The latest maximum in the Arctic, according to NSIDC numbers, is April 2 (in 2010), so still a long ways to go before we can be sure. There has been a trend toward later maximums over the satellite record and recent years in particular have tended to be later than normal. I think this is because the ice extent is lower - i.e., farther north in the colder, darker regions. Ultimately, the maximum is sun-limited, so with the edge farther north there is actually more potential for late growth. As long as the sun hasn't made things too warm, the ice edge can shift around quickly due to thin ice growth and/or winds. So still a lot of potential to grow ice, especially if cold weather is coming to the Bering. Walt Meier NASA Goddard

FYI: While the single-day number is now below 2007, NSIDC uses a 5-day trailing average, which is still not quite below the 2007 5-day average. The reason we do this is because the sea ice data includes noise due to weather effects and other things, so there will be day-to-day ups and downs that are due to the characteristics of the data and not real changes in ice. Five days averages out this synoptic scale noise and, in our view, gives a more robust estimate. For those wondering about the NIC estimates (as can be seen here: http://nsidc.org/data/masie/, NIC produces operational ice analyses, focused on using many data sources of varying quality and quantity to detect as much ice as possible, even small concentrations. NSIDC's passive microwave data may miss some low concentrations (it uses a 15% concentration cutoff), particularly during melt. So it's not unusual for NIC/MASIE to show more ice, though it's more than in other years because the low concentration ice is scattered over a much larger area. An important point is that NIC/MASIE, while picking up more ice, is produced via manual analysis and the data quality and quantity varies. So the product is not necessarily consistent, particularly from year-to-year. NSIDC's product is all automated and consistently processed throughout the record. So there may be some bias, but the bias is consistent throughout the timeseries. This means that comparison of different years, trend values, and interannual variability are more accurate using NSIDC. Hope this info helps. Walt Meier NSIDC