As of October 20, 2017, NOAA@NSIDC Releases Sea Ice Index, Version 3.0
Text below from SII webpage explains the changes in versions.
NOAA@NSIDC is pleased to announce the release of Sea Ice Index Version 3 (V3). V3 changes the way the monthly average area and extent data values are calculated. The way monthly average area and extent images are created remains the same as in V2. All daily data remain the same as in V2.
The V3 method simply averages daily extent values, while the V2 method derived monthly average values from the gridded monthly-average concentration field. The change is in response to questions about what seemed to be an inconsistency between daily and monthly values. When users summed daily values, and then divided by number of days in a month to get a monthly average value, that number was different, and sometimes quite different, from the monthly average numbers we presented. Both the V2 and V3 methodologies are valid and defensible ways of representing passive microwave-derived sea ice concentration data, but the goal of this change is to better match the understanding of the user community as the product evolves through time.
Values will change slightly. V3 monthly average areas are slightly greater than those from V2. Conversely, V3 monthly average extents are slightly less than V2. Trends will change slightly as well. These are small changes, as the following images for March, over the entire time series, illustrate. Extent is on the left and area is on the right. Windnagel et al. (2017) offers additional analysis on why V3 areas are higher than V2 and V3 extents are lower than V2, along with a simple example. With the update to V3, there are no considerable differences in conclusions that can be made about the overall trends in sea ice area or extent.
Details are in this report: Sea Ice Index Version 3 Analysis. NSIDC Special Report 19.
The Sea Ice Index version 3 dataset is available here
Several posts at this blog compare results from MASIE and SII. Walt Meier, head of Sea Ice Index (SII) and colleagues did a comparison of the two datasets, published in October 2015. SII went from version 1 to v.2 in 2016, and then in January 2017 from v.2.0 to v.2.1, with some changes in past values.
One of the biggest discrepancies in the past with previous SII versions arose in the Month of October, and it is timely that version 3.0 appears now. The dataset is now available with past calculations according to v3, and the past can be compared in advance of 2017 monthly results.
For comparison, the graphs below show October monthly averages for MASIE compared to SII v.2 and SII v.3., 2007 through 2016.
It seems clear that SIIv.3 is a big improvement relative to MASIE. And as well, October 2017 results are coming in with the two datasets tracking quite closely.
MASIE: “high-resolution, accurate charts of ice conditions”
Walt Meier, NSIDC, October 2015 article in Annals of Glaciology.