2019 vs 2012

[Sterks]

Why are you commenting if you don't understand the theory?

I know enough about ENSO to correctly identify which years were el nino influenced and which were not.

It is not your theory, thus you are completely incorrect. I stated clearly you must look at a two-year period, you responded discussing only a one-year period making your response irrelevant.

And your statement is an opinion which you state as fact, so doubly incorrect.

Of couse Niños elevate the tide of AGW and I agree some direct impact can be observed (such as 2016 a year with no June sun, a stormy July and a GACCY August, where shear warmth of the NH post Niño was also noticeable pulling a Nr 2 in area). I can also understand for the same reason 2000 low post 98.

But the rest is weak stuff. It’s there, it affects sea icesomehow, but the numerology you manage I don’t follow. Anyway, if you look at a period of two years, the probability of catching a Niño is pretty high.
Drop it here anyway.

[binntho]

Someone said there is a way to mute people... I don't like jerks.

Go to Profile, Modify Profile, Buddies/Ignore List, Edit Ignore List. It's not easy to find! I've only ever ignored one poster, and I certainly won't ignore you Killian but feel free to ignore me whenever you feel like it!

[Killian]

8/09/2012 = 5.04M km sq. after a large 150k km sq fall.
8/08/2019 = Stands at 5.33M. Loss of > 290k km sq. needed to set a daily record for this date.

...Call it 105k+/-10k.

And we ended up with... 5.239 or a drop of 95k. Nice. Within range. But, that's the 3rd time I didn't go with my gut and ended up wider of the mark than I would have. Ah... trusting your intuition is a hard thing to learn to do. 

8/10/2012 = 4.94M km sq. after a 100k drop, the last of the GAC -100+ days.
8/09/2019 = Stands at 5.24M. Loss of > 300k km sq. needed to set a daily record for this date.

Call it 95k+/-10k.

Analysis: The loose ice in the Beaufort gyre just doesn't move much of late. I suspect this is due to the winds being opposite of the gyre currents, so call that a draw. From the mid-Alaskan shore to the Strait it's one side of a cyclone and from there further into the Chukchi it's the other side of the cyclone. Call it a draw. Winds comeing off of Siberia over the western half of the ESS and over the Laptev are toward the ice pack, and stronger than some recent days. Winds headed out over Svalbard into the Barents are going to be strong, but the ice there doesn't repsond much to winds, either. At least not recently. And over the Kara they'll be toward the CAB.

So... more momentum and with higher winds over the most mobile ice off of parts of Siberia, we get a nice little sub-100k day of ASIE, noted above. Could be a pretty low day, so don't be surprised if I'm high by quite a bit, particularly if the Atlantic side gets into motion.

The 11th looks really interesting...

-------------------------------

Daily Changes Needed to Exceed 2012 low on Aug. 10.
(Related to effect of GAC and it's import vs. 2019's melt cycle.)

8/10/2012 stood at 4.94M km sq.
2019 needs a drop of > 299k km sq. for a record low on this post-GAC date. Final day! Thank the gods!  This number does tell us something of interest, however, if we bring 2016 into the picture. The magnitude of 2019 so far is clear when we consider '19 is +200 to 2012, but -420 to '16. So far, as if we didn't know, '19 really is unusual because it has been so normal. The only outstanding element so far is the sheer amount of heat in the system; two straight all-time hottest months and a June with record insolation, I believe, and the scientific finding that June insolation correlates well with Sept. minima.

That heat's not an Arctic phenomenon, it's very much a global one. But imagine if the huge outlier that is 2012 had never happened and 2019 came along? We'd be treating it like the end of all ice, like some thought back in '12.

I believe 2016 pumped so much energy out of the oceans and to the surface that it initiated/pushed a number of feedbacks into gear/higher gear. For '17 and '18 I think maybe this exhibited in the Arctic primarily as low winter ASI accumulation.

415ppm, folks, is looking deadly.

Daily Changes Needed to Exceed 2012 Record Low on Sept. 15.
(Related to comparison of 2012 vs 2019's melt cycle.)

9/15/2012 stood at 3.18M km sq. on this date.
2019 needs an average daily drop of > 55.65k km sq. for a record Sept. low. (37 days)

[binntho]

<snip>
But imagine if the huge outlier that is 2012 had never happened and 2019 came along? We'd be treating it like the end of all ice, like some thought back in '12.
</snip>

So true!

[SimonF92]

2012 has the advantage peripherally, 2019 has the advantage toward the core

[Random_Weather]

For some Folks not uninteresting


Data: ftp://ftp.remss.com/sst/daily/mw_ir/v05.0/netcdf/ (need for registry)
Red: 2019 warmer
Blue: 2019 cooler


https://seaice.uni-bremen.de/data/amsr2/asi_daygrid_swath/n6250/netcdf/

Red: 2019 more SIC
Blue: 2019 less SIC

[binntho]

For some Folks not uninteresting

For me very interesting!

[Glen Koehler]

Thanks Random, very interesting views

[D-Penguin]

QUOTE from Killian on: Today at 10:22:24 AM
"9/15/2012 stood at 3.18M km sq. on this date.
2019 needs an average daily drop of > 55.65k km sq. for a record Sept. low. (37 days)"

If the date of minimum extent for 2019 is extended by 8 days compared to 2012 because of the additional heat energy in the Arctic system:-

2019 would need an average daily drop of > 45.76k km sq. for a record September low (45 days)

Presumably the soot deposits from current forest fires in the Arctic region will increase to some degree top melting during the period of diminishing insolation until the September equinox adding to side and bottom melt.

Whatever the outcome of the present melting season, the onset of the 2019/20 re-freezing season will be very interesting.

[Neven]

For some Folks not uninteresting

Very nice, RW! I wish I knew how to do that kind of stuff.

[Random_Weather]

Neven,

Thats very easy, you just need to download: https://www.giss.nasa.gov/tools/panoply/

I wonder that most people here dont know about?

[DrTskoul]

Actually first mentioned in 2013: https://forum.arctic-sea-ice.net/index.php/topic,384.msg16282.html#msg16282 version 3.

I guess after a while these useful links get buried/forgotten

[Killian]

QUOTE from Killian on: Today at 10:22:24 AM
"9/15/2012 stood at 3.18M km sq. on this date.
2019 needs an average daily drop of > 55.65k km sq. for a record Sept. low. (37 days)"

If the date of minimum extent for 2019 is extended by 8 days compared to 2012 because of the additional heat energy in the Arctic system:-

2019 would need an average daily drop of > 45.76k km sq. for a record September low (45 days)

No way to know how long the melt season will go, of course. I base that number on the first day JAXA reaches 3.18 bc the comparison is with the 2012 low, but the next day is likely a few k km lower, though also 3.18 on their chart. Anything from the 15th to 17th would likely be ok. And, sure, if it goes on another week or two, that changes things quite a bit.

Note, however, that bc most of the melt happens before Sept, the daily average needs to be about 20k per day remaining by Sept. 1 to be close to a new record.

[Killian]

2012 has the advantage peripherally, 2019 has the advantage toward the core

Suggests higher chance of record low area/volume than extent, maybe?

[Killian]

8/10/2012 = 4.94M km sq. after a 100k drop, the last of the GAC -100+ days.
8/09/2019 = Stands at 5.24M. Loss of > 300k km sq. needed to set a daily record for this date.

Call it 95k+/-10k.

Or not. -130k. 25k beyond my range ain't bad, is it? LOL...

The 11th looks really interesting...

At least that is still true...

8/11/2012 = 4.89M km sq. after a 50k drop. Exhausted after the GAC, I guess.
8/10/2019 = Stands at 5.11M. Loss of > 220k km sq. needed to set a daily record for this date.

2019's 130k drop helped tighten things up a bit. Another 100k+ day would go a long way towards making this look like a horse race again.

Analysis:

There are so many cyclonic forces going on right now, it's impossible to sort out what in the world is going to happen. To make it even more ambiguous, despite all the cyclones/anti-cyclones, most of the wind direction at the surface is going *across* the ice edges for the most part for most of the 11th - according to NullSchool. What in the world does wind across the ice edge do, tickle it to death? The exception for the day seems to be the area of the Fram Strait and Svalbard, so we may get some export. I don't think that will affect etent much.

The one thing that might keep the loss over 100k is the surface heat, though not esceptionally high, is over a very large area, with, as others have noted, particular incursion over the Laptev where there isn't much ice near shore, but a lot of mushy stuff. Coming with the winds and heat is... soot. Holy melt, Batman, if that stuff gets onto the ice! Of course, if it stays up in the air it will block sunlight and maybe cool the area it's over.

Let's go with momentum and heat: 110k+/-10k.

Zero confidence in this prediction.

-------------------------------

Daily Changes Needed to Exceed 2012 low on Aug. 10.
(Related to effect of GAC and it's import vs. 2019's melt cycle.)

8/10/2012 stood at 4.94M km sq.
2019 needs a drop of > 299k km sq. for a record low on this post-GAC date. Final day!

Ding-dong the GAC is dead, the GAC is dead, the GAC is dead!

Got a -130k drop and ended up 170k higher. I said a good while back if 2019 was below or near 2012 on this date there'd be a good chance of a new record. 170k higher adds a bit less than 5k to the daily average melt needed. Not small, but within reach.

Anywho... that's the end of this run.

Daily Changes Needed to Exceed 2012 Record Low on Sept. 15.
(Related to comparison of 2012 vs 2019's melt cycle.)

9/15/2012 stood at 3.18M km sq. on this date.
2019 needs an average daily drop of > 53.61k km sq. for a record Sept. low. (36 days)

For perspective, on Sept. 1 the average melt needed to the 15th is about 20k. 2019 needs to eat up 23.61k/day over the next 21 days. Well, hell, if Sept can chew up the equivalent of 20k/day for 15 days, August can chew up 23 or 24 per day over 21.

I think we have ourselves a horse race.

[Killian]

Anyway, if you look at a period of two years, the probability of catching a Niño is pretty high.
Drop it here anyway.

As I've mentioned, and so needs to be checked.

[D-Penguin]

8 or 10 days added to the average minimum date could be the deciding factor as to whether or not 2019 produces a new record minimum.

If 2018/19 and 2020 show progressive extensions of the minimum date, 20-20 vision is not required to see what happens next! SSTs and advection regardless of solar insolation in the Arctic.

What happens in the Arctic does not stay in the Arctic.
More importantly:-
What happens outside the Arctic does not stay outside the Arctic!!!

[AndyW]

The interesting feature at the moment is that the tongue of ice into the east siberian sea, which had older ice due to failure to melt last few years,  has now almost melted out it seems.

Big loses extent wise always favourable to the Russian side.   

[uniquorn]

Not sure it this is helpful but here attempting to overlay unihamburg amsr2-uhh 2012 onto 2019, aug1-10.
2012 lower ice concentration has been colorised purple and is overlayed at 60% transparent to allow 2019 ice edge to show through. 2012 0% concentration (usually dark blue) has been set to fully transparent.
7days/sec is quite fast but helps to identify which year is which.

Concentration shouldn't be seen as meaningful as 2012 has been enhanced somewhat to bring out the ice edge and the overlay of white on white distorts concentration nearer the pole.
tech note: edge detect was too messy in this case

[Sterks]

Not sure it this is helpful but here attempting to overlay unihamburg amsr2-uhh 2012 onto 2019, aug1-10.
2012 lower ice concentration has been colorised purple and is overlayed at 60% transparent to allow 2019 ice edge to show through. 2012 0% concentration (usually dark blue) has been set to fully transparent.
7days/sec is quite fast but helps to identify which year is which.

Concentration shouldn't be seen as meaningful as 2012 has been enhanced somewhat to bring out the ice edge and the overlay of white on white distorts concentration nearer the pole.
tech note: edge detect was too messy in this case

This week, while the pacific edge was being severely affected by the GAC in 2012, it has been an immobile and relatively cold week for this edge in 2019. Extent has continued dropping but note that 2012 started higher in extent.
Next week is the one where 2019 won’t really follow the pace, although it will follow a vigorous pace for the time of the year given the weather.

[TeaPotty]

Not sure it this is helpful but here attempting to overlay unihamburg amsr2-uhh 2012 onto 2019, aug1-10.
2012 lower ice concentration has been colorised purple and is overlayed at 60% transparent to allow 2019 ice edge to show through. 2012 0% concentration (usually dark blue) has been set to fully transparent.
7days/sec is quite fast but helps to identify which year is which.

Concentration shouldn't be seen as meaningful as 2012 has been enhanced somewhat to bring out the ice edge and the overlay of white on white distorts concentration nearer the pole.
tech note: edge detect was too messy in this case

This week, while the pacific edge was being severely affected by the GAC in 2012, it has been an immobile and relatively cold week for this edge in 2019. Extent has continued dropping but note that 2012 started higher in extent.
Next week is the one where 2019 won’t really follow the pace, although it will follow a vigorous pace for the time of the year given the weather.

Sterks is on a mission!
Even as 2019 and 2012 remain tightly competitive, lol 😂

[Sterks]

Not sure it this is helpful but here attempting to overlay unihamburg amsr2-uhh 2012 onto 2019, aug1-10.
2012 lower ice concentration has been colorised purple and is overlayed at 60% transparent to allow 2019 ice edge to show through. 2012 0% concentration (usually dark blue) has been set to fully transparent.
7days/sec is quite fast but helps to identify which year is which.

Concentration shouldn't be seen as meaningful as 2012 has been enhanced somewhat to bring out the ice edge and the overlay of white on white distorts concentration nearer the pole.
tech note: edge detect was too messy in this case

This week, while the pacific edge was being severely affected by the GAC in 2012, it has been an immobile and relatively cold week for this edge in 2019. Extent has continued dropping but note that 2012 started higher in extent.
Next week is the one where 2019 won’t really follow the pace, although it will follow a vigorous pace for the time of the year given the weather.

Sterks is on a mission!
Even as 2019 and 2012 remain tightly competitive, lol 😂

Yeah, whatever. Go back to lurking if you don't really have anything to positively contribute with.

[uniquorn]

This week, while the pacific edge was being severely affected by the GAC in 2012, it has been an immobile and relatively cold week for this edge in 2019. Extent has continued dropping but note that 2012 started higher in extent.
Next week is the one where 2019 won’t really follow the pace, although it will follow a vigorous pace for the time of the year given the weather.

As a relative newcomer to the forum, with 2012's reputation, it's amazing to see how much ice is left in the chukchi.

[HapHazard]

Looking at the most recent extent graphs in the data thread, using a cycling analogy: It kinda looks like the 2012 GAC amounts to a breakaway while 2019 is the peloton, going steady. Will 2019 chase it down? Did 2012 break too early & lose steam? Very interesting!

Here's my graphical scientific representation of 1 potential outcome (exaggerated for clarity):

[BenB]

SST update for 11 August 2012 vs 2019. The next few days will see warm, humid air and warm waters push north into the Laptev sector of the CAB. Meanwhile, northerlies will transport/export ice into the Chukchi, on one side of the Arctic, and into the Barents/Greenland seas on the other. I would expect area declines to pick up, but extent may hold up better initially. We could also see more areas of open water near the Pole as the ice is pushed apart.

[peterlvmeng]

SST update for 11 August 2012 vs 2019. The next few days will see warm, humid air and warm waters push north into the Laptev sector of the CAB. Meanwhile, northerlies will transport/export ice into the Chukchi, on one side of the Arctic, and into the Barents/Greenland seas on the other. I would expect area declines to pick up, but extent may hold up better initially. We could also see more areas of open water near the Pole as the ice is pushed apart.

Good comparison!

[philopek]

Looking at the most recent extent graphs in the data thread, using a cycling analogy: It kinda looks like the 2012 GAC amounts to a breakaway while 2019 is the peloton, going steady. Will 2019 chase it down? Did 2012 break too early & lose steam? Very interesting!

Here's my graphical scientific representation of 1 potential outcome (exaggerated for clarity):

This was what one of the guests posted back in spring and I think the steady decline in 2019 could indeed be key feature of this season.

Also in contrast to other opinions i think there is a lot of dispersed ice left to melt and the easy ice is not gone yet by far.

Weather, as it always has been, will define the outcome when it comes to 1st or 2nd while I think that for a 3rd place an extreme weather on the ice-friendly side would be needed and that's simply not in sight yet and in 2 weeks we are gonna see the 2016 low valu if things continue just a little longer like they do now.

[Sterks]

Yes, there's a lot of ice to go, on the order of 1M km2 to maybe 1.5M km2. That is a lot.
But not 2M as in 2012.

[petm]

It says a lot that we're in the 2nd week of August and this is still a very open question, even without any major weather event (or at least no GAC).

[JUST ICE]

Has anyone given any consideration to the fact that 2012 was a leap year?. Because of the additional "leap day" we are, in essence, comparing 2019's todays with 2012's tomorrows. In the race to the bottom 2012 has been given a one day head start but it's just an illusion of our imperfect calendar.

I know NSIDC has considered this as their Chartic graph shows 2012 dats points (and those of all leap years) offset by one day. After leap day, that is.

[Glen Koehler]

Post 1 of 2
     As per request I added an alternate version of the 2019 projected minimum based on extending the melt season to date anomaly vs. 2011-2018 average.  Thus if 2019 reduction in that measure to date was 5% higher than in 2011-2018, then the remaining projected losses to minimum would also be 5% higher instead of using the 2011-2018 average losses from latest observation date to annual minimum.









FWIW - August  12,  2019 has more ice on the Russian side which is surprising with the persistent high temperatures in coastal Siberia earlier this summer.  But 2019 has less ice in core of what used vto the MYI triangular fortress from norther Greenland  to North Pole and Down to Ellesmere Island.  Even though 2019 does not look to break the 2012 records for Extent and Area, it seems that the condition of the CAB is substantially weaker.  Also the high concentration/low spread CAB ice in 2019 is perched closer to the Fram Strait exit ramp.  I guess too late this year for it to get shuttled out of the Arctic and who knows where that ice will be next June.   But that location seems to be more bad news for ASI.



EDIT -- Dang - forgot to update 2019 projected min Extent and Areas based on 2011-2018 remaining losses in the tables.  Can't fix right now, away from computer.  Values in text section above tables are the correct ones.  The yellow shaded Extent in table should be 3.80 & 114%.  Yellow shaded AREA should 2.60 & 117%.
   Thus, since August 4 Extent got closer to 2012, but Area fell farther behind.


 

[Glen Koehler]

Post 2 of 2





The 2019 average index calculated by current melt season anomalies is marked with * and is shaded green.  The projected 2019 minimum Extent, Area, Volume, and Thickness values based on current melt season anomalies appear to less reliable predictors than the 2011-2018 average remaining losses.  The average ratio for based on the 2019 melt season anomaly estimates is included in the table to show relative position, but to avoid assigning two ordinal ranks to a single year, only the 2019 minimums based on 2011-2018 average remaining losses are assigned a rank vs. other years.  Likewise, the color coding for 1st, 2nd, etc. lowest values does not include the 2019 melt season anomaly values.

EDIT -- Dang part 2.  Away from computer and can,t fix right now.  Forgot to edit the yellow shaded 2019 values based on 2011-2018 average remaining losses in table below.  Extent ratio should be 1.14 and Area ratio should 1.17.  COMBINED AVERAGE stayed at 1.08. 


The index value for each year is based on its average of ratios to the minimum value observed in 1979-2019 for each measure, not by the average ordinal ranking.

Observations/Speculation: 
-- 2012 stands alone as the lowest overall with 1st place ordinal rankings for all four measures.  (Technically, 2019 minimum thickness was lower, but only by insignificant margin.) 

-- 2019 is second to lowest, with one 1st, and three 2nd place rankings.  The estimated 2019 minimums for Extent and Area are substantially larger than for 2012.  The 2019 and 2012 minimums for Volume and Thickness are similar.
      The *2019 minimum Extent, Volume, and Thickness estimates based on the current season anomaly are lower than the estimate based on 2011-2018 average losses from latest observation date to minimum.  But the Area estimate is higher.  While 2019 still finishes second to 2012 for the average ratio, the gap is much narrower. 

-- 2016 and 2011 are close to each other for 3nd lowest overall ranking, followed by 2010 and 2017 in a tie for 4th place. 

-- The sequential rankings of 2010 (#5), 2011 (#4), and 2012 (#1) suggest that the 2012 minimum record may have been the culmination of a three year sequence of predisposing bad melt years vs. being entirely due to conditions in 2012. 
 
-- Except for 2007, there is a high degree of congruence between the 2D measures (Extent, Area) and the 3D measures (Volume, Thickness). 

-- Nine of the 10 lowest ranking years have been in the last decade.  All of the top ten ranks, except 2007 at #9, have occurred in 2010-2019.  At rank #12, 2014 is the only year in the most recent decade to not be in the top ten for lowest index overall value. 

[Phil.]

Has anyone given any consideration to the fact that 2012 was a leap year?. Because of the additional "leap day" we are, in essence, comparing 2019's todays with 2012's tomorrows. In the race to the bottom 2012 has been given a one day head start but it's just an illusion of our imperfect calendar.

I know NSIDC has considered this as their Chartic graph shows 2012 dats points (and those of all leap years) offset by one day. After leap day, that is.

It's better to use the Chartic approach and plot with respect to 'day of year' rather than calendar date.

[sja45uk]

It's better to use the Chartic approach and plot with respect to 'day of year' rather than calendar date.

Surely that just changes when the day jump occurs to the year after the leap year instead of the year of the leap year, and any use of day numbers as time axis labels would be bad.

[petm]

True about leap years, except... What's actually happening is the days are shifting relative to the astronomical calendar (e.g. summer solstice) every year. Leap years just reset this drift, in effect jumping back about 3/4 of a day relative to the previous year. Actually it's even worse, since to stay aligned with the astronomical calendar, the leap year has to be skipped 3 out of every 4 centuries.

I have the feeling that using day of year vs. date just makes it easier to plot Feb. 29th.

https://en.wikipedia.org/wiki/Summer_solstice

[gerontocrat]

LEAP YEARS -

I simply disappeared 29 Feb from the data from 1980 onwards.
Not a big deal as Feb 29 is at the end of the season so area / extent daily changes are minimal. Though in one year one of the minima or maxima was on Feb 29. Too bad.

In 2020 29 Feb will not exist in my data. So a day's holiday for me on March 1.

But 2020 will be a pain from March 1 onwards as I use the date to count the number of days so far in millions of "=offset(expression)" formulae. My spreadsheets, like topsy, just grew, and I failed to make a standard parameter sheet so one change would fix everything.

Shocking to be so cavalier.

[petm]

You murdered a whole day!?! :-0

[gerontocrat]

You murdered a whole day!?! :-0

Mass murderer!

One day every 4 years . 9 to date. Double figures next year!

[Killian]

Things have really calmed down compared to 2012, as we see on the charts, but here are the recent trends in numbers.

Current JAXA '19 - '12 extent: @ +370k.

Average JAXA extent change for August: @ -74k.
Average JAXA extent change for August prior to last seven days: @ -91k.
Average JAXA extent change over the last seven days: @ -44k.

Strong downward trend.

Same numbers for 2012:
Average JAXA extent change for August: @ -102k.
Average JAXA extent change for August prior to last seven days: @ -114k.
Average JAXA extent change over the last seven days: @ -81k.


Average JAXA daily extent change required to equal 2012 low: @ -51.5k.
(Average JAXA daily extent change required to equal 2016 low: @ -25.79k.)

Average JAXA daily extent change for Sept. 2012: @ -20k.
Things really slow down in September, so we'd need to see +11.5k more melt per day than we've seen the last week for the rest of the month. That +370k has to be gone by September, pretty much.

------------------------------------------------------------------

NSIDC Data

Current NSIDC '19 - '12 extent: @ +421k. (+330k 5-Day Avg.)
(Numbers are raw daily numbers, no averaging.)

Average NSIDC extent change for August: @ -67k.
Average NSIDC extent change for August prior to last seven days: @ -86k.
Average NSIDC extent change over the last seven days: @ -33k.

Strong downward trend here, too.

Same numbers for 2012:
Average NSIDC extent change for August: @ -102k.
Average NSIDC extent change for August prior to last seven days: @ -114k.
Average NSIDC extent change over the last seven days: @ -81k.

(Wow. I thought I'd made a copy/paste error, but no: Those NSIDC and JAXA 2012 numbers are that similar.)

Average NSIDC daily extent change required to equal 2012 low: @ -53.6k.
(Average NSIDC daily extent change required to equal 2016 low: @ -35.7k.)

[AndyW]

The Atlantic side near Spitzbergen shows more ice than the last couple of years currently.

In regards to 2012 I don't think 2019 is comparable at all due to the unusual weather in 2012  which 2019 has not had. 2019 is now following the general trend for August (so far) but will still be a very low year.

Personally I always ignore 2007 and 2012 as outliers and just look at the general trend.

Once the Arctic gets down to 2012 ice extent levels  repetitively, like it has compared to 2007 now, then that is worth noting.

[seaice.de]

Difference of August 22nd 2012 and 2019 (AMSR2 data so far until 5:18 UTC).