Author(s): Juha Karvonen; Istvan Heiler; Jari Haapala
Linked Author(s):
Keywords: No Keywords
Abstract: Small-scale ice drift in high temporal resolution can be utilized in ship routing and ice model validation. This kind of information is not available from the remote sensing instruments, such as SAR-based instruments, typically used for operational sea ice monitoring, because both their spatial and temporal resolutions are relatively low. In a national project ShipsensorNet the use of ship and coastal radars have been studied for local ice drift estimation. One principal idea in the project was to study the possibilities of a network of ships radars and coastal radars as an information source for navigation. One part of this work has been to study the estimation of ice drift and related parameters from radar data. For these experiments we have collected data from one coastal radar and some ship radars during the winters 2007-2008and 2008-2009. The ice motion estimation has been performed using an algorithm based on local phase correlation between radar image pairs. Before the drift estimation, efficient preprocessing of the data is required, because radar data is often noisy. Also the ship motion, in the case of a ship radar, must be compensated. We have performed ice drift estimation for some interesting radar image time series, and also computed some derived parameters for the drift vector fields, such as ice drift vector field divergence or convergence, ice drift vector field rotation (curl) and strain. The algorithms, and results of the computations are presented and their relation to ice models and usefulness for ice navigation is discussed.
Year: 2010