SandyS. I'm not suggesting that there's one problem, just that something is wrong.

Nullis in verba’s comment of Aug 11, 2015 at 10:14 PM (foot page 17) is very rich in content and logically correct. He rightly points out the statement

"you acknowledge that scientists predicted warming. And warming is what we observed."

commits the fallacy of "confirming the consequent”. But it could be said of more content-rich retrospective “predictions”. Try this statement

From 73 climate models the prediction is that between 1979 and 2012 average global temperatures would rise by between 0.3 and 1.5C, with an average estimate of 0.8C. Most were within the 0.6 to 1.2C.

The actual rise was 0.2C, (as Roy Spencer showed), but suppose the rise had been 0.7C and there was a slight acceleration post 2000 in line with the emissions accelerating? It still commits the fallacy of "confirming the consequent”, so still does not prove anything. But it better distinguishes CAGW than the banal statement. It is the same with real predictions. If hurricanes making landfall had increased in frequency & intensity post Katrina in 2005, it would not prove it is human caused, but it would be so unusual, it would be unlikely to random.
But how do we distinguish between the more scientific and the less scientific statements? Is climatology progressing towards more precise empirical statements that support CAGW, or degenerating into ever more bland and dogmatic statements? Consider the following statement

CAGW ⊂ AGW ⊂ GW ⊂ CC
Where
GW = the global warming hypothesis
CC = Climate Change all types
⊂ = "a subset of"
I explain it in more detail here, where there is also a diagram of concentric rings illustrating the same point.

The key elements of progression are empirical statements that combine increasing empirical content and greater relevancy to the catastrophic anthropogenic global warming hypothesis. By combining these elements, the CAGW hypothesis is given stronger support from the real world. I cannot find examples of this happening, but plenty of the opposite. What do others think?

Kevin Marshall

You have just demonstrated why the formal logic of propositional calculus should not be applied to science.It is too simplistic, which is why higher-order logical systems were developed.

When you formulate a hypothesis, , and use it to predict the outcome of observation or experiment you are making the statement "If P then Q". No scientific hypothesis can never be proved correct by experiment because, as you say, there is always the possibility that another effect has caused the observed result.
This is why Popper introduced the concept of falsifiability.

The CMIP5 models apply " If P then Q" to projected temperatures. They are based on the same physical model and data as HadCRUT and their output is surface temperaures. The correct comparison between models and observation is with HadCRUT4 like this.
Note that HadCRUT4 shows a rise of -0.6C between 1979 and 2012, well within the projected range.

Roy Spencer's comparison between model outputs and radiosonde/satellite data is comparing surface projections with atmospheric temperatures at an unspecified height above the surface. Lapse rate will mean that his figures will be lower than the model output regardless of any other factors.Anything he induces about the projective capability of the models is therefore meaningless because he is not comparing like with like.

I don't know where you get your graphs from, but that graph isn't figure 11.5 from AR5 WG1. If you tool over to these pages Figure 11.9you will find the information on this the graphs in there that shows far from being well within the projected range HadCRUT4 is at the bumping along at the very bottom of the projected range. Quite why you don't understand that the period 1976 to 2012 has already passed and adjustments have been made to bring the past projections in line with the observations I don't know.

Martin A, geronimo

The problems with formal logic come when one tries to apply it to the real world.

Strict application of the propositional calculus would make all science impossible, not just climatology. It would also invalidate any system using evidence, from law to engineering.

The graph I cited is updated from IPCC to include Hadcrut4 to Aprli 2014. I got it here . The most recent version also includes other datasets adjusted to give a common baseline throughout.

You will be pleased to know that the HadCRUT4 monthly figures for 2015 to date average 0.68C, which will put 2015 well within the projected range of uncertainty defined by the models.

Don't get too hooked on RSS, it is looking unreliable. I put up this graph earlier. Note that all the datasets except RSS show warming since 1998. That includes the other satellite dataset, UAH. RSS is the only one still showing a decline.

You're misleading yourself again EM. When trends are close to zero you should not be surprised that small changes in the data can change the sign of a trend.

The graph you cited was identified itself as an IPCC graph, it isn't. In fact the graph doesn't exist in the IPCC AR5. Of course SkS have been trying desperately to overcome the awful performance of the models. AFAICT the "updated" graph you pointed me too included the Cowtan and Way 2014 dataset and can't be compared to the HadCRUT 4 dataset which was used by the IPCC . It is clear to me that you're getting your information from the kidz at SkS.

I'm really surprised you're persisting here given the number of massive howlers you've made on this thread.

BTW have you worked out the blunder you and Dana (or should I just say Dana) made in calculating the 1910 - 1940 TCR yet? It's pretty simple really, but I suppose we'll have to wait for Dana to figure it out.

BTW updating your forecasts post hoc is only practised by climatologists as far as I'm aware.

Michael hart

The test is statistical significance.

Consider the GISS data. The 95% confidence limits for the global annual averages are +/-0.09C. The threshold for 95% significance is a difference between the start and end of a trend of 4SD+. In this case the difference would need to be 0.18C.

This is the linear trend for GISS from 1998 to date. I have included 95% confidence limits.

The difference between the beginning and the end of the trend is 0.21C. This is more than 4SD, a 95% significant difference.

This is not a "close to zero" trend. I chose 1998 deliberately, because it also indicates that the "pause", if it ever existed, is over.

Martin A

Your cockerel example illustrates my point. It is possible to falsify the cockerel's hypothesis (If P then Q) by experiment. Stop the cockerel crowing and see if the Sun still rises. If the Sun does not rise (If not P not Q) then the cockerel is correct. If the Sun rises ( If not P then Q), then " If P then Q" has been falsified.

To test the hypothesis that increased greenhouse gases generate global warming is more difficult. You can demonstrate the effect on the laboratory scale, but do so on a planetary scale needs a duplicate planet on which greenhouse gas concentrations have been kept constant. If that control planet warmed as much as our Earth, global warming would be falsified.

The resources for such an experiment are not yet available. How do you distinguish between the fallacy of "confirming the consequent" and genuine cause and effect?

"If P then Q" must be shown to be false.You might do so in two ways.

Firstly by demonstrating that the mechanism linking cause and effect is not sound (If P not always Q).

Secondly by demonstrating that another cause better explains the observations (If R then Q).

This is how science operates. It is known in formal logic as modus tollens or "denying the consequent"

Entropic Man, what part of your own words don't you understand? You pointed out, correctly, that the satellite observations and surface thermometers are technically not measuring the same thing. So by your own words, that limits the inferences you can draw about the differences between them. And then you go ahead and do it anyway. You know...comparing apples and oranges. (*)

The satellite trends are much closer to zero, so it is not remarkable that UAH could be slightly positive and RSS slightly negative. Nothing you have written shows that either is better than the other. And the recent update of UAH methodology has actually brought it closer to RSS, so your claim that RSS is spuriously low is still unsubstantiated.

I know you read Roy Spencer's blog (or, more correctly, I know you comment there) so I recommend you read what he has said himself about the differences.
http://www.drroyspencer.com/2012/05/our-response-to-recent-criticism-of-the-uah-satellite-temperatures
http://www.drroyspencer.com/2015/01/why-do-different-satellite-datasets-produce-different-global-temperature-trends/

I'll add that the satellite measurements have radiosonde measurements as an independent source of verification. So I would advise you to stick to arguing that they are different beasts to surface thermometers, and not go wandering off down statistical dark alleys without a candle.

*(The curious might also wonder if the growing disparity between them might be the true signal of increased CO2 concentrations. Unfortunately for some, that disparity is not big enough for alarmista to claim the sky is falling on our heads. And some no longer trust the way the surface data is updated, to put it mildly. Whatever, the satellite data measurement can at least claim to be physically closer to the location where outgoing radiation occurs.)

Geronimo

"Sigh"

Go to Ed Hawkins' post.

Read the legend below the graph.

"Updated version of IPCC AR5 Figure 11.25 with HadCRUT4.3 (black) global temperature time-series. The CMIP5 model projections are shown relative to 1986-2005 (light grey). The red hatching is the IPCC AR5 assessed likely range for global temperatures in the 2016-2035 period. The blue lines represent other observational datasets (Cowtan & Way, NASA GISTEMP, NOAA NCDC, ERA-Interim, BEST). "

Now look again at the graph. Ed Hawkins' post has his most recent updated graph including HadCRUT4 to April 2015.

"BTW updating your forecasts post hoc is only practised by climatologists as far as I'm aware."

What a silly thing for an engineer to say. In any field techniques are improved by iteration Climate models are no different.

Note that Roy Spencer's UAH model, used to derive atmospheric temperatures from microwave emissions, is now in Version 6.0.

EM. Go to your post and look at the graph you pointed to.

Of course in any field improvements are made by iteration. What are you daft? Forecasts are made perfect by updating them post hoc.

Still waiting for massive penny to drop on you, and your hero, Dana's, opportunity to fiddle the TCR for 1910 -1940.

Harold W

From Cowton et al 2015, Robust comparison of climate models with observations using blended land air and ocean sea surface temperatures

"Global mean temperatures from climate model simulations are typically calculated using surface air tem-
peratures, while the corresponding observations are based on a blend of air
and sea surface temperatures."

Geronimo

"Still waiting for massive penny to drop on you, and your hero, Dana's, opportunity to fiddle the TCR for 1910 -1940."

Make your point clearly without all your usual bullshit.

Entropic Man Aug 15, 2015 at 11:55 PM
You have just looked at the example, not at my reasoning.
In the blog article I quote Richard Feynman speaking in a lecture on The Scientific Method. It is at about 5.10 here.

You cannot prove a vague theory wrong. If the guess that you make is poorly expressed and the method you have for computing the consequences is a little vague then ….. you see that the theory is good as it can’t be proved wrong. If the process of computing the consequences is indefinite, then with a little skill any experimental result can be made to look like an expected consequence.

Feynman is speaking about an extreme form of Popperian falsification. If you accept it, climatology will always be a non-falsifiable vague theory, so therefore can never be a science. This is why I propose we should look to classify empirical statements about climate into which area of the theory they support – from the (potentially) policy-justifying CAGW to the banal support for the global warming hypothesis. I would also grade statements according to their quality and relative lack of ambiguity.
It should be noted that Popper developed the idea of falsification as the demarcation of science from pseudo-science as a reaction to both Freudian Psychoanalysis and Marxism. Both claimed to be scientific, but could always explain any event in terms of the theoretical paradigm. In climatology we have the same issues. In particular, there appears to be admitted circumstances where mitigating through cutting carbon emissions would not be desirable. My view is that from a climate change perspective it is only when you can demonstrate reasonable empirical support for the potential non-trivial and adverse consequences of rising GHGs that you should start considering policy, hence the need to distinguish between the trivial and non-trivial and the strength of the evidence.