For climate analysis we consider the average monthly extents for March and for September of each year in the satellite record, and the differences (the melt extent). Though we would prefer a longer record, these are currently the most popular data. Several observations:
March averages (annual maximums) do not vary greatly: 15.48 M Km2 is the average extent, with a range of 16.45 to 14.43 M Km2. 2/3 of the years are between 15 and 16M.
September averages (annual minimums) vary much more: 6.40 M Km2 is the average, with a range of 7.88 to 3.63 M Km2. Standard deviation is +/- 1.07 M Km2.
Note: The largest September extent (7.88) in the record occurred in 1996, the same year of the smallest melt extent: 7.25. And the smallest September extent (3.63) occurred in 2012, due to the largest melt in the record, 11.8M. The March extents of those two years were nearly the same.
The Arctic ice extent time series appears to consist of three periods:
1979 to 1996 Annual minimums mostly above average
1997 to 2006 Annual minimums around average
2007 to 2014 Annual minimums below average
|1979 to 1996||15.8||7.2||8.6|
|1997 to 2006||15.3||6.2||9.0|
|2007 to 2014||15.0||4.5||10.5|
Since 2005 the combination of below average March extents, combined with above average melts has produced September extents below 6 M Km2 each year.
It is now evident that 2012 was an outlier (probably due to the unusual storm activity). That year’s melt of 11.8 was 28% above the average melt of 9.09 and more than 1 M km2 larger than the second largest melt in 2008.
The pivotal decade was 1997 to 2006, preceded by slightly declining extents, and followed by much lower extents. What any of this has to do with CO2 and air temperatures is not obvious.
Data is here: