Saturday, April 05, 2014

Climate Nuts vs the IPCC

In one recent climate thread on Google+, I was informed, by three different people, that global warming would:

Create two billion climate refugees

Flood most of the world's large cities

Destroy civilization.

The simplest rebuttal to such claims is the latest IPCC report. The IPCC, after all, is doing its best to persuade people to support action to slow global warming, so unlikely to minimize the problem. But if you look not at the rhetoric but at the factual claims, the impending climate catastrophe looks like a wet firecracker. For instance:
There is no evidence that surface water and groundwater drought frequency has changed over the last few decades, although impacts of drought have increased mostly due to increased water demand.

Economic losses due to extreme weather events have increased globally, mostly due to increase in wealth and exposure, with a possible influence of climate change (low confidence in attribution to climate change).

Some low-lying developing countries and small island states are expected to face very high impacts that, in some cases, could have associated damage and adaptation costs of several percentage points of GDP.

... most recent observed terrestrial-species extinctions have not been attributed to recent climate change, despite some speculative efforts (high confidence).

With these recognized limitations, the incomplete estimates of global annual economic losses for additional temperature increases of ~2°C are between 0.2 and 2.0% of income ... .

Oddly enough, precisely the same evidence can be used against nuts on the other end of the spectrum, such as the gentleman in one thread who claimed that the IPCC had a budget in excess of twelve billion dollars. Pointing out to him that the actual budget, available on the web, came to about eleven million dollars had no effect. If, as he confidently believes, the IPCC is a massive fraud, inventing global warming out of thin air for its own sinister purposes, why should I be so naive as to believe their account of their budget? Pointing to webbed data on global temperature, which has increased by about  one degree C since 1910, would be no more effective, since he can, and probably would, claim that the data are fake. 

What he cannot explain away so easily is the modesty of the IPCC's claims. If, as he (correctly) believes, they are trying to scare people into doing something about global warming and if, as he (incorrectly) believes, they are unconstrained by actual evidence and science, why don't they tell a better story? Two degrees of warming over the next 86 years, a foot or two of sea level rise, are not very impressive threats, despite all the rhetorical efforts of the IPCC and its supporters. Why don't they make it ten degrees and ten feet? Twenty feet? A hundred feet? Why don't they tell something closer to the story that their nuttier supporters want to believe?


Eric said...

I think the answer to your question is that the rats are, if not yet actually leaving the sinking ship, at least eying the hawser cables.

Anyone in the IPCC with a shred of scientific integrity left has to know what it means that GAT trend has fallen outside the 95% confidence range for their entire sheaf of models.

I think we'll see two contradictory responses to this over the next 12-18 months. On the one hand, heroic efforts to add epicycles to the models that explain away the absence of warming. On the other, subtle positioning for a future attempt to deny that they ever made catatrophic oredictions at all.

The backpeddling in AR5 is the second kind if response.

Handle said...

One thing I would like to see is a particular test for retrodictive overdetermination of the results General Circulation Models.

I want them to run the best, most sophisticated GCM, using the latest scientific understandings, at the highest level of granularity the world's best supercomputer will allow, but with all the inputs relying on measurements at least a decade old.

So, for example, if certain regressions are run on temperature data, the data set should stop at 2004. If the climate sensitivity estimator is required as an input, it should rely on a method of estimation that doesn't smuggle in any data of more recent vintage.

And the question would be, when we run this updated GCM(2004), how good a job does it do in terms of predicting the actually observed average trends over the past decade.

If the updated GCM(2004) meets this 'validation' test, then it would be interesting to see where it differs from the old GCM(2004) that were run in 2004, and which proved to be so biased to the upside.

If, one the other hand, the GCM(2004) does not meet the test, then anything that relies upon the results of that GCM's simulations, such as the IPCC report and countless scientific papers, should be rendered highly suspect and presumptively invalid.

Perhaps someone out there knows how they have updated the old GCM(2004) to GCM(2014) in way distinct from those related to collecting more recent data and adjusting regression coefficients, which, as I indicated above, poses a hazard of overdetermination.

That is to say, obviously our models and understanding of the global climate was incorrect in 2004. Have we updated that model in any fundamental way about the way energy is circulated besides collecting more data that makes the historical fit look better? Have we tested for forecasting validation in this way?

Tibor said...

Again, I think a webcomic is relevant. this time however, it is not xkcd, but SMBC:

Most of those who argue about things online belong to the "nutty" part of the rectangle/circle. And they are also often the most vocal ones. Since they don't care about arguments, it is best to ignore them once one finds out who they are.

jdgalt said...

Why could it possibly matter that the alarmists aren't exaggerating the number of degrees of potential warming, if they're exaggerating the likely consequences (and thus the alleged need for government to DO SOMETHING!!! which is always the real point of these phony emergencies)?

jimbino said...

Most of those problems could be solved by putting contraceptives in the world's water.

It's not that the gummint needs to do something, but that it needs to stop doing something, namely, encouraging folks to breed by public support of education and tax-favors. It's crazy to be pouring billions into increasing demand for water and other scarce resources.

Trouble is, most commenters, including David, have responded to gummint incentives at are part of the problem, not the solution.

Old Odd Jobs said...

Tee hee hee! He says "gubmint" because only inbred hicks are suspicious of government. Ah, what a smug fool.

Will McLean said...

As far as I can tell, 2 degrees of warming is not what IPCC projects based on business as usual. That's 4.5 degrees. Also, they say there is some reason to fear that 4 degrees may be worse than twice as bad as 2 degrees.

Shaddox said...

> "What he cannot explain away so easily is the modesty of the IPCC's claims. If, as he (correctly) believes, they are trying to scare people into doing something about global warming and if, as he (incorrectly) believes, they are unconstrained by actual evidence and science, why don't they tell a better story?"

Perhaps the IPCC knows that deliberately minimizing the threat of global warming will lead people like David Friedman to place more trusts in their raw data. Perhaps they're attempting to optimize between convincing some people there's a real threat (by stating the threat as the conclusion of their raw data) and convincing others that their raw data is reliable (by not overstating the conclusion of their raw data).

Of course, I'm kidding, but you really can't argue effectively against people with views like that.

Brad Williams said...

Handle, I noticed you didn't specify what it would mean for the GCM(2004) to do a "good job." This is something I wonder about the models. Since there are an infinite number of functions which fit any given data set perfectly, finding one doesn't, in itself, tell us that it correctly models reality and will be reliable in the future. I think the only way to judge that a program models reality well would be to already know that its code matches the patterns of the physical system. One could never prove that a model is good by running it ...though you could prove that it is bad.

David Friedman said...

"I think the only way to judge that a program models reality well would be to already know that its code matches the patterns of the physical system. "

Suppose a program does a good job of predicting data not available when it was constructed. We have two possible explanations:

1. It succeeded by pure chance.

2. It succeeded because it models reality well.

Suppose you start with prior probabilities for both explanations. After each successful prediction, you should, by Bayes Theorem, raise the relative probability of explanation 2, since the conditional probability of a successful prediction given 2 is much higher than it is given 1.

Bruce said...

What he cannot explain away so easily is the modesty of the IPCC's claims

The problems with arguing with nuts is that there is usually no end to the nuttiness. Each oddity of the nut's current theory -- contradictory facts, implausible theory, inconsistent premises -- can be explained away by an extra ad hoc hypothesis. If you have as many independent variable as data points, the model fits perfectly.

This approach generates rich and clever conspiracy theories, and it often makes for fun entertainment. But apart from its entertainment value, the exercise is largely wasted. You won't convince the nut, and no one else was bothered by his initial claims to start with.

Jim Rose said...

exaggeration is common in politics, so voters adjust for it.

politicians adjust for this voter adjustment and the circle spirals to the bottom

Brad Williams said...

David, to apply Bayes' rule, we'd need to be able to estimate the probability that the hypothesis of high climate sensitivity is true, plus the probability that a future temperature will be above Y if the hypothesis is true. Can we estimate either? I doubt it. For one thing, arbitrary postulates, aka wild speculation, cannot have a probability assigned. But to get anywhere with Bayes' rule, we'd have to assume that the hypothesis is not arbitrary, that it is grounded in some positive evidence, and we can at least say something like "we have 10% confidence that climate sensitivity is greater than 2C." I take it that some people very familiar with the state of climate science, such as Lindzen and Curry, are doubting that the high-sensitivity hypothesis is anything more than baseless speculation.

Here is a thought experiment about the value of arbitrary models: Let supercomputers randomly generate a trillion models tomorrow that fit all known data perfectly. Wait ten years, then pick the best performer. Question: Does that model tell us about reality? Can we peer inside it and find scientific knowledge?

David Friedman said...


As it happens I have an old blog post which I think shows the difference between your thought experiment and the real world equivalent, which uses humans rather than a supercomputer.

Brad Williams said...

If I understand your other post: everything else being equal, those who have done good work tend to keep doing the same kind of work well. That is certainly true, and it's why we justifiably trust specialists, from barbers to doctors. But that all goes out the window when everything else is not equal, e.g., when the specialists disagree, or there is reason to suspect their motives. Then every inch of "probability" for their models and theories must be earned.

Jon said...

I would suggest that IPCC understates the severity of the problem for very good reasons. Take a look at what Michael Mann went through for generating the "hockey stick" graph. It was ultimately used in an IPCC report and shown without error bars. The fossil fuel industry came out in force, using the Senators that they own to bludgeon Mann. In fact Mann's hockey stick has been verified in many different ways, including recently when a Koch funded group that sought to rebut Mann in fact ended up confirming his conclusions. Mann went through the ringer, and though the fossil fuel industry is wrong in service to profits of course, other scientists are simply gun shy. Petroleum lost the battle, but won the war. They prevent scientists from saying anything but what they can confidently assert for fear of Inhoffe and all the others that simply do what their corporate masters require. Science itself is a probabilistic enterprise. Scientists are just not the kind of people comfortable with drawing confident conclusions, as conservatives tend to do. This feeds the propensity for understatement of the problem, and makes the alarm expressed in the recent IPCC report all the more alarming.

In fact in a number of key areas IPCC estimates have been shown to understate the problem. Take the polar ice sheets. The IPCC assessments predicted that we could see summers without ice by 2100. We desperately need that ice because it reflects solar radiation, unlike water which absorbs it. No ice and we are facing run away warming. Look at the current projections. By 2020, or possibly even prior to that, we could have summers with no ice. This is a catastrophe.

See here for other areas were IPCC has understated the problem.

We are in dire straights here people, and the continued downplaying of the issue here and at other right wing places is not helping.

Mac Muir said...

As I understand it Michael Mann's "Hockey Stick" graph is based on tree rings of six carefully selected pine trees.The raw data of which he has conveniently "Lost." In his graph both the Medieval warm period and the little ice age, both well documented events in both Europe and China, have disappeared. Thus it might be appropriate to refer to Mann and his followers as "Climate History deniers." It is likely that the extra tree ring growth in Mann's work is more a result of the fertilizing effect of the extra CO2 rather than an extra temperature rise.

As I explained in a comment to a previous post: The Earth may be warming (After all Climate is always changing.) but any rise now is NOT being caused by extra CO2.

Mac Muir said...

A Further note: It is much more likely that graph 7.1 (page 202) of the IPCC report AR1 is valid than Michael Mann's hockey stick.

Jon said...

Mac take a look at the wiki entry on Mann's hockey stick. It's a good summary and shows that it has been verified by multiple independent lines of evidence.