Coherence onset and two-dimensional deformation of landfast sea ice at Cambridge Bay, Nunavut

Introduction: Landfast sea ice governs winter travel and nearshore ecology across the Canadian Arctic Archipelago, and the calendar of its formation is shifting under Arctic amplification. Interferometric SAR has been applied at Cambridge Bay to retrieve two-dimensional displacement (Chen et al., 2021) and its interannual variation (Zhu et al., 2023). In both, coherence selects measurable pixels rather than serving as a measurement. But ice loses coherence while drifting and regains it once fixed, so its seasonal course may record when the ice consolidated and whether it stayed that way.

Methods: Sentinel-1 IW SLC scenes for winters 2018–2019 and 2020–2021 were processed as consecutive 12-day pairs at 80 m spacing, with descending S1A (HH) and ascending S1B (VV). Coherent ice area was tracked at 12-day intervals from mid-December to late April, with onset defined as the coherent area sustained above 50%. SBAS displacement, referenced to fixed land, was decomposed into east–west and vertical components at four points 1.2–8.2 km offshore. Networks are consecutive, precluding phase closure. Precision from an independent land control is ±1.6 and ±0.7 cm, rising to ±4 cm in the 2020–2021 ascending series.

Results:  In 2018–2019, the ice consolidated once, on 30 January, remaining coherent to early April. In 2020–2021, it consolidated 49 days later, on 20 March, and only after an earlier consolidation failed: the coherent area reached 55% in late February, collapsed to 16% in early March, then recovered, with decay underway by mid-April. Coherence in that winter fell from 0.78–0.81 at 12 days to 0.54 at 24, constraining measurement to the 12-day repeat. Deformation was faster in the intermittent winter: at the outermost reliable point, horizontal motion rose from −0.43 to −1.18 cm/day, and vertical motion reversed sign, from -0.18 to +1.42 cm/day.

Conclusions: Coherence resolves not only when landfast ice forms but whether formation succeeds, capturing a mid-winter break-up and refreeze invisible to deformation measurements. The intermittent winter was also the more mobile, though two seasons cannot separate this from coincidence. Snow metamorphism and early-growth dielectric change also cost coherence over immobile ice, so dates are late-biased; longer records and independent draft data are needed to bound this.