Introduction - Computational River Dynamics - page 8

in the longitudinal section. Because of the complexity of channel geometries, the width-
averaged 2-D models are preferable to the idealized vertical 2-D models in natural
situations. The horizontal 2-D models, which also are often called depth-averaged
2-D models, study the horizontal distributions of the depth-averaged quantities of
flow and sediment. 1-D models are widely applied in the simulation of long-term
sedimentation processes in long channels, 3-D models are often used in local fields
with strong 3-D features, and 2-D models are in between them.
Based on flow states , flow and sediment transport models are often categorized as
steady, quasi-steady, or unsteady. Steady models do not include the time-derivative
terms in flow and sediment transport equations, but consider temporal changes in
bed elevation and bed-material gradation. Quasi-steady models divide an unsteady
hydrograph into many time intervals, each of which is represented by a steady flow
discharge. Quasi-steady models are often used in the simulation of long-term fluvial
processes in rivers, but they cannot be used in cases with strong unsteadiness, such as
tidal flow in estuaries and flash floods in small watersheds. Unsteady models are more
general and can be used to simulate unsteady fluvial processes as well as steady and
quasi-steady processes.
As for the number of sediment size classes simulated, sediment transport models can
be uniform (single-sized) or non-uniform (multiple-sized). Uniform sediment models
represent the entire sediment mixture using a single-sized class, whereas non-uniform
sediment models divide the sediment mixture into a number of size classes and study
the behavior of each size class. Because sediments in natural rivers are usually non-
uniform in size and experience interaction among different size classes, non-uniform
sediment models are more realistic.
In accordance with sediment transport modes , sediment transport models are often
grouped as bed-load, suspended-load, and total-load models. Many early developed
models considered only bed-load or suspended-load transport. Because sediment may
change from bed load to suspended load or vice versa depending on flow conditions,
total-load models are more preferable.
Based upon sediment transport states , sediment transport models are classified as
equilibrium (saturated) and non-equilibrium (unsaturated). In many of the early mod-
els, it is assumed that the actual sediment transport rate is equal to the capacity of
flow carrying sediment at equilibrium conditions at each computational point (cross-
section or vertical line). The models based on this local equilibrium assumption are
called equilibrium transport models. However, alluvial river systems always change in
time and space due to many reasons; therefore, absolute equilibrium states rarely exist
in natural conditions. The local equilibrium assumption is not realistic, particularly
in cases of strong erosion and deposition. Non-equilibrium sediment transport mod-
els renounce this assumption and adopt transport equations to determine the actual
bed-load and suspended-load transport rates. Non-equilibrium transport models are
being more widely applied in river engineering these days.
In terms of numerical methods , flow and sediment transport models are categorized
as finite difference, finite volume, finite element, finite analytic, or efficient element
models. Since each of these numerical methods has its advantages and disadvantages,
numerical models based on all them exist in the literature. The choice of a specific
model depends on the nature of the problem, the experience of the modeler, and the
capacity of the computer being used.