Author(s): Bruce J. Macvicar; Alexander N. Sukhodolov
Linked Author(s): Bruce MacVicar, Alexander Nicholas Sukhodolov
Keywords: Double (space and time) averaging; ensemble averaging; field studies; open channel flow turbulence; streams and rivers; velocity measurements
Abstract: Shear velocity and von Kármán’s coefficient can be considered as scaling parameters for the characterization of velocity and turbulence distributions. However, few studies have assessed their accuracy in rivers. We apply a finite ensemble averaging procedure to obtain statistically representative profiles for evaluating the effects of sampling period duration. Results show that uncertainty in point statistics can be minimized by estimating sampling requirements from turbulent or morphologic scales, but such estimates do not consider non-turbulent (i.e. hydrologic) variation. Measuring at 25 Hz, durations of 3–4 minutes with at least three repetitions at each sampling point were required to reduce uncertainty to within ±10%. Sampling durations to achieve similar accuracy in other cases will vary with the number of sampling points in the profile, flow characteristics, and the nature of non-turbulent fluctuations. Overall, sampling recommendations from a consideration of scaling parameter uncertainty will greatly exceed recommendations from other studies.