The ability to assess flow hydraulics and the movement of substrata is important to understanding the complexity of habitats and variation in spatial and temporal conditions in stream and river ecosystems. The concept of flow competence (i.e., flow necessary to mobilize the streambed) has been applied to analyses of streambed stability and generally is based on the assumption that measures of rock size composing the substrata can be used to estimate the magnitude of flow that would mobilize the bed material. However, analyses have had limited success owing in part to the expectation that simple linear expressions accurately predict threshold entraintment (i.e., initiation of bed movement) over a broad range of hydraulic conditions from low-gradient sand-bed rivers through high-gradient boulder-bed rivers. Examination of the literature revealed ambiguity of terminology and use of equations without the proper adherence to limitations and assumptions inherent in the derivation of the equations. We present the derivations of the founding equations of flow competence, and discuss their applications, assumptions, and limitations. We also examine the efficacy of a stream stability index based on the ratio of estimated bottom boundary shear stress applied to the bed during bankfull conditions, and the critical entrainment shear stress calculated from the size of the surface bed material. Data from 33 gravel- to boulder-bed river sites were used to evaluate the stream stability index as a predictor of bed movement. The stability index indicated that a bankfull discharge would not generate critical threshold entrainment for 28 of the 33 sites, with most sites requiring a 2-fold increase in bottom boundary shear stress to mobilize the riverbed gravel and cobbles, and nearly an order of magnitude increase in shear stress to mobilize boulders. We conclude that other factors affecting the possible range of variation in flow velocity, particle size distributions, bed packing, and momentum exchange from collisions between saltating particles and those at rest on the bed may lead to departures between predicted and actual streambed movement. Hence, flow-competence approaches to predicting streambed instability should be used only to objectively bracket ranges of variation in disturbance and streambed movement. Moreover, the assumptions and limitations of equations for such indices should be investigated carefully before elaborate statistical analysis of watershed or streambed variables are undertaken to explain deviations from theoretical estimates, or to explain why a particular index does or does not predict bed movement.
