August Arctic Sea Ice Outlook Contest

The Sea Ice Prediction Network (SIPN) announces the call for contributions for the Sea Ice Outlook August report.

The Sea Ice Outlook provides an open process for those interested in arctic sea ice to share ideas about the September minimum sea ice extent. The monthly reports contain a variety of perspectives—from advanced numerical models to qualitative perspectives from citizen scientists. A post-season report will provide an in-depth analysis of factors driving sea ice extent this summer as well as explore the scientific methods for predicting seasonal ice extent.

For the August Outlook report, we particularly encourage regional Outlooks and spatial forecasts and maps; both regional and pan-Arctic outlooks will be accepted.

We also welcome any field- or ship-based updates on ice conditions in the different regions and input on which observations are most useful for improving models. Sea ice scientists, volunteer observers and oceanographers are collecting observations routinely by ship throughout the Arctic Ocean. Such observations include sea ice morphology (e.g., concentration, ice type, floe size, thickness, snow, melt pond characteristics, topography), meteorology (surface measurements) and oceanography (temperature, salinity, upper ocean temperature). Which observations are most useful to improving model predictions or validating these models? Are there particular in-situ observations that we should highlight in our efforts to coordinate data collation?

To help with ship-based observations, a SIPN-related effort, IceWatch, provides a framework and a software tool (ASSIST) to help with standardized, ship-based ice observations. Detailed information and data from past and ongoing cruises can be accessed at


Due to the Heartland Conference in early July, we missed be able to poll readers during that time. The July SIPN report is here and the graph from it below.

Figure 1. Distribution of individual Pan-Arctic Outlook values (July Report) for September 2014 sea ice extent (labels on the bar graph are rounded to the tenths for readability. Refer to the Individual Outlooks at the bottom of this report for the full details of individual submissions).Figure 1. Distribution of individual Pan-Arctic Outlook values (July Report) for September 2014 sea ice extent (labels on the bar graph are rounded to the tenths for readability. Refer to the Individual Outlooks at the bottom of this report for the full details of individual submissions).

The contest is to forecast the September monthly average Arctic sea ice extent (in million square kilometers). NOTE: Not the minimum, the average.

For a primer, see the WUWT Sea Ice Reference Page

This year, a model forecast from NOAA CFSv2 says that we might see a minimum something around 6.0 million square kilometers, however that is not a validated model.


JAXA Arctic Sea Ice Extent -15% or Greater:

Japan Aerospace Exploration Agency (JAXA) – International Arctic Research Center (IARC) – Click the pic to view at source

Arctic temperature is running slightly below normal:

Arctic Temperature:

Mean Temperature above 80°N

Danish Meteorological Institute – Click the pic to view at source


My best guess is about 5.6 million sq km for the September average. You can add your prediction to the poll. Few seem to think it will be a repeat of 2012, and the only issue now would be if we have some inclement weather like that year that broke up the ice.

The poll runs 24 hours and closes at 9AM PDT Tuesday, as the deadline is close of business that day. The average of the top 5 responses (which usually exceeds 50% of the vote) will be submitted as WUWT’s estimate, unless there is a single number voted in that exceeds 50% of the vote.


newest oldest most voted
Notify of



I’m feeling optimistic this year, so I go for 5,8 million. Look at the sea ice volume anomaly change during July, according to PIOMAS…comment image
Yes, it is calculated, not measured, but still it looks promising.
Both area and extent are slightly above last year’s values by this date, and more or less by the same ammount, so the ice seems similarly packed, and if volume is greater…

D.J. Hawkins

I’m in for 5.5.

John F. Hultquist

NOTE: Not the minimum, the average.
average = mean ??
The mean is biased by the outliers that, in this case, will be the minimum. Thus, go low.



Steve from Rockwood

5.5 and I haven’t been right yet.


Meier was quoted recently as saying the ice was thin and spread out, and thus could drop noticeably if it got blown around.


“This year, a model forecast from NOAA CFSv2 says that we might see a minimum something around 6.0 million square kilometers, however that is not a validated model.”
Not only is it not validated, it seems to have been performing quite poorly this year. Until recently, the model was still predicting above average summer minimum. It continued to do so well after a naked-eye test indicated that above average was absurd.

5.5 with minimum of 5.0. Unit = Gm²

Michael Jennings


Sorry, correct unit is Mm².


“Arctic temperature is running slightly below normal”
A couple of things that I think are worth pointing out about the DMI temperature plot:
1) It only analyzes temperatures above 80 deg latitude. Even as we near the summer sea ice minimum, a substantial amount of ice exists below 80 deg: Local weather conditions can easily make for above average temperatures at those lower latitudes
2) They mark 0 deg C on the plot. As everyone knows, this is the freezing point of pure water. However, seawater’s freezing point is a couple degrees lower due to the salt content, so don’t expect the melt to stop when the red line drops below the “freezing” line in the next couple of weeks.
3) I recall reading at one point (although I can’t find the source, maybe somebody else can locate it?) that the DMI plot uses lat/long-weighted averaging rather than area-weighted averaging. The result is that it places undue weight on the far north latitudes, where it is undoubtedly cooler, skewing the whole plot towards a cold bias in terms of absolute temperature.
In short: by all means use the DMI plot as a tool to help guide your guestimate of sea ice minimum, but please use with care and do not simply assume that since it says the arctic temperature is about to drop below freezing that we are essentially at the annual minimum.

Back in June, the UK Met Office were forecasting 4.1.
Given more multi year ice this year, and below average temps, I go for slightly higher than last yr, around 5.4

Walter Allensworth

A couple of tenths below the 2000’s average is my call – about 5.2 mk squares

CC Squid

I recommend the following article at “… The graph of the Five year moving average of the Arctic Sea Ice Index [red line] is inversely correlated to the Atlantic Multidecadal Oscillation [AMO] five year moving average shown in green….
Arctic sea ice could continue to recover for next 30+ years of negative AMO
The natural ~60-90 year Atlantic Multidecadal Oscillation [AMO] has been in its positive warming phase since 1976 and after ~30+ years of warming is “pausing” and transitioning to its ~30-45 year negative phase [cooling]:…”


I remember earlier this year a study that said they could predict the final ice total based on the number of melt ponds by a certain date (I think it was July). I never heard anything after and wonder if they made a prediction or not and, if so, how accurate it is.



See - owe to Rich

I’m sticking with 5.2 just like I said back in June when the WUWT consensus was 6.0 IIRC.
BTW JAXA have a data file at , which used to be up to date, but it does not have 2014 data. Does anyone know if their 2014 data are available to the public?

Bill Illis

The NSIDC September average is tracking to 5.4M km2. [NSIDC does some post-processing to this number which adjusts it by +/- 0.1M].

See - owe to Rich

OK, I can answer my own question: .


I doubt that it gets below 5 million square km, but it’s always possible since the wind and currents will determine the final number. The multi-year ice growth looks good though, and if it continues for another year or two there should be steady recovery. It’ll be interesting to see how the alarmists explain a pause in their only tangible indicator that the Earth is warming.
Does anyone else imagine Hannibal Lecter calling NSIDC and asking “Well Serreze, has the ice stopped screaming?”


in Figure 1. there exists a clear prediction breakpoint between 5.1 and 5.4. I say 5.3 million square km (5.25 rounded up).

Accuweather September forecast for Arctic Bay, Canada, shows the average temp drops to minus 1 degree on the 7th, and steadily drops for the rest of the month. Does anyone know how this compares to other years?
For that matter, does anyone know of research comparing Arctic ice melt to temperature?

Eustace Cranch

It’s not the coverage area, it’s the thickness!
*ducks and runs*

bit chilly

5.5 million km2


A 97% consensus of 2 people predict there will be ice in my G & T.


I’ll keep guessing 6 million and some day I’ll be right – especially once the AMO heads south.

Ulric Lyons

I think that the melt rate is just about to speed up for a week or two because of negative NAO conditions, then slowing down late August and through September, giving a minimum of around 4.9 sq km.

Alec aka Daffy Duck

‘Cam on’ want to know the melt pond prediction:
‘prediction that the September 2014 sea ice minimum will be 5.4
million km2, close to the level in 2013. The figure is likely to fall
within the range 4.9-5.9 million km2.’
Read more at:
An awfully big range,

Kelvin Vaughan



The Arctic seems to be i a thickening mode in the last few years, in most part due to the low summer temps above 80 N. This growth in multiyear ice means the ice cap has a thicker edge than previous years, combined with another low temperature summer ice melt has slowed. So my prediction for Sept average will be a small increase on last year 5.5-5.6.

Ron Clutz

There’s another way to look at this. In recent years, MASIE ice extent goes down 2 to 3 M sqKms in August, and loses another 0.5 to 1.5 M sqKms in September.
For 2014 to achieve an average melt season, the minimum would occur on day 255, and in the 5 weeks between now and then, the losses would need to average 2.2% of max extent each week–in other words this year would see from now on a weekly melt of 0.34 M sqKms, or a total additional loss of 1.7 M.
That scenario would result in a 2014 minimum of 5.1 M sqKms. That extent would match last year; anything higher would signify a rising trend.

The CBC’s prediction 2 years ago was 0 (zero) or up to 1 (one) depending on the definition of “ice free”. They have been rather quiet as we approach the date of this momentous minimum ice extent.
As an aside, British Columbia no longer has “forest fires”. I leave it as an exercise for the reader to figure out why 😉

I’ll say 5.2 and I say that because the chicken entrails told me so.


The summer temperatures 80° and above have been a little cool, but the winter was much warmer than normal. A uniformed conclusion from that would be that the ice would have gotten a little less thick than normal. On a side note: does anyone know of a location of the DMI data in a tabular form? Looking at their site they have all the graphs going back many years, but I only found links to the non-modled data, which I am not up to converting to average temperature.


Are we talking absolute extent minimum? Or minimum with 15% or greater ice, as per the above graph?


New AMO (#wxAMO on twitter) data is out today for July and shows a jump to .244 My April 1 call for Arctic Sea Ice is below 2013 and sticking with that since the AMO is still rising. I may bust on my 2015 and 2016 forecast unless the AMO heads back to neutral or Negative.
AMO Data 2009 to July 2014
2009 -0.037 -0.143 -0.139 -0.109 -0.040 0.143 0.249 0.174 0.079 0.186 0.091 0.104
2010 0.061 0.199 0.309 0.447 0.481 0.469 0.472 0.547 0.471 0.346 0.257 0.229
2011 0.165 0.130 0.077 0.114 0.174 0.200 0.113 0.169 0.166 0.085 -0.049 -0.023
2012 -0.043 0.026 0.048 0.102 0.186 0.322 0.397 0.453 0.470 0.351 0.187 0.163
2013 0.150 0.138 0.180 0.159 0.123 0.068 0.212 0.216 0.277 0.369 0.150 0.059
2014 -0.040 -0.020 -0.059 -0.072 0.021 0.084 0.244
Speaking of the AMO new paper out today says the AMO controls Arctic Sea Ice and now Green house gasses.
Who would of thought that.

Peter Sable

Do we get to vote on the moving Gaussian mean (or Hamming, or Hanning) looking back 30 days as of Sept 30?
[ this is signal processing sarcasm humor in case you missed it ]


Any chance we could just use the median guess as our submission? I believe that statistic would be least biased by outlier guesses (and certainly better than an ad-hoc method like the average of the top 5 vote getters).


I’ll be conservative and guess 5.1 and if it’s more than that I’ll be happy.


Guinea Hen entrails tell me 6.0

Bill Illis

njsnowfan says:
August 11, 2014 at 3:26 pm
New AMO (#wxAMO on twitter) data is out today for July and shows a jump to .244 ..
There is a huge problem in various temperature series for ocean SSTs in the current methodology being used (including the AMO index).
The scientists running the systems have left a huge amount of seasonality in the anomaly data. Not so much for the global ocean but for smaller regional areas.
This is more than a rookie mistake. Something more than a mistake is involved here since it is extremely easy and straightforward to correct for seasonality but they f it up.
For example, here is the weekly SST anomalies for the AMO region in a scatter chart by week covering 2004 to 2014. The AMO index always goes up in July to September now. I mean really. And it always goes down in the January to May period. There is +/-0.2C of seasonality left in the numbers.
I’ve noted this before. Someone has to fix it. The monthly AMO index is useless now.


Chilean earthquake triggered icequakes in Antarctica

Pamela Gray

I studied my crow’s feet and they are growing, so I pegged it at 5.8.


The representative statistic should be median.
I again am not able to enter zero.
Laugh all you want. We will see who is laughing when I kayak to the North Pole this September. The real North Pole. Things have been really bad up there. The polar bears are few, and no one has seen a penguin colony in forever.
The only ice up there at the end of the melt will be in my highball glass as I celebrate.

Ah, Houston, Arctic Ice 14, we have a problem here.
All of the discussions about previous agency predictions and all of the personal predictions above – valid, invalid, or correct or incorrect – have been for Arctic 2014 sea ice minimum amounts. (Granted, we need to clarify if it is sea ice area, sea ice extents, corrections, and all the other …
Further, our most gracious host here states he will take our answers, average them (again, a prudent skeptical question was raised about how that average will be made) …
BUT. The submission table above does not ask that question.
Rather, the highlighted submissions table clearly asks “What will the September monthly average Arctic sea ice extent be?” A different question altogether!
Thus, I strongly urge the question be revised to become: “What will the September minimum Arctic sea ice extent be this year?


TheLastDemocrat says:
August 11, 2014 at 8:45 pm
Dear LastDemocrat, please I’d be happy to loan you my kayak, it’s sea kayak, with a foot steering ruddr and sealed compartments. Just what you’ll need for your journey to the North Pole.
P.S. Is your wife cute?

Bill H

5.9 was my guess..
Reasoning is the air temp above 80 lat in most areas is 4-6 deg f below freezing already only areas close to 80 lat are bordering the 0 C range. The ocean currents are now cold and no major storms to break up ice are on the horizon. Not seeing much to disrupt the Ice formed already.

Mr Green Genes

John F. Hultquist says:
August 11, 2014 at 9:37 am
NOTE: Not the minimum, the average.
average = mean ??
The mean is biased by the outliers that, in this case, will be the minimum. Thus, go low.

A self-fulfilling prophesy. I looked at the numbers and went for 5.5.

An interesting image you and your readers may wish to peruse before casting your final vote?