Global warming

[mhaze]

My point remains that despite niggling about the precise amount it is almost universally admitted that humans have significantly increase atmospheric CO2 levels. The CO2 increase should, due to many separate reasons, have an impact on the biosphere beyond climate, and that we should have some agreement that reducing atmospheric CO2 is a good idea, despite disagreement about the extent of climatic impact.

One minor point about your first comment here. There has been an assertion that over the 8000 years of the Holocene erase, approximately 40 ppm of CO2 was introduced into the atmosphere gradually - thus in the absence of human activity, the level in about year 1800 would have been 240 ppm.

Research has shown this to be false, and has also shown that during the Holocene there was considerable rapid fluctuations in CO2 level. This is not to say that we are not now contributing approximately 100 ppm, just that the climate is dynamic and the natural level of CO2 does vary, roughly say between 2xx and 3xx. There is no baseline, static level which is "right".

The second point you make can be vigorously debated on several levels as I am sure you are aware. You can't have an international agreement to control CO2 when the Asian countries producing the huge brown clouds are excluded and when those clouds are known to constitute more than 50% of the problem in those areas. That makes no sense. You are then attacking one thing, a possible non problem and ignoring a known problem.

We could wind up as successful at understanding and controlling "the CO2 problem" as we have with controlling those dangerous freons, proven scientifically to be responsible for the ozone hole.

 

[a_unique_person]

That is like saying that if I made careful notes on the last 1000 responses of a mechanical Las Vegas slot machine, and carefully programmed a computer to produce those responses, that the computer would then produce a correct answer for the 1001th try.

Obviously, that is wrong. The model must accurately represent the physical realities and their interactions. The very concept of parameterizing models to match history is flawed.

Valid Exceptions: Such models may well be of use in educational environments or in looking at limited scenarios and trying to understand them.

Or when there is no alternative. I would have thought that this situation warrented that.

Your analogy also fails. The slot machine produces random results, the climate incorporates physical reactions within restraints. As Hansen has publicly demonstrated, his model's prediction from ten years ago was correct, even for a model that is not as advanced as the models they are using today.

There is a lot of work put into validating the models.

http://ipcc-wg1.ucar.edu/wg1/Report/AR4WG1_Print_Ch08.pdf

Have you read it yet?

 

[mhaze]

Or when there is no alternative. I would have thought that this situation warrented that.

Your analogy also fails. The slot machine produces random results, the climate incorporates physical reactions within restraints. As Hansen has publicly demonstrated, his model's prediction from ten years ago was correct, even for a model that is not as advanced as the models they are using today.

There is a lot of work put into validating the models.

http://ipcc-wg1.ucar.edu/wg1/Report/AR4WG1_Print_Ch08.pdf

Have you read it yet?

Where exactly do you find the word "validate" used in connection with "models"?

 

[a_unique_person]

Where exactly do you find the word "validate" used in connection with "models"?

This chapter assesses the capacity of the global climate
models used elsewhere in this report for projecting future
climate change. Confidence in model estimates of future climate
evolution has been enhanced via a range of advances since the
IPCC Third Assessment Report (TAR).

 

[Walrus32]

I still don't see "validate".

 

[PogoPedant]

I still don't see "validate".

Indeed. And as we all know, if a paragraph does not use the word 'validate', then it cannot possibly be about any form of 'validation'.

Just like the last sentence cannot possibly be sarcastic.

 

[Walrus32]

Not sarcastic at all. A question was asked, and it wasn't answered.

Vincent Gray, "Spinning the Climate", 27 September 2007:


"...The first draft of the 1995 Report had a Chapter 5 "Validation of
Climate Models" as in the First Report. I pointed out that it was wrong since no climate model has ever
been "validated", and they did not even try to do so. They thereupon changed the word "Validation" to
"Evaluation" no less that fifty times.

Perhaps I should explain what is meant by "validation". It is a term used by computer engineers to
describe the rigorous testing process that is necessary before a computer-based model can be put to use. It
must include successful prediction over the entire range of circumstances for which it is required. Without
this process it is impossible to find out whether the model is suitable for use or what levels of accuracy
can be expected from it."

 

[CapelDodger]

Indeed. And as we all know, if a paragraph does not use the word 'validate', then it cannot possibly be about any form of 'validation'.

Just like the last sentence cannot possibly be sarcastic.

Ironic, not sarcastic. Sarcasm is what the lower orders do .

From his response, I don't think walrus32 does irony. Nor comprehension in any serious sense.

 

[a_unique_person]

Not sarcastic at all. A question was asked, and it wasn't answered.

Vincent Gray, "Spinning the Climate", 27 September 2007:


"...The first draft of the 1995 Report had a Chapter 5 "Validation of
Climate Models" as in the First Report. I pointed out that it was wrong since no climate model has ever been "validated", and they did not even try to do so. They thereupon changed the word "Validation" to "Evaluation" no less that fifty times.

Perhaps I should explain what is meant by "validation". It is a term used by computer engineers to describe the rigorous testing process that is necessary before a computer-based model can be put to use. It must include successful prediction over the entire range of circumstances for which it is required. Without this process it is impossible to find out whether the model is suitable for use or what levels of accuracy can be expected from it."

Vincent Gray is one more of the lunatic fringe who has no idea what the current state of the art is in research, since he retired 20 years ago and has not been actively researching and publishing in that field. This self proclaimed 'expert reviewer' is one of the more forgettable particpants in the process.

An example of his insight.

"Delete "change". The word has an unfortunate commotation, as it is defined legally by the Framework Convention on Climate Change as restricted to" human activity that alters the composition of the global atmosphere". The IPCC tries to alter this definition by a footnote to the "Summary for Policymakers" (page 3) but this leads to confusion as the
public may not notice this and assume that you are referring only to the redtricted definition. You should therefore avoid using the term"climate change" altogether to avoid this confusion
[VINCENT GRAY (Reviewer’s comment ID #: 88-28)]"

Ad nauseum.

Finally, have you read the chapter? It's quite comprehensive, and discusses the limitations of the models and just what they can be expected to produce under the constraints they operate under.

 

[CapelDodger]

One minor point about your first comment here. There has been an assertion that over the 8000 years of the Holocene erase, approximately 40 ppm of CO2 was introduced into the atmosphere gradually - thus in the absence of human activity, the level in about year 1800 would have been 240 ppm.

Research has shown this to be false ...

That is not the case.

... and has also shown that during the Holocene there was considerable rapid fluctuations in CO2 level.

Nor is that the case.

This is not to say that we are not now contributing approximately 100 ppm ...

Best not, really, since it's based on direct measurements over the recent past.

... just that the climate is dynamic and the natural level of CO2 does vary, roughly say between 2xx and 3xx.

Such variations are claptrap. We know that we're pumping billions of tons of CO2 into the atmosphere every year and the year-on-year increase of CO2-load varies between 0.5 and 2.5 ppm pa. You're claiming that in other periods of the Holocene CO2-load varied between 120 and 360ppm over short periods. Or possibly between 240 and 720ppm; your post isn't terribly precise.

This is not credible.

There is no baseline, static level which is "right".

There's normal. Forget right and wrong. "Normal" does not include huge CO2-fluxes for no apparent reason that just happened to stop when precise measurement became available. Which was about the time that fossil-fuel consumption got into its stride - a very apparent reason for a large CO2-flux.

The second point you make can be vigorously debated on several levels as I am sure you are aware. You can't have an international agreement to control CO2 when the Asian countries producing the huge brown clouds are excluded and when those clouds are known to constitute more than 50% of the problem in those areas. That makes no sense. You are then attacking one thing, a possible non problem and ignoring a known problem.

We could wind up as successful at understanding and controlling "the CO2 problem" as we have with controlling those dangerous freons, proven scientifically to be responsible for the ozone hole.

"Asian Brown Clouds". At least we've moved on from the peril being yellow.

When it all goes to <Rule 8>, of course, we'll be able to blame the Chinese. And justify our own inaction in the process.

The Chinese are to blame for the ozone hole, after all. (Irony)

 

[Schneibster]

Well, I'm back, and nothing has changed except the climate.

 

[Schneibster]

One minor point about your first comment here. There has been an assertion that over the 8000 years of the Holocene erase, approximately 40 ppm of CO2 was introduced into the atmosphere gradually - thus in the absence of human activity, the level in about year 1800 would have been 240 ppm.

Research has shown this to be false, and has also shown that during the Holocene there was considerable rapid fluctuations in CO2 level.

What research, performed by whom, and published in what peer-reviewed journal of geophysical or climate research? Linky-poo, perhaps? CD seems to have tired of looking over articles with incestuous citations; I am not tired. I haven't been here in a while, and it looks like there's plenty of work for the weed-whacker.

 

[CapelDodger]

Well, I'm back, and nothing has changed except the climate.

Hey dude, long time !

The weather has changed around these parts, of course, but not very much considering that it's autumn now. Piers Corbyn's predictions of a stormy August and no late summer have been soooooooo frickin' busted that ... it'll make no difference at all to those who believe in him .

 

[Schneibster]

Good to see you too, CD.

 

[mhaze]

Not sarcastic at all. A question was asked, and it wasn't answered.

Vincent Gray, "Spinning the Climate", 27 September 2007:


"...The first draft of the 1995 Report had a Chapter 5 "Validation of
Climate Models" as in the First Report. I pointed out that it was wrong since no climate model has ever
been "validated", and they did not even try to do so. They thereupon changed the word "Validation" to
"Evaluation" no less that fifty times.

Perhaps I should explain what is meant by "validation". It is a term used by computer engineers to
describe the rigorous testing process that is necessary before a computer-based model can be put to use. It
must include successful prediction over the entire range of circumstances for which it is required. Without
this process it is impossible to find out whether the model is suitable for use or what levels of accuracy
can be expected from it."

Not in defense of AUP's wrongness, but it was probably the case that he did not really understand the meaning of the phrase as it is used in computer science. Hence my question - but it looks like we've settled it, no the models are not validated. Already knew that, but just wanted to bring it out for discussion.

Shall we go back to Hansen's discussion of the models and note from his words their status in comparison to "validation"? Perhaps it is simpler to just leave this minor topic as closed.

 

[CapelDodger]

Not in defense of AUP's wrongness, but it was probably the case that he did not really understand the meaning of the phrase as it is used in computer science. Hence my question - but it looks like we've settled it, no the models are not validated. Already knew that, but just wanted to bring it out for discussion.

Shall we go back to Hansen's discussion of the models and note from his words their status in comparison to "validation"? Perhaps it is simpler to just leave this minor topic as closed.

I imagine the topic would be opened again fast enough if the models ever become invalidated, but that's a vacuous hypothesis (pertaining to circumstances that won't occur). The Hansen et al 1988 model hasn't run out of projection yet, and there are others which started up later. Twenty years of not being invalidated is a robust track-record. No conceivable track-record will be good enough for some, but that hardly counts as invalidation.

The models, and they're continued survival through the next twenty years of potential invalidation, will continue to be trumpeted by the likes of me. I know when I've picked the right horse, and I know when losers are weaseling about it not being a proper horse-race.

The models, or Singer's Salvation Cycle. My money's on the models.

There's Solar Cycle Twenty-Something in the race as well, but that'll fade early. The 60-80 year Dick-Cycle (no offence to Dr Dick) of Arctic ice-extent will barely leave the stalls. The Iris Theory scratched before the race even began. (Bluetongue virus, apparently. Any old excuse, if you ask me.)

 

[Schneibster]

Cycle 24 /pedant.

 

[mhaze]

The models, or Singer's Salvation Cycle. My money's on the models.
There's Solar Cycle Twenty-Something in the race as well, but that'll fade early. The 60-80 year Dick-Cycle (no offence to Dr Dick) of Arctic ice-extent will barely leave the stalls.

A truly amazing predictive capability.

With it, what need have you of unvalidated models?

 

[Schneibster]

There's only one way to validate a model, and that's run it and see if it agrees with the real world. So far, so good. What's your problem, hazy, not patient enough to see it through?

 

[CapelDodger]

Perhaps I should explain what is meant by "validation". It is a term used by computer engineers to
describe the rigorous testing process that is necessary before a computer-based model can be put to use. It must include successful prediction over the entire range of circumstances for which it is required.

Including the successful prediction "at this point everything goes <Rule 8>-shaped". (Alliteration. Opposite of ship-shaped .)

Without this process it is impossible to find out whether the model is suitable for use or what levels of accuracy can be expected from it.

What kind of models are you acquainted with?

I haven't worked with models much, I'm more a systems guy; my idea of validation, in computer terms, is 100% accuracy given any input. Even if the accurate output is "Sorry, system cannot cope, please try later". I leave such matters as traffic-modelling to those that do it, and use their results to specify a safe "Sorry, we're full" load while there's still some leeway. Then price the job on that basis.

Often one's faced with demands for 100% avalability, which means chasing down that long tail of the graph at silly expense. Sometimes silly expense is available - not so much these days, but time was ...

Of course when it came to the InterNet the established traffic-models turned out to be useless. I warned people about that, but did they listen? Did they heck as like .