A comparison between Envisat and ICESat sea ice thickness in the Southern Ocean

Abstract

Abstract. The crucial role that Antarctic sea ice plays in the global climate system is strongly linked to its thickness. While field observations are too sparse in the Southern Ocean to determine long-term trends of the Antarctic sea ice thickness (SIT) on a hemispheric scale, satellite radar altimetry data can be applied with a promising prospect. The European Space Agency's Sea Ice Climate Change Initiative project (ESA SICCI) generates sea ice thickness derived from Envisat, covering the entire Southern Ocean year-round from 2002 to 2012. In this study, the SICCI Envisat Antarctic SIT is first compared with an Ice, Cloud, and land Elevation Satellite (ICESat) SIT product retrieved with a modified ice density algorithm. Both data sets are compared to SIT estimates from upward-looking sonar (ULS) in the Weddell Sea, showing mean differences (MDs) and standard deviations (SDs, in parentheses) of 1.29 (0.65) m for Envisat − ULS (− denotes “minus” and the same below), while we find 1.11 (0.81) m for ICESat − ULS. The inter-comparisons are conducted for all seasons except for winter, based on the ICESat operating periods. According to the results, the differences between Envisat and ICESat SIT reveal significant temporal and spatial variations. More specifically, the smallest seasonal SIT MD (SD) of 0.00 m (0.39 m) for Envisat − ICESat is found in spring (October–November), while a larger MD (SD) of 0.52 (0.68 m) and 0.57 m (0.45 m) exists in summer (February–March) and autumn (May–June). It is also shown that from autumn to spring, mean Envisat SIT decreases while mean ICESat SIT increases. Our findings suggest that both overestimation of Envisat sea ice freeboard potentially caused by radar backscatter originating from inside the snow layer and the Advanced Microwave Scanning Radiometer for EOS (AMSR-E, where EOS stands for Earth Observing System) snow depth biases and sea ice density uncertainties can possibly account for the differences between Envisat and ICESat SIT.

Evaluation of Antarctic sea ice thickness and volume during 2003–2014 in CMIP6 using Envisat and CryoSat-2 observations

“…Consequently, comparing the trend of development in SIV based on satellite observations with that derived from the CMIP6 models might also have inherent uncertainty. Second, the SICCI Antarctic SIT dataset is classified as experimental data by the data creator because the satellite-derived observational data have notable uncertainty (Wang et al 2022). For example, owing to limited understanding of the snow layer on Antarctic sea ice, substantial errors might be added in subsequent retrievals of SIT (Maksym and Markus 2008).…”

“…However, 21 CMIP6 models exhibit a negative tendency that contrasts with the CS2 measurements. The anomaly sequence of Antarctic SIV shows a strong rising tendency over the CS2 period (2011.01-2014.12), as evidenced by Wang et al (2022) and Liao et al (2022). Specifically, CS2 data show a trend of 45.13 km 3 /mon in Antarctic SIV, whereas the trend computed from the CMIP6 MMM is only 1.82 km 3 /mon, suggesting that the CMIP6 models do not sufficiently capture this strong increase.…”

Sensitivity of NEMO4.0-SI3 model parameters on sea ice budgets in the Southern Ocean

“…In addition to several parameters that have an impact on the simulation of the SIC budget, the SIV budget also shows a high sensitivity to snow density, which is also one of the parameters that leads to a high uncertainty in the satellitederived sea ice thickness (e.g. Liao et al, 2022;Wang et al, 2022b). However, considering the simple approach to snow in the current NEMO4.0-SI 3 model (e.g.…”

The Impacts of Optimizing Model‐Dependent Parameters on the Antarctic Sea Ice Data Assimilation

“…The understanding of Antarctic sea ice variability holds significant scientific and socioeconomic importance, owing to the crucial role that Antarctic sea ice plays in the Earth system (Turner & Comiso, 2017). Nonetheless, the present sparsity of sea ice observations poses a challenge in achieving a comprehensive understanding of the Antarctic sea ice system (e.g., J. Wang, Min, et al., 2022; Worby et al., 2008), and numerical models currently exhibit notable limitations in adequately capturing the variations in Antarctic sea ice (e.g., Shu et al., 2020; Tsujino et al., 2020). Consequently, data assimilation has emerged as a valuable approach, as it synergistically combines information from both observations and simulations.…”