Smooth

NASA’s Goddard Institute for Space Studies (GISS) has updated their global surface temperature estimate to include November 2013. It turns out that this most recent November was, globally, the hottest on record:

Greg Laden posted about it (and other things) recently in his continuing efforts to let people know what’s really happening to the globe (it’s still heating up) as well as spreading the word that “earth” includes a lot more than just the atmosphere. He featured this version of the graph (provided by “ThingsBreak” but prepared by Stefan Rahmstorf):

Of course this means that the fake skeptics must come out of the woordwork. Referring to the smooth (the red line on the graph), here’s what Paul Clark had to say about it:

It’s not clear how this red line was obtained. The red line is not described on the poster’s page. The graph comes from, what Laden describes as, “climate communicator” ThingsBreak. What on earth is a “climate communicator”?!

It seems to be some type of smoothed moving average. Five year spline perhaps?

Problem is, the red line is roughly in the middle of the blue line, except at the end. At the end, the red line is not in the middle at all, but is down at the beginning, and up at the end, of that final 10 year period. It’s shooting right up at the end!

How can that be? I therefore find this line to be completely made up, and a case of wishful thinking.

---

Here’s something about which every honest participant in the discussion of man-made global warming should think. Carefully. Namely, this: Paul Clark complains that it’s not clear how the red line (the smoothed version of the data) was obtained. Furthermore, it doesn’t seem right to him. How does he react?

Did he acquire in-depth knowledge of smoothing techniques? (I can tell you for a fact: no he didn’t.) Did he consult a disinterested expert? (Apparently not.) Did he, oh I don’t know, maybe ASK how it was obtained? (Nope.)

You see, those are some of the ways an actual scientist might proceed. The guiding principle being this: LEARN MORE ABOUT THE SUBJECT *BEFORE* YOU OPEN YOUR MOUTH.

It seems that’s not Paul Clark’s way. He doesn’t think the smooth (red line) looks right, but with little to no effort at all to find out about it, he declares that it is “completely made up, and a case of wishful thinking.” I declare that Paul Clark’s opinion is completely mistaken, and just about as clear a case of the Dunning-Kruger effect as you’re likely to find.

Here’s something else worth thinking about: suppose I wanted to make the slope at the end artificially large. What smoothing method — other than “force it by hand” — could do that?

---

Rahmstorf used a smoothing method based on MC-SSA (Monte Carlo singular spectrum analysis, Moore, J. C., et al., 2005. New Tools for Analyzing Time Series Relationships and Trends. Eos. 86, 226,232) with a filter half-width of 15 yr. I get a very similar result using my favorite method (a “modified lowess smooth”) with about the same time scale.

My modified lowess smooth is in agreement with Rahmstorf’s MC-SSA smooth. Here’s just the modified lowess smooth (in red), a plain old plain-old lowess smooth (in green) for those who don’t trust me to modify anything, and a spline smooth (in blue):

One of the things I like about my own smoothing program is that it also calculates the uncerainty of the result. Here are the three smooths I computed, together with dashed red lines to show the range 2 standard deviations above and below:

The three methods are in agreement, within the limits of their uncertainty. Clearly.

Now let’s take the range of the modified lowess smooth which we plotted in the previous graph, and add some other smooths set to about the same time scale for smoothing: an ordinary moving average in black, a Gaussian smooth in green, and a 6th-degree polynomial (as used by Paul Clark himself) in blue:

The moving-average line stays within the range indicated by the modified lowess smooth, but that’s easy because the moving averages don’t extend to the ends of the time series, we lose years at both the beginning and end. The Gaussian smooth stays within the range indicated by the modified lowess smooth except at the end, when the Gaussian smooth levels off. Is Paul Clark wondering why that might be? Does he know enough about smoothing in general, and about Gaussian smoothing specifically, to have expected that? I did.

Perhaps most interesting is the 6th-degree polynomial, which wanders outside the modified lowess range, not just at the beginning or end but in the middle as well. What’s really interesting is why it wanders outside the range, because it happens for different reasons at different times! The 6th-degree polynomial fit smooths too much in the middle of the time span, but smooths too little near the endpoints. Is Paul Clark wondering why that might be? Does he know enough about smoothing in general, and about polynomial fits specifically, to have expected that? I did.

Ordinarily, this is where I would launch into a technical discussion of smoothing. Why do certain methods tend to go one way more than another? What should one expect near the endpoints of the time span? How do smooths with longer time spans compare to those with shorter time spans? Why is the Gaussian smooth questionable near the endpoints? Why do high-degree (and 6 is a pretty high degree) polynomial fits really really suck as smoothing methods, especially near the endpoints of the time span. Yes, they really suck, and the reason is actually quite interesting.

But I’m not gonna. At least not yet. It’s not my job to educate ignorant Dunning-Kruger victims about smoothing techniques.

But here’s an offer for Paul Clark: Come to this blog, find this thread, and post a comment in which you admit — without a bunch of caveats or excuses or bullshit — just admit in no uncertain terms that you don’t know enough about smoothing to know how valid Rahmstorf’s MC-SSA smooth is or why your 6th-degree polynomial choice is a really really sucky choice. You don’t have to weep and moan, just simply admit that you don’t know enough about this topic to justify your opinion. You don’t have to admit anything else, just that you’re ignorant about smoothing methods. Don’t clutter the comment up with unrelated stuff, if you want to spew about other things put that in a separate comment. Just a single, simple admission of ignorance on this topic.

If you’ll do that, Paul Clark, then I’ll do a blog post on smoothing. Or maybe two. Maybe even three — it’s a topic of great interest for me. How ’bout it, Paul? All you have to do is admit that you’re ignorant of the subject, and I’ll educate you.

In case that offer isn’t acceptable, here’s another. Paul: I’ll blog about the topic and you don’t even have to admit anything. But if you want me to supply some lessons without you admitting your ignorance — pay me. Cash American.

Fire Down Below

Australian prime minister Tony Abbot got elected saying, among other things, that global warming science was a bunch of crap.

Now he says that “Climate change is real, as I’ve often said, and we should take strong action against it…” Why the amazing massive ginormous flip-flop? Because Abbot is feeling the heat. So are a lot of Australians as they suffer through tremendous bushfires devastating huge areas of New South Wales. Australia has always been prone to fire, but the scale of this event is astounding. So too is the timing — it isn’t even summer yet down under. But it’s absolutely clear that “fire season” has been getting longer in Oz, starting earlier and ending later. And the reason for this very early outbreak: an extra-hot and extra-dry winter, exacerbated by — you guessed it — man-made climate change. Global warming.

The ICP report

Many of you are probably aware of a “report” which is intended to contradict the IPCC (Intergovernmental Panel on Climate Change) report. Its authors call it the “NIPCC” report for “Non-governmental International Panel on Climate Change.” It’s supposed to represent the very best that so-called “skeptics” have to offer.

Bob Tisdale pisses on leg, claims it’s raining

Global warming deniers really hate the fact that a proper comparison of computer model projections to observations does not show that “models fail.” But they love faulty comparisons.

Double Standard

Since 1975, global average surface air temperature has increased at a rate of 0.17 deg.C/decade (estimated by linear regression using either the NASA GISS or HadCRUT4 data sets). But the rate of increase hasn’t been perfectly constant over that entire time span.

As a matter of fact, there’s a 15-year time span during which the rate is notably different. Fifteen whole years!!! By at least one calculation, the difference is “statistically significant.”

Seasonal Niño

Most of the discussion surrounding Kosaka and Xie (2013, Nature, doi:10.1038/nature12534) has focused on how much, according to their research, might natural variability have altered the recent global warming trend. But some of their results that haven’t received much attention might turn out to be the more interesting aspects of that paper.

el Niño and the Non-Spherical Cow

Most people who follow climate science are aware that one of the natural factors which affects global temperate is the el Niño Southern Oscillation (ENSO). It’s a mode of natural variation in the tropical eastern Pacific ocean which is indicated by sea surface temperature in that region, as well as patterns of atmospheric pressure, surface winds over the ocean, even precipitation over a much larger region. As such, ENSO — whatever its cause (and it’s been happening naturally for a long time) — has far-reaching affects on weather over a large area, and a notable impact on global temperature. When the ENSO is in its “high” state (called simply “el Niño“) our climate tends to be warmer, but when it’s in its “low” state (referred to as “la Niña“) earth tends to be cooler.

Computer simulations of earth’s climate system can actually reproduce the ENSO phenomenon, showing that it might be a straightforward consequence of the circulation of wind and ocean currents. But those computer models which do recreate ENSO do so in apparently random fashion, as though it were a stochastic (i.e. random) phenomenon. This too is, as far as we know, correct — ENSO is a natural mode of variation within earth’s climate system rather than a response to some forcing agent. And, the ENSO state of the world has proven unpredictable on anything but the shortest time scales, just as we expect from a stochastic phenomenon.

Because of its impact on global temperature, ENSO can cause large fluctuations of the global temperature trend on time scales of a decade or even longer. Computer models of global climate simulate this too, but again, they do so at random times which don’t necessarily match the timing of the ENSO pattern we see in the real world. One of the consequences is that if you run multiple computer simulations of earth’s climate, then average the results, the simulated ENSO events get scattered throughout time and end up being averaged out, so that the model average ends up looking like it doesn’t have a strong ENSO impact even though the individual model runs do.

While the ENSO phenomenon has a potent impact on global temperature, it’s one of those phenomena which doesn’t create a long-term trend. It can and does cause temperature to go up and down and up and down and down and up and down and up, so that short-term (a decade or even longer) trends are profoundly affected, but on longer timescales (30 years or more, which we usually associate with “climate”) the ups and down mostly cancel each other and the long-term trend impact is minimal. But the short-term effect confounds the global temperature changes which are due to other influences, in particular to climate forcing agents like the man-made increase in greenhouse gases in the atmosphere. If we could account for the influence of ENSO on global temperature, we could isolate the influence which is due to other factors and, we hope, better understand how those other factors affect earth’s climate.

There have been many attempts to do so. One approach is to estimate global temperature as a simple function of climate forcing and ENSO through a regression approach; perhaps the best-known example is Foster & Rahmstorf (2011), which found that when the impact of natural factors (volcanic eruptions, solar variations, and ENSO) is removed, the trend in global temperature has been remarkably steady since 1979 (when satellite observations of atmospheric temperature begin). Another strategy is to use a climate model — not a climate simulation like most computer models are, but a simple mathematical model — which includes the affect of ENSO. I did so using a 2-box energy balance model here and found again that there’s nothing mysterious or inexplicable about the most recent pattern of global temperature, it is still following the path we expect — which means that it is still showing the warming influence of man-made CO₂ and other greenhouse gases.

The bottom line is that those who claim that global warming has “stopped” or even “paused” are deluding themselves. The phrase “global warming” refers to climate change, including temperature increase, which is caused by mankind, and that has continued unabated. In fact, if it weren’t for the continued warming due to human activity, natural variations (like ENSO) would have brought about a notable cooling over the last decade or so. But earth hasn’t cooled during that period, not even at the surface where we notice it most immediately, and that’s because the man-made component — global warming — has continued.

Those attempts have some severe limitations. For one thing, they’re linear models, in which the impacts of various factors (man-made greenhouse gases, ENSO, natural climate forcings) are additive, but while that is often a good approximation, the real world is nonlinear. For another, they’re based on global models which don’t capture the differences of the response of various regions, or subsystems, or their interactions (although there are some attempts to extend the scope of regression models, for example here and here). It’s reminiscent of the old joke about a physicist analyzing bovine behavior by starting with a simple model: “Assume a spherical cow.”

It might be very useful to run a computer model simulation in which the ENSO is constrained to follow its known historical behavior, so we can see how it might have affected actual history rather than a gereric “earth system.” That’s exactly what was done in a new paper by Kosaka and Xie (2013, Nature, doi:10.1038/nature12534) which investigates the impact of the tropical Pacific sea surface temperature on global temperature change. That’s the region whose variations are often referred to as the el Niño Southern Oscillation (ENSO).

The new research uses multiple runs of a coupled ocean-atmosphere computer model to simulate global temperature changes in response to climate forcing when the sea surface temperature (SST) in the el Niño region follows its historically observed values. They also estimate temperature change when those SST are not so constrained. In this way they hope to estimate the impact of actual el Niño/la Niña fluctuations on observed temperature.

They computed 10 runs each of three different scenarios. The “HIST” model runs use historical data for climate forcing, to estimate the average temperature change (and other variables) simply due to climate forcing. The “POGA-H” model also uses historical data for climate forcing, but constrains sea surface temperature in the tropical eastern Pacific (the ENSO region) to follow their historical values. This doesn’t constrain global temperature because this region covers only 8.2% of earth’s surface. Finally, the POGA-C runs constrain tropical east Pacific sea surface temperatures but do not follow historical data for climate forcing, instead holding climate forcing fixed at its 1990 values.

The HIST runs reveal that climate forcing causes reasonably steady warming since the 1970s and especially since about 1992, but that ENSO can still cause natural cooling for periods of a decade or more so that even though the man-made influence continues to cause warming, it is cancelled by ENSO cooling and results in a “hiatus” of global temperature increase:

In individual HIST realizations, hiatus events feature decadal La-Niña-like cooling in the tropical Pacific⁶ (Extended Data Fig. 2), …

In the 10 HIST model runs those kinds of events happen haphazardly, so that when they’re averaged the mean behavior shows steady recent warming:

However, When the model is run using historical data for SST in the ENSO region, the last decade’s la Niña dominance causes sufficient cooling to cancel out most of the warming due to climate forcing in the last 10 years:

If we take the difference between the POGA-H models (with ENSO constrained to follow historical data) and the HIST models, we see the estimated influence of ENSO on global temperature history:

This is, according to the new research, how ENSO has modified global temperature since 1950. The influence is clear: a pronounced recent ENSO-induced cooling which has cancelled the continued global warming due to man-made CO₂, leading to the “hiatus” in the increase of global temperature.

We can compare the POGA-H models’ average to observed temperature history (according to HadCRUT4 as the authors use) to see how well the models reproduce what has actually happened:

The agreement is outstanding, including over the last decade and more, with correlation between the models and observations of 0.93. Using just the data since 1970, when — according to the authors — sea surface temperature data is more accurate, the correlation is an impressive 0.97. This is powerful evidence that the recent slowdown in global temperature increase is not a slowdown of “global warming,” i.e. man-made climate change, it’s simply partial cancellation of global warming by the natural fluctuations due to ENSO.

Some people have not just misunderstood this new research, they seem to have bent over backwards to misunderstand. Probably the most nonsensical example comes from Judith Curry. She looked at the POGA-C results (in which ENSO was constrained by historical data but climate forcing was held fixed):

Her first mistake — quite an embarrassing one really — was to assume that this was the influence of ENSO on global temperature history. This quite misses the point, that one of the strengths of the new approach is that it allows climate forcing and ENSO to interact in a nonlinear manner. The actual estimate of the influence of ENSO, according to the new research, is shown in the graph labelled “POGA-H minus HIST.”

But her bigger mistake, which is so embarrassing that, in my opinion, she should actually admit how wrong she was and apologize, is cherry-picking the lowest value of the POGA-C average and its 2nd-highest value and calling that the influence of ENSO. Yes, folks, as hard as it may be to believe, Judith Curry actually took this as the estimated influence of ENSO on recent global temperature:

On that basis, she estimate the ENSO influence to be 0.4 degr.C, and the net global warming to be 0.68 deg.C, then declared that natural variation was responsible for more than half of recent global warming.

As the graph labelled “POGA-H minus HIST” shows, the influence of natural variation, at least that part of it from ENSO, has been cooling, not warming, and if we want to assign a percentage we should say that natural variation has been responsible for about negative 25% of global warming. Not only did Judith Curry execute one of the most blatant, most obvious, and most ludicrous examples of cherry-picking, she couldn’t even get the sign of the influence right. That’s what I’ve come to expect from her.

A Möurnful Application of Care and Skill

Not long ago WUWT reported on a paper published by Nils-Axel Mörner and Albert Parker. It’s their attempt to argue that sea level isn’t accelerating, and that sea level rise is nothing for Australia to worry about. They do so by disputing any data that disagrees with their thesis, mentioning as many confounding factors as possible (whether relevant or not), and presenting a sorry excuse for proper “analysis.”

Observe Closely

As strange as it may seem … I expected better, even from Anthony Watts’ blog.

That’s right. I said it.

There’s a post by Nils-Axel Mörner on WUWT which attempts to dispute concerns about sea level rise. In this particular case it draws attention to the Marshall Islands.

Misunderstanding Period Analysis

A post by Willis Eschenbach introduces something called the”periodicity transform” and extolls its virtues, suggesting that it’s superior in many ways to Fourier analysis. The method is very interesting, and in some situations quite useful, but it’s not the glorious improvement Willis seems to think it is — nor does Fourier analysis suffer from the flaws Willis seems to think it does. It’s also a recent method, but like most such innovations the core idea is not entirely new — there are strongly related methods that have been around for quite some time.