Cycle plots of Arctic sea ice

The downward trends in minimum sea ice area, extent and volume have become stunningly obvious to anyone who sees images such as the top row in this blog's long-term graphs page. But what about trends in annual maximum, or the in-between seasons? Cycle plots (Cox 2006) provide an unorthodox but information-rich visualization that tracks changes in the ice for each month across a series of years. Figure 1 shows a cycle plot of monthly mean Arctic sea ice extent and area, based on NSIDC data from November 1978 through August 2012 -- the full period of their satellite record.

Figure 1

Each squiggly line in Figure 1 follows the mean sea ice extent or area for a particular month, across all years covered by these data. The vertical axis indicates ice cover, from zero to about 16 million km². In August 2012, area reached the lowest point yet observed, 2.83 million km², with September (usually the lowest month) yet to come.

The monthly values zigzag from year to year, but their longer-term trends remain clear. Figure 2 summarizes these trends as simple regression lines for each month. All 24 downward lines pass tests for statistical significance at the .001 level.

Figure 2

Trends in northern sea ice volume (PIOMAS estimates) are even more clear. But for a post later this winter (austral summer) I'll take a cycle-plot look at the unclear patterns of Southern Hemisphere sea ice area and extent.

Comments

[Apocalypse4Real]

Larry,

This is great stuff and thanks for posting it.

BTW did Romm get your permission to post your graphs elsewhere on the web? Just thought it would be proper/ethical for him to do so.

[r w Langford]

Thanks Larry, your tables and graphs amaze and educate, me

[Artful Dodger]

Thanks, Larry.

Your charts are like the 'Tai Chi' of Arctic Sea Ice graphs:

"The stillness in stillness is not the real stillness. Only when there is stillness in movement can the spiritual rhythm appear which pervades heaven and earth."

You are one of the "Seven Immortals" ;^)

Cheers,
Lodger

[Ghoti Of Lod]

Larry,

Once again your charts are stunningly clear. I can't wait for the next ones.

[Artful Dodger]

Looking at Fig.2 Arctic Linear Trends, it seems very apparent that September is the odd man out.

The other Months all seem to show a 30yr trend where the start of the monthly series matches the end of the previous month's series.

This of course suggests that the seasons have expanded outward by a month in the Arctic.

However, September stands alone from the other months. To me, I would expect the Sep extent trend line to be placed about where the Sep area trend line is plotted in Figure 2.

Does anyone have an interpretation for what we're seeing here?

Cheers,
Lodger

[Wipneus]

The other Months all seem to show a 30yr trend where the start of the monthly series matches the end of the previous month's series.

I think that is only true for the "melting" months. For September and following it is definitely not the case.

[. ..]

The melting months are now about 1 month earlier than at the start of the series, while the freezing months are now about 1 month later than at the start of the series.

[Artful Dodger]

Hi Wipneus,

...it does seem to hold for Oct - Jan as well, it's just harder to see on the cycle plot during the freeze-up months.

Hi Espen,

Agreed. :^)

I guess we could ask a Polar bear, they'd know.

Cheers,
Lodger

[Peter Ellis]

No it doesn't hold for Oct - Jan

For the melt months, the END of one month's series is about the same as the START of the next month's series. For the re-freeze months, the START of one month's series is the same as the END of the next month's series.

For the month in between (September), neither pattern holds: because they can't both be true!

[Artful Dodger]

Yes Peter, that's a good way of expressing it. I also agree with your conclusion that September is the in between month. Makes good sense. Thanks!

Cheers,
Lodger

[Kris]

Record lovers,

As there is now an open water field with a surface greater than the Mediterranean, much more water gets evaporated, and as a consequence we might expect much more of precipitation in and around the Arctic.

Of which this 5 September Barrow precipitation record is one of the many proofs. For instance, at the other side, Tiksi, the situation is pretty much the same.
And btw, the Barrow precipitation record is just one of the many precipitation records in Alaska since mid August.

More water vapour in the air and more rainfall in and around the Arctic not only will induce a climate change but will be devastating too for the permafrost as well as the tundra.

So, this is a big deal.

[crandles]

>"All 24 downward lines pass tests for statistical significance at the .001 level."

That is clear. :-)

How many pass a test for downward acceleration?

[Jim Williams]

Lodger, the easiest way to see it is by moving September to the other end of the graph. It's both a melting and freezing month, and therefore neither.

[Artful Dodger]

Hi Jim,

"Oh, East is East and West is West, and never the twain shall meet,

Till Earth and Sky stand presently at God's great Judgment Seat;

But there is neither East nor West, Border, nor Breed, nor Birth,

When two strong men stand face to face, though they come from the ends of the earth!"

Who knew Kipling was an Ice Watcher? ;^)

Cheers,
Lodger

[Kevin McKinney]

For the melt months, the END of one month's series is about the same as the START of the next month's series. For the re-freeze months, the START of one month's series is the same as the END of the next month's series.

For the month in between (September), neither pattern holds: because they can't both be true!

Not quite--September's START is at least close to the END of October. And it's no more an 'in-between' month than is March, at the other end of the annual cycle.

Speaking of which, it's interesting how clearly the cycle plot shows the anomalous trajectory of 2012 is, especially round about April--and then, of course, the contrast to the current situation.

[Kevin McKinney]

IJIS is still falling, slowly--today is 3,601,875 km2.

Will it manage to break below 3.6? Seems likely.

[Djprice537]

It seems we are finally approaching the sea ice minimum extent for 2012.

http://www.ijis.iarc.uaf.edu/en/home/seaice_extent.htm

[Djprice537]

It seems that the cyclone of 2012 contributed a great deal to the dramatic year if you look at the acceleration of the loss in early August which set us on different trajectory.

Will denialists use this to explain away the new record as a simple weather anomaly?

[Djprice537]

KM....

I can't imagine that it won't. What are the two weak cyclones doing?

[Dan]

Incidentally, if any of you want to try something similar using excel, I went looking for a how-to guide earlier this year having admired Larry's previous graphs like these. I followed the instructions on this site

http://www.exceluser.com/dash/cycleplot1.htm

and they worked fine.

[L. Hamilton]

"BTW did Romm get your permission to post your graphs elsewhere on the web? Just thought it would be proper/ethical for him to do so."

Good question. Definitely it's thoughtful when someone asks about re-posting one of my graphs, as Doc Snow (a well-respected participant here under his other name) recently did for an essay on Ironies of the Republican Convention,
http://doc-snow.hubpages.com/hub/Republican-National-Convention-Ironies-2012
or Tom Gray did for a blog post at It's Burning,
http://itsburning.blogspot.com/2012/09/indispensable-arctic-sea-ice-graph.html

When asked, I typically say You're welcome! just please mention the source.

I hadn't seen this Climate Progress post until A4R brought it up,
http://thinkprogress.org/climate/2012/09/05/799761/death-spiral-watch-experts-warn-near-ice-free-arctic-in-summer-in-a-decade-volume-trends-continue/
but don't mind when other people re-post my graphs in good faith -- as Romm did in this case. That's the wonder of the internets, and the reason I put my name in small print within each graphic, knowing they might wander far from home.

I would protest if one of my graphs got misrepresented, or deceptively altered in some way. So far I haven't experienced that personally, although I've seen it done to other people's work.

[Jim Williams]

You can tell that Winter is setting in when the Antarctic Anomaly starts to fade and the Arctic Anomaly starts to bloom.

http://www.esrl.noaa.gov/psd/map/images/fnl/sfctmpmer_01b.fnl.html

[Remko Kampen]

Meantime another remarkable low was featured in the basin, pressure down to 970 hPa and storm force winds, hurricane force gusts.
From Jeff Masters' blog:
---
Huge storm pummels Alaska
A massive low pressure system with a central pressure of 970 mb swept through Alaska on Tuesday, generating hurricane-force wind gusts near Anchorage, Alaska that knocked out power to 55,000 homes. Mighty Alaskan storms like this are common in winter, but rare in summer and early fall. The National Weather Service in Anchorage said in their Wednesday forecast discussion that the forecast wind speeds from this storm were incredibly strong for this time of year--four to six standard anomalies above normal. A four-standard anomaly event occurs once every 43 years, and a five-standard anomaly event is a 1-in-4800 year event. However, a meteorologist I heard from who lives in the Anchorage area characterized the wind damage that actually occurred as a 1-in-10 year event. A few maximum wind gusts recorded on Tuesday during the storm:

McHugh Creek (Turnagain Arm)... ... ..88 mph
Paradise Valley (Potter Marsh)... ... 75 mph
Upper Hillside (1400 ft)... ... ... ... 70 mph
Anchorage port... ... ... ... ... ... ... .63 mph

The storm has weakened to a central pressure of 988 mb today, and is located just north of Alaska. The storm is predicted to bring strong winds of 25 - 35 mph and large waves to the edge of the record-thin and record-small Arctic ice cap, and may add to the unprecedented decline in Arctic sea ice being observed this summer.
---

[George Phillies]

I may have missed the prior announcement but the new 9/2 PSC ice volume curve is up

http://psc.apl.washington.edu/wordpress/research/projects/arctic-sea-ice-volume-anomaly/

showing progressive recovery of the anomaly.

[L. Hamilton]

Unfortunately none of the indicators seem to have yet found the floor. DMI down -50k for 9/6:

http://img.photobucket.com/albums/v224/Chiloe/12_Climate/sea_ice_DMI_this_date2.png

[Seke Rob]

Re George Phillies | September 07, 2012 at 15:35

Likely I'm having another of my frequent "lost in translation" moments (my favorite Bill Murray movie), but Plz call me again on your "progressive recovery", when the anomaly hits -5,000 cubic km, which is about where the Arctic was last time on day 349 of 2010, briefly. Probabilites of this to happen?

[L. Hamilton]

Including PIOMAS, down another 188 km^3 from 8/25 to 9/2. Here's a bar chart with the new update:

http://img.photobucket.com/albums/v224/Chiloe/12_Climate/sea_ice_PIOMAS_min_to_date.png

[Doctor__Stanley]

"showing progressive recovery of the anomaly"

...yes, it might now be less than 3 standard deviations below the (sharply declining) linear trend.

I don't know whether George Phillies intended it this way, but when I see the word "recovery" in this context I seem to automatically interpret it as denial/wishful thinking. Either that or ironic humor.

[Artful Dodger]

Hi Larry,

I think the last data point in the PIOMAS dataset, Day 246, is Sep 3, 2012.

2012	246	3.407

A quick count of rows in the data file shows that all years, including leap years, have 365 days. PIOMAS skips Feb 29, with Day 59 being Feb 28 and then Day 60 is Mar 01.

But I'm sure you knew this, just putting it in the record!

Cheers,
Lodger

[L. Hamilton]

Ah, you're right, my routine corrects automatically for leap years but that makes it ahead one day on PIOMAS. I'll fix that soon.

[Chris Biscan]

http://iup.physik.uni-bremen.de:8084/amsr2/index.html

the bread winner is back!

[Seke Rob]

By the looks of this map http://iup.physik.uni-bremen.de:8084/amsr2/arctic_AMSR2_nic.png from above link, anyone who can walk on water, never has to touch ice once going from Baffin to Beaufort, by NWP route.

[L. Hamilton]

OK, fixed the PIOMAS date glitch, thanks to Lodger for pointing that out. I work in Julian dates (days since 1/1/1960) but the tricks in converting from other forms such as decimal dates or the PIOMAS no-leap-years convention sometimes gives me squirrelly results. Anyway, now that bar graph should more appropriately say the last day is 9/3.

http://img.photobucket.com/albums/v224/Chiloe/12_Climate/sea_ice_PIOMAS_min_to_date.png

In case anyone's interested, here's the Stata code definging Julian dates (edate) from the PIOMAS-provided year and day of year (doy):
replace edate= mdy(1,1,year) + doy -1
replace edate = edate+1 if year/4==int(year/4) & doy>=60

[Espen Olsen]

Seke Rob,

I have never seen the CA like that before, it is like a ghost city!

[Protege Cuajimalpa]

Hi, L. Hamilton:

Talking about graphs, I like the Gompertz curve want that you made for Search, so I want permission to use it.

http://www.arcus.org/files/search/sea-ice-outlook/2012/06/pdf/pan-arctic/hamilton.pdf

What I want to do is criticize the graphs of NSIDC, that always use straight lines in their tendencies, when it is obvious that we are having a collapse of the Arctic Sea Ice. I like your graph because it predicts a 2 million km2 at 2020 (I expect that is going to be less, if PIOMAS is right and volume makes a difference in 8 years (2020).

[L. Hamilton]

Protege, you're quite welcome to use the graph. Can you send me a link if you publish somewhere?

A linear fit to minimum extent, area or volume should be rejected now by any statistician because the data are so obviously curved. Still the linear models hang around partly as a legacy, and partly because there's no agreement on what should replace them -- as in our discussions of Gompertz vs. exponential, or whether curve-fitting in general is misguided.

[Protege Cuajimalpa]

Thanks, L. Hamilton.

I was thinking on doing just a Spanish version of my document, partly because I have to make and effort if I write it in English.
But here is an idea: Why we don’t establish the goal, thinking of this Sea Ice Blog as a community, to make NSIDC change their graphs. I will love to see that they start this September with a curve tendency, instead of the straight line tendency. I don’t care if they use a moving average, a quadratic equation or what ever method they like to use, as long as they show that the melting is accelerating. They also don’t need to forecast several years, that is the reason to avoid an agreement in the curve choose. I believe that given that we are about to have a new record, this will be a great goal. The exposition that NSIDC documents will have in October, will have more diffusion that any document that we can do as individuals. So making them adopt a new standard in their graphs, could be a way to make a change in the global perception on what is happening in the Arctic Sea Ice.

[L. Hamilton]

Protege, I'd be interested to see a Spanish version as well, if you don't mind sharing it.

As for NSIDC, I think it's likely their presentations will continue to evolve, partly in response to the growing public sophistication and appetite for information that is visible on this sea ice blog.

[Sheridanmayo]

I see the IMS domino has fallen - no more straws to cling to for Watts et al.

[George Phillies]

and based on the last few years, the anomaly will likely heal up to -7 or-7.5 thousand, and then flatten out until next year, when it will fall again.

[LRC]

Saw this: http://minnesota.publicradio.org/collections/special/columns/updraft/archive/2012/09/arctic_sea_ice_reaches_yet_ano.shtml
Now I may be getting it totally wrong but he seems to feel that with open Arctic weather systems in stead of coming south and impacting our climate farther south could start doing what happens around lakes and create a micro climate that moves from water to land back to the origin water source. If he is at all right, than that would mean clouds would be generated in the Arctic dump its contents and rain or snow then move back the system over the Arctic.
If that were to happen, that would greatly effect the precipitation patterns all over NH and require a strong jet stream to pull any system out of the Arctic especially if a virtually self sustaining cyclone has gotten going.

[Protege Cuajimalpa]

Hi, Larry.
Of course, I will share the document with you and with anyone at this Blog. Not something that this community will not know, but it will be good to make some people think about Mexico's future.

[Kevin McKinney]

"Will IJIS break below 3.6 million?" I asked.

Sure enough:

The latest value : 3,595,781 km2 (September 8, 2012)

And I'm going to go out on a long, long limb here and say that Steve Goddard is now officially wrong in his prediction of an early end to melt season...

;-/

[L. Hamilton]

"an early end to melt season..."

Well, of the 4 daily indicators I watch (DMI, CT, NSIDC, IJIS), 3 reached a new low this morning, and the 4th (DMI) remained flat.

[Artful Dodger]

There's 75 K of soup in the middle-right of this MODIS square, just waiting for the first good breeze to come along.

http://lance-modis.eosdis.nasa.gov/imagery/subsets/?subset=Arctic_r04c04.2012252.terra.367

This is not sea ice. It is see-ya ice.

[Noley]

We all look at ice loss, but has anyone tracked the rate of increased extent, area and volume after the melt season ends? I'm a total amateur, but it seems that viewed over time it would be an interesting data set. Might even show a warmer winter (in relative terms, of course).

The winter ice cover drives a lot of NH weather and it would be interesting to see if there is any influence as the ice grows back, maybe especially because it is all first year ice. Is there any correlation, for example, between rate of area and extent and weather patterns in the NH? Is NH winter weather different with a preponderance of FYI as opposed to MYI?

Just wondering.

[Espen Olsen]

Noley,

There is but it is an equation with more than 1000 unknowns!

[Noley]

Thanks, Epsen. That would certainly make it a bit challenging! It was just something I was wondering about.