| A method of describing the large-scale dynamic and thermodynamic behaviour of pack ice in terms of its thickness distribution (a probability density function of ice thickness) was developed earlier and used in a numerical model of the Arctic ice cover. The main purpose of the present research was to extend this thickness distribution theory by including ridged and level ice explicitly, and to investigate the behaviour of the ridge redistribution function and the role of the various parameters which describe the ridging process. Calculating ridged and level ice thickness distributions separately allows more insight into the formation, distribution, and evolution of ridged ice and enables more comprehensive comparison to observations.; This more general theory was incorporated into a large-scale sea ice model, termed the multi-level model, and used to simulate the Arctic ice cover using 7 years of observed forcing data (1979-85). Evolution of the ice cover over a 21 year period, beginning with a uniform level ice cover, was studied by repeating this 7 years of forcing three times. This length of time was necessary to allow the ridged ice statistics to reach a seasonal equilibrium. Computed thickness and ridge statistics were compared to submarine sonar measurements from 1958 and 1970. Compactness fields based on SMMR sattelite observations from 1980 to 1983 were also compared to the model results. In addition, a number of sensitivity studies were conducted to investigate the role of various physical processes and their parameterization in the model. Of particular interest here was effect various parameters describing the ridging process and the functional form of the so-called ridge redistributor had on the computed thickness and ridge statistics. Finally, a similar extension of the simpler and widely used two-level sea ice model was developed, with the results of large-scale simulations compared to those of the more complete multi-level model. |
