CGRG Bibliography of Canadian Geomorphology
Search Results
Author : Roy, M.L.; and Roy, A.G.
Date : 2007.
Title : Characterization of turbulent flow in shallow gravel-bed rivers: sampling guidelines, interpolation and spatial structure.
Publication : CANQUA Ottawa 2007. Canadian Quaternary Association Conference, June 4-8, 2007. Carleton University, Ottawa, Ontario, Canada.
Issue :
Page(s) :
Abstract
Recently, flow turbulence has been recognized as a variable affecting fish behavior and energy expenditures. In gravel-bed rivers, where depth is often limited, the most common way to measure turbulence is to use high frequency velocity sensors at a point. Spatial interpolation can then be useful to link fish movements to hydraulic habitat. In this study, we provide guidelines for the sampling and the interpolation of velocity point measurements at the reach scale. The objectives were (1) to estimate an optimal density of measurement that would present a tradeoff between an extensive sampling strategy and reasonable interpolation accuracy and (2) to investigate the downstream and transverse structure of turbulent flow using geostatistics in morphological units in order to adjust the sampling grid accordingly. The study sites consisted of four river sections, two riffles and two pools, with contrasting velocity ranges, depths and bed roughness. Velocity time series of 100 s were sampled systematically every 25 cm using two acoustic Doppler Velocimeters (25 Hz). From the velocity time series, several mean and turbulent flow parameters were estimated (mean flow velocity, RMS, Reynolds shear stress and turbulent kinetic energy). We tested the influence of the sample size on the interpolation model accuracy was done by successively subtracting sub-samples of 5 % up to 95% of the total number of points and by computing interpolation error estimators Densities varying between 4 points/m2 and 5.5 points/m2 represent a tradeoff between minimizing sampling effort and maximizing interpolation accuracy. Mean bed roughness presents a significant linear correlation with hydraulic heterogeneity of the sections and with interpolation errors. Nevertheless, the sections presenting a rougher bed did not show higher density threshold, suggesting that some part of the variability in turbulence statistics in the section varies over a scale smaller than 0.25 m. Furthermore, at the optimal sampling density, interpolation errors related to turbulent intensities are comparable to those of mean flow velocity (10 to15 %) and TKE-related errors are slightly higher (20 %). Mean vertical velocity and shear stress are not well characterized using a planimetric sampling scheme as errors range from 30 to 50 % and 65 to 100 % respectively. Directional semivariograms showed a more important spatial variability in the transversal direction comparatively to the longitudinal direction. The overall coherence of spatial patterns extended over a distance twice as long in the downstream than in the transverse direction. The opposite was observed for a few variables and the magnitudes of the difference between the directions varied substantially between the sections. Therefore, systematic sampling could be more suitable to a wide range of sections than an asymmetric sampling. These results could provide tools for a future investigation of the relations between turbulent flow and fish habitat use.
Bibliography of Canadian Geomorphology