Experimental investigation on flow resistance of high gradient streams.

High gradient streams are characterised by steep slopes, coarse bed material with nonuniform size distribution, flash floods and flow depths of magnitude comparable to the bed material size. Another characteristic of high gradient streams is development of armor layers with cobbles and boulders. In these streams, it is often difficult to measure water discharge or flow velocity directly, particularly at high flowsinstead, it is necessary to use the indirect methods in which flow velocity and discharge is calculated using a flow resistance predictor. The problem of predicting flow resistance in high gradient alluvial channels with sufficient accuracy is of great interest to hydraulic engineers and is essential for the effective design of water-resources related projects. Flow resistance in natural alluvial channels is largely influenced by bed material particle size, bed configurations, channel cross-section shape, longitudinal slope and flow depth. Although resistance analysis is a classical component of the study of free surface flow, the process of flow resistance is complex for high gradient streams due to the composite processes of flow and limited field and laboratory data and hence determination of flow resistance is still a major task. During the past decades several flow resistance relationships have been presented and tested against data in high gradient streams. The application of such equations has not been fully evaluated under different flow conditions. In the recent literature, flow resistance is mostly quantified in terms of Manning roughness coefficient and dimensionless Darcy-Weisbach friction factor. The present study is aimed to review the recently developed high gradient stream flow resistance predictors, with the purpose of evaluating their accuracy and applicability, and to introduce the most appropriate equations. Considering the general forms of Keulegan friction law and Manning-Strickler equation, the selected flow resistance relationships are generalized and rewritten into the two common forms of semilogarithmic and power-law functions. The generalized equations illustrated that flow resistance is a function of relative submergence R/D84 and for a given mean hydraulic radius (or mean flow depth), the channel slope S and the bed material roughness parameter D84 are required to determine mean flow velocity. In this project, using 75 sets of accurately measured laboratory flume data, the flow resistance coefficient was investigated. The main purpose of the study was to propose the most appropriate equations for high gradient streams with coarse-bed material. The data indicated flow discharges ranging from 8.0 to 64.0 lit/s and measured mean velocities in the range of 0.26 to 0.9 m/s. The type of flow was sub and super-criticalthe relative submergence (d/D84) varied between 0.74 and 4.78 indicating small to intermediate scale roughnessthe width to depth ratio (W/d) varied from 22 to 92. Based on measured data, best-fit flow resistance formulas were suggested for estimating flow resistance coefficient (8/f)1/2 and mean flow velocity of natural coarse-bed rivers. Using field data from 16 high gradient rivers in England, the accuracy of the proposed formulas were investigated. Two different sets of accurately measured hydraulic and bed material data, one from a laboratory flume and the other from a typical reach of a coarse gravel bed river in IRAN, were also used to investigate the applicability of selected mountain stream flow resistance relationships.