Testing the ice-water discrimination and freeboard retrieval algorithms for the ICESat-2 mission

The ICESat-2 mission will provide routine estimates of sea ice freeboard from profiles of surface heights acquired by its photon-counting lidar: the Advanced Topographic Laser Altimeter System (ATLAS). In this paper, we describe and test procedures devised to separate returns of ice from open water - a crucial step in the estimation of local sea levels for freeboard calculations. The two data sets used in these tests, one each from winter and summer, were acquired by a Multiple Altimeter Beam Experimental Lidar (MABEL) implemented to support pre-launch development of retrieval approaches. Our approach first identifies likely open water returns using surface photon and background count rates as proxy indicators of apparent surface reflectance. Since these measured rates are noisy estimates of expected reflectance, relative surface heights are used to refine the selection of the candidate sea surface samples. Results show that winter freeboard distributions are consistent with expected regional variability, and the nearly identical repeat-track freeboard distributions during summer show retrieval consistency. From coincident lidar coverage, variability of sea level samples identified in the MABEL and Airborne Topographic Mapper (ATM) lidar profiles are 1.6 cm and 2.6 cm, with the overall difference between the distributions at 0.00 ± 0.15 m. This demonstrates the viability of the algorithms. The parameters used in these procedures will serve as a baseline and as a guide for understanding algorithm reliability. It is expected that they will be adjusted post-launch to reflect the on-orbit performance of ATLAS.