Journal of Hydraulic Engineering

Eleventh Hunter Rouse Lecture recounts the writer's 40 years of theoretical and experimental researches concerning the inception, development, and damaging action of vortices. It was found that a vertical vortex is normally formed in an unstable environment, in the presence of a diverging or converging flat jet that subsequently feeds it by momentum transfer. A tentative explanation of the intermittency of swirls at intakes was obtained proving the theoretical viability of vortices whose intensity varies more or less periodically along their axes. The inspection of intermittent vortices that develop spontaneously behind a weir or under a sluice gate led to the establishment of a universal Strouhal law, able to predict the frequency of the oscillations that can be excited within a restrained fluid body by a free current flowing along it, and, as a consequence, the frequency of vortex shedding and the structural vibrations that this phenomenon may induce. Applications that define the causes of revetment damages in hydraulic structures are made.

The literature abounds with apparently conflicting relations for predicting the reference near‐bed concentration of suspended sediment or the entrainment rate of noncohesive bed sediment into suspension. Most of these relations have been developed for uniform material. In this paper, seven such relations are compared against a common set of experimental data for which direct measurements of near‐bed concentration are available. The relations put forward by Van Rijn and by Smith and McLean are found to perform best. A new relation for uniform material, having similar predictive capabilities, is developed using the same experimental data. The new relation is generalized to handle sediment mixtures with the aid of field data. The generalization accounts for the hiding effect due to the presence of grains of different sizes. A test against an independent field data set indicates that the empirical fit can provide reasonable estimates of the sediment entrainment coefficient for beds covered with nonuniform material.

Laboratory experiments are carried out to determine the increase in bed shear stress and local scour over the zone of suction associated with a bed‐type river intake in a channel with a mobile bed. Bed shear stress is estimated first by observing the movement of indicator grains whose threshold shear stress in uniform flow was already known, and second by observing the rate at which local scour takes place. Evaluation of the two sets of results indicates that, at the onset of scour when particles are lifted from the bed by increased bed shear stress, they do not strike the bed again until they are clear of the scour zone. The final scour depth is found to be directly proportional to the difference between the initial bed shear stress because of suction and the threshold bed shear stress for the bed grains in the presence of suction.

This paper examines and reviews published results relating to local scour around submarine pipelines. It highlights the limitations of existing methods for estimating scour depth at the pipeline. Based on experimental results, the study proposes an empirical function relating the amount of gap flow through the scour hole for given flow conditions. With the aid of this function, it is possible to predict the maximum Scour depth at submarine pipelines for given flow and geometric boundary conditions. Published results suggest that the maximum equilibrium scour depth occurs when the pipeline is just lying on a plane bed and subjected to a pure unidirectional current. The undisturbed bed shear stress is equal to the critical shear stress for sediment entrainment. This condition implies that there is no general sediment transport away from the pipeline. The predicted maximum scour depth using the iterative method proposed in this study compares well with experimental results.

A theoretical formulation describing leakage from clogged channels that partially penetrate surficial aquifers is reported in this paper. Leakage characteristics are parameterized by a reach transmissivity, which is defined as the volume flow rate out of the channel per unit length of the channel per unit drawdown. An expression is developed to relate the reach transmissivity to the transmissivity of the formation, mean channel width, distance of drawdown measurement from the channel centerline, ratio of drawdowns on both sides of the channel, and local reach transmissivity associated with clogging. This theoretical expression, and associated limitations, are verified using a fine-scale numerical model. The proposed formulation was field tested at two channels. The measurements of reach transmissivity were found to vary linearly with the drawdown ratio, as predicted by the theory. Furthermore, the aquifer transmissivity obtained by fitting the theoretical formulation to the measurements is in close agreement with independently estimated values.

The use of derived distributions to predict flood frequency is relatively new and requires refinement and improvement. This paper reports another effort to obtain and test a derived flood‐frequency distribution, applicable to small watersheds in which overland flow is considered an important runoff component. Watersheds are conceptualized as first‐order streams with two symmetrical planes. The kinematic wave theory is used to obtain expressions to compute peak discharge and time to peak, as functions of effective precipitation variables and watershed parameters, for four different runoff cases. These cases are formulated according to whether or not there is concentration on the overland flow plane and in the stream. Some of these expressions are subsequently improved using regression analysis and kinematic wave simulated results. The probabilistic model for rainfall characteristics and the infiltration model are taken from results presented previously in the literature for other derived flood‐frequency distributions. The usefulness of the derived distribution approach is tested and demonstrated through application to two real watersheds.

Kinematic and diffusion waves are reviewed prompted by the continuing controversy regarding their nature and applicability. Kinematic waves are shown to be nondiffusive, but to undergo change in shape due to nonlinearity. This latter feature gives kinematic waves the capability of steepening, eventually leading to the formation of the kinematic shock. Kinematic wave solutions using finite differences are shown to possess intrinsic amounts of numerical diffusion and dispersion. These numerical effects are artificial and tend to disappear as the grid size is refined, making the solution dependent on the choice of grid size. Kinematic wave theory can be improved by extending it to the realm of diffusion waves. In this way, the diffusion inherent in many practical runoff computations can be accounted for directly in the modeling, rather than as an afterthought. The use of a kinematic wave method is indicated for small catchments, in cases where it is possible to resolve the physical detail without compromising the deterministic nature of the model. Conversely, the unit hydrograph is advocated for midsize catchments, where the kinematic wave method may prove difficult to implement. The dynamic extension to kinematic and diffusion models shows promise, particularly for modeling channel and flow conditions in which the Vedernikov number is substantially different from zero.

Uniform flow with natural air entrainment on spillway chutes is considered. Based on the experimental data of Straub and Anderson, the air-concentration profiles are first investigated. A similarity plot is presented, which unifies the data pertaining to various chute slopes. First, an expression for the average air concentration is deduced. The latter increases almost linearly with the chute slope. Second, the ratio between the air-water mixture flow depth and the pure water depth is analyzed. Based on previous findings, a simple formula is deduced for wide chutes. According to this relation, the bulkage of uniform aerated flow depends mainly on the chute slope, moderately on the roughness coefficient and only insignificantly on the uniform flow depth. These findings may then be brought in relation to those previously established and pertaining to the average air concentration.