Satellite-based analysis of ocean-surface stress across the ice-free and ice-covered polar oceans

Abstract

Ocean-surface stress is a critical driver of polar sea-ice dynamics, air-sea interactions, and ocean circulation. This work provides a daily analysis of ocean-surface stress on 25 km Equal-Area Scalable Earth (EASE) grids across the ice-free and ice-covered regions of the polar oceans (2011-2021 for the Arctic, 2013-2021 for the Antarctic), covering latitudes north of 60° N in the Arctic and south of 50° S in the Antarctic and Southern Ocean. Ocean-surface stress is calculated using a bulk parameterization approach that combines ocean-surface winds, ice motion vectors, and sea surface height (SSH) data from multiple satellite platforms. The analysis captures significant spatial and temporal variability in ocean-surface wind stress and the resultant wind-driven Ekman transport, while providing enhanced spatiotemporal resolution. Two sensitivity analyses are conducted to address key sources of uncertainty. The first addresses the fine-scale variability in SSH fields, which was mitigated using a 150 km Gaussian filter to smooth 3 d SSH datasets and enhance compatibility with the other monthly product, followed by linear interpolation to achieve daily resolution. The second investigates uncertainty in the ice-water drag coefficient, which revealed that variations in the coefficient have a proportional influence on the computed ocean-surface stress under the tested conditions. These uncertainties are most pronounced during winter, with median values reaching 20 % in the Arctic and 40 % in the Southern Ocean. Validation efforts using ice-tethered profiler velocity records revealed weak to moderate correlations with satellite-derived stress (rCombining double low line 0.4-0.8) between observed surface velocities and satellite-derived estimates (Ekman + geostrophic) at daily resolution, with significantly improved agreement when averaged to weekly means. This dataset is publicly available at 10.5281/zenodo.15534576 (Liu and Yu, 2024).