Home | Contact ST  

Feature Article

Measurements of Ice Parameters In the Beaufort Sea
Using the Nortek AWAC’s New Ice Measurement Capability To Support Engineering and Climate Studies

By Bruce Magnell
Senior Physical Oceanographer and Systems Engineer
Leonid Ivanov
Physical Oceanographer
Woods Hole Group Inc.
East Falmouth, Massachusetts

and
Eric Siegel
Northeast and Great Lakes Region Manager
NortekUSA
Annapolis, Maryland


Measurements of sea ice presence, ice keel draft and ice thickness are important for climate studies as well as design and maintenance of coastal and offshore structures. Measuring long-term trends in total ice thickness is important for validating climate models and observing trends in climate change. Measurements of maximum ice keel depths are required for planning installation and maintenance of underwater infrastructure, such as cables and pipes, in shallow water.

Common ice measurement methods include remote sensing from satellites and aircraft, in-situ ice buoys, and upward-looking sonar. Single beam, upward-looking sonar instruments may be deployed on the seafloor or on a subsurface buoy. These instruments acoustically measure the distance to the bottom of the ice and can estimate ice thickness by subtracting the ice keel location from measurements of water depth (from an onboard pressure sensor). Such sonars have been used widely and successfully for climate and engineering studies.

Building on the success of such upward-looking sonar systems for ice measurements, Nortek (Rud, Norway) recently introduced new firmware and measurement methods to make observations of sea ice with the company’s AWAC acoustic Doppler current profiler (ADCP).

One of the first commercial applications of this new measurement capability was a project to measure ice parameters in Alaska’s Beaufort Sea. The goals of the project were to measure ice thickness during stationary fast ice periods and to determine ice block thickness and size distribution during partial ice coverage periods.

Acoustic echo amplitude from three measurement bursts of the AWAC vertical AST in November 2008 (top panels) during ice formation. The lighter color represents higher acoustic echo strength and indicates the ice/water interface. Analysis of AST echo provides calculation of ice thickness (blue area in bottom panels, depth in meters).

New Measurement Methods
The Nortek AWAC has long been used for routine measurements of ocean current profiles and directional waves using the method known as acoustic surface tracking (AST). In the traditional “wave tracking” mode, the AWAC employs a dedicated narrow, upward-looking vertical acoustic beam to measure the range to the ocean surface, using AST to track the water-air interface by locating the peak amplitude of the acoustic echo, typically at a rate of two to four hertz.

Nortek has now introduced a new capability for the AWAC that allows automatic ice tracking. The AWAC finds the range of the acoustic echo indicating the location of the water-ice interface by locating the leading edge of the peak of the returned acoustic signal. Because the AWAC applies both the max peak and leading-edge algorithms to the AST echo returns, the AWAC does not need to be programmed to switch between wave measurements and ice measurements at a predetermined date. Thus, the device can be mounted on the seafloor or deployed on a subsurface buoy to measure the directional wave field during open-water conditions and also measure ice formation, thickness and maximum keel draft during ice-cover conditions. The AWAC also functions as an ADCP to measure the full current velocity profile.

The AWAC comes in three acoustic frequencies (one megahertz, 600 kilohertz and 400 kilohertz), providing ice, wave and current profiling capabilities over ranges from 10 to 100 meters. Data conversion software provides standard ice measurement statistics, such as mean ice thickness and maximum ice keel draft.


Error Analysis
In order to estimate ice thickness accurately, known sources of instrumental and environmental error associated with the method of using upward-looking sonar measurements and pressure must be corrected in post-processing.

Estimating Sea Surface Elevation. To determine the free sea surface height above the bottom, the measured absolute water pressure must be corrected for atmospheric pressure fluctuations and then scaled by the average density of the water column.

Empirical distributions of mean and maximum ice drafts for October 2008 based on the ASL (blue) and the AWAC (red) data. Click to enlarge.

All pressure sensors have an inherent accuracy limitation, which is typically expressed as a fraction of the full-scale (FS) pressure rating. The AWAC pressure sensor has an error of approximately ±0.5 percent of FS, which equates to ±0.25 meters possible error for a 50-meter FS pressure sensor.

While this error is not known in advance of deployment, it is generally manifested as a constant offset, so it can be compensated in post-processing by applying an offset to match the free sea surface height measured by the pressure sensor with that measured by AST.

The AWAC’s pressure sensor measures absolute pressure, but the data are presented relative to a nominal value of atmospheric pressure at the time of sensor calibration. The data have to be corrected for fluctuations of the actual atmospheric pressure relative to that nominal pressure.

These fluctuations, which are associated with the passage of high and low pressure zones in the atmosphere, occur on daily to weekly time scales and can have an amplitude equivalent to a water height of ±0.5 meters if not corrected. A time-series record of barometric pressure is required to make this important correction. Since atmospheric pressure variations are associated with large-scale synoptic weather systems, atmospheric pressure measurements obtained within a few tens of kilometers from the AWAC typically result in residual errors after correction of less than ±0.05 meters.

The water depth is estimated from the corrected pressure data using values of water density and acceleration due to gravity. Water density is calculated using measured temperature and salinity estimated from proxy information. For example, when ice is present it may be assumed that the minimum water temperature is near the freezing temperature, which depends on salinity. The salinity can be estimated based on this assumption.

A combination of measured water temperature and estimated salinity yields a slowly time-varying estimate of water density and a water depth correction factor related to water density. It is reasonable to assume that the water column beneath the sea ice is fairly well mixed, so vertical stratification (less dense water near the surface) probably does not contribute any significant error in the density calculation.

Estimating Sea Surface Height. The AWAC’s acoustic range measurement technique uses an estimate of the vertically averaged speed of sound to convert acoustic travel-time measurements into distance estimates. The speed of sound is calculated in the instrument from user-input salinity and measured temperature at the instrument’s location. Salinity plays a relatively minor role in sound velocity, so the uncertainty in salinity is not expected to contribute significantly to errors in the acoustic range data. Temperature is important for sound velocity, however, so an uncorrected temperature change of -2.5° C to +2.5° C would result in a depth error of 0.15 meters. By using the measured AWAC temperature to correct the speed of sound, combined with an estimate of salinity, it is possible to reduce the residual error due to density uncertainty to less than ±0.05 meters.

An approximate value of the residual error may be estimated by comparing the pressure-based measurements to the AST ice draft measurements during times when there was open water—for example, just before ice formation or after ice breakup. An offset can be applied to the pressure-based sea surface height estimate to zero out any difference at this time, leaving only small, time-varying errors. As a practical matter, a residual uncertainty (among all variables) of ±0.05 meters to ±0.1 meters remains.


Beaufort Sea Measurements
In August 2008, two Nortek one-megahertz AWACs with ice measurement capability began a one-year deployment in the Beaufort Sea near Prudhoe Bay. The instruments were deployed at a water depth of about 12 meters to measure sea ice coverage throughout the seasons.

The AWACs were configured with enough internal memory and external batteries to measure current velocity, surface wave height and direction while under open water—and ice thickness while under ice cover—for a one-year period. Every 30 minutes throughout the deployment, the AWAC used AST to sample the ice/water interface at a rate of one hertz for 512 samples. For each ping, it recorded the entire high-resolution echo amplitude profile. This high-resolution burst measurement provided information about the mean ice thickness, maximum ice keel draft and ice-block movement, as well as the acoustic characteristics of the water column.

At one AWAC site, an ASL Environmental Sciences (Victoria, Canada) ice profiler, Model IP-5, was also deployed for comparison. The ASL IP-5 uses an upward-looking acoustic echosounding technique similar to that of the AWAC, but the instrument has a narrower beam and lower acoustic frequency (420 kilohertz). A comparison of ice thickness during the period of stationary fast ice shows excellent agreement between the two similar acoustic measurement methods, with no bias and a mean difference of less than 0.05 meters.

Ice formation began in mid-October 2008, when air temperatures were routinely below -10° C. Following a three-week period when thick floating blocks of multiyear ice were present, the ice became solid and its keel draft varied only slowly. The thickness increased almost linearly until it peaked in mid-May at about 1.8 meters. This represents an increase in ice thickness of approximately 0.01 meters per day. The ice began to thin rapidly in late May and ice breakup occurred in early July, with the breakup and transition to open water lasting only a few days.

The 512-point time series of ice thickness in each ice sampling burst provided detailed information on keel draft, from which information on the size and frequency of individual solid blocks of ice could be derived. An ice block analysis was performed to study the time interval for block passage and average thickness. The current velocity data from the AWAC was used to estimate the horizontal length scales of the ice blocks in the direction of movement. This information was used to calculate the distribution of ice block size and mass, assuming symmetrical shape in the horizontal.

Individual ice blocks and their parameters were identified in the AWAC and ASL data for October 2008. Results showed that 70 percent of the ice blocks had a keel draft greater than two meters and 10 percent had an ice keel greater than three meters.


Conclusions
Results from the first long-term, commercial deployment of the Nortek AWAC under the ice in the Beaufort Sea, along with a comparison between the AWAC and a colocated ASL ice profiler, have demonstrated that the AWAC provides high-quality ice keel draft measurements.

With proper data processing and corrections for atmospheric pressure, temperature, salinity and tilt, the AWAC’s independent estimates of free sea surface height when ice is not present agree with the ASL ice profiler, typically to within ±0.1 meters or better. Detailed time series of ice keel depth, in combination with the AWAC’s current profile data, permit estimates of the size and movement of floating ice blocks to be made. The AWAC’s combination of current profiles, directional wave measurements and accurate, detailed ice keel draft measurements makes it a valuable new option for ice studies.


Acknowledgments
This work was performed for BP Exploration Alaska Inc. (Anchorage, Alaska) by Woods Hole Group Inc. under Contract No. BPM-06-00027, WRN 9489, NSU 831762. Woods Hole Group thanks BP Alaska for permission to publish the results of the study and gratefully acknowledges the help provided by numerous organizations and persons, especially the logistical support and air freight services provided by BP Alaska and shore and vessel support at Prudhoe Bay provided by Alaska Clean Seas.

Atle Lohrmann and Torstein Pedersen from Nortek AS also provided assistance in configuring the instruments and contributed valuable insight on the details of the AWAC signal processing and error analysis.



Bruce Magnell is a physical oceanographer and ocean system engineer specializing in designing measurement programs, selecting instrumentation, evaluating instrument performance and data quality, and performing physical process studies.

Leonid Ivanov is a physical oceanographer who is mostly focused on various types of oceanographic data analysis and interpretation.

Eric Siegel is a physical oceanographer and the global segment manager for academics at Nortek. He collaborates with clients to develop new applications and innovative oceanographic measurement solutions.




-back to top-

-back to to Features Index-

Sea Technology is read worldwide in more than 110 countries by management, engineers, scientists and technical personnel working in industry, government and educational research institutions. Readers are involved with oceanographic research, fisheries management, offshore oil and gas exploration and production, undersea defense including antisubmarine warfare, ocean mining and commercial diving.