The seasonal cycle of sea ice extent in Antarctica has been fairly stable over the length of the satellite record.  There is a slow growth of sea ice from a minimum of ~3x10⁶ km² in February to a maximum of ~19 x10⁶km² in September, before there is a relatively rapid fall in the Antarctic spring.

But this year something different is happening.

Below is Tamino's image for the Southern Hemisphere sea ice extent, the red line is 2016 up to 16 November 2016.

From January up to September the sea ice extent in 2016 follows all previous data.

But what happened in September?

After that date Southern  Hemisphere sea ice extent dipped, and reached historic lows by the end of October.

I noticed something was up at the time and, on the 10th September, I blogged that Antarctic Sound was ice free.

This is the image I used

At the end of September in a post called Low Antarctic Sea Ice Extent, I blogged about how the sea ice in the Weddell Sea was being pushed away from the Antarctic Peninsula. This led to some surprising open water regions. This is the MODIS image I used:

In a post called Western Weddell Sea ice factory I pointed out that this would lead to a lot of sea ice growth.

Another pointer to the pack being looser this year is that in August 2016 in the Eastern Weddell Sea there was a rare sighting of the Weddell Polynya (actually over a relatively shallow region called Maud Rise).

I also pointed out in four posts how polynya were forming in the Amundsen sea, and that the polynya were then developing. I noted that there was very rapid sea ice grown in the Dotson Getz Polynya, and finally that The Prince Gustav Channel is opening. That last one is a bit of a surprise given what I have seen the last couple of years.

As it is currently, I stand by what I said at the beginning of October:

I think the Weddell Sea pack ice is more mobile this winter.

To untangle what the sea ice extent observations are telling us, I did some calculations. The next plot shows the difference between the daily ice extent, in each grid cell, for 5 August 2016, and the mean, for each day, based on historical data for the 5 year period 1989-93.

On 5 August the sea ice extent was pretty much as we expected it - in that it was within the normal range of sea ice extent. White shades mean that 2016 has no deviation from the 1989-1993 mean.

It's the dark shades of red and blue you should take notice of.

Now you are orientated, this is how the difference from the mean sea ice extent changed from 17 August to 21 November 2016.

During September the sea ice gets compressed against the Antarctic Peninsula. The same winds moved and loosened the sea ice in the Weddell Sea as I wrote in my post Low Antarctic Sea Ice Extent. This is why 2016 starts to deviate from the previous observations.

At the same time there is also  a compression of sea ice south against the continent in two clear areas: the Cosmonauts Sea, and the Ross Sea.

Overall it looks like their are three large sea ice anomalies: One in the Ross Sea, One in the Cosmonauts Sea and finally one in the Bellingshausen and Weddell Sea I have written about previously.

So that is the answer. The sea ice is at a historic low in the Southern Ocean because in three locations from September 2016 onward, the sea ice has been pushed to the south and away from the peninsula by the winds. This has been superimposed on top of the expected usual seasonal retreat.

The next job is to look at the mean monthly surface pressure fields for September-November and see how much they deviate from the mean field. This will tell us about the winds that have been shifting the sea ice about.

Above is the sea ice extent anomaly for a short period, I put the year to date from 1 January to 21 November 2016 on a separate post: The Antarctic 2016 sea ice anomalies.