Antarctic Sea Ice Extent and Area

[kassy]

So we are 11 days into the melt season with quite a head start. Slowly onto 2024...

[kiwichick16]

its still theoretically possible for area gains , as in 2007 and 2014, but if we are past the peak , that headstart is over 600,000 sq kms

[kiwichick16]

2016 , the year in 2nd place currently , was in the last year of an El Nino event , while we are just going into the next El Nino now .

2023 on track to be the hottest year on record .

https://www.abc.net.au/news/2023-09-11/global-temperatures-pass-1-5c-above-pre-industrial-levels/102836304

[gerontocrat]

JAXA ANTARCTIC SEA ICE EXTENT:  16,891,361 KM2 as at 17-Sep-2023

- Extent gain on this day 6k, 7 k more than the average loss on this day (of the last 10 years) of 1k,
- Extent gain from minimum on this date is 14.94 million km2, 0.72 million km2, (4.6%) less than the 10 year average of 15.66 million km2.

- Extent is 1st lowest in the 45 year satellite record, which makes daily sea ice extent lowest for 212 days this year

- Extent is  816 k LESS than 2002
- Extent is  956 k LESS than 2017
- Extent is  927 k LESS than 2018
- Extent is  1,184 k LESS than 2022
- Extent is  1,712 k LESS than the 1980's Average
- Extent is  1,710 k LESS than the 2010's Average

- On average  99.3% of ice gain from minimum to maximum done, and 5 days to maximum

Projections. (Table JAXA-Ant1)

Average remaining freeze (of the last 10 years) would produce a maximum in Sept 2023 of 16.99 million km2, 1.01 million km2 below the 2002 record low maximum of 18.01 million km2, which would be 1st lowest in the satellite record.
___________________________________________________________
The question that still remains is whether the 2023 maximum has been reached at 16.99 million km2 on the 7th September
___________________________________________________________
Click on image for full-size

[gerontocrat]

Nico Sun's daily sea ice area shows a 19k drop, the 4th drop in a row. That's enough for me. Time to open the melting season thread - which gives me a quandary - see next post(s)

[gerontocrat]

Below are the unadjusted results for the melting season from NSIDC Antarctic 5 day trailing average data
The first quandary that the melting season gives me is impossible results, i.e. for some years negative Antarctic Sea Ice Area at the minimum. That can be fixed by making the lowest possible result zero.

The 2nd quandary is that in theory all but one of the Antarctic seas can approach close to zero sea ice. The exception is the Weddell Sea due to its unique position on the continent, as was explained to me by Nico Sun some time ago. (I will post more about this when I have made some additional analysis of sea ice area in the Weddell Sea)

Which just goes to show that, (in common with all models), when the data moves outside model limits, my little model collapses.

NSIDC ANTARCTIC SEA ICE AREA (5 day trailing average):  12,972,425 KM2 as at 17-Sep-2023

- Area loss on this day 71k, which is 61 k more than the average loss on this day (of the last 10 years) of 10k,
- AREA loss from maximum on this date is 0.40 million km2, 0.04 million km2, (10%) less than the 10 year average of 0.44 million km2.

- AREA is 1st lowest in the 45 year satellite record, with daily sea ice area lowest for the day for 197 days this year

- AREA is  740 k LESS than 2016
- AREA is  886 k LESS than 2017
- AREA is  2,328 k LESS than 2006
- Area is  1,141 k LESS than 2022
- AREA is  1,548 k (10.7%) LESS than the 1980's Average
- AREA is  1,658 k (11.3%) LESS than the 2010's Average

- On average 3.3% of ice loss  from maximum to minimum done, and on average about 160 days to minimum

Projections. (Table NSIDC ANTARCTIC AREA-1)

Average remaining melt (of the last 10 years) would produce a sea ice area minimum in Feb 2023 of 0.15 million km2, 1.11 million km2 below the 2022 record low minimum of 1.23 million km2.
_______________________________________________________________
Click on images to enlarge

[gerontocrat]

NSIDC ANTARCTIC SEA ICE AREA (5 day trailing average): 2023 MAXIMUM (provisional)

The Sept '23 maximum at 13.37 million km2 is about one week earlier than average, not a good sign for the melting season.

The maximum is lower than the previous record low max in 2017 by 0.79 million km2, below the longterm linear trend by 1.56 million km2, and below the 2022 maximum by 0.90 million km2.

click graph to enlarge, click table twice for fullsize

[kassy]

The first quandary

For which years? Recent ones or older years? How many out of the set? Adjusting to zero would make it look neater but it is also interesting to see where or how much the simple model breaks.

The second...they can in theory but since you report on the numbers that will only be a problem if they approach that in the data and if that happens the model is not really our main concern?

Or maybe the problem is the fact that the historic years melt ice which is now not there so another correction would be to deduct the change in starting amount from them. It would be slightly better then just capping the end result at zero if still not very precise.

I think that it would be best to run things as usual assuming not too much breaks down at first.

[Renerpho]

For which years? Recent ones or older years? How many out of the set? Adjusting to zero would make it look neater but it is also interesting to see where or how much the simple model breaks.

2005, 2006, 2007, 2009, 2012, 2014, 2020, 2021, 2022 (9 out of the last 18 years), with the lowest being -0.70 million km² (projected from 2014).

Gero is giving details in the table he posted (this one).

If you insist that your projections aren't predictions, gero, then I suggest to not make any adjustments, but to add a very bold and very loud note that explains the issue. I already tried to warn people that the projections are useless this time beyond December.

[oren]

A model that breaks the extent/area into the 5 different regions, extrapolates each of them with a cutoff at zero, and then sums them up, would be mostly free of the negative ice problem. Of course, there would be added complications at calculation.

[Renerpho]

A model that breaks the extent/area into the 5 different regions, extrapolates each of them with a cutoff at zero, and then sums them up, would be mostly free of the negative ice problem. Of course, there would be added complications at calculation.

Maybe. I have two issues with this:
1. Is that data even available for the early years?, and
2. Can that be easily reproduced by someone else?

I know this isn't a nice thing to think about, but gero won't always be here to do these posts, and the easier they are to reproduce, the higher the chance that someone else can take over.

[oren]

Regional NSIDC data is available but it's not very easy to process due to their file format and due to the variable Pole Hole (at least in the Arctic, I assume the Antarctic doesn't have this problem).

[gerontocrat]

Regional NSIDC data is available but it's not very easy to process due to their file format and due to the variable Pole Hole (at least in the Arctic, I assume the Antarctic doesn't have this problem).

My NSIDC data has been adjusted for the Pole Hole - which only affects the total and the Central Arctic Region.

Yes, there is no worry about a pole hole in the Antarctic - there is a big fat ice sheet in the way (until it completely melts, not a worry for me).

And yes, doing the 5 seas seprately is not a technical problem - just a drag. Doing it for the Arctic's 14 seas/regions - a very big drag and my computer is already complaining about the analyses by individual Arctic regions/seas by area, extent and volume.

[Phil.]

As shown by Gero there is quite a difference in the various seas.  The three biggest regions are all at or close to their lowest values, the other two are unexceptional .

[John_the_Younger]

The first quandary that the melting season gives me is impossible results, i.e. for some years negative Antarctic Sea Ice Area at the minimum.

A model that uses percentages rather than absolute area would prevent negative projections.  If an area has 2 million km2 but used to have 4 million km2, projected losses of 2.1 million, arithmetically would give a negative area projection, but projectively losing 52.5% would give a 0.95 million km2 projection.  How about merging an arithmetic and percentage model: using the above example, subtract half the 2.1 loss and half the 52.5% loss: 2 - 1.05 - 0.525 = 0.425 million km2 projection.

When minimum area is historically low, on the other hand, I tend to think that ice growth potential is greater than historical growth.  When a sea used to grow from 4 to 8 million km2, now that it starts with 2, I'd expect it to grow more than 4 [how much it used to grow] but less than 6 [to reach the old maximum]...  How to model that???

[gerontocrat]

At least the projections do not produce sea ice of less than zero.

JAXA ANTARCTIC SEA ICE EXTENT:  16,893,069 KM2 as at 18-Sep-2023

- Extent gain on this day 2k, which is 6 k less than the average gain on this day (of the last 10 years) of 8k,
- Extent loss from maximum on this date is 0.06 million km2, 0.18 million km2, (76.3%) less than the 10 year average of 0.24 million km2.

- Extent is 1st lowest in the 45 year satellite record , which makes  sea ice daily extent lowest for 213 days this year

- Extent is  1,289 k LESS than 2015
- Extent is  969 k LESS than 2016
- Extent is  921 k LESS than 2017
- Extent is  2,253 k LESS than 2006
- Extent is  1,496 k LESS than 2021
- Extent is  1,173 k LESS than 2022
- Extent is  1,715 k LESS than the 1980's Average
- Extent is  1,711 k LESS than the 2010's Average

- On average 1.5% of ice loss  from maximum to minimum done, and about 155 days to minimum

Projections. (Table JAXA-Ant1)

Average remaining melt (of the last 10 years) would produce a minimum in Feb 2023 of 1.00 million km2, 1.13 million km2 BELOW the 2022 record low minimum of 2.13 million km2 which would be 1st lowest in the satellite record.
_______________________________________________________________
Click on image to enlarge

[gerontocrat]

NSIDC ANTARCTIC SEA ICE AREA (5 day trailing average):  12,890,832 KM2 as at 18-Sep-2023

- Area loss on this day 82k, which is 79 k more than the average loss on this day (of the last 10 years) of 3k,
- AREA loss from maximum on this date is 0.48 million km2, 0.04 million km2, (8%) more than the 10 year average of 0.45 million km2.

- AREA is 1st lowest in the 45 year satellite record, with daily sea ice area lowest for the day for 198 days this year

- AREA is  848 k LESS than 2016
- AREA is  953 k LESS than 2017
- AREA is  2,400 k LESS than 2006
- Area is  1,166 k LESS than 2022
- AREA is  1,630 k (11.2%) LESS than the 1980's Average
- AREA is  1,749 k (11.9%) LESS than the 2010's Average

- On average 3.4% of ice loss  from maximum to minimum done, and on average about 159 days to minimum

Projections. (Table NSIDC ANTARCTIC AREA-1)

Average remaining melt (of the last 10 years) would produce a sea ice area minimum in Feb 2023 of 0.08 million km2, 1.18 million km2 below the 2022 record low minimum of 1.23 million km2. This result verges on the impossible
_______________________________________________________________
Click on images to enlarge

[gerontocrat]

I have decided what to do on the projections of the minimum.

I will apply the same methodology to each of the 5 Antarctic regions, the summary table being the sum of the 5 regions. Not such a big task, just time-consuming to set it up.

[kiwichick16]

good luck Gero  ......what are the odds of this being a normal / average melt season , given where the ice area is currently, on top of the sea and air temps we are seeing at the moment?

i'm not betting on it

[Phil.]

good luck Gero  ......what are the odds of this being a normal / average melt season , given where the ice area is currently, on top of the sea and air temps we are seeing at the moment?

i'm not betting on it

In other years it would take a month or two from now to reach this level, it would be hard to see the ice area not reach very low levels next year.  Most of the ice I would think be left in the Weddell Sea which is about the only place there is any MYI.

[kiwichick16]

+1 Phil   ......its the equinox at Scott Base tomorrow  ...... so 12 hours plus daylight time on the sea ice .....or on open water

https://dateandtime.info/citysunrisesunset.php?id=6620787&month=9&year=2023

[gerontocrat]

good luck Gero  ......what are the odds of this being a normal / average melt season , given where the ice area is currently, on top of the sea and air temps we are seeing at the moment?

i'm not betting on it

The idea of much of what I post is to show the deviation of current sea ice from the result given by an average year.

Meanwhile,  the table and graphs of the JAXA (ADS) 2023 maximum of 16.99 million km2 is:-
- lowest in the 45 year satellite record,
- 1.62 million km2 below the longterm linear trend,
- 1.13 million km2 below the 2022 maximum,
- 1.08 million km2 below the previous record low maximum in 2017.

click graph to enlarge, click table twice for full size

[gerontocrat]

JAXA ANTARCTIC SEA ICE EXTENT:  16,897,549 KM2 as at 19-Sep-2023

- Extent gain on this day 4k, which is 10 k more than the average loss on this day (of the last 10 years) of 6k,
- Extent loss from maximum on this date is 0.09 million km2, 0.05 million km2, (38.2%) less than the 10 year average of 0.14 million km2.

- Extent is 1st lowest in the 45 year satellite record , which makes  sea ice daily extent lowest for 214 days this year

- Extent is  1,290 k LESS than 2015
- Extent is  991 k LESS than 2016
- Extent is  886 k LESS than 2017
- Extent is  2,238 k LESS than 2006
- Extent is  1,456 k LESS than 2021
- Extent is  1,160 k LESS than 2022
- Extent is  1,711 k LESS than the 1980's Average
- Extent is  1,692 k LESS than the 2010's Average

- On average 0.9% of ice loss  from maximum to minimum done, and about 154 days to minimum

Projections. (Table JAXA-Ant1)

Average remaining melt (of the last 10 years) would produce a minimum in Feb 2024 of 1.01 million km2, 0.94 million km2 BELOW the 2023 record low minimum of 2.13 million km2 which would be 1st lowest in the satellite record.
_______________________________________________________________
Click on image to enlarge

[kiwichick16]

@ Gero  ....absolutely  .....really difficult to pick a real trend from the numbers  ......perhaps the recent La Nina years have disguised the trend over the last couple of years, but adding more heat every year will eventually overcome the noise

[gerontocrat]

This is hopefully the last time I will post this ludicrous projection of -ve sea ice area values (but I don't know what to do about JAXA sea ice extent, as there is no area data

SEE NEXT POST

NSIDC ANTARCTIC SEA ICE AREA (5 day trailing average):  12,803,689 KM2 as at 19-Sep-2023

- Area loss on this day 87k, which is 100 k more than the average gain on this day (of the last 10 years) of 13k,
- AREA loss from maximum on this date is 0.57 million km2, 0.14 million km2, (31%) more than the 10 year average of 0.43 million km2.

- AREA is 1st lowest in the 45 year satellite record, with daily sea ice area lowest for the day for 199 days this year

- AREA is  955 k LESS than 2016
- AREA is  1,038 k LESS than 2017
- AREA is  2,478 k LESS than 2006
- Area is  1,248 k LESS than 2022
- AREA is  1,727 k (11.9%) LESS than the 1980's Average
- AREA is  1,847 k (12.6%) LESS than the 2010's Average

- On average 3.3% of ice loss  from maximum to minimum done, and on average about 158 days to minimum

Projections. (Table NSIDC ANTARCTIC AREA-1)

Average remaining melt (of the last 10 years) would produce a sea ice area minimum in Feb 2023 of -0.02 million km2, 1.28 million km2 below the 2022 record low minimum of 1.23 million km2.
_______________________________________________________________
Click on images to enlarge

[gerontocrat]

NSIDC ANTARCTIC SEA ICE AREA (5 day trailing average):
Replacing projections of the minimum with the sum of projections for each of the 5 Regional Seas

Starting with the Weddell Sea.
Why? Because of its unique position, there is a large area of ice tucked away in the far SE of the Weddell embayment, still largely protected from significant melt with a sustantial proportion of MYI. The Weddell Sea is the last refuge of summer sea ice in the Antarctic

So I have produced a projection of the 2024 minimum using the same formulae as for the overall total. The results show that far from projecting an extremely low minmimum, the projection for 2024 is 0.75 million km2, a small increase of 47k km2 over 2022.
EDIT: Whoops. I forgot to attach the projections table Now attached

Also the linear trend remains as an annual increase. See minima graph attached.

This graph is in complete contrast with the current projection of the 2023 annual average daily sea ice area, which shows very large decreases in the average in the last two years, with the 2023 average expected at more than 500k km2 below the linear trend, which is now -ve, i.e. annual sea ice area losses. See 3rd graph

I think what is happening is that first year ice has become the first casuality of the temporary? permanent ? change in the Antarctic. Sea ice maxima decline and in the summer sea ice losses accelerate. But that change has not yet reached the ice refuge of the Weddell. When it does......

I will be doing the same exercise for each of the other 4 seas. It will take some time.

[gerontocrat]

NSIDC ANTARCTIC SEA ICE AREA (5 day trailing average):
Replacing projections of the minimum with the sum of projections for each of the 5 Regional Seas

Continuing with the Indian Ocean Region (i.e. going clockwise, as do the Circumpolar winds & currents).

So I have produced a projection of the 2024 minimum using the same formulae as for the overall total. The unadjusted results show in all years, -ve sea ice area is projected. This is entirely due to the record low and very early maximum in 2023. (Maxima tables and graphs yet to be produced). So I have changed that to a minimum of 2,000 km2, as there is almost always a bit of ice hiding in a sheltered corner. See the projections table attached

Also the linear trend is for a very small annual increase. See minima graph attached.


The 2023 annual average daily sea ice area, shows very large decreases in the average in the last two years, with the 2023 average expected at more than 350k km2 below the linear trend, which is -ve, i.e. annual sea ice area losses. See 3rd graph

I think what is happening is sea ice maxima decline and summer sea ice losses acceleration.

I will be doing the same exercise for each of the other 3 seas/regions. It will take some time.

[gerontocrat]

NSIDC ANTARCTIC SEA ICE AREA (5 day trailing average):
Replacing projections of the minimum with the sum of projections for each of the 5 Regional Seas

Continuing with the West Pacific Ocean Region (i.e. still going clockwise, as do the Circumpolar winds & currents).

So I have produced a projection of the 2024 minimum using the same formulae as for the overall total. The unadjusted results show that for nearly all years, -ve sea ice area is projected. This is entirely due to the record low and very early maximum in 2023. (Maxima tables and graphs yet to be produced). So I have changed that to a minimum of 2,000 km2 where indicated, as there is almost always a bit of ice hiding in a sheltered corner. See the projections table attached

Also the linear trend is for a small annual increase. See minima graph attached.


The 2023 annual average daily sea ice area, shows very large decreases in the average in the last two years, with the 2023 average expected at more than 200k km2 below the linear trend, which is for small annual sea ice area gains. See 3rd graph

I think that in all regions what is happening recently is generally sea ice maxima decline and summer sea ice losses acceleration.

I will be doing the same exercise for each of the other 2 seas/regions. It will take some time.

[gerontocrat]

NSIDC ANTARCTIC SEA ICE AREA (5 day trailing average):
Replacing projections of the minimum with the sum of projections for each of the 5 Regional Seas

Continuing with the Ross Sea (i.e. still going clockwise, as do the Circumpolar winds & currents).

The Ross Sea's position makes it more vulnerable than the SE quarter of the Weddell sea to sea ice melt, and is apparent in the volatility of the data.

So I have produced a projection of the 2024 minimum using the same formulae as for the overall total. The unadjusted results show that for all years bar 1, -ve sea ice area is projected. This is entirely due to the record low in 2023 and high area losses in recent years. (Maxima tables and graphs yet to be produced). So I have changed those less than zero sea ice years to a minimum of 2,000 km2 where indicated, as there is almost always a bit of ice hiding in a sheltered corner. See the projections table attached

Also the linear trend is for a 4k annual decrease. See minima graph attached.


The 2023 annual average daily sea ice area, shows very large decreases in the average in 2023, with the 2023 average expected at more than 450k km2 below the linear trend, which is for small annual sea ice area gains. See 3rd graph

I think that in all regions what is happening recently is generally sea ice maxima decline and summer sea ice losses acceleration.

I will be doing the same exercise for the last region. It will take some time.

[gerontocrat]

NSIDC ANTARCTIC SEA ICE AREA (5 day trailing average):
Replacing projections of the minimum with the sum of projections for each of the 5 Regional Seas

Finally, the BellinghausenAmundsen Sea

The BellinghausenAmundsen sea in 2023 went against the general trend - increasing sea ice area, not decreasing.

I have produced a projection of the 2024 minimum using the same formulae as for the overall total. The unadjusted results show that for most years, a sea ice area above zero is projected, due to a strong recovery in sea ice area in 2023. (Maxima tables and graphs yet to be produced). Those years with less than zero sea ice years are auto-changed to a minimum of 2,000 km2 where indicated, as there is almost always a bit of ice hiding in a sheltered corner. See the projections table attached

Also the linear trend is for a 5k annual decrease. See minima graph attached.

The 2023 annual average daily sea ice area, shows a modest reovery in the average in 2023, with the 2023 average expected at less than 50k km2 above the linear trend, which is for a modest annual sea ice area loss. See 3rd graph

I will be summing the data of the 5 regions to produce a modified prediction,but tomorrow is another day

[gerontocrat]

JAXA ANTARCTIC SEA ICE EXTENT:  16,825,018 KM2 as at 20-Sep-2023

- Extent loss on this day 73k, which is 84 k less than the average gain on this day (of the last 10 years) of 11k,
- Extent loss from maximum on this date is 0.16 million km2, 0.01 million km2, (3.7%) more than the 10 year average of 0.16 million km2.

- Extent is 1st lowest in the 45 year satellite record , which makes  sea ice daily extent lowest for 215 days this year

- Extent is  1,404 k LESS than 2015
- Extent is  1,035 k LESS than 2016
- Extent is  921 k LESS than 2017
- Extent is  2,352 k LESS than 2006
- Extent is  1,503 k LESS than 2021
- Extent is  1,230 k LESS than 2022
- Extent is  1,770 k LESS than the 1980's Average
- Extent is  1,786 k LESS than the 2010's Average

- On average 1.0% of ice loss  from maximum to minimum done, and about 153 days to minimum

Projections. (Table JAXA-Ant1)

Average remaining melt (of the last 10 years) would produce a minimum in Feb 2024 of 0.92 million km2, 1.03 million km2 BELOW the 2023 record low minimum of 2.13 million km2 which would be 1st lowest in the satellite record.
_______________________________________________________________
Click on image to enlarge

[gerontocrat]

NSIDC ANTARCTIC SEA ICE AREA (5 day trailing average):  12,737,156 KM2 as at 20-Sep-2023

N.B. Projections are based on the results for the 5 regional seas

- Area loss on this day 67k, which is 83 k more than the average gain on this day (of the last 10 years) of 16k,
- AREA loss from maximum on this date is 0.64 million km2, 0.22 million km2, (52%) more than the 10 year average of 0.42 million km2.

- AREA is 1st lowest in the 45 year satellite record, with daily sea ice area lowest for the day for 200 days this year

- AREA is  1,021 k LESS than 2016
- AREA is  1,101 k LESS than 2017
- AREA is  2,540 k LESS than 2006
- Area is  1,339 k LESS than 2022
- AREA is  1,796 k (12.4%) LESS than the 1980's Average
- AREA is  1,918 k (13.1%) LESS than the 2010's Average

- On average 3.1% of ice loss  from maximum to minimum done, and on average about 157 days to minimum

Projections. (Table NSIDC ANTARCTIC AREA-1)
The projections in this post are after making projections for each of the 5 regions. When individual years in individual regions produce a less than zero result for the minimum, this is automatically adjusted to zero.
As a result the attached table does not add up across as the “remaining melt” column is before adjustments for -ve sea ice. Tomorrow I will add columns to show the melt adjustments and have a think on what those adjustments should be.

Average remaining melt (of the last 10 years) would produce a sea ice area minimum in Feb 2023 of 1.01 million km2, 0.25 million km2 below the 2022 record low minimum of 1.23 million km2.
_______________________________________________________________
Click on images to enlarge

[oren]

Thanks for all the efforts you put into the new calculation Gero. I think it better represents the physical reality of the Antarctic, where the seas are mostly separated geographically and ice does not move freely between them.
BTW, I am still quite amazed with this sentence. "AREA is 1st lowest in the 45 year satellite record, with daily sea ice area lowest for the day for 200 days this year". And each day it grows by one more. Crazy year.

[Ktb]

Gerontocrat, any chance you've got the 365 trailing charts handy? They have to look absurd right now

[gerontocrat]

JAXA ANTARCTIC SEA ICE EXTENT:  16,802,153 KM2 as at 21-Sep-2023

- Extent loss on this day 23k, which is 25 k less than the average gain on this day (of the last 10 years) of 2k,
- Extent loss from maximum on this date is 0.18 million km2, 0.02 million km2, (12.9%) more than the 10 year average of 0.16 million km2.

- Extent is 1st lowest in the 45 year satellite record , which makes  sea ice daily extent lowest for 216 days this year

- Extent is  1,414 k LESS than 2015
- Extent is  1,095 k LESS than 2016
- Extent is  910 k LESS than 2017
- Extent is  2,398 k LESS than 2006
- Extent is  1,537 k LESS than 2021
- Extent is  1,249 k LESS than 2022
- Extent is  1,759 k LESS than the 1980's Average
- Extent is  1,820 k LESS than the 2010's Average

- On average 1.0% of ice loss  from maximum to minimum done, and about 152 days to minimum

Projections. (Table JAXA-Ant1)

Average remaining melt (of the last 10 years) would produce a minimum in Feb 2024 of 0.90 million km2, 1.05 million km2 BELOW the 2023 record low minimum of 1.95 million km2 which would be 1st lowest in the satellite record.
This projection probably exaggerates the melt in the Weddell Sea in particular
_______________________________________________________________
Click on image to enlarge

[gerontocrat]

Gerontocrat, any chance you've got the 365 trailing charts handy? They have to look absurd right now

OK - attached, + annual daily average

[Chris83]

  Worth mentioning in the context
 Incredibly large heat anomaly today (check map on CR)
This is the data for previous days

Highest anomaly in #Antarctica in 2023

 +5.39c (14.09.2023).   (highest for both poles in fact in 2023)
+5.34 (15.09.2023)
+5.34 (16.09.2023)
+5.29 (22.09.2023)
+5.04 (13.09.2023)
+5.06 (17.09.2023)
+4.84c (20.09.2023)
+4.49c (04.07.2023) (12.09.2023)
+4.06c (18.09.2023)
+3.7.  (11.09.2023)

Data climate reanalyser

[RoxTheGeologist]

Lack of sea ice typically leads to high temperature anomalies - the ocean warms the air.

[gerontocrat]

JAXA daily sea ice extent has been lowest for the day for 217 out of the 265 days that have passed so far this year, and given average melt would stay that way for the remainder of the year. Definitely a unique event.

JAXA ANTARCTIC SEA ICE EXTENT:  16,714,738 KM2 as at 22-Sep-2023

- Extent loss on this day 87k, which is 93 k less than the average gain on this day (of the last 10 years) of 6k,
- Extent loss from maximum on this date is 0.27 million km2, 0.12 million km2, (73.8%) more than the 10 year average of 0.16 million km2.

- Extent is 1st lowest in the 45 year satellite record , which makes  sea ice daily extent lowest for 217 days this year

- Extent is  1,469 k LESS than 2015
- Extent is  1,193 k LESS than 2016
- Extent is  1,000 k LESS than 2017
- Extent is  2,481 k LESS than 2006
- Extent is  1,647 k LESS than 2021
- Extent is  1,336 k LESS than 2022
- Extent is  1,823 k LESS than the 1980's Average
- Extent is  1,924 k LESS than the 2010's Average

- On average 1.0% of ice loss  from maximum to minimum done, and about 151 days to minimum

Projections. (Table JAXA-Ant1)

Average remaining melt (of the last 10 years) would produce a minimum in Feb 2024 of 0.80 million km2, 1.15 million km2 BELOW the 2023 record low minimum of 1.95 million km2 which would be 1st lowest in the satellite record.
This projection probably exaggerates the melt in the Weddell Sea in particular
_______________________________________________________________
Click on image to enlarge

[gerontocrat]

When the projections for the minimum sea ice result in a less than zero minimum for any year in any of the 5 regional seas this is changed to the 21st Century minimum.

More explanation later

NSIDC ANTARCTIC SEA ICE AREA (5 day trailing average):  12,623,608 KM2 as at 22-Sep-2023

N.B. Projections are based on the sum of the results for each of the 5 regional seas

- Area loss on this day 48k, which is 56 k more than the average gain on this day (of the last 10 years) of 8k,
- AREA loss from maximum on this date is 0.75 million km2, 0.35 million km2, (88%) more than the 10 year average of 0.40 million km2.

- AREA is 1st lowest in the 45 year satellite record, with daily sea ice area lowest for the day for 202 days this year

- AREA is  1,151 k LESS than 2016
- AREA is  1,211 k LESS than 2017
- AREA is  2,573 k LESS than 2006
- Area is  1,496 k LESS than 2022
- AREA is  1,882 k (13.0%) LESS than the 1980's Average
- AREA is  2,040 k (13.9%) LESS than the 2010's Average

- On average 3.0% of ice loss  from maximum to minimum done, and on average about 155 days to minimum

Projections. (Table NSIDC ANTARCTIC AREA-1)
The projections in this post are after making projections for each of the 5 regions. When individual years in individual regions produce a less than zero result for the minimum, this is automatically adjusted to the 21st Century minimum.
Average remaining melt (of the last 10 years) would produce a sea ice area minimum in Feb 2023 of 1.11 million km2, 0.05 million km2 below the 2022 record low minimum of 1.23 million km2.
_______________________________________________________________
Click on images to enlarge

[gerontocrat]

I attach the two versions of the projections table, the 2nd modified to use the 21st Century minmum when a result produced a less than zero sea ice result.

Explanations (with luck) tomorrow

[oren]

When the projections for the minimum sea ice result in a less than zero minimum for any year in any of the 5 regional seas this is changed to the 21st Century minimum.

For consistency, value after cutoff should be equal to the cutoff threshold. Otherwise a projection of 1k remains 1k, while a very similar projection of -1k can become 100,000k.
So values below 21st century minimum (not below zero) should be changed to the 21st century minimum.
But I think this will make the projection too conservative as it does not allow for further losses even when current numbers are much lower than historical numbers for the same date. OTOH assuming results of 2k as previously is too unconservative.
Perhaps I would go with: projected values below 75% (or 50%) of the 21st minimum (for each sea separately of course) are changed to 75% (50%) of the 21st century minimum.

[gerontocrat]

When the projections for the minimum sea ice result in a less than zero minimum for any year in any of the 5 regional seas this is changed to the 21st Century minimum.

For consistency, value after cutoff should be equal to the cutoff threshold. Otherwise a projection of 1k remains 1k, while a very similar projection of -1k can become 100,000k.
So values below 21st century minimum (not below zero) should be changed to the 21st century minimum.
But I think this will make the projection too conservative as it does not allow for further losses even when current numbers are much lower than historical numbers for the same date. OTOH assuming results of 2k as previously is too unconservative.
Perhaps I would go with: projected values below 75% (or 50%) of the 21st minimum (for each sea separately of course) are changed to 75% (50%) of the 21st century minimum.

I wish to emphasise that in making these projections all I am trying to do is present an average and reasonable range of possible outcomes using outcomes from the last 10 years and current sea ice data.

In my view it is also important to keep it simple, i.e. a methodology easily explainable is best. Often complexity does not add value, only confusion.

I agree with Orens comments, and have started by applying it to JAXA Antarctic sea ice extent in my next post.

At the end of a melt season insolation reduces quickly and ocean heat start to dissipate, slowing down melt. In the Antarctic, especially the Weddell Sea, by the end of the melting season the sea ice that is left is mainly MYI landfast ice and much harder to melt. So I have chosen a cut-off using a maximum 25% reduction to the current record 21st Century minimum to each year, i.e. Oren's 75% option. 

That can be changed by a change to the paremeter in the spreadsheet and might be required. Not only is current sea ice extent nearly 2 million km2 below the 2010's average, a look at graph Ant-2 in the next post suggests at the moment the melting season is about one month ahead of schedule.

[gerontocrat]

JAXA ANTARCTIC SEA ICE EXTENT:  16,656,408 KM2 as at 23-Sep-2023

- Extent loss on this day 58k, which is 39 k more than the average loss on this day (of the last 10 years) of 19k,
- Extent loss from maximum on this date is 0.33 million km2, 0.16 million km2, (92.1%) more than the 10 year average of 0.17 million km2.

- Extent is 1st lowest in the 45 year satellite record , which makes  sea ice daily extent lowest for 218 days this year

- Extent is  1,488 k LESS than 2015
- Extent is  1,271 k LESS than 2016
- Extent is  1,031 k LESS than 2017
- Extent is  2,525 k LESS than 2006
- Extent is  1,703 k LESS than 2021
- Extent is  1,337 k LESS than 2022
- Extent is  1,854 k LESS than the 1980's Average
- Extent is  1,967 k LESS than the 2010's Average

- On average 1.1% of ice loss  from maximum to minimum done, and about 150 days to minimum

Projections. (Table JAXA-Ant1)

Average remaining melt (of the last 10 years) would produce a minimum in Feb 2024 of 1.53 million km2, 0.42 million km2 BELOW the 2023 record low minimum of 1.95 million km2 which would be 1st lowest in the satellite record.
Note that the projected minimum for each year is constrained by a maximum 25% reduction to the current 21st Century record low
_______________________________________________________________
Click on image to enlarge

[kiwichick16]

 1 million sq  kms below the previous record low  for this day , 2002

[Stephan]

1 million sq  kms below the previous record low  for this day , 2002

... and weeks ahead compared to the 1980s-2010s average.
The actual value (16.66 M km²) was reached on Nov 8-Nov 10 in all decadal averages. Therefore 2023 is six weeks in advance...

[gerontocrat]

I have settled on a cutoff value of 67% of the 21st century minimum when making the minima projections for the 5 regions which are summed for the projection attached to this post. I was very surprised as to how insensitive the projection is to a change of parameter is once a cutoff is applied.

NSIDC ANTARCTIC SEA ICE AREA (5 day trailing average):  12,590,315 KM2 as at 23-Sep-2023

N.B. Projections are based on the sum of the results for each of the 5 regional seas

- Area loss on this day 33k, which is 39 k more than the average gain on this day (of the last 10 years) of 6k,
- AREA loss from maximum on this date is 0.78 million km2, 0.39 million km2, (100%) more than the 10 year average of 0.39 million km2.

- AREA is 1st lowest in the 45 year satellite record, with daily sea ice area lowest for the day for 203 days this year

- AREA is  1,166 k LESS than 2016
- AREA is  1,270 k LESS than 2017
- AREA is  2,564 k LESS than 2006
- Area is  1,516 k LESS than 2022
- AREA is  1,891 k (13.1%) LESS than the 1980's Average
- AREA is  2,080 k (14.2%) LESS than the 2010's Average

- On average 3.0% of ice loss  from maximum to minimum done, and on average about 154 days to minimum

Projections. (Table NSIDC ANTARCTIC AREA-1)
The projections in this post are the sum of the projections for each of the 5 regions. When individual years in individual regions produce a less than zero result for the minimum, this is automatically adjusted to a minimum of 67% of the 21st Century minimum.
Average remaining melt (of the last 10 years) would produce a sea ice area minimum in Feb 2023 of 1.13 million km2, 0.07 million km2 above the 2023 record low minimum of 1.06 million km2.
_______________________________________________________________
Click on images to enlarge

[gerontocrat]

I applied a cutoff of a maximum 33% drop in the 2024 projected minimum from the 2023 record low minimum,. This produces results that only show two things - the current sea ice extent is in totally unprecedented territory, and this makes the projection model useless.

I think I have to abandon it for the time being.

JAXA ANTARCTIC SEA ICE EXTENT:  16,630,880 KM2 as at 24-Sep-2023

- Extent loss on this day 26k, which is 1 k more than the average loss on this day (of the last 10 years) of 25k,
- Extent loss from maximum on this date is 0.35 million km2, 0.08 million km2, (31.2%) more than the 10 year average of 0.27 million km2.

- Extent is 1st lowest in the 45 year satellite record , which makes  sea ice daily extent lowest for 219 days this year

- Extent is  1,433 k LESS than 2015
- Extent is  1,266 k LESS than 2016
- Extent is  977 k LESS than 2017
- Extent is  2,569 k LESS than 2006
- Extent is  1,708 k LESS than 2021
- Extent is  1,390 k LESS than 2022
- Extent is  1,857 k LESS than the 1980's Average
- Extent is  1,964 k LESS than the 2010's Average

- On average 1.7% of ice loss  from maximum to minimum done, and about 149 days to minimum

Projections. (Table JAXA-Ant1)

Average remaining melt (of the last 10 years) would produce a minimum in Feb 2024 of 1.31 million km2, 0.64 million km2 BELOW the 2023 record low minimum of 1.95 million km2 which would be 1st lowest in the satellite record.
Note that the projected minimum for each year is constrained by a maximum 33% reduction to the current 21st Century record low
_______________________________________________________________
Click on image to enlarge

[oren]

I really don't think the updated projection is useless.

[gerontocrat]

I really don't think the updated projection is useless.

I am glad you think so.

OK, I will stick with the 67% cutoff, and attached are some of the regional results for NSIDC sea ice area and other data, starting with the Weddell Sea which deserves a closer look.

A look at the Weddell Sea

The 1st graph attached is the Weddell Sea Ice Area for 2023 to date.
This demonstrates that extremely low sea ice area during the year does not guarantee a record low maximum, but surely it guarantees that at the maximum a large percentage of sea ice will be much thinner than usual. That must give this melting season a high chance of rapid melt, greatly increasing Albedo Warming Potential (AWP) and from that ocean heating.
Climatically, the extremely low sea ice during the freezing season is far more significant than the resulting maximum.

Equally, a fast melt and high ocean heating must be more significant on the following freezing season than the one day minimum. You can see in the 2nd graph that in 2022 the minimum sea ice area was at the 1980's average, but during the 2022 melting season was mostly well below average.

However, setting up the day by day projections plume to demonstrate this given the use of the 67% constraint is not an easy task and leaves the question as to when during the melting season to apply the constraint. This is logically towards the end of melting as open water approaches the large area of hard to melt thick MYI landfast ice in the SE of the Weddell sea.

The 3rd graph shows the annual average daily sea ice area of the Wedddell sea - actuals to Sept 23, estimates to year end. This does show that 2023 will almost certainly end at a record low, probably 500,000km2 below linear trend. That is a lot of missing ice.

The last graph is the projections of the 2024 minimum with the 67% constraint applied. This produces an average minimum of 0.78 million km2, 30,000 km2 above the 2022 record low.
However in my view the annual daily average sea ice for 2024 is very likely to also be extremely low.

So what matters? A one day maximum and a one day minimum or what happens during the freezing and melting seasons?


more regional seas later
click images to enlarge