[1] Satellite passive microwave observations document an overall downward trend in Arctic sea ice extent and area since 1978. While the record minimum observed in September 2002 strongly reinforced this downward trend, extreme ice minima were again observed in 2003 and 2004. Although having three extreme minimum years in a row is unprecedented in the satellite record, attributing these recent trends and extremes to greenhouse gas loading must be tempered by recognition that the sea ice cover is variable from year to year in response to wind, temperature and oceanic forcings.

Abstract. In an introductory section we review the physical processes influencing the formation and evolution of melt ponds on sea ice during the Arctic summer. As melt progresses, the changing properties of the surface interact strongly with the surface heat balance. The small interannual variability of the seasonal ice extent suggests an interannual variability of the surface heat balance of _+ 1 W m -2 or less. The interannual variance of atmospheric forcing represented by the transport of moist static energy into the Arctic is an order of magnitude greater. This appears to contradict the notion of a highly sensitive sea ice cover and emphasizes the need to generate albedo as an important internal variable in interactive models. Observations of melt ponds are needed in order to derive improved relationships between surface albedo and parameters such as the amount of snow, the onset and termination of melting, the ice thickness distribution, and ice deformation. Here classified (National Technical Means) imagery is used to measure melt pond coverage as it evolves over a summer on ice surrounding a drifting buoy. Local variability of pond cover is greatest at the beginning of the melt season, that is, pond coverage from 5% to 50% depending on ice type, as previously found by Russian investigators. An important distinction is found in the temporal change of pond cover: it decreases with time on thick ice, and it increases with time on thin ice (eventually leading to the disappearance of thin ice at the end of summer). An attempt to relate pond coverage to ice concentrations derived from passive microwave data proved unsuccessful. IntroductionIn the central Arctic, summer starts in late May when increasing solar radiation rapidly warms the snow that covers the ice. Melting, which starts in early June, lowers the albedo and further enhances the absorption of short-wave radiation. By mid-June a significant fraction of the pack ice is covered by melt water ponds. As melting progresses, the ponds deepen, their area diminishes, and much of the thin ice melts completely so that more of the dark ocean surface is exposed. Freeze-up occurs in late August or early September.The progression of melt controls the albedo of the ice surface. Thickness and mass balance of sea ice are extremely sensitive to the summertime short-wave radiation balance, making surface albedo a powerful tuning parameter in sea ice models. While the processes affecting the surface albedo are qualitatively known, the available data on regional and temporal differences are sparse and unsatisfactory for validating efforts toward modeling the albedo as an internal variable derived from surface processes such as available snow, melt rate, ice topography, thickness distribution, and ice deformation. At the same time, there is an acute need for an improved representation of sea ice processes in models.The computational requirements of fully interactive models of the global climate system for the prediction of, for instance, the effects of increased greenhou...

Arctic sea ice data from a variety of historical sources have been synthesized into a database extending back to 1850 with monthly time‐resolution. The synthesis procedure includes interpolation to a uniform grid and an analog‐based estimation of ice concentrations in areas of no data. The consolidated database shows that there is no precedent as far back as 1850 for the 21st century's minimum ice extent of sea ice on the pan‐Arctic scale. A regional‐scale exception to this statement is the Bering Sea. The rate of retreat since the 1990s is also unprecedented and especially large in the Beaufort and Chukchi Seas. Decadal and multidecadal variations have occurred in some regions, but their magnitudes are smaller than that of the recent ice loss. Interannual variability is prominent in all regions and will pose a challenge to sea ice prediction efforts.

Arctic sea ice extent and area in September 2002 reached their lowest levels recorded since 1978. These conditions likely resulted from (1) anomalous warm southerly winds in spring, advecting ice poleward from the Siberian coast (2) persistent low pressure and high temperatures over the Arctic Ocean in summer, promoting ice divergence and rapid melt.

In September 2002, Arctic sea ice extent reached a minimum unprecedented in 24 years of satellite passive microwave observations, and almost certainly unmatched in 50 years of charting Arctic ice [Serreze et al., 2003]. Again, in September 2003, ice retreated to an unusually low extent, almost equaling the previous year's minimum (Figure l). The Sea Ice Index (http://nsidc.org/data/seaice_index/), an easy‐to‐use source of information on sea ice trends and anomalies, assists in observing these minima. The Sea Ice Index is intended for both researchers and the scientifically inclined general public.