Crowd-Source Prediction of Mean September Sea Ice Extent (July update)

Each June, July and August, the SEARCH Sea Ice Outlook (SIO) collects predictions for the mean extent of Arctic sea ice in September. These predictions come mainly from scientists but also some other people, drawing on a variety of modeling, statistical or subjective methods that each contributor describes in written documentation. The first round of SIO predictions, submitted by June 7, have just been published online. Twenty-three contributors suggest values ranging from 3.4 to 6 million km², with a median of 4.1.

In a post last month on this blog, I invited readers to submit their own predictions. By my count a total of 116 predictions were posted by the June 7 deadline, along with some interesting discussion. My original post had an ambiguous title, leading some people to offer predictions for the one-day minimum instead of the September mean, so I view this first round as an experiment. But from here on we’ll look at how that experiment came out.

 Figure 1

The ASI blog estimates ranged from zero to 6.6 million km², with a median of 3.0. The interquartile range (IQR), spanning the middle 50% of predictions, goes from 2.5 to 3.75. Figure 1 graphs the ASI median and IQR in context with observed September ice extent from 1979 to 2012. Collectively, ASI blog participants expect a substantial decline past last year’s record-breaking 3.61 million km².

Most ASI predictions were posted here in early May, within a few days after the original post. They consequently are not informed by developments including a slow start to the melt season that became evident later that month and early June. Early submissions might be a second factor that biased predictions toward low values. Figure 2 graphs the 116 individual predictions by date submitted, with a lowess regression curve confirming that the earliest predictions tend to be lower.

Figure 2

The day/month ambiguity and early predictions make the first iteration experimental, but in that spirit we can still compare our crowd-source results with those of the SEARCH SIO. Figure 3 visualizes both distributions as box plots, with boxes showing the median and IQR of each group. Outliers, which are predictions more than 1.5IQR beyond the first or third quartile, are marked individually in these plots. The ASI distribution is both lower and more dispersed, as I would expect. Continuing this experiment, it will be interesting to follow both distributions over the July and August iterations, and compare their evolution with the real September ice extent.

Figure 3

 Some of the ASI predictions reflect detailed analysis, and some are admittedly just guesses. By the same token, SIO contributions vary substantially in their basis and the expertise of their authors. The SIO distinguishes some contributors as “public” (in contrast to scientists), but if we set aside the 5 public contributions to the June SIO, the median stays the same. Resistance to outliers or distribution tails is one attraction of using medians as simple summaries for these data.

SIO also distinguishes between three broad types of methods: modeling, statistical or heuristic. Figure 4 breaks down the June SIO predictions by method. From this small sample it seems that statistical approaches give more pessimistic and also more consistent results. Are statistical predictions in general likely to be more realistic, or less so? That evaluation, and how our ASI crowd source predictions fare, should be topics for post-season discussion.

Figure 4

The next SIO deadline is early July. You are all invited to submit, as comments to this post, your best guess for the mean September extent of Arctic sea ice (NSIDC). Your numerical prediction should be in the first line of the comment, followed by at least a sentence or two explaining the basis for your prediction — whether pure intuition, elaborate calculations, whatever you’ve got. Predictions are not bound by what you did or did not guess last month. As before, general discussion is welcome too.

Results from this crowd-source experiment will be research data in their own right. Individuals can submit their own predictions directly to SEARCH, of course. The analysis here will focus on collective skill and uncertainty rather than individual. I will summarize the results after each cycle, possibly forming the basis for a research paper as well as future blog posts. The whole process should be completely transparent, because the raw data — predictions submitted here as comments — remain public and accessible to anyone.

Comments

[Neven]

Great post, Larry, thanks a lot. I was just a tad later with the SEARCH 2013 Sea Ice Outlook: June report blog post, but I'm keeping this one at the top of the main page.

[Paddy]

4.2 million square kilometres.

This is an adjustment upwards of 0.5 million square kilometres from my prior prediction of 3.7 million square kilometres, based on the slow start to the melt season. However, it would still be the second lowest on record...

[jdallen_wa]

3.25 million square KM.

Higher side of my previous prediction. Mostly Heuristic, but founded on my understanding of trends and the rather unpleasant condition of this years ice.

[Kevin McKinney]

3.25 If memory serves me correctly, that was my previous guesstimate. Repeating it reflects my assessment that the slow start will not strongly affect the September outcome.

The original rationale was a guesstimate that 2012 did indeed represent a regime change, and that the process of ice loss would end up taking us 10% below the 2012 value.

[Greg Wellman]

3.8 million sq km
We don't generally get back-to-back records and this year is starting slow, but PIOMAS has the total volume lower than last year, so we're not talking "recovery".

[dominik lenné]

5.0 e6 km²

PIOMAS volume looks a little bit better than 2011. Practically all SIE curves look a little bit better than 2011. Supposing continuous weather change tending to more melting - which I did not check - I just take the 2011 value. Jaxa leads to 4.6 or so, dmi.dk yields 5.0 or so, norsex 5.4 or so, so I jump right in the middle and take five.

[Chris Biscan]

So this wasn't published to SIO?

This first month might bust worse than anything Watt's collective group put out in June.

Even before the slow start. The odds of smashing 2012 by this much were extremely next to nil.

On Americanwx I predicted a single day min of 4.0-4.25 km2.

I thought this site was strictly science. The numbers above look like the opposite of Watts.

Hopefully after this Summer an adjustment will be made.

If I am wrong because a 2007 like dipole takes over so be it.

If I am wrong because average weather happens from here on out compared to 2007-2012 mean then I will fully admit that I missed something.

But my feelings are that average weather will bring us to a 2007 single day min.

But it could easily be like 2011 or 2008.

[jdallen_wa]

@Chris Biscan - We're not intending to be the opposite of Watts - its our honest assessment of how we think conditions will play out over the next 3 1/2 months. We may be wrong; our numbers may be too low. However, that's not how some of us see it. We'd be dishonest if we stated otherwise.

We'll happily learn (and celebrate) if our pessimism is unfounded.

Let's see what the Ice does.

[Robertscribbler.wordpress.com]

2.8 million square kilometers.

There's a lot of ice at the edge that will melt as summer progresses (Kara, Hudson Bay, Baffin Bay, CAA, Beaufort).

The Laptev and Russian side of the CAB are taking a pounding in June, while the remaining thick ice is shoved along toward the Fram by PAC 2013. So, in my view, this CAB thinning will make it far more vulnerable to melt come end summer.

[Robertscribbler.wordpress.com]

Oh, and one more thing. North Pole Cam 1 shows open water now.

[Paul Beckwith]

Goose-egg, nadda, finito, zero for my prediction. I just blogged about it a few days ago: http://www.sierraclub.ca/en/AdultDiscussionPlease

Based on overall system behavior (both Arctic + global systems), and the numerous feedbacks and unusual recent behavior. Feedbacks such as spring snow cover collapse, sea ice fracturing + thinning + increasing melt interface area, + fragmentation, algae + melt water pool darkening, increase in cyclone frequency, duration, amplitude behavior, methane bubbling degrading ice + extremely high methane levels in Arctic atmosphere and localized warming, higher water temperature, increased deep water upwelling behavior, increased wave action on ice rheology, rapidity of response of system in paleorecords, polar vortex fracturing, jet stream breakup, + extent of extreme weather whiplashing/weirding around globe, change in cloud behavior, average height decrease, change in mechanical breakup, increased ice advection into Atlantic, less buttressing, higher rate of ocean heat transport into Arctic basin, cyclonic storm surge causing increased mixing of basins, and many others...

Mostly the cyclones...+ recent unprecedented observations a) March fracturing, b) cyclone thinning MYI + north pole ice, c) persistence of high sea ice area

[Jai Mitchell]

My guesses

May - 4.3
June - 2.8

[Jai Mitchell]

2.8 million square kilometers

Reduced from 4.3 in June, I thought that there would be significant lagtime in the warming. Now I see that there are center-pack losses that will effectively double the melt rate once Siberian land temperatures melt surface sea ice.

[Kevin O'Neill]

2.9 Mkm^2

This is the same value I posted in May. Based on PIOMAS volume trends along with northern hemisphere snowcover trends and their correlation with SIE. Estimate is biased a little low on the guess that NH snow cover will plummet in June and July - and those two months have the highest correlation to September SIE.

[Chris Biscan]

You guys know you're talking about September monthly extent?

Not single day?

Do you know how hard favorable the pattern has to be in middle to late September to keep the ice from exploding in growth?

There is no signal any ice has been melted from that cyclone.

There is also a think called solar insolation that is about to peak in a week and start dropping soon after.

Very little time does the central arctic towards the Greenland side get enough sunny days for any substantial melt.

All of the factors you guy talk about are not new.
Except this year the Kara and Beaufort have ice and higher albedo.

Precious energy needed to melt more ice.

Weather is a huge driver in this. It's wasted weeks of important melt time.

Solar insolation can't be made up.

[Robertscribbler.wordpress.com]

That storm's been boring a hole through the CAB all month long. The thinning is in all the concentration monitors and CICE/HYCOM is showing thickness falling quite rapidly. That CAB ice looks worse than it ever has in June. And a storm thinning the central ice so much in June is quite unprecedented.

[sofouuk]

3.5 Mkm^2

approximate extrapolation of the downward trend, nudged upwards due to the slow start.

@Chris Biscan: please just let people post their guesses. the last thing this thread needs is off-topic debate

[Paul Beckwith]

@Chris Biscan I cannot go any lower than nil.

[Chris Biscan]

@Robertscribbler.wordpress.com:

1. It is impossible to tell how thick ice is on concentration charts. Just because ice is broken into smaller floes by the wind doesn't mean it melted. The buoys say no melt. And they said lots of snow fell over those areas. Which is a negative feedback. These are indisputable facts. It may or may not have a huge effect on melting later. But we know it didn't melt the ice so far.

2. Hycom is worthless. I have no idea why people here use it. It horribly over-estimates the ice thickness. There are published papers stating this. I see many people using it to say the ice has melted there knowing it's worthless.

3. The ice where there is snow on top of it still hasn't melted at all. I'm sorry.

@sofouuk

Any debate that will prevent this blog from going down the Watt's et al route is important.

[Hans Gunnstaddar]

4.35 which is precisely identical to my first prediction for the exact same reasons, which are in the brief history available there have not been more than three consecutive years of ice extent minimum reduction, and no two consecutive years of minimum ice extent records set. The rest is dead reckoning as a rebound from 2012.

Although standing by my prediction above, I am keeping a watchful eye out for the idea put forth by some that due to so much 1st year ice in such a slushy condition, there could be a flush out the Fram Strait the likes of which we haven't experienced before and if so, hope it would shock the rest of humanity into coherence.

[sofouuk]

@Chris Biscan - don't be ridiculous, and there is no need to question people's intelligence. we know there is also a think called solar insolation. you do not have a unique understanding of what is going on. any prediction that is approximately in line with the decade-long downward trend is essentially as good as any other, because we don't know what the ultimate effects of the cyclone will turn out to be, or, as you admit yourself, what the weather will do for the rest of the summer. there is no point pretending otherwise

[Dromicosuchus]

I can't fathom how we could possibly get a monthly average extent of zero for September (particularly since by that point both melt and freezing are going on at different points in the Arctic, and if everything's melted in the hot bits...well, there's nowhere to go but up), but that's your business, Paul, not mine, I guess.

Anyway. I'm kind of flummoxed, frankly. On one hand, so far there's little of the massive fragmenting on the perimeter that led to the spectacular Pacific-side collapse we saw in 2012, and I'm just not really ready to believe that ALL of that ice could melt out again so perfectly this year. On the other hand, that recent cyclone has bashed a huge swath of ice in the central Arctic to bits. Melt will proceed more slowly there than it would have if the fragmentation had occurred perimeter, but it will happen. I'm just not sure whether the slow speed of melt there will be overwhelmed by the combination of melt from both the inside and the outside, or vice versa. There's also the matter of whether the March fragmentation will have a significant effect, leading to a repeat of 2012's Pacific collapse. Goodness knows I've no answers to that.

I suppose my guess is for an average extent below 2007's (thanks to the greatly-reduced volume and central cavity) and possibly below 2012, but if so not too far below. Probably a wee bit above. Shall we say...3.8 million km^2? I don't recall what my previous prediction was, but it was definitely lower than that, as I was expecting a faster start to the melt season and a May drop in volume that never happened. Mind, I would also expect an absolutely bizarre distribution of ice in the final reckoning, with a huge bite taken out of the Atlantic side, battered but intact ice above Greenland and the Canadian archipelago, and fragmented, thin, but still unmelted ice over towards the Pacific side.

[Chris Biscan]

Looking at these predictions I am clearly not being ridiculous.

So people know that solar insolation has been way low this year vs the 2007-2012 period so far.

They know more snow exists on the ice than in those years so far.

They know a big contributor to the epic melts was heat in the water. Do you know how far behind this year is? Yeah, soooooo far!

2013 had higher volume as of the last piomas update over 2011 and 2012. It's all but certain the next one will be double that in terms of difference or even higher.

It's taken a JJA NAO pattern like we have never seen back to at least 1895 to make this happen.

We can attribute some of it to feedback's but we don't know how much.

but we know a positive or neutral NAO = more ice.

p://www.cpc.ncep.noaa.gov/products/precip/CWlink/pna/nao.sprd2.gif

It's June 15th. Basing the sea ice min on the most recent year's is over. We are a quarter of the way into the melt season. And it's been a complete 180 of recent years. if 2012 started like up to this point and still had the same weather after wards it would of came no where close to the record low it got to.

I don't have to pretend otherwise I can chop off a quarter of the melt season and know that a whole lot of melt now has to be crammed into a smaller time frame that will have to be like 2007 to get back to 2012 or below.

There isn't a magic energy tree coming to shake some energy here and there to play catchup.

[Paul Beckwith]

@Sam Yates Re: zero ice for September;
2 possibilities:

a) "The scars are still visible from my double lobotomy", as a commenter said on steven goddard's site Real Science.

b) The Arctic ocean is blue at the end of August or earlier. The stormy cyclonic driven surface water has lots of mixing with deeper saltier warmer water and inflow from the Atlantic/Pacific and the cyclonic driven extreme wave action delays any September freeze-up to October.

[Neven]

Right now I'm thinking of something right between 2007 and 2012: 3.95 million km2 for the September mean sea ice extent.

But it's very difficult. On the one hand I agree with Chris Biscan that precious melting time has been wasted. A good start is half the work, as the Dutch say.

On the other hand it's just mid-June and considering the awful weather for ice melting so far, 2013 isn't doing all that bad. And everything that happens in the Arctic, always has two faces. That persistent cyclone might not have caused much of a direct extent decrease, but it definitely has had an unprecedented effect on the ice pack, from the Central Arctic to the Siberian coast, that could make a difference in August.

I've been doing a bit of research yesterday, looking at previous blog posts from 2010 and last year, and one stroke of weather that is good for melting, will put 2013 firmly in second spot, I believe. But it will take a long period of that to catch 2012.

I might go for a lower number next month, but right now I'm sticking to 3.95 million km2.

September minimums for the last 7 years (million km2):

2005: 5.57
2006: 5.92
2007: 4.30
2008: 4.73
2009: 5.39
2010: 4.93
2011: 4.63
2012: 3.61

[Kevin O'Neill]

Chris B -- most volume is lost by bottom melt - not top melt. The ratio is basically 2:1.

From Steele et al, Mechanisms of summertime upper Arctic Ocean warming and the effect on sea ice melt:

Our analysis shows that top melt dominates total melt early in the summer, while bottom melt (and in particular, bottom melt due to ocean heat transport) dominates later in the summer as atmospheric heating declines. Bottom melt rates in summer 2007 were 34% higher relative to the previous 7 year average. The modeled partition of top versus bottom melt closely matches observed melt rates obtained by a drifting buoy. Bottom melting contributes about 2/3 of total volume melt but is geographically confined to the Marginal Ice Zone, while top melting contributes a lesser 1/3 of volume melt but occurs over a much broader area of the ice pack.

Typically bottom melt is restricted to the marginal ice zones - but guess what? The CAB now looks like a marginal ice zone. The cyclone has dramatically fractured and opened up the pack ice.

It's easy to fall into a trap thinking sunlight hitting the ice is responsible for most of the melt - but this notion is false. Ocean water does most of the melting. Summer insolation is more effective when there is open water near ice. Water can hold far more energy than air.

BTW, sea ice can definitely melt when it has snow cover. In fact, during certain times of the year snow cover is essential to melting; it acts as an insulator keeping cold air temps away from the ice and allows warmer ocean water to eat away at the bottom. Ocean heat melts the ice and it doesn't lose its heat simply because snow fell on the ice.

[jdallen_wa]

@Chris Biscan

No, you aren't being ridiculous.

Your arguments about energy input would be more persuasive except for three things that nag me.

1) The incredibly mangled state of the ice, as illustrated so marvelously by werther.

2) Its significantly reduced volume. You also pointed out HYCOM tends to *overestimate* ice thickness; considering how thin it shows most of the extent to be, in view of what we see in MODIS, the ice is leaps and bounds more vulnerable than it has been at any time recently observed.

3) Total system enthalpy. For sure, insolation north of 70 is key, but it has lots of help now it never had previously. The total heat in the ocean, and inflows of heat from continental river flow, the North Atlantic Drift, and the amazing spikes in circum-polar temperatures recently have been continuously upping the total energy in the system, *even with* over-winter losses.

So, in short, our "Heuristic" estimates do have substance, and are just as rational as yours, even if not quantified. It will be a matter of seeing how the forces play out over the next 12 weeks.

[wanderer]

I will go with 3.0 million square km.

In my opinion we have seen the weak ice conditions with the fragmentation event and early area loss in April, but cold weather "used" these events in favor of the ice, so there was a small recovery and the actual melt was delayed a bit.
So overall, we have now a little more ice than in the years before, but the ice is (in my opinion) in a weaker condition.
Yes, melt was ahead in several regions in 2012, but when you look at the ice thickness and make a day-to-day comparison, you see that we had more MYI in 2012 and we certainly didn't have (and never had before) such a little thickness so close to the pole).
Actually, now that I look at it, it's just Beaufort and Kara where we had significantly more ice last year at this day.
So if weather doesn't stay that unusually cold (i think it was the coldest melt season start ever on record), we will catch up to 2012 and probably have a new record.

[wanderer]

PS: @ solar insolation.
Yes, extent is higher, but not that much, and when you look closely on MODIS, worlview, you see lots of little gaps or tiny holes, that probably won't all be recorded in the data...

[Glenn Tamblyn]

I'm going for 2.75

Reasons.
Even with any question marks over HYCOM, any errors they have must at least be reasonably consistent. And they have shown a massive rate of change in the CAB. This at a time when there was lots of cloud and cold weather. Ergo, most of the melt has been from the bottom and fragmentation of the ice, degrading of the fresh water lens etc.

Most of the peripheral basins will go, even if they have started slowly.

Now basic thermodynamics says that whole eastern hemisphere region will melt out earlier than last year. And not all of that is extremely high latitudes so still getting sunlight and exposure to air masses from Siberia/Canada/Alaska.

So warmer water to generate weather systems.

Complete melt of the Eastern hemisphere right to the pole is plausible and perhaps earlier than we might think. Or that there are large swathes of open water within the eastern hemisphere. All this allows weather/winds/waves to build up a head of steam before slamming into the bastion above Greenland/Ellesmere which must still be weakened due to the earlier cracking events.

So larger transits out through the Fram/Nares, into the Beaufort, or even simply breaking the bastion up into separated chunks. All because the ice is more easily mobile.

Imagine a storm like last year by but this year it doesn't hit ice till the other side of the pole.

[DavidR]

WAG 2.75M sq km
- PIOMAS decline currently averaging 800 CuKm per year, this is primarily due to water rather than air temperature. Water temperature is rising.
- Thickness is low so more area is likely to melt.
- Broken ice means more ice surface area in contact with water so higher melt rates.
- this also allows more heating of water between smaller floes compounding the effect.
Alternatively we may see higher extent but lower area as less contiguous ice remains.

[Sam]

3.0 million sq. km +0.5/-1.0

I am revising my average slightly upward based on the late cold, and broadening the downside based on the atrocious condition of the ice in general.

Though it seems unlikely, should we get a strong anticyclone in August, we could see the remaining multi year ice flushed into the Atlantic and near zero ice conditions.

[Neven]

Just to be clear: Comments are for predictions only (with a bit of explanation).

For further debate of the methods/data/observations various people use for their prediction, and their criticism of those of others (which I find mightily interesting and stimulating, so please continue), I urge you to use the latest SEARCH post. Or one of the ASI updates.

[Neven]

Oh, and the Forum, of course.

[Kate]

2.75 or 2.5 depending on cloud cover and storm activity. The ice is mush, there is a lot of melt around the arctic to add warm water to the basin, the storm has done damage and I think there are more storms to come.

It is early to see such damage and where that damage is located. I've watched the Fram strait ice on Cryosphere for years. I've never seen it so thin and washed out.

Greenland melt is high for this time of year - imagine what it will do up north to the last shreds of MYI!

[Climate Changes]

I'll stick to the figure I posted on the first crowd-source prediction post last month:P
2.82 MKm2 and NP ice free.

[Fufufunknknk]

2.2

I wish the poll was about the lowest 30 day moving period than a slightly arbitrary period like September.

I hear what Chris is saying but I think it is like Moore's law. Remember back in 1998 when people had proved that Moore's law would end because of physics? I think he is right about a lot of what he says but I think the mechanisms have changed.

For example, I think the stormy conditions are more permanent, a la the Allmans:

They call it Stormy June
But July's just as bad.
Lord, and August's worse
And September's all so sad.

[Climate Changes]

2.82 MKm2 and NP ice free.

Sorry I've just realised figure goes first :\ Reason, based on observation of this years early, widespread break up of the ice followed by the ongoing ice churning cyclon/s. I second Kevin O'neill's words too but if anything may make me change my mind it will be the behaviour of ENSO in the next 2 months.

[MaGa]

2.8 Mkm^2,

slightly less than last time. The ice is in such a bad condition from Svalbard to the pole over to the New Siberian Islands that it will melt quite early. From the pacific side, there is so much heat input melting the thin first year ice. And some of the last thick ice is lost through Fram straight. Together that makes the rest of the old ice very vulnerable at the end of the melting season. I expect a rapid drop!

[Phil263]

I am going for 5.5- 5.7 M sqkm (September average).

My estimate is based on both statistics and "hunch".
Statistics: we are now in mid -June, i.e. two months into the melting season and this year is looking more and more like 2009 with an SIE just below 2009 (IJIS). SIA anomaly is currently at -0.720 compared to -0.511 at the same date in 2009, this is close enough and is a much smaller anomaly than for any year since 2005 except 2009. Let's not forget however that 2009 had a very fast melt in July and nothing presages that we will have a similar thing happening in 2013.
Hunch: This year has been colder than usual since March in the Northern Hemisphere and there is no reason to expect that general weather pastterns will change before Autumn when the minimum has passed.

I do not believe that this higher minimum ( if it does eventuate)should be interpreted as a Northern SI recovery , just a bounce back from an abnormally low level last year: a normal short term fluctuation due to weather patterns. If 5.7 million is a high minimum year now, we should compare with high minimum years in the 1990s which would have been around 7.5 million. In this respect, we are still on a downward trend.

[Kate]

@Phil263
I don't think land changes are the biggest factors here, not in this scenario. I may be a novice and inexperienced but I can see a trend and it's not up

[GeoffBeacon]

2.2 million square kilometres as before

This blog tells me of a slower start but

1. Ice scattered by the storm into places where it will melt.

2. The posts here have taught me that cloudy weather can cause more melting than sunny weather.

3. All my previous reasons.

Also one irrational reason. The SEARCH people are professional climate scientists.
Given the previous track records of climate scientists, it's got to be quite a bit worse.

(But I would have more confidence in SEARCH than the UK Met Office, who are just beginning to panic about the effect of the Arctic melt on UK weather. I must find out if they have asked Jennifer Francis.)

[Ac A]

I stick with 3.8 mil sq km, just as before!

Alex

[L. Hamilton]

Kate and Phil263, can you give me a point estimate to go with your uncertainty ranges? I need point estimates for the basic stats.

Regarding discussion vs. just-prediction posts, discussion is welcome too, explaining why you think your number is reasonable. This thing will be most interesting, though, if everyone's predictions reflect what they seriously believe is going to happen with the ice, not shifted by tactical thinking.

[anthropocene]

3.1 (same as last month)

Hunch rather than any mathematical process.
Expecting 'dead cat bounce' because of change of state of ice. expecting type of weather to make little difference - therefore to be consistent have to stay with original number even if this summer is going to have cool weather. I don't believe the start to melting has been that slow - it's just the numbers used to measure ice melt aren't showing it (yet) e.g. PIOMAS is a model - if the ice state is a new state then how can it model it accurately? However there is less confidence in the guesstimate than last month. It seems the melt is on a knife edge. We could be seeing some new (unexpected) negative feedbacks coming in due to the new state of the ice.

[Staxx]

2.0-2.5 Mkm2 Up from last time due to slow start,not as pessimistic as Paul although he is concise with the reasons with which I agree.

PS DNFTT

[Arcticicelost81]

2,0-2,5

In 2010 and 2011 had a quick start, but a slow finish.
In 2012, on the contrary - a slow start, but a quick finish.
2013 looks even more brutal than in 2012, which will cause a very fast finish in July and August 2013.

http://s017.radikal.ru/i402/1306/d7/50b8d70e84e6.jpg

From PIOMAS forecast that will be saved only ice off the coast of Canada and Greenland.

[Shared Humanity]

I am not going to make an estimate because I am a lurker and have no idea what the minimum will be. We do, however have a winner......

I wish the poll was about the lowest 30 day moving period than a slightly arbitrary period like September.

A 30 day minimum is a 30 day minimum. As a 35 year professional in manufacturing and very comfortable with charting processes, it is simply not correct to have the month of September as the period to track. I have seen charts on this blog that show much faster drops in SI in the later months of the calendar year. This would suggest that the minimum is moving towards October. This may be due to a lengthening of the melt season or a delay in the freeze but it is real and has been evident for many years.

I'll go back to lurking.

[Shared Humanity]

Also, I believe this lengthening of the melt season or delay of the freeze (or perhaps a combination) is one of the single biggest reasons for the deterioration of the remaining sea ice. There is simply less time during the freeze to strengthen the ice. This has the effect of weakening MYI and FYI. The fracturing that occurred during the winter, I believe, is evidence of this. This fracturing occurred in FYI and MYI.

[Shared Humanity]

One last comment.......

While I don't have a clue what the minimum will be, I am absolutely certain that it will fall within the group range. (somewhere between nada and 5.9 M sqkm) I do believe the estimates have such a wide spread because the Arctic is becoming increasingly chaotic and unpredictable. This wide spread has nothing or little to do with ideology or stupidity (my own excepted) but has everything to do with the fact we are in uncharted territory and there are far more variables to consider. Two decades ago, no one talked about persistent cyclones because these simply had no bearing on the way the ice behaved. This is only one of the new variables that everyone is trying to get a handle on.

[Shared Humanity]

One last, last comment. This time I mean it.

Everyone, keep up the good work.

Each of us brings a unique perspective (a subjective understanding of the thing studied) to the discussion. It cannot be otherwise as our perspective is influenced by the unique knowledge we have accumulated in our lives. As we share our perspectives, we are able to develop a deeper understanding of the Arctic. This is true because all learning is dialogue. If you want to understand this, read works by Lev Vygotsky, a brilliant Russian psychologist who transformed our understanding of learning.

This magnificent website has allowed this dialogue to occur and is, I believe, the reason it is the absolute best place to develop an understanding of what is happening.

Thank you Neven.

[Shared Humanity]

Oh, Hell. I've done it again. This is because I am in an area where I feel competent.

If you want to know how you can determine if real dialogue and learning is occurring, look for evidence of a contentious debate. It is this sparring of unique perspectives that is the single most prominent feature of learning in groups. The key is to not allow this debate to become personal or learning will cease.

We are not wattsupwiththat. Take a gander at the discussions that occur there. I have. It is one commenter after another supporting the genius of the other.

[Mignonette]

4.4 million km^2.

I think the issues raised by Chris Biscan cannot be dismissed lightly. The cold and cloudy weather in June could influence the sea surface temperatures in the late season. We may end up with a lot of thin and fragmented ice which did not have enough time to melt.

It's human nature to overcompensate. The estimates for 2012 were too high, but that does not mean we should assume the worst case scenario for 2013.

That said, I fully acknowledge that we are in a new Arctic regime - Fram export is the great unknown.

[Henry1]

I'll go 4.5 million sq km.

It's true that the water melts the ice much more than the air. But we need sunlight to heat that water and it hasn't been nearly as abundant as recent years in this pattern so far. Additionally, we need to melt the ice first to let that sun heat up the water which is taking a long time in the Beaufort, Chuchki, and East Siberian Seas.

It should go fast once those areas start to melt out, but my guess is we run out of time to end up like last year.

[Arjen Costerus]

Long time lurker here out of hiding. Enjoying the blog since my first trip to Baydaratskaya Guba 3 years ago.

My point: A number of people here are going on about the fact that the temperatures are below average at the moment and they give me the idea that this means that there is not a lot of energy up there to get rid of the ice.

But I see a tremendous amount of energy up in the Arctic. Not in heat, but in mechanical energy that is shearing the ice in ever smaller pieces. The ice is highly mobile this year, so all these pieces must be grinding along each other and breaking up in smaller pieces all the time. The cyclone may not provide a heat budget to melt the ice, but that does not automatically mean that it does not give energy to the ice pack. Just not the energy that we normally associate with the melt season.

[OldLeatherneck]

2.9 - 3.1

I still believe that we will break the record of 2012, although not by as much as I originally thought (2.85) or that others seem so certain of.

Using the current SIE as a means of projecting the final numbers, in September, is a very risky proposition this year. The cyclone is acting like a blender in a bowl of crushed ice, spreading crunchy remnants across the entire perimeter.

While much of the MYI in the CAA will be reduced and flushed out the FRAM, I believe that it is resilient enough to hang around for a few more years, at the minimum.

Shared Humanity has a valid point when talking about using the average of the lowest 30 days as opposed to the September mean. However, if we as a community were to change our metrics at this time we will be lambasted by WUWT and his cohorts for changing metrics just to suit our "Hidden Agenda."

[Shared Humanity]

OldLeatherneck....

I refuse to take credit for the suggestion that we should be using the average of the lowest 30 days.

This was a suggestion made by Fufufunknknk.

[Doomcomessoon]

Not that much has changed since May so I think a "copy and paste" will do perfectly fine.

"IJIS minimum extent: 1,458 km^2.
CT minimum area: 0,968 km^2.
NSIDC september extent: 1,7 km^2.

The ice in the main pack will need a little bit of time to weaken in the spring and early summer, but once it gets going it will be really dramatic. The amounts of MYI are not impressing and will melt away surprisingly fast in August when everything else is gone."

Great work Larry, those statistics are quite telling.

[fryingpan136]

I am going to stick with 2.7 million km^2.

It seems low given current area, but I am swayed by the melting from within which should soon start. The HYCOM thickness models show many thin areas near the pole. This should split the ice cover and increase the ice/water boundary. Only the thickest ice that doesn't go thru the Fram should remain in September.

[r w Langford]

3.2 mKM2
Two unknowns that may be sleepers are the intensity and depth of the halocline after the stormy weather and the degree of warm water mixing that has resulted near the surface. These two parameters may elicit rapid melting throughout the fractured ice region which has been greatly extended. One other new factor is summer cyclone intensity. With a weaker jet stream the incidence of severe storms in the arctic may increase thus contributing to greater melt rates. My prediction is a guesstimate as adequate new data is not available for these parameters.

[Boa05att]

~3.2M km^2

(same as last time)

This is based on Wipneus' analysis of PIOMAS volume decline by grid cell, which still seems to me to be the best way to relate the exponential extrapolations of sea ice volume decline to declines in its extent.

[dominik lenné]

Question to Neven:
jaxa, dmi.dk and norsex deliver quite different values for the same year. Which will be the reference for our guesses in the end?

[Kate]

L. Hamilton
Not exactly sure what you mean, but if you want just one number then I'd say 2.75

It is only a hunch, from seeing so much data for a long time. Things up north have changed and I guess I'm thinking 'cliff' year is soon.

[NegativeNAO]

4.5

Graphs and reality are jagged and toothy. Trends hold, but there are always new teeth to come and bite us along the way. We'll give 2013 "new tooth" status and let it go at that.

[wayne]

We have unprecedented remote sensing coverage of Arctic Sea ice, amazing assets. I read often the basic assumptions about extent.
Which is there is more ice now than 2012, therefore a lesser melt is expected. This is based on simple straight forward projections, however the ice is anything but simple, anything but straight forward, sea ice is complicated. Even the data from mass buoys- as simple as presented- is more complex. I'd like to think that there are false echoes. http://imb.crrel.usace.army.mil/irid_data/2013E_thick.png
Something like fresher water mixed with ice confusing the bottom read. Extent may be presenting itself falsely as a cooler season as well, the relentless counterclockwise pushing aside from the usual clockwise rotation of the Beaufort gyre also affects extent greatly, equally confuses a comparison with 2012

http://iabp.apl.washington.edu/maps_daily_arcticbasin.html

what will likely happen is a blitz opening of ice when the gyre matches the circulation with a more regular anticyclone. This will
happen because the air will be warmer than the sea and ice .
Then again there are more complexities to consider, I applaud Whether's close up approach.

[Anthonywobrien]

3 +/- 1

I am so torn. Following the past numbers I would expect a dead cat bounce. Looking at the current numbers, I would expect a bounce.

However looking at peoples comments about the state of the ice a new record low looks likely.

I realy do not know, but much above 4 or less than 2 would be a surprise.

I doubt we will see zero this year, but it is not very far off. When it does happen I expect it will be earlier.

[Surveyork]

4.3 million km^2

Source: Intuition + Slow start of melting season.

[Talha Muhammad]

3.0 million km^2 +/-.2 million
I think the major determinant factor is volume and even though piomas is more than previous two years it is on the russian side and I don't think that will affect the final area/extent. And I think the ice is thin and spread enough that the area and extent will fall off a cliff.

[Dromicosuchus]

3.8 million km^2

Sorry, Neven, I skimmed through the article and missed your (bolded, no less!) injunction to include the prediction in the first line. I'm reposting this, then, with a link to my previous comment w/ rationale: http://neven1.typepad.com/blog/2013/06/crowd-source-prediction-of-mean-september-sea-ice-july-update.html?cid=6a0133f03a1e37970b0191035ca127970c#comment-6a0133f03a1e37970b0191035ca127970c

[Phil263]

Kate and Phil263, can you give me a point estimate to go with your uncertainty ranges? I need point estimates for the basic stats.

Hi Larry

Make it 5.6 m spkm September average then.
Cheers
Phil

[Patrice Monroe Pustavrh]

Still staying at 4.1 mio km2.
It is just a guesstimate.

[Neven]

Question to Neven:
jaxa, dmi.dk and norsex deliver quite different values for the same year. Which will be the reference for our guesses in the end?

"The next SIO deadline is early July. You are all invited to submit, as comments to this post, your best guess for the mean September extent of Arctic sea ice (NSIDC)."

[Account Deleted]

2.0 M km2. To melt something, grind it and then heat it, rather than heat it and then grind it. The second worked for last year, why should the first not work for this. And please, North Pole out of the ice, big news and big concern for people and governments.

[Sushi]

2.5 million square kilometres

Unchanged.
Heuristic.

My belief is that a slow start to the melt season will be of limited import due the fact that the ice age and thickness render it highly vulnerable to melt. Include cyclonic effects, open leads, and ice discharge via the Fram and I think we will see a significant abrupt reduction in extent as the season progresses.

My estimate is influenced by the rate of Greenland ice melt (data available here: http://nsidc.org/greenland-today/

This too had a slow start but is rapidly exceeding the 1981-2010 average melt extent. I suspect we will see the same phenomena with arctic sea ice extent.

The choice of sea ice extent is a poor metric. Should there be significant melt and floe breakup there is likely to be an increased distribution of the remaining ice across the arctic. There may be a significant reduction in ice volume but that remaining volume may be distributed in such a way that the extent metric delivers a misleading interpretation of the actual ice conditions.

[Werther]

I see no reason at all to change my own expectations on mean September SIE at 3,28 Mkm2.
See the Forum for my reasons if you like.
(See: Short to Medium Term Arctic Sea Ice Conditions Discussion)

[LRC]

Ice: 2.9
I think it could be even lower. The only thing that could really mess up the readings is the PAC 2013. It could create so much slush and spread it out over such a large area that we could end up with a major bump up. If we do get a major bump up this year then next year could see a very high cliff.
In my mind even volume needs to be relooked at because all the ice up there has very little strength left in it.

[Manolo El Lobo]

Ice about and around 3.0
This is a mighty game of guessing, so many variables and quite some unknowns. I like Kevin O'Neill comments about warm water and ocean circulation being decisive in these predictions. Seems we do know little about it. How much comes in ? How does it circulate ? How does it affect the frozen bottom of methane hydrates? We sure will find out soon enough...

[michael sweet]

I am going to raise my guess to 3.0 Mkm2.

The slow start to the season means less heat from the sun has been absorbed. The ice is thin and that will eventually mean less extent. I expect a big melt later in the season. The NOAA map shows some increased sea temperatures under the ice (and some cold areas in the Beaufort area). A lot depends on how much ice blows out the Fram Strait. Greenland is melting about the same as last year.

[dominik lenné]

@ Neven : Thanks for the extra work - could have looked it up myself.
---
"The next SIO deadline is early July. You are all invited to submit, as comments to this post, your best guess for the mean September extent of Arctic sea ice (NSIDC)."

[dominik lenné]

4.5 e6 km² - correctig myself as I got it wrong, which data source is the reference.

[wayne]

Sea Ice conditions in the Spitzpergen to Severmaya Zemlya axis is very precarious, has been all winter, also we note melt ponds everywhere circumpolar wise: http://earthdata.nasa.gov/labs/worldview/index.html?map=-958013.190435,-24076.349154,-548413.190435,180467.650846&products=baselayers,MODIS_Terra_CorrectedReflectance_TrueColor~overlays,arctic_coastlines_3413&time=2013-06-08&switch=arctic

of interest is the blue green pond zones coinciding with high pressure ridging, the air is warmer than the surface there, its even clearing at the Pole:

http://psc.apl.washington.edu/northpole/NPEO2013/WEBCAM1/ARCHIVE/npeo_cam1_20130616082146.jpg

note the ice crystal reflections, similar deposits happened further south at 75 N 3 weeks ago.

Cloud cover gives the impression of no melting, the overall cyclonic flow of sea ice over the Arctic Basin also misleads in suggesting cooler weather not affecting anything. But melting has something to do with thermal rays, ice density, and sea ice variances in thickness much different then say 10 years ago. There is nothing here that suggests not another record year for a melt, there is plenty of confusing signals suggesting otherwise. Must read though the fog of appearances.

Sorry Werther, mijn geheugen met naam spelling niet zo goed

[jsaenz]

It seems to me that the numbers in text " from 3.4 to 6 million km2, with a median of 4.1." do not match the errorbars in Figure 1. Is there any kind of unit conversions that explain the difference? Am I missing something? Thanks.

[JackTaylor]

3.30 km^2

Range +0.45 / -0.50

Based on color blind perception of history.
http://webpages.charter.net/swf-test/gimg/predictA13.png

[JackTaylor]

3.30 Mkm^2 = correcting myself.

[Llosmith57]

4.25 as of now looks likely. Ice conditions will rule in the end. Extent could easily be more or less depending on weather conditions between mid August and early September. A big cyclone combined with low Fram totals means high extent while a big high crunches the ice into a pile and lowers extent. We're starting behind and will stay behind for the balance of the melting season unless things change.

[Bert van den Berg]

4.2

Just looking at Figure 1, and expecting an uptick this year (~5 year cycle), upticks are about .5.

No change in my thinking since last month.

[Dan Green]

3.85 m km2 for NSIDC Sept avg, which was my guesstimate last month.

Peeking through the lance-modis clouds, the ice looks like crumble topping on the Eurasian side of longitude 0deg / 180deg. Not hard to imagine a load of it whizzing through Fram if atmospheric conditions stay as they are. On the other hand, if it stays cloudy and cool and less onsolation gets through...so a cagey midway number from me.

[Dan Green]

, insolation. I'll get my coat.

[Werther]

“Sorry Werther, mijn geheugen met naam spelling niet zo goed”

It doesn’t matter Wayne, as long as your message comes through…
(should see my grammar after having a good Chardonnay...)

[james cobban]

2.8 MKm^2
Same estimate as May, based on the weak and well-churned state of the ice.

[logicman]

2.5

But it could go as low as 2.0

I base that on my interpretations of the current state of the entire Arctic, as I just described in my blog, together with scientific reports of recent amplification factors.

Given extensive melting around the Arctic coasts and islands, the mainly 1st year ice in the central basin will be unimpeded in moving through Fram Strait and other warmer waters such as the Barents and Kara seas.

The melt is well under way and is headed for a cliff.

Around The Arctic June 2013

[Mtough]

2.8 for me as well, the same as I had for May.

Yes it's a slow start but the cliff approaches (if we haven't already fallen off the edge 0.825 today and going south) I still don't see too much upside. The cold should have made a huge difference but I don't think it has and there are polynyas popping up in some strange places.

The new paradigm is everything's new and what was a negative feedback ain't necessarily so.

Based (as always) on the wisdom of others and a little twist of lateral thought (not necessarily in a good way- but I try)

[Remko Kampen]

2.0 M.
I don't want to win my bet on WUWT just yet... (at 1.5M, and time to 2015).

[Fairfax Climate Watch]

0.3 square km for September mean.

It looks like there could be a giant hole right in the middle of the ice. If this happens, I would expect some very strange things to happen with Arctic ocean circulation!

There are still 2 full months of very high melt potential for whatever bizarre changes happen to take their toll.

Compare this year so far with the previous years for concentration and thickness:
http://www.fairfaxclimatewatch.com/blog/2013/06/arctic-sea-ice-crowdsource-projections.html

[Fairfax Climate Watch]

edit: [million] square km

[NeilT]

2.5 Mkm^2

Range - .5 Mkm^2 + 1.0 Mkm^2

I've updated my outlook mainly because NSIDC has a history of showing virtually invisible ice as solid extent. It's looking like we may get a lot of feathered edges around a badly melting pack. This will show as extent.

Area will be a very different story due to the fractured ice.

Insolation doesn't switch off after the solstice. It continues and some of the heaviest melting can often be seen in July, when the top of the ice is being pounded by the sun and the temperature above the ice is rising constantly.

From what I can see the pack is breaking up and melting over 60% of its area now and webcams are showing that every time the clouds go away, the near shore melt is very rapid and long lived. It should go over a cliff in July and August. The wildcard will be September storms.

Whatever happens I don’t expect it to be dull.

[Remko Kampen]

What's the greatest drop in sea ice area (CT) ever? And the greatest increase of negative anomaly?
Seems yesterday was a candidate.

[Wynn Dego]

1.75 million km2

The number is a WAG, but my reasoning is that as volume decreases, eventually a point is reached where the ice is so thin, that its resistance to break-up and melting is greatly diminished, and it disappears much more rapidly area-wise than before. I think this is the year, and it's likely the first of a 2 or 3 year descent to zero, i.e. 'ice free summer'.

[Jdean Dingler]

"It is impossible to tell how thick ice is on concentration charts. Just because ice is broken into smaller floes by the wind doesn't mean it melted. The buoys say no melt. And they said lots of snow fell over those areas. Which is a negative feedback."

Is is a negative feedback? Doesn't this insulate the ice against heat loss into the open air and allow the warmer ocean temps to more effectively warm the ice?

[Mignonette]

@Remko Kampen: It's easy to get excited about a single day's drop in one of the data sets but that does not say anything about the final outcome for the melting season. Cryosphere Today has the bad habit of alternating precipitous drops with periods of slowdown even when reality is probably much smoother.

There is no reliable indication yet that volume can keep up with the 2012 and 2011 values so I stick with a conservative rather than an aggressive estimate (4.4 Mkm^2).