The 2017/2018 freezing season

[Sterks]

I dare open this thread with the image below. Drift predictions for the 25th, which are similar 26 and 27th.
High pressure will dominate Arctic, broad clear skies with little movement.
Will this strange melting season bring a minimum in August?
Also attached the ensemble average of pressures next five days. This should decrease extent, but the isobars are not very tight...

[Pavel]

So the cold melting season comes to the end and we begin the warm freezing season 

[Jim Hunt]

Will this strange melting season bring a minimum in August?

I'd hazard an (educated?) guess. No.

Which metric does your "minimum" refer to?

[Sterks]

Not an specific one, but is IJIS extent ok? since this spinned off from the discussion of that thread.
In any case, it will be interesting to see if the minimum happens before or very early September, since the weather does not seem very aggressive.

[oren]

Minimum could come 5th-10th of Sept, but very unlikely in August. I think this thread is too early...

[gerontocrat]

Minimum could come 5th-10th of Sept, but very unlikely in August. I think this thread is too early...

When I see 3 (three) extent gains in a row then I will change my spreadsheets from melt to freeze. Feb 29 is going to be a pain to deal with - I see a fudge coming.

[Sterks]

Minimum could come 5th-10th of Sept, but very unlikely in August. I think this thread is too early...

We'll see
Anemic melting this last stretch of August if you ask me...

[Ned W]

Every JAXA daily minimum so far has been during the two week interval of Sept 7-21.

The linear trend in JAXA extent over the past week is -22690 km2/day.  This is the earliest in the season that such a slow week has occurred.  However, several other years show a similar 7-day slowdown starting soon. 

On average, a slowdown like this occurs 11 days before the minimum ... but it ranges from 28 days before minimum (in 2005) to 2 days after minimum (in 2008).

The fascinating thing, though, is that an early slowdown like this is anticorrelated with an early minimum.  Years with an early slowdown tend to have late minimums, not early ones:



The data are noisy and the standard error is large, but the p-value is highly significant (0.012) at a=0.05.

Based on that model, one would expect the 2017 minimum to be somewhere between days 257-271 (i.e., September 13-27) with a best estimate of day 264 (Sept 20).



Make of it what you will.

[Daniel B.]

Every JAXA daily minimum so far has been during the two week interval of Sept 7-21.

The linear trend in JAXA extent over the past week is -22690 km2/day.  This is the earliest in the season that such a slow week has occurred.  However, several other years show a similar 7-day slowdown starting soon. 

On average, a slowdown like this occurs 11 days before the minimum ... but it ranges from 28 days before minimum (in 2005) to 2 days after minimum (in 2008).

The fascinating thing, though, is that an early slowdown like this is anticorrelated with an early minimum.  Years with an early slowdown tend to have late minimums, not early ones:



The data are noisy and the standard error is large, but the p-value is highly significant (0.012) at a=0.05.

Based on that model, one would expect the 2017 minimum to be somewhere between days 257-271 (i.e., September 13-27) with a best estimate of day 264 (Sept 20).



Make of it what you will.

Yes, I would not correlate it with an early minimum just yet.  Many factor go into the exact timing of the minimum, with local weather being one of the largest.  Perhaps the recent slowdown can be correlated with the early snows and mass buildup in Greenland.  This year has been tracking quite closely with last year, which had a minimum of Sept. 10.  Without putting too much into an analysis, that is my guess.  However, I would not put too much faith in predicting a particular day, or even week.

[greatdying2]

Although the melt has slowed down, particularly surface melt and storm-associated melt due to nice weather, this thread is too soon. OTOH, I'm not sure it even makes sense to have separate melt and freeze threads...

[Ned W]

. OTOH, I'm not sure it even makes sense to have separate melt and freeze threads...

If nothing else, it ensures a switch to a new thread every six months.  That at least keeps it from growing out of control. 

[Sterks]

Although the melt has slowed down, particularly surface melt and storm-associated melt due to nice weather, this thread is too soon. OTOH, I'm not sure it even makes sense to have separate melt and freeze threads...

In your opinion.
In my opinion, it doesn't hurt to speculate in parallel threads during the transition, as usual. And this season has proven colder than expected once and again.

While I still expect some extent drops in the next few days, this weekend may bring another pause. Additionally, temperatures for next week, see ASIG, while GFS the least reliable source, for the last days of August seem falling.

Is there a lot of bottom melt? Or most of the ice bottom-melting at this stage is gone because it was much thinner than usual? No winds, no movement, bottom and lateral melt must be very slow.

[Neven]

As soon as the melting season is over for real, I'll change the title to '2017/2018 freezing season'. 

My guess - without looking at too many graphs and maps lately, I must admit - is that the low concentration stuff within the ice pack is filling up, and there's not enough weather to stir things at the periphery. As soon as that's done, melting on the periphery will keep things going very slowly into September.

I think end position 2 will be difficult to attain. Maybe top 3?

[Sterks]

Every JAXA daily minimum so far has been during the two week interval of Sept 7-21.

On average, a slowdown like this occurs 11 days before the minimum ... but it ranges from 28 days before minimum (in 2005) to 2 days after minimum (in 2008).

The fascinating thing, though, is that an early slowdown like this is anticorrelated with an early minimum.  Years with an early slowdown tend to have late minimums.
....
The data are noisy and the standard error is large, but the p-value is highly significant (0.012) at a=0.05.

Based on that model, one would expect the 2017 minimum to be somewhere between days 257-271 (i.e., September 13-27) with a best estimate of day 264 (Sept 20).

Make of it what you will.

Nice.
 Yes, it may happen, at least you have the numbers, I just have the hunch this year will be slow and early, adding noise to the stats. Been a weird melting season.

[Cid_Yama]

There is always somebody trying to rain on the melting season early.  They are ALWAYS wrong.

Every party needs a pooper, that's why we invited you. Party pooper.

It's only August 22.  2016 didn't fall off the cliff until the last couple days of August.

And we have far less volume.

[Coffee Drinker]

According to GFS, the freezing season should start next week in parts of the Canadian archipelago.

Night temperatures of -10C will be expected around Baffin Island.

[Sterks]

Every party needs a pooper, that's why we invited you. Party pooper.

I am not going to respond to your insult.
My speculation has easily been put in question by cold numbers (see Ned's above) and experience (see Neven's above), but it is not affected by insults.
We can make of this, 1 Cross-firing of ad-hominems WUWT Style, 2 A dead thread until the freezing season is well-acknowledged 3 Somewhere to place facts of starting symptoms of the pack refreezing even before the minimum, be it on August or October, ... or not refreezing if you wish
I prefer 2 or 3

[Ned W]

Keep in mind that the current "melting season" thread was started on January 1 and had dozens of posts in January and February, and over 100 posts before the end of the freezing season.  There's absolutely nothing wrong with starting this thread now.

It makes sense for them to overlap, because the freezing season starts in some parts of the Arctic while the melting season is still finishing up in other parts.

[Shared Humanity]

Keep in mind that the current "melting season" thread was started on January 1 and had dozens of posts in January and February, and over 100 posts before the end of the freezing season.  There's absolutely nothing wrong with starting this thread now.

It makes sense for them to overlap, because the freezing season starts in some parts of the Arctic while the melting season is still finishing up in other parts.

And the three measures (extent, area and volume) behave differently during the transition.

[echoughton]

I Love this forum. SO much incredible talent and science. Yet it usually settles in on opinion. LOL.
With the wild slowdown and near-future weather forecast...it may still be possible that my 4.75-5 mil sq K prediction could happen.

[Pavel]

Looking everywhere on the SSTs one can expect sluggish ice growth both in Inner Basin and peripheral. The extent curve can dip lower than 2016. Not only in summer the cannonballs could appear

[Ned W]

I used Wipneus's file of daily PIOMAS volumes aggregated by region to make this chart, which shows the start of the freezing season for each basin, for each year (2000-2016):



The Y axis is day of year.  There are big and consistent differences in terms of the onset of freezing from basin to basin (note the positions of the diamonds, which are the median date of the start of the freezing season).

Medians, sorted from earliest to latest:
GrnLS   237   25-Aug
Baffn   247   4-Sep
CAB   256   13-Sep
CAA   257   14-Sep
Hudsn   257   14-Sep
Beauf   263   20-Sep
KaraS   264   21-Sep
Laptv   265   22-Sep
Baren   266   23-Sep
ESS   267   24-Sep
Chukc   270   27-Sep
Okhot   311   7-Nov
Berng   312   8-Nov

Comments, corrections, suggestions for improvement etc. are welcome.

Annoying details:  I defined the start of the freezing season as the date of the first increase in volume after the minimum.  If a basin hits the same minimum repeatedly during a single year, it's the day after its first minimum (maybe it would be better to have used the last, but no big deal). A couple of random weird outliers (noise) were deleted.

[jplotinus]

According to GFS, the freezing season should start next week in parts of the Canadian archipelago.

Night temperatures of -10C will be expected around Baffin Island.

What is source for -10° temps "around Baffin Island"? Weather forecast for Clyde River, which is at middle of east coast of Baffin, has temps above 0° throughout next nine days, with highs above 10° for some of them.

[Sterks]

@Ned nice job, for further discredit of my opening post
For the last metric, I was expecting area to really start reaching bottom, but not really. As unusual as this season usually has been. The area curve should be getting flat now and extent keep dropping, not the opposite.

[oren]

Ned very interesting. Off the cuff, most areas minimum is determined by export into them. Best shown by the Greenland Sea. The CAB is really the best indicator for real freezing onset.
In addition I suspect your criteria are picking up weirdness somehow. Hudson refreezing on day 220? Bering refreezing before Chukchi? Possibly small transitory blips cause this. Maybe add a threshold criterion of an absolute or percentage gain to mark a day as past the minimum. And/or use the last time at the minimum rather than the first.

[magnamentis]

@Ned nice job, for further discredit of my opening post
For the last metric, I was expecting area to really start reaching bottom, but not really. As unusual as this season usually has been. The area curve should be getting flat now and extent keep dropping, not the opposite.

@Ned nice job, for further discredit of my opening post
For the last metric, I was expecting area to really start reaching bottom, but not really. As unusual as this season usually has been. The area curve should be getting flat now and extent keep dropping, not the opposite.

thanks for opening this thread now. the price for thinking or even acting ahead of the main stream seems to be high enough once more, i, like yourself, find it interesting to speculate the transition now that the season once more has proven it's "off any norm status" sooner or later we probably need a rule for opening such threads, until then i enjoy this and more to come. IMO it would even be possible to discuss the impact on much earlier states of the melting season on the coming freezing season. perhaps some people planned to open the freezing thread themselves ?

[Ned W]

Ned very interesting. Off the cuff, most areas minimum is determined by export into them. Best shown by the Greenland Sea. The CAB is really the best indicator for real freezing onset.
In addition I suspect your criteria are picking up weirdness somehow. Hudson refreezing on day 220? Bering refreezing before Chukchi? Possibly small transitory blips cause this. Maybe add a threshold criterion of an absolute or percentage gain to mark a day as past the minimum. And/or use the last time at the minimum rather than the first.

Great, this is exactly the response I wanted.

* Yes, it's quite possible that export (er, "import" in this case) is actually what produces the first rise after minimum in many basins in many years.

* Yes, some of the "weirdness" you identified was due to my poor choice of using the first increase after the first instance of the minimum.  I've now re-done the graph using the first day of increase after the last instance of the minimum:



The results are somewhat different, including addressing both the points of weirdness that you identified (though in 2000 the Bering still shows its first increase only three days after the Chukchi). 

Here are the updated medians:

GrnLS   239   26-Aug
Baffn   250   7-Sep
CAB   256   13-Sep
CAA   258   15-Sep
KaraS   265   22-Sep
Beauf   266   23-Sep
Laptv   268   25-Sep
ESS   269   26-Sep
Baren   270   27-Sep
Chukc   273   30-Sep
Hudsn   274   1-Oct
Okhot   311   7-Nov
Berng   312   8-Nov

I'm going to continue to think about this and see if I can come up with further improvements.

[Sterks]

Ned very interesting. Off the cuff, most areas minimum is determined by export into them. Best shown by the Greenland Sea. The CAB is really the best indicator for real freezing onset.
In addition I suspect your criteria are picking up weirdness somehow. Hudson refreezing on day 220? Bering refreezing before Chukchi? Possibly small transitory blips cause this. Maybe add a threshold criterion of an absolute or percentage gain to mark a day as past the minimum. And/or use the last time at the minimum rather than the first.

Great, this is exactly the response I wanted.

* Yes, it's quite possible that export (er, "import" in this case) is actually what produces the first rise after minimum in many basins in many years.

* Yes, some of the "weirdness" you identified was due to my poor choice of using the first increase after the first instance of the minimum.  I've now re-done the graph using the first day of increase after the last instance of the minimum:



The results are somewhat different, including addressing both the points of weirdness that you identified (though in 2000 the Bering still shows its first increase only three days after the Chukchi). 

Here are the updated medians:

GrnLS   239   26-Aug
Baffn   250   7-Sep
CAB   256   13-Sep
CAA   258   15-Sep
KaraS   265   22-Sep
Beauf   266   23-Sep
Laptv   268   25-Sep
ESS   269   26-Sep
Baren   270   27-Sep
Chukc   273   30-Sep
Hudsn   274   1-Oct
Okhot   311   7-Nov
Berng   312   8-Nov

I'm going to continue to think about this and see if I can come up with further improvements.

Greatness! Thanks Oren for suggesting the improvement.
We don't have a NASA cluster yet a nice list of what to expect region by region in the coming weeks based on past years, that's more than expected. Thanks Ned.
Of course there is the unpredictability component.

[Cid_Yama]

According to GFS, the freezing season should start next week in parts of the Canadian archipelago.

Night temperatures of -10C will be expected around Baffin Island.

What is source for -10° temps "around Baffin Island"? Weather forecast for Clyde River, which is at middle of east coast of Baffin, has temps above 0° throughout next nine days, with highs above 10° for some of them.

Thank you for jumping on that.  We don't need any disinformation.

[Coffee Drinker]

According to GFS, the freezing season should start next week in parts of the Canadian archipelago.

Night temperatures of -10C will be expected around Baffin Island.

What is source for -10° temps "around Baffin Island"? Weather forecast for Clyde River, which is at middle of east coast of Baffin, has temps above 0° throughout next nine days, with highs above 10° for some of them.

Thanks for the correction and sorry for the confusion. I mean Ellesmere Island not Baffin. Looked mostly at the station Eureka and the GFS 2m temp forecast from GFS.

Baffin is indeed very far from freezing conditions at this stage. Sorry again.

GFS has backed off now a little but still freezing weather expected from next week. Will be exciting to see the first new ice of the year.

[Archimid]

Thank you Ned W. I took the liberty to use ASIG's regional map and added the median day for the beginning of the freezing season to help me visualize it better.

[Ned W]

Very nice, Archimid!  The map definitely helps visualize this.

I've been messing around with this some more, and thought this was a useful comparison:



The first day of refreeze is calculated as before.  The first week represents the start of the first seven-day period of continuous increases in volume after the last day at the minimum.

The graph shows the means, but medians are probably better, and they're what Archimid used on the map, so here they are:

Okhot: first week = 336 (02-Dec) first day = 311 (07-Nov)
Berng: first week = 336 (02-Dec) first day = 312 (08-Nov)
Beauf: first week = 276 (03-Oct) first day = 266 (23-Sep)
Chukc: first week = 289 (16-Oct) first day = 273 (30-Sep)
ESS: first week = 272 (29-Sep) first day = 269 (26-Sep)
Laptv: first week = 277 (04-Oct) first day = 268 (25-Sep)
KaraS: first week = 276 (03-Oct) first day = 265 (22-Sep)
Baren: first week = 297 (24-Oct) first day = 270 (27-Sep)
GrnLS: first week = 249 (06-Sep) first day = 237 (25-Aug)
CAB: first week = 264 (21-Sep) first day = 256 (13-Sep)
CAA: first week = 268 (25-Sep) first day = 258 (15-Sep)
Baffn: first week = 280 (07-Oct) first day = 250 (07-Sep)
Hudsn: first week = 304 (31-Oct) first day = 274 (01-Oct)

[jplotinus]

According to GFS, the freezing season should start next week in parts of the Canadian archipelago.

Night temperatures of -10C will be expected around Baffin Island.

What is source for -10° temps "around Baffin Island"? Weather forecast for Clyde River, which is at middle of east coast of Baffin, has temps above 0° throughout next nine days, with highs above 10° for some of them.

Thanks for the correction and sorry for the confusion. I mean Ellesmere Island not Baffin. Looked mostly at the station Eureka and the GFS 2m temp forecast from GFS.

Baffin is indeed very far from freezing conditions at this stage. Sorry again.

GFS has backed off now a little but still freezing weather expected from next week. Will be exciting to see the first new ice of the year.

Thanks for your clarification. In 2017, Arctic sea ice extent has been as enigmatic as ever. I think all any of us can do with confidence is observe and try to interpret as best we can. The sea ice is telling us something, for sure.

[Ned W]

Right now it seems to be telling us that summer is ending and fall is on the way.  With all due respect to the ice, I could really use another month of summer!  Not ready for fall yet...
 

[Dharma Rupa]

Not ready for fall yet...
 

Fall is the second best season!   I've have enough of the heat and humidity.

[Niall Dollard]

Given that frazil ice can start to form when air temps are down to -6 C, we are not far off that figure according to nullschool at 85N 77W.

[jplotinus]

Given that frazil ice can start to form when air temps are down to -6 C, we are not far off that figure according to nullschool at 85N 77W.

Alert, at the north coast and of Ellesmere, is now experiencing below 0° temps on a daily basis, but not yet down to -6°
Since the Arctic summer has been 'cool', I wonder whether the autumn will revert to mean by being relatively warmer?

[Dharma Rupa]

Given that frazil ice can start to form when air temps are down to -6 C, we are not far off that figure according to nullschool at 85N 77W.

Alert, at the north coast and of Ellesmere, is now experiencing below 0° temps on a daily basis, but not yet down to -6°
Since the Arctic summer has been 'cool', I wonder whether the autumn will revert to mean by being relatively warmer?

I think you want to watch this for an answer:

[Quantum]

I would be gobsmacked if the melting season ended in August. Though I suppose stranger things have happened to the ice.

[Shared Humanity]

Given that frazil ice can start to form when air temps are down to -6 C, we are not far off that figure according to nullschool at 85N 77W.

Alert, at the north coast and of Ellesmere, is now experiencing below 0° temps on a daily basis, but not yet down to -6°
Since the Arctic summer has been 'cool', I wonder whether the autumn will revert to mean by being relatively warmer?

Arctic winters have been anomalously warmer for more than a decade. Yes, the 2016 winter was a record breaker but 2015 was ridiculously warm as well except when compared to 2016. You have to go back to 2004 before you find a winter where temperatures tracked with the 1958-2017 mean.

I am absolutely convinced this is the result of an emerging new Arctic climate regime and would be very surprised if winters do not continue to be anomalously warm.

[Niall Dollard]

A useful eye witness account from Pen Hadow's Arctic Mission which gives an on the spot idea of conditions at about 80 N around 4 hours ago. Very early stages yet for sea water freezing. (Even if they dont make 90N still they are still gathering useful data on the Arctic)

A missive from expedition co-leader and Bagheera skipper Erik de Jong: 'It's shocking to see how open the water is, almost no ice of any significance, therefore we can sail practically the course we want and at full speed. It is pretty cold out here, the air is a few degrees below freezing, and the water is a couple degrees below freezing. Due to the salinity of the seawater, it remains liquid, but it is border line. Fresh water patches are frozen, and so is the bottom of our fresh water tank! The pick up line for fresh water is  frozen shut and there is no more water coming out of the faucet. The prediction is that warmer waters from the south are moving in shortly, so it should remedy itself over the next few days. In the meantime, we have our backup supply in portable barrels  in the galley and I've hung a hose connected to the water pump, so that water can be used again. This will work for a week to 10 days without rationing usage. If the tank remains frozen till that time, there are some other tricks we can pull to get the water out of the tank. During the science experiments we are often anchored to an ice floe so that we don't move to fast and that we don't have to keep an eye on where the boats are drifting. Many of these floes have had several inches of fresh snow on them. We are collecting that in buckets and are able to add to our water supply that way too.'

[Sterks]

Thanks Niall. I read in the dedicated ASIF thread that they are approaching the pack through the Beaufort bite, and if this is 80 N , they are just at the edge of broken ice.
Is there a web where they are posting regular updates and their location?

[Niall Dollard]

Hi Sterks

This link will get you regular updates of their current location:

http://www.arcticmission.com/follow-arctic-mission/

[weatherdude88]

Our first candidate for NSIDC arctic sea ice extent minimum is 8.25.2017. The NSIDC arctic sea ice extent metric is up 92 thousand Kilometers squared the past two days. The JAXA sea ice extent metric also recorded an increase today. 

ftp://sidads.colorado.edu/DATASETS/NOAA/G02135/north/daily/data/N_seaice_extent_daily_v2.1.csv

[Sterks]

Hi Sterks

This link will get you regular updates of their current location:

http://www.arcticmission.com/follow-arctic-mission/

Thank you!

[Ned W]

Are first candidate for NSIDC arctic sea ice extent minimum is 8.25.2017. The NSIDC arctic sea ice extent metric is up 92 thousand Kilometers squared the past two days. The JAXA sea ice extent metric also recorded an increase today. 

ftp://sidads.colorado.edu/DATASETS/NOAA/G02135/north/daily/data/N_seaice_extent_daily_v2.1.csv

It's unlikely, but of course possible that this was the minimum.  If it turns out that way, Sterks can take a bow and the rest of us will applaud. 

 


-----------
Edited to add:  If this was in fact the minimum (which it wasn't  ) then we collectively did abysmally badly in the June/July/August JAXA polls.  # of predictions above 4.75:
June: 1 of 122
July: 1 of 122
August: 3 of 105

[Daniel B.]

Are first candidate for NSIDC arctic sea ice extent minimum is 8.25.2017. The NSIDC arctic sea ice extent metric is up 92 thousand Kilometers squared the past two days. The JAXA sea ice extent metric also recorded an increase today. 

ftp://sidads.colorado.edu/DATASETS/NOAA/G02135/north/daily/data/N_seaice_extent_daily_v2.1.csv

It's unlikely, but of course possible that this was the minimum.  If it turns out that way, Sterks can take a bow and the rest of us will applaud. 

As a reminder, 2016 saw a large rebound of 170 thousand km2 on Aug. 28, only to decline for another 10 days, reaching a minimum on Sept. 7.  By as Ned says, it is possible.

[Pavel]

Yesterday snowfalls in some parts of Siberia are month earlier than usual. Snow depth in north-west Yakutia is 10cm what is the biggest ever for 27 of August, according to a russian wheather TV

[Sterks]

The inverted dipole that has seized the Arctic will be pulling air from the Atlantic/Asia and compacting the ice North of Barents, so I would expect more extent loses. But the winds across the Arctic become Northernlies at the broken ice edge of the Pacific side, where one would expect those sudden refreezes like last year would happen first. So who knows.

[jplotinus]

The inverted dipole that has seized the Arctic will be pulling air from the Atlantic/Asia and compacting the ice North of Barents, so I would expect more extent loses. But the winds across the Arctic become Northernlies at the broken ice edge of the Pacific side, where one would expect those sudden refreezes like last year would happen first. So who knows.

Yup. "Who knows"? A second small uptick per JAXA on 8/28.  😯