Jim Steele brings the Arctic to Antarctica

There's a very odd article at WUWT by Jim Steele about Antarctic sea ice (archived here, latest here). He started off the article saying he'd just read a new paper about Arctic sea ice. The paper, by Neil Swart and colleagues was published in Nature at the end of January.

The introduction to the paper sets the scene:

Internal climate variability can mask or enhance human-induced sea-ice loss on timescales ranging from years to decades. It must be properly accounted for when considering observations, understanding projections and evaluating models.

The scientists were looking at trends in Arctic sea ice in recent years. What they were looking at in particular was the extent to which internal climate variability can affect the trend. That is, how much of the ups and downs in Arctic sea ice could come from internal variability compared to the underlying decline from enhanced greenhouse warming.

The main message from the paper, I think, is that Arctic sea ice decline is not necessarily underestimated by climate models. The recent big dips of 2007 and 2012 could be natural variability. It's difficult to tell.

What is the trend?

The authors described the trends in minimum September ice extent as follows:

Arctic sea-ice extent was lost at a considerably higher rate from 2001–2007 than in the preceding decades (Fig. 1), which caught the attention of scientists and the public alike2. In contrast, from 2007–2013 there was a near-zero trend in observed Arctic September sea-ice extent, in large part due to a strong uptick of the ice-pack in 2013, which has continued into 2014.

I admit at first I thought this way of looking at the ice extent a bit unusual. Particularly when the extent in 2012 was the lowest ever recorded. Here is Figure 1 from the paper, so you can see what I mean. I've written on the chart, the 2012 record low summer minimum.

Figure 1 | Arctic September sea-ice extent anomalies. Sea-ice extent anomaly relative to 1980–2000 from observations (red) and 102 realizations from 31 CMIP5 models (grey), along with the CMIP5 ensemble mean (black). Linear trends are fitted to the observations over 2001–2007 (green) and 2007–2013 (blue). The CMIP5 ensemble mean is calculated such that each model has a weight of 1. Observations extend to 2014.
Source: Swart15

For the rest of the paper the authors were talking about a near zero trend for seven years. Why not look at the trend from 2001 to 2012 I wondered. Why cut it off at 2007? I figured there was a very good reason. What the authors say about it is this (my emphasis):

By deliberately cherry-picking these periods we will demonstrate how using short-term trends can be misleading about longer-term changes, when such trends show either rapid or slow ice loss.

So it looked as if they deliberately but arbitrarily picked the start and end periods to prove a point. It wasn't because they necessarily see a distinct change in trend in 2007. I think that's important. The reason it's important is that deniers are now claiming that Arctic sea ice decline has been flat for seven years, when that's clearly not the case.

Here's a chart of Arctic sea ice minima to show you what I mean. The animation compares the trend from 2006 to 2014 with the trend from 2007 to 2014. Just one year makes all the difference between a positive and negative slope.

Data source: NSIDC U Colorado

The trend accelerated about seventeen years ago

Being a fan of the blog of one of the authors, Ed Hawkins, I decided to pay him a visit and see if he had something more to say on the subject, And I was in luck. He wrote an article just a few days ago. His explanation was rather neat. Ed wrote:

Over the past 35 years, the extent of September sea-ice has reduced by about 35% overall. But, this decline has not been smooth. The linear trend over the second half of this period is larger than over the first half, suggesting an acceleration, and in 2007 and 2012 the summer extent was dramatically lower. 

That cleared up one point for me. That being that the recent linear trend has been steeper than it was in the first 17 years or so of monitoring sea ice extent. In light of that, here are two more charts. This time the comparison is more in line with what Ed wrote about a change in the trend:

• linear trend over the whole period
• split trends for the periods 1979 to 1996 and 1997 to 2014.

Data source: NSIDC U Colorado

So it looks to me that neither Ed Hawkins nor his fellow authors are postulating that there has been a zero decline/trend in the past seven years.

Be wary of extrapolating a long term trend from short term dips and bounces

What the paper was exploring, in part, was the possible behaviour of Arctic sea ice as it disappears over the next few decades. In particular, they showed that you cannot judge the long term trend by short term fluctuations - whether they be big dips as occurred in 2007 and 2012, or hikes as has happened in 2013 and 2014. That's the reason they cherry picked specific start and end dates for different trend lines. They were wanting a trend line close to zero and they got one by carefully selecting the start and end year. By starting at a low in 2007 and ending high in 2013, they were able to draw a trend line that was close to zero. But they are not arguing that there was in fact a shift to a zero trend from 2007. (That won't be known for some time yet, if it happens.) This period was picked purely for the sake of the exercise.

Ed made an analogy with a bouncing ball, for which he gave credit to Richard Betts. He wrote:

Imagine a ball bouncing down a bumpy hill. Gravity will ensure that the ball will head downwards. But, if the ball hits a bump at a certain angle it might move horizontally or even upwards for a time, before resuming its inevitable downward trajectory. This bouncing ball is an analogy for the behaviour of Arctic sea-ice. 

That analogy (and my previous interpretations) seem to be confirmed by a chart Ed Hawkins had on his blog, showing how in the short term the September ice extent bounces around. Over the long term though, there is an unmistakable decline:

The paper goes into quite a bit of detail, exploring the extent to which internal variability can affect the growth and decline of Arctic sea ice. Based on their analysis of particular climate model runs, they were able to demonstrate that internal variability could have played quite a part in the recent sharp declines and similarly could, in the future lead to extended "pauses" in the ongoing decline, and bounces.

When will the Arctic be "ice free"?

Ed explains in closing that the Arctic will most probably become ice free sometime in the next thirty five years. He wasn't any more definite than that. He said it depends on emissions, of course. And I got the impression he rejected the notion of an ice free Arctic for at least ten years.  By the way, ice free doesn't really mean no ice. It means less than one million square km of ice in summer.

Ed wrote (my emphasis):

The expected outcome is that the long-term decline in Arctic sea-ice will continue as global temperatures increase. There will be further bounces, both up and down. Individual years will become ice-free sometime in the 2020s, 2030s or 2040s, depending on future greenhouse gas emissions and the natural variations. But, even at the bottom of the hill the ball will continue to bounce – not every year will have zero ice in summer. The bounces will become smaller if global temperatures continue to increase, and other summer and autumn months will also become ice-free. However, in future scenarios with sharp reductions in greenhouse emissions, the Arctic sea-ice recovers slightly as global temperatures decline. 

There is quite a lot to digest in the Nature commentary, helped by Ed's blog article.

Jim Steele misses the point completely

Jim Steele of course didn't understand much at all about the paper. Neither its purpose nor its findings. He wrote:

The paper might be better titled a Statistical Justification For The Pause In Arctic Sea Ice Melt as they concluded, “Thus, pauses in sea-ice loss, such as seen over the past eight years, are not surprising and are fully expected to occur from time to time.” In other words, we should still trust the models and ignore skeptics who cherry-pick the current pause and thickening of sea ice.

Except that if there's a "pause" it's a two season "pause". Not what most people would call a pause. You could say the paper was a gift to ignorant deniers like Jim Steele. Even though right up front the authors say they deliberately cherry-picked those periods, deniers gloss over that and ignore the 2012 record minimum and jump into assuming there really has been a flattening of Arctic sea ice over the past seven years.

Jim Steele is very silly

Jim then went all silly. He wrote (my emphasis):

However the “elephant” mired in the thickening Arctic ice was, if this paper was truly anything more than an excuse for the lack of an Arctic sea ice death spiral, and the “background of long-term radiatively forced sea-ice decline” is a global phenomenon, then why wasn’t their analysis extended to the condition of global sea ice and Antarctica? Why cherry-pick just the Arctic?

Huh? I can see you reading that again and again, thinking I must have misquoted Jim Steele. I didn't. Does he really and truly think that summer ice in the Arctic hasn't already declined a huge amount?  And did he really did expect a paper on Arctic sea ice to be about ice at the opposite end of the world?

One might as well ask: why didn't the authors write about the snows of Kilimanjaro, or precipitation in the Simpson Desert.

It gets much worse than that

Wait for it.

Jim posted a chart from the Nature paper, which showed the probability of a pause as a function of pause length - for Arctic sea ice decline - and, yes - you guessed it. He applied it to sea ice around the continent of Antarctica.

Can you believe it? Yes, you probably can. If you know Jim Steele by reputation then you can accept it only too well. Even if you don't, you probably by now know the reputation for pseudo-science crankery at WUWT. That means you know that it's all too possible for a "guest essayist" to not know the difference between the Arctic and Antarctica.

Jim wrote:

Without access to their models, I can’t directly ascertain their statistical probability of a pause in Antractica’s hypothesized sea ice decline, but their Figure 3B (below) suggests the probability is zero. The black line represents the modeled probabilities of a increasing pause‑lengths based on observational data. A probability of a 30‑year pause (or increase in sea ice) between 1979 and 2013 is clearly zero.

That was just puffing himself up. I doubt he'd be in a position to "ascertain their statistical probability" any more than he could accept climate science. He certainly wouldn't be able to assess the probability of a pause in Antarctica's sea ice decline by looking at their Figure 3B or 3C for several reasons, including:

1. Figures 3B and 3C relate to the Arctic, not Antarctica
2. Figure 3B is about the probability of seven year pauses, not 30 year pauses. And, contrary to what Jim wrote, it doesn't show anything with zero probability, although RCP8.5 gets close around 2070
3. Figure 3C shows that in the Arctic, the probability of a 30 year pause in the Arctic depends on the emissions scenario
4. Sea ice around Antarctica isn't declining - it's growing.

Jim also misrepresents the charts. The black line in Figure 3C is based on historical RCP4.5 models, not on observations.

Here are Figures 3B and 3C from the paper. I've annotated the charts with the relevant pathway. First Figure 3B, which is about the probability of a seven year pause in the decline of September sea ice in the Arctic, over a 21 year rolling window:

Figure 3 | Probability of a pause in September Arctic sea-ice extent. b, Probability of a 7-year pause over a 21 year rolling window. Source: Swart15

Next Figure 3C, which is about the probability of a thirty year pause in the decline of Arctic summer sea ice. I've similarly put labels on the curves. The future curves relate to probabilities from 2066 to 2100. The historical relates to RCP 4.5 CMIP5 experiment over the period 1979 to 2013.

Figure 3 | Probability of a pause in September Arctic sea-ice extent. c, Probability of a pause as a function of pause length in the Historical-RCP4.5 experiment over 1979–2013 (black), and in the future over 2066–2100 under the RCP2.6 (blue), RCP4.5 (cyan) and RCP8.5 (red) experiments. The horizontal dashed line represents a probability of p = 0.05. A pause is a period with a trend ≥0. Only ice extents ≥1 x 106 km2 are considered.
Source: Swart15

You can read yourself the details of the findings. About Figure 3C, for example, the paper states:

Pause lengths of up to 32 years are seen in the CMIP5 ensemble over 1979—2013, and pauses of 20 years or less occur with p > 0.05. Pauses of longer duration become more likely towards the end of the century under RCP2.6 and RCP4.5 (Fig. 3c). Under RCP2.6, even pauses of 35 years occur with a probability p > 0.4 in the CMIP5 simulations over 2066–2100. Under RCP8.5, pauses of all lengths become less likely over 2066–2100 than they were in the historical period. Clearly the emissions scenario is an important factor in the future evolution of Arctic sea-ice, which we now consider in more detail.

Jim really didn't understand the paper or the charts he showed. He wrote:

Their study was strictly a statistical analysis, independent of the various causes that might be attributed to the changing sea ice patterns at either pole. So it doesn’t matter how many hypothetical reasons may be conjured up to explain Antarctica’s growing sea ice.

Yes, it was a statistical exercise, looking at probabilities. Where Jim went wrong was where he claimed it was "independent of various causes". It wasn't. It was dependent on the known physics and chemistry of climate science, determining internal variability and the greenhouse effect.

The very fact that the charts show various different Representative Concentration Pathways should have told Jim that at least some of the "causes" were considered - that of different greenhouse gas emissions. If he knew anything about the CMIP models then he'd have known that the exercise also accounts for internal variability. The science is built into the models themselves.

Poor Jim. He dived deeply into his mess. It really does look as if he thinks models of the Arctic September sea ice extent can be used for Antarctic sea ice for all years and all months. He put up a monthly chart of Antarctic sea ice and wrote:

Based on their analyses, the models’ inability to predict a 30‑year trend in growing Antarctica sea ice "despite strong anthropogenic forcing", can not be explained by CO2 driven models or random variability. 

Ha. Once again, "their" analysis was of Arctic sea ice, not Antarctic sea ice. They are completely different. Jim wrote that nonsense so he could sound all sciency and say:

As I have argued before Antarctica sea ice growth is a better indicator of climate change and there are very good reasons to believe the loss of Arctic sea ice is better explained by ocean and atmospheric oscillations.

I got curious and wish I hadn't. I'm not about to rebut his very long article. You can read it here if you want to. (If you're not up with the science I wouldn't recommend it. You may not be able to tell the good science from the bad. Jim Steele is not an expert in climate science nor an expert in sea ice. To my knowledge he's never published a science paper in his life. He's been involved in one or two land rejuvenation projects. He used to be a part time manager of a field office for a university in California until he retired to take up climate science denial full time. He is becoming an expert in the art of pseudo-science and the Gish gallop.)

The bottom line, according to Jim Steele, is that the Arctic summer sea ice decline has nothing to do with global warming. The extending sea ice in Antarctica is a much better measure, according to Jim. His reasoning was convoluted and wasn't grounded in science. Something to do with the circumpolar currents down south keeping out warm water. If Arctic summer ice were growing and Antarctic ice declining, would he be arguing that it was the Arctic sea ice that was a better measure of global warming?

Antarctic and Arctic summer sea ice trends

Here is a chart showing the ice extent in the months where it is the lowest in the southern and northern hemispheres. September up north and February down south.

Data source: NSIDC U Colorado

According to the NSIDC, in the Arctic, the ice extent last September was the sixth lowest on record at 5.02 million square kilometers. The rate of loss (for September) over the period since 1979 is 829,000 sq km a decade. In Antarctica the ice was the highest in the satellite record at 3.82 million square kilometres. The rate of gain (for February) over the period since 1979 is 129,000 sq km a decade.

From that you can see that the Arctic is losing a lot more ice than Antarctica is gaining, when it matters. That is, as far as albedo goes.

Why is sea ice increasing in the southern hemisphere?

There is no clear cut answer to this question. There's a recent paper in GRL by M‐È Gagné et al, which describes some of the ideas that have been considered. I'll just list them below:

• ozone depletion - 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

The paper states that "model simulations forced with all known anthropogenic and natural forcings show on average a small but significant negative trend in Antarctic SIE, although the intermodel spread between such trends is large". So what is observed (an increase) isn't consistent with the average of the models. However that doesn't mean that the increase in sea ice proves global warming isn't happening. It clearly is happening.

I won't go into any more detail on this topic. You can read the paper here, it's open access. It looks to me as if the question is still open.

From the WUWT comments

I wrote most of this article a day ago. I went back today (archived here) expecting to see several comments pointing out the wrongness of extrapolating results of a paper on the September sea ice in the Arctic to sea ice in the southern hemisphere.

There was one. That's all I found. One comment, which is the second last of 112 comments. Written by Barry a full twenty hours later:

February 4, 2015 at 2:47 pm
The study was of ARCTIC sea-ice (perhaps they should have put it in all caps in the title), so you can’t infer anything about Antarctic sea ice from the graph shown. We all now the northern hemisphere is warming faster than southern because it has more land mass and thus less heat capacity on/near the surface.

This is so obvious that I can only guess it is yet another baiting tactic by Mr. Steele, and I will anxiously await for him to “hook” me with his reply.

The other comments were mainly nonsense and conspiracy theories of the type you'd expect at WUWT. For example Will Nitschke seems to think there aren't any scientific papers on southern hemisphere sea ice:

February 3, 2015 at 7:06 pm
I suspect if they’d also done an analysis of Antarctica they would never have got their paper published as that would have called into question their statistical assumptions. The speculations are always hidden in the assumptions. The statistical calculations afterwards merely give a veneer of mathematical certainty.

---

Ed Hawkins: "Arctic sea-ice decline erratic as expected" blog article at Climate Lab Book, 28 January 2015

Swart, Neil C., John C. Fyfe, Ed Hawkins, Jennifer E. Kay, and Alexandra Jahn. "Influence of internal variability on Arctic sea-ice trends." Nature Climate Change 5, no. 2 (2015): 86-89.  doi:10.1038/nclimate2483 (pdf here)

Gagné, M‐È., N. P. Gillett, and J. C. Fyfe. "Observed and simulated changes in Antarctic sea ice extent over the past 50 years." Geophysical Research Letters (2014). DOI: 10.1002/2014GL062231 (open access)