Channel Adjustment and a Test of Rational Regime Theory in a Proglacial Braided Stream

The upstream reach of the Sunwapta River, Alberta, provides a useful quasi-experimental field case of channel adjustment in a proglacial stream. Historically, the formation of a proglacial lake deprived the river of its coarse sediment supply for several decades and lead to a dramatic decrease in braiding intensity close to the lake while braiding intensity increased further downstream. This response to the reduction of gravel input is consistent with previous experimental results. Subsequent construction activity and channelization close to the lake have contributed to the continuation of these temporal and spatial trends in channel pattern. The current state of adjustment of the river morphology can be explained, in the context of these historical changes, using rational regime equations. The study reach has no tributaries and bed material size decreases twofold along the reach while width and braiding intensity increase, yet channel slope decreases by only 10%. The absence of any significant change in discharge downstream along the reach allows testing of regime equations under conditions in which discharge is held constant. The current downstream trends in slope and fluctuations in width are predicted reliably from rational regime equations, but not by the existing empirical hydraulic geometry relations. The rational equations incorporate the effect of grain size and slope on channel width and the effect of width and grain size on channel slope. The regime equations are successful even though they were devised for single channel gravel streams. The small (10%) decrease in slope along the reach, despite a halving of median grain size, is attributed to the counteracting (positive) effect on slope of the downstream increase in braiding intensity and width. The downstream increase in braiding intensity must be largely the result of decreasing grain size. This confirms the influence of grain size on channel pattern thresholds and demonstrates, using spatial transitions in channel pattern, that channel pattern predictions based on stream power alone are inadequate