Controls on spatial variations in flow resistance along steep mountain streams

Detailed channel and water surface surveys were conducted on 15 mountain stream reaches (9 step‐pool, 5 cascade, and 1 plane‐bed) using a tripod‐mounted Light Detection and Ranging scanner and laser theodolite. Reach‐average velocities were measured at varying discharges with dye tracers and fluorometers. Multiple regressions and analysis of variance tests were used to test hypothesized correlations between Darcy‐Weisbach friction coefficient, f , and potential control variables. Gradient ( S 0 ) and relative grain submergence ( R h / D 84 ) individually explained a low proportion of the variability in f (R 2 = 0.18), where R h is hydraulic radius, D 84 is the 84th percentile of the cumulative grain size distribution, and R 2 is equal to the coefficient of determination. Because channel type, grain size, and S 0 are interrelated, we tested the hypothesis that f is highly correlated with all three of these variables or a combination of the above variables with flow period (a categorical variable) or dimensionless unit discharge ( q *). Total resistance correlated strongly (adj‐R 2 = 0.74, 0.69, and 0.64) with S 0 , flow period, wood load (volume of wood/m 2 of channel), q *, and channel type (step‐pool, cascade, plane‐bed). Total resistance differed between step‐pool and plane‐bed and between cascade and plane‐bed reaches. Significant differences in f in step‐pool and cascade reaches were found at the same values of flow and S 0 . The regression analyses indicate that discharge explains the most variability in f , followed by S 0 when discharge is similar among channel reaches, but that R h / D 84 is not an appropriate variable in these steep mountain streams to represent variations in both resistance and discharge. Results also indicate that the forms of resistance among channel types are sufficiently different to change the relationship of the control variables with f in each channel type. These results can be used to further the development of predictive equations for high‐gradient mountain streams.