Chang, 1979), are sensitive indicators of valley-slope changes (Schumm, 1986). The gradient of bedrock channels is a semi-independent variable and is not directly determined by the hydraulic regime (Howard, 1980). Factors including the physical characteristics of rocks and tectonic movements determine the channel gradient of the bedrock channels. In alluvial channels, with increasing shear stress (slope-depth product) high sinuosity streams transform to braided streams (Begin and Schumm, 1984). Therefore, even minor changes in the gradient of large rivers, like Narmada, are likely to upset the balance between process and form, and both the extrinsic and intrinsic geomorphic thresholds (Schumm, 1987) will be exceeded. Hack (1973) defined a stream-gradient index (SL) which is the product of slope of a reach times the length from the headwaters divide. Variations in gradient index reflect the longitudinal variations in discharge, but more commonly the lithologic or tectonic controls on channel slope of a given reach (Hack, 1973; Bull and Knuepfer, 1987). Fig. 5 shows that the SL indices for the Narmada River range between 71 and 797. The high values of the index (SL = 797) for the reach between Handia and Punasa suggest that this may be the most actively downcutting part of the Narmada River, followed by another reach downstream of Rajghat (SL = 774), implying tectonic or lithologic control. The abrupt changes in the gradient and SL values at the study site can be related to both changing bed resistance (Miller, 1991b) and tectonic uplift (Wohl et al., 1994). Large rivers, owing to their relatively low channel gradients, are most significantly affected by the minor changes in slope induced by active deformation (Schumm, 1986). The Narmada River, with a very large discharge and an average gradient of 0.00072 is likely to be significantly affected by any tectonically induced increase in the channel gradient. Theoretical calculations based on average width of the undivided reach (735 m) and peak discharge at Handia (46,000 m3 s-') indicate that, at very low slopes, even very small amounts of tectonic uplift in the river bed level will so increase the slope as to result ir major changes in the power per unit area of bed. A local rise in bed level by only 20 cm, for example, could raise the power per unit area by a factor of 30. At the present site, this effect will be, obviously, downstream of the axis of uplift. Experimental studies and field observations confirm that a change of valley-slope will cause a change of channel pattern and dimensions (Schumm, 1986). Ouchi (1985), on the basis of experimental studies, inferred that domal uplift across a meandering river

V.S. Kale et al. / Catena 26 (1996) 85-98

NARMADA RIVER

P = Punasa

R = Rajghat

V.S. Kale et al. / Catena 26 (1996) 85-98

offsets increase in the hydraulic depth caused by incision, and allows an initial increase in specific energy due to an increase in gradient to decrease with time (Simon, 1992). It, therefore appears that under the given water and sediment discharge conditions as well as channel perimeter lithology, a catastrophically disturbed bedrock channel will establish a width, depth, gradient and plan configuration that results in a minimum stream power (Chang, 1979; Simon, 1992) or stream power per unit length (Yang and Song, 1979). Therefore, the formation of multiple channels in bedrock and associated local increase in channel width and channel capacity for the Narmada River might represent an important means of reducing specific energy, under high-energy conditions induced by block or domal uplift. The association with the major faults and the fact that strong neotectonism and moderate seismicity characterise the Son-Narmada-Tapi (SONATA) lineament zone (Ravi Shankar, 1991) strongly imply tectonic control on the SL values as well as on the formation of multi-channel pattern at the study site.

5. A hypothesis for the development of multi-channel bedrock patterns DISTANC F R O M SOURCE I N K M

Fig. 5. Profile of the Narmada River showing values of Hack's (1973) strearn-gradient index (SL) in gradient-meters.

leads to the development of reticulate or anastomosed channel patterns. According to him, on the upstream side of the uplift, as a result of damming effects, there will be inundation of flood-plain and channel avulsions, leading to the development of reticulate or, in some cases, anastomosing channel pattern. Anastomosing channel patterns observed along downcutting streams in south-central Indiana were ascribed to avulsion by Miller (1991a). Burnett and Schumm (1983), Schumm (1986) and Gregory and Schumm (1987) described examples illustrating the development of a braided pattern downstream of the rise and anastomosing pattern upstream of the rise, as a result of uplift and consequent reduction of valley-slope above the axis of uplift. Braided channels are the result of high bed-load transport on steep gradients with high width-to-depth ratios and erodible banks (Schumm, 1986; Miller, 1991b). At the present site, although the gradients are steeper than upstream, the banks are more resistant with the channel cut into bedrock. Further, estimates based on hydraulic modeling suggest that the annual peak floods on Narmada are competent to move the bed sediments up to 2.5 m in intermediae-axis (Kale et al., 1994). Such a condition would not favour the development of a braided channel pattern for the bed sediments are entirely reworked. The raising of channel bed increases the specific energy (sum of pressure and velocity heads). At the present site, the raising of the channel bed will increase the specific energy downstream but will decrease upstream. Channel widening in conjunction with incision, concentrated at locations of maximum boundary shear stress and stream power per unit area, is the most effective mechanism of energy dissipation in a catastrophically disturbed system (Simon, 1992). In other words, channel widening

The high SL index values and the occurrence of major faults up and downstream of the study reach, the local setting of the anomalous reach and the tectonic history of the area all suggest that the multi-channel pattern development was initiated by block or domal uplift. Neotectonic movement of the fault-bound block would have raised the bed level of the Narmada River, leading to a damming response in the upstream sub-reach (Fig. 6). This is indicated by the occurrence of deep water, quiet flows and predominance of fine-grained sediment in the upper sub-reach. The tectonic uplift of the channel bed increases the channel gradient as well as the stream power, greatly increasing the energy conditions downstream of the uplift. It appears that the channel responded to this catastrophic disturbance by channel widening and incision, allowing specific energy to decrease with time. This is an interesting contrast to anabranches in alluvial channels

UPLIFTED BLOCK I

DEEP CHANNEL

............. KNlCK

CHERT BRECCIA

'.

\

@ = PUNGHAT

/

- FAULT I

@) = BHENSWARA

Fig. 6. Sketch of relationships between Narmada River profile and structural/tectonic features of the study reach.

V.S. Kale er a[./ Catena 26 f 1996) 85-98

which appear to locally increase unit stream power and enhance sediment through put (Nanson and Knighton, 1996). In the downstream sub-reach of the study site, increased gradients would have resulted in slope adjustment by degradation. This is suggested by steeper channel slopes and numerous rapids in the lower sub-reach (Fig. 6). The knick point and a deep gorge located downstream of the multi-channel reach, provide evidence of intense high-gradient conditions. With a rise in the bed level, overbank flows during extreme floods, devoid of significant bedload and having sufficient energy to erode, must have facilitated rapid incision (particularly below the uplift) along deeply weathered joints and weak zones (lineaments) in the underlying granite and gneissic rocks. This led to the development of anabranches and establishment of the multiple channels in bedrock. Thus, a reduction of specific energy was achieved by channel widening and by degradation in bedrock. The lithology of the granite and its fracture system favoured the development, under the influence of occasional intense floods, of the multi-channel pattern. Lack of disruption for the Narmada drainage, diverting it to new locations, indicates a relatively slow uplift of the fault-bounded block.

6. Conclusion A multi-channel pattern occurring on the regionally downcutting Narmada River indicates probable local effects of tectonic deformation and domal uplift, as well as the influence of structure and flood processes. The Narmada River, which throughout its course is deeply incised in bedrock or alluvium, deviates from its general pattern and suddenly widens its channel at the study site, giving rise to a multi-thread pattern in rock. Large rivers, like the Narmada, owing to their relatively low channel gradients, are likely to be significantly affected by any tectonically induced increase in the channel gradient. The raising of the channel bed increases the specific energy downstream of the uplift. Using the concept of minimization of stream power, Simon (1992) argues that degradation accompanied by widening is the most efficient means of fluvial energy dissipation. Anabranching usually concentrates stream power (Nanson and Knighton, 1996) and braiding disperses it across a wider channel width. The Narmada appears to be minimizing the stream power by adjusting its channel geometry (width, depth, gradient and plan configuration). It is generally accepted that downcutting streams are disequilibrium systems (Bull and Knuepfer, 1987; Miller, 1991b) and, hence, a distinctive channel geometry determined by the local geologic and tectonic controls implies that the channel is attempting to attain an equilibrium channel-bed form. The development of the multi-channel pattern, caused by a variable external to the system (tectonism), provides an excellent example of a river near its pattern threshold (Schumm, 1987). The response is. broadly analogous to that occurring in alluvial rivers, but the conditions of flood energy and channel boundary properties (lithology) are more extreme, resulting in different timescale for the equilibrium adjustment. Acknowledgements This paper is contribution No. 34 of the Arizona Laboratory for Palaeohydrological and Hydroclimatological Analysis (ALPHA), University of Arizona. The research was

V.S. Kale et al. / Catena 26 (1996) 85-98

97

supported by Indian Department of Science and Technology Grant ESS/CA/A3-04/90 to V.S.K. and S.M. The authors thank S.N. Rajaguru, Y. Enzel, L. Ely, Avijit Gupta and Vivek Kale for fruitful discussions. Reviews by G.C. Nanson and E.E. Wohl were very helpful in preparing the final version of this paper.

References Ashley, G.M., Renwich, W.H. and Haag, G.H., 1988. Channel form and processes in bedrock and alluvial reaches of the Raritan River, New Jersey. Geology, 16: 436-439. Baker, V.R., 1978. Large-scale erosional and depositional features of the Channeled Scabland. In: V.R. Baker and D. Nummedal (Editors), The Channeled Scabland. National Aeronautics and Space Administration, Washington, DC, pp. 8 1-1 15. Baker, V.R. and Costa, J.E. 1987. Flood power. In: L. Mayer and D. Nash (Editors), Catastrophic Flooding. Allen and Unwin, London, pp. 1-21. Baker, V.R. and Pickup, G., 1987. Flood geomorphology of the Katherine Gorge, Northern Territory, Australia. Geol. Soc. Am. Bull., 98: 635-646. Begin, Z.A. and Schumm, S.A., 1984. Gradational thresholds and landform singularity: Significance for Quaternary studies. Quat. Res., 21: 267-274. Brahman, N.K., 1990. Seismic hazard at Narmada Sagar Dam. Current Sci., 59: 1209-1211. Braun, D.D., 1983. Lithologic control of bedrock meander dimensions in the Appalachian valley and ridge province. Earth Surf. Process. Landforms, 8: 223-237. Bretz, J.H., 1923. The Channeled Scabland of the Columbia Plateau. J. Geol., 31: 617-649. Bretz, J.H., 1928. The Channeled Scabland of eastern Washington. Geogr. Rev., 18: 446-477. Bull, W.B. and Knuepfer, L.K., 1987. Adjustment by the Chanvell River, New Zealand, to uplift and climatic changes. Geomorphology, 1: 15-32. Burnett, A.W. and Schumm, S.A., 1983. Alluvial-river response to neotectonic deformation in Louisiana and Mississippi. Science, 222: 49-50. Chang, H.H., 1979. Minimum stream power and river channel patterns. J. Hydrol., 41: 303-327. Duly, G.H., 1964. Principles of underfit streams. U.S. Geological Survey Professional Paper, 452A. pp. 1-67. Fairbridge, R.W. (Editor), 1968. Encyclopedia of Geomorphology. Reinhold, New York, pp. 90-93. Gamer, H.F., 1966. The derangement of the Rio Caron, Venezuela. Rev. Gomorphol. Dyn., 2: 50-83. Gamer, H.F., 1967. Rivers in the making. Sci. Am., 216: 84-94. Gamer, H.F., 1974. The Origin of Landscapes. Oxford University Press, New York, 734 pp. Gregory, D.I. and Schumm, S.A., 1987. The effect of active tectonics on alluvial channel morphology. In: K. Richards (Editor), River Channels, Environment and Process. Blackwell, Oxford, pp. 41-68. Hack, J.T., 1973. Stream-profile analysis and stream gradient index. U.S. Geol. Surv. J. Res., 1: 421-429. Harden, D.R., 1990. Controlling factors in the distribution and development of incised meanders in the central Colorado Plateau. Geol. Soc. Am. Bull., 102: 233-242. Howard, A.D., 1980. Thresholds in river regimes. In: D.R. Coates and J.D. Vitek (Editors), Thresholds in Geomorphology. Allen and Unwin, London, pp. 227-259. Kale, Vivek S., 1989. Significance of Riphean stromatolites from the Kishangad ( f Bijawar) group, Dhar Forest inlier, Central Narmada Valley. Himalayan Geol., 13: 63-74. Kale, V.S. and Shingade, B.S., 1987. A morphological study of potholes of Indrayani Knick Point, Maharashtra. In: V.S. Datye et al. (Editors), Explorations in Tropics. Professor K.R. Dikshit Felicitation Volume. Professor K.R. Dikshit Felicitation Committee, Pune, pp. 206-214. Kale, V.S., biy, L.L. Enzel, Y. and Baker, V.R., 1994. Geomorphic and hydrologic aspects of floods on the Narmada and Tapi Rivers in central India. Geomorphology, 10: 157-168. Kelsey, H.M., 1988. Formation of inner gorges. Catena, 15: 433-458. Kirkby, M.J., 1972. Alluvial and nonalluvial meanders. Area, 4: 284-288. Leopold, L.B. and Wolman, M.G., 1957. River channel patterns: Braided, meandering and straight. U.S. Geological Survey Professional Paper, 282B.

V.S. Kale et al./ Catena 26 (1996) 85-98 Lorenc, M.W., Barco, P.M. and Saavedra, J., 1994. The evolution of potholes in granite bedrock, W Spain. Catena, 22: 265-274. Miller, J.R., 1991a . Development of anastomosing channels in south-central Indiana, Geomorphology, 4: 22 1-229. Miller, J.R., 1991b. The influence of bedrock geology on knick-point development and channel-bed degradation along down-cutting streams in south-central Indiana. J. Geol., 99: 591-605. Nanson, G.C. and Kinghton, D.A., 1996. Anabranching rivers: their causes character and classification. Earth Surf. Process. Landforms, in press. O'Connor, J.E. 1993. Hydrology, hydraulics, and geomorphology of the Bonneville Flood. Geological Society of America Special Paper 274, pp. 1-83. Ouchi, S., 1985. Response of alluvial rivers to slow active tectonic movement. Geol. Soc. Am. Bull., 96: 504-515. Rajaguru, S.N., Gupta, A., Kale, V.S., Ganjoo, R., Ely, L., Enzel, Y. and Baker, V.R., 1995. Channel form and processes of the flood-dominated Narmada River, India. Earth Surf. Process. Landforms, 20: 407-421. Rathburn, S.L., 1993. Pleistocene cataclysmic flooding along the Big Lost River, east central Idaho. Geomorphology, 8: 305-319. Ravi Shankar, 1991. Thermal and crustal structures of SONATA: A zone of mid continental rifting in the Indian Shield. J. Geol. Soc. India, 37: 21 1-220. Rust, B.R., 1978. A classification of alluvial channel system. In: A.D. Miall (Editor), Fluvial sedimentology. Can. Soc. Pet. Geol. Mem., 5: 187-198. Seidl, M.A. and Dietrich, W.E., 1992. The problem of channel erosion into bedrock. Catena Suppl., 23: 101-124. Schumm, S.A., 1985. Patterns of alluvial rivers. Annu. Rev. Earth Planet. Sci., 13: 527. Schumm, S.A., 1986. Alluvial river response to active tectonics. In: Active Tectonics. Geophysics Study Committee, National Academy Press, Washington, DC, pp. 80-94. Schumm, S.A., 1987. Some applications of the concept of geomorphic thresholds. In: D.R. Coates and J.D. Vitek (Editors), Thresholds in Geomorphology. Allen and Unwin, London, pp. 473-486. Shepherd, R.G. and Schumm, S.A., 1974. Experimental study of river incision. Geol. Soc. Am. Bull., 85: 257-268. Simon, A., 1992. Energy, time and channel evolution in catastrophically disturbed fluvial systems. Geomorphology, 5: 345-372. Wohl, E.E., 1992. Bedrock benches and boulder bars: Floods in the Burdekin Gorge of Australia. Geol. Soc. Am. Bull., 104: 770-778. Wohl, E.E., 1993. Bedrock channel incision along Piccaninny Creek, Australia. J. Geol., 101: 749-761. Wohl, E.E., Greenbaum, N., Schick, A.P. and Baker, V.R., 1994. Controls on bedrock channel incision along Nahal Paran, Israel. Earth S u d Process. Landforms, 19: 1-13. Yang, C.T. and Song, C.C.S., 1979. Theory of minimum rate of energy dissipation. J. Hydraul. Div. ASCE, 105(HY7): 769-784.