By Christopher Monckton of Brenchley
The monthly satellite lower-troposphere temperature anomaly from Remote Sensing Systems, Inc., is now available.
Taking the least-squares linear-regression trend on this dataset (the bright blue horizontal line through the dark blue data), there has now been no global warming – at all – for 17 years 5 months.
Would readers like to make a projection of how many mainstream media outlets will report this surely not uninteresting fact?
It shows that the Hiatus hernia for true believers in the New Religion continues.
My own prediction is that the number of media reporting 17 years 5 months without any global warming will be approximately equal to the number of general-circulation models that predicted such a long Pause notwithstanding ever-rising CO2 concentration.
Print out the graph as a postcard and send it to the editor of a newspaper near you that has shut down democratic debate by announcing that it will refuse to print any letters at all from “climate deniers”.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
The sneakily pseudonymous “JBJ” asks what my scientific credentials are. Only a Socialist would ask such a question. A scientist would look at the quality of my published research and form a view that way. If “JBJ” were to read the nonsense originally posted on this thread by the snidely pseudonymous “Village Idiot”, and my detailed replies to each of the “Idiot’s” pseudo-scientific fatuities, it would be able to form a view on which of us was right, scientifically speaking. It does seem that the trolls are becoming more desperate every day as their precious global warming theory collapses for lack of – er – global warming.
Bruce says:
February 8, 2014 at 12:03 pm
If four centimeters of sea level rise in 17 years & five months could submerge a cay, then it was really more of a sand bank than an island. And that’s taking a high guess for Caribbean MSL rise during that period.
rgbatduke says:
February 8, 2014 at 8:48 am
Well & tightly argued, as usual.
It has IMO long been apparent that GCMs aren’t adequately skillful tools upon which to base economic & energy decisions affecting the lives of billions at a cost of trillions.
The infantile state of climate modeling now reminds me of the premature attempt in the 1970s to wage a war on cancer, at a cost of billions (tens thereof inflation adjusted), before the basic science involved was adequately understood. Throwing money at cancer didn’t work, but wasted resources which could better have been expended elsewhere.
Monckton of Brenchley says:
February 8, 2014 at 2:10 pm
So what are your scientific credentials then??? You didn’t answer my question!
Monckton of Brenchley says:
February 8, 2014 at 2:10 pm
Are you saying socialists cannot be scientists? Your true colours are coming out!
“It does seem that the trolls are becoming more desperate every day as their precious global warming theory collapses for lack of – er – global warming.”
Another point Mr. Moncton … I do not believe in AGW … you are very presumptuous to say the least.
rgbatduke says:
February 8, 2014 at 8:48 am
” It would be extremely wise to assemble a correctly done counteranalysis of the GCM data (which is, fortunately, freely available) and use that analysis to reject the failed models and thereby quantitatively moderate the egregious predictions and “confidence” assertions of AR5′s SPM.”
Well, I guess after waiting for a decade or more for some mainstream climate scientist to do something about the hopeless models, eventually someone has to take up the task. Skepticism used to be a lot easier – you didn’t have to prove some alternative, just logically deconstruct the status quo or show the math and stats were incorrectly applied. So, I agree the job has to be taken up or we will be flagellated with meaningless rainbow spaghetti for decades to come. While we are at it, we better start out with an overhaul of the thermometric record which has been the main front of tinkering CAGW climate scientists trying to hold the trace up to match the trajectory of the spaghetti. How on earth can one hope to improve the physics of climate models without correcting the fiction of the temperature record upon which they are being tested? We could chop off a few tenths right now, based on the idea that if these guys with their belief system were tinkering with the record all these years, they certainly wouldn’t have been subtracting temps. Submerging the mid30s-mid40s by several tenths to erase this otherwise-still-standing record warm decade could easily be re-visited.
I’m led to recall Donald Trump’s intervention in getting the skating rink in Central Park going after decades of failure by NY authorities.
http://news.google.com/newspapers?nid=1899&dat=19860610&id=OPJGAAAAIBAJ&sjid=dfMMAAAAIBAJ&pg=4023,1181043
He took on the task and succeeded and at the same time gave an insight into why he is so successful. He went to Toronto and asked who is the best hockey rink designer there is, found him, and this guy came down, threw out all the old badly functioning system and set up one that since has given joy to hundreds of thousands of New York skaters. Donald Trump, who was well used to being hated for his success, got a new wave of hateful static, of course.
The current crop of corrupted leading CliSci’s are not going to get the thing going in the right direction no matter how far out their projections have been. They’re busy trying to rationalize the fiasco. They’re chasing the delinquent heat like a pinball in three dimensions. They are checking up aerosols (did I spell that correctly) and other confounding avenues, but not to find alternatives to their GHG theory of warming. No. Rather it is to put the dyed-in-the-wool theory on life support. You are right that we don’t know if we will end up with a hot future or not. We do know that the theory as expounded has been proven wrong, though. I gasp to think what would have happened if they were wrong but the climate did continue to heat up. We would have covered the earth with solar and windmill power and when it turned cooler would have had a self-fulfilling prophesy that would have destroyed civilization, freedom and put Greenpeas on every plate. Yeah, for goodness sake let’s do something.
C. Monckton: one other thing. We are doing the CAGW theory a favor by extending the horizontal temp further into the past. They counter with slightly reduced linear slope, arguing it is still rising. Let’s start at 2005, or whatever it is and proceed with a cooling trend.
The Contributions of JBJ
#1 – That regression is statistically pointless!
#2 – Look at this graph http://cdiac.ornl.gov/trends/temp/vostok/graphics/tempplot5.gif
I don’t think it would be too difficult to find a date which shows there has been no warming for the last 300,000+ years!
#3 – “In the meantime, the “Idiot” should be more circumspect in future when attempting to sneer about matters of which it has little understanding.”
And your scientific credentials are what???
#4 – So what are your scientific credentials then??? You didn’t answer my question!
#5 – Are you saying socialists cannot be scientists? Your true colours are coming out!
=====
That’s everything JBJ ‘contributed’ in this forum.
The head teacher running the local schoolhouse, is always “The Principle” (etc).
He may well have strong principles, but he would probably be called “The Principal”
Nick Stokes says:
February 7, 2014 at 2:06 pm
You can see that easily here. The regions of near zero trend are marked in brown. You can click for details. It gives CI’s (85%) and t-statistics
Nick – your triangular diagram is really clever. What software did you use to produce it? It neatly summarises pages and pages of data which I have tabulated in Excel spreadsheets. I also have one or two financial spreadsheets which could also benefit from such clarity. My only complaint would be the similar colours for small positive and negative trends, but I guess these are customised to taste?
Regards
Richard
Richard Barraclough observed @ur momisugly February 8, 2014 at 8:09 pm;
aarrrggghh … thank you. 😉
Monckton of Brenchley said February 8, 2014 at 2:10 pm:
“The sneakily pseudonymous “JBJ” asks what my scientific credentials are. Only a Socialist would ask such a question.”
In as much as the issue of “credentials” comes up here from time to time (too often!) I have a couple times posted my favorite quote on the issue, from Noam Chomsky, who many would consider to be a socialist I believe! So do you mind one counter-example? Here is what Chomsky said:
“Generally speaking, it seems fair to say that the richer the intellectual substance of a field, the less there is a concern for credentials, and the greater is the concern for content.”
Richard Barraclough says: February 8, 2014 at 8:18 pm
“Nick – your triangular diagram is really clever. What software did you use to produce it?”?
Thanks, Richard. There are R programs that I run at home. Some download (update weekly) and process the data. Others calculate the regression coefficients and the CI’s etc. There’s a bit on the computation of CI’s here. It takes about an hour to make all the triangle and timeseries plot files.
The rest is Javascript. That does all the responses to clicking, including re-calculating the data that pops up when you click, and also the dynamic picturing of trend lines.
I’ve written a few posts, with links listed in this most recent one.
rgbatduke says: February 8, 2014 at 7:39 am
” Makes cherrypicking easier than ever before (just kidding, kinda, maybe:-)!”
Actually, I titled the first post “A cherrypicker’s guide”. My hope was that, while it made cherrypicking easier, it also made it easier to spot. But thanks for the kind remarks.
Solar models show that this period would be about the same temperature, or slightly cooler see: http://lasp.colorado.edu/sorce/news/2012ScienceMeeting/docs/presentations/S2-03_Scafetta_SORCE.pdf Until people are really understanding the basis of the cycles then you can’t take chances with global warming. In this case it is better err on the side caution.
Ted Clayton says:
February 8, 2014 at 4:42 pm
The Contributions of JBJ
Thanks Ted … I’m glad you enjoyed them 🙂
The “pause” (already a rationalization) is a perfect fit to Taleb’s Black Swan Theory:
http://en.wikipedia.org/wiki/Black_swan_theory
“Identifying a black swan event -Based on the author’s criteria:
1)The event is a surprise (to the observer).
2) The event has a major effect.
3) After the first recorded instance of the event, it is rationalized by hindsight, as if it could have been expected; that is, the relevant data were available but unaccounted for in risk mitigation programs. The same is true for the personal perception by individuals.
Brian T. Johnston said @ur momisugly February 8, 2014 at 10:26 pm;
The Scaffetta PDF is interesting. Not compelling, but a worthwhile read. Factors other than planetary gravity could cause effects. Magnetospheres, eg.
Can’t take chances with global warming? And CO2 did it? And combustion has to be curtailed?
Let’s assume for the moment that CO2 wasn’t from the start Activism & Politics.
Our population now relies on a smooth-running ‘technological civilization’. There are several ‘linch-pins’ that have to be in place, to support this phenomenon. The economy, and energy, eg.
In the USA, more than 3/4 of the economy is “consumer spending”. Without it, the economic house of cards comes down, and with it civilization.
We jerk society around too much, our support-system could fail.
So, we err on the side of caution, to minimize big-picture risks.
His suggestion of conducting a proper statistical analysis based upon the models is interesting. It would be an enormous task. My one fear is that, by the curse of intercomparison, much the same errors are propagated throughout the models, and I am not sure how a statistical analysis of their output would take account of this.
I don’t think it would be that enormous a task. Steve McKintyre has the data here:
http://www.climateaudit.info/data/models/cmip5/GLB/
and the entire point is to NOT compare the models to each other or consider their shared errors, but to take each model, one at a time, and apply the ordinary rules of hypothesis testing to the following null hypothesis:
This GCM is a perfect representation of climate dynamics.
in comparison with the actual time evolution of the climate. Chapter 9 of AR5 is (apparently) morphing over time to partially acknowledge the problem; box 9.2 is an entire apologia in three parts for the “fifteen year” hiatus (their terminology) where it is only 15 years because they terminate the record in 2012. However, the fundamental problem with Chapter 9 is a very simple one. They base all of their predictions and projections and assessments of probable accuracy on two things:
a) The Multimodel Ensemble. Seriously, what are they thinking? Yes, they acknowledge in section 9.2.2.3 that model runs are not iid samples from a distribution. They present many learned papers that address the fact that model runs are not iid samples from a distribution and hence “creates challenges for how best to make quantitative inferences of future climate” (an understatement if there ever was one!). As it is, the multimodel ensemble mean is the next best thing to a completely mean-ingless quantity as exists on Earth. And while they do indeed do a credible job of listing some of the reasons that this is true (the models aren’t independent — they share lineage and hence possible bias, the models don’t all contribute the same number of model runs to the overall result, the models do not all score equally well in the training set validation) while ignoring one of the most important of the others — there is no reason to believe that intermodel errors are a uniformly distributed random variable with a symmetric effect on the mean performance and so there is quite literally zero reason to expect that averaging over models will give a better result than simply using the best model.
In fact, we know perfectly well that the best model will usually outperform the mean performance of an ensemble average of the best model and a bunch of worse models, especially in this precise context. Treating the results of the worse models as noise, one can apply the concepts of information theory and predict the degradation of the information due to the admixture of noise from the inferior models. To give a lovely example in physics, to the extent that Hartree-Fock models usually outperform Hartree models in capturing correlation/exchange energy, averaging ten Hartree models to one Hartree-Fock model simply guarantees a worse performance in a systematic way.
b) Validation on the training set. This is truly inexplicable to me. Forget about things such as figure 9.2c, which (if I understand the figure) indicates in simple graphical form that nobody could possibly miss that the MME mean fails badly to correctly represent the climate — consider figure 9.8a, which basically plots CMIP5 against the actual climate across all of the time span of HADCRUT4. Are they kidding me? The training set period is clearly marked — the agreement even of the meaningless mean within the training set is indifferent at best and the performance of the individual models is for the most part terrible — one only preserve the illusion of good performance by means of the mixing of TRAINING SET FITS which are more or less constrained to give the right interval increase in temperature and hence cannot much deviate in the mean from HADCRUT etc. If one examines the rest of the graph, hindcast performance is much worse than even this. For example, it completely fails on the interval from roughly 1900 to 1940 — it completely misses the first half of the twentieth century, both one model at a time and in meaningless aggregate. No wonder they can state with “confidence” that natural variation is unimportant — they build models that eliminate a completely natural interval of strongly increasing temperatures and replace it with a far-too-smooth, much warmer curve reaching back to 1870 or so and missing most of the interesting bounces in between, claim that the MME mean is validated on the training set (where it is not, not one model at a time, which is the only way a meaningful hypothesis test/validation can be performed) and then assert that because the models didn’t need to use natural variation on the training set it still is unimportant on the future prediction.
Hell, the MME mean failed in the past predictions for a 30-40 year interval — why wouldn’t it fail in the future as well?
The point is that it is really, really easy to fix this. Simply test each model against HADCRUT4! Most of the models have the data from their individual runs available somewhere or the other — the part generated by the perturbed parameter ensembles discussed in 9.2.2.2. This is really the only part that matters. If a given model is used to generate 100 PPE runs, and these runs collectively spend 98% of their time above the actual measured climate across the entire trial period (or fail any one of a number of other elementary tests, such as having approximately the correct autocorrelation time and variance), presto chango, the model has now officially failed a simple hypothesis test with a p-value of 0.02 or less!
When an individual model fails and individual model hypothesis test, the only sane thing to do is to remove it from the MME! I mean jeepers! In figure 9.8a, which is basically the full span version of SPM 1.4 without the last-second adjustment voodoo, one can clearly make out multiple colored threads that spend all of their time above even the MME mean, let alone the actual HADCRUT4 temperature far beneath that mean. One can see that most of the threads are spending all of their time above the actual mean.
How many threads? According to box 9.2: 111 out of 114 realizations, and that is as of 2012, not as of the present. Box 9.2 is basically openly acknowledging safely in a science section where no policy maker will ever find it that CMIP5 is badly broken, we don’t know why, we cannot even narrow the possible reasons why down to only two or three, and the best that we can do is note that the models themselves sometimes one at a time spend 17 year spans too warm so that this doesn’t disprove them!
Come on, this is pure piffle. You can’t have it all ways — ignore the weighting of the models in terms of PPE, ignore their shared lineage and biases, refuse to apply any sort of hypothesis test to the models one at a time before admitting them into the ensemble to be used to make critical decisions regarding the expenditure of the energy of the entire human species for decades into the future, the first such collective expenditure in human history after the end of the cold war and one that diverts that energy away from things like establishing world peace and prosperity in favor of averting a hypothetical disaster, create a MME mean that by itself fails a hypothesis test against the data everywhere outside of the training set where it is constrained not to fail and then claim any sort of statistical high ground or knowledge at all!
This is absolutely shameful. There really is no question about this. What they do is indefensible, and its indefensibility will be obvious to the entire world in another three years no matter what unless they get lucky and warming resumes at a frenetic pace! Which even they recognize is now not likely gonna happen. It could, but it is literally not probable at this point and they know it!
rgb
As I have not yet seen an answer, I will repeat my question. My Lord Monckton, would you please be kind enough to explain what I am missing here?
—
I’m afraid I don’t quite understand this:
Sure enough, in the 8 years 9 months from 1 January 2001 to 30 September 2009 (my speech was on 14 October) the RSS dataset shows a statistically-significant cooling of 0.16 K, equivalent to 1.87 K/century of cooling.
The “Idiot” also says I said in that same speech that all global warming had stopped since 1995. And so it had – all but a statistically-insignificant 0.4 K warming over the period from January 1995 to September 2009.
Why is 0.16K/45 months (0.0035/month or 0.43/decade) statistically significant but 0.4K/117 months (0.0034/month or 0.41/decade) statistically insignificant?
—
I’m not clear on how cooling of 0.43K per decade is significant, but warming of 0.41K per decade is insignificant. Can anyone on this thread clear this up for me?
TonyG:
At February 9, 2014 at 10:23 am you ask
OK, I will try. I suspect nobody answered because your question reveals there is so much you do not understand that an attempt at an answer is daunting.
The short answer is that significance is a function of the variance of the data and is often expressed as a percentage confidence. But I feel sure that is only words and not information for you.
This link may help. If not then get back to me so I can try to explain it.
http://mathbits.com/MathBits/TISection/Statistics2/correlation.htm
Richard
richardscourtney says:
OK, I will try. I suspect nobody answered because your question reveals there is so much you do not understand that an attempt at an answer is daunting.
Seems to me that it’s better to help someone understand than to leave them ignorant, especially when they’re asking for help.
The short answer is that significance is a function of the variance of the data and is often expressed as a percentage confidence. But I feel sure that is only words and not information for you.
No, I actually get what you’re saying – it seems that I was missing some of the context, and that was the cause of my lack of understanding. I was simply plotting the amount of change over the same period, instead of relating it back to the data. So my simplification was not a valid comparison.
I’m learning, and have been from a few years – but it’s really easy to miss things when I only have a chance to read in snippets. I wish I had more time to devote to understanding things.
This link may help. If not then get back to me so I can try to explain it.
http://mathbits.com/MathBits/TISection/Statistics2/correlation.htm
Thank you for the link. I understood the concept of the correlation coefficient, but only in a very broad sense (I’m aware of it and of its importance, but not the details of how it’s calculated and such). That fills in a lot of details for me. I’ll be keeping that window open & studying it a bit more.
TonyG says:
February 9, 2014 at 10:23 am
These quote are from Monckton of Brenchley February 7, 2014 at 5:05 am
By way of acknowledging your request, my initial remark would be that the “0.4 K warming over the period from January 1995 to September 2009” is termed insignificant, because much stronger warming was supposed to be occurring.
What warming took place during the Pause, is deemed insignificant, in comparison to the unrealized predictions of computer climate models.
I speculate … but it’s a start. 🙂
[Ah – Richard is here, but mine requires less homework. ;]
Monckton of Brenchley said in part February 8, 2014 at 3:50 am
“…..There was, however, a typo in one of my answers. Though the “Idiot” did not spot it, others did. I had referred to “a statistically-insignificant 0.4 K”. The figure should have been 0.04 K.”
Not unlikely, TonyG (February 9, 2014 at 10:23 am), this is why your question was unanswered.
Bernie Hutchins
“…..There was, however, a typo in one of my answers. Though the “Idiot” did not spot it, others did. I had referred to “a statistically-insignificant 0.4 K”. The figure should have been 0.04 K.”
And then there’s that. Leaving aside coefficient of correlation, even on my super-simplified approach that makes a HUGE difference! I missed the correction – thank you for pointing it out.
I am grateful to Bernie Hutchins for answering “TonyG’s” question. I had omitted a zero, and a value that should have read “0.04 K” read “0.4 K” by mistake. Nearly as bad as the IPCC printing that all the ice in the Himalayas would be gone by 2035 when they meant 2350: except that their error was deliberate and mine accidental.
And I am also grateful to Professor Brown for elaborating his suggestion of a project to evaluate the climate models statistically, one by one, comparing them with the real-world outturn. If he would like to say how many scientists the task would need, and how long it would take, and how much it would cost, it might be possible to find funding for the project.
The Professor would be an admirable lead scientist for the project that is his idea. His rigor and clarity of thought are exceptional.
Nick Stokes
Thanks for the links. I’ve never used R, so I’ll schedule myself for a bit of education next week
Cheers
Richard