Title New Bayesian algorithm for sea ice detection with QuikSCAT
Author Rivas, M.B.; Stoffelen, A.
Author Affil Rivas, M.B., Royal Netherlands Meteorological Institute, De Bilt, Netherlands
Source IEEE Transactions on Geoscience and Remote Sensing, 49(6 Part 1), p.1894-1901. Publisher: Institute of Electrical and Electronics Engineers Geoscience and Remote Sensing Society, New York, NY, United States. ISSN: 0196-2892
Publication Date Jun. 2011
Notes In English. 24 refs. GeoRef Acc. No: 309674. CRREL Acc. No: 65006423
Index Terms backscattering; detection; ice; microwaves; radar; remote sensing; statistical analysis; Antarctica; Arctic Ocean; algorithms; Bayesian analysis; microwave methods; probability; Quick Scatterometer; QuikSCAT; radar methods; satellite methods; sea ice; winds
Abstract The authors propose a new sea ice detection method for a rotating Ku-band scatterometer with dual-polarization capability, such as SeaWinds on the Quick Scatterometer (QuikSCAT), based on probabilistic distances to ocean wind and sea ice geophysical model functions (GMFs) and evaluate its performance against other active and passive microwave algorithms. All the methods yield similar results during the sea ice growth season but show substantial differences during the spring and summer months. A detailed comparison based on high- resolution synthetic aperture radar and optical imagery shows that major discrepancies relate to newly formed, low- concentration, and water-saturated sea ice species. The new GMF-based algorithm for sea ice detection with QuikSCAT improves on the misclassification scores that affect other algorithms and provides daily sea ice masks at a 25-km resolution for use in ground processors that require the effective removal of sea ice contaminated pixels all year round.
URL http://hdl.handle.net/10.1109/TGRS.2010.2101608
Publication Type journal article
Record ID 91384