Thomas Lavergne, Norwegian Meteorological Institute, EUMETSAT Ocean and Sea-ice Satellite Application Facility
Sea-ice motion from the OSISAF
Figure above: Sea-ice motion in Ross Sea from 14-16 August 2010. The colour map scales with displacement in km during the two days.
The sea-ice motion processing at the OSISAF is based on the Continuous Maximum Cross-Correlation algorithm (CMCC, Lavergne et al 2009, which allows sub-pixel accuracy in the retrieved ice motion vectors.
The product is based on low-resolution satellite microwave imagers, both active and passive, such as the Special Sensor Microwave Imager and Sounder (SSMIS), and the Advanced SCATterometer (ASCAT).
The same algorithm, applied on the same instruments was recently deployed for processing sea-ice motion vectors over the Antarctic regions (Figure 1).
The result is a 48-hours ice drift product (each vector recording the displacement of a sea-ice parcel during roughly 48 hours) with a grid spacing of 62.5 x 62.5 km on a polar stereographic projection. A new such product file is available every day around 06 UTC from the OSISAF servers.
Figure above: example sea-ice motion from 12th to 14th August 2010 in Ross and Lazarev seas. The colours scale as vorticity and reveal the position of a deep low-pressure travelling over sea-ice during these two days.
This new ice motion product is directly targeted for operational monitoring of sea-ice motion through visual analysis e.g. at the National Ice Services, as well as through data assimilation into coupled ocean-ice forecasting models. The archive of products can also be relevant for validating these models and estimating their parameters, as well as for case studies.
The rather crude resolution of the product grid, directly imposed by the rather low-resolution of the satellite images, is compensated by the total geographical coverage. By comparison, fine-resolution ice motion products, e.g. based on Synthetic Apperture Radar (SAR) images cannot cover the whole Antarctic sea-ice on a daily basis.
A potential usage of the OSISAF ice drift product, particularly relevant for Antarctic operations, could be tracking of low-pressure over sea-ice (Figure 2).
As of today, the main limitation of the OSISAF ice drift product is the yearly ‘summer observation pause’, from November 1to April 1, during which it is challenging to process quality-checked ice motion vectors. To maintain our focus on being an operational service, ice motion vectors will not be distributed until further progress is made on this front. The next operational ice motion product will be available on April 1 2014 from the OSISAF website.
Validating the motion vectors
Validating the motion has been central to our research efforts. The main reason why it took three years for starting the Southern Hemisphere monitoring is the lack of recent, long-lasting, high-quality buoy trajectories for validation purposes. It is true that the summer pause in our processing did not help for finding validation matches, however the lack of buoy trajectories in the Southern Hemisphere has delayed significantly the delivery of the operational product.
To compensate, we reverted to comparing our product with that delivered in near real-time by the Danish Technical University (DTU, href=”http://www.seaice.dk”), which is based on SAR images and validates excellently against buoys in the Arctic. We found a good agreement between the two products (about 5 km RMSE on 48-hour drift vector components).
Interested users are invited to browse an archive of products for year 2013 at the OSISAF website http://osisaf.met.no. The same website holds a Product User Manual and a validation report.