Classification of Sea Ice Types in the Arctic by Radar Echoes from SARAL/AltiKa

An important step in the sea ice freeboard to thickness conversion is classification of types, since type affects snow depth and density. Studies using Ku-band CryoSat-2 have shown promise distinguishing FYI MYI based on parametrisation radar echo. Here, we investigate applying same algorithms that success for measurements acquired by SARAL/AltiKa at Ka-band. Four different classifiers are investigated, i.e., threshold-based, Bayesian, Random Forest (RF) k-nearest neighbour (KNN), data from five 35 day cycles during Arctic mid-winter 2014–2018. The overall performance shows highest accuracy 93% (Bayesian classifier) 39% (threshold-based classifier). For all algorithms, more than half cover falsely classifies as FYI, showing difference surface characteristics attainable Ka-band compared due scattering mechanisms. However, high (above 90%) estimated three supervised (KNN, RF Bayesian). Furthermore, leading-edge width parameter potential discriminating open water (ocean) when visually with reference data. Our results encourage use waveform parameters further validation ice/open edges discrimination types combining Ka- Ku-band, especially planned launch dual-frequency altimeter mission Copernicus Polar Ice Snow Topography Altimeter (CRISTAL) 2027.