A Configurable Multi-Sensor Tripod
For the Study of Near-Bottom Ocean Processes
R. Birch
1
, P. R. Hill
2
, M. Clarke
1
, D. Lemon
1
, & D. Fissel
1
1
ASL Environmental Sciences,
1986 Mills Road,
North Saanich, BC, Canada, V8L 5Y3
asl@aslenv.com
Abstract- A cost effective and highly configurable scientific
platform, dubbed Norton, has been developed for the scientific
study of near-bottom ocean processes. The basic platform is a
collapsible lightweight aluminum tripod that can be
transported in the back of a pick-up truck, and is easily
accommodated on most vessels. Ballast is supplied in the
form of large ship anodes bolted to the base of each leg. The
desired sensors are attached to the tripod legs and bracing. A
typical deployment might utilize the following sensors:
· Acoustic Doppler Current Profiler for current profiles and
directional wave data.
· Marsh McBirney electromagnetic flow sensor for
measurements close to the bottom, supporting both wave
orbital burst sampling and average current measurements.
· Optical Backscatter sensors (OBS) for sediment
concentration.
· Sector-scanning sonar to supply bed-form images; an
Imagenex variable frequency scanning head has been
integrated into a self-contained housing containing a control
module, data logger and power supply.
Other sensors can be added as desired. Norton is usually
deployed by lowering with a line from the ship. An acoustic
release is used to let go of the tripod once it is on bottom. A
customized ‘tilt-pinger’ provides confirmation to the crew
that the tripod is upright. Recovery can be via a pop-up buoy
and/or a ground-line.
I. INTRODUCTION
As part of the Georgia Basin Geohazards Initiative, the
Geological Survey of Canada is studying erosion and
sediment transport on the Fraser River delta. The
Geological Survey of Canada has had considerable success
on Canada’s east coast using a platform termed Ralph [1],
[2]. When a similar requirement for sediment transport
measurements on the Fraser River delta became necessary,
ASL was contacted to build a suitable instrument platform.
While its origins are obscure, there seems to be a trend in
naming bottom tripod/frames after characters from the old
Honeymooners TV series. The New Zealand National
Institute of Water & Atmospheric Research called theirs
Alice, after the lead female character, and as mentioned, the
Geological Survey of Canada on the east coast named
theirs Ralph, after the lead male role. The Ralph platform is
quite big and burly, much like the actor Jackie Gleason
who played that role. When the Canadian west coast
version was built, it was more lightweight and mobile, and
we named it Norton after Ralph’s scrawny and somewhat
idiosyncratic sidekick. Characteristically, Norton’s
aluminum legs buckled slightly during the latest recovery,
after encountering a much larger mass (the recovery
vessel).
2
N
atural Resources Canada,
Geological Survey of Canada,
P.O. Box 6000, Sidney, BC, Canada, V8L 4B2
phill@NRCan.gc.ca
GSC Publication No. 2003086
II. NORTON TRIPOD CONFIGURATION
A. The Tripod
The Norton platform is an aluminum tripod with a base
dimension of about 3 m and an overall height of 2.2 m.
Aluminum was chosen for its light weight and resistance to
corrosion. Large ship anodes are added to the base of the
legs for stability as well as corrosion protection. A mid-
section brace adds strength as well as a means of attaching
instruments. The legs are hinged and with the mid-section
removed they collapse, making the platform easily
transportable (it fits in the back of a pickup truck).
Fig.1. Norton being recovered from its inaugural
deployment off the Fraser River delta.
B. Sensor Suite
1234
presented at Oceans 2003, San Diego
Norton is easily configurable with whatever sensors are
required. Instruments such as OBS sensors and Doppler
current profilers are generally attached to the tripod legs
whereas the scanning sonar and electromagnetic flow
sensors are normally mounted in the open area directly
under the tripod.
At present the instruments are stand-alone with each having
its own power supply and data logger. Time
synchronization is used to later merge the data streams.
This approach provides a highly configurable suite of
various sensors that can be optimized for a particular
application, and it also ensures simple field assembly and
operation.
Instruments that have been used on Norton to date include:
• Coastal Leasing MiniSpec current/wave gauge
. This
instrument is equipped with a Marsh McBirney flow
sensor for measuring near-bottom currents, and in
burst mode measures the wave orbital velocities. It
also has a Paroscientific pressure sensor for wave
height measurements. When combined with the
orbital velocity data, wave directions can also be
determined (PUV method).
• Imagenex 881A Scanning Sonar
. The multi-frequency
(175-1000 kHz) scanning sonar head was integrated to
a custom-built data logger/power supply. A large
alkaline battery pack and up to 1 GB of data storage,
allow the instrument to be deployed for 8-10 months
using an hourly scan rate.
• OBS Sensors
. Norton is usually outfitted with two
D&A Optical Back Scatter (OBS) sensors, at different
heights above bottom. The OBS sensors are stand-
alone using an Applied Microsystems CTD platform
for power and data logging. Sediment samples are
usually collected from the site to allow post-survey
conversion of the OBS sensors output to suspended
sediment concentration units.
• Acoustic Doppler Profiler
. Doppler current profilers
can be added to provide current profile data, as well as
(directional) wave data.
C. Deployment / Recovery
To date Norton has been deployed twice off the Fraser
River delta, once in 10 m water depth, and most recently on
a tidal bank that dries at low tide. On both occasions, the
Canadian Coast Guard Hovercraft Siyay was used for
deployment and recovery. At the 10 m depth site, the
tripod was deployed with a ground-line and an acoustic
release-activated pop-up buoy at the end of the ground-line.
During recovery the pop-up buoy failed to surface and the
ground line was snagged with a grapple. In the
oceanographic community there is often some doubt as to
whether a bottom frame is upright after deployment. ASL
has developed a tilt-pinger that pings only if it is within 20°
of vertical, providing assurance that the tripod was
deployed upright.
On the tidal bank, the deployment was relatively easy. The
hovercraft settled onto the sand bank at low tide and the
crane was used to swing the tripod over the side. The field
team then disembarked to measure the tripod heading,
collect sediment samples, take photographs, and measure
the bedforms. When the tripod was recovered a month
later, the tide was higher but the top of the tripod was
above water and easily recovered.
III. TWO DEPLOYMENTS ON ROBERTS BANK, THE
FRASER RIVER DELTA
A. Environment, Background Oceanography
The Strait of Georgia separates Vancouver Island from the
BC mainland. The Fraser River enters the ocean at
Vancouver and many of its suburbs are located on the
subaerial delta. The main channel is dredged and the
Steveston Jetty was built along the northern portion of the
channel extending 5 km into the Strait. This and other river
management projects have deprived some areas of the delta
front of much of the sediment formerly received during the
peak spring discharge.
Tides within the Strait are mixed, mainly semi-diurnal,
with ranges up to about 3-5 m. Tidal current speeds can
exceed 1 m s
-1
. The Strait has limited fetch for wave
generation with the largest waves affecting the delta
coming from the west-northwest (approximately 80 km
fetch in this direction). Wave heights are generally less
than 0.5-1 m [3]. Significant waves heights up to about
3.4 m height, and about 7.5 second period, have been
recorded in the deeper water off the delta (Fisheries and
Oceans Canada station #108; http://www.meds-sdmm.dfo-
mpo.gc.ca/meds/Databases/Wave/).
Fig. 2. 2001 Landsat image of the Fraser River delta
showing the 2002 site (
■=Norton) in 10 m water depth,
and the 2003 locations in 8 m (
■=ADCP) and 0 m
(
■=Norton) depths. Roberts Bank is outlined in yellow.
1235
presented at Oceans 2003, San Diego
Roberts Bank is located just south of where the main
(South Arm) Fraser River channel enters the Strait of
Georgia. The substrate of Roberts Bank is primarily sand
with a mean grain size of about 0.125 mm [4]. The sand
becomes finer-grained towards the shore.
It is generally thought that much of the slope of Roberts
Bank is being eroded, with the possible exception of the
northern portion of the Bank. There is evidence for sand
from the South Arm becoming incorporated in a clockwise
gyre across the north half of Roberts Bank [5]. Numerical
model results indicate the formation of a clockwise eddy on
the Bank during weak flood tides, apparently caused by the
Steveston Jetty [6]. The Steveston Jetty provides partial
protection from winter storm waves from the west-
northwest. The dominant tidal currents and breaking waves
can generate a net longshore current over Roberts Bank [7].
B. Winter 2002 Deployment on Roberts Bank
The Norton tripod was deployed in 10 m water depth off
the edge of Roberts Bank (
■ in Fig. 2 above). Previous
multibeam surveys show outcropping beds aligned roughly
north-south in the 5-15 m depth range, indicative of net
erosion [8]. Further offshore (20-30 m depths) were
subaqeous dunes aligned northeast-southwest,
perpendicular to the predominant direction of tidal flow.
The currents were tidally dominated with flood tides
(northerly) stronger than ebb, up to 1 m/s versus 0.4 m s
-1
respectively. The peak tidal flows re-suspended bottom
sediment on both the flood and ebb.
Several storm events were recorded with significant wave
heights up to 1 m, and 5 second peak periods. The
scanning sonar (set to 30 m range and default gain and
frequency settings) showed high suspended sediment
concentrations just trailing the storms and the emergence of
long wavelength (c. 10 m), low-relief, irregular patches [8].
C. Preliminary Results from the Winter 2003 Deployment
on Roberts Bank
The 2003 deployment of Norton was at a location on
Roberts Bank where bedforms were expected to consist of
small ripple features about 2 cm high with a 5-10 cm
wavelength. The Imagenex scanning sonar settings were
optimized for this deployment using tank tests.
A series of three ripples were created in the laboratory
using fine-grained sand to simulate the expected bedforms
on the Bank. Frequency and gain settings were then varied
until optimal sonar images were obtained (Fig. 3). Echo
returns at approximately 1.6 m range are from the tank
wall. The final settings chosen were: 1 MHz frequency, 5
m range, and 18 dB gain.
Fig. 3. Imagenex sonar image (left) of test ripples (about 1-
2 cm height, about 10 cm length) in tank. Red range rings
are 1 m increments. Three sand ripples are shown starting
at 1 m range (first red ring).
The Norton tripod was deployed on Roberts Bank (
■ in
Fig. 2) at low tide (Fig. 4 below).
Fig. 4. Deployment of Norton on Roberts Bank March 23,
2003 using the Canadian Coast Guard Hovercraft Siyay. A
radar reflector and flashing light were mounted on a pole
above the tripod.
A 1200 kHz RD Instruments WH ADCP™ with Waves
firmware was deployed in 9 m water depth just offshore of
the edge of the Bank (
■ in Fig. 2). The ADCP was to
measure the ‘offshore’ waves so that we could determine
the wave attenuation over the Bank. During the
deployment, winds were blowing from the northwest and
waves about 1 m high could clearly be seen breaking on the
edge of the bank, reducing in height as they moved onto the
Bank.
Using the settings previously determined in the laboratory,
the Imagenex scanning sonar head was suspended from the
tripod at a height of 45 cm above bottom.
1236
presented at Oceans 2003, San Diego
The scanning sonar often recorded radial patterns, with a
generally low reflectivity background (Fig. 7 below).
These images occurred frequently and nearly always during
a lower high tide (consecutive high tides are different
heights due to the diurnal inequality), and not when
successive low tides were of similar height (more purely
semi-diurnal tide). OBS levels (0.3 m off bottom) were
low at these times (about 10 FTU, up to 20-30 max).
Currents were generally weak with speeds of about
At the time of deployment (Fig. 5 below) the bedforms
consisted of sand ripples about 1-2 cm high, and about 5-10
cm wavelength, similar to the ones fabricated in the test
tank. The ripples were not continuous, nor linear, and far
more complicated than the ones created for the tank test.
20 cm s
-1
.
Fig. 5. Bedforms at time of deployment: ripples about 1 cm
high, with about 7 cm wavelength. Overall trend in ripples
was approximately northeast-southwest.
The first images from the scanning sonar (after the tide had
risen enough to submerge the instrument) show ripple-like
features, particularly in the northwest quadrant with
wavelengths of the same magnitude as seen visually (Fig.
6). The ripple features are oriented northeast-southwest.
The reversed-grey color option best shows the ripple
features.
Fig. 7. Radial-pattern sonar image (image 040909).
Another interesting occurrence was high suspended
sediment concentrations at diurnal time scales, centered on
the lower low tides. Sonar images such as the one below
were recorded just prior to, and immediately following, the
emergence of the sonar head from the water as the tide was
falling.
The suspended sediment signal from the OBS sensor 30 cm
above bottom clearly shows the diurnal nature of the signal
(Fig. 9 below). The ‘dropouts’ in the middle of the peaks
are when the sensor was in air at low tide. Waves from the
west added to the high OBS readings on the 21
st
.
Fig. 6. First image (image 03231900) from the scanning
sonar after the tide had submerged the sonar head. Tripod
legs and instruments show up darkest, with ripple-like
features evident, particularly in the northwest quadrant.
Parallel lines are used to highlight ripple features having a
NE-SW orientation.
Ripple-like features were occasionally evident on sonar
images at other times, but their characteristics (wavelength,
orientation, etc) did not vary significantly from those
observed initially. Several ‘storms’ occurred during the
deployment, but significant wave heights in the deepwater
increased to only about 1 m, and the sonar images
following the storms were not noticeably different than
before.
Fig. 8. High suspended sediment concentration during
moderate westerly winds and near low tide (image
04051200).
1237
presented at Oceans 2003, San Diego
Fig. 9. Uncorrected OBS readings from the sensor 30 cm
off bottom, showing high suspended sediment
concentrations at lower low water.
These OBS diurnal peaks occurred only during the lowest
low tides, the lower low tides that occurred when the
diurnal inequality was most pronounced. Lower low tides
are accompanied by strong ebb and flood currents, up to 50
cm s
-1
ebb and 125 cm s
-1
flood, as recorded by the ADCP
located in 8-10 m water depth off the edge of the Bank
(Fig. 10). Note that the flood tide associated with the
preceding higher low tide is negligible.
Data from the Norton sensors help explain the processes
causing these high suspended sediment values (Fig. 11).
Starting at high tide, suspended sediment values are low
and currents weak (near slack). As the tide ebbs and the
water level drops, the current speed increases to 40-50 cm
s
-1
, and suspended sediment levels increase. As low tide
nears, the water level falls below the level of the sensors
which are then exposed to the air for up to 7 hours. During
this period of low tide, we expect that suspended sediment
levels are low, due to the ‘slack’ current. As the tide turns,
the flood waters submerge the sensors. The flood tide is
already running fairly quickly, and OBS levels are high.
As slack high tide is approached, the current weakens and
OBS levels fall.
The high suspended sediment levels occur during the
strong ebb/flood currents associated with the lowest of the
low tides. Sediment re-suspension occurs when the current
speed (as measured by the sensor 80 cm above bottom)
reaches about 30-40 cm s
-1
. This is consistent with a 25 cm
s
-1
threshold current for re-suspension of fine sand
(0.125 mm), based on the Hjulstrom Curves.
Fig. 10. ADCP current profile data for April 22, 2003 (north component of velocity is in mm s
-1
). The speed direction time
series for a mid-depth is shown in the lower portion, along with the pressure time series.
1238
presented at Oceans 2003, San Diego