Estimating sea ice thickness distributions and modeling their evolution in time

The true nature of the sea ice thickness distribution in polar oceans continues to elude the Earth sciences because data sampling requirements are large and data collection capabilities are small. Thorndike (1992), Maykut (1986) and Rothrock (1986) have described the thickness distribution mathematically and have developed techniques for modeling its evolution in time using dynamics, thermodynamics and random processes as the underlying physics to help drive what is largely considered a stochastic process. Lipscomb (2001), Hunke (2001) and Kwok (2000), among others have done more work to model the evolution of the sea ice thickness distribution, but these efforts often lacked sufficient data to validate the resulting models. Remotely sensed imagery collected by satellites has already started serving as a substitute for in situ measurements to estimate sea ice motion and ice thickness parameters, most recently as part of the RADARSAT RGPS work undertaken by Kwok and others (2000). In a similar manner, this paper´s author makes use of data derived from satellite imagery, namely sea ice stage of development and concentration data contained in digital sea ice charts available from the National/Naval Ice Center in Washington, DC. With the advent of digital sea ice charts, based on satellite imagery interpretation, which can be accessed by geographic information systems (GIS), the tedious task of extracting quantitative information related to ice thickness from sea ice charts has become greatly simplified. The author both develops techniques for estimating the ice thickness distribution from digital sea ice charts and presents the fundamental physics required to model the time evolution of the distribution. Results of the model are then compared to subsequent estimated thickness distributions derived from sea ice charts. The work here constitutes an intermediate step toward true validation of sea ice thickness models.