The composition, structure, and dynamics of the streambed surface play defining roles in many hydraulic, geomorphic, and biologic processes.  The streambed surface can coarsen via more than one mechanism.  Current stream bed morphodynamic models incorporate only one armoring mechanism, namely divergence-driven sorting, but substantial armoring is observed under conditions for which this mechanism is not effective.  Current models require changes in the mean bed elevation to produce changes in surface grain size, even though armoring is observed with negligible bed elevation change.

Both sediment-feed and sediment-recirculating flumes have been used to study bed armoring.  In one, sediment is fed into the flume, typically at a constant rate and composition, and the sediment transported beyond the downstream end is removed.  In the other, sediment transported to the downstream end is recirculated to the upstream end, providing a periodic sediment flux boundary condition.  While neither exactly simulates the behavior of a natural river, each represents an end member of a continuum of behavior that stems from the dual nature of real rivers—that is, over long time scales, a river adjusts to the external supply of water and sediment from its watershed and therefore behaves as a sediment-feed system, whereas recent evidence suggests that over the duration of a flood, the armor layer in a gravel-bed river may persist with the bed and transport exhibiting behavior similar to a sediment-recirculating flume (Wilcock & DeTemple, 2005).

Armoring is observed in both flume configurations despite their different transport boundary conditions.  In the externally-controlled feed case, the system evolves such that the flux and composition of sediment leaving the system matches that entering the system upstream.  As the bed adjusts, the surface coarsens as larger grains overpopulate the bed in order to overcome their inherently higher resistance to transport and match their fractional composition in the feed.  This is mobility-driven sorting.

In contrast, transport in a recirculating flume has little or no divergence in the fractional transport rates at the reach-scale and, therefore, no apparent opportunity for mobility driven sorting.  Despite this, surface-coarsening has been observed in sediment recirculating flumes (Marion & Fracarollo, 1997) with a possible sorting mechanism being kinetic sieving (in which small grains are deposited in pockets left by entrained large grains, but not vice versa, a process similar to the classic Brazil nut sorting of agitated mixtures).  Even though there is little or no divergence in the mean transport, we have observed in our experiments persistent sorting waves which produce spatial and temporal transport variability and local, mobility-driven sorting.  In a transport system with no divergence in the mean transport and no bed scour or aggradation beyond the grain scale, substantial armoring can occur via both kinetic sieving and local mobility sorting.  Neither mechanism is adequately represented in current armoring models.

A complete framework for modeling vertical sorting has been proposed by Parker et al. (2002).  This model requires size dependent vertical exchange functions, specification of which requires unprecedented detail in the vertical grain size distribution, which we are collecting for the plane-bed mixed-size case.  Prediction of physical and biological stream response to natural or engineered changes in water and sediment supply requires an ability to predict the frequency and rates of transport and streambed armoring.  The development of a general armoring model will support evaluation of stream disturbance and restoration scenarios.

Major accomplishments:

Papers
DeTemple, B., and P. Wilcock (2006), Observations of armor formation in a sediment-recirculating flume, in River, Coastal, and Estuarine Morphodynamics: RCEM 2005, edited by G. Parker and M. H. Garcia, pp. 1049-1058, Taylor & Francis Group, London.

Wilcock, P. and B. DeTemple (2005), Persistence of armor layers in gravel-bed streams, Geophysical Research Letters, 32, doi:10.1029/2004GL021772.