Probabilistic assessment of ice rose diagrams for ice drift in the beaufort sea

Abstract

The intention of this paper is to introduce the ice rose diagram, which is a graphical tool that makes it possible to display the ice drift speed, the occurrence frequency, and the drift direction at a particular location for a given observation period. The ice drift data is obtained from subsurface local measurements with an acoustic Doppler current profiler (ADCP) which forms the basis for construction of the ice rose diagram. The data is collected during different years from 2010 to 2017 by ADCP instruments, which are installed by means of mooring buoys in the Beaufort Sea. Probabilistic models are introduced in order to fit the data for ice drift speed and drift direction. It is found that the sea ice in the Beaufort Sea tends to drift with different speeds during the winter (i.e. the growth season) versus during the summer (i.e. the melting season). The drift speed is generally lowest during the growth season and highest during the melt season. The Weibull probability model provides the most appropriate fit for the data of ice drift speed during the growth season. For the melting season, a mixed distribution provides the best fit for the ice drift speed. Regarding ice drift direction, a mixture of von Mises distributions provides the most appropriate fit for both the growth and the melting seasons.