2016 SIPN Sea Ice Outlook: June report

The first Sea Ice Outlook of this year has been published. The SIO is organized by the Sea Ice Prediction Network (as part of the Arctic research program 'Study of Environmental Arctic Change', or SEARCH), and is a compilation of projections for the September 2016 Arctic sea ice extent, based on NSIDC monthly extent values. These projections are submitted by professionals as well as amateurs (public outlooks).

Here's the summary for the June report:

The distribution of Outlooks for statistical and dynamical models have a median extent that is 0.15 million square kilometers different from each other, with a slightly larger spread for the statistical models than the dynamical models. The same sort of spread difference was seen in 2015. Separating out fully-coupled dynamical models from ice-ocean models shows that fully coupled models give a slightly lower forecast with a median of 4.57 versus 4.62, yet have a larger spread (Figure 2). The median forecast of heuristic forecasts is 4.0 million square kilometers and the spread is small. However, actual contributions to the two polls show a diverse range from 2.9 to 5.3 million square kilometers.

The overall width of the distribution of Outlooks has narrowed compared to last year: the interquartile spread across all types of contributions is 0.53 million square kilometers, which is a decrease of 0.27 million square kilometers since last year. The overall range is also reduced to 3.4 to 5.23 million square kilometers.

And here's the figure showing all the projections (click for a larger version):

 The September minimums for the last 11 years (in million km2, found here):

    2005: 5.57
    2006: 5.92
    2007: 4.30
    2008: 4.73
    2009: 5.39
    2010: 4.93
    2011: 4.63
    2012: 3.63
    2013: 5.35
    2014: 5.28
    2015: 4.63

There has also been a poll on the Arctic Sea Ice Forum with 105 people voting. The most voted for bin was 'between 3.75 and 4.0 million km2' and the average was 3.59 million km2. The median, however, was 3.81 million km2, which is almost half a million km2 lower than the June SIO median of 4.28 million km2. If either of these two medians - or any number between them - turns out to be correct, 2016 will end up second lowest on record.

Not that I have any favourites, but the predictions I pay most attention to, are those from the PIOMAS team (4.2 million km2), the CPOM team (4.5 million km2) that has melt pond distribution as its basis (see here), and those from people who regularly comment here and on the Forum, such as Chris Reynolds (4.6 million km2) and Rob Dekker (3.8 million km2).

One other prediction I find interesting, is that by Ionita-Scholtz and Grosfeld (4.7 million km2), because I met the former at EGU2016 where she told me she's researching how sea ice loss affects atmospheric patterns (for instance, winter weather in Europe due to sea ice loss in the Barentsz-Kara region). It'd be cool if she could predict these things in advance, so I know when to start digging the plants out and put them in the attic. ;-)

It'll be interesting to see how these predictions evolve as the melting season progresses. It's very difficult to do a prediction based on May data alone, but we'll know more after June and July. In the meantime there's a new poll up on the Arctic Sea Ice Forum, where you can vote on what you think the NSIDC 2016 Arctic SIE September minimum will be. Voting closes on July 12th. Things are about to get exciting/interesting in the Arctic, so be sure to vote!

Read the entire 2016 June Outlook Report, a summary of current conditions and more detailed predictions here.

Comments

[D]

Slater - persistence at 4.3 is pretty much in the dead center of the predictions. It looks like the citizen scientists didn't change the mean much at all.

[wayne]

It seems quite possible , the Met office has a new sea ice model, rather than couple of guys at the cafeteria calculating on a napkin. 3.6 million seems pretty all right. The NASA guys have definitely a model, 5.2 seems off, I like RASM the best. But if any of these have a model animation output readily available to the public to peruse, please show!

[wayne]

OK, here is my Excel, not napkin calculation, from yesterday assume 50,000 km2 a day 2016 loss till minima: about 4,437,072 would be the Minima. Now the average for an alleged "cold year", 2013, was a daily loss 55,838 km2 a day. If we apply a 50,000 a day loss for 2012 we get 4,513,480, minima really was 3,177,455 with daily average melt of 73,781 km2 a day.

The average daily loss for 80's and 90's was 54,990 and 57,530 respectively. Let us assume worse case scenario 56,000 km2 a day loss and the graph above really starts to be accurate at NCAR and 4 more projections, 25 science centres or groups are seriously off. What exactly makes their models go off kilter?

[Hans Gunnstaddar]

There's an easily discernible trend in the stats relative to the record minima years. Each new record is followed by higher minima's until it peaks, then starts going lower and continues until a new record occurs, then starts the cycle over again.

Follow it:

2007 new minima sept. record
2008 higher than 2007
2009 higher than 2008
2010 lower than 2009
2011 lower than 2010
2012 New minima sept. record
2013 higher than 2012
2014 About the same as 2013 (tad less, and this is the only year out of sync, but only a by .07, otherwise the trend is unmistakable.
2015 lower than 2014
2016 June SIO median of 4.28 would be lower than 2015
Which means based on this sequence of post minima peaks followed by a descent until a new minima record occurs, 2017 should set a new minima record.

Ok, now maybe I'm getting carried away, but if we take the difference between 2007 & 2012 is -.67

2012's minima -.67 = 2.96, call it an even 3, which is what it should come in at approx. in 2017.

[wayne]

I'll bring it down to best case for sea ice 56,000 km2 a day for 71 days, assuming September 15 is minima. This makes worse case 4,000,000 km2 loss from July 5 to minima. About 4,500,000 would be best scenario for 2016. Assuming lesser than 2012 daily average drop say 65,000 km2 a day we get a minima =3.9 million. Leaving for same as 2012 76,000 a day, giving 3,084,000 km2. a new record.
4.5 million best case scenario makes 18 groups or institutes out of 30, in the correct range, of which NCAR would be correct in the more modest than 2012 scenario, and none predicting a new record, which is still in my estimate the most likely outcome.

[wayne]

I am not ready to concede acknowledge or even consider that sea ice minima extent will be achieved with daily losses less than the average 80's, putting into question many models grasp on sea ice reality. Especially if you look at temperatures in the Arctic now , one case in point near but away from Beaufort sea is Cambridge Victoria Island with unrecognizable temperatures in the 20s for days, similar temps to Southern Canada https://www.weather.gc.ca/city/pages/nu-15_metric_e.html

Yesterday Kugluktuk (On Arctic mainland coast) had a high of 29.2 C. If only London UK could have such Arctic weather, they would be happy and confused.

[Cato Uticensis]

Good news from PIOMAS in the meantime. Actually just a confirmation of what DMI has been showing in the last few weeks. Not surprising in itself: weather conditions in the Arctic have been characterised by persistence of low pressure systems and in spite of high-pressure conditions developing on Beaufort and CAA, the CAB, ESS and Laptev remain under the influence of LPs. The decrease in ice extension these days is in my view related with the quick decrease in Hudson, Baffin and Beaufort. It will be much more difficult to get these decrease rates in the next weeks as Hudson will be gone and provided the current weather forecasts are confirmed.

[wayne]

Cato,

I seem to recall that 2012 Laptev wasn't so melted at this date, surely less than now:

http://igloo.atmos.uiuc.edu/cgi-bin/test/print.sh?fm=07&fd=06&fy=2007&sm=07&sd=06&sy=2012

2016 is also "potentially" well ahead of 2012 because of Baffin and Hudson Bay 2016 lag. 2012 didn't have a massive dipole like 2007 as well. A single persistent cyclone neatly moving along the transpolar current very well is neither a good example or excuse for 2016 not being top 3.

Finally is there anyone out there thinking that from now on, 56,000 km2 average daily loss is too pessimistic? I got a parka for sell at Victoria station 2nd washroom down. We wont need so many of them now.

[Cato Uticensis]

Hi Wayne, actually I'm just a beginner here and I do not pretend to be able to provide a sensible forecast for the ice extension in September. The long-term trend is very clear and non-questionable. I just try to see the glass half full, that's it. In terms of weather conditions, after a June more or less in line with 2015, we are having a cooler July. Moreover the volume is showing a decent performance in the last weeks and the thickness maps provided by DMI show much better conditions for the CAB than in 2014 and 2015. I've created a picture comparing the last 4 years but I can't find a way to attach it, sadly :( Anyway, only time will tell. Based on Neven's precious analyses and considering the current trend at synoptic level in the Arctic I expect that 2nd half of July and August could be less favourable to ice retreat than in 2015 and 2016. Just my feeling.

[Cato Uticensis]

"less favourable than 2015 and 2012, of course..."

[Artful Dodger]

Wow, Wayne. 20.2°C at CYCB today at solar noon, and still more than 3 weeks until 1st "sunset" (well, "1st moment of twilight").

So how do you say "scorchio" in Inuinnaqtun? The "Qimmiit" must be suffering in that heat. Hope they get a polar dip to cool off!

Cheers,
Lodger

P.S. Nevin (if you don't already own one) I think we've found your hat! ;^)

[D]

The ECMWF shows a stunning heat wave developing on the Atlantic side of the Arctic according to the 850mb temps. Those temps under high pressure over an area where the sea ice has retreated the most will warm the open waters to possibly unprecedented temperatures over the next 10 days if the forecast verifies.

The DMI SST map as well as NOAA's and the reanalyzer show temperatures in the Barents and Kara seas are already way above normal.

The build up of heat on the Atlantic side will affect this year and also 2017 because the heat pulse around Svalbard runs deep.

-Fish

[wayne]

Artful,

That is a nice photo, I don't recognize the parka, must be either Russian Inuit or Innupiat Alaskan, I almost thought Greenlandic but the kamiks are wrong. Think of the Polar bears as well,
both free moving dogs and polar bears will have a swim . Remember the polar bear in a zoo cage in summer? It suffers and has a huge near permanent frown, most times Polar Bears smile on sea ice.

Lets compare Cambridge UK weather forecast

http://www.bbc.com/weather/2653941

With Cambridge Nunavut Canada:

https://www.weather.gc.ca/city/pages/nu-15_metric_e.html

Its warmer in the Arctic!

Cato:

We need billions of beginners to get aware, welcome to the club!

"we are having a cooler July. "

That is very much questionable. If you have sea ice in Arctic summer outside the North Pole sun. There is a diurnal temperature variance associated with top of sea ice temperature. Top of ice now should be between -1.8 and 0 C. The lower midnight sun at latitudes 70 to 80 doesn't warm the ice as much as local apparent noon. Surface temperatures are largely influenced by top of ice at night.
Presence of sea ice in great extent moderates any warming, unless overwhelming. Cloudy conditions flatten the temperature variance.
Newly opened water usually is quite cold, about -1.8 C. The only thing increasing surface temperatures is warming of ice and water, but at a snails pace. I am not very confident about the models capacity to calculate accurately surface temperature because there are a great deal of possible surface variances. Its complex, and we have very few buoys to study.

D,

Russian side temps are equally hot. Particularly Eurasia. The only cooling thing are clouds. In the next few days they will vanish a little more.

[Neven]

The ECMWF shows a stunning heat wave developing on the Atlantic side of the Arctic according to the 850mb temps. Those temps under high pressure over an area where the sea ice has retreated the most will warm the open waters to possibly unprecedented temperatures over the next 10 days if the forecast verifies.

The DMI SST map as well as NOAA's and the reanalyzer show temperatures in the Barents and Kara seas are already way above normal.

The build up of heat on the Atlantic side will affect this year and also 2017 because the heat pulse around Svalbard runs deep.

What a coincidence, Fish. I just wrote about that this evening for the latest PIOMAS update:

"Although clockwise drift speed has been higher in the Central Arctic during the first half of this year, it doesn't seem to have resulted in as much export through Fram Strait as during the 2006-2016 period. I'm not sure how important this is, as this year the heat has decided to come to the ice, instead of the other way around."

[D]

Neven: Of course, I read what you wrote before I wrote that comment but I added a number of details that were not in your post.

The Arctic is affected by the dynamics of the ocean to more than 300m deep and the dynamics of the atmosphere up to about the 5mb level in the stratosphere. Most of the papers on the Arctic and the discussion here focus on the ice and the 30 meters above and below it. I entertain myself by looking at all kinds of obscure information about the oceans and the atmosphere. Sea ice is just one of the things I follow.

[Rob Dekker]

The SIPN report shows predictions made based on May data.
My entry was 3.8 as you can see in the report.

Meanwhile, the June numbers are in, so I would like to share my updated projection for September sea ice extent.

As many of you know, my method is based on estimating the 'albedo' effect of Arctic amplification during the melting season, and uses 3 variables to determine how much heat the Arctic is exposed to :
- Land snow cover
- Ice area
- Ice concentration
http://neven1.typepad.com/blog/2013/07/problematic-predictions-2.html

Regressed over past (1992-2015) extent, I obtain between 4.0 and 4.1 M km^2 as the most likely outcome of the September sea ice extent this year. That is 200-300 k km^2 more than my May data outlook, which is mostly caused by reduced ice concentration in June, due to the persistent lows over the Arctic during the month.

This is what this projection method predicted for the past 24 years :
http://i1272.photobucket.com/albums/y396/RobDekker/JunePredict_zpsquedrtdc.png

By itself, this projection is not so interesting. Just another guy with another projection for Sea ice extent in September.

But more importantly, the standard deviation over this prediction is 340 k km^2, which is pretty darn 'tight', which means that we can say something about the probability of 2016's ranking.
Based on these numbers :

There is a 92% chance that 2016 will end up in the top-3 (only 2 out of 24 years showed a larger difference between 2011 and this 2016 projection).

There is 66% chance that 2016 will be second place (after 2007 but shy of 2012).

There is an 8 percent chance of 2016 beating the 2012 record in September.

You can humiliate me if 2016 will beat all records, and if it will not make the top-3. Until then, I claim that average weather will get us to 4.0-4.1 or so, and that is bad enough.

[wayne]

Is Probably the best forecast out there Rob...

But I must point out that there is nothing average about this year, ECMWF is showing a stronger presence of Anticyclones, thus reduction in albedo. It is pretty much what I expected, namely the Lows will hug open water areas. However the situation is newish, a substitution dipole where a High hangs more over the Pole area, and a Low over Beaufort. The CAA coast will never be so open when tides are not so strong, has already happened. Despite contrarian winds the Beaufort Gyre is still clockwise as well. Quite remarkable, the sea ice off Banks Island was more open 1 month ago, there is compaction, here we see similarities with 2007, Sea ice is lead the melted warmed up Beaufort despite the weather.

[P-maker]

Rob,

No need to humiliate anyone here. On the contrary, you deserve all our respect for providing your evidence in such a clear way.

I’m happy you have allowed an 8 % window to go lower than extent observed in 2012.

One aspect, which may eventually help to improve your – already excellent – modelling work, could come from the fact that this year is also clearly a “post-ENSO” year.

If you compare your estimates for 1998 and 2007 (also clearly post-ENSO years) , you will see that your estimates were roughly a quarter and half a M sq. km respectively higher than observed.

Assuming a 10 year doubling time for such an accumulated post-ENSO heating signal, 2016 could well be the year that your estimate turns out to be about 1 M sq. km too high.

Should we end up near 3 M sq. km by September this year, it may be valuable to try to include such a factor in your model.

I still consider 2012 to be an outlier due to the GAC.

[AbbottisGone]

Slater - persistence at 4.3 is pretty much in the dead center of the predictions. It looks like the citizen scientists didn't change the mean much at all.
Posted by: D | July 06, 2016 at 00:37
<<

Is this meant to be a type of propaganda against the meaningfulness of bothering to argue online?

Philosophy is always a slippery slope that we all run away from, including- I suspect- "D"!

[I'll let this one stand, but don't go the ASIF way; N.]

[Rob Dekker]

Thanks guys !

Wayne, June was definitively dominated by low pressure (low temp) over the Central Arctic, and thus below-average melting. That resulted in an increase from 3.8 (May data) to 4.1 (June data) in my method. I am not very good with these weather maps, but are you saying that there are higher pressure systems settling over the Central Arctic in July ? Or will we continue to see low pressure systems dominating ?

P-maker, I don't see that ENSO signal in my residuals. 1998 was estimated a little bit too low, and 2007 was estimated too high.
The largest excursions from my estimate were in 2001 (estimated too low), 2004 (estimated too high), 2006 (estimated too low) and 2010 (estimated too low).

If you know of any effect that explains 2001, 2004, 2006 and 2010, then please let me know.

2007 was estimated too high, but there I know the cause : a sustained dipole over the Arctic, causing compaction on the Pacific side and ice export on the Atlantic side.

2012 was estimated too high too, but not by much.
That suggests that there was a LOT of heat absorbed in the 2012 melting season, and the GAC just topped off the losses that were largely expected based on this method of prediction.

[Rob Dekker]

My method gives only a single number for the September ice extent, based on three variables (snow cover, ice area and ice concentration).
It does not say anything about how that ice in September is distributed.

Dr. Slater's method is based on only one variable AFAIK (ice concentration) but it DOES provide an indication of how ice is distributed. The latest distributed ice extent from Dr. Slater's model is here :

http://cires1.colorado.edu/~aslater/SEAICE/prob_map_reg_2016.gif

Now, the thing that is concerning about this projection is it suggests that a large chunk of ice by be (with a 60-70% probability) will be cut-off from the main pack by the end of August.

If that really happens, we may see a repeat of 2012, with "flash-melt" events of isolated ice fields away from the pack with the lightest of storms at that time. Which is why I believe that my projection has more downside potential than upside potential...

[wayne]

Rob,

"but are you saying that there are higher pressure systems settling over the Central Arctic in July ?"

Yea , that was foreseen and happening as of now:

http://www.ecmwf.int/en/forecasts/charts/medium/mean-sea-level-pressure-wind-speed-850-hpa-and-geopotential-500-hpa?time=2016071000,0,2016071000&parameter=Wind%20850%20and%20mslp&area=North%20hemisphere

Anticyclones may be generated if the surface is cooler for a wide area. AS opposed to Cyclones, usually created by planetary waves mainly above much warmer sea water. The current ongoing Pole High should hang out over the main pack, or what is left of it. But may vary in location (as ECMWF predicts), given multiple reasons due to General circulation. If by any chance it settles over the Arctic Ocean Gyre for a significant length of time (its forecasted close to it in about a week), no need to calculate if 2016 will be all time lowest extent in September.

[D]

Actually, my point was that Chris Reynolds and Rob Dekker are using approaches that are as scientific as the so-called experts.

Rob's approach appeals to me. His forecast is much more plausible the ridiculously high forecast by the NASA group. Chris Reynolds method reflects his training in engineering and is based on an in depth review of published research. His forecast is well-reasoned and within the range of the possible although I personally think it is a bit on the high side.

What bothers me is the so-called experts who come up with forecasts outside of the range of reason given what we know right now about extent, area and volume. The so-called citizen scientists are doing just fine.

[wayne]

There was 1 sea ice weather station which reported +6 C , #48594 , about 81 N. two others near the North Pole reported +2C. Very plausible, likely highly accurate, very warm, especially near the Pole. True to location, near North Pole stations seem to be +2 C 24 hours a day. Interesting for many reasons, DMI reports too low about 1 to 1.5 C for 80N, that is what I would suspect given that models may under calculate surface temperatures, as revealed here:

http://eh2r.blogspot.ca/2016/06/the-models-may-be-calculating-sea-ice.html

There is also a thin veil of clouds where the Pole High is, again intriguing, but no one there to report what it is.

[Colorado Bob]

Some really clear images coming in from the Terra and Aqua satellites today and yesterday
Can alone tell me if this baby has a higher albedo than just open ocean ?

Huge algae bloom North of Russia

Aqua/MODIS
2016/193
07/11/2016
08:40 UTC

https://lance.modaps.eosdis.nasa.gov/cgi-bin/imagery/single.cgi?image=crefl2_143.A2016193084000-2016193084500.2km.jpg

[Colorado Bob]

Sorry, "anyone" , not "alone".

[Colorado Bob]

Some really clear shots of Greenland coming in today , the Northeast coast , blue water every where with lots of glacial milk in fjords …………….

Aqua/MODIS
2016/193
07/11/2016
05:30 UTC

https://lance.modaps.eosdis.nasa.gov/cgi-bin/imagery/single.cgi?image=crefl2_143.A2016193053000-2016193053500.250m.jpg

[Colorado Bob]

The top half of Greenland , truly amazing images –
Aqua/MODIS
2016/193
07/11/2016
08:45 UTC

https://lance.modaps.eosdis.nasa.gov/cgi-bin/imagery/single.cgi?image=crefl2_143.A2016193084500-2016193085000.4km.jpg

[Jim Hunt]

Bob - Have you ever explored NASA Worldview? See e.g.

http://go.nasa.gov/29LPlor