Range compensation in pack ice imagery retrieved by coastal radars

Abstract

Finnish Meteorological Institute has instrumented coastal radars with radar servers for coastal ice field monitoring. The servers capture the radar data and processes one image per revolution with given parameters. The image time series provide real time high temporal and spatial resolution data on ice characteristics and are accompanied by kinematic products. The data are used in research and provided to end users, mainly icebreakers. A basic problem in the use and analysis of radar data is that the intensity of the ice signatures decreases with range. When uncompensated this effect decreases the usefulness of the images as navigational aid. More importantly, the compensation is required when the images are classified or quantitative ice parameters are retrieved, especially for ridging. The intensity decrease is mainly due to the physical properties of the radar, but also the ice area that is shadowed by ice ridge sails increases with distance. The range compensation problem is approached by three different methods. The observed decrease of intensity can be used the make a range dependent scaling for any image. Secondly, order statistical methods not dependent on absolute intensity values can be applied for more universal approach. Third, homomorphic filtering provides a method independent on assumptions concerning on the characteristics of the ice field. The three methods are compared for selected cases.