Retracker-dependent radar freeboard correction methods for satellite radar altimetry-based sea ice thickness estimation

Sea ice thickness is derived from its freeboard measured by satellite radar altimeters. However, the radar freeboard, which is initially estimated freeboard by interpreting the observed waveform, needs correction before sea ice thickness estimation so that it coincides with the height of the snow-ice interface from the sea surface. The so-called radar freeboard correction is thus an essential procedure for sea ice thickness estimation from satellite radar altimeter data, such as those from the CryoSat-2 mission. Today, most studies do the correction taking into account a slower wave propagation speed in the snow layer on sea ice under the assumption that the main scattering horizon is the snow-ice interface. However, while several recent studies have raised questions on that assumption, there is also a possibility that a retracker, which is an algorithm that estimates radar freeboard from waveform, has systematic bias. Accordingly, this study revisits the conventional way of doing the radar freeboard correction. First, we directly compare the CryoSat-2-derived radar freeboards from different retrackers with reference airborne freeboard measurements to introduce alternative correction methods for each retracker. Then, those correction methods are combined with a recently developed methodology where snow depth, sea ice thickness, freeboard, and ice draft are retrieved simultaneously. In order to compare the performance of different correction methods, including the conventional light speed correction, retrievals are done using the updated methodology, and those results are assessed using various reference datasets. Those are snow depth and freeboard from airborne observation, ice draft from mooring observation, and freeboard from satellite laser altimeter observation. In addition, the correction methods are combined with another independent retrieval method that estimates snow depth and sea ice thickness by combining satellite laser and radar altimeter measurements. Lastly, the consistency between the results from the two retrieval methods is examined for each radar freeboard correction method.