Abstract
This study presents Arctic sea ice drift fields measured by shipboard marine X‐band radar (MR). The measurements are based on the maximum cross correlation between two sequential MR backscatter images separated ∼1 min in time, a method that is commonly used to estimate sea ice drift from satellite products. The advantage of MR is that images in close temporal proximity are readily available. A typical MR antenna rotation period is ∼1–2 s, whereas satellite revisit times can be on the order of days. The technique is applied to ∼4 weeks of measurements taken from R/V Sikuliaq in the Beaufort Sea in the fall of 2015. The resulting sea ice velocity fields have ∼500 m and up to ∼5 min resolution, covering a maximum range of ∼4 km. The MR velocity fields are validated using the GPS‐tracked motion of Surface Wave Instrument Float with Tracking (SWIFT) drifters, wave buoys, and R/V Sikuliaq during ice stations. The comparison between MR and reference sea ice drift measurements yields root‐mean‐square errors from 0.8 to 5.6 cm s−1. The MR sea ice velocity fields near the ice edge reveal strong horizontal gradients and peak speeds > 1 m s−1. The observed submesoscale sea ice drift processes include an eddy with ∼6 km diameter and vorticities <–2 (normalized by the Coriolis frequency) as well as converging and diverging flow with normalized divergences <–2 and >1, respectively. The sea ice drift speed correlates only weakly with the wind speed (r2 = 0.34), which presents a challenge to conventional wisdom.
Key Points
Marine X‐band radar data are processed to yield sub‐km scale sea ice velocity fields
Root‐mean‐square errors between radar and reference sea ice drift measurements are ∼1–6 cm s^–1
Contrary to conventional wisdom, sea ice drift and wind speed are only weakly correlated (r^2 = 0.34) near the ice edge
