| Abstract: | This paper investigates variability in bedload transport and channel morphology for 11 replicate experimental runs in an approximately 1:50 braided river model. The experiments, each of 90 h duration, were carried out in a 20 × 3m tilting flume. All the experiments started with the same initial conditions. Bedload transport was measured at 5 min intervals in a collection drum at the exit from the flume. The model showed reasonable hydraulic similarity when compared to prototype rivers. Results show that mean bedload transport rates for the 11 runs vary in the range 0·98 to 1·49gs?1 (mean + 1·21, coefficient of variation 11 per cent). Within-run transport rates commonly vary from close to zero, to two and occasionally three or four times the mean rate. Within the bedload series, several irregular phases of transport intensity can be observed, but time series analysis of the data show little underlying serial structure (an AR(2) autoregressive model is appropriate). Channel patterns are narrow/braided, are established quickly and remain relatively stable throughout the runs, although channel widths increase between 20 and 103 per cent over the 11 runs. Channel behaviour varies from aggradational to transitional between aggradation and degradation. Time-averaged bedload transport rate is weakly correlated with braiding intensity. In general, these results demonstrate that for a given set of controlling variables, bedload transport and channel morphology can be approximately replicated. |
