The Thomas and Copeland methods are sophisticated, multi-layer approaches to bed mixing and armoring. They are also complicated, doing a lot of math behind the scenes, providing results that can be difficult to interpret. Because HEC-RAS computes transport capacity for each grain class independently, these methods limit erosion while computing transport that reflects the subsurface gradation, which is generally observed in rivers.

However, a simplified two-layer active layer method (figure in Sorting and Armoring) is also included in HEC-RAS. A simple active layer approach has obvious disadvantages including less vertical resolution and no explicit armoring factor. Use it with caution. However, it is a more intuitive and transparent method, it can form a coarse or fine active layer, and, with an appropriate exchange increment, it may be preferable in some cases for modeling mobile armor systems (Gibson and Piper, 2007).
Hirano (1971) is often credited with introducing the "active layer" approach for sediment transport modeling, though HEC was doing similar work at the same time. This approach divides the substrate into an active (mixing or surface) layer, available for transport, and an inactive layer that has no influence on the computations for a given time step.

Selecting an Active Layer Thickness

Active Layer Exchange Incremnts

As the bed aggrades and degrades the sediment passes the active layer resets to a specified thickness (e.g. the d₉₀), and the layers pass material between them. If the bed erodes, the inactive layer sends sediment, an "exchange increment" to the active layer, to restore it. If the bed deposits, the active layer resizes and sends an exchange increment to the inactive layer.

The gradational composition of the erosion exchange increment is straight forward. The material the inactive layer sends to the active layer has the gradation of the inactive layer, and is mixed with the active layer sediment.
The depositional case is more complicated because the deposited material has a different gradation than the active layer. Three basic options have been proposed.

1. Mixed active Layer/Deposition Gradation. Adds the deposited material to the active layer, mix them, and send an exchange increment with the fully mixed gradation to the inactive layer. This method assumes the depositing material and the active layer are mixing fully at tight temporal scales.
2. Ambient Active Layer Gradation: Remove the exchange increment from the active layer and send it to the inactive layer first, then add the deposited material to the active layer and mix them. This method assumes static stratigraphy, that the deepest material, which would have the active layer gradation, would become inactive.
3. Deposited Material Gradation: Early work on dynamic armoring recognized that actively transporting channels maintain an armor layer even in depositional environments. But the previous two methods would cause depositing beds to fine, burying the cover layer (Parker et al. 1991a,b). So this method sends the deposited material directly to the inactive layer. This maintains a coarse cover layer, but does not allow any gradational evolution in the active layer.

Toro-Escobar et al (1996) tested these hypotheses and found that the was not composed entirely of the active layer or bed load gradations, but it also wasn't proportional (e.g. fully mixed). They found that method 3 was closer to their observations than method 2, but was not sufficient. So they proposed a depositional exchange increment during deposition was composed of 30% active layer material and 70% deposited material. This maintains a coarse cover layer better than methods 1 and 2 above, but also allows gradational evolution.
HEC-RAS follows this approach, passing depositional exchange increments to the inactive layer that are 30% active layer material and 70% depositional material. For example, if HEC-RAS deposited 10 tons of material in a time step (assuming the active layer remained the same thickness) it would transfer 3 tons from the active layer to the inactive layer and add 7 tons of the deposited material directly to the inactive layer. It would then mix the remaining 3 tons of the deposited material into the active layer.

Surface Based Transport Equations and the Active Layer Method: The Thomas and Copeland methods subdivide the active layer to address competing principles in sediment transport:

1. Transporting sediment generally has the same gradation as the subsurface layer, so transport capacity must be computed based on subsurface gradations.
2. The cover layer regulates erosion.

These bed mixing and armoring algorithms allow HEC-RAS to attribute the transport potential of classic transport functions based on subsurface gradations, computing transport capacity based on these gradations, without over predicting erosion.

In the years since these algorithms were developed, a different conceptual approach to graded bed transport has emerged: surface-based transport functions. Parker (2008) suggested that basing transport functions on surface gradations automatically account for dynamic armoring processes.
Wilcock and Crowe (2003), which is included in HEC-RAS, is a surface-based equation in this tradition. It accounts for inter-particle interactions like hiding and sand-dependent gravel transport explicitly, but builds the armor layer regulation into the equation implicitly, by basing transport on armor layer gradations. Therefore, when using the Wilcock and Crowe (2003) equation in HEC-RAS, always use the active layer Mixing algorithm. Using Thomas or Copeland with a surface-based transport equation double counts armoring and miss-matches conceptual frameworks.