Error analysis of real-time remotely sensed microwave sea-ice motions in the Western Arctic Ocean

An algorithm used to composite SSM/I 85.5 GHz imagery and derive sea ice motion was adapted for operational testing at Fleet Numerical Meteorology and Oceanography Command (FNMOC). A feature tracking technique was applied to a 6-month period, with data provided by FNMOC and the Naval Research Labora...

Full description

Bibliographic Details
Main Author: Carsten, David M.
Other Authors: Durkee, Philip A.
Language:en_US
Published: Monterey, California. Naval Postgraduate School 2012
Online Access:http://handle.dtic.mil/100.2/ADA386446
http://hdl.handle.net/10945/9399
Description
Summary:An algorithm used to composite SSM/I 85.5 GHz imagery and derive sea ice motion was adapted for operational testing at Fleet Numerical Meteorology and Oceanography Command (FNMOC). A feature tracking technique was applied to a 6-month period, with data provided by FNMOC and the Naval Research Laboratory (NRL). Ice motions are detectable using the SSM/I motion algorithm, and fields of SSM/I motion vectors are qualitatively consistent with coincident fields of in situ buoy motion vectors. Accuracy of the SSM/I motion vectors relative to buoy motion vectors increase significantly with buoy speed. No correlation between SSM/I and buoy motion vectors is observed for speeds below 3 cm/s and correlation increases significantly above 5 cm/s. The results are very sensitive to compositing techniques used to combine SSM/I passes into a single sea ice representation. FNMOC data was composited using a "drop-in-the-bucket" technique while NRL data was composited by a bilinear interpolation technique. Significantly poorer results were found with ENMOC composited data.