Environmental Engineering Reference
In-Depth Information
Froehlich (2003; in press) installed a set of micro-
phones encased in steel pipes, and recorded the
signals generated by gravel collisions with the pipes.
He was able to quantify the relation between the
number of cumulative gravel-pipe interactions and
cumulative bed-load discharge captured in sediment
basins. Mizuyama
et al
. (2003; two papers in press)
and others installed a similar system, consisting of a
single pipe containing a microphone deployed on a
Sabo-type dam, designed to retard the propagation
of debris fl ows. Mizuyama
et al.
(2003) found good
correlation between counted impacts and bed-load
transport rate at intermediate- to high-transport
rates, with lower correlations at very low transport
rates and at extremely high transport rates.
Hinrich (1970) modifi ed the Grenoble sensor to
use a hydrophone instead of the microphone, and a
brass plate instead of a steel plate. Hinrich (1970)
also installed a hydrophone on an Arnhem sampler
(Hubbell 1964) and used it to verify the sampler
data. Although Hinrich (1970) could recognize
incipient motion, he was unable to calculate trans-
port rates. Anderson (1976) based his microphone
system on that of Johnson & Muir (1969), and sug-
gested that moving sand generates noise dominated
by frequencies above 38 kHz, based on directionality
arguments relating to the microphone that he used.
Anderson also observed 15- and 6-minute periodicity
in the acoustic record. Richards & Milne (1979)
modifi ed Anderson's (1976) system to allow fre-
quency analysis and in two fi eld sites, observed that
the Froude number of the fl ow may impact the
sensor volume, and that the scatter in the acoustic
amplitude was much higher in sand-bed streams than
in gravel-bed streams.
In the marine literature, Thorne and colleagues
(see, for example, Thorne
et al.
1984, 1989; Thorne
1986a,b, 1987, 1993; Thorne & Foden 1988;
Voulgaris
et al.
1995) began with a hydrophone
recording the noise generated by glass spheres in a
rotating drum, then created a theoretical relation
based on the Hertz law of contact, and ultimately
created a fi eld platform where the agreement of
acoustic signals with video recordings and compari-
sons with Doppler velocity transducer current meas-
urements led the authors to conclude that second-scale
temporal variability of gravel transport is dominated
by turbulent bursting events.
Barton (2006) and Barton
et al.
(2005, 2006,
in press) have expanded upon this work, examining
the effectiveness of a hydrophone for fl uvial bed-load
monitoring. Their hydrophone was mounted in near-
bank slack waters of the Trinity River, California,
USA, providing protection from impacts with sedi-
ment and debris, and separation from turbulent
noise. Continuous data were collected concomitant
with manual bed-load measurements using pressure-
difference samplers (Fig. 2.4). Barton
et al.
(2006)
found a signifi cant relation between bed-load trans-
port and the noise generated by the process; the
acoustic signals were exploited to predict the bed-
load discharge between pressure-difference sampling
measurements. Smith (Graham Matthews and
Associates 2006, 2007, 2008) has continued this
work, collecting data at the same location on the
Trinity River.
Rickenmann (1997), Rickenmann
et al.
(1997),
Rickenmann & Fritschi (in press), and Hegg &
Rickenmann (1998, 2000), building on earlier work
by Bänziger & Burch (1990), have shown the effec-
tiveness of accelerometer and geophone (velocity
transducer) installations (mounted beneath a metal
plate installed on the bed) for long-term bed-load
monitoring in the Swiss Alps. Bogen & Møen (2003)
and Møen
et al.
(in press), using a system similar to
that of Rickenmann (1997), but with different fre-
quency sensitivity, have shown that an accelerometer
with a narrow frequency band is heavily infl uenced
by sediment grain size, and that with appropriate
calibration, a wideband accelerometer may be able
to account for changes in the grain size. Richardson
et al.
(2003) also mounted an accelerometer beneath
a steel plate, and found that although the relation
between sediment impact rate and transport rate was
nonlinear (particularly at high transport rates), the
relation was consistent with theory based on shear
stress.
Govi
et al.
(1993) counted impacts recorded by
geophones (velocity transducers) buried in the stre-
ambed immediately upstream from a weir, and were
able to establish streamfl ow discharges correspond-
ing to initiation and cessation of bed-load motion,
but did not calculate transport rates. Burtin
et al.
(2008) used a high-density seismic array in the
Himalayas to monitor the bed-load fl ux qualitatively
in the narrow and deeply incised Trisuli River,
Nepal. Although they were unable to separate con-
tributions to the seismic signal completely owing to
turbulence in the fl ow, they were able to record a
hysteresis loop in the seismic rating curve, indicating
