Relationship of bed and bank resistance to total flow resistance in a high gradient stream, Fraser Experimental Forest, Colorado, USA

The relative influence of bank resistance versus bed resistance on the total flow resistance in a high gradient system has been a source of debate, but because of the difficulty of making measurements in these systems has rarely been explored. Nine step-pool and five cascade reaches were surveyed over five stages in Fraser Experimental Forest, Colorado, using a combination of a laser theodolite and LiDAR (Light Detection and Ranging). The LiDAR was used to capture banks and channel geometry at low flows, whereas the water surface and bed data were collected with the laser theodolite at both low and high flows. Reach-averaged mean velocity was measured using fluorometers and Rhodamine WT dye tracers. The roughness created by the bed and banks were calculated using the standard deviation of the residuals of a longitudinal profile regression (σbank and σbed). In addition, hyperspectral and spectral analysis are used to characterize the bed, banks, and water surface during both base and bankfull flows. The three methods of characterizing the roughness created by the bed and banks are used to ascertain relative influence of the banks versus the bed on total flow resistance. These data are used to evaluate three hypotheses: i) step-pool bed morphology has a distinctively different spectral signature than cascade bed morphology; ii) the bed and bank resistance can be evaluated using either standard deviation of the residuals or by hyperspectral analysis and related to total flow resistance; iii) potential controls on bank resistance include both gradient and grain size, which differ based on channel type (step-pool versus cascade). Preliminary results indicate that both step-pool and cascade channel morphologies can be evaluated using spectral analysis despite the short reach lengths (≤ 30 m). Evaluation of σbed indicated that there is no significant difference between means for step-pool (0.145) versus cascade (0.163) reaches, but the boxplots show a greater variability in the values of σbed for step-pool reaches. Furthermore, there was no significant difference in the value of σbank for step-pool (0.656) versus cascade (0.692) reaches. Both gradient and grain size are significant controls of σbank. The larger grain sizes led to higher values of σbank, most likely from larger boulders forcing flows towards the banks creating an irregular pattern of erosion. At bank filling flows the bank resistance was found to be significantly related to total flow resistance, with flow resistance increasing as σbank increased.