Classification of Arctic Sea Ice Type in CFOSAT Scatterometer Measurements Using a Random Forest Classifier

The Ku-band scatterometer called CSCAT onboard the Chinese–French Oceanography Satellite (CFOSAT) is first spaceborne rotating fan-beam (RFSCAT). A new algorithm for classification of Arctic sea ice types on measurement data using a random forest classifier presented. trained National Snow and Ice Data Center (NSIDC) weekly age concentration product. Five feature parameters, including mean value horizontal vertical polarization backscatter coefficient, standard deviation coefficient copol ratio, are innovatively extracted from orbital time to distinguish water, first-year (FYI) multi-year (MYI). overall accuracy kappa type model 93.35% 88.53%, respectively, precisions FYI, MYI 99.67%, 86.60%, 79.74%, respectively. Multi-source datasets, daily EUMETSAT Ocean Sea Application Facility (OSI SAF), NSIDC age, (MYIC) provided by University Bremen, SAR-based released Copernicus Marine Environment Monitoring Service (CMEMS) have been used comparison validation. It shown that most obvious difference in distribution between results OSI SAF mainly concentrated marginal zones FYI MYI. Furthermore, compared with type, area derived more homogeneous less noise, especially case younger multiyear ice. In East Greenland region, identifies pixels as lower MYIC values, showing better identification areas mobility conclusion, this research verifies capability monitoring classification, spatial homogeneity detectable duration classification. Given high processing speed, forest-based can offer good guidance FY-3E/RFSCAT, i.e., dual-frequency (Ku C band) WindRAD.