High-resolution sea ice drift and deformation from sequential SAR images in the Transpolar Drift during MOSAiC 2019/2020

Sea ice deformation is a crucial process in the polar climate system and, thus, it is an important cross-cutting theme for all disciplines of the interdisciplinary research expedition MOSAiC. Because sea ice deformation is highly localized and intermittent, drift and deformation with a high spatial...

Full description

Bibliographic Details
Main Authors: von Albedyll, Luisa, Hutter, Nils
Format: Dataset
Language:English
Published: PANGAEA 2023
Subjects:
Online Access:https://doi.pangaea.de/10.1594/PANGAEA.958449
https://doi.org/10.1594/PANGAEA.958449
Description
Summary:Sea ice deformation is a crucial process in the polar climate system and, thus, it is an important cross-cutting theme for all disciplines of the interdisciplinary research expedition MOSAiC. Because sea ice deformation is highly localized and intermittent, drift and deformation with a high spatial and temporal resolution and a large spatial coverage are required for a comprehensive description of the sea ice dynamics. We provide a regularly gridded, high-resolution drift and deformation dataset that can be used for several potential applications. Drift fields were obtained from Sentinel-1, HH polarization SAR images acquired in enhanced wide mode. These had a pixel resolution of 50 m in Polar Stereographic North projection (latitude of true scale: 70 N, center longitude: 45 W). We used an ice-tracking algorithm introduced by Thomas et al. (2008, 2011) and modified by Hollands and Dierking (2011) to derive drift from sequential pairs. Typically, the time between two scenes was one day, with a few exceptions of 2-3 days, and the size of the scenes was on average 200 x 200 km. Images are available for the entire study period, except for the time between 14 January and 15 March 2020, when the ship was north of the latitudinal coverage of the satellite. The resulting drift data set was defined on a regular grid with a spatial resolution of 700 m. Next, we calculate the spatial derivatives from the regularly spaced drift field following von Albedyll et al. (2021). Divergence, convergence, shear, and total deformation are then derived from the spatial derivatives of the velocity field. To reduce noise in the divergence fields, we filter the drift data with a directional filter that detects the direction with the smallest variation at each pixel and smooths along, but not across this orientation, with a 1-d kernel. The direction is chosen to minimize the standard deviation in a neighborhood of 7 pixels. This way, noise is reduced while preserving the strong gradients in the velocity field that are indicative of ...