The 2017 melting season

[Sterks]

In terms of drift, hycom predicts some movement, until the 25th, when the forecast is shown below. 26th and 27th similar frigid quietness. The broad highs seem not to set much in motion.
This is excellent to cool things down. If the highs bring clear skies, the better for cooling down ocean and ice.

[subgeometer]

That's a really nice animation ATeam. Those DMI SSTs sure are noisy. Interestingly it shows the blob I circled also appearing for one day about 2 weeks ago with what appears to be exactly the same shape

[be cause]

Hi Sub. I think your blob appears on the anomalous frame in A-teams gif .. the 20-08 frame appears in the middle and the end of the show .. the one in the middle would be causing the problem .. b.c.
  Certainly an interesting melting season is winding up . I was certainly not expecting this much ice at this stage . All the snow seems to be significant in protecting what little ice we had . how will all the snow affect the freezing season .. there must be quite a crust atop much of the remaining ice .. b.c.

[Niall Dollard]

Recent trend has been unlike any of the past few years. NSIDC now has extent level with 2011.

If it continues at this rate, the NSIDC min looks like it could be around the 5 million km² mark.

Edit: not appearing in the legend but 2017 is the light blue line.

[F.Tnioli]

...
If it continues at this rate, the NSIDC min looks like it could be around the 5 million km² mark.
...

It shouldn't, definitely not at this rate at least.

[iceman]

I've been wondering for a while what's up with the ESS bulge that survived thus far in the face of a very early opening of the ESS, and major advances on the Chukchi CAB front recently. My take is that it could soon disappear despite its longevity up to this point.
   ....

It does look vulnerable, though PIOMAS has only the outer fringe in the lowest-thickness band.

   ....
Not much wind lately or in the forecasts, perhaps enough to "pack up things" a little before coldness takes over ...

Concentration in the lobe may be low enough that light southerly winds won't cause much compaction.  Bottom melt is likely to predominate.  So this could be a good test of whether thin ice can hold out in the last month of the melting season, under what might be near-neutral conditions.

[A-Team]

near-neutral conditions.

The best forecast for the next few days is probably at the new ESRL site. Net melt continues over the entire area, with an extensive periphery losing at a moderate 2.6 cm/wk rate. Overall ice volume could also be estimated though the product does not cover the whole Arctic Ocean.

Very little action in sea ice fraction is foreseen to Aug 26th on the ESS lobe, whereas the Beaufort arm above Banks Island declines markedly without melting out its extent boundary, though that trend looks stronger already on UH AMSR2 by August 21st.

https://www.esrl.noaa.gov/psd/people/amy.solomon/seaice.html

[greatdying2]

Someone should drop a buoy (or 10) in that polynya to help figure out what's going on.

[gerontocrat]

Someone should drop a buoy (or 10) in that polynya to help figure out what's going on.

The budgets for environmental science are being hacked. The satellite record is at extreme risk. The number of buoys next year is questionable.

[meddoc]

The budgets for environmental science are being hacked. The satellite record is at extreme risk. The number of buoys next year is questionable.

I think Data Reliability has already suffered some heavy Blows this Year.
All that Soup, being accounted for as Extent.
When comparing PIOMAS and Hycom Numbers just don't add up, regarding recent Years (2012 onwards).
SIE at NSIDC seems  to have flattened out the recent Days.

Everything's fishy.
And meanwhile, Fakestream Media still blubbing about 2100...

[Lord M Vader]

Seems like the "Garlic Press" is under way by judging from ECMWF 00z op run... In addition it looks like a strong HP-dome is going to cover most of the Arctic basin beginning in about 48-72 hours. Some late melting potential should be able over the PAcific side and perhaps some minor oe for the Atlaantic sector. But the most serious thing should be the "Garlic Press" if it pushes a lot of MYI towards the CAA.

[Bill Fothergill]

Plea for help guys! I need to trace a link to one our our member discussing the current 15% or more measure for extent/area with NSIDC ? ...

G-W, I was one of the people in communication with the NSIDC on this subject. I've sent you a Personal Message with the relevant emails appended.

[Shared Humanity]

The budgets for environmental science are being hacked. The satellite record is at extreme risk. The number of buoys next year is questionable.

I think Data Reliability has already suffered some heavy Blows this Year.
All that Soup, being accounted for as Extent.
When comparing PIOMAS and Hycom Numbers just don't add up, regarding recent Years (2012 onwards).
SIE at NSIDC seems  to have flattened out the recent Days.

Everything's fishy.
And meanwhile, Fakestream Media still blubbing about 2100...

Is it data reliability or is the extent measure simply capturing the behavior of the pulverized and more mobile sea ice? There is a four decade trend towards increased dispersion at minimum which I think is a result of the destruction of thick MYI.

[Daniel B.]

...
If it continues at this rate, the NSIDC min looks like it could be around the 5 million km² mark.
...

It shouldn't, definitely not at this rate at least.

I think he was referring to the current rate after the recent slowdown.  If it were to continue through to mid-September, that situation would occur.  I would agree with you that the minimum will be below the 5 million mark, as these slowdowns do not tend to last.  Still, I may have to change my prediction of 3th or 4th lowest to 5th or 6th.

[Thomas Barlow]

Has anyone figured out what 15% ice cover (least of the extant values), would be like if you are in a boat? ie. How far apart would icebergs be, roughly?

[Daniel B.]

Has anyone figured out what 15% ice cover (least of the extant values), would be like if you are in a boat? ie. How far apart would icebergs be, roughly?

Check out the graphics here.

http://seaiceatlas.snap.uaf.edu/glossary

[Thomas Barlow]

Thanks. I still can't tell how far apart they are (on average)?

[gerontocrat]

I always thought the 15 percent rule originated from the days of sail as a measure of when it would be dumb to venture further into the ice. But I have never found a confirmation of this.

[Bill Fothergill]

I always thought the 15 percent rule originated from the days of sail as a measure of when it would be dumb to venture further into the ice. But I have never found a confirmation of this.

I thought the same as you, i.e. anything more than 15%, and you don't even want to think about going there.

However, there is also another reason for that particular value. During one of my dialogues with the NSIDC team, I was given the following response...

"15% concentration is useful for some marine navigation. However, the contour was originally developed since this threshold provided the best agreement between passive microwave remote sensing on space-paced platforms and aerial overflight work in early studies (e.g. Cavalieri et al. 1991: Aircraft active and passive microwave validation of sea ice concentration from the Defense Meteorological Satellite Program Special Sensor Microwave Imager. Journal of Geophysical Research: Atmospheres, 96, 21989 – 22008, doi: 10.1029/91JC02335.)"

[Richard Rathbone]

I always thought the 15 percent rule originated from the days of sail as a measure of when it would be dumb to venture further into the ice. But I have never found a confirmation of this.

I thought the same as you, i.e. anything more than 15%, and you don't even want to think about going there.

However, there is also another reason for that particular value. During one of my dialogues with the NSIDC team, I was given the following response...

"15% concentration is useful for some marine navigation. However, the contour was originally developed since this threshold provided the best agreement between passive microwave remote sensing on space-paced platforms and aerial overflight work in early studies (e.g. Cavalieri et al. 1991: Aircraft active and passive microwave validation of sea ice concentration from the Defense Meteorological Satellite Program Special Sensor Microwave Imager. Journal of Geophysical Research: Atmospheres, 96, 21989 – 22008, doi: 10.1029/91JC02335.)"

15% is the accuracy at the lower end of the concentration scale, which is why the match works best at this level. A measurement of above 15% almost certainly has ice in it, one below could well be open water.

[magnamentis]

I always thought the 15 percent rule originated from the days of sail as a measure of when it would be dumb to venture further into the ice. But I have never found a confirmation of this.

I thought the same as you, i.e. anything more than 15%, and you don't even want to think about going there.

However, there is also another reason for that particular value. During one of my dialogues with the NSIDC team, I was given the following response...

"15% concentration is useful for some marine navigation. However, the contour was originally developed since this threshold provided the best agreement between passive microwave remote sensing on space-paced platforms and aerial overflight work in early studies (e.g. Cavalieri et al. 1991: Aircraft active and passive microwave validation of sea ice concentration from the Defense Meteorological Satellite Program Special Sensor Microwave Imager. Journal of Geophysical Research: Atmospheres, 96, 21989 – 22008, doi: 10.1029/91JC02335.)"

15% is the accuracy at the lower end of the concentration scale, which is why the match works best at this level. A measurement of above 15% almost certainly has ice in it, one below could well be open water.

thanks for explaining this and especially good to know that it's not arbitrary

[Andir]

At this time, The Arctic Sea Ice has a very small volume, but in relation big extent/area . Is my feeling right, that with this constellation the  minimum for extent will be more late this season.

[A-Team]

Arctic Sea Ice has a very small volume, but in relation big extent/area with this constellation the  minimum for extent will be later this season.

You are quite correct that the freezing season has not begun -- the best available evidence, which seems to be at the links below, shows net melt everywhere out through August 27th. It is better to wait on NSIDC's call than go charging off on divisive new forums out of synch with the scientific community.

I am skeptical that anyone here is running more sophisticated models than ESRL. Most people here are not running any model at all and indeed lack access to the necessary computing resources. It is usually better to defer to those with far greater skill sets -- unless you can specifically document problems with their products and link to your conflicting data.

Meanwhile I received an update today on the wonderful new ESRL forecast page from Dr. Janet M Intrieri, deputy branch chief scientist at NOAA.

https://www.esrl.noaa.gov/psd/forecasts/seaice/ new page
https://www.esrl.noaa.gov/psd/forecasts/seaice/model_validation/ hindcast validation
ftp://ftp1.esrl.noaa.gov/RASM-ESRL/ModelOutput all-important new archive with netCDFs
https://www.esrl.noaa.gov/psd/forecasts/seaice/skill_assessment/ 2016 freezeup skill

https://www.esrl.noaa.gov/psd/people/amy.solomon/seaice.html all-in-one old page

"These 0-10 day, experimental sea ice forecasts are produced by the NOAA Physical Sciences Division from a fully coupled ice-ocean-atmosphere model called RASM-ESRL. RASM-ESRL is run daily and posted online at 2 UTC [seeking clarification on 2:00 vs 14:00].

The model is initialized with the NOAA Global Forecast System (GFS) analyses and the Advanced Microwave Scanning Radiometer 2 (AMSR2) sea ice concentrations. The model is forced at the lateral boundaries by 3-hourly GFS forecasts of winds, temperature, and water vapor.

RASM-ESRL, adapted from the Regional Arctic System Model (Maslowski et al. 2012) was modified for short-term, weather-scale forecasts and includes the following model components: the Weather Research and Forecasting (WRF3.5.1; run with 40 vertical levels) atmospheric model; the Parallel Ocean Program (POP2) model; the Los Alamos Community Ice Model (CICE 5.1, Hunke et al. 2013); and the NCAR Community Land Model (CLM4.5).

All components, run at 10 km horizontal resolution, are coupled using a regionalized version of the CESM flux coupler (CPL7), which includes modifications (Roberts et al. 2014) important for resolving the sea ice pack response to weather events.

Other model optimizations include: a bulk double-moment cloud microphysics scheme for droplets and frozen hydrometeors (Morrison et al. 2009), running ensemble forecasts initialized with GEFS ensemble members, and extending the model domain to include the Bering Strait and Svalbard.

https://www.ncdc.noaa.gov/data-access/model-data/model-datasets/global-ensemble-forecast-system-gefs 21 ensemble members

New for 2017:

Extended domain includes Bering Strait [confirming they mean Barents Sea here]
Ice thickness initialization uses spring CryoSat-2
Weekly sea ice concentration updates with AMSR-2
Ensemble runs using GEFS members (in progress)
Updated website with new guidance products, meteograms, and cross-sections

We greatly appreciate feedback on products you find useful or requests for any additional fields you would like to see on the web page. Model output is continually being evaluated.  Please visit the site for updated information on skill, ensemble performance, and comparisons with observations as the freeze-up season progresses."

[gerontocrat]

Meanwhile I received an update today on the wonderful new ESRL forecast page from Dr. Janet M Intrieri,

Any info on anything regarding the Antarctic. Seems to be the poor relation as regards development, and yet maybe of greater significance in the long-term. Hoping I am wrong on  both counts.

[Sterks]

Since the ESRL postprocessed total melting has the limitation of being cut off the at the rather meager value of 0.35 cm/day, I can't but suppose that the computed melt is much higher at many ice edge locations.
This tool can be validated, no better moment than September. During those few days when the pack goes white in the concentration maps and compactness has a fast increase, this tool will have to show mostly zero or negative melt (refreeze) in general inside the pack.

[Brigantine]

I'm watching that little remnant of ice around the eastern tip of Baffin Island (Davis Strait map on CIS).

Last year it was gone in early September. This year it seems to be melting a week behind schedule, but then the water around (/ underneath?) it seems warmer too.

Any guesses when that will be gone?

Also that ice island at 67*25' N 63*30' W (in CIS maps) has been there for a long time. Is it real? What's known about it?

[A-Team]

total melting has the limitation of being cut off at 0.35 cm/day,  the computed melt is much higher at many ice edge locations.

That is a good one to write them about ... questions are encouraged.

Looking at the ESRL ice thickness chart, that too has a cut-off, at 1.6 m (or greater). There could be a good reason for that, or just some choice made in map-making. (The netCDF files might have more.) At any rate, each thickness class is scored below. The four columns are percent of colored pixels, pixels counts of each palette bin, mid-range thickness, and product volume.

The latter would need normalization by pixel area to get at cubic km (assuming an ~ equal area projection, this one looks like Lambert azimuthal). However it would only be feasible to compare with Piomas (which aren't available yet for this date) for the ice thickness categories at 1.5 m and below. To do this comparison at all accurately it would be necessary to re-grid both to a common resolution (which seems vastly higher at ESRL). It is probably feasible to animate their ice thickness map and others from the archive, though that does not seem to go very far back.

There's a lot to look at here with 37 products. I will be asking about rainfall as we have that Aug 5th eyewitness at the pole, earlier observations by JayW, and lightning data over the Beaufort.

[greatdying2]

Meanwhile I received an update today on the wonderful new ESRL forecast page from Dr. Janet M Intrieri, deputy branch chief scientist at NOAA.

Fantastic

[Lord M Vader]

In about 48 hours, a radical shift to more southerly winds seems to impact the Atlantic sector. Given the void of these winds it will be highly interesting to see how the sea ice there will react when warm and salty water will hit it.

In addition, these winds will kick the "Garlic Press" into high gear.

To be continued!

[crandles]

SIPN 2017: August Report is out

https://www.arcus.org/sipn/sea-ice-outlook/2017/august

the median Outlook value for September 2017 Arctic sea ice extent of 4.5 million square kilometers with quartiles of 4.2 and 4.8 million square kilometers (See Figure 1 in the Overview section, below). These values are unchanged from the July Report, which is consistent with the moderating impact of summer 2017 Arctic weather.

Method 2017___Med   Q1   Q2   N
Heuristic_______4.3   3.7   4.7   4
Statistical______4.8   4.5   5.0   16
Mixed_________4.6   4   4.6   3
Dynamical_____4.4   4.0   4.7   14
All___________4.5   4.2   4.8   37
 

[Thomas Barlow]

At this time, The Arctic Sea Ice has a very small volume, but in relation big extent/area . Is my feeling right, that with this constellation the  minimum for extent will be more late this season.

It is possible that the season could go longer, hard to say.
It's going to be bad, no matter how you slice it.
All the models are very fuzzy-edged, none has any claim to fame over the others, but looking at several different sources, and judging by SST anomalies, that could place it in the later-recovery category.
However, one thing seems clear, the Arctic Ocean itself, may be in the worst state (volume) on record, as I show here ---> https://forum.arctic-sea-ice.net/index.php/topic,119.msg126019.html#msg126019

[Daniel B.]

SIPN 2017: August Report is out

https://www.arcus.org/sipn/sea-ice-outlook/2017/august

the median Outlook value for September 2017 Arctic sea ice extent of 4.5 million square kilometers with quartiles of 4.2 and 4.8 million square kilometers (See Figure 1 in the Overview section, below). These values are unchanged from the July Report, which is consistent with the moderating impact of summer 2017 Arctic weather.

Method 2017___Med   Q1   Q2   N
Heuristic_______4.3   3.7   4.7   4
Statistical______4.8   4.5   5.0   16
Mixed_________4.6   4   4.6   3
Dynamical_____4.4   4.0   4.7   14
All___________4.5   4.2   4.8   37
 

One caveat in the report, "There is more divergence in early August, and in the past week the ice retreat has slowed considerably; this means that forecasts starting from the end of July may likely end up being lower than the actual September extent."

[numerobis]

I'm watching that little remnant of ice around the eastern tip of Baffin Island (Davis Strait map on CIS).

Last year it was gone in early September. This year it seems to be melting a week behind schedule, but then the water around (/ underneath?) it seems warmer too.

Any guesses when that will be gone?

Also that ice island at 67*25' N 63*30' W (in CIS maps) has been there for a long time. Is it real? What's known about it?

A friend just back from two weeks camping on the beach just outside Qik had big chunks of ice on many of her pictures, so I suppose there really is the ice that CIS claims there is. That said, there are also icebergs around there -- but icebergs tend to be far bigger.

She reported it was quite cold (as I'd seen from forecasts). Cold enough to snow a bit in camp, and to start to cover the mountains. South Baffin has been colder than "average" all summer, one of the few blue spots on the global temperature anomaly maps.

In other news: The Eureka ice maps are showing movement through Nares the past few days. And the narrow almost-a-passage through the NWP closed off yesterday.

[greatdying2]

In addition, these winds will kick the "Garlic Press" into high gear.

Yep. Wipneus has just posted a nice animation on his home brew thread showing how the CAA is now smashed and ready...

Also definitely worth keeping an eye on that big low in the long-term forecast on the Atlantic side. Things may get interesting again soon.

[meddoc]

In addition, these winds will kick the "Garlic Press" into high gear.

Yep. Wipneus has just posted a nice animation on his home brew thread showing how the CAA is now smashed and ready...

Also definitely worth keeping an eye on that big low in the long-term forecast on the Atlantic side. Things may get interesting again soon.

Those 2 Lows should not get ripped in half. The Vortex will have difficulty to form.
Year by Year it is forming ever later & desintegrating ever earlier...

[Gray-Wolf]

Yes it will be interesting again as we watch the PV try and form up?

Just watched Jen Francis at a climate meet the winter before last putting forward a "Two to Tango" notion where PDO positive actually helps extend the triple R ( ridiculously resilient Ridge) into the basin ensuring early ,fast ice loss on the Pacific side ( so allowing a lot of energy into the waters) and she thinks this is also what is throwing the PV out of kilter?

Last year i thought it was the cold dome over the WACCy snow in west Siberia that was causing issues but I'm pretty convinced by her proposition that sometimes 'naturals' augment the AGW signal and so push things faster/harder?

Poor Neven! Once upon a time ice watching was just a summer thing....... now we have to watch all year!!!!

[Shared Humanity]

The past 2 winters were riveting and I have been visiting this site daily, year round, due to this.

[Sterks]

total melting has the limitation of being cut off at 0.35 cm/day,  the computed melt is much higher at many ice edge locations.

That is a good one to write them about ... questions are encouraged.

It turns out that there is a tab dedicated to thickness change (snow and ice), and the scale does not cut off the different melts (bottom, lateral, top). In this case I cropped the right-side palette scales to fit the screenshot into the 700x700, but they were identical to the scale for the left-upper melt map (ice bottom melting).
In fact, bottom melt is forecasted of the order of 1 cm/day and up to 2 cm/day in the very edge of the pack.
It surprises me how much bottom melt varies day to day, and how dependent it is of the changing winds in the next days.
And it surprises me the almost absence of noticeable lateral melt in the model forecasts. Nothing going on inside the "soup"? Even with these temperatures I would have expected to see something under the white bin, of -0.3 /day (funny that in the total melt graph, the researcher sets up an Upper cut-off of close to 0.3/day, if we combine this bottom melt map with the total melt map, we almost get the complete picture of it)
As soon as it gets "quiet", bottom melt is reduced, that is nicely reflected in the forecast.
Can we consider the "2017 Surface Melting Season" finished?

[Sterks]

Plenty of compacting winds for the next five days, warm from the Atlantic too, that will bring losses of extent for the rest of August.
I don't see a garlic press happening though from the Hycom drift forecasts. But at any rate the Hycom forecasts are not at its height in reliability, momentarily

[A-Team]

Sea surface temperatures (SST) in the 'upper meters' are shown for the last 31 days in the animations according to DMI's methods as explained below. (Note an earlier post above looked at anomalies (SSTA) relative to 1985-2001, not literal temperatures. The first time derivative, or daily difference, though is the same for either since the mean is a constant.)

The first animation uses their palette which encompasses colors appropriate to much more extreme variations than we have in August such as might be encountered at lower latitudes or in different seasons.

This causes the temperatures of interest to ice melt in the Arctic Ocean proper to suffer from congested colors, unresolvable in the green near 0ºC. The second animation stubs in a different palette (four ramps, interdigitated) that resolves these key temperatures better.

If the ice should move over warmer waters in coming days, then bottom melt might be expected. That's better left to ESRL than to Hycom at five days out.

The maps cannot be rotated -45º into 'Greenland down' or re-sized arbitrarily without degrading either the palette or data. However it is very easy to construct and vary the new palette to taste as DMI, like UH and UB, serves up clean maps (discrete map colors matching palette tiles).

The web page http://ocean.dmi.dk/satellite/index.uk.php shows the sea surface temperature (SST) and anomalies derived from infrared measurements from the polar orbiting satellites. One interpolated field is constructed daily. Only nighttime SST observations are used for the interpolation because these are more representative of the temperature in the upper meters of the water column. One image per day is shown for the last 30 days.

The SST anomalies have been calculated with respect to a mean, which has been derived from observations from 1985 to 2001. They are SST climatology monthly values from the Pathfinder project and temporal interpolation is used between the two nearest months to obtain this days SST climatology. For more information on the monthly climatology, see: www.nodc.noaa.gov/sog/pathfinder4km/

Gaps in the observations due to clouds are filled using a 3-dimensional Optimal Interplation technique (see Høyer and She, 2007). The interpolation scheme uses statistics, which are derived locally and provides the "best possible" estimate of the SST observations, assuming steady state statistics. The mean error of the gridded SSTs is about 0.5-0.7oC.

The observations of the sea surface temperature are based upon observations from up to 10 different satellites, measuring in Infrared and Microwave wavelengths. The observations are obtained from the Ocean and Sea Ice SAF project (www.osi-saf.org) and from Group on High Resolution Sea Surface Temperature (www.ghrsst.org).

[Sterks]

That's great A-Team.
The only doubt I have is how DMI satellite-based SST maps can display above-zero SSTs under the ice (notably at the Beaufort appendix).
But I assume that it is normal noise, and also may be caused by the current thinness and sparseness of ice over there.

[Lord M Vader]

Sterks: check the EPS daily average, not the 1-5 day period. However, the "Garlic Press" isn't as pronounced as it was a few days ago.

[Thomas Barlow]

Model correction .... or the slippery slope?
The former I think.

[Sterks]

Sterks: check the EPS daily average, not the 1-5 day period. However, the "Garlic Press" isn't as pronounced as it was a few days ago.

Thank you
The multiple averaging of the EPS across simulations and over 5 days must smear the isobars through CAA. These are clear when looking at the daily ECMWF.

[gerontocrat]

JAXA data says average extent loss per day from 17th to 21st Aug was less than 20,000 km2 per day.
In last 4 days gone up to 50,000 km2 per day, just below average for time of year.
The NSIDC graph seems to mirror that - so not a model correction?

[A-Team]

the "Garlic Press" isn't pronounced

It takes wind from the north to apply a force to the main ice pack to push it through the CAA channels (which additionally need to be somewhat open). However there hasn't been any weather for quite a while this summer nor is any foreseen to August 30th by ESRL.

The lower animation, UH AMSR2 3 km, shows mobility but very little net change over the last two weeks in the CAA garlic press area; purple is 100% concentration. The upper animation shows ice thickness and ice motion for the next five days as foreseen in plotRASM10 and plotRASM10 from the ESRL REB_plots.2017-02-25.tar.gz entry in the archives.

ftp://ftp1.esrl.noaa.gov/RASM-ESRL/ModelOutput

[Thomas Barlow]

JAXA data says average extent loss per day from 17th to 21st Aug was less than 20,000 km2 per day.
In last 4 days gone up to 50,000 km2 per day, just below average for time of year.
The NSIDC graph seems to mirror that - so not a model correction?

Yes, could be a slippery slope instead.
Maybe because July, early Aug. may have had more peripheral ice than usual, collapsing quickly as mentioned --> https://forum.arctic-sea-ice.net/index.php/topic,1834.msg121998.html#msg121998

[Thomas Barlow]

Based on this, there is some chance that if 2017 follows the 2016 rate, but season goes as late as 2007 did, then the 2nd worst extent on record could happen.

http://nsidc.org/arcticseaicenews/charctic-interactive-sea-ice-graph/

[greatdying2]

I think the earlier speedup and the slowdown were simply caused by weather, stormy on the Pacific side, but more recently, great weather for ice preservation. I guess ice is slow to melt when it just sits in its own meltwater -- needs a bit of stirring.

Now, the weather is starting to moving again. I expect the speedup to continue for a few days at least, somewhat on the Russian side and especially on the Atlantic side, where low concentration ice is about to meet some wind and waves.

A speed-up, at least for today. Ice retreating along the ice edge facing ESS,  in the CAA and near Svalbard.

[A-Team]

ice is slow to melt when it just sits in its own meltwater -- needs a bit of stirring.

The animation at bottom shows another re-paletting, this time to grayscale tiles spaced by 10 units on the 8-bit [0 255] scale. This allows quantitative averaging of SST over the 30 days that stays within the color space. The mean is shown as well as the 0ºC-and-below bounding isotherm.

As Sterks notes, it is a bit mysterious how the SST under the ice is determined at all, why it varies day to day, and if this variation is just noise then why it doesn't smooth out more upon averaging.

(There's ample opportunity for confusion here as a freshwater floe would melt laterally in contact with ample 0º+ water, whereas 36 psu saltwater needs -1.8ºC to freeze, yet the Beaufort Gyre dome has been somewhat freshened in the past to say 26 psu, and sea ice has variable extents of brine depending on its age.)

DMI, as with about half our maps, does not provide the underlying netCDF data grid. If they had, averaging could have been done directly in excel or alternatively in equal area pixel maps generated  in the grayscale provided by Panoply. We will be needing Panoply in any event to re-project .nc files from the new ESRL mega-resource into our standard stereographic map for overlays and comparisons.

https://www.giss.nasa.gov/tools/panoply/ much improved version 4.8.2 netCDF open source