8256

• UCREFRP Catalog Number: 8256
• Author: Holly, F. M. and R. Ettema
• Title: Sediment Imbalance in Rivers
• USFW Year: n.d.
• USFW - Doc Type: Simulation Possibilities and Problems
• Copyright Material: YES

Iii"..

420 BIOLOGICAL REroRT 19

available, and computational river hydraulics is
cursed by the lack of any consensus as to which of
them is most appropriate as a general predictor.
Order-of-magnitude differences in predicted trans-
port rates are not uncommon, see for example
Vanoni (1975).
The suspended-load source term, as well as
bed-sediment relations, become enormously more
complicated when cohesive sediments are dealt
with.
The interested reader is referred to Holly et al.
(1990) or Fan (1988) for more background on the
equations used and numerical methods employed
for their solution. For the f\tissouri River appijca-
tion described in the next section, an implicit
fInite-difference technique is used for all but the
advection-diffusion equation, which is solved by
the method of characteristics.

Missouri River Example

As discussed by Sayre and Kennedy (1978), the
Missouri River reach between Gavins Point Dam
and Blair, Nebraska, has experienced severe deg-
radation in the past 50 years. The bed level has
lowered more than 2 m near Sioux City, Iowa,
potentially compromising the structural integrity
of bridge-pier foundations, reducing the effIciency
of power-plant cooling-water intakes, and consid-
erably changing the natural habitat through in-
creased velocities, lowering of oxbow lake levels,
and decrease in vegetative cover. This degradation
seems to be the river's response to a series of
interventions by man: (1) reduction of the sedi-
ment load entering the reach below Gavins Point
Dam through closure of upstream dams, (2) regu-
larization of annual flood hydrographs through
storage and release cycles in upstream reservoirs,
and (3) channelization of the river through spur-
dike and artificial-cutoff construction, to form a
reliable navigation channel and to reclaim flooded
lands along the river.
Whatever the benefits accruing from these inter-
ventions, there also have been structural and envi-
ronmental costs that may not have been foreseen
when river engineering began in earnest in the
early part of this century. Concerned with the pos-
sible future course of the river's continuing re-
sponse to these interventions, the U.S. Army Corps
of Engineers has been carefully monitoring the
progress of the degradation and has conducted nu-
merical modelling efforts to better understand it.

In 1981 the Omaha District of the U.S. Army
Corps of Engineers commissioned the Iowa insti-
tute of Hydraulic Research to construct a numeri-
cal mobile-bed model of the degrading reach ofthe
Missouri River, with two objectives: identify the
root cause of the degradation, if possible, and
forecast the future course of the degradation.
The fIrst-generation mobile-bed simulation code
IALLUVIAL was developed for this study (Holly
and Karim 1986). The CHARlMA code, whose
equations are outlined above, is a third-generation
descendant of IALLUVIAL (Holly et al. 1990). Re-
sults presented further on were obtained using the
earlier IAILUVlALcode, but have been repro-
duced using CHARlMA.
The study reach is shown in Fig. 2. The compu-
tational model extends from Gavins Point Dam
some 505 river kilometers (rkm) down to Rulo,
Nebraska. Nine tributaries, including the Platte
River, are represented as local water and sedi-
ment inflows. The upstream boundary condition
of the model is water and sediment inflow through
Gavins Point Dam; the downstream boundary
condition is the stage-discharge rating curve at
Rulo. The initial channel cross sections are those
reported by the U.S. Army Corps of Engineers,
representing the natural channel in 1960 from
Gavins Point Dam to Ponca State Park, and the
channelized sections (roughly rectangular, with a
180-m width) as of 1960 for the remainder ofthe
model. The initial bed-sediment distributions are
from U.S. Army Corps of Engineers measure-
ments. It should be noted that the natural reach
above Ponca State Park had already been sub-
jected to at least 5 years of effect from the previous
closure of Fort Randall and Gavins Point dams.
In an initial calibration and verifIcation mode,
the model was run for the 1960-80 period in a
postdictive simulation. The simulation adopted
repeated annual inflow hydrographs typical of the
release schedules from Gavins Point Dam and of
tributary springtime flooding. Figure 3 shows lon-
gitudinal profIles of bed level and water-surface
elevation resulting from this simulation from
Gavins Point Dam to Omaha, with the observed
changes in water-surface elevation shown for com-
parison. The model captured the major features of
the river's response, including the relative inac-
tivity of its channel downstream of Omaha. The
only calibration that was done, other than use of
the Total Load Transport Model (TLTM) trans-
port predictor that is known to be appropriate for
this stretch of the Missouri River (Karim and