As discussed above, the sediment transport calculations for bed change, sorting, and layering may be computed on a coarser subgrid resolution than the hydraulics grid due to computational and memory restrictions. Since the flow and sediment transport utilize different subgrid curves it is necessary to transfer information from one set to another without losing information or causing numerical artifacts. In order to simplify the computations, the sediment transport subareas (for cells) and sublengths (for faces) are defined by grouping hydraulic subareas and sublengths. The approach followed here is to simply group the high-resolution subregions based on a target number of grouped subregions and also similar bed elevations (i.e. inflections in the area-elevation and length-elevation curves). The figure below shows an example where high-resolution subareas are grouped for two different cases.

Figure 3 25. Schematic showing the grouping of high-resolution subareas for use in sediment transport.

The algorithm begins by computing a grouping number nG = round(Mh/Ms) in whichround(X) is the rounding function, Mh is the number of hydraulics subareas, and Ms is the number of sediment subareas. In both cases the input hydraulics curve has 6 subareas and the desired sediment curve is set to have 3 subareas. A first attempt at grouping the hydraulics subareas is done using the grouping number starting from the lowest elevation. Since the grouping number is calculated using a rounding function, the last group of subareas may contain fewer or more than the rest in order to maintain the number of sediment subareas). This is illustrated in Figure 8a. In some situations the first grouping may lead to a poor description of the elevation-area curve inflections. An improvement of the sediment curves is done further grouping subregions on the high-resolution curve based on similar elevations. This process is illustrated in Figure 8b. Grouping the hydraulics subgrid regions into sediment subregions greatly simplifies the computations since each subregion may be treated separately.