Arctic sea ice maximum extent was present for at least two weeks at about 14.9 million km2

From Polar Bear Science

Dr. Susan Crockford

US National Snow and Ice Data Center says the Arctic maximum extent for this winter peaked at 14.88 mkm2 on 25 February, but in fact this amount of ice coverage lasted for at least two weeks (22 February – 8 March), with very slight variation. Just a little something they all left out of their announcements, for some reason.

Here are the NSIDC maps I saved to my own archive:

NSIDC scientists calculate an average to two decimal places in order to arrive at a single peak date.

However, that doesn’t negate the fact that sea ice covered between 14.8 and 15.0 mkm2 of the Arctic for at least two weeks until the 7th or 8th of March, and that’s what is biologically significant. Keep in mind that for long-term comparative (climatological) purposes, the average extent for the entire month of March is used (see graph below), not the peak extent, and we won’t know that until early April.

Graph above from NOAA Arctic Report Card 2021 ‘Sea Ice’.

From the NSIDC report (22 March 2022):

On February 25, 2022, Arctic sea ice likely reached its maximum extent for the year, at 14.88 million square kilometers (5.75 million square miles), the tenth lowest extent in the satellite record. [See their Table 1, below] This year’s maximum extent is 770,000 square kilometers (297,000 square miles) below the 1981 to 2010 average maximum of 15.65 million square kilometers (6.04 million square miles) and 470,000 square kilometers (182,000 square miles) above the lowest maximum of 14.41 million square kilometers (5.56 million square miles) set on March 7, 2017. Prior to 2019, the four lowest maximum extents occurred from 2015 to 2018.

The date of the maximum this year, February 25, was fifteen days earlier than the 1981 to 2010 average date of March 12. Only two years had an earlier maximum, 1987 and 1996, both on February 24. This year is the second earliest date on the satellite record, tying with 2015, which also reached its maximum extent on February 25.

The ice growth season ended with near average sea ice extent in the Bering Sea, above average in Baffin Bay and off the coast of south-eastern Greenland, and below average in the Barents Sea with a narrow open-water wedge north of Novaya Zemlya. Extent was well below average in the Gulf of St. Lawrence and the Sea of Okhotsk.

Since the maximum on February 25, extent has dropped about 390,000 square kilometers (151,000 square miles), with losses primarily in the Sea of Okhotsk and the Barents Sea. These losses have been offset by gains in the Bering Sea, Baffin Bay, and the Labrador Sea.

Postscript: at 20 March, ice extent had only dropped to 14.7 mkm2 (see below).

5 24 votes
Article Rating
Newest Most Voted
Inline Feedbacks
View all comments
March 24, 2022 6:28 am

NSIDC scientists calculate an average to two decimal places in order to arrive at a single peak date.

Why do that? What is wrong with just reporting the peak or peaks that occurred during the season be it one, two or twenty? We can be thankful that these guys don’t “calculate” baseball standings I suppose.

March 24, 2022 6:38 am

Looking at table 1 it looks like we are going towards a new ice age because the ice extent has increased for four years. If the trend continues the whole earth will be frozen over soon. /sark

Last edited 2 months ago by Lars Silen
Tom in Florida
Reply to  Lars Silen
March 24, 2022 7:04 am

What doesn’t show is that 2020 maximum was 15.05 km^2, so the extent has increased for 5 years. If that trend continues by 2100 there will be an additional 8.47 km^2 of extent bringing the total extent to just over 23.5 km^2. (using climatology extrapolation)

Reply to  Tom in Florida
March 24, 2022 8:57 am

The uncertainty of sea ice extent is greater than the annual variations since 1980.

Sea Ice Uncertainty .png
Reply to  Thomas
March 24, 2022 9:47 am

Do you know where those red lines are coming from in that graph?

Last edited 2 months ago by bdgwx
Reply to  bdgwx
March 24, 2022 10:51 am

I put them there, based on the information at the link.

Reply to  Thomas
March 24, 2022 11:13 am

Thanks. The new link is

What that is saying is that the sea concentration in any specific grid cells is known to with ± 15% where the concentration can range from 0 to 100% with 0% having no evidence of ice and 100% having the cell completely ice covered. This is why the SII product defines the ice edge as those grid cells with 15% or less ice concentration.

That ± 15% figure is not the uncertainty of the sea ice extent itself, but only the concentration in a particular grid cell. Per Meier & Stewart 2019 the uncertainty on wintertime Arctic sea ice extent maximum is ±0.034e6 km2 with the value ranging from 0.03e6 to 0.07e6 km2. What you want to do is redraw your red lines so that they are ±0.07 around the extent figure in the graph.

BTW…the astute reader will recognize that 0.034e6 km2 figure. It is actually cited in the blog post!

Last edited 2 months ago by bdgwx
Reply to  bdgwx
March 24, 2022 11:55 am

It does not seem possible for overall uncertainty to be less than the certainty per grid cell.

M&S 2019 evaluated uncertainty in two ways, “(1) absolute uncertainty via comparison of extents from several different products, and (2) relative uncertainty by comparing extent variability with different inputs and processing.

But neither of those methods consist of comparing the satellite measurements so some other more accurate measurement, like an aerial survey, for example.

Reply to  Thomas
March 24, 2022 12:22 pm

Thomas said: “It does not seem possible for overall uncertainty to be less than the certainty per grid cell.”

I’m not even sure what that would mean. Concentration is a measure of the ice coverage within a cell. Extent is a measure of the surface area of all grid cells contained within a perimeter bounded by cells transitioning from < 15% to > 15% concentration. They are two fundamentally different measurements. A single cell cannot have an extent nor can the perimeter ring have a concentration so a comparison of the uncertainty for each would not have any useful meaning that I know of.

Reply to  bdgwx
March 24, 2022 12:33 pm

A single cell can have an extent of ice coverage from 0% to 100%, but I think I see what you are saying. If we assume the main body of ice is without significant gaps, then all the uncertainty is at the edges. Large uncertainties at the edges could have a small affect on the certainty of the overall extent, so my red dashed lines on the chart are probably wrong. Thanks.

Reply to  Thomas
March 24, 2022 12:58 pm

Exactly. For example, at 14.88e6 km2 and assuming the extent is circular (it’s not) that makes the perimeter ring radius sqrt(14.88e6 / 3.14) = 2176 km. That is the radius at which the perimeter cells transition from > 15% coverage to < 15% coverage. If if the transition from 100% to 0% coverage occurs over a very short distance such that it is the 1% for every 1 km. Then with a 15% error in coverage we can expect a 15 km error in the radius. A 15 km error in the radius around 2175 km translates to 0.2e6 km2 of error in the surface area bounded by that radius.

Reply to  bdgwx
March 24, 2022 1:45 pm

Makes sense. Thanks again.

Robert B
Reply to  bdgwx
March 24, 2022 5:04 pm

“accuracy of the median sea ice extent edge position for Sea Ice Index products has not been rigorously assessed. It would be difficult to do so, because ice edge is not a well-defined parameter. For our purposes, it is where source data grid cells transition from greater than 15 percent to less than 15 percent concentration. Operational services usually speak of a marginal ice zone of varying width over which concentration transitions from more than 90 percent to 0 percent. Spot checks of the sea ice edge position using a 15 percent concentration cutoff against NIC ice charts show that when there is a broad, diffuse ice edge, the NRTSI and GSFC products sometimes do not detect sea ice where the concentration can be as high as 60 Percent”

If the concentration is unknown to 15% of 15%, the edge is the extent where the concentration is less than 15+/-2% if its detected as under 15. The uncertainty should be the size of regions measured between 13 and 17%. I’m guessing that its much greater than 10% of the ice cover. This would be an over estimate as each cell is very unlikely to have been over or under estimated in any year. I’m guessing that the much smaller estimate is that 4% dived by the square root of the number of cells in the that 13-17% region. The issue being that its fine of the errors are perfectly random, with zero systematic error such as correcting for changes in orbit of the satellite.

That percentage within 13-17% is important because we know that any systematic error that led to sea ice increasing would be identified and corrected for. Not so much when its the other way. They pretty much say its for tracking the response to climate change.

Reply to  Robert B
March 25, 2022 6:48 am

I read that as a literal and straight up ±15%. In other words 30% could be 15-45%.

For a circular distribution to have a 10% difference on its area would require about a 110 km difference on the radius. And if 110 km represents the 15% error mark then 110 * (100 / 15) = 733 km would be the transition zone from 100% to 0% concentration.

Robert B
Reply to  bdgwx
March 25, 2022 12:53 pm

It couldn’t be that bad?

I was thinking that, apart from proof reading, I should have ended it with something along the lines of

They have put a lot of effort into a complex calculation of ice extent, which completely dwarfs the effort to calculate if its meaningful.

Reply to  Robert B
March 25, 2022 7:24 am

Thanks Robert. I tend to agree with your analysis.

NSIDC has this to say on the subject.

“NSIDC does not have error bars on the time series plot shown in the “Daily Image Update” and the daily time series plot (usually labeled “Figure 2”) because we strive to keep the images concise and easy to read. Plus, the error bars would be quite small compared to the total extent values in the images.

We estimate error based on accepted knowledge of the sensor capabilities and analysis of the amount of “noise,” or daily variations not explained by changes in weather variables. For average relative error, or error relative to other years, the error is approximately 20,000 to 30,000 square kilometers (7,700 to 11,600 square miles), a small fraction of the total existing sea ice.

For average absolute error, or the amount of ice that the sensor measures compared to actual ice on the ground, the error is approximately 50 thousand to 1 million square kilometers (19,300 to 386,100 square miles), varying over the year. During summer melt and freeze-up in the fall, the extent may be underestimated by 1 million square miles; during mid and late winter before melt starts, the error will be on the low end of the estimates. It is important to note that while the magnitude of the error varies through the year, it is consistent year to year. This gives scientists high confidence in interannual trends at a given time of year.

The absolute error values may seem high, but it is important to note that each year has roughly the same absolute error value, so the decline over the long term remains clear. NSIDC has high confidence in sea ice trend statistics and the comparison of sea ice extent between years.”

Error bars would be too messy and confusing and they don’t matter anyway because, they assume, errors are consistent from year to year.

Still, “scientists have high confidence” so there is nothing to be concerned about. : )

Reply to  Lars Silen
March 24, 2022 9:16 am

Add perspective to another thing lost in the climate wars, along with cycles. Why can’t someone do a climate test of the general population and reporters on the start of the satellite record with respect to fears of a coming ice age and cyclical lows of long run ocean cycles?

Boulder Skeptic
Reply to  Lars Silen
March 24, 2022 10:50 am

Yes, Arctic sea ice volume reported by BPIOMAS (can be viewed on the WUWT Sea Ice Page) is a sine wave trend, not a linear trend. Granted these BPIOMAS data are a model updated monthly from the University of Washington, not a real-time observation. Here’s the trendline I created and put on their data (green line). Sine fit looks pretty good to me with the 90-year period and +/- 6.1 (1000 km^3) magnitude from 0 anomaly shown. Their linear trend (blue line) is nonsense.

Screen Shot 2022-03-24 at 11.42.22 AM.png
Reply to  Boulder Skeptic
March 24, 2022 6:56 pm

Would you please share the declining sine wave function you fit?

Boulder Skeptic
Reply to  bigoilbob
March 25, 2022 7:37 am

My excel spreadsheet with the raw data and my curve fits is here.

Last edited 2 months ago by Boulder Skeptic
Reply to  Boulder Skeptic
March 25, 2022 9:18 am

Thank you very much. I had no issues with your fit, just curious.

Reply to  Boulder Skeptic
March 25, 2022 12:49 am

Yes, definitely, any true physical scientist would see the sinusoidal nature in the extent right away, and in the global temperature readings too.

A political scientist/activist, not so much.

Reply to  Lars Silen
March 26, 2022 5:42 am

When I look at the comment’s votometer, I’m not sure that your sarcasm has been understood 🙂

Ireneusz Palmowski
March 24, 2022 7:04 am

Ice extent and surface temperature on the Great Lakes on March 23, 2022.comment image
Forecast.comment image

March 24, 2022 7:11 am

10th lowest would clearly indicate that the ice is not disappearing . The fudging of the dates to call the maximum in February rather than into March with same /more ice is yet another BS tactic to convince the uneducated that this should be something of concern. Next up, predictions of low summer ice … and those will fail too.

Ireneusz Palmowski
March 24, 2022 7:18 am

The sea surface temperature is falling.comment image

Bruce Cobb
Reply to  Ireneusz Palmowski
March 24, 2022 7:39 am

Clearly, SSTs need Life Alert.

Janice Moore
Reply to  Bruce Cobb
March 24, 2022 11:18 am

😄 “… and I CAN’T GET UP!!”

Reply to  Ireneusz Palmowski
March 24, 2022 1:33 pm

Note the “warming hole” in the North Atlantic. That’s cooling for those who can still use that word–mostly outside the UK.

NOAA SST-NorthAtlantic GlobalMonthlyTempSince1979 With37monthRunningAverage.gif (880×481) (
(No updates)

Bloke down the pub
March 24, 2022 7:22 am

Has anyone compared the ice extent maps that NSIDC produce with their graphs? I know I’m only using a mk1 eyeball but there have been a number of occasions this winter where the missing ice appeared to be far less than the graphs were suggesting.

March 24, 2022 7:29 am

Lie by omission is still a lie, and such is prosecutable under Federal and State’s laws.

March 24, 2022 8:25 am

Dr. Susan Crockford,

Where are you getting those maps and why do they not match the data NSIDC publishes?

Richard Page
Reply to  bdgwx
March 24, 2022 8:50 am

Dr Crockford has clearly said they are from the NSIDC and from looking at the labels on them, I would assume they had been put up daily or weekly. Presumably when they archive the maps they only keep a few of them that show key points in their data. How many are in previous years archives, for example?

Reply to  Richard Page
March 24, 2022 11:34 am

RP said: “How many are in previous years archives, for example?”

The daily data for those graphs goes back to 2006.

Reply to  bdgwx
March 24, 2022 9:31 am

I think I figured it out. The data in those graphs come from USNIC via their MASIE product. The confusion is probably caused by the NSIDC processing the MASIE grids and hosting maps on their own website. MASIE (USNIC) and SII (NSIDC) are different datasets from different institutions with different goals, methodologies, and definitions of extent. That’s why there is some disagreement between them. Note that MASIE is designed as an input for numerical weather prediction while SII is designed for climate research. MASIE methodologies change to benefit numerical weather prediction whereas SII, like all climate datasets, uses a temporally stable methodology so that past, present, and future data can be compared. One other notable difference is that MASIE only goes back to 2006 whereas SII goes back to 1979.

Last edited 2 months ago by bdgwx
Reply to  bdgwx
March 24, 2022 10:38 am

Time for a cup of tea now Sherlock?

Reply to  Mr.
March 24, 2022 10:57 am

I am more shocked how CO2 continues to rise while the ice remains. Bdgwx needs to update some models 😉

Reply to  Derg
March 24, 2022 11:21 am

There is no need to update any models. The model(s) of the link between CO2 and Arctic sea ice extent adopted by the scientific community are consistent with the observation of the wintertime maximum in Arctic sea ice extent in 2022. The breakdown is that contrarians are testing models (like those that say CO2 is the only thing that can modulate Arctic sea ice extents) that the scientific community does not advocate for.

Reply to  bdgwx
March 24, 2022 6:24 pm

Lol…sure they are. According to your models when will the Arctic be ice free?

Reply to  Derg
March 25, 2022 6:55 am

In AR6 the IPCC expects the first essentially sea ice free (< 1e6 km2 extent) summer will occur around 2050. This is the earliest prediction I have seen to date. Prior to AR6 the expectation was around 2070 and in the 1990s the consensus was 2100.

Reply to  Mr.
March 24, 2022 11:23 am

It wasn’t that hard. My main concern is that Dr. Crockford is supposed to be the expert here so I’m going to give her the benefit of the doubt and assume she knows this. So why didn’t she mention it?

Last edited 2 months ago by bdgwx
Cheshire Red
March 24, 2022 8:33 am

Less than 5% drop on the 1981-2010 average. Irrelevant.

March 24, 2022 8:41 am

Dr. Susan Crockford,

I wanted to see if the 2 week near max period was unusual. What I did was take the 5-day average NSIDC data from here and backed into a value of -0.15e6 km2 of which max-0.15e6 km2 yielded 15 days in 2022. I then applied this rule to all years in the NSIDC record. Here is what I saw.

Year – Days Near Max
1995 – 50
1992 – 49
2015 – 34
1981 – 33
2008 – 30
2016 – 29
2017 – 29
2012 – 28
1997 – 28
1999 – 27

2022 – 15

The average days near max is 19 and the median is 18. In other words the 2022 number of days near max is not exceptional. In fact, if anything its length is below average.

Reply to  bdgwx
March 24, 2022 9:50 am

That’s nice, but sea ice volume is continuing to climb and probably will for another month. As I mentioned in my comment in the last sea ice article, extent has a lot to do with weather conditions along the edges of the ice sheet. You get a blocking pattern setup with a persistent wind out of the wrong direction and it will erode large areas of sea ice that isn’t that thick to begin with as the season changes. No wind and it will slowly dissipate. Also remember open water gives up more heat than ice covered water.

We are coming out of a solar minimum and the polar jet this NA winter has been pretty wavy. That’s why there has been so many abnormally cold and warm air masses push across the northern hemisphere. Lot’s of storms and lots of sea ice erosion (and a tremendous amount of northern hemisphere snow). Every winter is different. One of these years the Beaufort Gyre is going to let loose and with all our windmills, solar panels and unicorns, people are going to be really cold in their homes. I’m glad I live pretty far south.

Reply to  rbabcock
March 24, 2022 12:25 pm

The author is discussing extent and the number of days that extent hovers around the maximum so that’s what I analyzed. I have no problem with analyzing volume though. Are you thinking the behavior of the volume is exceptional or out of the ordinary this year?

Reply to  bdgwx
March 24, 2022 6:56 pm

No, just sea ice extent isn’t that much of a measure of most anything since it depends a lot on weather and whether the polar jet is zonal or meridional than “global temperature”. One huge storm can break up large areas of sea ice and mound it up on the lee side. Same amount of sea ice afterward, much less sea ice coverage.

Reply to  rbabcock
March 25, 2022 4:20 pm

Polarportal’s chart is perfect, except that its resolution is somewhat weak.

Here is an alternative representation, using NSIDC’s monthly data:

comment image

And here the same stuff over the year in anomaly form, NSIDC’s daily data:

comment image

ASIE just crossed the 2016-2021 average, moving a bit down. Could be worse.

Anyway, these daily considerations are somewhat boring.

No one should forget that 2002 was not only

  • the year with the least September extent level; it had also a pretty good ice growth season,
  • but also a year with an Antarctic extent above the 1981-2010 average for most of the year.

And that is the reason why I prefer a sorted list of yearly extent averages:

Arctic (ascending)

2016: 10.16
2020: 10.17
2019: 10.21
2018: 10.35
2017: 10.40
2012: 10.42
2007: 10.50
2011: 10.51
2021: 10.57
2015: 10.59

Antarctic (here is the descending sort more interesting)

2014: 12.73
2013: 12.47
2015: 12.37
2008: 12.21
2010: 12.06
2009: 12.00
2012: 11.98
2004: 11.94
2003: 11.92
1995: 11.75

2021 is at position 26 of 43; 2016 at 39, and 2017 at 43.

No idea by the way of what happens to the Antarctic:

comment image

The deviations in the 6 last years look simply tremendous.

March 24, 2022 9:17 am

They have to keep Griff in play somehow.

March 24, 2022 10:08 am

They didn’t leave it out. They think they already had to covered. Remember, this? Arctic Snow Depth, Ice Thickness, and Volume From ICESat-2 and CryoSat-2: 2018–2021 – Watts Up With That?

Because it was rotten, no good, thin ice the extent did not count.

BTW back in Feb. Tony Heller was saying:
Arctic sea ice extent is the highest in the past thirteen years and sixth highest this century.
“sixteenth lowest extent” | Real Climate Science

Reply to  rah
March 24, 2022 10:59 am

Rah don’t confuse us with facts…there is fear to peddle. Why can’t you feel for the plight of the polar bears and Santa?

Ron Clutz(@ronaldrc)
March 24, 2022 10:23 am

Yes, it was an early daily peak for Arctic ice this year. Of course, one day doesn’t signify much, the March monthly average is a better climate indication. The notable thing about 2022 Arctic ice is the low ice extent in Sea of Okhotsk. Okhotsk ice peaked at 850k km2 on day 63, compared to an average peak of 1085k km2. Of course that basin is outside the Arctic circle, has no polar bears, and is the first to melt out every spring. No drama there.

Last edited 2 months ago by Ron Clutz
March 24, 2022 10:48 am

I reckon all these Arctic sea ice reports should carry a standard warning label like stock brokers have to display –
“Disclaimer – this is not advice etc etc”


Because every year some numpties believe stories about global warming melting the Arctic, and set off to sail around the general area.

Of course they then get stuck in ice (who could have guessed?) and have to be rescued by taxpayer-funded naval assets.

March 24, 2022 11:21 am

Odd, I recall a story from either the late 1800s or early 1900s where a group took a trip to the arctic circle and it was over 100°F above the arctic circle during their trip. I should hunt down that article and ask someone to explain that event…

Janice Moore
March 24, 2022 11:25 am

“… for some reason.”


THANK YOU, Dr. Crockford, for your data-driven analysis so generously shared here.

Grateful for you!

Ireneusz Palmowski
March 24, 2022 2:35 pm

Current ice extent in the Arctic.comment image

Pat from kerbob
March 24, 2022 9:34 pm

Looks like a good year for those Rekya guys in Iceland that use icebergs to make vodka.
Lots of ice around

March 25, 2022 12:29 am

Am I wrong to secretly hope/pray that the summer ice totally disappears? There will be weeping and gnashing of teeth for a few months but when people realize that nothing bad will come of it, and in fact when the productivity of the sea goes off the charts with the Sun shining down on an ice free Arctic 24hrs a day in the summer they’ll see nothing but good things result.

Ireneusz Palmowski
Reply to  PCman999
March 25, 2022 12:39 am

Moreover, the water from the sea ice as less saline will be at the surface and will lower the sea surface temperature.

Reply to  PCman999
March 25, 2022 9:44 am

For the summer sea ice to totally (0 km2) disappear would likely require all of the land ice on Greenland to disappear.

March 25, 2022 1:34 am

Lies, damned lies ……..

Gerald Machnee
March 25, 2022 7:36 am

New Headline: It is higher than it was 16 years ago.

Robert Wager
March 25, 2022 10:24 am

the tenth lowest extent in the satellite record. ” So clearly things are not getting worse!

%d bloggers like this: