Authors
Chris Picard (CIRES,University of Colorado Boulder Department of Geography), Waleed Abdalati (CIRES)
Abstract
Previous estimates of freeboard-derived sea ice thickness (SIT) in Fram Strait have relied on modeled or climatological snow depths. These methods have been leveraged due to a lack of direct snow depth observations, and the accuracy of these estimates is one of the key uncertainties in calculating SIT. However, the recent cooperation between NASA and the ESA to acquire near-coincident observations of ICESat-2 lidar (IS2) and CryoSat-2 radar (CS2) freeboards provides a means to directly calculate snow depth over sea ice. The efficacy of this technique, referred to Cryo2Ice, has yet to be examined on a regional scale outside of the central Arctic. In this research, I use Cryo2Ice within Fram Strait, a region of heterogeneous surface topography, to estimate snow depth and SITs over November to April, 2019-2024. I first compare the differences in Cryo2Ice estimates using IS2 with three publicly available radar products: the ESA Baseline E Level 2 product (ESA-E), the Alfred Wegener Institute Level 2 SIT product (AWI), and the Lognormal Retracker Altimetry Model (LARM) product. The optimal IS2 and CS2 pair for Cryo2Ice is determined by comparing SIT distributions with those from upward looking sonar estimates (ULS). Then, I examine the extent to which SITs derived using modeled and climatological snow depths overestimate thickness when compared to SITs derived using Cryo2Ice. Finally, a sensitivity analysis is completed to determine how SIT estimates vary based on changing snow density, ice density, and binning radii.
The LARM product consistently estimates lower freeboards, larger snow depths, and thinner SIT than the ESA-E and AWI products. I find that the optimal Cryo2Ice pairing is IS2 with CS2 LARM, as estimates yielded a modal thickness of 1.90 m, in close agreement with the ULS modal peak of 1.86 m. Thicknesses derived using modeled and climatological snow depths can lead to estimates between 10 cm and 99 cm larger than SITs derived using freeboard differencing from Cryo2Ice. The results indicate that SIT estimates are particularly sensitive to changing snow and ice densities, with binning radii having relatively insignificant impact on estimates. Thus, the findings of this research emphasize the need for judicious choice of the radar product for a given study, to exercise caution when using independent snow depth products, and to strongly consider what parameters (i.e., densities) are being assumed in calculations. This research provides a first examination of the applicability of Cryo2Ice to estimate SIT distributions within Fram Strait, a lower latitude region characterized by variable ice types. Importantly, the results of this research provide confidence in Cryo2Ice as a technique to estimate regional snow depths and SITs outside of the central Arctic.