A Geophysical Investigation of the Arctic Sea Ice Surface

Thesis (Ph.D.)--University of Washington, 2016-06

The oldest records of the Arctic sea ice pack illustrate a frozen, yet dynamic icescape composed of hummocks and weathered ridges draped in thick snow. In recent decades, the effects of climate change have transformed this image: the Arctic sea ice pack is younger, thinner, and more dynamic. As a result, the properties of its surface are changing and impacting its ice mass balance. This work investigates the recent geophysical changes of the Arctic sea ice surface, giving emphasis to snow, melt ponds, and sea ice surface topography through the three following papers: (1) interdecadal changes in spring snow depth, (2) seasonal evolution of melt ponds, and (3) the spatial scaling of melt pond distributions. In the first analysis, recent in situ and airborne observations were used to extend the snow climatology to the contemporary period. Through this, we were able to identify the interdecadal change in spring snow depth distributions, and found that snow has thinned by 37 ± 29% in the western Arctic and 56 ± 33% in the Beaufort and Chukchi seas. The decrease was attributed to later autumnal sea ice formation. During the peak snowfall period in autumn, snow falls into the ocean and melts due to the absence of sea ice. In the second analysis, an algorithm was developed for identifying melt ponds in high-resolution satellite images of a Lagrangian site. The site was composed mixed sea ice types, allowing for a comparison of seasonal melt pond evolution between first-year and multiyear sea ice undergoing the same forcings. Surprisingly, melt ponds formed three weeks earlier on multiyear sea ice than first-year sea ice. Nearly half of the snow on the multiyear sea ice was optically-thin, which likely contributed to early melt pond formation. The uniformity in melt pond formation, drainage, and distribution was inversely proportional to the level of sea ice deformation; melt pond uniformity increased with decreasing sea ice deformation. The third analysis investigated the spatial scaling of melt pond distributions at two sites with homogenous (undeformed first-year) and heterogeneous (mixed deformation and age) sea ice. The relationship between small-scale variability in melt pond geometries and aggregate-scale estimates of melt pond fractions was examined. The results revealed that: (1) melt pond geometry is most variable before melt pond drainage at the heterogeneous site, but after melt pond drainage at the homogenous site, (2) aggregate-scale estimates of melt pond fractions in homogenous and heterogeneous sea ice sites are larger than previously recognized, ranging from ~70 km2 to ~480 km2, and (3) aggregate-scale estimates of melt pond fractions may be dependent on the composition of sea ice types and stage of melt pond evolution.