Seasonal ice -i.e. sea ice that does not survive the melting season- occupies an expanding fraction of the Arctic at the expense of multiyear ice. A system able to measure ice thickness and heat fluxes between the ocean, ice, and atmosphere, not only in thick ice, but also withstanding thin ice conditions is therefore highly relevant.

As a step forward in this direction, we have been developing an autonomous lightweight prototype buoy, called "Ice-T" (for "Ice-Thickness") since 2005. Ice-T provides real-time measurements of ice thickness (discriminating changes occurring at the base and at the top of the ice layer), thermal profiles through the ice and snow, atmospheric pressure, air temperature, ice drift, and horizontal current, temperature, and salinity at the base of the ice. The Ice-T prototype has been deployed with success in already formed ice, but being a floating instrument, it can in principle be deployed in open water (before ice formation), making the study of thin ice possible. This would have the advantage to monitor the initial stage of ice formation, or regions where ice cover is intermittent, such as the very important polynya regions where dense water forms. This capability has not yet been tested in the field.
The concept of the instrument complies with two main constraints: (1) minimal cost, as sea-ice is a highly destructive medium (2) versatility, the Ice -T float acting as a platform that could host additional sensors (in particular atmospheric sensors, an important aspect of the current OPTIMISM project). The constraint of minimal cost, which is our main risk mitigation strategy, has led to a concept of relatively small-sized, resistant, easily deployable instrument. The counterpart is that the instrument can embark only a limited additional payload.
The buoy is autonomous for one year and its data is real time transmitted through the iridium satellite communication system.

Figure 1
Figure 1

Ice-T buoy schematics

The Ice-T system includes two floats: a buoyant surface float, caught in the ice sheet, and a weighting subsurface body (hereinafter the « fish ») 5m below, connected with a cable for transmission of data and energy (Figure 1).
The surface buoy is equipped for satellite real-time data transmission (iridium system), includes a thermistor chain on its side to measure temperature profiles within the snow and ice layers (a thermistor is reserved for air temperature measurement), and an inclinometer to correct temperature gradients should the surface float be trapped tilted in the ice. Additional sensors include an atmospheric pressure sensor (used also for reference in the computation of ice thickness), and a GPS receiver for positioning and to estimate the ice drift.
The fish is equipped with an upward looking sonar altimeter, a pressure sensor, and an inclinometer/compass. These are designed to measure the ice keel draft and therefore the ice thickness, assuming the sea water density, ice density and snow load to be known.
The accuracy on ice thickness measurement is on the order of 2 cm. Most of the error is owed to uncertainties in the snow load and ice density. Some of these parameters are (or will be) available from the instrument, reducing the uncertainty. Recently, the fish has been equipped with a high-accuracy conductivity (salinity) and temperature sensors, which makes it possible to determine the freezing temperature of sea water, which depends on salinity, and therefore the heat available to melt the ice, knowing the in situ temperature.

Estimating ocean currents: the poor man's current meter.

The "fish" is freely hanging in the water (a deliberate choice to reduce the size of the instrument and therefore ease the deployment logistics) and will therefore tilt and rise under the action of drag and lift forces caused by ocean currents relative to the ice.
A substantial amount of work has been devoted to finding a shape and mass repartition that ensure a correct hydrodynamical behavior of the fish and enhance the response of the fish at low speeds. Since its tilt is measured very accurately, numerical simulations, confirmed by towing tank experiments, indicate that we should be able to measure horizontal currents to within an accuracy better than 2 cm/s, which is the accuracy of a manufactured Doppler current-meter. The relationship between velocity, tilt and vertical displacement of the fish is accurately known and can be updated for any in situ density (which affects viscosity and Archimede's forces) thanks to a numerical simulator.
The buoy therefore provides relative ocean-ice currents (and hence absolute ocean currents since the ice drift is measured from GPS data) at no extra cost. The surface current is of great interest to study the ice drift. It is also a critical parameter for the sea ice mass balance as it is needed, together with the in situ temperature and freezing point to estimate the ocean-ice heat flux.

Tilt of the fish as a function of velocity (right), simulated and measured during towing tank experiments (left)
Tilt of the fish as a function of velocity (right), simulated and measured during towing tank experiments (left)

The thermics of the float has also been thoroughly studied (finite element modeling) so that to ensure that the instrument does not perturb the temperature field that it is supposed to measure. The electronics of the float take advantage of the capability provided by the iridium network, in particular the duplex mode, and one can upload at any time a new software: this allows for instance to increase the sampling frequency if a particularly interesting phenomenon is being observed, or on the contrary move to an energy saving mode to extend the lifetime of the mission.


The Ice-T project has been developed by F. Vivier and A. Lourenco at LOCEAN since 2005, with contributions from many in the marine techniques and instrumentation team at LOCEAN. Starter funds for this project came from Institut Pierre-Simon Laplace (IPSL).
As of 2006, the electronics was developed by DT-INSU (A. Guillot), and development pertaining to the hydrodynamics were done in collaboration with the Fluid Mechanics Laboratory at Ecole Centrale de Nantes.
After preliminary in situ tests in 2006, the prototype was completed early 2007 and succesfully deployed in March-April 2007 in the Storfjorden (Svalbard) as part of the “IceDyn” campaign (project 1058 funded by IPEV).

Validation of the prototype


The prototype was deployed in March/April 2007 in the Storfjorden (Svalbard) for a 6-week period. as part of the “IceDyn” campaign (funding IPEV), nearby “Vagabond”, a polar yacht wintering in the frame of the EU DAMOCLES program. This validation test was successful: the float fully worked during the 6 weeks of the experiment, providing accurate ice thickness observation and thermal profiles. These observations also demonstrate the capacity of the instrument to measure changes in the snow load, either through underwater pressure data (see example in the figure below), or in the evolution of the vertical thermal gradient. Finally, observations demonstrated the good accuracy of Ice-T velocity estimates by comparison with ADCP data (RMS difference smaller than 1cm/s ).

Thermal profiles in the air,snow,ice and water layers (with ice thickness acoustically measured superimposed, thin black line). Note that a heavy snow fall is detected on April 14 by a lowering of the float by nearly 2cm (visible as well as in the temperature profiles)
Thermal profiles in the air,snow,ice and water layers (with ice thickness acoustically measured superimposed, thin black line). Note that a heavy snow fall is detected on April 14 by a lowering of the float by nearly 2cm (visible as well as in the temperature profiles)