Journal of Hydraulic Engineering

New field data relating to bend migration rates with bend radius of curvature to channel width ratios $\left(r_{m}w\right)$ for the Beatton River confirm the form of the relationship the writers described in 1975, although there is more variance than our initial results suggested. Recent data from forest successions on the floodplains of a variety of migrating rivers show that channel migration is a discontinuous process within any single bend. This means that predictions of migration rates for individual bends, based on short‐term measurements using erosion pins or aerial photographs, are highly suspect. These variations in migration rate are likely a response to fluctuating channel widths over time because of a short‐term imbalance between the rate of cutbank erosion and the rate of point‐bar sediment accumulation. Finally the examination of new evidence for channel incision on the Beatton River shows that the river is downcutting at a rate of about 2 mm/yr, not at 10 mm/yr as reported earlier.

The results of laboratory experiments on the clear‐water local scour of cohesionless bed sediment at cylindrical piers are presented. Based on the data, equilibrium depth of local scour is related to the particle size distribution of the bed sediment, mean particle size of the bed sediment, relative to the pier diameter, flow depth relative to both the pier diameter and the mean particle size of the bed sediment. A formula for estimating the maximum depth of local scour is also given.

The response of a gravel bed reach to imposed steady flows is examined under nonequilibrium conditions, where the bedload inflow is zero. Experimental research to date and general mathematical models for nonequilibrium alluvial conditions are reviewed. An experimental program was designed to examine the transient bed response for different reach lengths. Also, measured nonequilibrium bedload transport rates were compared with comparable local equilibrium capacity rates at selected time intervals and distances for a common mean flow velocity. Differences occurred between nonequilibrium transport rates and comparable equilibrium capacity rates at a maximum near the beginning of the reach but diminishing towards the downstream end of the local scour hole. This spatial variation of the transport rate deficit exists because the flow requires a finite length of bed to erode sufficient bed material to satisfy its equilibrium transport capacity. Consequently, mathematical model performance will be poor in the local scour hole region.

A theory is developed to predict the bed material load for steady fluid and sediment flow. The theory is based on the fact that the total power of the fluid and sediment flow is constant. The power of the sediment flow increases as the power of the fluid flow decreases. A solution is obtained for the flow stage with maximum sediment transport. The theory is supported by results of field measurements.

The vertical distribution of the suspended sediment in two‐dimensional, uniform, and fully developed turbulent flows under ice cover is formulated using the results of the $k‐\epsilon$ turbulence model of Launder and Spalding. Three sets of distribution curves each corresponding to a particular value of the ratio between the ice cover roughness and bed roughness are presented in a form similar to the distribution curves of H. Rouse for free‐surface flows. For the same value of the ratio between the shear velocity and the fall velocity of sediment particles, the relative concentration is less in ice‐covered flows than in free‐surface flows. Under similar conditions, the relative concentration of suspended sediment decreases as the ratio of the ice cover roughness to bed roughness increases. The agreement between the present formulation and the measurements found in the literature is favorable.

Depending on assumptions and data used in analysis, a probable maximum flood (PMF) can vary over a wide range of values each associated with a certain degree of conservatism. This paper presents a procedure which potentially can be used to determine the relative conservatism of a PMF estimate. By knowing how conservative the PMF is in the range of possible estimates, the engineer will be able to make a better decision as to how it should be used in designing the structure.

Experiments on the suspension of sediment were performed in flume with and without sediment layer on the bottom. For movable bed runs three‐dimensional sand ripples were found to affect seriously the features of suspended sediment. The horseshoe vortices and the resulting bursts in the lee of the ripples entrain intermittently a large amount of sediment into the outer layer of flow. Details of the suspension process were studied using a sand ripple model made of plaster. The result revealed that the vertical component of turbulent eddies is responsible for the suspension of sediment, and the statistical distribution of the component was found to be the Gaussian law. The statistical distribution of the concentration fluctuation was observed to be highly skewed, and the reason was determined to be the fact that the concentration increases with the power of eight against the vertical fluid velocity. The period of the intermittent ejection of sediment agreed reasonably with those reported for natural geophysical flows. Thus, the suspension of sediment in flows with rippled beds was proven to be significantly correlated with bursting phenomenon behind sand ripples.

The present paper is a study of the methods for calculating flow in branch channels. In order to obtain the unsteady flow in the channels without branching, this paper provides the non‐equidistant difference schemes on interior mesh points and the corresponding difference schemes on boundary mesh points. The branching of channels is a usual natural phenomenon, The calculation of the unsteady flow at branching is an important problem to be solved. In 1957, Stoker (5) proposed a calculating method, but it can be applicable only to the simple circumstance when the branch‐point is basically stationary. In fact, the branch point of tidal river (or some inland rivers) is movable within a somewhat large area depending on the variation of tide height (or flood height). For the movable branch point, the implicit and semi‐implicit methods will be provided in this paper. Both methods have already found their use in practical problems.

The measurement of air concentration and velocity obtained in the laboratory are compared with the recent prototype measurement obtained on Aviemore spillway. This comparison suggests that if a characteristic depth such as the depth at which the air concentration is 0.9 chosen then the averaged air concentration over this depth is only a function of the slope. Similarly the air concentration distribution is also determined by the slope. The velocity distribution determined in the laboratory and in the field are also compared and it is suggested that water velocity measurements are consistent with a universal velocity distribution. These with the Darcy friction factor computed from laboratory results lead to a simple method for predicting the design parameters in the uniform aerated flow at the end of a long spillway.

This paper presents the results of an experimental study on the structure of turbulent flow near groin‐like structures. Based on experimental observations, the deflected flow has been analyzed using the model of the three‐dimensional turbulent boundary layer. The flow bounded by the separating stream line, the groin, and the adjacent bank has been analyzed by treating it as a shear layer in which the velocity profiles have been found to be similar. The amplified bed shear stress near the groin has also been analyzed using the similarity idea. Even though the majority of the present experiments were done with the thin‐plate groin, some observations were also made on a groin with a semicylindrical hose.