38
ERDC TR-04-1
the OHWM may be easier in some nonalluvial settings compared to alluvial
channels.
Hydraulic Models
The most widely used hydraulic model in the United States is HEC-RAS,
which is an updated version of the once widely used HEC-2 model (U.S. Army
Corps of Engineers 2002). HEC-RAS is a one-dimensional flow model that
predicts flow velocity, width, and depth for a given discharge if the cross
sectional area, gradient, and Manning's roughness value are known. Conversely,
the HEC-RAS model can calculate the discharge necessary to reach a certain
stage in the stream channel, which is especially useful for calculating the
discharge of past flow events where the high water mark is preserved (O'Connor
and Webb 1988). The high water mark for a given flow can be identified from
hydrologic flow indicators such as the highest presence of silt, flotsam, or trim
lines. Such hydrologic high water marks may be representative of the OHWM
when found in conjunction with geomorphic or vegetation indicators, such as the
height of the active floodplain or the lower level of perennial vegetation. As a
one-dimensional model, HEC-RAS largely assumes that discharge is constant
through time and space, conditions that rarely hold in nature, especially in arid
climates. Sheetflood and debris flow conditions are especially difficult to model
using HEC-RAS because of rapidly expanding flow and highly variable viscosity
of flow, respectively.
Two-dimensional models are more capable of accounting for the spatial and
temporal variations in flow conditions associated with arid-region river systems,
but the increased analytical complexity associated with two-dimensional
modeling has, to date, precluded its wide acceptance as a predictive tool in
watershed management. Two-dimensional hydraulic models have been
developed that are capable of characterizing several features and processes
typical of arid-region rivers such as rapid width adjustments (Darby 1998),
erodible banks (Mosselman 1998), multiple threaded braided channels (Lane and
Richards 1998), debris flows (O'Brien et al. 1993), flow around obstructions
such as boulders (Crowder and Diplas 2000), and transmission losses (El-Hames
and Richards 1998). However, two-dimensional models are not yet capable of
modeling all of these conditions and their complex interactions simultaneously,
as occurs during natural flows in ephemeral channels. Two-dimensional models
can accurately recreate conditions observed during known flows, but the exten-
sive calibration necessary with actual conditions precludes their use for predic-
tive purposes. With time and increasing computing power, these models hold
promise to be of much greater value than at present.