MODELING THE SENSITIVITY OF CHANNEL ADJUSTMENTS IN DESTABILIZED SAND-BED RIVERS

Comprehensive empirical data of the response of unstable streams over a range of environmental conditions are unavailable. In this study, as a substitute for empirical data, a physically based numerical model o f channel evolution is used in a range of numerical simulation experim ents designed to predict the sensitivity of channel response to change s in control variables. The scope of the study is limited by the scope of the numerical model which applies to straight, sand-bed streams wi th cohesive bank materials that have been destabilized by sediment sta rvation and evolve towards equilibrium through bed degradation followe d by channel widening. Results are presented for stable and unstable c hannel conditions. Stable channel depths are most sensitive to channel discharge, though the critical threshold shear stress for the entrain ment of cohesive bank materials and discharge are both significant in determining the width. The sediment load, channel gradient, bank mater ial cohesion, size of failed bank material aggregates and the initial bank height have sensitivities an order of magnitude smaller than disc harge for both width and depth. Variations in bed material characteris tics within the sand-size range are found to have little impact on sim ulated stable channel morphology. For unstable channels, the relative dominance of parameter sensitivities is examined in the context of an empirical-conceptual model of channel evolution proposed by Thorne and Osman (1988), to highlight the relationships between parameter domina nce, time, and the processes and forms characterizing individual stage s of channel evolution. Rates of change with time of width and depth s ensitivity parameters for five tested independent variables (discharge , sediment supply, channel gradient, bank material cohesion and bed ma terial size) are found to vary as a function of time, such that differ ent stages of channel evolution are characterized by variations in the relative dominance of tested variables. The results support the hypot hesis proposed by Thorne and Osman (1988) that the critical bank heigh t required to initiate mass-wasting and widening may be regarded as a geomorphic threshold.