Computer models provide evidence in the same way as clairvoyants and fortune-tellers do.

Computer models can provide guidance as to what is likely to happen. The accuracy of the prediction is down to how tightly constrained the problem is and how well the relationship between inputs and the result is understood.

An example is the simulation of the airflow over a wing, lift from a wing or potential interference between two tracks on a PCB knowing the characteristics of the agressor signal. All fairly well understood, constrained problems. Having said that the engineer who relies blindly on what the models say might happen is doomed to failure.

I should have added the mechanism of 'climate' is so badly understood the 'models' are almost completely worthless.

I should also have added that although the examples I gave above are relatively simple, the software/ algorithms to solve the problems are really fairly complex if you want an accurate prediction (complex enough that companies make good money writing it).

> But I don't think such predictions are the same thing as providing "evidence".

No, in most engineering applications you need to test your wing or PCB to get the 'evidence' that it's going to do what you expect it to.

Climate models (GCMs) do indeed provide evidence.

That the climate is not sufficiently understood to enable its future behaviour to be predicted reliably.

Rhoda, does that mean that evidence of ignorance is the same as ignorance of evidence? :)

More seriously, did Trenberth re-examine the output from a previous model run and find the missing heat skulking at the bottom of the oceans, or was it just assumed to be there? If the former is the case, I wonder if a description been given of the path taken by the heat.

Having said that, I have myself used molecular modelling results to lend credence to the probable structure of a class of compounds. In the publication I think I used a phrase like "in support of this, a molecular model...", not "evidence" . I would describe it as a similar situation to the engineering cases described above in that it confirmed what I already hoped for/expected for other reasons. The real evidence came later.

Some of the models also provided something somewhat 'unexpected' in that the calculations reached an interesting unusual state where the programs found it very difficult to finally converge to an energy minimum/precise structure. As a result, what should have taken hours/days would run on for months if I let them. To this day I am not exactly sure what, if anything, that told me about either the models or the molecules.

Jul 8, 2014 at 11:53 AM | rhoda

the clue to all that heat which is going into the ocean

1. All other aspects of the model are correct (unlikely in a complex chaotic system)
2 The missing heat isn't going somewhere else
3 The measurements are accurate (not proven)
4. The heat actually is missing

?

I have seen various definitions of 'evidence', including:

- A thing or things helpful in forming a conclusion or judgment
- Something indicative; an outward sign
- Ground for belief or disbelief

Model output seems to fit each of these. That is not to say it is 'proof'. Just like lines of evidence in a court case may be false or may still point to the wrong verdict. Models are tools, they serve a purpose. They are not perfect and their outputs need to be interpreted and treated with caution. But that doesn't prevent them from being useful.

If you reject models, what do you propose in their place, if (although this seems unlikely) you want to understand the likely course of the climate.

rhoda
Sorry I missed the sarc tag! It did strike me as an odd post from you.

Raff,
If tried in a Scottish court the worst case verdict the CO2 could expect when charged with the evidence of the models against reality would be Not Proven

As previous posts have pointed out current climate models don't meet the first two of your points, and are designed to confirm the belief of the modeller, your third point.

If you reject models, what do you propose in their place, if (although this seems unlikely) you want to understand the likely course of the climate.
Jul 8, 2014 at 8:27 PM | Unregistered CommenterRaff

If I point out that something does not work, that does not put me under an obligation to propose some alternative that will work. In any case, if there are firm grounds for believing that the task is fundamentally impossible, it would not make sense for me to attempt to come up with such a thing.

It seems to be common amongst AGW believers to imply that an objection to something is not valid unless an alternative to the thing is proposed at the same time.

Sandy, you, and it seems Ridley, are confusing evidence with proof. Models are used everywhere in engineering and science and their output is generally used as evidence in support of some proposition or other. On its own though, such evidence is not proof of the proposition. For proof, some sort of validation (a confirmation that the model is probably correct) is necessary.

Ridley seems to have tripped himself up by using the word oxymoron, thinking that it made him look clever and in failing to use it correctly looking ignorant and foolish.

Models have their uses and if one wants to understand what might happen to future climate they seem to be an essential tool, along with studies of past climate. If you don't want to understand the likely course of the climate, then you need no models, no research, no scientists. That seems to be the position of your host and other readers. If you do want to understand climate but reject models, what do you propose in their place?

Big Oil, I didn't notice your post until after posting. But I'd put you in the category of no wanting to understnd climate from what you say.

Raff - I do not know why you assume that.

Observing that a thing is not understood or thinking that it is probably impossible to understand it are quite different things from not wanting to understand it.

Suppose the discussion were about the detailed evolution of the English language from its ancestors instead of about climate. If I pointed out that this is not understood and it is probably incapable of being understood, would you infer that I therefore don't desire to understand how English evolved?

Raff:

How to you get from "no models" to "no research, no scientists"?

In place of (or rather, in preference to) models, I propose scientists doing research.

Raff
Models in engineering are of well understood systems and physical properties. Despite this it is rare that the final product is actually that of the original design and modelling. As a retired test and reliability engineer I can tell you from personal experience despite all the systems put in place to ensure reliability and long life neither are guaranteed no matter how many times the models are run.

For Climate models they are dealing with a multiple input chaotic system, and as we've agreed with Entropic man there are unknown unknowns,as well as known unknowns, i'm not confusing evidence and proof what i'm saying is that climate models are so far from the actuality that they shouldn't be used for any sort of prediction or planning. They certainly shouldn't be used to gamble the nations future wellbeing on their "predictions"

rhoda +100 for the "evidence" proving the heat was missing so must be in the oceans.

Raff your quite right one could have evidence that seemed to prove something but turned out not to be proof, and indeed we do, according to climate scientists there is plenty of evidence of the existence of malign climate change, only they can't prove it because their "evidence" comes from models and fertile imaginations, not observations.

In reality however the rest of us take "evidence" to mean proof. In the case of "evidence" from models it's still not evidence, or proof, because it doesn't come from the physical world it comes from the equations and personal biasses of the modellers who put the models together. It doesn't matter whether the models are right what comes out still isn't "evidence" which is a physical thing lying around waiting to be found.

Sandy, Geronimo, you and Ridley et. al. are just confused by the word 'evidence', but it is not a difficult word. A dictionary and normal usage will help you. Most things have evidence for and evidence against. Decisions are made on the balance of evidence. Proof is another concept entirely.

What is the first thing people do if they want to understand a system, complicated or simple? They create a model, either mental or pen and paper or in a computer. Does the model get it 100% right. Sometimes yes and sometimes no. But it helps to visualize and to predict. It is a tool.

We know that climate is changing and will continue to change. This itself is a risk for our nations' future wellbeings, yet you give me no idea of how you intend to try to assess that risk, let alone to mitigate. Please tell me what you propose in place of models. Crossed fingers?

Jolly Farmer, see the link below and the lack of any complaint from the regulars about cutting research. Oh the lucky country...
http://www.bishop-hill.net/blog/2014/6/4/hitting-back-at-scientivists.html