Using ICESat-2 Satellite Altimeter Data to Improve Understanding of Thermodynamic and Dynamic Sea Ice Characteristics in the Ross Sea

The fine spatial resolution (~11 m of footprint) and precision (~2 cm of vertical precision) of NASA's ICESat-2 (IS2) satellite altimeter data allow us to detect sea ice deformation features, such as pressure ridges and leads. Moreover, the thermodynamic and dynamic sea ice processes can be investigated by monitoring sea ice deformation features for a large area over a regular period. However, the IS2 sea ice products (ATL07/ATL10) have critical problems in further accurate monitoring of sea ice: (1) spatial sparsity, (2) dark lead misclassification, and (3) incapability of thin/gray ice detection. In this study, I address these issues and improve the monitoring of sea ice freeboard in the Ross Sea using IS2 data. In Chapter 2, a geostatistical approach, kriging, is employed to address the spatial sparsity issue and interpolate the freeboard values where the IS2 tracks are not directly available. The freeboard maps retrieved from the kriging method exhibit the weekly distribution of sea ice freeboard and persistent polynya areas in the Ross Sea. In Chapter 3, instead of the original ATL07 decision tree algorithm, a data-driven neural network (NN) model is used to classify sea ice surface into open water, thin/gray ice, and snow-covered/thick ice. The NN approach shows a better performance in delineating the sea ice surface than the original ATL07 product, especially the separation of thin ice covers. In Chapter 4, the ability of IS2 to capture the thermodynamic and dynamic ice growth is assessed using in-situ sea ice mass balance array (SIMBA) buoys deployed in the central Arctic. The modal freeboard of the IS2 measurement agrees with the thermodynamic ice growth, but the mean and median IS2 freeboard involve the redistribution of sea ice thickness induced by sea ice deformation. Finally, in Chapter 5, taking advantage of IS2 data for monitoring dynamic sea ice processes, the spatiotemporal variations of pressure ridges in the Ross Sea are investigated. The southeast sector of the Ross Sea is highly deformed by a strong easterly wind and thick ice inflow in 2019-2020, but the transition in the atmospheric condition in 2021-2022 affects the distribution of deformation features. This result emphasizes the importance of atmospheric forcings on the thermodynamic and dynamic sea ice conditions in the Ross Sea.