Predicting channel patterns

The proposed distinction between meandering and braided river channel patte rns, on the basis of bankfull specific stream power and bed material size, is analysed and rejected. Only by using regime-based estimates of channel w idths (rather than actual widths) has discrimination been achieved, and it is argued that this procedure is unacceptable. An alternative is to explore the patterning processes underlying the marked pattern scatter on bankfull stream power/bed material size plots. Of the f ive sets of patterning processes, large-scale bedform development and stabi lity is seen as especially important for meandering and braiding. For grave l-bed rivers, bedforms developed at around or above bankfull stage appear i mportant for pattern generation, with braiding relating to higher excess sh ear stress and Froude number. There seems to be an upper threshold to both meandering and braiding which is achieved at extreme discharges and steep g radients, as on steep alluvial fans, rather than for the rivers with availa ble flow data here considered. For sand-bed rivers with greater excess shea r stress, the equivalent upper plane bed threshold may occur below bankfull , with bed material mobility and bedform modification occurring over a wide r range of sub-bankfull discharges. Sand-bed channel margin outlines appear to be less perturbed by bedform effects than gravel bed planforms, and the y may have naturally straight or sinuous planforms. Bedform relief may neve rtheless lead to some being designated as braided when viewed at low flows. It is concluded that the use of a single-stage stream power measure and bed material size alone is unlikely to achieve meandering/braiding discriminat ion. (C) 2001 Elsevier Science B.V. All rights reserved.