Evaluation of statistical properties of dune profiles

A set experiments for dunes developing from a flattened sand bed was obtained in a narrow 0.44-m-wide and 12-m-long glass-sided open channel. The sand in use was a coarse uniform sand of D50=0.85-mm and was exposed to a series of steady and uniform flow conditions. The chosen flow depths generated practically 2D dunes in flow direction over the length of the channel. Spatial sand-bed-elevation profiles were recorded on the centreline of the flume over a distance of 6-m, roughly every 23-sec over the time of development. The recorded dune geometries were evaluated with both the discrete and the continuous approach. For the discrete approach, discrete height and length values are obtained for dunes. For the continuous approach, the second-order distribution moments, standard deviations, were used to obtain characteristic height and length values of the dune field. It is shown that a lack of clear definitions for the discrete approach results in a wide range of averaged dune geometries, depending on how thresholds were set during the analysis process. The continuous approach provides more objective results, but interpreting the results of the analysis requires careful consideration. For this paper, the analytical results of applying both approaches are compared. A preliminary correlation between both approaches is discussed. The physical relationship between the flow field and characteristic height and length values, as obtained through the continuous approach, is not yet clear.