"noisy" when scattering material in the water is low. Low SNR was a problem
during the fixed-bed hydrodynamic experiments, as discussed in Chapter 9. To
solve this problem, the manufacturer recommended inserting into the water
particles having a density near that of water and a mean diameter of 10 to 20
microns. Neutrally buoyant particles remain in suspension without additional
stirring, and the size recommended provides a relatively strong echo per unit of
concentration. The manufacturer recommended using hollow glass spheres with
a mean diameter of about 10 to 20 microns. However, this special seeding
material could only be purchased in large quantities at a cost of 0 for 20 kg.
After much trial and error, it was determined that the natural soil which
exists in the Vicksburg, MS, area is an excellent soil to use as a seeding material.
Therefore, prior to each fixed-bed hydrodynamic experiment, two 20-ℓ buckets of
dry soil were distributed throughout the testing region of the facility.
Immediately prior to the experiment, a hand-held rake was dragged through the
water to stir the soil up into the water column. This increased the SNR readings
from as low as 5 db up to 15 to 25 db even before waves were generated. After
waves were generated, the SNR reached values ranging from 30 to 50 db, which
is more than adequate to ensure high quality velocity data.
It should be noted that for the moveable-bed LST experiments, the sand used
to construct the beach provides sufficient scattering material while waves are
generated to produce SNRs of 30 to 40 db, and therefore natural soil does not
need to be added as a seeding material.
Quality of current data
adversely affected by the presence of air bubbles in the flow. The speed of a
sound wave is changed by the presence of bubbles, resulting in a significant error
in the velocity measurement, Hughes (1993). This is a problem encountered in
the LSTF when making velocity measurements in very shallow water, since
breaking waves force air bubbles down into the water column, sometimes to the
depth of the ADV sensor. If this occurs, erroneous spikes appear in the velocity
time series. However, these spikes are removed during postprocessing using a
GEDAP filtering routine, developed specifically to remove spikes with these
characteristics. Waves that break farther offshore do not cause a problem,
because the air bubbles do not generally penetrate down to the depth of the
The manufacture's performance specifications for the ADVs indicate an
accuracy of +1 percent of the measured velocity, or +0.0025 m/sec for velocities
less than 0.25 m/sec. However, it is very difficult to verify these specifications in
a facility that holds 1,350 m3 of water, because the water rarely is totally
motionless. Nonetheless, tow tank tests have been carefully conducted using the
bridge and to tow all 10 ADVs through the water at an average speed of 0.2 and
0.3 m/sec. A stopwatch was used to measure the elapsed time as the bridge
moved at a constant speed over a distance of 20 m. These tests were conducted
for both the x- and y-axes of the ADVs and also were used to verify that the
81
Chapter 8