Click images for more details



Recent comments
Recent posts
Currently discussing

A few sites I've stumbled across recently....

Powered by Squarespace

Discussion > Are goal posts moving?-How long until climate models are definitely WRONG?

The Pause-How long until climate models are definitely WRONG?

As 'climate scientists' react to the [15, 16 year?] pause in global warming, the von Storch interview in Der Spiegel reminds me - is there a post here tracking all of these declarations?

My vague recollection is that Gavin Schmidt has said 12 or 15 years...

Hans von Storch says another 5 years (is that 21 years? or 20? total).

And climate modeler with Lawrence Livermore National Labs (in California), Ben Santer, has said 17 years - which is now close....

For those of us interested in documenting whether and how much) 'goal posts' are moving, place you links and sources HERE, please.

Jun 21, 2013 at 8:25 PM | Unregistered CommenterOrson

The models will never be admitted to be wrong. They will be continually modified to meet reality at least in hindcast. It's not a reversal, it is merely a process of continual imrovement. And we sceptics will never be right. Even when the sensitivity is adjusted to a realistically low figure, we will still be deniers. Nobody will accept that this is what we said all along, for we are anti-science.

Rhoda's position: All the models have 5.35ln(c/c0) programmed in. This number is wrong, or inappropriate feedbacks are used. That is why they are all wrong. When the modellers and the CAGW crowd fix that, as they will, they won't tell us what they did. They WILL keep telling us they have a unique understanding of the problem, just as they do now.

Jun 21, 2013 at 8:36 PM | Registered Commenterrhoda

"The Pause-How long until climate models are definitely WRONG?"

In any proper modelling exercise, that sort of question would have been built in to the validation process.

Climate models have never been validated, because of the impossibility of doing so when the system being modelled is not understood from many aspects and where only a brief and incomplete record of its behaviour is available.

Hind testing, said by the MO to confirm their models, does not provide validation if only because it involves reproducing climate over the same period as acquisition of data used in constructing the models and for "parameterisation".

The question is related to the weather/climate issue. I commented, on another thread, that, to tell which is weather and which is climate, you need to have statistical models of both. Since statistical models of neither are available, there is no way to answer the question (other than making it up). In fact, the question is meaningless.

It is slowly dawning on me that I have probably stumbled across the fundamental equation of climate science (CS):

CS = bullshit × (bollocks + physics)

bollocks = nonsense;
bullshit = stuff made up to convey the desired impression and which cannot immediately be shown to be false.

Jun 21, 2013 at 10:10 PM | Registered CommenterMartin A

Martin A
I understand the bollocks and the bullshit bit but I'm not sure where you get the physics from.
Without going all the way with AlecM I am rapidly coming to the conclusion that they have either never got the physics right in the first place or that the input has been some sort of guesswork (what you call "parameterization", I think) couched in "physicsey" language.
Robert Brown has said most eloquently (I paraphrase): if all the models are supposedly based on the same laws of physics and they are all wrong you can't draw any conclusions from them. I would go further and say that if they differ to the extent they evidently do then they cannot be based on the laws of physics.

Jun 22, 2013 at 10:23 AM | Registered CommenterMike Jackson

Good thread title. Climate science seems to be "goalposts on rails" at the moment. Have you noticed how they dont seem so keen to plot temperature graphs these days? There is a tendency to talk less about warming and more about "climate disruption" and weather events.

Jun 22, 2013 at 11:03 AM | Registered CommenterPaul Matthews

I think that Martin's equation is almost right, but I would add an extra term to account for the low pass filters and "continuity resets" in the integration: SQR(SFA) - the square root of sweet f**k all. In other words, what you have is a random number generator, or no long term natural vatiability. However, when you multiply it all by Rhoda's term - 5.35ln(c/c0) - that is all you will see as the integration progresses.

Slingo has stated that models are the only way to study the effect of GHGs on the climate, and this in itself is an admission that there is no empirical evidence for AGW. As the models can not be verified (due to uncertainties in the variables, coupling, non-linearity, chaos, etc) they can only be validated by comparing their predictions with reality. Retrospective forecasts/predictions/projections can not do this, for reasons that have been explained ad nauseum.

So, to answer the question - how long until the models are proved to be wrong - in my opinion they already have been given their performance in this century when compared with reality. Furthermore I am not at all surprised - for all of the reasons that Dr. Brown has stated far more eloquently than I could.

Jun 22, 2013 at 11:04 AM | Unregistered CommenterRoger Longstaff

MJ - perhaps I should have put a weighting factor in front of 'physics'. It would have been non-zero as there are undoubtedly some aspects of climate science that make correct use of the laws of physics.

Jun 22, 2013 at 11:04 AM | Registered CommenterMartin A

There is a tendency to talk less about warming and more about "climate disruption" and weather events.
Jun 22, 2013 at 11:03 AM Paul Matthews

Or to redefine the meaning of "warming" to mean things other than "rising temperature".

Jun 22, 2013 at 11:11 AM | Registered CommenterMartin A

Paul Matthews

Climate science seems to be "goalposts on rails" at the moment.

Nice one. I'll raise you: "Goalposts on Ice". In a pink tutu?

I can see right up your bum :,)

Jun 22, 2013 at 11:19 AM | Registered CommenterHector Pascal

" perhaps I should have put a weighting factor in front of 'physics'."

They do numerically solve the Navier Stokes equations for some physics in moving the fluids around. The problem is this bit is largely irrelevant in determining what the global temperature will be. The temp is determined by the bit of code that Rhoda describes.

Jun 22, 2013 at 5:59 PM | Unregistered CommenterRob Burton

RB - Yes.

In computer and telecommunication performance modelling circles, the kind of situation you refer to is termed "over-modelling".

It refers to the tendency to model some parts of a system in minute detail, even though a much cruder model of those parts would make essentially no difference to the final results, because the dominant errors are due to difficulties in characterising and modelling other parts of the system, whose behaviour is poorly understood.

Jun 22, 2013 at 6:40 PM | Registered CommenterMartin A


I can see the merit in using a full blown GCM for a weather forecast (remembering that the initial observations and data you feed into it are probably even more important than the model.)

For climate, I'm sure a correctly defined simple model will be more than adequate. As repeatedly pointed out though we don't understand what influences climate yet.

Jun 22, 2013 at 10:09 PM | Unregistered CommenterRob Burton

The models have never been right. This is typical climate science.

I can make any model be right at the start. Meaningless.

How long before the models are proved right?

Even if we knew every physical process perfectly parameterised, we still do not know how those processes overlap and interact.

The distances (in time and space) and masses and size of previous events within our solar system create echoes and rhythms which are impossible to model.

We are not a god.

Models are wrong when they start. Only validation can make them right.

Validation is impossible in human time-scales.

Uncertainty is what drives science on, modelling in climate science has had none.

Take away uncertainty from the scientific process and then you just have faith and dogma.

Jun 24, 2013 at 7:50 AM | Unregistered CommenterJiminy Cricket