However, figure 1 in that link indicates that the 95% confidence interval includes 0.0;

Not by 2011 it doesn’t. Try again.

To me, this suggests that observations are well below predictions.

So firstly you’ve backed off from “much, much greater” to “well [above]”, and you’re indicating your opinions are driven by personal suggestion rather than analysis. We’re making some progress.

Next, in a noisy system one needs to compares statistically significant trends so as not to be led astray by the noise. You are not doing so when you argue:

…each of the separate models used by the IPCC to construct their ensemble had predicted that T would rise over the interval 2000-present.

and you go on to (apparently ignore any confidence intervals) and compound the error by comparing either endpoints or statistically non-significant trends over an interval such as 2000-2011:

Each of the resulting 56 projections indicated that T would increase, rather than remain unchanged as has been observed.

and again when you compare endpoints rather than trends over a longer interval:

that is, no change over the average T in the interval 1979-2011!

If you want to do an honest assessment, then ask yourself what were the predicted trends (with confidence intervals) from the IPCC models – and what has the observed warming trend been over a period long enough to demonstrate statistically significance? Once you do that is one “much, much greater” the other? (Hint: the RealClimate article provides the numbers.)

Thirdly, one needs to compare apples to apples. The real world has confounding natural variation effects that are not incorporated into the models such as volcanic eruptions, ENSO and AMO. What does the temperature record look like when you remove most of their influence? Would you claim that model trends are “much, much greater” than the trend in this underlying signal?

Fourthly you cite Hansen’s 1988 model as some kind of support. It’s true that his predictions were high – the model assumed forcings that were too high, and the model itself is much simpler than today’s models and its climate sensitivity is believed to be too high because it didn’t capture important interactions.

Strangely enough, though, that’s just one model – which scientists have acknowledged was too high. You claimed all of them were “way, way higher” – a claim that you have not demonstrated (apparently because you are cherry-picking noise instead of analysing the underlying signal).

And then you compound your error by asserting that a statistically significant trend is actually random:

Similarly, one should expect that over a 20 year interval (1979-1998) one might observe a random warming period.

(Why do you think scientists test for statistical significance in the first place? What is the significance level over that 20 year period, and how exactly does it support your contention that it is likely “random”? If the significance level is not high enough for you, what about a 30 year period? And why do you ignore the influence of other forcings that provide a very good explanation for periods such as 1940-1970 – do you really think all the scientists have done is look at some temperature curves and go “yeah, that must be due to CO2”?)

…the model projections diverge from the observed T changes under even the best model.

Yep, you naively assume that projections are expected to model the “noise” of observations and then claim that means they are useless – entirely predictable fallacious claim. (What were the red dashed lines on that figure for, and why would they be relevant to your claims? Why is it said that all models are wrong – but some models are useful? How would one honestly determine whether a model is useful or not?)

…the only indication that humans are causing significant global warming via CO2 emissions is based on climate models.

“Skeptics” keep saying that, and they keep being wrong.

If Gleich had merely revealed that he had received the “strategy” memo, without vouching for its authenticity, he couldn’t reasonably be criticized.

How charmingly naive.

“…but in two years, or five, or maybe another ten – depending on how those ‘random’ confounders behave – the inescapable truth that the consensus was right all along will come knocking at your door.”

Bernard, we may be near a point of agreement. You think that the global T is about to resume its climb and begin to match the models. I suspect the next decade or two will trend flat or down as the quiet sun exerts its effects. We might both agree that the observed T over the next decade should either validate or disprove the current climate models.

I will change my opinion if the data change. I hope you will do the same.

A proper understanding of the “precautionary principle” does not suggest that we ought to impoverish humanity by abrogating the energy sources which have lifted us out of preindustrial poverty, unless we are quite certain that the alternative is worse. The failure of the climate models suggests that we must instead wait for credible evidence.

You need to read someone such as Schumaker, or if you can bear a walk on the wild side, Greer.

Your logic basically distills to the notion that we need to spend the capital, sell the farm, so that everyone can eat caviar.

But now we’re wandering from the subject at hand and toward the reasons why you don’t like the implications of the (hard) science in the first place. Let’s stay on topic.

The occurrence of processes known to be significant in climate cycles such as the Pacific Decadal Oscillation (PDO), Atlantic Multidecadal Oscillation (AMO), and el nino-southern oscillation (ENSO) cannot be explained or predicted by the current climate models.

Yet again you are confabulating fundamentals with confounders.

Consider genetics, where the basics of codons, genes, and such are understood, but where many of the finer epigenetic phenomena are not. Or ecological processes, where trophic relationships and growth asymptotes are understood, but many of the higher and more complex interactions amongst system components are not. Or even general physics, where gravity, electromagnetism and thermodynamics are understood at a broad level, but where the minutæ are far from being elucidated.

In each of these disciplines the fundamentals permit us to well describe the essential processes of the respective fields, even as there is much refinement left to achieve. Similarly, the fundamental principles underpinning climatology are solid, and well describe its essential processes, including the action of ‘greenhouse’ gases and how human carbon dioxide emissions will warm the planet.

This truth might be unpalatable to you, and you might like to deny it, but doing so won’t change the science, nor will it alter the future trajectory of the planet’s climate. You can go on harping about uncertainty and how the experts don’t understand their own fields, but your claims are mere smoke in the wind.

The upshot is that climatology is as ‘mature’ as those other disciplines. Your protestations to the contrary are a sign of a lack of understanding of the science, and/or an attempt to sway from the truth people who do not have a technical understanding of the subject.

You might manage in the short term to retain an illusion of knowing better than the professionals in science, but in two years, or five, or maybe another ten – depending on how those ‘random’ confounders behave – the inescapable truth that the consensus was right all along will come knocking on your door.

I hope that when that time comes you have the good manners to answer with grace that knock. Unfortunately by the time you acknowledge that climatology was ‘mature’ all along, the global climate itself probably won’t have the slightest amount of time for your apology.

“When a house is burning down, do you demand a fire fighting model to predict the degree of damage before using water to fight the fire?”

No, but I do want to know whether there is in fact a fire before I flood the house with firehoses.

In the present case, the only indication that humans are causing significant global warming via CO2 emissions is based on climate models. If the models are inadequate, the proposition that decreasing fossil fuel burning will benefit the climate becomes untenable. A role for fossil fuels in global warming hangs on the only period of global warming observed since WWII, namely 1979-1998. Since existing global warming models have not been verified as able to predict natural (i.e., non-CO2 based)warming, we have no basis for deciding whether the warming from 1979-1998 was attributable to anthropogenic CO2 or natural variability.

A proper understanding of the “precautionary principle” does not suggest that we ought to impoverish humanity by abrogating the energy sources which have lifted us out of preindustrial poverty, unless we are quite certain that the alternative is worse. The failure of the climate models suggests that we must instead wait for credible evidence.

Figure 3 of that same linked realclimate posting indicates that observed T are below all of the high, medium and low T projections of Hanson from 1988.

Parenthetically, by “all the models” used by IPCC, I had intended to indicate that each of the separate models used by the IPCC to construct their ensemble had predicted that T would rise over the interval 2000-present. I have seen an IPCC graph presenting the central tendency of each of 17 or 22 models under each of 3 climate assumptions over the 2000-2011 period (sorry,I didn’t save the link). Each of the resulting 56 projections indicated that T would increase, rather than remain unchanged as has been observed.

Lotharsson further argues that “in a noisy system with an underlying trend,” a 15 year period of flat T should be expected to occur randomly. I do not dispute this point; indeed, I embrace it. Similarly, one should expect that over a 20 year interval (1979-1998) one might observe a random warming period despite an underlying trend of no T change. The warming over the interval 1979-1998 has nonetheless become the “evidence” cited to support the notion of AGW (1940-1979 showed cooling, and 1998-2011 is flat). You can’t have it both ways.

Finally, Lotharsson argues that Spencer & Braswell should not be cited to support the proposition that current IPCC models do not adequately model ENSO events; Lotharsson links to a skepticalscience posting purporting to debunk S&B. However, the link in fact supports my assertion that IPCC models do not accurately model ENSO events. Examine each of their 3 figures: the model projections diverge from the observed T changes under even the best model. I rest my case.

Here’s the difference: Dan Rather claimed to be a journalist who had examined the authenticity of the Bush documents and was vouching for their authenticity. If Gleich had merely revealed that he had received the “strategy” memo, without vouching for its authenticity, he couldn’t reasonably be criticized. And since others would assume the Heartland “strategy” memo was true despite Heartland’s denial (as is actually happening even now, despite Gleich’s admission), Heartland would be damaged rather than Gleich.

When you buy insurance, how predictable does the potential loss have to be?

Over what time frame is the melting of the Greenland ice sheet an acceptable risk? In other words, over what time frame would a 7 meter rise in sea level be acceptable?

…all of the climate models used by the IPCC have predicted warming much, much greater than that actually observed.

All of them? You know this…how? What’s your definition of “much much greater” and where does it fit with this analysis which suggests your sweeping claim is not supported by the data?

None of these models predicted the leveling off in global temperatures observed over the last 15 years.

Engineers and scientists alike well know that in a noisy system with an underlying trend, that is precisely what one expects every now and then. And indeed, the results are entirely within the envelope of outcomes projected by the models – and Meehl at al (2011) demonstrate in a model a mechanism that leads to this kind of observation.

As the recent Spencer & Braswell paper demonstrated, none of the IPCC models accurately predict the T changes observed before and after an el nino event.

And as recent peer review demonstrated, the recent Spencer and Braswell paper was so deeply flawed that it should not have been published in a peer-reviewed journal and anyone who relies on it indicates they know not what they are talking about.

Might be worth finding better sources – or sticking to the memos on this thread.