Characterizing flow resistance in high gradient mountain streams, fraser experimental forest, Colorado.

摘要

High gradient mountain streams dissipate energy when water flows over poorly sorted grains in the bed and banks and over bedforms such as steps and pools, creating a constant alternation between supercritical and subcritical flow and causing energy dissipation through hydraulic jumps. Mountain streams (bed slope ranging between 0.02 and 0.19) differ from their low gradient counterparts by having large boulders that are of the same order of magnitude as the depth of flow, low values of relative grain submergence (Rh/D84, where Rh is hydraulic radius and D84 is the 84th percentile of the cumulative grain-size distribution), armored beds, and wood that commonly spans the entire width of the channel. The complex interaction of the different forms of flow resistance in steep mountain streams has made it particularly challenging to quantify flow resistance, usually represented by the dimensionless Darcy-Weisbach friction factor (ff). This research focuses on studying controls and interactions among different forms of resistance in step-pool, cascade, and plane-bed reaches on two different streams, where a reach is a length of channel 100-101 m in length with consistent channel morphology. The project is divided into three parts: (1) identify specific controls on the total flow resistance throughout the channel network using statistical analysis; (2) investigate specific variations and controls in relation to stage within each reach by analyzing at-a-station hydraulic geometry; and (3) quantify and evaluate interactions among the individual flow resistance components that contribute to total flow resistance.;Detailed channel and water surface surveys were conducted on 15 mountain stream reaches (nine step-pool channels, five cascade channels, and one plane-bed channel) using a tripod-mounted Light Detection and Ranging (LiDAR) scanner and laser theodolite. Reach-average velocities were measured at varying discharges with dye tracers and fluorometers. Results indicate that gradient is a dominant control for both total ff and the individual components of ff, which were divided into grain (ffgrain), form (ffstep), wood (ffwood), and spill resistance (ffspill). A second strong control on values of ff was discharge, with values of ff decreasing with increasing discharge. Spill and form resistance contributed the greatest amount towards total ff at low flows, whereas wood contributed a larger proportion at high discharges. The contribution of grain resistance was small at all flows, but generally decreased with increasing discharge. Methods for calculating the components of resistance were found to have large sources of error. Grain resistance was typically under-estimated at lower discharges, because methods assuming a semi-logarithmic velocity profile become invalid at base flows. A new method of calculating grain resistance is suggested for lower flows, by dividing the characteristic grain size between those elements that protrude above the water surface (D90) and those that are still submerged (D50).;Methods for calculating wood resistance were also found to have high sources of error and cause the values of ffwood to be overestimated. An attempt is made to calculate form resistance created by adverse pressure gradients around the step bedforms at high flows. Commonly, this effect is ignored in favor of lumping the remaining component of resistance into spill resistance. Although spill resistance still made up the largest amount of the total at the lowest flows, ffstep made a significant contribution at bank-filling discharges and further work in the flume and field needs to be done to understand the contribution of form drag around steps. Interactions between components of resistance also indicate that an additive method of resistance partitioning is not appropriate in these higher gradient streams.;Wood significantly affected the values of flow resistance throughout each channel type. The presence of wood increased resistance within each reach. Steps with wood are significantly wider and have greater drop heights than boulder steps. Wood also was significantly related to grain resistance, causing values of ffgrain to be smaller than in reaches without wood. The increase in resistance from wood, as well as the larger steps, caused reduced velocity, increased depth and therefore decreased ffgrain.;The detailed analysis of these high gradient reaches shows the large amount of complexity inherent in these channel types, which makes developing predictive equations of ff difficult. This analysis was undertaken to better understand the complexity and to help in determining appropriate methods for calculating ff. The dominance of gradient as a control on both total ff and its components is useful to understand because this is a metric that can be used to remotely predict these characteristics, as the resolution of remote data improves with time.