Fourth shortest melt season in Arctic

[xmetman]

Rather bizarrely the melt season in the Arctic was the fourth shortest in the Sea Ice data series that started in 1979. The melt started on March 21st of this year and ended on the 7th of September, which made the melt season 169 days long, the shortest since the 166 days of 1997. The short season is down not just to a very late maximum, but also an early minimum, but why that should occur in a season that saw the second lowest maximum is slightly puzzling.  The spring maximum was 13 days later than average, and the autumn minimum was 4 days early than average.

[Alastair B. McDonald]

Yes, it all depends on the weather which is as we all know, or at least as I keep saying, is chaotic. It makes it difficult to detect the signal of anthropogenic global warming. What would trend lines look like on your two charts?

[Alastair B. McDonald]

Oh! I see those are trend lines not averages. So the melt season has increased by 0.1 days in an average of 185.5, and the freeze season has increased by 0.3 days from an average of 277.2 (but I think that should be 177.2). Interesting that they do not sum to 365.25.

[xmetman]

Alistair 

I did add a linear trend to the Arctic chart, and if you look it's almost horizontal.

In the blog I added this extra chart:-

In the Antarctic as you can see there has been a  lengthening of the freeze season by 6 days and a reduction of the melt season by 11 days since 1980.

[Paul Garvey]

Just weather conditions Bruce. 

[Alastair B. McDonald]

The Antarctic trends fit with the increase in Antarctic sea ice until this year. It is strange that the Arctic trends do not fit with the decrease in Arctic sea ice. It's a pity Graham Davies is not here. He might have some useful ideas.

[Paul Garvey]

The main trend is very clear, however. A warming Arctic means a clear trend over time of less summer ice. Weather conditions in some years overall (like this one - very stormy) or at times during the Arctic melt season alter the rate of melting. The Antarctic is a wholly different kettle of penguin. *>))

[Alastair B. McDonald]

Bruce,

I thought that you might be interested in this paper which has just been published: http://www.the-cryosphere.net/10/2027/2016/tc-10-2027-2016.html "Sea-ice indicators of polar bear habitat."

They report that "In all 19 regions there is a trend toward earlier sea-ice retreat and later sea-ice advance. Trends generally range from −3 to −9 days decade⁻¹ in spring and from +3 to +9 days decade⁻¹ in fall, with larger trends in the Barents Sea and central Arctic Basin."

It seems strange that they get a different result to you.

Cheers, Alastair.

[xmetman]

Hi Alastair

If we are using the same SI data, I can't see how I could misconstrue the season length when it's simply the difference in days between each maximum and minimum.

The linear trend that I've added to each graph shows little change since 1980. 

All I can conclude is that they may be using different data (maybe MAISIE) or the linear trend that I use is not as sophisticated as whatever method they are using.

I would love for someone else to download the same data and look at it in a spreadsheet and tell me where I'm going wrong, so far no one has.

With apologies to Pete Postlethwaite (in the film Brassed Off), I may only have a basic understanding of statistics but I can knock out a bloody good graph!

Bruce.

[Alastair B. McDonald]

Hi Bruce,

They say " We define the dates of sea-ice retreat and advance in a region as the dates when the area of sea ice drops below a certain threshold (retreat) on its way to the summer minimum or rises above the threshold (advance) on its way to the winter maximum. The threshold is chosen to be halfway between the historical (1979–2014) mean September and mean March sea-ice areas." So they are calculating it differently.

I think I now have the answer. They are considering regions where for part of the year there is no ice cover. Thus the time between end of ice formation and the start can increase as the Arctic warms. Since you are considering a region (the whole of the Arctic) where ice is there all year round, then there is no change. In other words, the time during ice expansion and ice retreat must alway add up to 365 days, and the two times will typically be 365/2 = 182.5 days.

Alastair.

[xmetman]

That makes sense! 

One day that may happen to the whole of the Arctic.