Monday 11 May 2015

An Early Warning for Summer Sea Ice Crashes?

Just a quick post on on of the elements I've looked at while trying to improve prediction for the summer. This comes with a caution, because the method outlined only 'predicts' two recent sea ice summer loss events, the 2007 and 2012 crashes. Time will tell whether it is of any use in the future.

Compactness is the ratio of area / extent. Looking at compactness in the Arctic Ocean and the Peripheral Seas (Beaufort, Chukchi, East Siberian Sea, Laptev), and using two periods, June average and Late June average (10 to 30 June); here is a graph of the four indexes.

The Peripheral region shows a greater variation of compactness for both June and Late June, so I will focus on that. The Late June shows a greater variation than for the whole of June, and shows the possible utility of this idea. For those suspecting a link, this is indeed related to the 'June Cliff' in area.

My suggestion is to keep an eye on 10 to 30 June average compactness for Beaufort Sea, Chukchi Sea, East Siberian Sea, Laptev Sea. In 2007 and 2012 compactness for late June fell to around 0.7, unusually low for the 1979 to 2014 period. 2007 and 2012 were both years in which new records of low September extent were set. But it is not new records per-se that I am looking for, rather years of very low September extent. So in summary:

If June 10 to 30th average compactness for the Peripheral Seas of the Arctic Ocean Basin is around 0.7 for a given year, a large summer melt can be anticipated, but is not guaranteed, over the following summer.

For completeness: Here are the new records in NSIDC September Extent and the Peripheral Late June index.

I did say likely, not guaranteed. ;)


Neven said...

Great post, Chris. I've become convinced that May and June are very important for what I have now dubbed 'melting momentum'. Regional CAPIE is extremely useful in assessing how much of this momentum is built up in the key regions you denote, so I hope you'll write about this occasionally.

If I was able to put a number to my visual analysis of initial sea ice state (right after the freezing season) and this second phase of melting momentum through melt ponds, I'd be submitting a forecast to the SIPN Sea Ice Outlook.

Anonymous said...

Is the low compactness a sign of high melt-ponding?

Jim Hunt said...

Chris - Would I be correct in assuming that you're using NSIDC extent and Cryosphere Today area in your calculations?

Anon - Assuming the answer to the question above is yes, then the answer to your question may also be at least partly yes.

I'm sure there's at least anecdotal evidence that the CT area metric is susceptible to counting melt ponds as reduced concentration ice. I'll see if I can dig out something more substantive.

Chris Reynolds said...

Neven, Jim,

I don't use Cryosphere Today, I think NSIDC Area is better and I use Wipneus' area and extent calculations which are based on NSIDC gridded concentration. So this isn't CAPIE, it's compactness.


Not necessarily. I have looked at this:
And have found that most of the decline in June Area is from grid boxes with concentration between 0.796 and 0.400. This is more associated with the region near the ice edge than in the middle of the pack. So what is going on is not just melt ponding. I suspect that thinner ice and river discharge play a role.

Anonymous said...

Thanks Chris, Jim. I got so behind on posts on here, I decided to skip a load and revisit older ones as they get referenced in future ones, if you see what I mean.

BTW, didn't mean to be unnamed anonymouse, hit enter and submitted before I wrote my name, sorry.


Anonymous said...

That's OK Adam, your tactic is what I do with Realclimate and Open Mind.

Chris, problems logging in.

Jim Hunt said...


So your metric of choice is the bright red line on Wipneus' compactness chart?

Chris Reynolds said...


Not really, I haven't got one metric, but that is indeed compactness, and the bright red line is the one to follow. All I am saying is that concentration for the peripheral seas seems to be a potential indicator of a coming large loss summer.

BTW - I didn't want to say so at Neven's or on the other links (too public), but you will lose the 'CT Area below 1M km^2 by 2022' bet.

Jim Hunt said...

1) OK - Thanks Chris

2) If so then at least my money will go to a good home!

In the meantime the ad homs have started at Breitbart, I've just been "reported" to the management as a "spammer", and nobody wants to take the other side of my wager any more.

I don't suppose you fancy a flutter do you?

Chris Reynolds said...


I'm quite happy fleecing denialists, but would rather not fleece you.

Jim Hunt said...

I'm quite happy with the odds you are offering, so feel free to fleece away!

Nightvid said...

So first you say "likely, not guaranteed", then you proceed to OUTRIGHT ASSERT that Jim will be wrong about < 1 M km^2 CT area by 2022 with absolutely zero reservation, as if you are absolutely certain??

Chris Reynolds said...


Two entirely different lines of reasoning.

1) The contents of this post do not indicate enough confidence to be anything but suggestive.

2) My reasoning behind the 'slow transition' idea is solid and detailed. I am so confident I have bet Jim: I pay him £1000 if he is correct. But if I am correct he only pays £1, which I waive in advance.

Those asymetric odds are intended as a test of my confidence. We will not see CT Area drop below 1M km^2 before 2022 as a result of AGW related processes.