One of the fastest Arctic sea ice growths on record!

The Arctic sea ice extent reached a minimum a few days ago, to much consternation as usual at WUWT (here and here and here). Deniers don't much like reports about sea ice. It's a difficult measure for them. When there are open seas in the Arctic it's a stark reminder of just how much we are affecting the climate.

Below is a chart of Arctic sea ice extent, comparing 2016 ice extent to the average of recent decades and to the years 2007, 2012 and last year. This chart is from the National Institute of Polar Research. Click to view enlarged.

Figure 1 | Arctic sea ice extent for 2016, 2015, 2012 and 2007. Source: National Institute of Polar Research

Some of the information you can get from the above chart is:

• This year the ice extent was very low again, second only to 2012.
• For each of the past four and a half decades, Arctic sea ice extent has been shrinking all year around. The biggest shrinkage has been in the summer/autumn months.
• This year the ice extent was the lowest of all between April and July.
• The melt pattern can be very different in different years. In 2012, which has the record lowest extent of all, the extent was above the 2000's average between March and May 2012.

There are a number of places that provide information on sea ice. The National Snow and Ice Data Center (NSIDC) has a chart that compares this year's extent with that of 2012 and with the 1981-2010 average (which as you can see from Figure 1, is well below the 1980s average). Below is a copy of the latest chart, which also shows two standard deviations from the 1981-2010 mean.

Figure 2 | Arctic sea ice extent from June to October. The chart shows ice extent for 2016, compared to 2012 and the average for 1981-2010 with the grey bars showing +/- 2 standard deviations. Source: NSIDC

As I've pointed out on the Figure 2 chart, if you just go by statistics with no accounting for a downward trend, you'd expect only 2.5% to be below the grey shading on the chart. This year, the sea ice extent was below that 2.5% boundary for almost all the time from June this year. (Almost all the time in fact, as you can see in Figure 1, but the Figure 2 chart only shows from June to October.)

Anthony Watts discovers a non-discrepancy

Over at the climate conspiracy blog, WUWT, Anthony Watts has just discovered that the National Snow and Ice Data Center (NSIDC) uses a five day trailing average for monitoring sea ice extent (archived here). (To be more accurate, it wasn't Anthony who discovered anything. One of his readers noticed that the minima from daily extent data was different to the NSIDC report of the date of the minimum. Anthony just jumped at what he thought was a chance to show those blinky scientists that deniers can't be hoodwinked. Except they can. Deniers hoodwink each other all the time. It defines them.)

Given that Anthony Watts posts about Arctic sea ice so often you'd have thought he'd have known about the NSIDC use of the five day average by now. But no. He doesn't. He devoted an entire article to the subject, with the headline: "Discrepancy in NSIDC press release vs. data puts turning point for end of Arctic ice melt 3 days earlier."

What he meant by that is that the date NSIDC gave for the minimum was 10 September. That's using their normal five day trailing average, which puts the minimum at 4.14 million square kilometers (1.60 million square miles). If they'd used the daily data it would have been 4.08 million square kilometers on 7 September, three days earlier. Anthony didn't bother checking back to see how it was reported in previous years. (Hint for Anthony: They have used the five day trailing average since 2012. Before that they used a five day centred average.) Instead Anthony wrote that he couldn't find the answer to his puzzle:

It is a puzzle. Perhaps whoever wrote the NSIDC press release looked at their 5 day average value in their Chartic interactive graph instead of the raw data? When using that tool, the data rounds up to 4.14 as you can see:

Rather than check for himself, it looks as if Anthony is going to be big and brave and, after taking a deep breath and whatever else gives him courage, is going to ask the NSIDC after the weekend. He finished his article with this:

NSIDC is closed for the weekend, so perhaps we will get an answer to this puzzle on Monday.

You don't have to wait till Monday, because the answer is there on the Internet. Plain as day. If you can't be bothered going to the prominently displayed FAQ, then all you have to do is look at the charts, like Anthony did. These are plotted using the value of the five day average. You can see how past years minima line up with the past reports, using the archives at NSIDC. (Use the sidebar on the right to get the past reports for September in any year.)

If you are feeling more adventurous you can download the data for yourself, plot the five day trailing averages, and see how they line up with past archived reports of minima.

Anthony Watts deviates and fails statistics

Anthony Watts is either lazy or incompetent. He's also not shy about showing his ignorance. He's written a lot of articles about Arctic sea ice lately. In one of them he wrote (my emphasis):

While Romm and others wail about “death spirals” since 2008 when the term was first coined by Mark Serreze, here we are 9 years later, and Arctic Sea Ice extent has been just slightly below two standard deviations for most of the melt season, and isn’t close to the 2012 event caused by a large storm which broke up and dispersed ice.

Three points about that. Well, four if you think of how silly it is of Anthony, who is one of the dumbest people in the climate blogosphere, to criticise two of the smartest people when it comes to climate stuff. The other three points are:

• Arctic sea ice extent was the lowest in the record from May 2016 to July 2016, below that of 2012.
• If the ice extent is below two standard deviations, it's below at least 97 1/2 % of what would be expected.
• Every decade that passes, the average that the standard deviations come from, is dropping (see Figure 1 above).

We're facing "one of the fastest ice growths on record" - duh!

In his latest article, Anthony did a Denial Depot, quoting "climate hoaxer" Paul Homewood. Anthony wrote:

Paul Homewood passes these point on via email:

1. Earliest minimum since 1997 – shows how cold it is there
2. This year extent was 22% above 2012, despite two massive storms
3. Thickness is way up on 2010 and 2011
4. Already extent is above 2007, as well as 2012, for this date
5. We are looking at one of the fastest ice growths in September on record.

Here's my take on this:

1. The early minimum shows how warm is was there early in September, as much as it shows how cold it is there at any time of the year! The minimum this year was four days later than the minimum in 1997, and one day earlier than the minimum in 2015 (and 2010). For some trivia - the latest minimum was in 1989 and the earliest was in 1987.
2. His point 2 shows what an extraordinary melt took place in 2012. This years minimum was 4.083 m sq km, which is indeed 22% more ice extent than the lowest in 2012. It's also 7% less than 2015, 18% less than 2014 and 19% less than 2013.
3. There's no link to or evidence provided for point 3.
4. "Already" - I mean "ALREADY"? Not even a week has passed and they are talking about "well up". Sheesh. Look at the chart! On September 15 this year, the ice extent was measured at 4.387. In 2007 it was measured at 4.167. In 2007 the minimum was reached on 14 September!
5. This is the funniest of the lot. I bet the rate of freezing can't beat the rate of freezing after the melt season in 2012. Any takers? Denial Depot did it well - here and here. 

How deniers see it vs how the rest of the world sees it:

For the slow: the year of the fastest "recovery" from the minimum extent was, not surprisingly, the year of the lowest extent - 2012. The next fastest was the year of the next lowest extent - 2007.

From the WUWT comments

There are some real classics in the comments to the various WUWT articles. Did you know that a melting Arctic means the world is getting colder? Well, that's what some of the climate conspiracy theorists are arguing.


Greg takes the view that just because the Arctic sea ice trend has been down since the 1970s, it doesn't mean that the Arctic ice is melting, or it doesn't mean it's melting because the world is getting warmer. Or something like that.

September 16, 2016 at 12:13 am
A linear trend is bullshit in anything but a trivial linear system. That is what the mentally challenged pseudo-scientists in climatology fail to realise.
Just because you load any set of data into excel and fit a “trend” does not mean a trend has any physical meaning or reveals any understanding of what is happening.
It is an arbitrary model.

If Gabro is still around in 60 years time, it's extremely likely he or she will see an ice-free Arctic. It will probably be virtually ice free in summer by the 2030s.

September 15, 2016 at 4:40 pm
Ton,
Your prognostication that sea ice will keep decreasing is an opinion, presumably based on nothing more than extrapolation. The trend is your friend, until it isn’t any more.
We have millennia of data demonstrating the fact that, just in the Holocene alone, sea ice has fluctuated far more than “observed” since 1979. What goes down, must go up. Sooner or later. And sea ice follows the same ~60-year cycle as the ocean oscillations. If you wanted to be scientific, you’d wait for at least that long before making not just a prediction but asserting a certainty that isn’t in evidence.

MarkW doesn't believe that this is the second lowest minimum in the record. He's hanging on to the years of extraordinary melts in 2007 and 2012.

September 16, 2016 at 6:35 am
It’s been 4 years since the lowest and 9 years since the second lowest. Yet they keep telling us that next year is the year the collapse will begin.
Denial, it’s not just a river in Africa.

Remember how twenty years of satellite records of tropospheric temperature used to be enough to "prove" to science deniers that global warming stopped or wasn't happening? Well, take a look at this comment. RobRoy wrote how neither 37 years nor 150 years is sufficient to draw any conclusions:

September 16, 2016 at 10:13 am
“The record lowest extent in the 37-year satellite record…
37 years??
We’ve only just begun to measure Arctic sea ice extent.
Given this scant 37 years of satellite data,
We really know nothing of the history of Arctic
sea ice extent.
Any reference to “history” on this subject is a thin argument at best.
(150 years of temperature data hardly describes “history” either.)

Richard Keen fails critical thinking 101.  Maybe he has a serious case of confirmation bias.

September 15, 2016 at 4:15 pm
“It was a stormy, cloudy, and fairly cool summer,” said NSIDC director Mark Serreze. “Historically, such weather conditions slow down the summer ice loss, but we still got down to essentially a tie for second lowest in the satellite record.”
Thus spoke the “Arctic is Screaming” Serreze.
So what he’s saying is that ice loss is disconnected from warming, and blows his whole line about “Warming” causing ice loss. Perhaps he’s admitting that there’s so many causes that it looks randon to us mortals – or chaotic if you prefer.

Chris tries to set him straight, but my guess is Richard Keen is too far gone and will never recover.

September 16, 2016 at 10:22 am
“So what he’s saying is that ice loss is disconnected from warming, and blows his whole line about “Warming” causing ice loss.”
No, he is not saying that at all. He is saying that the ice loss would’ve been worse, if not for the stormy and cloudy weather, which is atypical for that time of year in the Arctic. It is unlikely that those same storms and cloudiness will occur in the coming years, which will mean the rate of ice loss may well accelerate.

If you thought what Richard Keen wrote was bad, how about the twisted illogic of TLMango: Arctic ice melting means cooling don't you know:

September 15, 2016 at 4:46 pm
We know that temperature data since the 90’s has been
rolled to create the false impression that every year is the
hottest ever. There is a cooling on the horizon which is
evidenced by the melting in the arctic. The earth expels
it’s heat at the poles and as the oceans give up their heat
even more ice will melt.

qbagwell is another one who doesn't think the 37 year satellite record is long enough to draw conclusions. (Does he or she think the satellite troposphere record is long enough for anything?) Thing is, Arctic melt not the only sign of warming. It needs to be seen in context of the very rapid warming of the planet as a whole, and the increase in greenhouse gases.

September 15, 2016 at 8:58 pm
37 years of data and we’re supposed to draw long term conclusions? If this year is 22%more than the 2012 minimum I wonder why he expects the coming years to be lower minimums? What trend would he attribute that to? Just looking at the history it would appear that the likelihood if higher minimums would be better than lower ones in the coming years. Unless of course you believe this is all being manually controlled by the trace gas CO2, then of course you might wonder why it’s 22% above 4 years ago with CO2 levels higher now than then? To me , it doesn’t appear to be alarming at all.

That's not atypical of the illogic of dim deniers. Can you see the logical fallacy? What qbagwell is arguing is that there should be no year to year variability within a long term trend. That reality should be a perfectly straight line with no variations, otherwise global warming is a hoax. That's a bit like arguing that if a man is not 170 cm tall (or whatever the average height of a man is these days) then he is not a man.