9500

• UCREFRP Catalog Number: 9500
• Author: Pitlick, J. and R. Cress
• Title: Downstream Changes in the Channel Geometry of a Large Gravel Bed River
• USFW Year: 2002
• USFW - Doc Type: American Geophysical Union
• Copyright Material: YES

PiTLICK AND CRESS: DOWNSTREAM CHANGES IN CHANNEL GEOMETRY
5W f-
?. _.
.
aoa .350 .. aao 250 z
a0 SO !aa
Distance (km)
Figure 5. Downstream trends in the subsurface sediment (bulk bed material) grain size distribution.
scaling variable, the bank-full width typically increases
faster downstream than the bank-full depth [Leopold and
Maddock, 1953; Church, 1992; Knighton, 1998]. Our data
exhibit the opposite trend. The slope of the overall regres-
sion equation for bank-full depth is much higher than the
slope of the overall equation for bank-full width (Table 3).
Over a total distance of 260 km, the bank-full depth more
than doubles, whereas the bank-full width increases by only
about 50% (Table 1). The downstream trends in channel
geometry of the Colorado River appear to be driven by the
particular combination of input variables; in this case the
discharge and grain size change slowly in relation to
the slope. The disproportionate increase in channel depth
helps satisfy the requirements of continuity for both water
and sediment, but a large increase in width is not necessary
if the input variables (discharge, total bed load, grain size)
are slowly varying.
[22] We conclude this section by focusing on downstream
trends in the bank-full shear stress, Tb, and the bank-full
Shields stress, T*b. Figure 8 shows the downstream trends in
Tb, with separate segments marked as in previous figures.
The overall trend indicates that there is roughly a 3-fold
2.5 -
1 •

v 2.0,--
n `

t 0

o

1.5 .:... n

4

1.0 ..
IS
0.5-- Y = 1.95 - 0.0020X
_.? r2 W 0.22 p = 0.02
/'? /?
0
0
T?ZT
. ; 7 - 1 7 TT-? . l i R T T
1
350 300 250 200 150 1
00
Distance (km)
34-7
decrease in Tb, from an average of 47 N/m_ in the upper-
most reach to an average of 17 N/m2 in the lowermost
reach. An exponential fit of the full data set gives the
relation lnTb = 27.6 - 0.00135 X, with r2 = 0.08 and p <
0.001. However, this trend is clearly skewed by the high
values in the lower reaches of the study area (rkm 140-
115); with these values removed, the relation is consider-
ably steeper, lnTb = 16,9 - 0.0030 X, and the correlation is
higher (rz = 0.47; p < 0.0001).
[23] Similar to the trends in surface DSO, there appear to
be steeper trends in Tb within several segments. In fact, the
400

200 '(7• •

-
•
75
?•
?
•
:?
too
__
• •
. •
m 80 '? r - ' # •?
60 •
400 350 300 250 200 150 100 50
400 350 300 250 200 150 100 50
Distance From Green River (km)
Figure 6, Ratio of the median grain size (D50) of the
surface sediment to the subsurface sediment at paired
sampling localities. The value indicated by the solid dot was
not included in the regression.
Figure 7. Downstream trends in (a) bank-full width and
(b) bank-full depth of the Colorado River. Open circles
represent cross sections in alluvial reaches; solid circles
represent cross sections in quasi-alluvial reaches.