Sediment sorting in the gravel-sand transition along rivers : a field and modelling investigation.

Gradual downstream fining in gravel bed rivers is often followed by a spatially rapid switch to a sand bed. These gravel-sand transitions (GSTs) can occur in rock types where abrasion rates are low. A common factor is declining shear stress towards base level causing size selective sorting. This study investigates the characteristics of visually abrupt GSTs in two streams of different size and morphology (AlIt Dubhaig, Scotland and Vedder River, British Columbia, Canada).
A one-dimensional numerical model of width-averaged size selective gravel sorting is enhanced to simulate gravel-sand mixtures. The updated model fails to generate a GST unless an abrupt break of slope is specified at the start of a run, although only one of the fieldsites exhibits this feature today. This finding suggests that additional processes are crucial to initiate a GST.
A qualitative method of assessing bed surface facies is developed. This is shown to be quantitatively accurate in predicting the bed surface sand range and, when combined with bulk bed grain size distributions (GSDs), indicates a that threshold exists for gravel bed sand content. Above this threshold the channel bed facies switches from gravel framework to sand matrix causing a non-linear relationship between bulk and areal sand content.
Laterally-distributed sampling shows alternation in width-averaged GSD along Vedder River above the GST, with gravel bar samples having higher D50 and lower sand proportion than those between bars. The channel bed exhibits a sandier GSD above the GST than would be indicated by inspection. The drop in D50 and increase in sand proportion across the GST is of similar magnitude to that associated with bars upstream although the change in grain size is extremely abrupt in surface appearance. Beyond the last gravel bar there is a much greater lateral variability in facies than either upstream or downstream. Point sampling of GSDs, which tends to be done on bars, may be inadequate to characterise the GST or positively misleading.
Evidence from subsurface probing investigations and bed surface sedimentology indicates a slowly prograding gravel front. The position of the front is dependent on near-bed hydraulics. A fine-gravel tracer experiment shows that the transport of these sizes in the GST reach is size selective, although this is not the case in the distal gravel reach.
Field characterisation indicates that the crucial processes missing from the model include: the overwhelming of a gravel framework bed by sand, as the threshold for sand storage is approached, leading to an increased availability of sand on the bed surface; and the lateral sorting of sediment into patches of different ambient grain size, further increasing the availability of the fine fraction.