The role of melt ponds in the evolution of Arctic summer pack ice albedos (INVITED)

The response of the Arctic sea-ice cover to climate change is deemed to contribute substantially to the disproportionate warming predicted by general circulation models (GCMs) for the Arctic as a result of enhanced greenhouse gas forcing. Satellite observations indicate reductions in Arctic summer ice extent during the past decade. Moreover, conceptual and numerical models suggest that ice-albedo feedback processes can amplify the ice cover's response to changes in radiative forcing. Hence, the state of the summer ice pack requires closer scrutiny in order to unravel the causes of diminishing ice extent and improve model simulations of sea-ice variability. In this context, seasonal reductions in large-scale ice albedo from >0.8 to <0.55 are of particular importance, since GCM simulations are highly sensitive to changes in ice albedo. Measurements during the Surface Heat Budget of the Arctic Ocean (SHEBA) Program in the summer of 1998 and near Barrow, Alaska in 2000 indicate that the seasonal albedo reduction is mostly due to formation of melt ponds at the ice surface. In contrast, bare ice and open water exhibit remarkably small variability in albedo during the course of summer. Measurements of meltwater transport through the ice cover indicate that the seasonal evolution of melt-pond coverage and depth is controlled to a large extent by the permeability structure of the underlying ice. During early melt, low ice permeabilities result in widespread flooding and pooling of water at the ice surface and within the snowpack. Subsequent increases in ice permeability foster drainage of surface meltwater. As water circulates through ponds and their porous margins during the latter part of the meltseason, the meltpools deepen and widen. In late August, pond freezing and new-snow accumulation terminate surface ablation. Our field measurements indicate that these changes in ice hydrology are closely linked to the seasonal evolution of ice albedo. In a simple attempt to assess how well GCMs perform at simulating this seasonal cycle, we compared measurements of ice albedo and meltwater distribution with albedos derived from two GCM parameterization schemes. While the total amount of shortwave radiation supplied to the ice between May and August corresponds to within 10% between measurement and simulation, daily deviations are substantial. These discrepancies indicate a gap in our understanding of the impact of small-scale ice processes on parameters such as ice albedo at the scale of GCM grid cells. Despite adequate performance in reproducing the total amount of solar radiation absorbed by ice during a typical summer, albedo parameterization schemes may not be capable of reproducing more complex, non-linear changes triggered by the impact of climate variability and change on ice processes at macroscopic or microscopic scales.