It is time to stop quivering in our boots in pointless fear of the future and just roll up our sleeves and build it.
- Ray Pierrehumbert

Friday, August 26, 2011

Uncertainty, Uncertainty, Uncertainty

It's peculiar that Judith Curry is criticizing the IPCC and climate community for understressing uncertainty. It's something one can imagine from the Wattses and McIntyres with their narrow focus on data, but it's incomprehensible from a member of the community.

It's hard to avoid this thought popping up as I watch the Schneider Symposium; probably over half the speakers have talked at length about uncertainty, and how to treat it in interfacing with the public and the policy sector. This includes a few high level IPCC muckety-muck types.

An interesting aspect that several have agreed upon is that the word "uncertainty" carries unfortunate semiotics; the public perceives "uncertainty" as meaning "confusion". A comment from the audience, after much talk about how to communicate the range of plausible outcomes, noted that the new draft AMS statement on climate change eschews the word "uncertainty" entirely, even though the prior statement used it many times. I was astonished to hear this greeted by general applause and enthusiasm!

I heard one speaker suggest turning it upside-down and talking about "certainty" as a quantitative measure, which makes little sense. The right word to use is "confidence", which in fact means exactly the same thing as "uncertainty" technically, despite appearing as its exact opposite in informal speech!

But the idea that this isn't something the community struggles with is nonsensical.

Turning the word upside-down to call the "uncertainty range" the "confidence interval" may not help the basic battle against agnotology. The forces of confusion keep insisting that action is counterindicated because uncertainty is broad. As I've always said, this is complete nonsense. The greater the uncertainty, the greater the need for a vigorous response. It's palpable but unstated in most of the sessions. People look at the uncertainties and are mostly concerned about cataclysms, not about false negatives. Schneider was quoted explicitly making the case that a false negative is far more dangerous than a false positive in this matter.

But it's a matter which is easy to distort. And so, those inclined to inaction stress uncertainty. This is odd enough. But then they criticize a community that is plainly obsessed with uncertainty of ignoring it!

We can use the actuarial concept of risk very nicely. Risk is cost-weighted probability. The risk spectrum in the climate matter is dominated by worst cases, not by best cases. And that is why uncertainty is not your friend. And if you advocate inaction, it is not your ally.


8 comments:

muoncounter said...

Apparently, Dr. Curry enjoys a touch of whimsy. The preliminary version of Curry and Webster 2011, entitled "Climate science and the uncertainty monster," is currently in revision for the Bulletin of the American Meterological Society. The phrase 'uncertainty monster' is given teeth here:
The “monster” is therefore the confusion and ambiguity associated with knowledge versus ignorance, objectivity versus subjectivity, facts versus values, prediction versus speculation, and science versus policy.

And all this time, I thought 'uncertainty' meant a statement of experimental accuracy and precision, something quantitative dealing with sigmas. Hardly a monster.

I'm reminded of another fictional monster:
"Beware the Jabberwock, my son!
The jaws that bite, the claws that catch!
Beware the Jub-Jub bird, and shun
The frumious Bandersnatch!"

David B. Benson said...

Many people will understand risk as it pertains to economic activity, bond yields and so on. Many people take out insurance against risks.

Here risk is taken as probability (whatever that might actually mean, ask an actuary) multiplied by cost (loss).

If this can be correctly applied to describe risks (or change of risk) due to climate change comprehension will increase, IMHO,

Grypo said...

David makes a good point. The problem is that uncertainty means undefinable risk. For those concerned about the economic risks, we need to have better answers than 'carbon tax'. This is especially the case when discussing the regressive nature of carbon taxes on poor people. This is why 1) always talk about either Earned Income Credits or Negative Income Taxes when discussing carbon taxes. 2) When people talk about GDP, always mention the Gini Coefficient. 3) And get real about alternatives. Negotiate. We'll get to all renewables, probably not within our lifetime, but if we want to there at all, we need to discuss alternatives that make sense in the present.

Andy S said...

In oil industry exploration economics a careful distinction is usually made between risk and uncertainty. Uncertainty describes the pdf of expected sizes of a discovery, invariably a log-normal distribution. Risk is the chance of failure, the dryhole case, when one or more wheels fall off your model.

Uncertainty is typically calculated by a Monte Carlo model; risk by a partly statistical, but mostly subjective, process of tummy-rubbing.

Much of the value in oil exploration is in the unusual, high-end discoveries in which everything goes right. I would expect that there's a similar model in Hollywood, where the blockbusters provide most of the profits. Similarly, the impact of the right-hand sides of the climate pdfs is disproportionately large.

Successful businessmen understand these concepts (although, admittedly, bond traders use the word "risk" rather differently). Business people who, for whatever reason, oppose climate science, tend not to feed the "uncertainty monster". They know better. They leave that confusionism to the politicians and their fellow-travelers, like Curry and Pielke Jr.

David B. Benson said...

Many people will understand One in five hundred year event such as this year's flooding of the Ohio-Mississippi-Missouri river system. Few will comprehend the underlying probability distribution function; eyes will glaze over, even if the pdf is graphed.

David B. Benson said...

Probability density function

[Was keying faster than thinking yesterday.]

Kooiti Masuda said...

Large uncertainty about the effects of greenhouse gas emission will be a reason for action rather than against it.

But, large uncertainty about the effect of other causes will detract us from action to decrease emission.

In terms of global mean temperature, we are confident that the green house gas emission will be dominant, unless there be a very big volcanic eruption or unprecedented change of solar activity.

But, it is local climate that matters to people. And it is not certain whether anthropogenic greenhouse gas emission will be a dominant factor of changes of local climate at any specific locations, though it seems certain that it will contribute somewhat.

Perhaps ENSO, NAO or AO, Pacific Decadal Oscillation and Atlantic Multidecadal Oscillation will be more important than enhanced greenhouse effect in adaptation to local climate by local people, though these oscillation will somewhat be modulated by global warming.

So, it is difficult to use uncertainty to sell mitigation of enhanced greenhouse effect.

David L. said...

"The greater the uncertainty, the greater the need for a vigorous response."
By that logic, we should put all our wealth into protecting ourselves from a deadly gamma ray burst from a nearby star rather than "controlling climate".

Such inverted logic is robbing the poor of funds for the most urgent global humanitarian projects, as detailed by the Copenhagen Consensus 2008. Providing micronutrients heads the list - globa warming mitigation comes in dead last on benefit/cost.

Similarly, to Fix the Climate research leads the way, while cap and trade or carbon tax come in last.

Let us restore rational scientific foundations with a priority for the poor, the widows and orphans.