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

34 - 8 PITLICK AND CRESS: DOWNSTREAM CHANGES IN CHANNEL GEOMETRY
Table 5. Results of the Analysis of Covariance (ANCOVA)
Comparing Downstream Relations for Bank-full Width and Bank-
full Depth in Alluvial and Quasi-alluvial Reaches'
Bank-full Width Linear Regression
Reach Type n a b SE, r2 p
Alluvial 72 205 -0.0016 0.00056 0.10 0.0057
Quasi-alluvial 60 187 -0.0022 0.00057 0.20 0.0003
Bank-full Width ANCOVA
ss",X nas, , F P
Disumcc - 0.0575 0.0967 0.59 0.44
reach tr pe
Reach type 1.339 0.0964 13.89 0.0003
Bank-full Depth Linear Regression
Reach Type n a b SEA, r2 p
Alluvial 72 7.1 -0.0033 0.00043 0.46 <0.0001
Quasi-alluvial 60 7.4 -0.0029 0.00041 0.47 <0.0001
Bank-full Depth ANCOVA
S$pK MS- F P
Distance « 0.0217 0.0537 0.40 0.53
reach type
Reach type 0.3361 0.0534 6.29 0.01
'Notation is the same as in Table 3; SS,,g is the regression sum of
squares, and MS,_ is the mean square of the residuals.
most notable changes in Tb occur in the same segments
where there is clear downstream fining (rkm 300-200 and
rkm 139-100); the reverse is also true, i.e. where Tb varies
slowly there is also little variation in grain size (rkm 370-
310 and rkm 185-140). The correlation between Tb andD50
is clearly important to our main hypothesis, as it implies that
the ratio of shear stress to grain size is balanced at both local
and regional scales. Figure 9 shows that there is indeed a
close correspondence between Dso and Tb in three of the
four segments. To evaluate these trends more carefully we
compared the regression relations for InD50 and b9Tb within
each of the segments listed earlier in Table 4. Formal
statistical tests of the relations for the first and third seg-
ments (Figures 9a and 9c) are not necessary because these
relations are not statistically significant (Table 4); in other
words, there is no significant trend in 1nD50 or 1nTb in either
of these segments. A t test of the equations for InDS0 and
111Tb in the second segment (Figure 9b) indicates that there is
no significant difference in the slopes of these relations (p =
0.49). Comparison of the equations for the lower segment
(Figure 9d) also indicates no significant difference in slopes
(p = 0.08), however, this result is not convincing, given the
scatter in the grain size data; if the two values of DSO at rkm
139 and 136 are removed, then the relations for 1nDso and
1116 are nearly parallel.
[24] The results presented above indicate that the change
in Tb within individual segments is roughly balanced by the
change in D50. It follows that the bank-full Shields stress,
Th, should be approximately constant from segment to
segment. A plot of the individual values of T*b (Figure 10)
suggests that this is indeed the case. A least squares fit of
the full data set gives the relation lnT*b = 0.048 - 4.7 x
10-05 X, with r2 << 0.01 and p = 0.86. If the data from rkm
139-105 are excluded, there appears to be a slight negative
trend in 1116, but the relation is very weak (r2 = 0.03; p =
0.07). Even with these data excluded, the difference in Tb
between upper and lower reaches is very small (0.050 to
0.045), thus we have no reason to reject the hypothesis that
Tb is constant downstream. For the reach as a whole, Tb
averages 0.049, which is about 1.5 times typical values for
the threshold for motion (0.03).
5. Discussion
[25] The results of this study show that the geomorphic
characteristics of the Colorado River vary systematically
downstream to maintain a constant bank-full. Shields stress,
Tb. The consistency in T*b holds for reaches that are bounded
by different sedimentary strata and potentially for reaches
that are influenced by recent tectonism. The Colorado River
has developed these geomorphic characteristics apparently
in response to regional-scale processes which supply water
and sediment in disproportionate amounts. The average
annual discharge of the Colorado River roughly doubles
through the study area, whereas the average annual sus-
pended sediment load increases by a factor of almost four.
Longitudinal measurements of bed material grain size show
that downstream fining is weak, suggesting that the supply
of coarse sediment from tributaries, hillslopes, and terraces
is high enough to replenish material worn down by abra-
sion, but not high enough to force significant changes in
Figure 8. Downstream trends in the bank-full shear stress of the Colorado River. Symbols are the same
as in Figure 7.