Thermageddon and the Golden Horseshoe

It appears that Greenpeace activist Robert Hunter (not the lyricist, I presume) coined the term “thermageddon” in earnest as the title of a book about global warming issued in 2003. It appears the word “thermageddon” has long seen use as a term of mockery among those who think that the idea of global warming is an amusing little joke (as well as a terrifying conspiracy — I’ve heard that paranoia is often inconsistent like that).

I learned these facts about “thermageddon” as I sat down to compose this posting. I write to respond (at considerable length) to this naysayer tweet, which at first glance appears unremarkable:

Gavin Schmidt recalls a snippet describing a bar called the Golden Horseshoe, from a novel of that name by Dashiell Hammett:

I was reading a sign high on the wall behind the bar:
‘Only genuine pre-war British and American whiskeys served here’
I was trying to count how many lies could be found in those nine words, and had reached four, with promise of more …”

I was struck by a similar realization in regards to this tweet.

Lies or Mistakes?

I don’t like to fling accusations if I can help it. I think very few of the naysayer crowd are active deniers, in the sense that “deniers” are willing to say things whether or not they believe them.

But as you’ll see, there’s an astonishing amount of wrongness bound up in what they think and say, so much that I see eight errors in this one tweet. And what’s interesting is that all the errors lean the same way — the way that would indicate that mt (myself) is off his nut as is anyone who suggests that climate change is a big deal.

I don’t think @AZComendador is stupid or crazy or dishonest or corrupt. I’m just suggesting that he/she sees what he/she wants to see and disregards the rest.

I think everyone (with the possible exception of myself) is biased about some things. I think we should forgive each other and try to leave people room to retreat.

So I won’t call these things “lies”, but “mistakes” is a bit too generous — if they were unbiased mistakes it’s hard to imagine them all leaning the same way. I will call them “misrepresentations”.

I think a fair-minded person ought to have done better, but fair-mindedness is hard to come by in these strange and distracted times. That said, I think the confusion is genuine, and I hope against hope to leave room for @AZComendador to reconsider.

The Precursor Misrepresentation

First of all, let’s look at the context. I once claimed that

“an important factor in our ongoing flirtation with an astonishingly severe global disaster is that non-serious arguments, purportedly about science, have been used as a proxy for a serious argument about how to manage the collective commons.”

@AZComendador, an unknown-to-me Twitter account, piped up criticizng me because I had not provided evidence that one should worry about “an astonishingly severe global disaster”.

There is a misdirection even in the precursor question.

It suggests that in the CO2 On Trial article I was arguing that there is reason to believe that a global disaster is possible. I was not.

I was writing for an audience willing to believe or even willing to concede for purposes of argument that such a disaster is possible. It was a premise, background for my argument.

I presumed that readers already had some idea that such argument existed. So I replied “I kind of thought referring to IPCC reports in the opening paragraph might cover that.”

Perhaps the most exasperating thing about the naysayers is their refusal to countenance any conversation which excludes their position, as if people who do not believe them have no right to converse. No.

The Misrepresentations:

Here’s that Tweet from AZC again: “AR5 found virtually nothing bad could be blamed on the 1/2 Thermaggedon that has already happened.”

Here are the eight mis-statements, all biased toward ignoring climate change, embedded in that brief tweet:

1) 2 degrees C is claimed to be “Thermageddon”

No, the international community has agreed that 2 C is the threshhold of risk. It’s not as if we thought 2.01 degrees would be catastrophic. If we had, 2.00 degrees would be an idiotic target, wouldn’t it?

If anyone is claiming that 2 C is thermageddic, it certainly isn’t any of the mainstream climate-focused experts, or else they’d be grossly negligent in not proposing much tighter bounds.

2) Impacts of a Given Warming are implied Instantaneous

No, many of the consequences of a given warming take some time to unfold. For instance, if emissions cease, as we understand it, temperatures will soon stop rising (there is some lag) and then the ice sheets will take some time to re-equilibrate and the deep oceans will also continue to expand. So we have already bought plenty of sea level rise. Similarly, species take time to go extinct, and ecosystems take time to collapse, as they hang on to the edge hoping for a cooler, more normal year.

Although by some measures we have just crossed the threshhold of a 1 C warming, nothing says we have already seen all the impacts of a 1 C warming.

3) Damage is Implied to be Linear with Temperature Change

The idea seems to be that since we now see the damage from 1 C (which we don’t, see point 2 above, but suppose we did) we can double it to estimate the damage due to 2 C.

No. Of course, this is wrong. Damage increases faster than proportionally. This is obvious by taking it to absurd limits.

This isn’t realistic, understand, but for purposes of argument, if temperature rises 30 C probably most wildlife dies. At 85 C, the oceans start to boil off and we all cook.

That is not what we expect, fortunately! (That would be a real “thermageddon”.) Bit consider the scenario. The damage would surely be more than 85 times the damage we see at 1 C warming. So the damage curve has to swing up faster than the temperature eventually.

Does it do so between 1 C and 2 C? There are a lot of reasons to think so, and (though I find this graph very complacent for rather complicated reasons) in fact even the economists are on board with that:

4) IPCC (AR5) Report is implied Contemporaneous with 1 C global warming


There is an argument that the world in 2015 will end up 1 degree C warmer than the preindustrial baseline temperature (and probably warmer than at any time since the Eemian, but that’s a bit off the thread here). But the argument in the tweet refers to observational evidence in the 5th IPCC report. The physical science part of that report was issued in 2013, back when global temperature was climbing so slowly that many people were calling it a “hiatus”. But the past two years have showed sharply increased temperature. So the report can’t possibly say anything about observed impacts of 1 C. They hadn’t happened yet.

In fact, to be considered for inclusion in the physical sciences section of the IPCC AR5 report, papers had to be submitted for publication by the end of July 2012 (and published by March of 2103).

And then there’s publication delays. Since these are observational papers, the pipeline includes collecting and collating observations, along with analysis, literature search, writeup, and (lay people can’t imagine how antediluvian this step is) unbelievably tedious formatting. So nothing in the observations feeding the AR5 attribution chapter could be collected much later than 2010.

5) Statistical Analyses are presumed Instantaneous

So based on the last section, at least we’re looking back no further than 2010? Nope.


In extracting trends from observations, you needs enough observations to do statistics. Typically, climate statistics are collected over a thirty year period. For rapid changes you might push this to look at a decadal time scale.

The more rare an event, the fewer data points there are, and the harder it is to track statistical changes.

(As it happens, rare events on earth have very peculiar statistics, too, as they seem to come bunched up for some reason. Americans will remember the intense hurricane seasons of 2004 and 2005 as a striking example. This makes the attribution problem even harder.)

So efforts at observational attribution have to look back a decade or so to have much chance to extract any statistics. If we’re lucky, we’d be looking at the decade from 2000–2010, centered on 2005. So we’re really looking back a whole decade for the attribution chapter.

6) CO2 is presumed to be on Trial

The tweet is formulated as pointing to a lack of proof. I have been arguing that this is very much like a court of law wherein CO2 emissions have perceived rights. There’s an implication that climate science is in the business of proving that bad things arise from human interference in climate.

No. Only one of 14 chapters of the WG I report resembles this pattern of argument. Most of climate science is not about determining the guilt or innocence of CO2.

The fact that weather prediction is possible a week in advance (when a century ago it was considered laughable to predict more than a day) demonstrates that the physics of the atmosphere is well understood, at least on short time scales. That understanding is enough to indicate that CO2 accumulation has a strong effect on the climate system.

We do not collect data to confirm things we already know with certainty.

As I argue, the questions we should be asking ourselves are at about what level our monkeying with this effect is risky. And so far, the vast predominance of evidence is that we are near this point. This is reflected in chapters 1, 2, 3, 4, 5, 6, 7, 8, 9, 11, 12, 13, and 14 of the most recent Physical Science Report (Working Group I) of IPCC .

My claim was that IPCC offers grounds for concern, not that severe damages had already been detected.

The tweet claims that I was wrong to suggest we are “flirting with disaster”. But it tries to shoot that down by pointing to places which are formulated as proof, not as plausibility arguments. That’s not responsive to my claim that we are taking a collective risk

7) Absence of Proof is conflated with Proof of Absence

No. There’s a difference between “not proved” and “not true”.

It’s true that the confidence that certain effects have already been detected is low. But the tweet seems to imply that means the effects don’t exist.

Does “virtually nothing bad can be blamed” just mean that blame has not been assigned? Does it mean “nothing bad can be proven”? Or does it mean “It’s proven that there is nothing bad”? If it’s the latter, it’s hardly grounds for objecting to my “flirting with disaster”, so it must be the former.

What’s more the playing field is tilted in formal attribution studies.

Naysayers are fond of focusing on chapter 10, which focuses on the “attribution” question. @AZComendador is doing exactly that. That is because that chapter is structured as being about proof.

There’s a long history of what I think is excessive attention to the “attribution” question. The jargony version of my opinion is that it applies frequentist thinking to what reasonably should be a Bayesian problem. The less jargony version is that the “attribution” questions are set up like clinical trials, where everything is presumed false until proven “true”. The peculiar “frequentist” definition of true (advocated explicitly by Michael Crichton) that is used in attribution studies means “rejecting the null hypothesis with 95% probability” which means, “if nothing were happening we would be very unlikely to see this due to random variation”, which in turn involves all sorts of assumptions about what random variation looks like.

That is a good way to set up drug trials. If you don’t have enough data to prove efficacy of a drug, you go out and get more subjects. You can buy data points. In earth science, we get one year per year, and no amount of horses or men can make the data collection go faster. So attribution problems are hard.

In most sciences, failed attribution is not noteworthy. Failing to find a signal in the noise is considered non-publishable. It’s very different from finding that there is no signal in the noise.

Suppose I look at severe rainfall events in Texas. Suppose I can see an uptick in recent years. Is it a random, natural variation? Or is it climate change? The purely statistical record may be ambiguous in this regard. We could be seeing an uptick that’s attributable to human-driven climate changes, but we don’t have enough data for the statistics to reach 0.95 rejection of the null. That does NOT mean it isn’t happening.

Suppose there’s a crime, and nobody is convicted. Does that mean there is no responsible party? Statistics may not be enough to tell us whether a trend is attributable to human-caused changes. Sometimes guilty suspects are released because “burden of proof” arguments are hard. That doesn’t mean they didn’t do it.

And One more

Finally, just as I write, I discover that “thermageddon” was not a recent coinage, but a risible title of an overwrought book some dozen years ago. Is this another hidden lie? Well, yes it is — it’s a dogwhistle for the denialist troops that I am an extremist alarmist.

I have discovered that there is this idea that there is an “extremist” “alarmist” “warmist” “campaign”, somehow organized, such that anything said by anyone about anything about climate is something that must be defended before climate change is to be taken seriously.

Anyone that takes climate change seriously is responsible for anything anybody ever said who takes climate change seriously.


Come on, be serious.

Of course, I think the coinage of “thermageddon” is laughable, and I doubt that the book is worthy of very much attention, but the mockery implies otherwise. I have been running into something like this in my arguments with @GaiusPublius01 who keeps speaking of a climate change “campaign” which is apparently poorly executed in some way, and to which I seem to be a party, so anything somebody else said is my fault if he doesn’t like it.

It’s incredibly tedious.

That makes eight.


The Golden Horseshoe 2015

(or mt’s version anyway)

Eight bits of misleading nonsense all biased to make global warming look trivial. Not bad for one tweet.

Also a fine example of the Law of BS:

In this case, a gross underestimate. In this case it took 2500 words to properly fisk a single tweet.

I think if I still have the Golden Horseshoe about I will award this year’s edition to @AZComendador, adding this remarkable tweet to previous years’ winners David Rose and Guy McPherson.


PS — Who is the genius who coined the law of BS, probably pictured above?


Update via William Connolley: Gavin Schmidt attributes the rediscovery of the Golden Horseshoe story to a blog called Some Are Boojums which also gave such an award to a climate science distortion.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.