Detection and identification of mechanisms and trends of sea ice drift variability in the Southern Ocean

De Jager, Wayne

Detection and identification of mechanisms and trends of sea ice drift variability in the Southern Ocean

Thesis / Dissertation

2025

Publisher

University of Cape Town

Department

Department of Oceanography

Faculty

Faculty of Science

Abstract

Antarctic sea ice regulates global climate processes by mediating exchanges of heat, momentum, and mass between the Southern Ocean and atmosphere. However, the logistical constraints of acquiring in situ sea-ice observations is challenging. Inferring sea-ice properties, such as concentration and motion, from satellite observations helps to address these data gaps, but the retrieval algorithms carry some limitations that complicate data acquisition. Temporal resolution limits are particularly pronounced when studying the effect of transient weather phenomena, such as polar and extra-tropical cyclones, which are major drivers of short-term sea-ice variability. This thesis presents a systematic investigation into the weather-induced rotation dynamics in Antarctic sea ice, with a focus on the feasibility of quantifying hourly-to-daily sea-ice concentration and drift variability. A new algorithmic approach that is sensitive to the detection of rotational sea-ice drift features from satellite data at spatial and temporal scales comparable to that of polar weather is proposed, yielding a novel index for assessing interannual dynamical trends. Comparing selected 48-hour satellite ice motion products in the Atlantic Sector between 2013–2020 indicates good agreement in detecting anticyclonic rotation but discrepancies for cyclonic rotation. Cyclonic rotation is shown to be more intense than anticyclonic rotation, with the mean intensity of the 95th percentile of cyclonic features shown to be 1.5–2.0 times larger than their anticyclonic counterparts. Cyclonic rotation is also associated with greater variability and observational uncertainties, suggesting that the cyclones engender more heterogenous motion compared to high-pressure systems but also reduce the capacity of satellite products to accurately detect this motion. The analysis was expanded from 1991–2020 over the entire Southern Ocean using the latest generation 24-hour satellite sea-ice drift product. Results demonstrate that sea-ice rotation has become more correlated with atmospheric rotation at daily timescales, suggesting an increasing responsiveness of sea ice to weather forcing over the past three decades. Notably, the increased correlation between ice and atmospheric rotational dynamics is evidenced despite no discernible trends in the atmospheric forcing, pointing to a plausible role of the ocean in modifying the rheology of pack-ice. While the Southern Annular Mode was considered, it does not directly explain the strengthened coupling. Uniquely, sea ice in the Weddell Sea experienced a sharp decline in the intensity of its cyclonic and anticyclonic components from 2002 onwards, which may instantiate that large synoptic systems are not the primary driver of ice rotation in this region. To further address the temporal resolution limitations associated with daily-averaged satellite products, this thesis also explored a novel method for deriving sub-daily sea-ice type and motion observation from passive microwave swath data. The ECICE algorithm was directly applied to AMSR-2 Level-1B footprint data to produce sub-daily young (YI), first year (FYI) and multiyear (MYI) sea-ice concentration (SIC) observations. Results indicate ice-type classification was inconsistent, primarily because the radiometric signature of YI overlaps considerably with that of FYI and MYI. The use of motion vectors derived from the same swath data revealed a significant sea-ice response to weather forcing in just ~9 hours, which cannot be resolved by traditional daily-averaged products. The role of ice advection, melt, and growth contributions to SIC variability was calculated, but two major challenges arose: (i) coverage limitations associated with the drift retrieval algorithm caused unrealistic melt and growth estimates in the marginal ice zone (MIZ); and (ii) SIC saturation in the mobilized ice interior yielded minimal SIC changes despite a significant and widespread drift response. Ultimately, an assessment of drift products against dedicated observational experiments is recommended to evaluate product performance, especially in the MIZ, before swath-based retrieval techniques can be improved and systematically implemented.

Keywords

Sea Ice

Southern Ocean

Reference:

Collections

PhD / Doctoral

Full item page