Cryosphere Today 2013 Arctic SIA daily minimum: July poll

[Jim Pettit]

Another big drop--160k--as the July cliff dive continues. Four days ago, 2013 lagged 2012 by 908k; that difference has been cut to 551k, with 2012 averaging just 51k each day over the next 11.

Yesterday's decrease was the 14th of 100k or more this month; even if there are no further drops of such size, this July will have tied for most century breaks in the past 18 seasons.

Area has fallen by 2.35 million km2 so far this month, the second largest July-to-date drop on record.

[BornFromTheVoid]

We're now 4th lowest for the time of year, and below the minima of 1980, 1983 and 1986.

[sydb]

Looks like I just missed the poll deadline being on Mountain time. I'll stick to my original 1.75-2M Km^2. I still think the advecting heat will cause a lot of bottom melt giving an excellent chance of a new record.

[Richard Rathbone]

Another big drop--160k--as the July cliff dive continues. Four days ago, 2013 lagged 2012 by 908k; that difference has been cut to 551k, with 2012 averaging just 51k each day over the next 11.

A dive just as the webcams show melt ponds galore. It could be a coincidence but I think not.

[Vergent]

A dive just as the webcams show melt ponds galore. It could be a coincidence but I think not.

There were not any melt ponds last year?

[Richard Rathbone]

A dive just as the webcams show melt ponds galore. It could be a coincidence but I think not.

There were not any melt ponds last year?

Do you have a reference for last year's melt pond distribution? My impression is that the distribution is very different this year.

[ktonine]

The emergence of melt ponds in June is typical and widely considered responsible for the June "cliff".

This year's delayed melt pushed widespread melt pond formation 7 to 10 days later than recent years - and is probably a significant factor in 2013  first falling behind and now catching up to the other post-2007 years.

It's the timing of melt pond formation - not the absolute number year over year - that becomes the issue.

[Vergent]

A dive just as the webcams show melt ponds galore. It could be a coincidence but I think not.

There were not any melt ponds last year?

Do you have a reference for last year's melt pond distribution? My impression is that the distribution is very different this year.

The Canadian Ice Service publishes a weakly declouded false color composite based upon MODIS daily composites. You can tell the melt ponded area by the darker blue color. Unfortunately, this was the first from 2012, and the most recent from 2013, and they are off by a week.

We will know when the next weekly composite comes out so we can do a true comparison.
 

[BornFromTheVoid]

After a drop of 175k on the latest update, here's where we stand with regard the previous minima.



We've dropped below 8 previous minima (79, 80, 82, 83, 86, 87, 88 and 96) and are within 1 million km2 of another 19.

Despite this year being slightly less dramatic than last, with just maintaining the loss rate of the last 10 days, we could be lowest on record and below 27 out of 34 previous minima before we even reach August.

[Jmo]

It seems it could well shape up to be the fastest 6-5 million km2  ("Days spent at 5.0M") drop ever recorded? looking like it will be 8 days, which is incredibly quick based on historical records (using Jim's CT SIA chart). And this in a year apparently not overly conducive to melt...

[Jim Pettit]

It seems it could well shape up to be the fastest 6-5 million km2  ("Days spent at 5.0M") drop ever recorded? looking like it will be 8 days, which is incredibly quick based on historical records (using Jim's CT SIA chart). And this in a year apparently not overly conducive to melt...

We'll know shortly whether area fell below 5 million km2 yesterday, though I imagine it did. If so, that'll mean 2013 spent fewer days than did 2012 in the 5 million range, the 6 million range, the 7 million range, and the 9 million range. And since I'm a betting man, I'll also go out on an Arctic limb and predict that 2013 will fall from 5 million to 4 million in less time than the two full weeks it took to do so last year.

FWIW, 2013 has already lost more area this month than it did in all of July 2012. This month's to-date loss of 2.67 million km2 is the third largest on record for July, and more than half a million km2--or 23.4%-- above the 1979-2012 average.

[Wipneus]

Well its down to 4.857 (-174k) and about 360k over 2012 (same day).

[Jmo]

I think that means # days spent at 5.0M in 2013 is the lowest on record (8 days?), and the big drop into the "Days at 4.0M" category sets it up for Jim's call above.   
The simple, macrolevel nature of this reporting is powerful, and the results this year somewhat alarming.   
https://sites.google.com/site/pettitclimategraphs/sea-ice-area#asiammdpdsb
I keep hearing many saying how this is not a "favourable" melt year (for want of better terminology).  It just doesn't seem to be reflected in reality
Under current state of the ice, could CT area be being over (or under) estimated?

[werther]

Another large SIA drop of 174K continues to give an indication of the vulnerability of the sea ice. It also marks the first day of the “GAC-2013 A“ Low.
This year is in striking distance of the ’07 day 204 number and is closing in on ’11 and ’12.

Wipneus, you took account for the leap year day to get 360K above '12?

[Wipneus]

Wipneus, you took account for the leap year day to get 360K above '12?

I compare days that have the same "normal" area (6.4004054), they happen to be the same day of the year (204).

[Nightvid Cole]

CAB area as of last update (to 07/24 data) is now below all but 2 of the pre-2007 years' minima!!!

[deep octopus]

Only up 98,026 km^2 from 2007.

[werther]

CT suddenly shows the high concentration +75% pack to be smaller than that on the same day 2012.
The difference is in the order 2.1 Mkm2 (13) – 2.5 Mkm2 (12) now. Just two days ago, I got around 3 Mkm2 for ’13, concluding there still was a long way to go!

What could be the cause?

For one, the cloudiness etc still erring the sensors this year, two; a fresh wave of spread pushing large areas from 75 to 60% (reds).
Illustrative for one is the drop where MODIS tile r04c04 is situated. That counts for about 500K going 75-60!

[Nightvid Cole]

CT suddenly shows the high concentration +75% pack to be smaller than that on the same day 2012.
The difference is in the order 2.1 Mkm2 (13) – 2.5 Mkm2 (12) now. Just two days ago, I got around 3 Mkm2 for ’13, concluding there still was a long way to go!

What could be the cause?

For one, the cloudiness etc still erring the sensors this year, two; a fresh wave of spread pushing large areas from 75 to 60% (reds).
Illustrative for one is the drop where MODIS tile r04c04 is situated. That counts for about 500K going 75-60!

High bottom melt amounts eating away at the thin spots. Concentration was already dropping in that area even before the storm hit, and there was a lot of solar heat input into the upper ocean there that will continue to gnaw quite heavily at the ice bottom through September. Judging by the HYCOM SST maps, it's going to be at least as bad as last year, if not far worse.

Expect ~50% concentration in that area by month's end. First area will fall off a cliff, then extent with a lag time of a week or so.

[ChrisReynolds]

2013 is showing a second cliff, 2008 had an anomaly cliff about this time (actually two weeks later - but not in June), but 2008 didn't have one earlier. And if things carry on as they are then 2013 could see the greatest July losses by a large margin.



I'm still sceptical about 2013 beating 2012 because of the drop in anomalies that happened after last year's August storm.



But this could be a late run towards a new record.

[Vergent]

The only ice of significance is the CAB. The Pacific influx will do in the Beaufort, and probably the CAA. The Atlantic influx will do in the Kara and Laptev. There maybe a little on the ESAS, but not much. The Chukchi is going going gone.

The 2013 CAB losses just blew by 2012. With all the open water in the Eastern CAB, and the thinness of this first year ice, it would be surprising if 2013 does not beat the 2012 1.8 Mkm^2 CAB record low.

Other than the CAB, where will there be significant ice in September?

Does anyone have a reason why 2013 should not be a record low CAB?

For me, the only question is "how low can the CAB go?" Everything else must be considered seasonal ice.

I think the CAB is headed for 1.5 M or lower. Add in 200k for scraps. that puts me in the 1.5 - 1.75 bin, one lower than I voted, before the melt ponds showed up.

Vergent

[Neven]

Two very big drop days, so I guess now we'll see slower days. Wipneus has had mostly slow days, and IJIS as well. CT SIA is playing catch-up.

[Richard Rathbone]

Two very big drop days, so I guess now we'll see slower days. Wipneus has had mostly slow days, and IJIS as well. CT SIA is playing catch-up.

Its the other way around. Wipneus area plays catch up with CT area. Currently CT has the CAB down to about 3.0 while Wipneus is still up around 4.1. CT treats melt ponds as being lower concentration than Wipneus, so all these melt ponds are showing up as reductions on CT area, but don't show up as reductions in Wipneus area.

[Jim Pettit]

CT SIA has definitely made up a lot of ground over the past month, as some had predicted. 2013 was 1.12 million km2 behind 2012 less than four weeks ago; that difference now stands at a relatively paltry 408k. And over the next six days, 2012 area dropped by just 214k, so that gap should be narrowed even further, even if things do slow down a bit as Neven predicts.

The average daily area drop so far this month has been a very robust 119k km2, 25% greater than 2012's July-to-date average daily decrease of 95k. Interesting, no?

[Neven]

Two very big drop days, so I guess now we'll see slower days. Wipneus has had mostly slow days, and IJIS as well. CT SIA is playing catch-up.

Et voilà: a 6894 km2 uptick.

[Phil.]

Two very big drop days, so I guess now we'll see slower days. Wipneus has had mostly slow days, and IJIS as well. CT SIA is playing catch-up.

Et voilà: a 6894 km2 uptick.

Is this a typo?

[Nightvid Cole]

Two very big drop days, so I guess now we'll see slower days. Wipneus has had mostly slow days, and IJIS as well. CT SIA is playing catch-up.

Et voilà: a 6894 km2 uptick.

I'm almost certain that that is merely the satellite sensor being fooled by the dense clouds of the cyclone. The CT concentration map is showing a fictitious region of near 100% concentration right where the thick clouds were. Within 1-2 days it will be gone and the CT CAB area will crash once more.

[werther]

We're too fixed on the daily numbers (include me). This is a continuing process and we might best concentrate on the week trend.
On the thick clouds fooling the sensors; I checked the last comparison set '12-'13 (showing 2407) against yesterdays version (2307). And the maps show creepy small boundary gains near Banks, North Slope, an arm in the Chukchi (getting over 15% again probably in a way  like what the Barrow cam shows), the ESAS, near Frantsa Yosefa and even in the Kara Sea.
On top of that, suddenly the main above 75% body has grown, direction Fram and Chukchi sector.

I have no indication that these changes were driven by clouds. As Neven has posted several times in the past, there is large daily change in these maps. And the cause isn't always clear.

But, looking at the Kara "fast ice" stronghold; it is showing signs of a nearing break-up now. See MODIS r03c05.

[Nightvid Cole]

We're too fixed on the daily numbers (include me). This is a continuing process and we might best concentrate on the week trend.
On the thick clouds fooling the sensors; I checked the last comparison set '12-'13 (showing 2407) against yesterdays version (2307). And the maps show creepy small boundary gains near Banks, North Slope, an arm in the Chukchi (getting over 15% again probably in a way  like what the Barrow cam shows), the ESAS, near Frantsa Yosefa and even in the Kara Sea.
On top of that, suddenly the main above 75% body has grown, direction Fram and Chukchi sector.

I have no indication that these changes were driven by clouds. As Neven has posted several times in the past, there is large daily change in these maps. And the cause isn't always clear.

But, looking at the Kara "fast ice" stronghold; it is showing signs of a nearing break-up now. See MODIS r03c05.

Compare the "suddenly high concentration" areas on your map of choice to the MODIS of the same day, overlay if needed.

The cloud locations often bear a 'striking resemblance' (for want of a better term) to those patches of >75% conc.

[Neven]

Two very big drop days, so I guess now we'll see slower days. Wipneus has had mostly slow days, and IJIS as well. CT SIA is playing catch-up.

Et voilà: a 6894 km2 uptick.

Is this a typo?

I'm pretty sure that's how you write 'voilà'. 

[wili]

Back to heading south a bit today, to 4.79573, down from 4.86399 yesterday (day 205).

[wili]

Now another slight bump up to 4.827. WTF?

[Vergent]

Now another slight bump up to 4.827. WTF?

During a storm noise like this happens, it has nothing to do with the rate of ice melting. If you look at the last frames of the animation, you will see fluctuating concentration in the western CAB. Also, when there is rapid ice movement, ice can get counted twice or missed. Expect the big losses to resume.

V

[Patrick]

And another uptick, ca. 45k.

[Neven]

My guess we'll see a couple of steep drops in the coming days, given the upticks, the day-to-day changes on UB and Wipneus SIC maps, and the overall state of the ice. The melting potential is rather large, I'd say, with 6-8 weeks of melting left.

[Steven]

And another uptick, ca. 45k.

It seems that melt pond drainage in the Arctic Basin has something to do with this.  Some of the winter snow melted recently, so it's possible that we reached a peak in melt pond formation a few days ago followed by a period with more drainage.  At least that's what I think from looking at the CT animation, and comparing it with the Bremen maps which are more robust against melt ponds:

http://arctic.atmos.uiuc.edu/CT/animate.arctic.color.0.html

Interesting to see the effect in the next days of the sea ice drift caused by the recent storm.

[wili]

"It seems that melt pond drainage in the Arctic Basin has something to do with this"

That is a very smart hypothesis that makes a lot of sense to me. Thanks.

[Steven]

wili, I was not claiming originality.  Others have alluded to this a few days ago, up in this thread.

The point is that the winter snow is melting later than in the previous years.  The strength of insolation is already on its way down.  So the albedo reduction caused by the melt pond formation may come too late for a complete melting of these areas.

[Wipneus]

CT area: -5k4 (after yesterday's -9k6).

Remarkable.

[BornFromTheVoid]

CT area: -5k4 (after yesterday's -9k6).

Remarkable.

We have had similar late July/early August plateaus before.

Some recent examples.
2005 only dropped 64k from day 207 to day 213.
2001 only dropped 35k from day 202 to day 209

They seem to happen roughly twice per decade

[werther]

Being into NCEP/NCAR today, I dumped your data in, BftV:




Remarkable! Both periods extremely low area loss, opposing atmospheric configuration.
Wonder if there are other properties involved?

[BornFromTheVoid]

Being into NCEP/NCAR today, I dumped your data in, BftV:




Remarkable! Both periods extremely low area loss, opposing atmospheric configuration.
Wonder if there are other properties involved?

A few other examples my help.

1997: Day 209 to day 216 (-34k)
1993: Day 207 to day 213 (-37k)

Other examples include 1987, 1985 and 1979, but I guess we're talking about completely different ice characteristics back then, though I suppose that's very much true even in the 90s!

[Richard Rathbone]

CT area bounces at the bottom of its cliffs.

The webcams have recently shown the North Pool form and drain, and the way CT area changes has followed that, suggesting that similar pools were forming and draining at the same time across the CAB.

Cliff as the Pool formed, and bounce when it drained.

[ChrisReynolds]

Richard,

It's worth looking at, 2012 didn't bounce at all (dead cat style), it just flopped there with anomalies level. 2010 shows a big rise some weeks after the cliff, 2011 bounces back a bit. But I agree this is probably due to pond formation then rapid drainage due to the ice at end of July being thinner than in early June (the typical time for a cliff - 2007, 2010, 2011, 2013, 2013).

Within two weeks we've had a cliff, then a bounce that's virtually wiped the cliff out. 

I can post a plot of anomalies, but as I'll be uploading graphs tomorrow for my July Status post I don't want to tonight.

[Richard Rathbone]

Richard,

It's worth looking at, 2012 didn't bounce at all (dead cat style), it just flopped there with anomalies level. 2010 shows a big rise some weeks after the cliff, 2011 bounces back a bit. But I agree this is probably due to pond formation then rapid drainage due to the ice at end of July being thinner than in early June (the typical time for a cliff - 2007, 2010, 2011, 2013, 2013).

Within two weeks we've had a cliff, then a bounce that's virtually wiped the cliff out. 

I can post a plot of anomalies, but as I'll be uploading graphs tomorrow for my July Status post I don't want to tonight.

I interpret the elasticity of the bounce as an indicator of how synchronised the draining is. Dead cat bounce if its a wide spread, strong bounce if its a narrow spread. Also if there is vigorous surface melt continuing then there will be new ponds forming as old ones drain and again it will be the dead cat bounce. Unfortunately CT only show 365 days on their regional plots and I'd really like to see about  a year and a half to get a good idea of how the development of the melt compares from last year to this year. Your plots are good for that for the whole arctic, but the bounce can be clearer on the regional anomaly plots when it happens at different times in different regions.

[ChrisReynolds]

Richard,

Having had the afternoon off work, rested and then got to really look at this...

I think the 17 - 18 July cliff is probably melt pond related as that's the quickest way to lose CT Area quickly. And the plateau in CT Area that leads to the 'bounce' in anomalies is probably due to divergent movement of the pack masking area loss. More detail in my latest blog post.
http://dosbat.blogspot.co.uk/2013/08/july-status-cryosphere-today-area.html

Sorry not to go into detail here, but it's late and I want some downtime.

[Richard Rathbone]

A link to detail is just as good, provided its not paywalled. I'd have thought that convergent movement would be better at masking area loss than divergent, but I'll read the detail and see if that changes my mind.

Edit.
I've read it and am not convinced. I think the other products might be fooled into treating divergence as area increase, but not CT. So if you were making that argument for Wipneus'  (AMSR2, ASI) area, I'd go along with it, but for CT (F17, NASA-TEAM) area I find divergence resulting in cracking which drains melt pools more plausible. Or a temperature drop resulting in lower top melt and no refill for slow draining pools or refreezing.

[ChrisReynolds]

Richard,

As I understand it Wipneus's work is at a higher resolution that CT Area seems to be. So I still think CT Area could be 'fooled' given the fragmented nature of the pack at this time of year.

Peter Ellis also disagrees with me. He thinks it's melt ponds too. To be clear I'm not saying melt ponds aren't a factor, just that I suspect divergence is a stronger factor. Certainly Beaufort and Chukchi show a cessation of ice edge retreat, I think I see a slowing

Peter has pointed out re-freeze, which I think he's seen on webcam. It is true that this recent period has seen a cooling, in the last 48hrs temperatures at 2012G have dropped to around -4degC which will cause refreeze. Using buoy air temperature, here are average temperatures for two periods:

2012G
pre 0.580119048
post -0.405806452

2012 H
pre 0.771301775
post -0.471190476

Pre is 17 to 23 July
Post is 24 to 30 July
To match periods on SLP plots in my blog post.

This is a small temperature deviation that doesn't seem to me to be likely to cause massive re-freeze.

In casting about trying to find a stronger basis for claiming divergence than just my fallible eye, the following isn't really satisfactory due to the length of time post 24 July. However I've got data from MASIE and applied Excel's 'SLOPE' function (slope of trend) to:

A) day 184 (earliest data) to day 204 (23 July 2013) and
B) day 205 (24 July 2013) to day 213 (latest data).

 (1) Beaufort   -5269.38   3547.90   -0.67
 (2) Chukchi    -8292.00   -3255.48   0.39
 (3) E_Siberi    -4190.58   -4826.53   1.15
 (4) Laptev     -8507.86   -5031.61   0.59
 (5) Kara       -19987.90 -9888.39 0.49

The first column is the slope (extent/day) of period A, the second is the slope of period B, and the final is the ratio of the two.

For all except the East Siberian Sea the slope during period B is much less than the slope for period A, during period B the weather was what I am claiming would lead to divergence. It's no more than suggestive but it's the best I can come up with.

[Richard Rathbone]

Take 10 pixels of arctic of which 7 are completely ice and 3 completely water.

Both CT and Wipneus will call that 7 pixels at 100% + 3 at 0%.

Now diverge the ice so that its equally spread out amongst the 10 pixels. I'm pretty sure that of the two methods CT is the most likely to say thats 30% which looks like water and call it 10 pixels at 70%, while and Wipneus is the one most likely to say that it looks like 30% water but 2/3 of it is probably melt ponds, and call it 90%. CT area isn't corrected much (if at all) for melt ponds, but Wipneus area is, and that makes it more vulnerable than CT to mistaking divergence for melt ponds.

Near the coast, the resolution could make a difference, but out in the CAB it won't and I don't think either of them are fooled to any serious degree by divergence. However, CT is protecting itself against that error by the way it treats melt ponds and Wipneus area isn't. CT area drops much faster. When one method is calls the CAB average as 95% and the other calls it as 70%, its the one at 95% thats at risk of overpredicting the area after divergence.

[ktonine]

If you read the studies done on the different SIC algorithms, they can differ by 20% or more depending on atmospherics, condition of the ice, etc.  In one area of the arctic on a given day the Bootstrap algorithm might read 20% higher than the NASA Team algorithm; in a different location over the same time period it might be 20% lower.

The uncertainties are so large during the melt period that even large differences are not significant - they're to be expected. 

Trying to compare SIE or SIA for:
A) an ice pack with marginal ice zones limited mostly to the perimeter of the pack; to
B) an ice pack with areas of open water spread throughout
very difficult.

Note, for instance, that PIOMAS states its uncertainties are for October:

The uncertainty of the  monthly averaged ice volume anomaly is estimated as ±0.75  10³ km3. Total volume uncertainties are larger than those for the anomaly because model biases are removed when calculating the anomalies. The uncertainty for October total ice volume is estimated to be  ±1.35 10³ km3 .

We can infer that the volume uncertainties during the melt season are even larger.