Agreement and Complementarity of Sea Ice Drift Products

Abstract

Changes in Arctic sea ice has raised questions about changes in sea ice drift patterns. Reduced sea ice coverage may open up the Arctic to further exploration of maritime activities, particularly during the summer months. Given such changes it is important to investigate differences between available sea ice drift products. Products based on synthetic aperture radar (SAR), radar scatterometer and radiometer are compared for both motion speed and direction within this study. Two Cband SAR and one L-band SAR product are used in the comparison. Differences in temporal and spatial resolutions of the drift estimates spanning from July 2010 until June 2011 are investigated. High temporal and spatial resolution was proven useful to fully capture the sea ice drift in the Fram Strait. For summer coverage, SAR data is a prerequisite and L-band is desirable. The two C-band SAR products have a mean speed correlation of 0.90 and exhibit high conformity, despite being generated by separate processing methods. The L-band SAR product and the scatterometer and radiometer products are to a lower degree in agreement with each other and the C-band SAR products, which may be attributed to the products' dependency on the temporal baseline. Depending on the choice of sensor or combination of sensors, the resulting twelve-month mean drift varies between 0.09 and 0.12 m/s excluding L-band SAR. The latter shows a particularly low drift of 0.05 m/s, which we attribute to an overrepresentation of slow ice.

Opportunistic evaluation of modelled sea ice drift using passively drifting telemetry collars in Hudson Bay, Canada

“…NSIDC drift speeds tended to overestimate slow collar drift (< 4.5 km h -1 ) and underestimate high collar drift (> 4.5 km h -1 ). This pattern is likely an effect of estimating a zero-bound variable, and is consistent with other satellite-based sea ice drift products (Johansson and Berg, 2016;Mahoney et al, 2019;Rozman et al, 2011;Sumata et al, 2014). As drift speeds approach 0 km d -1 , the probability of overestimation approaches 1, and as drift speeds increase, the range of values that below the drift speed (i.e., underestimates) increases.…”

“…Our estimates of drift speed bias are greater than estimated in studies of NSIDC and other drift products (Hwang, 2013;Johansson and Berg, 2016;Lavergne, 2016;Schwegmann et al, 2011;Sumata et al, 2014; but see Durner et al, 2017).…”

A Combination of Feature Tracking and Pattern Matching with Optimal Parametrization for Sea Ice Drift Retrieval from SAR Data

“…Peakedness/flatness of the cross-correlation function, estimated as the second order spatial derivative at the position of the identified maximum, may serve as a good candidate for that purpose [39]. Other improvements required in the future may include the following: rogue vector detection based on deformation [41]; proper filtering of small-scale drift anomalies that can be seen as spurious noise when computing the first order spatial derivative of the sea ice velocity field, i.e., sea ice deformation; different approximation of the first guess based on FT results that takes into account small-scale variations; improvement of the search for the highest MCC in 3D space (X,Y,rotation) using multivariate optimization technique.…”

Application of the Combined Feature Tracking and Maximum Cross-Correlation Algorithm to the Extraction of Sea Ice Motion Data From GF-3 Imagery

“…These images had a resolution of 10 m, a swath width of 100 km, and selective dualpolarization. Considering that the time interval between image pairs used for SIM extraction and the speed of motion of sea ice in the covered area affect the results of SIM extraction [40][41], GF-3 image pairs were selected that were acquired for different time intervals, regions, and seasons. The location and acquisition time of the selected GF-3 images are shown in Table 1 and Fig.…”