Environmental Engineering Reference
In-Depth Information
velocity and of instrument noise for measured data
from two gravel bed sites. The use of ADCP-measured
apparent bed velocity as a surrogate for bed-load
transport is a technique that shows considerable
potential for characterizing bed-load transport,
although calibration is required for each site, and
instrument noise is substantial.
tethered to maintain position). The “long average”
samples refer to these measurements (Table 2.3).
Furthermore, individual 5-minute ADCP measure-
ments contemporaneous with single VUV samples
are referred to as “5-minute averages”.
The apparent bed velocity was strongly correlated
with measured bed-load transport rate for the long
average Agassiz data and the Sea Reach data, and
less well for 5-minute averaged Agassiz data and
both Canoe Pass data sets (Fig. 2.6; Table 2.2).
Larger values of bed-load transport existed for the
Agassiz data than for the Sea Reach data for similar
values of apparent bed velocity; for particles travel-
ling at the same average velocity, the larger the par-
ticle the higher the mass-transport rate. In Canoe
Pass, similar bed velocities were measured in 2000
and 2001, despite lower bed-load transport rates
measured in 2001. Equivalent apparent bed velocity
despite lower bed-load transport in 2001 may have
resulted from use of a longer ADCP bottom-track
pulse length for ADCP bottom track measurement
that increased the infl uence of suspended scatterers
on apparent bed velocity. The variations in the
regression equations between sites suggested that the
relation between apparent bed velocity and bed-load
transport is site-specifi c, thus apparent bed velocity
must be calibrated for each site. Similar to the rela-
tions shown in Table 2.2, correlations of measured
bed-load transport and that calculated kinematically
with measured
v
b
varied for these data sets. Variations
resulted from differences in particle-size distribu-
tions, suspended-sediment concentrations, and
ADCP operating parameters.
All available data were plotted together using
non-dimensionalized bed-load transport rate,
g
*
,
correlated with non-dimensionalized apparent bed
velocity,
v
b
/
u
*, where
u
* is shear velocity calculated
2.2.1.2.1 Stationary boat studies.
Initial studies of
apparent bed velocity correlated the bed velocity
with bed-load transport rates measured by a physical
sampler and by dune tracking. The fi rst study was
conducted in 2000 (Rennie
et al
. 2002). Apparent
bed velocities were correlated with bed-load trans-
port rates, measured by concurrent physical bed-load
sampling, in the Agassiz gravel bed reach in the
Fraser River, British Columbia, Canada. This was
the fi rst indication that apparent bed velocity could
serve as a useful measure of bed-load transport.
Apparent bed velocity (
v
b
) and concurrent bed-
load transport rate (
g
b
) measured by physical sam-
plers were compared for fi ve data sets from three
reaches in Canada's Fraser River (Rennie & Villard
2004). Sea Reach and Canoe Pass were sand-bed
reaches near the river mouth. The third reach was
the gravel bed Agassiz site. A Helley-Smith bed-load
sampler (Helley & Smith 1971) was used for sand
and a VUV pressure-difference-type sampler (Novak
1957; Hubbell 1964; Cashman 1988) was used for
gravel. In the sand-bed reaches, measurements were
performed on the stoss sides of dunes to reduce
spatial heterogeneity. In the gravel-bed reach, several
5-minute VUV bed-load transport samples were col-
lected and averaged during a single ADCP measure-
ment (see Rennie
et al
. 2002). The ADCP samples
lasted between 2 and 112 minutes, (two 2-minute
samples were taken when the boat could not be
Table 2.3
Linear regression and functional relations for measured g
b
versus measured v
b
, Fraser River, Canada.
Location
N
r
2
Regression
Functional relation
95% CL
a
Agassiz long avg.
9
0.89
g
b
= 1.2
v
−
0.037
g
b
= 1.2
v
−
0.041
0.91-1.7
Agassiz 5 min.
13
0.52
g
b
= 2.0
v
−
0.059
g
b
= 2.6
v
−
0.088
0.60-7.8
Sea Reach
68
0.76
g
b
= 0.057
v
−
0.0007
g
b
= 0.062
v
+0.0005
0.062-0.062
Canoe Pass 2000
49
0.38
g
b
= 0.23
v
+0.001
g
b
= 0.36
v
−
0.00008
0.34-0.38
Canoe Pass 2001
15
0.42
g
b
= 0.090
v
−
0.0003
g
b
= 0.14
v
−
0.0004
0.0043-0.18
Non-dimensional
127
0.42
085
.
090
.
0.74-2.6
g
b
*
=
0 043
.
(
v u
)
g
b
*
=
0 045
.
(
v u
)
a
95% confi dence limits for functional relation slope.
From Rennie & Villard (2004).
