The Roaring Fork River at Basalt, Colorado, has a frequently mobile streambed composed of gravel, cobbles, and boulders. Recent urban and highway development on the flood plain, earlier attempts to realign and confine the channel, and flow obstructions such as bridge openings and piers have altered the hydrology, hydraulics, sediment transport, and sediment deposition areas of the Roaring Fork. Entrainment and deposition of coarse sediment on the streambed and in large alluvial bars have reduced the flood-conveying capacity of the river. Previous engineering studies have identified flood-prone areas and hazards related to inundation and high streamflow velocity, but those studies have not evaluated the potential response of the channel to discharges that entrain the coarse streambed. This study builds upon the results of earlier flood studies and identifies some potential areas of concern associated with bed-material entrainment. Cross-section surveys and simulated water-surface elevations from a previously run HEC?RAS model were used to calculate the boundary shear stress on the mean streambed, in the thalweg, and on the tops of adjacent alluvial bars for four reference streamflows. Sediment-size characteristics were determined for surficial material on the streambed, on large alluvial bars, and on a streambank. The median particle size (d50) for the streambed samples was 165 millimeters and for the alluvial bars and bank samples was 107 millimeters. Shear stresses generated by the 10-, 50-, and 100-year floods, and by a more common flow that just inundated most of the alluvial bars in the study reach were calculated at 14 of the cross sections used in the Roaring Fork River HEC?RAS model. The Shields equation was used with a Shields parameter of 0.030 to estimate the critical shear stress for entrainment of the median sediment particle size on the mean streambed, in the thalweg, and on adjacent alluvial bar surfaces at the 14 cross sections. Sediment-entrainment potential for a specific geomorphic surface was expressed as the ratio of the flood-generated boundary shear stress to the critical shear stress (to/tc) with respect to two threshold conditions. The partial entrainment threshold (to/tc=1) is the condition where the mean boundary shear stress (to) equals the critical shear stress for the median particle size (tc) at that cross section. At this threshold discharge, the d50 particle size becomes entrained, but movement of d50-size particles may be limited to a few individual particles or in a small area of the streambed surface. The complete entrainment threshold (to/tc=2) is the condition where to is twice the critical shear stress for the median particle size, the condition where complete or widespread mobilization of the d50 particle-size fraction is anticipated. Entrainment potential for a specific reference streamflow varied greatly in the downstream direction. At some cross sections, the bed or bar material was mobile, whereas at other cross sections, the bed or bar material was immobile for the same discharge. The significance of downstream variability is that sediment entrained at one cross section may be transported into, but not through, a cross section farther downstream, a situation resulting in sediment deposition and possibly progressive aggradation and loss of channel conveyance. Little or no sediment in the d50-size range is likely to be entrained or transported through much of the study reach by the bar-inundating streamflow. However, the entrainment potential at this discharge increases abruptly to more than twice the critical value, then decreases abruptly, at a series of cross sections located downstream from the Emma and Midland Avenue Bridges. Median particle-size sediment is mobile at most cross sections in the study reach during the 10-year flood; however, the bed material is immobile at cross sections just upstream from the Upper Bypass and Midland Avenue Bridges. A similar s