EM, I find it very odd that neither modeler Simon nor Martin nor anyone else can bear to address the fact that, once the forcings are adjusted (as you described much more clearly) temperatures have matched the models very well. Martin just uses a tactic often deployed of apparent deliberate misunderstanding to deflect the question despite him being in the best position to address this. The rest seem to base their rejection of models on nothing more than other people's rejection of models. There is no original thought.

To the very basic question of whether, if there were no models, we should (a) write models to see what we can learn, or (b) not write models, very few seem able to respond. Answer (b), that we shouldn't even try, appears to me and probably to everyone to be a crazy answer, so the reason people don't like to answer is that of course we should write models. Why wouldn't we if we stand to learn something. People like Martin will might say we can never learn anything from models that we cannot validate, but that would be equivalent to answering (b) and so maybe he wouldn't or maybe he would just have little support. But other people say we can only write them, we can't use what we learn from them for 'policy'. That too seems self evidently crazy. What a dilemma for a "skeptic"!

If people really want to "get real" about climate models they need to address this dilemma. I doubt they will.

Raff, when climate science addresses the fact that models are not able to produce anything that is of use, you might have a point. I doubt they will.

Raff, or the Joyce of Cork. It's a poor choyce at Bishop Hill today.

I remember reading a post on SOD on GCMs and the definition of climate. He suggested that the only basis for assuming that 30 years is climate would be if you could show that GCM runs of 30 years covered the entire range of weather outcomes. He was dubious that that had been demonstrated. It suggests that the modellers don't know that climate is a deterministic system but here we have EM asserting it without any evidence whatsoever.

Diogenes,
I don't think there is any dispute that the system is deterministic. It clearly obeys the laws of physics. The issue is the significance of the non-linearities and the sensitivity to initial conditions. It's clear that it is chaotic and that we can't specify the initial conditions with sufficient accuracy so as to make actual predictions beyond a period of days. However, it is also a boundary value problem and the general view is that the boundary values start to dominate when it comes to the overall average climate state. Of course, it is possible that the there are variations on scales longer than 30 years which means that averaging over only 30 years means that you're not capturing all of the possible variability. On the other hand, you do have to choose a timescale that maximises the chance of producing an average that reasonably represents the climate conditions while not choosing a timescale that is so long that you're actually averaging over a period over which the climate has actually changed. Given that rate at which we're changing the climate, 30 years is probably a reasonable time interval.

Martin A

merely that its control knobs have been twiddled to reproduce its training data.

That's why software (and hardware for that matter), including models, should be tested by people who want to break it.

Even well understood mechanical engineering project can still go wrong, as in the recent Nipigon Bridge failure

The future westbound bridge opened on November 29, 2015; both directions of traffic were shifted onto the new bridge to prepare the old span for demolition.

On January 10, 2016, the new bridge was closed to traffic after an expansion joint on the deck was lifted by 60 centimetres (24 in) after a winter storm,[5] resulting in the indefinite closure of the Trans-Canada Highway at the bridge
(wikipedia)

Which appears to be quite an important link Nipigon bridge partial closure slows $100-million of goods shipped daily

Nature likes nothing more than breaking stuff, especially CliSci™ models ;-)

It's my experience that those who didn't learn the principles of strictly adhering to the scientific method will never be able to appreciate their importance.

It's my experience that those who invoke the sanctity of the scientific method typically don't know what they're talking about. It's really just a rather pathetic appeal to authority. You can do better, can't you?

Martin A

Curious how you reject the holy Climate Science Formula, yet accept the Holy Ohm's Law. Clearly you only accept empirical formulae that tell you what you want to believe.

You still have this belief that radiative forcing is not measurable. In fact it can be calculated from theory, confirmed in the laboratory, and then confirmed by two independent techniques of measurement in the field

Firstly, the absorption spectra of greenhouse gases are well known from laboratory measurement. We know both the spot absorbtion frequencies and the effects of band spreading in a well mixed atmosphere. Forcing can be calculated from theory using this data and radiative physics. (Sorry, BH!)

Secondly, the outward radiation spectrum and intensities are monitored from orbit. At the absorption frequencies relevant to each greenhouse gas one can measure the decrease in energy output relative to the black body spectrum. Integrating these values gives the radiative forcing for the gas.

Thirdly, the intensity and spectrum of downwelling radiation is monitored from the ground. Once again calculating total energy at the emission wavelengths for each gas gives the radiative forcing for that gas.

Radiative forcing is predictable from theory and measured by three independent lines of empirical evidence. It exists as a measurable phenomenon, despite your lack of belief.

"Deterministic"

The climate system is a thermodynamic machine. Energy enters the system, moves around within the system and leaves the system in accordance with the laws of thermodynamics, the Navier-Stokes equations, the Stephan-Boltzmann equation and a number of other formulae governing the movement of energy around the system and the space enclosing it.

Given stable conditions the system will come to equilibrium. Change the conditions and the equilibrium will be lost until conditions stabilise, when a new equilibrium will eventually emerge.

On a planetary scale and over geological timescale these processes are deterministic. Repeat the conditions and the physics produces the same equilibrium. This is deterministic.

Over short timescales heat transfer from the tropics to the poles is turbulent and, in the mathematical sense, chaotic. This applies particularly under current warming conditiones when the whole system is out of equilibrium. While the underlying physics is deterministic, the heat transfer process itself is partially stochastic, making it difficult to get enough information for weather prediction.

Raff

I'm with Tamsin Edwards.

"All models are wrong, but some models are useful"

A model is perforce a simplified version of reality, so is unable to exactly simulate reality. It can, however, give useful information about outcomes.

From the earliest 1D models to the current GCMs that has applied. Over 40 years of modelling the underlying trend has been and continues to be correctly projected, within the confidence limits of the respective models.

The locals at BH and similar sites have a dogma that AGW is not happening. If the models say it is happening, then the models must be wrong. The inability of any model to match reality without the CO2 greenhouse effect just makes it worse.

This debate is not about the quality of models. It. Is about heresy.It is an affirmation of Bishop Hill's theology of climate change denial, by his acolytes.

Simon,

It's no surprise to me that you reject the scientific method as a logical fallacy.

I didn't say that. It's hard to take someone's appeal to the sanctity of the scientific method seriously, when it appears that they're comfortable making stuff up when responding to what someone else has said. Additionally, it's hard to take someone's appeal to the scientific method seriously when they are a regular on a site that largely rejects it. Come on, be serious.

Martin,

Well modelling virtually any bounded physical system is a boundary value problem: so what?

The issue is whether or not it is dominated by the initial conditions or the boundary conditions.

Some of the tenets of the gospel of climate "science". All matters of faith, with no objective basis:

◙ "weather" = (< 30 years, random and can't be predicted)

◙ "climate" = (> 30 years, deterministic and can be predicted)

◙ "we're changing the climate"

Did you actually read what I wrote? Do you want to read it again and respond one more time. Strawman games are a waste of time.

Simon,
I think you first need to read what I wrote again. I'll even repeat it

It's my experience that those who invoke the sanctity of the scientific method typically don't know what they're talking about.

Do you need me to explain it too? It seems pretty self-evident that nowhere did I reject the scientific method as a logical fallacy. I'm certainly not about to answer your loaded questions if you still maintain that what I said was somehow rejecting the scientific method. If you can't indicate that you can correctly interpret what others are saying, there's not much point in engaging further.

Simon Hopkinson, And then There's Physics

You seem to have moved on from climate models into scientific method and the philosophy of science.

May I ask each of you whether you think we should be investigating climate science using the approach of Popper or Kuhn.

Popper's approach of basic hypothesis testing is regarded as most useful in the early stages of a new field when a number of competing hypotheses are still viable.

Kuhn's approach better fits a more developed field in which a consensus has emerged around one hypothesis and work continues to fill in the details.