Journal of Hydraulic Engineering

Several studies have investigated the use of the bottom tracking (BT) mode of acoustic Doppler current profilers (ADCPs) for evaluating bedload transport. The raw apparent bedload velocity is usually noisy and contains erroneous data. This study investigates how bedload dynamics influence acoustic processes occurring at riverbeds (i.e., volume and roughness scattering). The accuracy of ADCP apparent bedload velocity measurements is analyzed in two sets of laboratory experiments using two ADCPs working at different frequencies [2 MHz RDI StreamPro (RDI Teledyne Marine, Cypress, Texas) and 3 MHz SonTek M9 (Sontek/Xylem, San Diego)], with a variety of sediment materials and different hydraulic conditions. Simultaneously, the velocity and surface concentration of the mobile sediments are measured using high-resolution cameras. Despiking and filtering are applied to the raw data, and the temporal average of the apparent bedload velocity is spatially normalized. The percentage of filtered erroneous velocity data from the ADCP time series demonstrates a strong correlation with the surface concentration of mobile particles. Velocities measured with the M9 matched the particle velocities measured by image velocimetry better than those measured with the StreamPro, which appeared to underestimate the bedload velocity by a factor of 2–4. This suggests that instruments with different acoustic frequency yield a different interpretation of the apparent velocity; instruments with lower acoustic frequency and larger acoustic sampling length are more affected by the fixed surface beneath the layer of moving particles. These results bear out the notion that filtered apparent bedload velocity can be used to estimate the spatial velocity of bedloads, but its dependence on a set of acoustic properties must be further investigated.

A new empirical equation for alternate bar height is introduced. It is assumed that the bars are formed under a steady and uniform flow; the stage of interest is that where bars have grown to their fully developed state. The equation is developed on the basis of dimensional considerations and all data available to the authors; the formulation also incorporates findings by stability analysis. The data result from a total of 191 flume experiments reported in 16 different works carried out from 1961 to 2014 using either sand or gravel as bed material. Both fully rough and transitionally rough flows were used in the experiments. Overall, width-to-depth ratios and relative depths varied from 3.5 to 54.4 and 3.8 to 191, respectively; the ratios of bed shear stress to critical bed shear stress ranged from 1.1 to 14.7. Data from the Naka River, Japan, are also used. The equation correlates bar height, normalized by flow depth, with the excess width-to-depth ratio ($B/h$) with regard to the smallest, or critical, value of $B/h$ at which alternate bars occur as well as with the relative depth and, in the case of transitionally rough flows, also the grain-size Reynolds number. It is shown that when compared with previous empirical equations, the present equation produces a considerably improved overall alignment of the data with the perfect agreement line and a significant larger (nearly double) percentage of data falling within the 20% error range. The results by the present equation are also substantially more congruent with those derived from existing theoretical and numerical analyses of the development of alternate bars, and more specifically, those resulting from a stability analysis of the phenomenon. As a by-product of this work, a first comparative evaluation of existing equations for alternate bar height is also presented.

Designing efficient rock-ramp fish passes with flows over a bottom with roughness on the same scale as the water depth requires a precise knowledge of hydrodynamics in order to avoid or limit characteristics unattractive for fish, particularly for small fish. This paper considered the numerical modeling of free-surface flow across a steep-sloped ramp covered with staggered surface emergent cylinders. Considering the importance of complex flow features for fish passage, computational fluid dynamics (CFD) was adopted because it is capable of predicting such features. Because of the longitudinal periodicity of the arrangement of the obstacles, cyclic boundary conditions made this fine simulation possible. Two computational meshes (coarse and fine) and two turbulence models [shear stress transport (SST) $k\text{-}\omega$ and Smagorinsky large-eddy simulation (LES)] were used. The SST $k\text{-}\omega$ coarse mesh model gives correct time-averaged values, the main flow unstationarities and is usable for rock-ramp fish pass design, but a fine model using LES turbulence closure can provide detailed flow characteristics in the wakes in order to provide possible rest zones, particularly for smaller fish.

The catastrophic failure of the spillway chute at Oroville Dam in February 2017 raised concerns throughout the water resources industry regarding design, construction, and maintenance practices for concrete spillway chutes, especially joints and cracks that could allow penetration of high-pressure water into a chute foundation. The independent forensic team investigation found that hydraulic jacking was the most likely cause of the initial chute slab failure, highlighting a need for better analysis of the hydraulic jacking potential of existing spillways and more resilient designs for spillways that operate under high-velocity flow conditions. This paper reviews the Oroville Dam event and findings and previous laboratory testing performed to evaluate uplift pressures and flow transmitted through spillway joints. A reanalysis of previous studies was used to develop relations between chute velocity, joint geometry, and uplift pressure transmitted into a joint. Uplift pressure head in these relations is expressed in a dimensionless manner, either as a percentage of the velocity head in the boundary layer at midheight of the offset into the flow, or as a percentage of the channel-average velocity head. The first approach is potentially more useful for prototype applications, but the second method provides the best fit to the available experimental data. Additional research is still needed to quantify rates of flow through open joints, confirm relations between uplift pressure and boundary layer velocities, and evaluate the effects of aerated flow.

The feasibility of using the velocity-based method for calculating the transverse mixing coefficient of the two-dimensional contaminant transport model was studied to substitute the concentration-based method in which the mixing coefficient is calculated from the concentration curves obtained via the costly tracer experiment. To calculate the transverse mixing coefficients, the hydraulic data and concentration data of the electric conductivity (EC) were collected at the confluence of tributary rivers in the Nakdong River, South Korea. The experimental results showed that the transverse mixing was controlled by secondary currents that were caused by the merging of the tributary rivers as well as the meandering of the main channel in the study area. The comparison of the flow-weighted transverse mixing coefficients estimated from velocity measurements with values estimated from concentration measurements revealed that a linear relationship with a fitted slope of 1.310 was found between the two values. This discrepancy occurred because the velocity-based method contains only shear effects due to the vertical velocity deviation, while the concentration-based method embraces other mixing effects, such as channel irregularities, storage zone effects, and turbulent diffusion, as well as the shear effect.