Buy

Books
Click images for more details

Twitter
Support

 

Recent comments
Recent posts
Links

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

Powered by Squarespace

Discussion > Let's get real about climate models

Raff, because I'm generous in nature, I would helpfully point out to you that a climate model which does not with both utmost precision and absolute accuracy, model every force and feedback in the system in the process of its output, but which successfully hindcasts, is wrong. If you believe that this is untrue and that the model is right, then you are without any shadow of a doubt deluding yourself.

I'm reminded of the Derren Brown illusion where he, apparently successfully, predicted a lottery result. I never saw it myself, but it was described to me. At the outset, Brown had 5? or 6? balls (I dunno, I don't do the lottery) lined up in a rack. As each number drawn in the lottery was called out, Brown revealed another of these balls, each in turn correctly indicating the number drawn.

If Brown had revealed the full sequence before any numbers had been called, I think I'd have been convinced that there really is magic. But of course he didn't. He progressively hindcast the results. The means by which he was able to do this is the sleight of hand, but the entire set-up is part and parcel of the illusion.

There are direct parallels between perpetuating the myth of GCM skill and perpetuating the myth that Derren Brown can see the future. With adjustable coefficients and careful tweaking, it is possible for GCMs to show in hindcast an illusion of what has become a known reality in precisely the same way Derren Brown is able to use sleight of hand to trick people into believing he has predictive powers.

But let's be clear, here. That people can be fooled into believing that GCMs can hindcast does NOT mean that magic really exists. Sorry to disappoint you, Raff.

Interestingly, Derren Brown claims that he was able to predict the lottery results by taking a sample group of lottery participants and selecting the mean of their number choices. This reminds me somewhat of the IPCC's averaging of GCM ensembles to produce a supposedly more accurate future global mean temperature projection. But I digress.

Jan 13, 2016 at 2:41 PM | Registered CommenterSimon Hopkinson

Mike, as I've probably said before, your (the "skeptics'") case rests entirely on such accusations of dishonesty and ulterior motives. It would be nice to find some logic behind that but my conclusion is that that is just the sort of people you are.

Simon, the not so artful dodger, I answered your question, please answer mine. In a "skeptical" model, if you change the solar forcing (for example) do you still get the same output? Why do you assume that changing the inputs to a climate model (e.g. intensity of the solar cycle) would not change the output?

Jan 13, 2016 at 3:09 PM | Unregistered CommenterRaff

LOL Raff! Your question is like this:

Raff: Simon, in a battle, which would win? A unicorn or a minotaur?

Simon: Raff, neither of them are real. They're both mythical.

Raff: Stop dodging the question, Simon!

Simon: Errr...

Jan 13, 2016 at 3:15 PM | Registered CommenterSimon Hopkinson

Just answer the question. It is probably not beyond you, though it might challenge your ideology.

Jan 13, 2016 at 3:38 PM | Unregistered CommenterRaff

Barf: I know you can be slow to catch on, but Simon is answering your question – there is no such thing as a “sceptical” model! Why do you refuse to see that? As I have asked before (yet, curiously, you refuse to answer), are you truly that thick, or are you being deliberately obtuse?

Jan 13, 2016 at 3:51 PM | Registered CommenterRadical Rodent

Oh another dodge. You guys are as slippery as a bar of soap in the shower. Let me rephrase the question: In any model that is influenced by solar forcing, if you change the solar forcing do you still get the same output? Why do you assume that changing the intensity of the solar cycle in a climate model would not change the output?

Jan 13, 2016 at 3:58 PM | Unregistered CommenterRaff

You still have not answered my question, then have the audacity to claim that we are being evasive. Sheesh.

Anyhoo… Apply a little logic: any influence that a model has been based upon will affect the outcomes of that model, should the influence be changed. You are asking the equivalent of: “Which is heavier, a ton of lead or a ton of feathers?” Now, please give us an example of a “skeptical” model that you have found, as you seem convinced that at least one exists.

Jan 13, 2016 at 4:59 PM | Registered CommenterRadical Rodent

Raff
You're plagerising my comments

Journal Entry Comment by SandyS on Mar 6, 2014 at 8:22 PM
It's a bit like getting hold of a bar of soap in a bath with all the threads on similar topics.

Which I feel sure I've used since, but can't find so perhaps it was elsewhere.

Jan 13, 2016 at 4:59 PM | Unregistered CommenterSandyS

I didn't know that sceptics had a "case" which rested on anything in particular, but I'm sure that if raff says so then it must be true — or more likely not.
To assume that the climate community (includes all those who stand to benefit either financially, politically, or socially from having us all believe in their claims) would not use every trick they can find to "give artistic verisimilitude to an otherwise bald and unconvincing narrative" ( WS Gilbert, The Mikado) is not realistic. The history of climate research from Hansen through Mann to Climategate and beyond demonstrates that while a few scientists still believe that facts are sacred and the data are to be believed regardless of how inconvenient they may be, a very large number along with most others involved in the Movement are not to be trusted to separate facts from conjecture from obfuscation from outright mendacity.
Don't take my word for it. By their words and their deeds shall ye know them!

Jan 13, 2016 at 5:31 PM | Registered CommenterMike Jackson

"Let's get real about climate models"

They are all rubbish, and only fit for computer gaming in Reality Climate Science. They should be sold in toy shops, and clearly labelled with a disclaimer, as producing no useable data. Otherwise Lara Croft Tomb Raider affianados would be claiming military and archaelogical expertise.

Jan 13, 2016 at 7:06 PM | Unregistered Commentergolf charlie

Mike (everyone in fact) why not try to go for one day without accusing anyone of anything? Try to formulate your objections to the varied evidence presented by climate scientists on the assumption that they are acting in good faith and that the measurements and evidence are no more or less than is claimed in the relevant research papers.

Ratty, no there is no "skeptic" model. I was just a rhetorical device - let's call it a "perfect model" if you dislike the implied ownership of "skeptic" model. So it is good that you recognize that a model's outputs depend upon its inputs - Simon seems to have trouble with that concept. Is it too much to expect you or anyone else to recognize that instead of comparing actual temperature graphs with plots from CMIP5 model runs that used inaccurate forcings, we should use plots from model adjusted for actual forcings?

Jan 13, 2016 at 7:45 PM | Unregistered CommenterRaff

it is good that you recognize that a model's outputs depend upon its inputs - Simon seems to have trouble with that concept.

Quite extraordinary. I mean in the twisted, unhinged from reality sense. You're really bizarre.

Jan 13, 2016 at 7:58 PM | Registered CommenterSimon Hopkinson

...instead of comparing actual temperature graphs with plots from CMIP5 model runs that used inaccurate forcings, we should use plots from model adjusted for actual forcings?

My lotus spreadsheet model (a simple table) is is excellent at predicting past HadCRUT4 numbers but it is completely (well almost completely) useless at predicting future HadCRUT4 numbers.

But when it has been adjusted with the new values, once they have become available, it once again predicts past values with complete accuracy

Jan 13, 2016 at 8:09 PM | Registered CommenterMartin A

… why not try to go for one day without accusing anyone of anything?
Obviously, you mean this to apply to everyone else, but not yourself. Please try not to appear too hypocritical.

Sceptics are well aware that there is no such thing as a “perfect model”; most are far too realistic to consider that possibility. While models might be able to help us to understand the processes in play, to date, they are certainly far short of being so reliable that they can be used to influence policy. Why can you not see the simple logic in that principle? You certainly are really bizarre.

Jan 13, 2016 at 8:17 PM | Registered CommenterRadical Rodent

He doesn't seem to be able to grasp the differentiation between predictability and actuality. It seems to me an odd thing for someone to be unable to conceive of. How very weird.

Jan 13, 2016 at 8:20 PM | Registered CommenterSimon Hopkinson

More wriggling. Do you recognize that instead of comparing actual temperature graphs with plots from CMIP5 model runs that used inaccurate forcings, we should use plots from model adjusted for actual forcings?

Jan 13, 2016 at 8:36 PM | Unregistered CommenterRaff

instead of comparing actual temperature graphs with plots from CMIP5 model runs that used inaccurate forcings, we should use plots from model adjusted for actual forcings?

Raff - could you please help me to understand why that would be a good idea?

Would it also be a good idea to adjust the model for other assumptions that, with new data available, would have been made differently? Coefficients computed from observational data, in parameterising difficult-to-model effects, for example.

Jan 13, 2016 at 9:28 PM | Registered CommenterMartin A

"Do you recognize that instead of comparing actual temperature graphs with plots from CMIP5 model runs that used inaccurate forcings, we should use plots from model adjusted for actual forcings?"


I'll predict next week's lottery draw, but can I wait until after it's happened to let you know the numbers?

Jan 13, 2016 at 9:37 PM | Unregistered CommenterNial

Martin A: I wonder if this is a suggestion that we should make our model graphs from graphs that reality has since provided us with? This will, of course, go to show how accurate the models were in the first place. Simples!

Jan 13, 2016 at 9:40 PM | Registered CommenterRadical Rodent

RR - I refer you to my Jan 13, 2016 at 8:09 PM posting.

Jan 13, 2016 at 9:52 PM | Registered CommenterMartin A

What is needed are temperature records that automatically adjust to fit the models. Making retrospective adjustments to temperature records is getting too obvious and time consuming.

Jan 13, 2016 at 10:16 PM | Unregistered Commentergolf charlie

gc
Better still, why not have temperatures that adjust themselves to fit the models?
Then everything would be right all the time. Or if that is too difficult just say that what the models say is what the temperature actually is. Then your local temperature could be set according to which model you happen to favour and the climate community could pitch for the contract for the US temperature or the EU temperature or the North Pole temperature. That would be a good one to have because you can play off WWF and the polar bears against the oil companies who want open water to drill in.
"Waddya mean, six metres of ice? The contract was for +2C to +4.8C for June to mid-August for five freakin' years! What kinda scam you tryin' to pull, pal?"

Jan 13, 2016 at 10:37 PM | Registered CommenterMike Jackson

Mike Jackson, I think you may have cracked it! It would be so much cheaper than all this pretence about 'climate science'.

Jan 14, 2016 at 12:35 AM | Unregistered Commentergolf charlie

...why that would be a good idea?
To get a better idea of how well the model worked? If you compare real outcome with projected outcome where the projection used incorrect forcings, it is not a like-with-like comparison. The point of the models is not really to project short term temperatures but instead to study the medium to long term where variations between assumed and actual forcings average out.

Jan 14, 2016 at 2:33 AM | Unregistered CommenterRaff

And there's the insanity that is faith in climate models to get right, on average, over the medium-long term what it can't get right in the short term. It is the wilful denial of the fact that climate models that don't accurately model all forcings or feedbacks, and that don't model many components of the climate system at all, can somehow transcend these fundamental flaws and deficiencies to become gnostic projections of future climate.

They say it is a mark of insanity to repeat an action over and over again, expecting the result to change. How many iterations does a GCM perform while computing its medium or even its long term projections, AFTER having failed so conclusively in the short term?

A GCM run is not a scientific experiment, it is merely a representation/discussion of the limitations of our understanding of the climate. There is no other purpose for a GCM than this. To treat a GCM run as if it were a scientific experiment, to treat its output as observational evidence, is a gross betrayal of science and the scientific method.

Jan 14, 2016 at 4:29 AM | Registered CommenterSimon Hopkinson