Mathemaical models : for a better undestranding of the dynamic of the icepack

Make a model for a better understanding. Everything must therefore be done to arrive at a better understanding of the secrets of the austral marine ice. Among the methods used to palliate the lack of observations from the field is the modelling of ice.

Models are representations of reality. Working from theoretical observations and studies, one tries to isolate the mechanisms governing the behaviour of the ice, mechanisms that are formalised by mathematical equations. That is what we call a model.

This method is practised, among others, at the Georges Lemaître Institute of Astronomy and Geophysics of Louvain-la-Neuve within the cell run by the Belgian climatologist André Berger. Hugues Goosse, a researcher and member of the team, explains: "Models are representations of reality. Working from theoretical observations and studies, one tries to isolate the mechanisms governing the behaviour of the ice, mechanisms that are formalised by mathematical equations. That is what we call a model. One then solves these equations thanks to electronic calculation capabilities, while hoping that the results obtained will be sufficiently close to reality. If this is not the case, one seeks to improve the model by taking into account what had been hitherto overlooked and by altering certain processes until a satisfactory result is obtained.

" This approach is in fact close to the classical scientific approach whereby one carries out laboratory experiments that enable the development of a theory, a theory that one improves in the light of subsequent experiments. Computer simulation in a way provides an additional instrument in the panoply of science. It must however be stressed that models can never replace observations. In effect, one must have field measurements in order to know how to represent the physical phenomena, for knowing the value of the major parameters such as the rigidity of the ice, the friction coefficient between the ice and the air, as well as for checking the results. Once a satisfactory model has been obtained, it can be used for performing all kinds of computer simulations. One of the ultimate goals of such experiments is an enhanced understanding of the behaviour of marine ice. Models are fantastic tools because they enable one to obtain a simulated value (the speed of the ice, for example) at each moment and in a very great number of points. It is therefore possible to carry out more detailed studies of the field observations than at the outset, which, in Antarctica, are relatively rare. One can also conduct experiments by changing the conditions in order to postulate hypotheses and analyse their implications. For example: if one wants to have an idea of the role of the drift of the ice on the ocean, it is, in a model, a fairly "simple" matter. It is "sufficient" to block the ice and to compare the results obtained from the mobile ice with those obtained by considering it to be fixed."

Such scientific instrumentation is not necessary for describing the formation process of marine ice. It has been known for a long time, in effect, that when salt water freezes at -1,8°C, it first cerates tiny crystals of ice that float on the surface of the sea - it is at this stage known as skim ice - while waiting to find itself transformed into platelets and needles of ice. When the water is rough, a sticky mixture of ice crystals forms beneath the surface; this is frazil ice. The sight of circles with raised edges resembling water-lilies is a familiar picture of seawater that is in the process of freezing. Joined one to another, these pancakes of ice leave thin strips of visible hardly-frozen seawater between themselves. This patchwork of course ends up by being consolidated, as does the frazil ice; if snow falls at that particular moment, it will form a thin carpet that will thicken with time and add to the ice floe.

Until now, the satellites that have been observing the Antarctic continent since the middle of the seventies have not provided accurate measurements of the thickness of the ice floe; the microwave or infrared radiometers that observe it from an altitude of 36,000 kilometres can determine the percentage of seawater covered by ice, but nothing more. It is however recognised that the thickness of the Antarctic ice floe varies from between 30 centimetres and several metres, with an average measurement of 1 metre - as against 3 metres in the Arctic Ocean. This is the result of the conditions that formed it, of the movements that it accomplishes, and of the inevitable deformations that ensue.

The thickness of the Antarctic ice floe varies from between 30 centimetres and several metres, with an average measurement of 1 metre - as against 3 metres in the Arctic Ocean.

Phenomena such a the relative stability of the properties of the ice floe or the presence of polynyas ("polynya" is the Russian word for "lake"), vast stretches of open water that here and there punctuate the areas of ice floe, for their part also remain something of a mystery. The polynya of the Weddell Sea - a 50,000km² lake - has been observed on three occasions between 1973 and 1976, but has not reappeared since. Although subjected to the bombardment of scientific computation, it has yet to deliver its secret. Some scientists attempt to explain this phenomenon by the contribution of the warmth of the ocean or by the thermodynamic effect of the wind: others advance the hypothesis according to which the Antarctic Ocean possesses in its entrails a selective memory that would be capable of producing from one year to the next - as though moved by its own momentum - an important event like the polynya of the Weddell Sea. The hypotheses stop there.