Dodgy statistics and IPCC Assessment Reports

Guest essay by Christopher Monckton of Brenchley

In the 19th century, British Prime Ministers used to say there were “lies, damned lies, and statistics”. In the 21st century, we may say there are frauds, serious frauds, and IPCC Assessment Reports.

Recall, for instance, the notorious graph in the Fourth Assessment Report that falsely indicated that the rate of global warming is accelerating and we are to blame. Using the same statistical dodge, one can show that a sine-wave has a rising trend.


In the Fifth Assessment Report, the IPCC still cannot bring itself to behave. My expert review of an earlier draft of that report opened with these words:

“To restore some link between IPCC reports and observed reality, the report must address – but does not at present address – the now-pressing question why the key prediction of warming in earlier IPCC reports have proven to be significant exaggerations. The IPCC’s credibility has already been damaged by its premature adoption and subsequent hasty abandonment of the now-discredited “hockey-stick” graph as its logo; by its rewriting its Second Assessment Report after submission of the scientists’ final draft, to state the opposite of their finding that no discernible human influence on climate is detectable; by its declaration that all Himalayan ice would be gone in 25 years; and by its use of a dishonest statistical technique in 2007 falsely to suggest that the rate of global warming is accelerating. But the central damage to its credibility arises from the absence of anything like the warming it had predicted.”

The IPCC have indeed addressed The Pause. But they have addressed it by using statistical prestidigitation to air-brush it out. As Bob Tisdale has pointed out, the very first graphs the reader of the Summary for Policymakers will see are in Figure SPM.1, which consists of three panels. Each of these panels exploits bogus statistical techniques to vanish the pause.

Here is what They did and how They did it.

The first of the three panels shows the global instrumental surface temperature record since 1850:


And what is wrong with that? It looks innocuous enough, but a mathematician would take one look at it and sniff. He would see two things obviously wrong with drawing any conclusion about dangerously-rising 20th-century temperatures from this graph.

First, there is the aspect-ratio dodge. For the x axis is in years and the y axis is in Celsius degrees of temperature change. One can choose any aspect ratio one wants. To make 20th-century global warming look worse, just stretch the graph northwards.

Not all climate extremists know that. In a debate with me on Roy Green’s radio show in Canada a few years ago, one of the pointy-heads at TheSmugBlog asked the audience, with that earnest desperation in his voice that is mandatory, “But don’t you see how serious it is that global temperatures are rising at an angle of 45 degrees?”

I had to explain to the poor sap, as gently as I could, that degrees of arc and degrees of temperature change are clean different things.

But it is Dick Lindzen, whose vast experience and profound knowledge allows him to put the climate scare into perspective as no other can, who has best illustrated the insignificance of 20th-century global warming.

His local paper, the Boston Globe, prints the previous month’s temperature movements in the city. He has superimposed on that record an orange band that shows the entire warming of 0.75 Cº over the 20th century.

Even allowing for the fact that a global annual average will change less than a regional monthly one, it is difficult to look at Dick Lindzen’s orange band and draw the conclusion that 20th century global warming was alarmingly beyond the bounds of natural variability.


The second statistical dodge in the IPCC’s first panel is the error-bars dodge. If you look carefully at the error-bars in the IPCC’s graph, you will see that they are absent. Let us remedy that absence:


Even today, the combined measurement, coverage, and bias uncertainties in the global terrestrial data are ±0.15 Cº. The uncertainties were far larger in the 19th century. Notice also how much less drastic and exciting the graph looks once the 2 σ uncertainty bounds are plotted.

There is a third dodge that is not directly evident from looking at the graph itself. All around the world the record-keepers have been rewriting the temperatures in the early 20th century to push them downward, so as to make the rate of warming over the century seem a great deal steeper than it was. Here, for instance, is New Zealand:


And Darwin Airport, Australia:


And the U.S. Historical Climate Network, before and after adjustment (this example and the next two are thanks to the vigilant Steven Goddard):


And the GISS record at Reykjavik, Iceland, before (left) and after adjustment (right):


And Santa Rosa, CA, this time with the trend-line added:


The effect of all these tamperings is to make it look as though there was more global warming in the 20th century than there was. Fortunately, there is not so much scope for the compilers of the terrestrial temperature records to tamper with what has happened since 1979, because the watching satellites now provide an independent record of global temperature change.

So to the second of the three mendacious panels in Figure SPM.1:


This graph is an illustration of a meme that has become a favorite with the apologists for Apocalypse: the most recent decade was warmer than earlier decades, so global warming is still getting worse (for the theology of the New Religion, standing common sense on its head, is that warmer weather is worse than cooler).

The priceless advantage of taking decadal averages, if one wants to magic the Pause away, is that it wipes out the entire trend of the most recent decade. One can dock off a further two years if, as here, one uses the decades 1991-2000, 2001-2010 etc. rather than 1993-2002, 2003-2012 etc. Finally, using decades docks off all the months of the current year. So this statistical dodge neatly erases the past 12 years 8 months of the Pause.


And, by what is perhaps more than a coincidence, the length of the Pause, taken as the longest period exhibiting a zero least-squares linear-regression trend on the mean of the monthly global mean surface temperature anomalies on the three terrestrial datasets, is – er – precisely 12 years 8 months.

There is another and more subtle dodge here. As we saw in the earlier graph of the uncertainties in the HadCRUt4 global temperature dataset, the error bars narrow toward the present. The way the IPCC has presented the decadal blocks on the graph exploits this to make it seem that the blocks in the 1970s, 80s, 90s, and noughties are much further apart than those in the 1910s, 20s, 30s, and 40s, implying without quite saying so that the rate of warming over the four most recent decades on the graph was significantly greater than the warming earlier in the 20th century.

Dick Lindzen, however, uses a graph that shows how little difference there is between the earlier and later periods of warming, even though it was only in the later period that we could have exercised much influence.


One panel shows the global temperature anomalies from 1895-1946. The other shows the anomalies from 1957-2008. Both cover 52 years. Both are plotted to an identical scale. Dick Lindzen asks his audiences whether they can tell which panel covers which period. It is not at all easy to tell.

Which brings us to the third panel. Here, the dodge is one of the newest in the arsenal of statistical shiftinesses on which the IPCC draws with such disfiguring frequency and relish. It is the use of colors, and bright ones at that, to try to suggest that the mild and beneficial global warming of the 20th century was grievous and alarmingly damaging.


And here the IPCC will find that it has made a mistake. Previously it has chiefly used bright colors in the red scale to indicate predictions of future planetary overheating. However, most people, on looking about them, will see remarkably little change as a result of 100 years’ warming. The trees are greener; the deserts have shrunk by quite a bit (the Sahara by 300,000 sq. km in 30 years); sea level is 8 inches higher; and that’s it.

Recoloring the graph in neutral tones would have been more scientifically adult:


Does the Earth really look that much different as a result of 0.7 Cº global warming over 100 years? Not really. Let us end with a God’s-eye view of the planet He has given us. Really, our stewardship has not left it in too much of a mess.





newest oldest most voted
Notify of

So why is the world spending trillions on attempts to reduce CO2 emissions? Why does Obama think regulating CO2 is so important that he will go it alone without Congress while equating CO2 to Arsenic and Lead?


Well written, and it would be nice if this were added into the IPCC report.


Well, since the IPCC got rid of the Medieval Warming Period, which was approximately 500 years in length, it can surely get rid of that annoying 20th-21st century “pause” of roughly 17 years in length.

Pablo an ex Pat

Whenever an assertion is made by the alarmist crew in the press a few minutes checking the background typically exposes it as being hyperbole.
We are relentlessly told that such and such an event is unprecedented, take the Colorado flooding, only to be told in the same article that it’s the worse they’ve seen in that area in x years. In the case of Colorado x = 37 years. A large tragic event ? Certainly. Unprecedented, err no, a similar occurrence happened in recent memory. Similar occurrences have likely occurred and unrecorded for thousands of years.
It’s the relentless hyperbole that weakens them, I hope that they continue to do it.

New paper in Nature Climate Change says IPCC uses statistical techniques ‘out of date by well over a decade’
Two papers using a modern statistical technique [cointegration] falsify AGW:

Silver Ralph

Let us end with a God’s-eye view of the planet He has given us.
I think you will find that your singular deity has to be neuter, not male.
Unless you are honouring the Egyptian Atum, who was certainly male. But then Atum’s hand was considered to be female, for obvious reasons (if you recall how the universe was formed). But despite popular opinion, I have not found any evidence that Atum went blind.
Or perhaps you are honouring the Primeval Adam, the primary deity of the Nazarenes, who was said to be hermaphrodite. Saul (St Paul) and others were, of course, Nazarene. But a hermaphrodite can hardly be male.
I would hate to think the deity currently in favour, is going to be offended by your incorrect gender attribution.

I am pleased to see them acknowledging the hockey stick was discredited.
But as an earlier commenter noted, the IPCC is not science, but Politics. So of course they are going to use polemics to get their point across. it does not matter if their point is false or not, only in getting it across.

Silver Ralph:
In attempt to avoid this thread being deflected by other proselytising atheists jumping in to support your religion, I write to point out that He is a gender neutral word when applied to an individual of unspecified gender. In other words, the excuse you used to present your nonsense was plain wrong so there is no valid reason for anybody to follow it up.

Pamela Gray

But if one were to include the error bars on that colorful globe to show statistically significant warming, the colors (or rather noncolors) associated with zero would dominate. And that is the greatest dodge of all.


With each year for which there is no meaningful increase in temperature, the odds that the IPCC projections are incorrect grow exponentially.
No wonder the alarmists sound a little desperate lately.

Here’s a good example circa Climategate that captures adjustments to raw data in Illinois rather nicely via “blink comparator.” Includes Wisconsin and Iowa as well (around 80 examples total).

@b>Silver Ralph, September 17, 2013 at 12:40 pm: “I would hate to think the deity currently in favour, is going to be offended by your incorrect gender attribution.”
Great Ghu the Grandfather God is offended by your weird effingness.

J Martin

One day I hope to see some of those graphs of climate fraud published in the worlds press and shown on TV. Maybe David Rose at the Daily Mail would be interested, or Christopher Booker, or James Delingpole.

Forgot to mention in my comment that my example elaborates on Christopher Monckton of Brenchley’s observation:
All around the world the record-keepers have been rewriting the temperatures in the early 20th century to push them downward, so as to make the rate of warming over the century seem a great deal steeper than it was.

United Nations = So united they can not agree on one important thing ever.
Useless Fools = So useless they can not agree on one valid fact ever.
Those who sign off on these lies should have all their degrees withdrawn and all the funds they by lies and deception aquired taken from them total.
High crimes aginst all mankind.


Silver Ralph says:
September 17, 2013 at 12:40 pm
“He” in English grammar is both the neuter & masculine gender pronoun, not necessarily connected to biological sex.

Bill Clinton supports the IPCC reports.
Al Gore supports the IPCC reports.
John Kerry supports the IPCC reports.
B. Obama supports the IPCC reports.
Common trait of these 4.

An Engineer

I don’t see anything odd with the first panel: both axes are linear and the upper and lower bounds appear to be chosen to allow the data to fill as much space as possible.
I don’t see anything odd with the second panel, the block widths are equal and the axes limits and scaling are the same as on the first panel.
On the last panel the only thing I find odd is that the contour interval has been tweaked so that it not constant across the full range of data; that would be worth looking into as the contours should be evenly spaced and cover the full range of data.

Bob Diaz

It appears that increased CO2 results in increasing frauds, serious frauds, and bad IPCC Assessment Reports. Thanks for your report.

Ralph of the Silver fingers:
As my nick suggest, spent time in Forward Operation Bases in combat zones.
Often the atheists (odly often from California) would come running during a bad ambush fire fight.
“You that Apache guy who knows how to pray.” ?
“YA” but ya see he may be busy like I am just now, got get us some ammo….”
You need to “help thyself” by knowing more.

Mumble McGuirk

If UKIP ever forms a government, what say we put His Nibs in for an OM?


Good points, but what are the ulterior motives of the controlled opposition – Monckton?

Pamela Gray

Richard you crack me up. My diety ensemble (I’m catholic-lolol) protects me like a female bear and lifts me up on her eagle wings. But Richard, you are free to be neutral.
I found great humor in Silver’s post. His post reminds me of my past. I found myself at odds with a number of elders during my long history as a Sunday School teacher and liturgical choreographer. Yet I was the one using proper gender in the original Arabic and Hebrew languages. One of these days you and I must share those stories. Like the time I choreographed a Holy Friday performance titled “Sophia’s Trial”, which was an interpretation of the crucifixion. We compromised. I got to put a female in the lead role of Sophia, but they made me change the title to “Wisdom’s Trial”.


I am pleased to see Monckton address the misleading graphs, statistics, etc., and wanted to add a thought I had when first seeing the third of the three panels that are at the end of this article. The scale has zero as being the turning point of the blues to the reds, but has not (even in the IPCC) there been an established “normal” warming? I was sure the argument was, “Yeah, the earth has been warming, but all this carbon has made it accelerate.” If that is the case, and that “normal” warming is, say, one degree per century, then it would follow that the point of transfer from blue to red in the scale should occur at the one degree mark, not at the zero mark.
In saying this, I do like how Monckton has used neutral tones, but I think a better reconstruction of this panel would be to shift the blue/red transfer point to wherever the “normal” warming per century is located – that would then highlight the “excessive” warming in red, keeping the “normal” warming in blues. Just by looking at the panel and imagining if (taking the one degree as “normal warming” for a moment) instead of the light blue/light yellow being the colors straddling the zero degree value they straddled the one degree value, the entire panel would suddenly become very dominantly blue and light yellow, with only a few smatterings of red.
Not being one who is experienced with graphical editing software, I would not know how to do this, but I am sure that one of you could and make an astounding rebuttal panel that put the light blue/light yellow moment at whatever the IPCC considered, “normal warming before human influence” and showed how much the warming is not a big deal at all.
Just a thought. Love this site, it is one of my top four that I go to. Keep up the good work everyone!


One can dock off a further two years if, as here, one uses the decades 1991-2000, 2001-2010 etc. rather than 1993-2002, 2003-2012 etc.

My eyeball tells me that the IPCC is using the decades 1990-1999, etc. If so, the IPCC is docking off three more years from The Pause, not just two.
Incidentally, Here’s a comment I posted earlier today at the end of another thread:
Here’s how to instantly undermine the psychological impact of the decadal chart. Someone here with charting skills should add a short-length block to the chart that shows the 3.5 years from 2010 thru June 2013. It will be below the block from 2000 thru 2009, indicating the warm has turned.
This block should be extended every six months to keep the follow-on chart up to date.

Simon W says:
“what are the ulterior motives of the controlled opposition – Monckton?”
Not sure of the point you’re trying to make. But it sounds slightly derogatory. So I ask: what are your ulterior motives?
Lord Monckton has been very straightforward throughout the entire debate, supporting his arguments with testable science. To label that as “ulterior motives” reeks of ad hominem desperation due to a lack of any credible science.
If it were not for ulterior motives, the climate alarmist crowd would cease to exist. They never admit to their true motives, do they? That is why they tuck tail and run away from any fair, moderated scientific debate. They cannot prevail based on science. Thus, their ad hominem attacks.
My apologies if I mis-read your comment. But if so, please explain it better.

Peter Miller

“would have been more scientifically adult”
The only thing ‘adult’ about the IPCC is one of Pachauri’s dodgy novels.

One question, one suggestion.
Suggestion first — it is highly instructive to plot the actual temperature by adding the “anomaly” to the supposed baseline temperature the anomaly is the anomaly of. That takes care of the “45 degree slope” problem once and forever. At typical high resolution plot scales, the entire 150 year warming is around one pixel high, so that the temperature graph is basically a straight line, with barely visible weakly trended noise. This eliminates a fraud straight out of the book “how to lie with statistics” — plotting the anomaly or difference instead of the actual quantity in question.
Although it is difficult to see, it is worthwhile illustrating the uncertainty that NASA itself acknowledges in the baseline temperature one is adding to the supposedly more accurate anomaly (where we could spend a few happy hours talking about whether one generally knows the mean of a distribution — this baseline — or its variance — the anomaly — more accurately, where I would have said that almost without exception it is the former known more, not less, accurately than the latter). This uncertainty is roughly 1 degree Kelvin today — anywhere from 13 to 15 C depending on the model used to estimate the actual mean. This makes the horizontal error bar on the baseline itself (on the absolute scale) roughly twice the size of the entire range of variation of the anomaly.
So do we know the variation of male height from the mean height of all males more accurately than we know the actual mean height of all males (from any given sample of measured heights of males randomly selected out of a crowd)? I don’t think so.
The question — that I have not had time to look at in the “leaked” report — is what they do with their statistical claims, the ones that were so horribly botched in AR4. In particular, do they persist in using language such as “likely”, “extremely likely”, and “possible” to describe e.g. probable warming in the future (or worse, give actual probability percentages for future expected warming given various warming scenarios)? If so, are those claims based on e.g. averages over sets of predictive model results from different general circulation models (GCMs)? Do the statistical ranges come from the assumption that these models somehow form independent and identically distributed samples drawn from some sort of distribution of GCMs that might enable them to use the Central Limit Theorem to make sound assertions of how probable it is that some sort of mean behavior is within some given range of the true, observed, or even expected behavior for any quantity those models might produce including global average surface temperature?
Because if they do, even implicitly, this goes beyond mere lying with statistics in the sense of presenting results that are accurately computed in a misleading way, it is an horrible error and abuse of statistics itself to promote any such conclusions.
The correct use of statistics in the general arena in which the GCMs live is to apply it to the models, one at a time, to determine if we can reject the null hypothesis: “This model is correct, given the observed climate since the model prediction was made” and if so, at what confidence level the rejection can be made. GCMs that fail to make a reasonable cut — for example, GCMs for which it is less than 5% likely that temperature trajectories as extreme as the actual, observed temperature series occur given the usual Monte Carlo spread in small perturbations of the model’s initial conditions and parameter set — should be summarily rejected and their results completely omitted from the report altogether. And by “as extreme”, I don’t mean trajectories that rarely dip down to overlap with the observed trajectory, I mean trajectories that vary over much the same range and with the same average trend.
This criterion would, I am confident, reject well over 90% of the GCMs that are currently being used as the basis for false statistical claims of probable future warming. Even after such a step, the mean of the remaining models could not be taken as being predictive of future climate behavior, but there would at least be a chance that one or more of the remaining models could be approximately correct, perhaps not in any sense we could ascribe a particular probability to at this time, but to the point where at least common sense tells us not to reject its predictions out of hand.
IMO, there is no possible way that one can make an assertion such as “over half of the warming observed post 1950 can be ascribed to human activity, e.g. increased CO_2” based on the data itself. Dick Lindzen’s graph above already suffices to show, even over the tiny time interval plotted, that late 20th century warming is not “unprecedented” or impossible to explain with natural variation, it almost exactly matches the non-anthropogenic warming of the first half of the 20th century. Only by comparing models that predict much less warming without CO_2 and much more warming with CO_2 can one make any such claim at all, let alone claim that it is somehow “likely”.
But when the claim is made on the basis of an ensemble of similar results, all of them obtained from models that fail the hypothesis test above, how then can we give it even a shred of predictive confidence? Am I certain that most of the planetary warming has been manmade because failed numerical models tell me that it is, at the same time they are telling me that the current climate anomaly should be 0.5C or thereabouts instead of the measured value of 0.15C or thereabouts?
The mean of a bunch of incorrect models is not a correct model. The standard deviation of the predictions of a bunch of incorrect models is not a predictor of how close the real climate is expected to get to the mean — quite the opposite! As it shrinks (widening the gap between model and reality in terms of the statistical variation) all it does is make us more certain that the model is wrong!
I very much fear that AR5 will attempt, one final time, to perpetuate this sort of completely incorrect use of statistics, and in doing so will set actual climate science back another five to ten years. There is nothing that refines the scientific mind like having one’s pet model rejected from a major scientific review for cause, because it just doesn’t work! Suddenly one is motivated to go back to the drawing board or else — the typical or else being loss of funding, failure to publish in the future, failure to get tenure, students who aren’t interested in working for you — or else you commit professional suicide, in other words.
The way things stand now, there seem to be no restrictions on what gets included in the construction of the final “average” predicted trend for the climate. Don’t want to hurt anybody’s feelings, after all. If your model predicted a whole degree C of warming over the last 15 years, who cares? It’s still in the hat, just as likely or unlikely as anybody else’s to be right, still weighted the same as somebody who might have actually written a model that WORKS and got most of “the pause” correct.
And that’s just stupid.


Silver Ralph- You are talking nonsense. Paul was no Nazorean!! He was not a Jew either. He was a member of Herod’s own large and rapacious family, they were Arabs who were pronounced Jews by fiat of the Romans and their High Priest puppets. This massively unpopular appointment and the Herodian penchant for ‘first cousin marriage’ in fact helped initiate the revolution against Roman rule in 66 AD. Nazoreans, the word comes from nezer- to separate, were like Jesus’ cousin John the Baptist, were sworn to poverty by Nazirite vows. They were part of the community that wrote the Dead Sea Scrolls. James the brother of Jesus, was one such and Paul, termite-like tried to undermine him in the hierarchy of the Jerusalem Church. Indeed, Acts is replete with snide remarks about those in Jerusalem, the chief enemy of course, was James the Just, Jesus’ brother ‘in the flesh’ as Paul writes in Acts. It seems clear that both Paul and Josephus
spent time among the Nazirites of Qumram and knew both James and John. They had to lie about this association to save their own hides, after their serial betrayals of friends and fellow travelers. Josephus especially. Paul was rich and well connected and used his Roman citizenship to get him out of trouble, often. Finally, even hint that a hermaphrodite was worshiped by these people is ludicrous. The sexual climate extremely fraught with political issues at the time and fooling around was *not acceptable behavior* and could get you killed.This stuff is OT but i couldn’t let this gross misstatement stand.

Dodgy statistics and IPCC Assessment Reports

Good job Christopher.
It is dodgy scientists. Statistics is just a neutral tool that they utilize to intentionally mislead.


Very well written piece!

Silver Ralph says:
September 17, 2013 at 12:40 pm
Let us end with a God’s-eye view of the planet He has given us.
I think you will find that your singular deity has to be neuter, not male.
I wonder if that’s intentionally written at this moment because the author is trying to advantage others’ sense of decorum? Surely, this has nothing to do with what Monckton was trying to communicate. Still, I’m given to understand that Christ called Him, Father.
Graph on the left is the most recent one. But, Christopher, you’re spot on about the chicanery the crazies are doing with artificially decreasing our past temps.
But, while we’re on that subject, I thought I’d share something I think is hilarious. A while back, the alarmists were trying to convince people of the power of CO2. At the time, I put several graphs together (using WFT) demonstrating very lengthy periods of time the earth hadn’t warmed, in spite of increasing aCO2! For instance, I did this one 1944-1977 It actually showed slight cooling. But, because of the lunatics’ fanatical obsession to make temps match their beliefs, it now shows warming. And, they’ve done this with all their data sets. In other words, they had cover for this recent pause, but, they erased it away! Clowns!


James Sexton says:
September 17, 2013 at 1:56 pm
In other words, they had cover for this recent pause, but, they erased it away! Clowns!

Old England

Here in the UK we are currently running at around 12-13 deg C in Southern England (in the countryside but not in the towns) – that’s about 5 degC to 6 degC BELOW normal September temperatures. Roughly a normal late October early November temperature. Suggests a long cold winter is a real possibility and the IPCC and their mouthpiece, the BBC, may find it a little difficult persuading the public that this ultra-cold is in fact warming.

Pamela Gray

DavidG what praytell is your background? I find your post fascinating and speaks of why I find the books so endlessly worth investigating.

I always enjoy Lord Monckton skewering the kebabs of falsity. Just posted this God here is neuter and theology trumps (dodgy) science.

I do hope that Professor Brown’s thoughtful and excellent comment will be elevated to a head posting in its own right.


Building on part of the above analysis, there are three quite interesting tests which could be carried out (e.g.) monthly using the “main” surface temperature series:
Starting “now” and working backwards:
1. What is the first month in a least squares fitted line with zero gradient: i.e. how many months show no gradient whatsoever?
2. What is the first month in a least squares line with a positive gradient, but is just within the bounds of a 95% confidence interval: i.e. how many months show no statistically significant warming, despite there being insignificant trend?
3. What is the first month in a least squares line (if one exists) with a statistically significant negative trend at the 95% level.
The third might not always produce an answer if the data is level, but it could be quite a tantalising measure if it does.
An appealing feature of the first two measures is that the first month could move backwards in time, if there are months added “now” which are either level with- or lower than very recent history.


RGB, as ever, shines his light into the dark spaces between Science and Stupidity. I agree with Lord Christopher; another headline post for the Duke of Pearls is in the offing!

Keith Minto

That second graph on the right at the top explains it clearly. A full cycle sine wave will have a zero trend, but if you only select two uplegs and one downleg, the trend must rise.This illustrates the lie in starting a temperature graph from a cool period, it is not a trend but part of a cycle.
Can we start the temperature graph from the MWP ?


“why is the world spending trillions on attempts to reduce CO2 emissions?”
“The world” doesn’t spend, hasn’t been spending. Rapacious politicians, OTOH…

Figures often beguile me, particularly when I have the arranging of them myself; in which case the remark attributed to Disraeli would often apply with justice and force: “There are three kinds of lies: lies, damned lies and statistics.”
– Autobiography of Mark Twain
for more info on the background of this quote, see Stephen Goranson’s post to the Mark Twain Forum 31 July 2002


If prof Brown provided a post in the future I hope he would include a good clear anomoly graph of world temp for the last 140 years ? and remember there are a lot ordinary folk who would like to understand the post as well.
At least provide a summary paragraph or two in lay person’s language. Thank you.

Werner Brozek

Jordan says:
September 17, 2013 at 2:38 pm
Here you are for 1 and 2 on several data sets. #3 does not exist yet.
On all data sets below, the different times for a slope that is at least very slightly negative ranges from 8 years and 11 months to 16 years and 10 months.
1. For GISS, the slope is flat since March 2001 or 12 years, 6 months. (goes to August)
2. For Hadcrut3, the slope is flat since April 1997 or 16 years, 4 months. (goes to July)
3. For a combination of GISS, Hadcrut3, UAH and RSS, the slope is flat since December 2000 or 12 years, 8 months. (goes to July)
4. For Hadcrut4, the slope is flat since December 2000 or 12 years, 8 months. (goes to July)
5. For Hadsst3, the slope is flat since August 2000 or 13 years, 0 months. (goes to July)
6. For UAH, the slope is flat since October 2004 or 8 years, 11 months. (goes to August using version 5.5)
7. For RSS, the slope is flat since November 1996 or 16 years and 10 months. (goes to August) RSS is 202/204 or 99% of the way to Ben Santer’s 17 years.
For this analysis, data was retrieved from,1,4,48,92%5D
For UAH: Since June 1995: CI from -0.006 to 2.422
For RSS: Since January 1993: CI from -0.025 to 1.966
For Hadcrut4: Since July 1996: CI from -0.004 to 1.389
For Hadsst3: Since January 1994: CI from -0.005 to 1.736
For GISS: Since March 1997: CI from -0.005 to 1.306

Gordon in Vancouver

I used to see a lot of the graph with the HCru 5 year running average, that was supposed to tell us a great deal when the red line was moving up. And I used to see a lot of that graph which shows the 150, 100, 50 and 25 year trends, that was supposed to tell us a great deal when each slope became steeper as the time period became more recent. Now I don’t see much of either graph, I wonder why certain scientists changed their mind about how valuable that information is?

Rud Istvan

IPCC is in a world of hurt, right where it and it’s acolytes like UK Met belong. It isn’t just the mendacious SPM opening as exposed here. It is through all of AR5 SOD, which it appears will not be much altered in the FOD. SOD obscures the fact that CMIP5 models still produce roughly constant UTrH, when the majority of newer observations using three methods (corrected Radiosonde, GPS dither, satellite) show a negative lapse rate feedback such that UTsH increases, but substantially less than constant UTrH. That means, as Nic Lewis has shown, the models necessarily run hot absent overstated aerosol corrections. SOD posits significantly positive cloud feedback, expressly attributed to unknown causes since observations show it is neutral or negative. Ditto why themodels necessarily run hot. SOD slightly widens the lowest sensitivity range but leaves the implicit expected ECS value at 3, when a host of recent studies say it must be less than 2. And so on.
Sticking to the few major points noted here (misleading manipulation and representation of data, wrong on humidity, wrong on clouds, wrong on sensitivity) and whatever else major turns up (grey literature, anyone?) will be a more powerful way to get at them than endless snipping on minutia. Concentrate fire like Nelson did at Trafalger. Break their line in the middle, and take out the flagships first. To paraphrase Nelson’s final battle signals, the world expects every man ( and woman) will do their duty.

rgbatduke on September 17, 2013 at 1:39 pm
One question, one suggestion.
[. . .]

– – – – – – – – –
Your latest product is good. Really good.
Here’s to clarity in physical science with no allusions to anything outside of demonstrable / observable ‘is-ness’. Thanks.
Hope to see much more of you around here.

M’lord, The “scientists” are required to make the adjustments to data and present graphs in the professional manner as prescribed in the IPCC protocol. Using this protocol, I found another problem with CO2 – unconstrained growth in human height (see


DavidG says:
September 17, 2013 at 1:47 pm
As you’re no doubt aware, The Acts of the Apostles was supposedly written by Luke, not Paul.


Hmm… Not terribly sure what to make of this essay. It’s an accusatory piece, to be sure. Which, thankfully and surprisingly, makes use of pedagogy and demonstration. But it’s still just an accusatory piece. (Or polemic if you prefer the word.)
The basic thrust of the argument is: That people, if they are not are their toes, can be caught out by ‘leading’ or ‘suggestive’ presentations. The basic mode of the argument is: A leading and suggestive presentation.
I’ve got not one issue with stating that one can, and perhaps should, use the same form of argument as their opponent. Certainly, if the opponent pounds the table that the mode is satisfactory, then they can hardly make legitimate disagreement with the conclusion when your returned use of it leads them into absurdity.
But such things are limited to running dialogues, interactive or written down in context. It has pedagogy in it, but it’s presentation is such that it isn’t terribly useful for those that aren’t on the right side of their confirmation bias already.
For example, if one already knows to have a care in analyzing a visual presentation of non-visual data, then the ‘scientifically adult’ coloring of the last graph is simply undesirable. Such graphs are presented in high contrast to assist looking for variance in the data values. Indeed, you could re-frame the same argument as stating that an ‘adult’ display would be to present a graph, of only green colors, to the colorblind. They might be able to sort out the differences in the low contrast that they perceive, but they’d have rather a time of it. In this one example the underlying argument, unstated, is that the range of values underlying the color assignments is simply too small to be useful. And this is a similar case for the underlying assumptions about the second graph.
I think it inarguable that the ‘interesting’ data ranges in a graph are precisely those that should be labeled on the axes. That is, after all, the only part we’re interested in. To make the contrary argument, that they should be presented in ‘some other’ scale or range requires stating that the ‘interesting’ range is uninteresting because there is a ‘more interesting’ range that ‘should’ be used. And that’s a legitimate argument, but only if there is data in that range. One can hardly argue that a different range is ‘more interesting’ for the data when there is no data in it.
A truly bad argument about ranges, and terribly common, is that ‘people are easily deceived by graphs.’ Which is simply a polite manner to put the claim: “People are incompetent at evaluating data.” And if we assume that is true, then what is the argument?
Is it that presenting data to non-experts is equivalent to Fraud? Is it that elected politicians are to incompetent at evaluating data to permit the use or advise of Science when crafting policy? Or is it that the incompetency, confirmation bias, and certitude of Truth, are so pervasive in Science that we cannot but call Science a religion for legal purposes?
There are any infinite number of possibilities. But the one it cannot be is that data is invalid as such. (Feel free to adjust the definition of Data if it doesn’t support your theory; that is a best practice in Science after all.)