"Both return to ask essentially the same question again and again."

The reason is that you and others continually cannot answer simple questions. Answer and I'll stop asking.

In contrast to you, Rhoda at least seems to accept that models could be useful if 'better'. But like you she cannot answer me what criteria would be used to judge them 'better' (or 'good' or whatever). Unlike you, she at least has some options for answering, whereas having ruled out models you are stuck with, "sorry guv, it's too difficult".

If you have opinions on this that are based on anything solid rather than hand waving, it is not apparent from your replies.

> Nial, can you describe [or illustrate] the probability distribution function produced by your model?

Yes. "Not Very".

There are _many_ more urgent problems facing mankind _now_ that need attention rather than getting our knickers in a twist over something that probably never will.

What criteria would be used to judge them "better".

Try here

https://pielkeclimatesci.wordpress.com/2012/04/11/guest-post-by-dan-hughes-on-climate-modeling/

Martin, are you a politician? Like you, they always find ways to avoid answering simple questions for which the answer is obvious but inconvenient to their ideology. Just answer the question.

Raff

Like you, they always find ways to avoid answering simple questions for which the answer is obvious but inconvenient to their ideology.

Where I come from the response to that is it takes one to know one.

raff I don't know whether you're not listening, or bothering to read the answers. The models are crap at forecasting, so if they consistently gave accurate forecasts they would be useful. There that was easy wasn't it, except of course they can't. Give accurate forecasts that is, because the climate is chaotic and forecasting the future states of a chaotic system is impossible.

"In sum, a strategy must recognise what is possible. In climate research and modelling, we should recognise that we are dealing with a coupled non-linear chaotic system, and therefore that the long-term prediction of future climate states is not possible.” IPCC TAR 14.2.2.2

As predictive ability is the hallmark of a "useful" model, and as the IPCC admits the future state of the climate can't be modelled it might follow, if one was to reason this out, that the models are useless, not useful.

Models could be described as educated guesses. The value of those models is based on their ability to guess correctly based on observation, not more guesses. Climate models are yet to demonstrate ability and are not yet useful tools. Until they can show some skill they remain useless. In few areas are the outputs from models used directly to impact upon the public. Medicine doesn’t do it, engineering doesn’t do it. Economics and financial areas are some of the few where guesses are acted upon and we all know how that can turn out. One might even call those fields as gambling.

So should we act upon climate models as the best guess? Well you can if you want to, I’m not interested. When I’m not sure about how to solve a problem or if the problem exists I set aside money to react quickly when either the problem or the solution becomes apparent. I don’t throw my money at the first shiny object that catches my eye. Doing nothing to wait for more data is a viable choice.

A common claim from warmists is that we must act. No, we don’t. We have free will and even if that means we make a mistake, we will act true to our theories, not yours. Only where the majority agree to act can the objectors be overridden. Warmists do not have that power or support. So it doesn’t matter how much warmists whinge that the models provide clear proof og CAGW, it isn’t their minds that need persuading. So are the models evidence? 400+ppm of CO2 and rising suggests not.

Mike J: "The trouble with that reference to TAR is that "the world has moved on" from those heady days when the people at the IPCC hadn't fully got the message and thought that they were supposed to be honest about the science of climate..."

Yes I know Mike, but it's still true. Not only is that true, but they're making their predictions on a chaotic system that interacts with (at least) two other chaotic systems, technology and society. I used to give lectures to marketing people on the value of forecasts and I (at least I believe I did) invented the term "static analysis" where people forecast the future behaviour of a market based on their knowledge of two parameters without regard to other parameters that might affect the outcome of their proposal. Malthus is the classic example, the maths was good, the rational was good, but he failed to take account of other parameters that interacted with his two. His were population and the ability to produce food. He ignored, or wasn't aware of, the changes in technology and society. He was wrong, but strangely he still has his supporters among "intelligent" people.

A propos of nothing I have a house in France, but couldn't live there permanently, I'm not suited for a rural lifestyle, although my house is in the centre of Cognac it's a bit slow for me.

Martin, no I have never constructed a model of any sort of distributed system. I have designed and built actual distributed embedded realtime systems though, if it helps.

You and others here admit (or don't contest) that they don't know what effect CO2 emissions will have in coming decades. You also don't contest that temperatures might rise by several degress (and that a change of 4 degrees C is comparable to exit from a glaciation event). So explain to me why ignoring this is better than studying it, even if that study involves models. You can repeat forever that models cannot be proven to be accurate, but all the while we study climate we improve our understanding. Even experiments that fail can improve our understanding (in answering why the fail). Saying it is too difficult and so we shouldn't do it is the response I expect from my daughter to her math problems, not from scientists and engineers. You offer nothing but sitting on our hands.

We know the risks even if we cannot quantify them exectly. And we already know the solution - cutting CO2. If you admit there is a risk, why are you implaccably opposed to solution?

And what is you "clear evidence that climate models for predicting future climate simply don't work"? And Geronimo, what would a good model look like to you. Would it need to predict volcanos or particulate emissions, solar cycles or El Ninos? Would its projections be fundamentally wrong if it failed to forecast these events.

I think that Raff is being highly provocative in claiming " You also don't contest that temperatures might rise by several degress ". That is an entirely unjustified presumption and just sets up a straw man.

The basic issue is that there is a theory about climate change based around CO2 and that various models (whether computer code or even spreadsheets) then help to assist with the development of the theory. If the models undergo adequate testing and verification they will then provide results in line with that theory and may allow people to develop the theory further. However to then claim the results are evidence is circular reasoning because all they are doing is showing what the theory does.

In order to be useful both the theory and the models need to be validated. This requires both testing against real data and confirmation that any parameters used are reasonable and in line with the scientific literature. The evidence to date is that validation where done at all is superficial.

http://blog.al.com/breaking/2014/05/what_happened_during_house_sci.html
http://judithcurry.com/2010/12/01/climate-model-verification-and-validation/
http://climateaudit.org/2008/02/03/curry-reviews-jablonski-and-williamson/

The issues with validation are now well exposed due to the failure of models to match the "pause" and I would expect that many models are now paying much more attention to it. However until the modellers do produce models that validate much better than the current crop then skeptics will continue to be unimpressed.

TinyCO2, July 11th 2.33 & Raff all posts.

Climate GCM's are WORSE than useless; they are expensive and dangerous as they have been used to make far reaching policy which effects millions in the under-developed world. Contributing to the practice of burning cow dung indoors with consequent impact on the health of millions. They have helped to suppress industry and economic activity in the developed world. In short they are a total disaster.

I think the ac suffix is similar to wick in English place names, indicating a settlement. Not to be confused with Wick in Northern Scotland which I think is derived from Uig - a port.

clivere, you say that my claim that noone disagrees that temperatures might rise by several degress, is an entirely unjustified presumption. It is actually a logical conclusion. To save you looking back, it originates from Martin's assertion that "we simply don't know", which noone disputed. To that, I commented (Jul 9, 2014 at 11:07 PM | Unregistered CommenterRaff):

However, if you really characterise your own position as "We simply don't know", then I still cannot believe you mean "we know nothing". That would imply that you think the probability of any outcome is equal - that two decades from now global temperatures have an equal chance of being zero Kelvin or infinite. I imagine you would agree that those are impossible extremes, so I have to assume you would consider some values more likely than others - you have some sort of probability distribution in mind.

So to repeat myself, as you have stated that your own mental model of the world cannot show that the effect of changing CO2 levels has a probability distribution that is infinitely narrow (in other words, that your model tells you exactly what will happen with no doubt), there is inevitably some sort of probability distribution to what your model actually predicts. And hence there will be tails of uncertainty. There is hence a non-zero probability that temperature will change as a result of extra CO2 by 4 degrees (say), or -4 degrees. The latter is comparable to entering a new glaciation. What does the former equate to? How big is the risk? Should we try to understand it and mitigate it? You don't know the probability of a 4C rise and it is difficult to see how you can argue that the risk is zero. How do you suggest we determine that probability - whether we have a problem - without models or paleo studies? Or do we just say, "its too difficult, let's ignore it"?

Micky, you are at least willing to nail you colors to the mast of "nothing happens", as opposed to "We simply don't know" from Martin. That is a bold claim, given what 150 years of research has told us about CO2 and radiation. And unless you can say with certainty that *absolutely* nothing will happen on doubling CO2, which as an extroadinary statement requires extraordinary evidence, your "nothing happens" is just the ceter of a probability distribution of unknown shape - unknown because you have pulled it out of thin air and have no calculations to back it up. So there are tails to the distribution that are non-zero at significant changes in temperature. But you cannot quantify them - how would you do that?

geronimo, you claim that a model of the climate should be able to model solar cycles, tectonic activity, social behavior (emissions) and ocean cycles. You seem to think that it is "outrageously, obviously, stupid" to question this. With the exception of ocean cycles, which we might hope to capture in a climate model, the mind boggles.