Downloaded from http://sp.lyellcollection.org/ by guest on September 25, 2021 Braided rivers: perspectives and problems C. S. BRISTOW 1 & J. L. BEST 2 1 Research School of Geological and Geophysical Sciences, Birkbeck College, Gower Street, London WCIE 6BT, UK 2 Department of Earth Sciences, The University, Leeds, West Yorkshire LS2 9JT, UK Abstract: Progress towards a fuller understanding of the dynamics and deposits of braided rivers demands an interdisciplinary approach to a host of unresolved problems. Although many advances have been made within recent years in interpreting the mechanics of flow, transport of sediment and sedimentary architecture of braided rivers many key issues remain to be addressed. In particular, several areas demand attention: the mechanisms of braid bar initiation; confluence-diffluence dynamics, the nature of sedimentary facies over a range of grain sizes and the influence of flow stage and aggradational regime upon the depositional architecture over a range of channel scales. This paper focuses upon these issues and highlights several areas of fruitful future interdisciplinary collaboration. Braided rivers form important topics of study for braided rivers may be fruitfully extended by many scientists and one of the primary aims of adopting such an interdisciplinary scope. this volume is to bring together work from many disciplines in an integrated approach to braided Zones of flow convergence and divergence rivers. For the geomorphologist braided fluvial systems are abundant within upland and pro- Braided rivers are characterized by 'having a glacial settings and are agents of considerable number of alluvial channels with bars and islands erosion and sediment transport. For engineers .between meeting and dividing again, and pre- the high rates of sediment transport, deposition senting from the air the intertwining effect of a and erosion combined with frequent channel braid' (Lane 1957). The division and joining of shifting and rapid bank erosion may pose con- channels are essential features of braided rivers siderable design problems both to within- and, whilst the bars within these rivers have channel structures, such as bridge piers (e.g. received attention from both geomorphologists Mosley 1982a; Sutherland 1986) and braidplain and sedimentologists, the areas of flow con- edge constructions such as roads and railways vergence and divergence have not been incor- Finally, for the geologist braided rivers form porated into braided river depositional models. important agents of deposition that have been The flow dynamics and morphology of channel responsible for the accumulation of many sedi- confluences have been studied by several mentary sequences that form valuable aquifers, researchers (e.g. Mosley 1976, 1982a; Best 1986, hydrocarbon reservoirs and sites for heavy 1987, 1988; Best & Roy 1991; Roy & Roy 1988; mineral accumulation. Because of these abun- Roy & Bergeron 1990; Roy et al. 1988), and dant and diverse applications, knowledge of the recent attention has highlighted the abundant mechanics and deposits of braided rivers is vital confluences within braided rivers (Ashmore within many areas and yet, when compared to 1982; Ashmore & Parker 1983; Klaassen & the wealth of literature upon meandering Vermeer 1988). However, the link between flow systems, they have been comparatively under- convergence and the downstream division of studied. This may, in part, be due to the diffi- flow has been neglected, despite the fact that this culty of measuring flow, sediment transport and transition is the area which may be of funda- morphology in the rapidly shifting braided river mental importance to the development of braid environment. Future progress in understanding bars (e.g. Ashmore 1991; Ashworth et al. 1992). the mechanics and morphology of braided rivers Although some depositional models of braided demands interdisciplinary collaboration and rivers are beginning to recognise and incor- calls for a more integrated approach across the porate confluence scour and fill (Cowan 1991; sciences than may have been present until com- Bristow et al. 1993; Bridge this volume; Huggen- paratively recently. This paper highlights some berger this volume; Seigenthaler & Huggen- specific areas upon which our knowledge of berger this volume), areas of flow divergence From Best, J. L. & Bristow, C. S. (eds), 1993, Braided Rivers, Geological Society Special Publication No. 75, pp. 1-H. Downloaded from http://sp.lyellcollection.org/ by guest on September 25, 2021 2 C. S. BRISTOW & J. L. BEST are less well understood and ignored in these erosional and depositional modification related models. In general, flow divergence is associated to changes in stage. At higher flow stages when with flow deceleration and sediment deposition the largest volumes of sediment are transported, and, once deposition has been initiated, the the channels are often scoured, bars may be sediment accumulation is likely to promote reduced in height or in some cases completely further flow division, deposition and bar forma- eroded. However, during falling stage maximum tion. Divergent flow may also impinge on the deposition occurs as discharge and flow com- bank at an increased angle leading to bank petence are reduced. Channel beds aggrade, the erosion, channel widening and a local increase in high stage bedforms may be modified and new available sediment, all of which are likely to lead bars may be formed or enlarged as sediment is to the development of a new braid bar (Carson deposited. As discharge continues to fall, bars 1984; Thorne et al. this volume). The deposits may become emergent and dissected by low of diffluence areas may therefore form the stage channels. Additionally, the nature of the foundations of braid bars but there are no known falling limb recession (rate and length of reces- descriptions of these deposits and their internal sion) will be important not only in the reworking structure. It is possible that some sedimentation of higher stage sediments, but also in the may occur by vertical or even upstream accre- deposition and spatial distribution of the finer tion in the diffluence areas in braided rivers grained sediments (silts and clays) which may (Ashmore this volume) and these must be mani- constitute discontinuous permeability barriers fested within the sedimentary record. Bed shear within braided alluvium. Classification of emer- stress has been shown to increase in shallow gent areas based on their low stage appearance flows over bar tops (Cheetham 1979) where a may be deceptive and care needs to be taken in coarse bed armour may form. In coarse-grained determining which areas are bars, scaling with braided rivers the bar heads may be charac- channel width, and those which are partially terized by coarse-grained sediments that are dissected bars or stranded collections of bars imbricated or laminated (Bluck 1979). There is a (Church & Jones 1982; Bridge 1985). Little data clear need for both an understanding of the fluid exists for the comparison of bar and channel dynamics of the diffiuence zone and how this morphology at different flow stages (but see may influence braid bar initiation and internal Mosley 1982b) and this is an area in which structure. controlled and correctly scaled flume models combined with field studies may contribute The influence of flow stage greatly to our understanding. The planform appearance of braided rivers can Channel hierarchies change radically with flow stage (see fig. 2 in Thorne et al. this volume). Indeed some authors The presence of a hierarchy of channels within (Doeglas 1962; Miall 1977) have proposed that braided rivers was first suggested by Williams fluctuations in discharge are a pre-requisite for & Rust (1969), who described three orders of braiding although this may often be discounted channel in addition to a series of levels within the as has been demonstrated by the modelling of river which represented active and inactive parts gravel-bed braided planforms in constant dis- of the channel system. These orders and levels of charge scaled flume experiments (e.g. Ashmore bar deposit may also be adjusted to the domi- 1982, 1991). Bluck (1979) suggests that bars may nant discharge of the alluvial system (see Thorne disappear at high flow stages, reforming as dis- et al. this volume). In the scheme proposed by charge falls, and similar observations have been Williams & Rust (1969) the entire river and reported by Smith (1974), Carson (1984) and active channels were termed the 'composite Gupta & Dutt (1989). This may imply that some stream channel' and the 'stream channel' respec- braided rivers act as single channels at bankfull tively adding two additional levels to the hier- stage and only adopt a characteristic braided archy. This system was modified by Bristow pattern on the falling stage. However, these (1987a) to a three fold hierarchy and Bridge (this observations appear to be fairly unusual and volume) suggests additional modifications to most braided rivers retain their bars at both high this view. If one accepts that the river can and low flow stage (Krigstrom 1962; Coleman operate as a single entity with channels within 1969; Smith 1970; Cant & Walker 1978; Collin- it and that there may be different scales of son 1970; Church & Jones 1982; Bridge et al. channels which depend upon total discharge and 1986; Bristow 1987a). Where bars exist for discharge fluctuations, then a threefold hier- periods of time in excess of a single flood event archy of channels is required. The first order they will experience a complex history of comprises the whole river (see fig. 2 of Thorne Downloaded from http://sp.lyellcollection.org/ by guest on September 25, 2021 PERSPECTIVES AND PROBLEMS 3 et al. this volume, for an image of the changes in discharge and discharge duration Brahmaputra River in full flood). Second order than a sand-bed river'. While a clear distinction channels are the dominant channels within the may be required when calculating sediment river whilst third order channels are primarily transport, there appears to be more morpho- low stage features which modify the bars logic similarity between sand and gravel bed deposited by the second order channels.