Friday, February 20, 2015

Black Swans, Black Cats


A couple of days ago, when describing the incoherence of the Luckwarmers, Eli cribbed from the Idiot Tracker  (v. 2010) about those wide eyed optimists
So the critical question then becomes: what is the lukewarmers' range? Consensus scientists estimate climate sensitivity at about 3C, but concede that it might be 1.5C, 4.5C or even higher (and very unlikely to be much lower). What range do lukewarmers think is plausible? 
So far, to my knowledge, no self-identified lukewarmer has been persuaded to answer this question. They will find it difficult. Because they have positioned themselves as participants in the scientific debate, they can hardly claim 100% confidence in X climate sensitivity, no error bars. If they are reasonable, they have to accept they they are as fallible as the rest of the scientific community, and although they think the climate sensitivity is 1.5C (say) it might be 1.0C, or 2.0C, or even (gasp!) 3.0C (where the consensus puts it).
Time has moved on Tom Curtis pointed out in the comments
Being fair to the lukewarmers, some including Nic Lewis and Steve Mosher have indicated uncertainty ranges (sort of). The first problem is that Lewis' uncertainty looks an awful lot like dogmatism given how tight his uncertainty interval is relative to that of the IPCC. The second problem is that Mosher's range looks more like a con job. 
Specifically, Mosher has said:  "ECS is not less than 1.2C ( or basically a no feedback value however you want to calculate it And the probablity of it being less than 3C ((Hmm I’ve prolly said 3.2 in a couple places) is greater than 50%." 
The IPCC does not give a PDF, but from their statements it is clear that any distribution maximally coherent with their claims must have a strong right skew. Assuming that their likely range is centrally placed, ie, that it is as unlikely to be below the range as it is above it, that means the modal value is around 2.5 to 2.8 C and the median is likely to be very close to, and possibly below 3 C. So, in an attempt to give a range for a purportedly distinct position, Mosher appears to simply redefine the IPCC as "lukewarmers".
Today, the twits guided Eli to a piece on skepticism by Nassim Taleb, keeper of the black swans. Black swans keepers are, of course, the opposite of Luckwarmers, who, in their way are the Émile Coués of the science blog world

Taleb looks at the effect of ignorance on estimates of probability
The introduction in general in any field with potential iatrogenics of any new element without available track record (hence model uncertainty) fattens the left tail. 
Some straight applications 
• Skepticism about climate models should lead to more precautionary policies in the presence of ruin. It is incoherent to doubt the mean while reducing the variance. 
Which pretty much kills Nic Lewis's squeezed prior and
• "Mitigating" policies aiming at reducing risks –say geoengineering– in fact are likely to increase such risk. 
• Conservatism is a dominant strategy in the tails.               
Taleb elaborates
In thin-tailed domains, an increase in uncertainty changes the probability of ruin by several orders of magnitude, but the effect remains small: from say 10−40 to 10−30 is not quite worrisome. In fat-tailed domains, the effect is sizeable as we start with a substantially higher probability of ruin (which is typically underestimated).
He also makes an interesting point about such issues as GMOs
For standard statistical theory doesn’t allow "acceptance", it only allows "failure to reject". Even when someone in prose says "accept that" he mathematically means "failed to reject at some significance level...", i.e., baring a tail event. Similarly, when someone is indicted, he is treated as innocent unless proven otherwise. This principle is adopted by scientific journals (remember from section 1.3 that statisticians are the "evidence" police and their evidence is "up to" a tail event that is not specified in impact). This is a very big thing and it is ironic. 
For the majority of biologists involved in the GMO debates against the precautionary principle don’t appear to be aware of the central fact that
evidence = ”barring a tail event”
and argue they have "evidence there can’t be a tail event".

7 comments:

THE CLIMATE WARS said...

My limited success hunting swan suggests that much as fat cats want thin tails appended to projections of fossil fuel forcing , the kitty in question has been running away from exponential projections of its behavior since first spotted by Hansen in 1989.

Unless it shows its spots soon, the mulliganns granted past models for the last four decades running may come due-- when the best fit remains a straight line after five decades and , and the sensitivity remains uncertain after ten, some reflection many be in order.

No use wasting bullets on snipe.

Fernando Leanme said...

I think you have a very fundamental flaw in your logic. To simplify the problem so you can see it, consider yourself driving on a single lane road at night.

You arrive at a fork. On one side you discern a path which takes you to what you consider a (more or less) bumpy landing. The other path, on the other hand, looks really dangerous. Your knowledge, training, political biases and what you think you know make you choose what you think is the better option.

Others cone to the same fork, and live in a mirror image of your world. They thnk your choice is much worse. They see risks and impossibilities where you see nothing.

The difference, you see lies way beyond the simple climate sensitivity issue. Once you pick a temperature range you have to develop a mental model of what that range implies. And this is where you tend to fall apart.

I have come to distrust everything because I see too much bullshit, and enormous arrogance comng from the fictitious 97 .%. The recent article in Scientific American, "Fantasy Islands" addressed the bulshit problem I brought up in my "Drowning Islands" post . But the arrogance factor hasn't been addressed. You do need to tone down and recognize you are hairless chimps.

Bryson said...

Fernando, I don't think risk management works the way that you think it works. Differing perceptions certainly lead to diffferent preferences. But that doesn't make the situaiton epistemically or morally symmetrical. We all live in the same world-- and 'mental models' of what some changes in temperature might mean are a long way from constituting a serious evaluation of risk.

Anonymous said...

Fernando: "I have come to distrust everything because I see too much bullshit, and enormous arrogance comng from the fictitious 97 .%. The recent article in Scientific American, "Fantasy Islands" addressed the bulshit problem I brought up in my "Drowning Islands" post. But the arrogance factor hasn't been addressed. You do need to tone down and recognize you are hairless chimps."

Clearly, you haven't calibrated your bullshit- and arrogance-o-meters in a long while. Perhaps this will help with your adjustments: http://www.nytimes.com/2015/02/22/us/ties-to-corporate-cash-for-climate-change-researcher-Wei-Hock-Soon.html

"Charles R. Alcock, director of the Harvard-Smithsonian Center, acknowledged on Friday that Dr. Soon had violated the disclosure standards of some journals.

“'I think that’s inappropriate behavior,' Dr. Alcock said. 'This frankly becomes a personnel matter, which we have to handle with Dr. Soon internally.'”

"...Many experts in the field say that Dr. Soon uses out-of-date data, publishes spurious correlations between solar output and climate indicators, and does not take account of the evidence implicating emissions from human behavior in climate change."

Bullshit much?

"...In a Senate debate last month, Mr. Inhofe pointed to a poster with photos of scientists questioning the climate-change consensus, including Dr. Soon. 'These are scientists that cannot be challenged,' the senator said." [emphasis added]

Arrogance much?

Fernando: "Once you pick a temperature range you have to develop a mental model of what that range implies. And this is where you tend to fall apart."

Your mental model was developed while wearing blinders when it comes to the negative costs of human-caused climate warming, while you only see the costs of attempts to avert such impacts. And this is where you tend to fall apart.

Tom said...

Logic and observations both lead to a preliminary 'prior', that feedbacks to the Earth's climate system are by nature negative, hence sensitivity to perturbations should be low. So far, every Lukewarmer I've read has a different estimate for both TCS and ECS.

Most lukewarmers stay away from WG2 and WG3 issues, other than offering fairly bland support for no regrets measures. Well, I label it bland because I think they are far more important.

Because of very real issues in how WG2 and WG3 matters have been presented, it is hard to find a starting point for debate.

However, I have done (amateur) projections indicating human energy consumption in 2075 will be six times greater than in 2010 (http://3000quads.com/2015/02/21/our-global-energy-future/) and that global investment in non-emissive energy infrastructure is only sufficient to maintain the status quo in terms of percentages of non-emissive energy sources.

Despite my dislike for the strategy and tactics of some in the consensus community, I am indeed 'alarmed' by this.

Even with a low sensitivity the effect of burning coal to produce the increased energy we will need to power the world will produce unwanted effects, both in terms of conventional pollution and climatic response.

THE CLIMATE WARS said...

There are few lessons in scientific humility more sobering than the failure of the four dozen odd peer reviewed estimates of climate sensitivity published since 1896 to converge on a single value.

Anonymous said...

Russell, I thought you were going to say the Hubble "Constant." Your point is really a red herring, and doesn't bear the constant repetition.