FLOOD01888

• County: Statewide
• Title: River Hydraulics
• Date: 10/15/1993
• Prepared By: US Army Corps of Engineers
• Floodplain - Doc Type: Educational/Technical/Reference Information

EM 1110-2.1416
15 Oct 93

(2) In the context of two-dimensional modeling for
river hydraulics, the study objectives usually require the
prediction of velocity or stage. Field measurements of
velocity must include the direction as well as the magni-
tude. Most two-dimensional models used for river
hydraulics compute vertically averaged velocities; there-
fore, the field data must be converted to vertical averages
for proper model-prototype comparisons. For most situa-
tions, it is adequate 10 use the average of the velocities
measured at O.2*depth and O.8*depth (French 1985).
Depth must also be obtained at the locations of the
velocity measurements. "Depth.. alone is of limited
value; one should also have the corresponding water
surface or bed surface elevation. Similarly, to calibrate a
model for stage prediction, one should have field mea-
surements of stage and the variation of stage with time at
many locations within the study area. Also, the dis-
charge(s) at the time(s) of those measurements must be
known.

C. Validation data. Validation data are field obser-
vations not used in calibration that are used 10 provide an
independent check on model performance (ASCE 1982).
The above considerations for calibration data also apply
10 validation data.

4-9. Data Development and Model Calibration

a. Geometry. An accurate geometric description of
the flow region is a primary requirement. "Accurate"
here means that the key flow controlling and conveying
features of the study area are appropriately represented in
the field data. The engineer should be aware of the
origin and veracity of the field data. Ideally, the area of
interest is described by a detailed digital terrain model or
contour map of adequate resolution for the study needs.
Refer 10 EM 1110-2-1003 and "Accuracy of Computed
Water Surface Profiles" (1986). Mosl existing model
data are, however, in the format of cross sections (HEC-
2). Direct use of HEC-2 style data for two-dimensional
or one-dimensional unsteady simulations should be tem-
pered by the following considerations: (I) the HEC-2
cross sections may not have been chosen to best repre-
sent the directinn and distribution of flow, (2) off-channel
storage areas (important for dynamic simulations) may
have been neglected when surveying the cross sections,
and (3) the sections may not be appropriate for the objec-
tives of the present study. Therefore, before using an
existing HEC-2 (or other one-dimensional steady flow)
data set, thoroughly check the data for conformance with
the needs of the present study objectives. The use of

4-6

cross sections 10 develop two-dimensional model input
requires that the sections be registered (located) on a
topographic map or aerial photograph and the conlours
filled in, usually by hand

e

b. Bottom roughness. In most two-dimensional
riverine situations, bottom roughness can be described in
the same fashion as would be used for a "traditional"
one-dimensional (HEC-2) analysis (refer to Chapter 6).
Due to the ability of the two-dimensional approach 10
incorporate spatial variation of roughness, aerial photo-
graphs or topographic maps can be used to identify
regions of uniform roughness, such as clumps of vegeta-
tion, changes in bed material or bed forms. As in the
one-dimensional approach, the roughness coefficients
selected from field inspection (which is essential for
successful modeling) will probably need to be modified
in the calibration process. Should the calibration process
indicate the need for values of coefficients that are out-
side the range suggested by good engineering judgment,
one should closely inspect the geometric data, flow data,
boundary condition specifications, and calibration data.
Most often it is flawed geometric data, or the manner in
which it is interpreted by the engineer and used by the
numerical model that is the cause of a poor simulation.

.
..

.
.
.

c. Turbulent exchange coefficients. Two-dimen-
sional flow models require turbulent exchange coeffi-
cients, often called eddy diffusivities, which represent the
internal shear forces created by the transfer of momen-
tum between faster and slower regions of flow by means
of turbulent mixing. This can actually be observed in
most rivers by watching surface boils and eddies move
about in the flow. These coefficients reflect, somewhat,
the energy losses that are described by the expansion and
contraction coefficients in one-dimensional models. The
values of these coefficients cannot be directly measured
nor observed. Calibrated expansion-contraction coeffi-
cients cannot be directly translated into values for the
turbulent exchange coefficients. Guidance on selection
of values for the turbulent exchange coefficients is pro-
vided in the documentation for two-dimensional models
(e.g., TABS-2, Thomas and McAnally 1985). These
coefficients primarily effect velocity distributions and
should be calibrated based on velocity distributions mea-
sured in the field. If measurements are not available,
information from pholographs (both ground and aerial) of
the flow or sketches of observed flow patterns can be of
use. Some flow situations such as a jet entering a still
body of water are momentum dominated. In these cases,
the exchange coefficients are very important. Most open

e

.
.

..

e