Key points are not available for this paper at this time.
In the Arctic, sea ice leads are crucial for the exchange of energy between the ocean and the atmosphere. Leads exhibit a wide range of widths, from several meters to tens of kilometers, with narrow leads being most abundant. Therefore, high-resolution observations are essential for monitoring leads distributions and variations. Here, we used an approach based on deep learning to generate an Arctic lead record with a spatial resolution of 40 m for wintertime 2016‒2023 based on Sentinel-1Synthetic Aperture Radar (SAR) images. The results indicate that our new lead record captures the spatial distribution pattern and the temporal variations in the lead fraction well, as one would expect from atmospheric and oceanic forcings and in agreement with two alternative pan-Arctic lead datasets. Meanwhile, the monthly Arctic lead fraction variations are consistent to the variations in sea ice drift. With our product the Arctic lead width distribution can be analyzed with a comparable high spatial resolution. Our results demonstrate that Arctic lead width distributions follow a power law function with leads narrower than 1 km accounting for approximately half of the total during winter. This study enhances the understanding of the spatiotemporal variations in Arctic sea ice leads and their width distributions, supporting climate change research.
Chen et al. (2025) studied this question.