Grain sorting in the morphological active layer of a braided river physical model

Abstract

Abstract. A physical scale model of a gravel-bed braided river was used to measure vertical grain size sorting in the morphological active layer aggregated over the width of the river. This vertical sorting is important for analyzing braided river sedimentology, for numerical modeling of braided river morpho-dynamics and for measuring and predicting bed load transport rate. We define the morphological active layer as the bed material between the maximum and minimum bed elevations at a point over extended time periods sufficient for braiding processes to re-work the river bed. The vertical extent of the active layer was measured using 40 hourly high-resolution DEMs of the model river bed. An image texture algorithm was used to map bed material grain size of each DEM. Analysis of the 40 DEMs and texture maps provides data on the geometry of the morphological active layer and variation in grain size in three-dimensions. Normalizing active layer thickness and dividing into 10 sub-layers we show that all grain sizes occur with almost equal frequency in all sub-layers. Occurrence of patches and strings of coarser (or finer) material relates to preservation of particular morpho-textural features within the active layer. For numerical modeling and bed load prediction a morphological active layer that is fully mixed with respect to grain size is a reliable approximation.

Morpho‐sedimentary characteristics of proximal gravel braided river deposits in a Froude‐scaled physical model

“…The basis of the approach is described in Carbonneau et al (2005a,b), where local image texture properties were derived from a 64 grey-level co-occurrence matrix (Haralick et al, 1973) and correlated to samples of homogeneous bed surface grain patches. Extensive calibration and parameter testing improved the accuracy and reliability of the textural analysis and mapping (see Leduc et al, 2015). The sampling window was 7 9 7 pixels (chosen based on median grain size and image resolution), and the best fit of the data was based on the entropy index.…”

“…Fifty-eight patches had uniform grain size, and an additional 152 samples randomly distributed across the surface had the full range of grain sizes and gradations. Half of the samples were used to develop the calibration, and the other half to validate the calibration and derive error statistics (see Leduc et al, 2015, fig. 1).…”

“…The relative error of the absolute grain size ranged from 0 to 100% but over half of the calibration set had an error less than 20%. The derived data are imagebased grain-size estimates of the median grain size, not strictly physically measured grain sizes (Leduc et al, 2015). Maps of this grain size were produced in SCILAB â using the ortho-rectified output images and mosaicked into a continuous map for each time period ( Fig.…”

“…For example, a Kolmogorov-Smirnov test of the distributions gave D statistics of 0Á110 (time 179 versus maximum surface) 0Á070 (time 179 versus minimum surface) and 0Á060 (maximum surface versus minimum surface), all of which show no significant difference at a of 0Á05. Thus, at the channel-reach scale, there is no statistical distinction between the grain-size distributions of the river bed, the minimum or the maximum surfaces (see also Leduc et al, 2015).…”

Simulating Laboratory Braided Rivers with Bed-Load Sediment Transport

“…Sediment gradation plays an essential role in the bed deformation process, which is especially important for the evolution of local units in natural [30] and laboratory braided rivers [31]. In addition, laboratory experiments [32] also show the influence of vertical grain sorting in the active layer, indicating that it is necessary to divide the riverbed into multiple bed layers in river simulation. This mixing layer concept has been adopted by many researchers such as Wu [29] and Van Niekerk et al [33].…”