Opening and closing of sea ice leads: Digital measurements from synthetic aperture radar

The coming opportunity to extract quantitative sea ice data from routine synthetic aperture radar imagery requires the development of automated image processing techniques. An algorithm is described for measuring the opening and the closing of leads by comparing two sequential synthetic aperture radar (SAR) digital images. The pair of images is classified into leads and ice, correcting for variation in brightness with range. Displacements are determined automatically on a regular 2-km grid (on a 100-km square scene) by cross-correlation. An algorithm groups together cells which cover a lead “event.” The lead area change is estimated for each cell group by counting lead pixels before and after deformation. All positive changes are summed to obtain the amount of opening, and all negative changes give closing. Errors are estimated by comparison with area changes determined from manually digitized lead boundaries and are about 0.2% of the area of the scene. For comparison, the measured opening and closing are in the range of 1 to 3% of the scene area. In ice models the openings and closings of leads are parameterized in terms of the large-scale deformation. SAR allows us to study these parametric relations by giving us access both to direct observations of opening and closing and to detailed kinematic data from which to estimate the mean deformation. In this image pair for which the deformation is nearly pure shear, we find relatively small amounts of opening and closing, consistent with previous observations, and with parameterizations in most models. The large values of opening and closing predicted by the Poisson-Gauss kinematic model are not in evidence.