Twenty-four gravel-bed rivers in the Rocky Mountain region of Colorado were selected for a detailed investigation of bed-material mobility and hydraulic geometry. The criteria for choosing the study reaches were: a nonbraided channel with self-formed bed and banks, evidence of quasi-equilibrium, minimal flow regulation, and a streamflow-gaging station record of at least ten years. In each river reach, several cross sections and longitudinal profiles of the riverbed, water surface, and bankfull elevation were surveyed. Samples of bed surface and subsurface material were collected for size analysis. The median diameter of the bed-material surface ranged from 23 mm to 120 mm. Bankfull discharges ranged from 0.70 m3 /sec to 255 m3/sec. Computed values of the dimensionless shear stress (τ*) showed that the threshold of particle motion was exceeded at flows slightly less than bankfull. The average value of τ* at bankfull was 0.046. Bankfull discharges were equaled or exceeded on an average of 8.1 days per year. The averaged computed value of τ* for the 100-yr flood discharges was only 0.068. Consequently, transport of bed-material particles is a relatively frequent occurrence. Large bed-material transport rates, however, are extremely rare. Empirical evidence obtained at these and other streams indicates that a stable, self-formed gravel-bed channel cannot be maintained at a dimensionless shear stress greater than about 0.080.
The 24 river reaches were divided into 2 groups, depending upon bank stability as indicated by bank vegetation. Separately determined bankfull hydraulic-geometry equations for each group were found to have the same exponents, but slightly different coefficients. An analysis of variance determined that 67% of the observations were within ± 15% to 20% of the predicted values for most hydraulic variables. The bankfull hydraulic-geometry equations for the Colorado rivers studied were compared with those of British gravel-bed rivers that had been similarly divided according to the extent of bank vegetation. No significant difference between the hydraulic geometries of Colorado and British gravel-bed rivers with thick bank vegetation was found. The comparison of hydraulic-geometry equations for those rivers with thin bank vegetation determined that there was no difference in the width-versus-discharge relations. No significant difference was found for the exponents of the depth, velocity, and slope equations; however, the coefficient values were slightly different.