Lead and Floe Detection From CryoSat-2 Radar and ICESat-2 Laser Altimetry

Differences in satellite sampling affect their ability to resolve small-scale features over Arctic sea ice. For CryoSat-2 (CS2) and ICESat-2 (IS2) these differences are driven by geometric (footprint resolution) and radiometric (radar or laser) sampling. Here we compare growth season (October-April) surface type densities (the detected densities of lead, floe, and ambiguous targets) from CS2 and IS2 products, Arctic-wide over a common mission period. We develop these products using standard and fully-focused CS2 sea ice processing, IS2 ATL07 sea ice height data, and IS2 ATL10 sea ice freeboard data. Our analysis shows agreement in the spatial distributions of lead and floe densities between products, but significant variations in magnitude. Average floe densities from CS2 standard and fully-focused processing are 40% and 41% respectively, but 91% for all IS2 products. The average lead density from CS2 standard processing is 45%, and below 10% for all other products. The factors causing ambiguous classifications and misclassifications differ between satellites; while CS2 is more susceptible to off-nadir ranging to leads, IS2 retrievals are complicated by variable apparent lead brightness at nadir, and the presence of ridged ice. We also investigate the impact of sampling on sea ice floe length estimates, which average 1.8–2.9 km. Finally, we assess the performance of CS2 and IS2 surface type classification along near-coincident CRYO2ICE orbits. Based on our results we encourage CS2 and IS2 data users to consider how satellite sampling impacts true geophysical retrievals, and to utilize both missions simultaneously to benefit from their complementary strengths.