In your face evidence of climate change

[ludo]

Guam said:

Agreed the localised chaotic effects due to micro climate impacts will average out across the data. However the fundamental flaws remain in both modelling excercises do they not

So you would agree that predicting the climate should be easier than predicting the weather (as the chaotic element is not relevant)?

The models are approximations, there is a good deal of uncertainty over the details, but I have not read anything in the litterature pointing out fundamental flaws. Perhaps you could give some examples?

Edited by ludo on Monday 4th August 12:41

[mondeoman]

ludo said:

Guam said:

Agreed the localised chaotic effects due to micro climate impacts will average out across the data. However the fundamental flaws remain in both modelling excercises do they not

So you would agree that predicting the climate should be easier than predicting the weather (as the chaotic element is not relevant)?

The models are approximations, there is a good deal of uncertainty over the details, but I have not read anything in the litterature pointing out fundamental flaws. Perhaps you could give some examples?

Edited by ludo on Monday 4th August 12:41

I think TB has already done that - the models ignore solar influence, cosmic ray influence, they have insufficient degrees of freedom to cover heat loss / heat gain etc in the atmosphere, they have too much weighting for CO2, they reduce the weighting of H2O, tectonics....

[nigelfr]

Blib said:

I freely admit that I'm not the cleverest bloke on here by a long chalk. Many of the posts and most of the graphs that TB, Ludo et al. place on the various threads about MMGW go flying over my head.

However, I have gleaned, rightly or wrongly, that the pro MMGW camp base their conclusions on powerful computer generated predictions based on current understanding of climate.

My question is this.

As far as I know, no one has much more than the most basic understanding of how our planet's complete climate operates. So, how can anyone extrapolate anything from the figures pumped into computers primed with an incomplete model with confidence?

Apologies if I've missed something fundamental here.

Thanks.

Let me try an analogy, with all the risks and perils that that entails: You hear that every night a drunk comes home from the pub. So you decide to make a model to predict the time he gets home. The first night you track his every weave and stagger and he gets home at a certain time. You put the data in your model and repeat. After a few weeks you can predict pretty confidently when he will get home, even though you don't know the exact path he takes. However, you can never say on a particular day the exact time of his arrivel home, because something unexpected may happen.

And then something unexpected happens e.g. they dig up the road, so he has to take another way home. He arrives home later than predicted. But, this is the important bit, if you're a competant modeller, you modify your model to take this into account (new observational data).

I'm sure you can pick holes in this analogy, but I picked it because the route that the drunk takes home is the weather, while his time of arrival is the
climate. On a second by second basis the model can't predict where he puts his feet (and it's not intended to) but it can eventually deliver a reliable ETA.

That's what's happening with the climate models which have been steadily refined for over 3 decades: they're modified to account for new data.

Now this is where the sceptics jump in and talk about fiddle factors and GIGO, but in proper, real world science, you have to be able to justify any changes made to the model. Another thing to remember is that there is not one model but many independant models that may differ in approach and detail, but mostly agree that the drunk gets home.

What is important is that the data be as complete as possible; which is why, for example, 3000 ocean floats are being launched to improve the quality of ocean data.

If you want a better answer to your concerns, why not ask someone at the Met Office? Your time is probably better spent doing that, rather than following the cut and thrust here. I contacted someone who explained very clearly the current state of play. I hesitate to give their name here, because I don't want to be responsible for them being spammed.

[Apache]

Hope it wasn't Michael Fish

[ludo]

Guam said:

ludo said:

ludo said:

Guam said:

Agreed the localised chaotic effects due to micro climate impacts will average out across the data. However the fundamental flaws remain in both modelling excercises do they not

So you would agree that predicting the climate should be easier than predicting the weather (as the chaotic element is not relevant)?

The models are approximations, there is a good deal of uncertainty over the details, but I have not read anything in the litterature pointing out fundamental flaws. Perhaps you could give some examples?

Good lord man I have repeatedly you cannot forecast with any degree of certainty (extrapolate if you will) where ALL the variables are not fully understood, nor their impacts fully understood, Solar Variation for example, EVEN if you average out the minor background noise (localised chaotic events) you cannot as yet forecast accurately what toonage of CO2 will result in what degree of warming or its abscense and the converse. Those are just two, when a major event (vulcanism) occurs what degree of theoretical change in the climate occurs and how sustained will it be, and to what degree (up or down) will it mask the trend?

The same inherent rational flaws as in Weather forecasting (just the variables may differ).

Comes back to Spurious correlation in raw statistical analysis, just because you seem to have a relationship at first run through the data doesnt mean there actually IS one!


Cheers

Sorry, that is just repeating your assertion. Can you provide e.g. a journal paper or even a reference to one that identifies a fundamental flaw in GCMs. I know your opinion, because I read it the first time. What I am asking for is some evidence to back up your opinion.

BTW, weather noise is larger in magnitude than climate trends, so "minor background noise (localised chaotic events)" sounds rather incongruous in this discussion. Also there have been studies where the observed effect of Pinatubo was compared with the GCM model prediction, and the results showed the models had useful skill (although of course not perfect).

[ludo]

mondeoman said:

ludo said:

Guam said:

Agreed the localised chaotic effects due to micro climate impacts will average out across the data. However the fundamental flaws remain in both modelling excercises do they not

So you would agree that predicting the climate should be easier than predicting the weather (as the chaotic element is not relevant)?

The models are approximations, there is a good deal of uncertainty over the details, but I have not read anything in the litterature pointing out fundamental flaws. Perhaps you could give some examples?

Edited by ludo on Monday 4th August 12:41

I think TB has already done that - the models ignore solar influence, cosmic ray influence, they have insufficient degrees of freedom to cover heat loss / heat gain etc in the atmosphere, they have too much weighting for CO2, they reduce the weighting of H2O, tectonics....

The models certainly do not ignore solar influence. They do ignore cosmic ray influence for the good reason that it has not been demonstrated that there is an effect, all that has been demonstrated is that there is a correllation that unfortunately breaks down if you look at recent data, and no unequivocal signal of cosmic rays in the cloud cover data. Too much weighting for CO2, how do they know there is too much weighting, that is mere assertion/opinion as it can't be directly measured. H20 similarly, we can't directly measure the "correct" amount. Tectonics, I take it you mean volcanos? The effects of volcanoes are incorporated in the forcings of GCMs, of course you can't predict when volcanos will ocurr, but you can attempt to predict what will happen if they do, and that is what the climatologists do by having predictions for a range of scenarios, some of which involve volcanic events.

So we have two errors, solar forcing and volcanoes are incorporated into GCMs, one red-herring cosmic ray theory is at too early a stage to be included and anyway the evidence is pretty flimsy (at this stage), for the weighting of CO2 and H2O, these are not fundamental flaws, they are issues where the parameters are uncertain, but the basic mechanism are understood.

[turbobloke]

ludo said:

I am not telling anyone how to behave, I am discussing the science.

Hardly, that website you linked to a page or so back was loaded with tendentious crap mixed in with the occasional blatant untruth, it hoped the reader would be propagandised already so not spurprising you go there.

hairykrishna said:

There's no point denying it. If you disagree with the all knowing Turbobloke and dare to suggest that you agree with the vast majority of people who actually study the climate full time then you obviously must be some kind of lentil munching hippy.

Vast majority eh, ask nigelfr about the Schulte litsearch he should just about have got on top of that by now.

All-knowing, of course not, knowing more than you and PH's True Believers, definitely, only an uninformed lightweight or mandacious zealot would regard knowledge and understanding as a bad thing and attempt to vilify it.

Others have studied climate science - you mean you haven't? Oh dear.

ludo said:

The models are approximations, there is a good deal of uncertainty over the details, but I have not read anything in the litterature pointing out fundamental flaws. Perhaps you could give some examples?

Why does it have to be in the literature, by which we can assume you mean the peer-reviewed literature? Must something get past estabkishment gatekeepers for you to give it credit (not that your judgement is automatically valid anyway)? The IPCC state openly they don't reject non-perr-reviewed science.(Procedures for the Preparation, Review, Accepetance, Adoption, Approval and Publication of IPCC Reports) they also happen to say that they will alter the science to fit the politicise Summary for Policymakers, which is a fascinating way to do science

As to the literature, ludo appears not to be up-to-date with bedtime reading.

ludo said:

Can you provide e.g. a journal paper or even a reference to one that identifies a fundamental flaw in GCMs.

Koutsoyiannis, D., A. Efstratiadis, N. Mamassis, and A. Christofides, 2008: On the credibility of climate predictions (Hydrological Sciences Journal, 53 (4), 671-684.)

Abstract: Geographically distributed predictions of future climate, obtained through climate models, are widely used in hydrology and many other disciplines, typically without assessing their reliability. Here we compare the output of various models to temperature and precipitation observations from eight stations with long (over 100 years) records from around the globe. The results show that models perform poorly, even at a climatic (30-year) scale. Thus local model projections cannot be credible, whereas a common argument that models can perform better at larger spatial scales is unsupported.

Extract: Remarkably, during the observation period, the 30-year temperature ... decreased by about 1.5°C, while all models produced an increase of about 0.5°C

Conclusion: At the annual and the climatic (30-year) scales, GCM interpolated series are irrelevant to reality. GCMs do not reproduce natural over-year fluctuations and, generally, underestimate the variance and the Hurst coefficient of the observed series. Even worse, when the GCM time series imply a Hurst coefficient greater than 0.5, this results from a monotonic trend, whereas in historical data the high values of the Hurst coefficient are a result of large-scale over-year fluctuations (i.e. successions of upward and downward ‘trends’. The huge negative values of coefficients of efficiency show that model predictions are much poorer than an elementary prediction based on the time average. This makes future climate projections at the examined locations not credible. Whether or not this conclusion extends to other locations requires expansion of the study, which we have planned. However, the poor GCM performance in all eight locations examined in this study allows little hope, if any. An argument that the poor performance applies merely to the point basis of our comparison, whereas aggregation at large spatial scales would show that GCM outputs are credible, is an unproved conjecture and, in our opinion, a false one.


The failures of computer climate models are many and well-known: data quality flaws (McKitrick and Michaels demolition of gridded surface data integrity), inadequate knowledge and understanding in detail of climate forcing mechanisms to allow credible midelling (see earlier post listing the approx half-dozen forcings the IPCC refer to but know little about in detail), inadequate spatial resolution, inadequate computing power to handle the modelling process - even more so if the mechanism detail was better known and the spatial resolution was improved. Vertical heat distribution is too rigidly paramaterised, negative feedbacks are omitted, etc etc.

To run a realistic climate model over all relevant spatial scales and obtain a 40 year projection would require 10 to the power 34 years of supercomputer time, this is 10 to the power 24 times longer than the current estimate of the age of the universe (Dr W Soon, Harvard Smithsonian Institute).

Climate models: chewing gum for the brains of True Believers and other uninformed gullible fools.

[ludo]

Guam said:

ludo said:

Guam said:

ludo said:

ludo said:

Guam said:

Agreed the localised chaotic effects due to micro climate impacts will average out across the data. However the fundamental flaws remain in both modelling excercises do they not

So you would agree that predicting the climate should be easier than predicting the weather (as the chaotic element is not relevant)?

The models are approximations, there is a good deal of uncertainty over the details, but I have not read anything in the litterature pointing out fundamental flaws. Perhaps you could give some examples?

Good lord man I have repeatedly you cannot forecast with any degree of certainty (extrapolate if you will) where ALL the variables are not fully understood, nor their impacts fully understood, Solar Variation for example, EVEN if you average out the minor background noise (localised chaotic events) you cannot as yet forecast accurately what toonage of CO2 will result in what degree of warming or its abscense and the converse. Those are just two, when a major event (vulcanism) occurs what degree of theoretical change in the climate occurs and how sustained will it be, and to what degree (up or down) will it mask the trend?

The same inherent rational flaws as in Weather forecasting (just the variables may differ).

Comes back to Spurious correlation in raw statistical analysis, just because you seem to have a relationship at first run through the data doesnt mean there actually IS one!


Cheers

Sorry, that is just repeating your assertion. Can you provide e.g. a journal paper or even a reference to one that identifies a fundamental flaw in GCMs. I know your opinion, because I read it the first time. What I am asking for is some evidence to back up your opinion.

BTW, weather noise is larger in magnitude than climate trends, so "minor background noise (localised chaotic events)" sounds rather incongruous in this discussion. Also there have been studies where the observed effect of Pinatubo was compared with the GCM model prediction, and the results showed the models had useful skill (although of course not perfect).

To the contrary YOU are the one putting the concept (along with others) it is YOUR responsibility to show where the criticisms are baseless,

Sorry that is not how science works. You can't prove anything using observations, only corroborate or disprove (see e.g. work of Popper). However, unless you can provide evidence of a specific fundamental flaw, the criticism is too vague to expect a refutation.

[AstonZagato]

ludo said:

Guam said:

ludo said:

Guam said:

ludo said:

ludo said:

Guam said:

Agreed the localised chaotic effects due to micro climate impacts will average out across the data. However the fundamental flaws remain in both modelling excercises do they not

So you would agree that predicting the climate should be easier than predicting the weather (as the chaotic element is not relevant)?

The models are approximations, there is a good deal of uncertainty over the details, but I have not read anything in the litterature pointing out fundamental flaws. Perhaps you could give some examples?

Good lord man I have repeatedly you cannot forecast with any degree of certainty (extrapolate if you will) where ALL the variables are not fully understood, nor their impacts fully understood, Solar Variation for example, EVEN if you average out the minor background noise (localised chaotic events) you cannot as yet forecast accurately what toonage of CO2 will result in what degree of warming or its abscense and the converse. Those are just two, when a major event (vulcanism) occurs what degree of theoretical change in the climate occurs and how sustained will it be, and to what degree (up or down) will it mask the trend?

The same inherent rational flaws as in Weather forecasting (just the variables may differ).

Comes back to Spurious correlation in raw statistical analysis, just because you seem to have a relationship at first run through the data doesnt mean there actually IS one!


Cheers

Sorry, that is just repeating your assertion. Can you provide e.g. a journal paper or even a reference to one that identifies a fundamental flaw in GCMs. I know your opinion, because I read it the first time. What I am asking for is some evidence to back up your opinion.

BTW, weather noise is larger in magnitude than climate trends, so "minor background noise (localised chaotic events)" sounds rather incongruous in this discussion. Also there have been studies where the observed effect of Pinatubo was compared with the GCM model prediction, and the results showed the models had useful skill (although of course not perfect).

To the contrary YOU are the one putting the concept (along with others) it is YOUR responsibility to show where the criticisms are baseless,

Sorry that is not how science works. You can't prove anything using observations, only corroborate or disprove (see e.g. work of Popper). However, unless you can provide evidence of a specific fundamental flaw, the criticism is too vague to expect a refutation.

You seem to be saying that we are not able to predict weather because it is too complex but that the incomplete models on MMGW work fine because they haven't been disproved yet.

[turbobloke]

Something for ludo nigelfr mattikake and hairykrishna to chew on from our politically ousted but scientifically respected climatologist Pat Michaels:


The basic rule of science is that hypotheses must be verified by testing their predictions against observed data.9 Hypotheses that cannot be tested can be useful, but they are not science. Hypotheses that are tested and fail must be modified and retested, or simply rejected. Science that relies upon hypotheses that have failed a comparison with reality is “junk science.” Acomputerized climate model, however sophisticated, is indeed nothing more than a hypothesis until it is verified by testing against reality. If it fails that test, and it continues to be used for a “scientific” assessment, that assessment then falls into the “junk science” category.

Both CGCM1 and HadCM2 make predictions of U.S. climate change based upon human alterations of the atmosphere. Those alterations have been going on for well over 100 years. Do the changes those models “predicted” resemble what actually occurred in the last century? The answer is clearly no. I compared observed U.S. annual temperature departures from the twentieth-century average with those generated by both the CGCM1and HadCM2 models. In both cases I used ten-year running averages to minimize interannual noise. This is a simple and common test. The modeled U.S. average temperature for 1991–2000is compared to the observed value. Then the comparison period is backed up one year, to 1990–99, and so on. This smooths out the effect of single years that are unusually warm or cold, such as occurs in a strong El Nino year (such as 1998) or after a large volcanic eruption (1992). I then examined the differences between the modeled and observed values for both the CGCM and HadCM2, versus the result that would obtain if I simply used the average temperature for the twentieth century to predict the observed values from year to year.

In fact, both models did worse than that base case. In other words, both climate models used in the USNA were worse than no model at all.

"worse than no model at all"

[ludo]

AstonZagato said:

You seem to be saying that we are not able to predict weather because it is too complex but that the incomplete models on MMGW work fine because they haven't been disproved yet.

No, climate prediction using GCMs is not based on predicting the weather, so the fact that "we can't even predict the weather" is entirely irrellevant.

The nature of science response was a different issue. You can't prove anything in observational science, so saying that "it is up to you to prove the models to be correct" (or words to that effect) ignores what is and what isn't possible in science. That is why the way science works is to propose new theories and to disprove old ones, the good theories survive, the bad ones are proven wrong. That is why it is up to the sceptics to provide evidence to back up their assertions of "Fundamental flaws".

BTW: GCMs are known not to be able to predict station level data, they only work on synoptic scales, which is why statistical downscaling is used for impact studies. See e.g.

http://www.realclimate.org/index.php/archives/2007...

[hairykrishna]

A computer model is not the same thing as a hypothesis; I don't care who said it. Computer models are built using a number of more basic hypotheses. A computer model that fails to exactly match reality is still useful as it can be used to develop more accurate computer models. CGCM1 to CGCM 2 to CGCM3 for instance.

I don't know enough about the computer models used by the IPCC to sensibly debate their reliability. I thought they were pretty good but I could be wrong.

Edited by hairykrishna on Monday 4th August 13:41

[ludo]

hairykrishna said:

A computer model is not the same thing as a hypothesis; I don't care who said it. Computer models are built using a number of more basic hypotheses. A computer model that fails to exactly match reality is still useful as it can be used to develop more accurate computer models. CGCM1 to CGCM 2 to CGCM3 for instance.

I don't know enough about the computer models used by the IPCC to sensibly debate their reliability. I thought they were pretty good but I could be wrong.

Edited by hairykrishna on Monday 4th August 13:41

GCMs tell you the (approximate) consequences of a set of assumptions about the underlying physics of the oceans/atmosphere and future forcing scenarios, nothing more. The climatologists know that, the media doesn't. Another point that is often missed is that the GCMs can only predict the forced component of the climate, they are unable (almost by definition) to predict the chaotic weather variation (which includes things like ENSO). Their predictions are generally accurate, within the known uncertainty of the predictions (i.e. the error bars - which are generally pretty broad).

[turbobloke]

Guam said:

TB Has just done so above

He has indeed and he's back for more...

Open communication from an IPCC reviewer, well wroth the effort to go through it:


My greatest success as an "expert reviewer" to the Intergovernmental Panel of Climate Change (IPCC) Working Group I (WGI) Science Reports was with the first draft of the 1995 Report.

There was a Chapter entitled "Validation of Climate Models"

I commented that this was incorrect. No Climate Model has ever been "validated" in the sense understood by computer engineers, and the Chapter included no discussion on how it should be done, let alone any of the necessary procedure, on any model.

They ageed with me. They changed the words "Validation", or "Validate" to "Evaluation" or "evaluate" no less than fifty times, throughout the Chapter. They have done so ever since. The word "validate" or "validation" does not appear anywhere in their Reports, and, notably, in the recently issued "Climate Change 2007: The Physical Science Basis: Summary for Policymakers"

One of the major objects of science is to simulate observable phenomena with a mathematical representation which can not only provide an explanation for the phenomena, buit also make it possible to predict future behaviour.

This task has a long history. "Stonehenge Decoded" by Gerald S Hawkins shows how before 1600 BC it was possible to build a system which would enable prediction of the movements of the sun and moon.

Ptolemy in the second century AD published the "Almagest" which predicted the movements of the planets with a system of "epicycles". Newton and Galileo replaced this with a better, simplified theory, and Eistein with a refined version. Nobody would even have heard of these people if there was not abundant positive evidence that their predictions actually work. Without them, we could never have sent rockets to the moon.

Let me spell out what is needed for "validation", the procedure without which no mathematical representation, or computer model, could possibly be capable of future prediction.

First, the model must be capable of simulation of past behaviour to a satisfactory level of accuracy. Computer models of the climate have usually failed to do this. Indeed, their only attempt has been on the so-called "global surface temperature anomaly record":which I showed, in my last Newsletter, to be subject to huge, unknown biases and inaccuracy because it is based on unrepresenrative and statistically flawed data. The claimed successful simulation of this flawed record could only be made by leaving out both consideration of these inaccuracies, and also one of the main "natural" contributors to the temperature record, the recently more frequent sudden warming peaks caused by the El Niño ocean oscillation behaviour.

The models are unable to simulate almost everything else.

They cannot explain why there has been no "warming" for the past eight years, even when measured by the unsatisfactory "surface record".

They cannot explain why there has been no warming at all on the Arctic continent.

They cannot explain why methane concentrations in the atmosphere are falling instead of rising. They even devote learned papers trying to find out why this behaviour is "anomalous".

A recent study by Douglass et al 2006 Geophysical Reserarch Letters 33 L19711 on the climatic effects of the eruption of Mount Pinatubo showed that it could only be expalined by a model iwith very low figures for "climate sensitivity" the basic parameter of the models.

The models therefore fail at the first requirement for "validation". They cannot reliably simulate past climate behaviour.

Suppose for one moment that Newton and Einstein had never lived and they were launching a rocket to the moon from Cape Kennedy. They ask the people who prepared the computer programme to guide the rocket "How reliable is it" Imagine if the reply was ' Our boys think it it is very likey to hit the moon, but we have no idea where"

We are taking all sort of drastic measures to damage our future energy policies based entirely on just such an "opinion" of partisan "experts".

The second important necessity for validation is successful predictioin of future behaviour under a variety of conditions to a satisfactory and measurable level of accuracy.

There has not been even a single attempt to meet this requirement for any computer model of the climate. They do not even discuss how it might be done.

The models are therefore worthless and should be discarded until validation has actually happened.

But why is it that so many people, not only Prime Ministers, US Presidential Candidates, Senior Economists, but also senior scientists and even winners of Nobel Prizes, seem to be convinced that those providing models have even MADE predictions, let alone provide a measures of their reliability.

It is even claimed that a large majority of scientists involved in climate research accept these false assumptions, and this claim might even be true.

Since the IPCC have accepted that no model has ever been validated, they have also accepted that they are unable to make predictions and they have never done so after the First Report (1990) The word "prediction" never appears anywhere in the recent IPCC Reports. The only thing the models can do is to provide "projections". This word implies that the figure obtained is purely a result of assuming that the data, parameters and equations in the model represent reality: but there is no evidence that they actually do.

How have they succeeded in fooling the world?

The answer is, that they have devised a whole series of tricky procedures designed to cover up the truth, and give the impression to casual readers and all but the intensive critic (which I claim to be) that they really have overcome the absence of validation, and provided definite figures which some people can pretend to be "predictions", and even provide what seems at first sight to be some measure of accuracy.

Their main tool is to pretend that they can replace scientific evidence with the opinions of "experts". The "experts" in this case are people who are mainly financed by Governments who promote the certainty of the greenhouse idea. I would not wish for a moment to suggest that these scientists could possibly be other than impartial, or that they could be influenced by pressure from their employers, even when they represent them at international conferences. But, all the same, most of them know that there might be undesi\rable consequences if any of them failed to endorse the value of models.

At this point let me say that the idea that scientific opinions can be influenced by employers is not just a myth. In my long scientifc career such pressure was applied to myself on several occasions, and on one of them, I was dismissed.when I resisted,

Because of the opinions I express in this newsletter I am sometimes accused of being influenced by mythical employers, For example Professor Neil Curtis, formerly from Victoria University of Wellington, and currently Patron of the New Zealand Association of Scientists, has accused me of being in the pay of oil companies. Vanessa Atkins, Greenpeace representative in New Zealand, says I am paid by Exxon, and the same accusation has been made recently on the "Real Climate" website.

I have never been employed by any oil company, or received finance from one. Campaigning for truth in climate science is not exactly financially rewarding. I might tell you about two of my recent contributions.

Last Year, I was invited to the Beijing Climate Center as a Visiting Scholar. I was welcomed by the Director General and I gave three well attended lectures. They paid my fare and accommodation Yet the Senior scientist there is Co-Chair if Working Group I resposible for the 2007 IPCC Report about to be issued.

The people in Beijing appear to be willing to listen to different points of view on climate change, but in New Zealand I could never be invited to address a meeting sponsored by NIWA, and Victoria University of Wellington now seems out of bounds. The Wellington Branch of the Royal Society .replies with an excuse. But I must admit I have recently address the Ohariu Branch of the Univesity of the Third Age, two Wellington Rotary Clubs, and a "Freedom Summit" Conference

Another recent source of income has been two book reviews in the Christchurch "Press" In the first I came down heavily on "The Weathermakers" by Tim Flannery, currently "Australian of the Year". He is a biologist with no knowledge of physics, since he thinks the greenhous efffect is caused by the heating of trace gases by the sun, instead of the more orthodox theory that they are heated from radiation by the earth. His only credit is that he demolishes the "hydrogen economy" because it ends up emitting more greenhouse gases than before. But that seems also to be true of "biofuels" so perhaps it does not matter

But. I digress. The "opinions" of the IPCC "experts" are graded in levels of "likeliness", and they are given spurious "probability" levels which bear no relationship to probability that most scientists recognise.

Besides the completely uncertain nature of the "projections" of climate models it is imposasible to provide a measure of their possible acuracy or reliabity. If there were such measures it would be possible to grade the models in order of success. Since this cannot be done all the models are given equal credence. They even hold occasional meetings to try and avoid too much difference between models, since too wide a "projection" might destroy the impression of plausibility.

It means also that the IPCC never has the embarassing task of telling any model maker that his model has a "failure" mark, since they have no way of marking them.The result is that the models are a free for all and the more extreme the "projections" are, the better some polticians or activists like them, and the better the chances for future funds. Many of the models can give low or even negative figures if you fit the right parameters, but the fate of those who have tried this is best not revealed.

Having managed to provided a half-way plausible "estimate" for a a model output, they then had to find a procedure to provide accuracy estimates of this figure, beyond that of the levels of "likelihood"

They do this by combining a restrictive choice of models with a restrictive choice of "emissions scenarios". Both of these are chosen so as to give a "range" of outcomes acceptable to the Governments who pay them. They carefully avoid outcomes that are too high ar too low. The "range" is then presented as if it were a scientific nmeasure of the accuracy of the combined model/scenario package.

The "scenarios" themselves are supposed to provide a range of plausible assumptions of what could happen to climate in the next hundred years.They do not have the confidence, however to carry out any checks to find out whether the assumptions are confirmed by what actually happens. This means they have no way of grading their plausibility. The scenarios are thus regarded as equally plausible, but like the animals in Orwells's "Aniumal Farm" some scenarios are more equally plausible than others.

My paper Gray, V R 1998 "The IPCC future projections: are they plausible" Climate Research 10 155-162 showed that the earlier scenarios were not plausible, and Chapter 7 of my book "The Greenhouse Delusion: A Critique of 'Clmate Change 2001" showed that the 2001 scenarios are also not plausible. They could not even get right the figures for the year 2000. So they cannot even predict the past.

Several senior economists have criticised the economic forecasting methods used by the IPCC, but with little response. One of these David Henderson, former Head of the Economics and Statistics Department of the OECD is addressing two meetings in Wellington next week, which I will be attending.

The First Draft of the 2001 IPCC Report had a "projection" graph for temperatures by the year 2100 which gave a maximum temperature rise of 4ºC. This figure must evidently have been considered to be not high enough, because the second draft, and the final one, had a figure of 5.8ºC which had been achieved by inventing an extra extreme "scenario". A1F1.

The latest "Climate Change 2007: Summary for Policymakers" gives the "projected" "Best Estimates" and spurious "ranges" for six different "scenarios". The most extreme one, which is still A1F1. gives a "Best Estimate" figure for a "projected" temperature rise by 2100 of 4.0ºC, with a "range" of 2.4ºC to 6.4ºC. You can bet your bottom dollar that the only figures anyone will quote is the 6.4ºC. All the rest, which go down to 1.1ºC, will be ignored.

The world is in the grip of "climate change" hysteria. Today"s BBC News gave an interview with the Mayor of San Francisco who is walking to work instead of taking the car. Next, perhaps, he will give up walking as well so that he exhales less carbon dioxide.

Down here in New Zealand they are so keen to get us to use public transport that they are scouring the museums to find 30s style railway carriages to put back into service to cope with the demand, and bus drivers currently have to ask passengers where the bus is supposed to go.

Cheers,
Vincent Gray

[CHJ]

mechsympathy said:

mattikake said:

And some 'facts':

http://www.nationaltrust.org.uk/main/w-chl/w-count...

If, as they claim, sea levels around the UK have risen on average 10cm how can the level have risen on 25cm around Liverpool?

it's all the hub caps thrown in the Mersey

[ludo]

Guam said:

ludo said:

Guam said:

ludo said:

Guam said:

ludo said:

ludo said:

Guam said:

Agreed the localised chaotic effects due to micro climate impacts will average out across the data. However the fundamental flaws remain in both modelling excercises do they not

So you would agree that predicting the climate should be easier than predicting the weather (as the chaotic element is not relevant)?

The models are approximations, there is a good deal of uncertainty over the details, but I have not read anything in the litterature pointing out fundamental flaws. Perhaps you could give some examples?

Good lord man I have repeatedly you cannot forecast with any degree of certainty (extrapolate if you will) where ALL the variables are not fully understood, nor their impacts fully understood, Solar Variation for example, EVEN if you average out the minor background noise (localised chaotic events) you cannot as yet forecast accurately what toonage of CO2 will result in what degree of warming or its abscense and the converse. Those are just two, when a major event (vulcanism) occurs what degree of theoretical change in the climate occurs and how sustained will it be, and to what degree (up or down) will it mask the trend?

The same inherent rational flaws as in Weather forecasting (just the variables may differ).

Comes back to Spurious correlation in raw statistical analysis, just because you seem to have a relationship at first run through the data doesnt mean there actually IS one!


Cheers

Sorry, that is just repeating your assertion. Can you provide e.g. a journal paper or even a reference to one that identifies a fundamental flaw in GCMs. I know your opinion, because I read it the first time. What I am asking for is some evidence to back up your opinion.

BTW, weather noise is larger in magnitude than climate trends, so "minor background noise (localised chaotic events)" sounds rather incongruous in this discussion. Also there have been studies where the observed effect of Pinatubo was compared with the GCM model prediction, and the results showed the models had useful skill (although of course not perfect).

To the contrary YOU are the one putting the concept (along with others) it is YOUR responsibility to show where the criticisms are baseless,

Sorry that is not how science works. You can't prove anything using observations, only corroborate or disprove (see e.g. work of Popper). However, unless you can provide evidence of a specific fundamental flaw, the criticism is too vague to expect a refutation.

TB Has just done so above, there is no need for two of us to be replicating the same data (boring in the extreme).

And as I have already said, GCMs are known to be unable to accurately reproduce station data and provided an URL for an article on RealClimate that explains why. So all TB provided was a paper that confirms that GCMs are not very good at doing something the modellers say the models cannot be expected to be good at.

[esselte]

ludo said:

Their predictions are generally accurate, within the known uncertainty of the predictions (i.e. the error bars - which are generally pretty broad).

What you're saying is black could be white so long as there's enough grey either side of them....

[ludo]

esselte said:

ludo said:

Their predictions are generally accurate, within the known uncertainty of the predictions (i.e. the error bars - which are generally pretty broad).

What you're saying is black could be white so long as there's enough grey either side of them....

[turbobloke]

esselte said:

ludo said:

Their predictions are generally accurate, within the known uncertainty of the predictions (i.e. the error bars - which are generally pretty broad).

What you're saying is black could be white so long as there's enough grey either side of them....

They are "worse than nmo model" which takes some doing.

[turbobloke]

Guam said:

So as I said earlier in the thread basically the GCM's are a set of guesses based on other guesses. Yep they are reliable in the extreme!!

Cheers

Ah but you'er not quite there yet. The impacts models feed off the GIGO from the climate models, so we got international policy based on a set of guesses based on a set of guesses that are worse than no guesses with dodgy data and polticisation to go.

You couldn't make this stuff up but the IPCC has no problems