Most of the Arctic sea ice is on land and other WUWT musings

Arctic sea ice from 1953

 Willis Eschenbach has been wondering about sea ice trends of the past few decades. He's written a couple of articles but seems to me to be more interested in hiding the trends than exploring them. In today's article (archived here, latest here), he has used HadISST data from the UK Met Office Hadley Centre. I don't know why he chose that over the more often cited Sea Ice Index from the National Snow and Ice Data Center.  I think he's meant his title to this latest article to be sarcastic, in the way that the Dunning-Kruger set use sarcasm: "The Awful Terrible Horrible Global Sea Ice Crisis".

Willis decided to look at the data from 1974 only because he found that for Antarctica before that time there was not good data. Then he said he removed the seasonal component, which looks like he deducted something from each month. Since Willis used HadISST data, let's look at what the authors of the authoritative text on the subject found in the 2003 paper by Rayner et al:

Figure 1 | Northern Hemisphere sea ice. (g) (top curves: January; bottom curves: July) from GISST3.0 (black dots), Walsh (light blue), NIC (green), GSFC (red), NCEP (dark blue) and HadISST1 (black). Source: Rayner03

Figure 2 | Southern Hemisphere sea ice. (g) (top curves: September; bottom curves: January) from GISST3.0 (black dots), Bristol (light blue), NIC (green), GSFC (red), NCEP (dark blue) and HadISST1 (black). Source: Rayner03

Here's Willis' result for what it's worth. It looks nothing like any chart of sea ice that I've seen. Willis has the minimum before 2010, whereas most data sets would show the combined minimum in 2011 for one thing (see Figure 4 and Figure 10 below):

Figure 3 | Willis' wondering about sea ice and his strange chart. Source: WUWT (where else?)

Why did he plot monthly and then do something that he said removed the seasonal trend? He could have just plotted the annual average, which would have got rid of any seasonal effect. (Another question is why he lumped all the ice together.) But let's put that aside for the moment and instead check his findings with NSIDC data.

For another comparison, below is monthly data from NSICD, adding together the sea ice extent in the northern hemisphere with that in the southern hemisphere (not that I think it has much if any meaning). The maximum total extent was in October 1980 and the minimum total extent was in September 2011. The maximum and minimum are marked with a blue horizontal line. February 2016 was the third lowest extent of all months. (Hover over the chart to see more detail.)

Figure 4 | Total of the monthly average sea ice extent = Arctic + Antarctic, with LOESS smooth (60 months). Data source: NSIDC

Summer sea ice trends

Below are some charts showing the trends for the average of the minimum month in the northern hemisphere, southern hemisphere and total of the two months. I've plotted the northern and southern hemispheres with the same scale on the Y axis, so you can compare them more easily.  As you'll see, the loss of Arctic sea ice is what dominates.

First is the Northern Hemisphere and Southern Hemisphere minimum sea ice extent plotted on the same chart:

Figure 5 | Northern and southern hemisphere sea ice extent average for the month of the summer minimum: Arctic (September) + Antarctic (February). Data source: NSIDC

Then the sum of the northern and southern hemisphere averages for the minimum month of each. This is a combination of different times of the year - the purpose is to show the trend of the minimum sea ice extent all up:

Figure 6 | Total of the sea ice extent average for the month of the summer minimum = Arctic (September) + Antarctic (February). Data source: NSIDC

Winter sea ice

In winter there's also been a decline in Arctic sea ice extent while around Antarctica there's been an increase.

First the trend for the northern hemisphere in February together with the southern hemisphere for September, when the sea ice at each location is at its maximum extent:

Figure 7 | Arctic sea ice extent average for the month of the winter maximum (February). Data source: NSIDC

Then the trend for the two added as a total. Again this is a total of the months with maximum extent, which means that it is the sum of averages for February in the north and September for the south.

Figure 8 | Total sea ice extent average for the month of the winter maximum = Arctic (February) + Antarctic (September). Data source: NSIDC

Annual sea ice

I don't think the annual sea ice trends tell us much, because the winter sea ice extent swamps the summer minima. I've included them for completeness. First the northern and southern hemisphere on the same chart:

Figure 9 | Northern and southern hemisphere sea ice extent annual average. Data source: NSIDC

And the total of both hemispheres:

Figure 10 | Total sea ice extent annual average = Arctic + Antarctic. Data source: NSIDC

What did Willis find and why?

Compare the above with what Willis decided was happening. He put up a chart and wrote:

A contemplation of this figure reveals some interesting aspects. First, there is no significant trend at all in the 40+ years of satellite data. In other words, it seems the Awful Terrible Horrible Global Sea Ice Crisis has been cancelled due to lack of evidence.

It seems to me there are two reasons for him deciding there was no trend. The first reason is that he combined the northern and southern hemisphere. That's a rookie error, or perhaps a denier error. If he knew more about the world he'd know that the sea ice around Antarctica isn't bound by land. It can spread out as far as the wind and ocean currents will carry it. It might or might not be very thick but it can spread far. By contrast, the sea ice in the Arctic has barriers that prevent it spreading out. It's a better indicator of what is happening in the region than is the sea ice around Antarctica. The point is that by only considering the total global trend in sea ice, Willis isn't able to see what's happening at each end of the world. He can't see the large decrease in Arctic sea ice because he hid it in the total.

The second thing is, I don't know what he's done with his seasonal adjustment. He didn't say. All he said was that he removed the seasonal component. Looking at his chart, he probably just deducted the same amount from each month averaged over the entirety of the record. He didn't need to do that. If he'd just plotted the annual average like I did above, he'd have removed the seasonal influence and got the trend. The results are:

The northern hemisphere is losing ice at a rate of:

• 840,000 sq km a decade in the summer minimum month
• 450,000 sq km a decade in the winter maximum month
• 510,000 sq km a decade averaged annually over all months

The southern hemisphere is gaining ice at a rate of :

• 120,000 sq km a decade in the summer minimum month
• 210,000 sq km a decade in the winter maximum month
• 190,000 sq km a decade averaged annually over all months.

Sea ice as a whole is losing ice at a rate of (note - different months for max and min):

• 690,000 sq km a decade for the summer minimum months taken together
• 240,000 sq km a decade for the winter minimum months taken together
• 320,000 sq km a decade averaged annually over all months

Now losing 320,000 sq km of sea ice a decade is a loss of 32,000 sq km a year. That seems rather a lot to my way of thinking. More important is what is happening in the Arctic, where on average, there has been a loss of 84,000 sq km a year in the month of minimum sea ice extent.  The difference between last year's minimum and the 1979 minimum is 2.5 million square kilometres, that's just over 30% of the area of Australia or just over 30% of the area of the contiguous USA. It's 19 times as big as England, and more than the size of Greenland. This opening up of the sea water in the Arctic has an impact on climate, not just in the immediate vicinity but over a much larger area.

Arctic sea ice and global surface temperature

Willis also wrote:

We’re pretty sure that the global average surface temperature increased from the mid-1970s to about 1998 (Figure 4 below). However, we see no sign of this in the global sea ice area data (Figure 3 above). Instead, ice area remained stable throughout the 1980s and the 1990s, while temperatures climbed.

How about we just look at Arctic sea ice extent. Antarctica is another story altogether. I'll even just show the annual average, not the trend in the summer minimum. I've put both the global surface temperature and annual Arctic sea ice extent on the same chart. The temperature anomaly scale is on the left hand axis and the sea ice extent is on the right.

Figure 11 | Arctic sea ice extent - annual average, and global mean surface temperature (GISTemp) from 1979 to 2015.

Do you see the same inverse correlation that I see? You don't really need the statistics, do you? (You can do it yourself if you want it :))

Sea ice trends in the Southern Ocean

How is it that sea ice in the Southern Ocean is increasing while the world as a whole is warming? Well, as indicated above, the extent of southern hemisphere sea ice isn't a great indication of cooling (or warming), because there's nothing stopping the sea ice from spreading out. This topic is still being researched. There have been various reasons put forward which I've discussed previously, including possibly:

• ozone depletion affecting winds - although there are conflicting views on this,
• ice shelf melting has apparently been ruled out as a substantial contributor,
• wind changes amplified by ice-ocean feedback processes is a possible factor,
• internal variability.

Reflected sunlight

Willis waffled on about reflected sunlight. I don't know where he got his 10% and 0.1 W m-2 from. I think it was probably from some article he wrote at another time. He wrote:

Next, remember from above that a change of 10% in the global sea ice area translates into a global average of 0.1 watt per square metre (W/m2) change in reflected sunlight. From inspection of Figure 3, the sea ice area varied by ± 1 Mkm^2 around an average of just over 20 Mkm^2. This is a change of ± 5%, and thus should be accompanied by a change of ± 0.05 W/m2 in reflected sunlight … in other words, far too small to be measured.

I prefer to get my science from the experts than from a denier blogger. A few years ago (2008) Dr Michael Winton of GFDL wrote a paper "Sea Ice–Albedo Feedback and Nonlinear Arctic Climate Change". In it he wrote about the impact of Arctic sea ice loss on various aspects of climate.

In the Arctic itself there are a number of forces operating, including the change in albedo meaning change in reflected short wave radiation, change in outgoing longwave radiation from the open ocean vs the ice covered ocean, and changes in ocean circulation which are in part related to changes in salinity or density.  Therefore there is an impact beyond the Arctic and the paper discusses how changes will affect global surface temperature. I haven't studied the paper in depth so I won't attempt to say more about it. Suffice to say that it is be more complicated than Willis' simplistic number.

Another more recent, if simpler, analysis was by Stephen R. Hudson of the Norwegian Polar Institute. He wrote just about the sea ice albedo feedback:

...the globally and annually averaged radiative forcing caused by the observed loss of sea ice in the Arctic between 1979 and 2007 is approximately 0.1 W m^-2; a complete removal of Arctic sea ice results in a forcing of about 0.7 W m^-2, while a more realistic ice-free summer scenario (no ice for 1 month and decreased ice at all other times of the year) results in a forcing of about 0.3 W m^-2, similar to present-day anthropogenic forcing caused by halocarbons. The potential for changes in cloud cover as a result of the changes in sea ice makes the evaluation of the actual forcing that may be realized quite uncertain since such changes could overwhelm the forcing caused by the sea ice loss itself, if the cloudiness increases in the summertime.

That 0.1 W m^-2 from 1979 to 2007 just for the Arctic is about double the 0.05 W m^-2 Willis worked out for global sea ice for the period 1974 to 2015.

A rather good paper is by Mark C. Serreze and Roger G. Barry. They go into a lot of detail about the causes of Arctic amplification, and the paper is very easy to follow. There are lots of diagrams. The one below illustrates albedo feedback in different seasons:

Figure 12 | Schematics of the effects of albedo feedback on the surface energy budgets over land and ocean, contrasting the situation for (left) an unperturbed Arctic and (right) in response to a positive (warming) climate forcing. The dotted line is an arbitrary isotherm in the lower troposphere; in a warmer atmosphere the isotherm is found at higher level in the atmosphere. Compared to land areas, albedo feedback has a small direct effect on surface air temperature over the ocean; the feedback is instead largely associated with a stronger net surface flux (NSF) into the Arctic Ocean column. With the onset of autumn, the heat gained by the ocean in summer is released back to the atmosphere. Net surface heat fluxes over land areas tend to be much smaller than over the ocean. SW = shortwave radiation flux, LW = longwave radiation flux, Sens. Heat = sensible heat flux, Lat. Heat = latent heat flux, NSF = net surface heat flux. Source: Serreze and Barry (2011)

Polar amplification of global warming is expected to happen in the Arctic and over Antarctica. The IPCC AR5 WG1 report suggests it will happen sooner in the Arctic (it is already happening there), and later in the southern polar region. From Chapter 5, Box 5.1: Polar Amplification:

In response to rapid atmospheric CO2 changes, climate models indeed project an asymmetric warming between the Arctic and Southern Ocean, with an earlier response in the Arctic and a delayed response in the Southern Ocean (Section 12.4.3). Above the Antarctic ice sheet, however, surface-air temperature can respond quickly to radiative perturbations due to the limited role of latent heat flux in the surface-energy budget of Antarctica.

This article probably raises as many questions as it answers. Yes, sea ice is declining in the Arctic. Because it's declining so quickly, despite the fact that it's not declining in the Southern Ocean, there is an overall reduction in total sea ice.  The loss of sea ice is important not just because the loss is a positive feedback (it reduces the amount of sunlight reflected while the amount of heat absorbed by the ocean increases), but because it affects ocean circulation - probably more so in the Arctic than in the Southern Ocean, which is a lot more open. (Don't take my word as being authoritative on that last bit.)

From the WUWT comments

As usual, I've finished the article before reading the comments. It'll be interesting to see if anyone challenges Willis. He's not got universal support at WUWT. Here's a sample from the latest archived version. It looks as if most people took not a scrap of notice of what Willis wrote, but just used the article to post their own "thoughts" about sea ice:

Gloateus Maximus picks him up on one point, though I don't know that his point hits the mark all that well:

April 6, 2016 at 10:59 am
The effect on albedo depends on where the sea ice is lost. A ten percent reduction in Antarctic ice would have a much greater effect than the same loss in the Arctic, since SH ice extends farther towards the Equator.

There's not much sea ice at all in summer around Antarctica. In winter time it does extend quite a bit, but then the sun isn't too bright in winter, even though some of the sea ice would get some sunshine. Below is a map from NSIDC showing the maximum in September 2014 (blue line) compared with a couple of years from the 1960s:

Figure 13 | September sea ice southern hemisphere. This image compares Antarctic sea ice extent for September 2014 (blue line) with extent for September 1964 (red line) and August 1966 (black line). The dotted ellipse marked A shows the eastern Weddell Sea and the dotted ellipse marked B shows the eastern Ross Sea. Source: NSIDC 

This next comment from george e. smith prompted the title of this article:

April 6, 2016 at 11:51 am
Most of the Arctic Sea Ice is on land. There is more land North of +60 deg. Latitude , than there is south of -60 deg. latitude.
G

george e. smith got a two or three bites, and he responded:

April 6, 2016 at 1:28 pm
I said Arctic. That’s what I meant; NOT Antarctic.
So ok I lied it’s not SEA ice.
If you want to calculate the land / sea areas north and south of 60 degrees latitude, be my guest. I’ve already done that.
Yes of course I mean calculate both the land and sea areas.
G 

Johann Wundersamer agrees with george e. smith, or thinks he does, that there's more ice on land up in the north pole than down in the south pole.

April 6, 2016 at 2:05 pm
Anyway as george points more of the polar ice in the NH is on land than in the SH.

Nearly 85% of the world’s permanent ice is locked in Antarctic glaciers.


Phil. thinks there could be a new minimum this year in the Arctic:

April 6, 2016 at 11:39 am
The Global metric is perhaps not the best for seaice given it consists of the sum of two out of phase oscillations. Given the current value for the Arctic is the lowest for the date it’s conceivable that the minimum this summer/fall will be a new record surpassing 2012. However, even if it all melts the global metric would only hit ~15million km^2 (CT data).

Bob Boder points out that a low winter maximum doesn't always lead to a low summer minimum:

April 6, 2016 at 11:52 am
Phil;
look back at the recent past, low maximums don’t usually have much to do with low minimums. plus if you look at the north pole sea ice volume is not all that low.

Arch conspiracy theorist Rud Istvan thinks that the Arctic sea ice is in recovery mode. ristvan wrote:

April 6, 2016 at 1:11 pm
Phil, there is an oft understated issue with sea ice extent. The usual metric is that if 15% of a pixel is ice, the pixel is counted as ice. DMI discontinued 30% when it changed to a finer land mask rather than recompute the historical 30% with the new mask, because it wasn’t showing Aectic sea ice deterioration for the following reason.

15% ice means an ‘ice’ pixel can be 85% open sea. Therefore how compact or dispersed the sea ice is, especially along the Atlantic edges, greatly effects extent. That is affected by winds, waves, and currents. So this year’s ‘least meximum’ probably is an artifact of the metric. The ‘warmer’ Arctic winter Serreze whined about last week was still ~ -25C accordingmto DMI. Sea ice forms at -2C.

So my expectation is that the summer minimum about first week October will show the steady recovery from the 2007 cyclic low (2012 was an unusual August Arctic cyclone issue) suggested by Akasofu in 2010, and revealed by early 20th century DMI ice maps and by Russian summer sea lane records.

Anthony Watts must have a bit more free time on his hands than usual, because he posted a rare "thought", and above some charts of unknown (and mislabeled) provenance, showing Arctic sea ice and temperature, which stopped in the year 2007 or so, wrote:

April 6, 2016 at 11:27 am
“Arctic temperature anomalies in the 1930s were apparently as large as
those in the 1990s and 2000s. There is still considerable discussion of the
ultimate causes of the warm anomalies in the 1920s and 1930s.” (IPCC AR5)

Here is the quote in context, from section 10.5.1.1 Arctic and Antarctic Sea Ice in the IPCC report, if you're interested (my paras and emphasis), with Anthony's quote mine in italics:

A question as recently as six years ago was whether the recent Arctic warming and sea ice loss was unique in the instrumental record and whether the observed trend would continue (Serreze et al., 2007). Arctic temperature anomalies in the 1930s were apparently as large as those in the 1990s and 2000s. There is still considerable discussion of the ultimate causes of the warm temperature anomalies that occurred in the Arctic in the 1920s and 1930s (Ahlmann, 1948; Veryard, 1963; Hegerl et al., 2007a; Hegerl et al., 2007b). The early 20th century warm period, while reflected in the hemispheric average air temperature record (Brohan et al., 2006), did not appear consistently in the mid-latitudes nor on the Pacific side of the Arctic (Johannessen et al., 2004; Wood and Overland, 2010). Polyakov et al. (2003) argued that the Arctic air temperature records reflected a natural cycle of about 50–80 years. However, many authors (Bengtsson et al., 2004; Grant et al., 2009; Wood and Overland, 2010; Brönnimann et al., 2012) instead link the 1930s temperatures to internal variability in the North Atlantic atmospheric and ocean circulation as a single episode that was sustained by ocean and sea ice processes in the Arctic and north Atlantic.

The Arctic wide temperature increases in the last decade contrast with the episodic regional increases in the early 20th century, suggesting that it is unlikely that recent increases are due to the same primary climate process as the early 20th century. In the case of the Arctic we have high confidence in observations since 1979, from models (see Section 9.4.3 and from simulations comparing with and without anthropogenic forcing), and from physical understanding of the dominant processes; taking these three factors together it is very likely that anthropogenic forcing has contributed to the observed decreases in Arctic sea ice since 1979. 

Javier asked Anthony where he got the charts from, but there's no reply yet. Strangely, a Google search for similar images brings up a lot of sheet music :)

April 6, 2016 at 2:05 pm
What is the source of that figure, Anthony?
It is not in the Mahoney et al. 2008 article:
http://seaice.alaska.edu/gi/publications/mahoney/Mahoney_2008_JGR_20thC_RSI.pdf
It seems labelled in Russian

References and further reading

Rayner, N. A., D. E. Parker, E. B. Horton, C. K. Folland, L. V. Alexander, D. P. Rowell, E. C. Kent, and A. Kaplan (2003), "Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century", J. Geophys. Res., 108, 4407, doi:10.1029/2002JD002670, D14. (open access)

Winton, Michael. "Sea ice–albedo feedback and nonlinear Arctic climate change." Arctic sea ice decline: Observations, projections, mechanisms, and implications (2008): 111-131. DOI: 10.1029/180GM09 (pdf here)

Hudson, Stephen R. "Estimating the global radiative impact of the sea ice–albedo feedback in the Arctic." Journal of Geophysical Research: Atmospheres 116, no. D16 (2011). DOI: 10.1029/2011JD015804 (open access)

Serreze, Mark C., and Roger G. Barry. "Processes and impacts of Arctic amplification: A research synthesis." Global and Planetary Change 77, no. 1 (2011): 85-96. doi:10.1016/j.gloplacha.2011.03.004 (pdf here)

Stocker, Thomas F., ed. Climate Change 2013: the Physical Science Basis: Working Group I Contribution to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, 2014. (available here)

Mahoney, Andrew R., Roger G. Barry, Vasily Smolyanitsky, and Florence Fetterer. "Observed sea ice extent in the Russian Arctic, 1933–2006." Journal of Geophysical Research: Oceans 113, no. C11 (2008). DOI: 10.1029/2008JC004830 (open access)

A longer view of Arctic sea ice: 1953 to now - HotWhopper, March 2016