It's only been six months since his previous wrong-headed column claiming that global warming has ended, but Michael Duffy has decided to write another one:
Last month I witnessed something shocking. Rajendra Pachauri, chairman of the Intergovernmental Panel on Climate Change, was giving a talk at the University of NSW. The talk was accompanied by a slide presentation, and the most important graph showed average global temperatures. For the past decade it represented temperatures climbing sharply.
As this was shown on the screen, Pachauri told his large audience: "We're at a stage where warming is taking place at a much faster rate [than before]".
Now, this is completely wrong.
You can see a video of Pachauri's talk here (it's 17 minutes in). Here's the graph that Pachauri showed (from the IPCC FAQ:
To quote from the FAQ:
Note that for shorter recent periods, the slope is greater, indicating accelerated warming.
Pachauri's statement is correct and it is Duffy who is completely wrong.
Duffy continues with:
For most of the past seven years, those temperatures have actually been on a plateau. For the past year, there's been a sharp cooling. These are facts, not opinion: the major sources of these figures, such as the Hadley Centre in Britain, agree on what has happened, and you can check for yourself by going to their websites. Sure, interpretations of the significance of this halt in global warming vary greatly, but the facts are clear.
The facts are indeed clear. Let's check what the Hadley Centre in Britain agrees with:
**Anyone who thinks global warming has stopped has their head in the sand.
The evidence is clear -- the long-term trend is that global temperatures are rising, and humans are largely responsible for this rise. Global warming does not mean that each year will be warmer than the last. Natural phenomena will mean that some years will be much warmer and others cooler.
You only need to look at 1998 to see a record-breaking warm year caused by a very strong El Niño. In the last couple of years, the underlying warming is partially masked caused by a strong La Niña. Despite this, 11 of the last 13 years were the warmest ever recorded.
Average global temperatures are now some 0.75°C warmer than they were 100 years ago. Since the mid-1970s, the increase in temperature has averaged more than 0.15°C per decade.
This rate of change is very unusual in the context of past changes and much more rapid than the warming at the end of the last ice age.
The rise in global surface temperature has averaged more than 0.15°C per decade since the mid-1970s. Warming has been unprecedented in at least the last 50 years, and the 17 warmest years have all occurred in the last 20 years. This does not mean that next year will necessarily be warmer than last year, but the long-term trend is for rising temperatures. ...
Over the last ten years, global temperatures have warmed more slowly than the long-term trend. But this does not mean that global warming has slowed down or even stopped. It is entirely consistent with our understanding of natural fluctuations of the climate within a trend of continued long-term warming.
And look at a graph of temperatures from the Hadley Centre for this year:
Far from being a "sharp cooling", the trend is a warming rate of 19 3.5 degrees C per decade. Of course, this doesn't mean anything, because changes in temperature on monthly or annual scales are weather and not climate.
Duffy:
Still, there's no doubt a majority of climate scientists agree with the view of the IPCC.
Today I want to look at why this might be so: after all, such a state of affairs presents a challenge to sceptics such as me. If we're right, then an awful lot of scientists are wrong. How could this be?
On the other hand, the explanation of why Duffy and other members of the pundit class such as Andrew Bolt get it wrong is quite simple: they don't understand basic statistics and refuse to learn.
Anyway, Duffy's explanation as to why he is right about the science and the scientists are wrong is given by this Richard Lindzen paper: climate scientists are conducting a scare campaign in order to get research funds and/or because they are environmental activists. Michael Tobis comments:
Which brings us back to Lindzen's diatribe. I am hoping the bizarre snipe at Ray P[ierrehumbert] was ghostwritten. It's inexcusable. I am confident that Ray has never publicly addressed scientific matters other than scientifically, and is in no reasonable sense fanatical. At best, Lindzen "approved the message". As such he is very clearly participating in the degradation of scientific conversation he claims to be bemoaning. As far as I am concerned, this one gross misstatement colors the credibility of the entire article. Perhaps others will find other similarly grotesque mischaracterizations elsewhere.
Werner Aeschbach-Hertig also comments:
Another critique of Lindzen is that science has deteriorated by moving from theory to simulation and modeling. I completely disagree: Numerical solution of complex equations is simply the logical way forward once your basic theory is known (which is the case in atmospheric and ocean dynamics) but you want actual solutions of the equations. The reason that modeling has become so large is simply that ever better computers have made this possible, a tool that was not available to scientists in the good old times conjured by Lindzen. An example of the power of numerical modeling is weather prediction, where models using the same core physics as climate models have achieved tremendous progress.
Duffy credulously repeats Lindzen's claims about "environmental activists":
Lindzen believes another problem with climate science is that in America and Europe it is heavily colonised by environmental activists.
Here are just two examples that indicate the scale of the problem: the spokesman for the American Meteorological Society is a former staffer for Al Gore, and realclimate.org, probably the world's most authoritative alarmist web site, was started by a public relations firm serving environmental causes.
If Duffy had bothered to fact check Lindzen, he would have discovered that:
We wish to stress that although our domain is being hosted by Environmental Media Services, and our initial press release was organised for us by Fenton Communications, neither organization was in any way involved in the initial planning for RealClimate, and have never had any editorial or other control over content. Neither Fenton nor EMS has ever paid any contributor to RealClimate.org any money for any purpose at any time. Neither do they pay us expenses, buy our lunch or contract us to do research.
And Anthony Socci who is a Senior Science and Communications Fellow with the American Meteorological Society, was Senior Science Advisor for the U.S. Senate Commerce Committee's Subcommittee on Science, Technology, and Space from 1991 to 1993. For some of this time, Al Gore was the chair, but I don't see how any reasonable person would think that this proves that Socci is an environmental activist.
Duffy offers one more argument for his thesis that he is right and the scientists are wrong:
Someone else who's looked closely at scientific journals (although not specifically those dealing with climate science) is epidemiologist John Ioannidis of the Tufts University School of Medicine in Boston. He reached the surprising conclusion that most published research findings are proved false within five years of their publication. (Lest he be dismissed as some eccentric, I note that the Economist recently said Ioannidis has made his case "quite convincingly".)
I think that Ioannidis makes a reasonable case for a kind of "winner's curse" in publishing biomedical research, but it's clear that Duffy doesn't understand the argument since he thinks it applies to research that finds evidence for anthropogenic global warming. If we test many different treatments for disease, then, even if they are ineffective, by chance 2.5% will be beneficial at the 5% level. If, say, 2.5% really are effective, half of the treatments initially found to be effective will be found to ineffective in follow-up studies. But what Duffy fails to understand is that the evidence for AGW doesn't rely on a single paper. Scientists have done follow-up studies and follow-up to the follow-up studies and confirmed rather than refuted the initial findings.
Why might this be so? Later work by Ioannidis and colleagues suggests that these days journal editors are more likely to publish research that will make a splash than that which will not. They do this to sell more copies of their publications and of reprints of papers in it. Ioannidis believes these publication practices might be distorting science.
Which means that journals will be more likely to publish research that overturns the scientific consensus, since that is what would make the biggest splash...
More like this
These people have such a hard time understanding the concept of a meaningful trend versus very short term fluctuations.
Not that they are trying....
Well, fairfax may be trying to "sell more copies of their publications" too, but it's cost them at least one - their f*wit factor has exceeded my pain threshold. Vote with your feet!
Note that for shorter recent periods, the slope is greater, indicating accelerated warming.
What does that prove? For shorter non-recent periods the slope is greater and in the wrong direction. Can we conclude anything more than the fact that we can't rely too much on shorter periods?
Over the last ten years, global temperatures have warmed more slowly than the long-term trend. But this does not mean that global warming has slowed down or even stopped.
Neither does it mean that global warming will continue unabated. These short term trends don't mean a whole lot of anything until they become part of a longer term trend which is yet to be seen.
> "Far from being a "sharp cooling", the trend is a warming rate of 19 degrees C per decade."
19C?! A typo, I hope!
These hardcore deniers of scientific reality are a fascinating bunch. They've created a near-impenetrable armour of delusion, Dunning Kruger and conspiracy theory. That same combination trickles down to the rag tag denial troops on the interweb, making them equally impervious to reason and science.
Fortunately, I think the climate deniers are becoming marginalised now. They're starting to enjoy the same credibility as 911 troofers, 'cdesign proponentsists', etc. Even 'peanut galleries', like Digg, no longer suffer an overwhelming avalanche of anti-science with every climate story.
Fix bayonets! We've got 'em on the run, lads!
I have a letter pointing out very similar points published on the SMH website (not sure if it will be in the paper).
At the risk of appearing slightly naive, don't the big red line and big purple line represent, well I guess you might call it a long term trend?
I must be missing something. Pachauri:
Met. Office:
And looking at the Hadley's own graphs of the same data, we can clearly see the plateau. The accelerated trend depends on the time period chosen: the yellow line in the plot goes back 25 years. Start at a later time (as the Met. Office's pdf does), and you get a smaller slope.
Now, I'm not denying that climate change is occurring, but the evidence suggests that the increase in global temperatures has slowed down over the last few years. It's clear from the data that these sorts of fluctuations do happen: I don't know of any reason why the increase should be monotonic (I'm not a climate scientist, but it seems reasonable to expect fluctuations: life's like that). But to claim that this data shows the increase is accelerating seems to be relying on cherry-picking the analysis to get what you want, so I hope there's a deeper analysis behind the claim.
And now Andrew accuses Tim of cherry-picking and predicting 19 degree per year temperature rises, and proves it by cherry-picking from this post. Irony, ahoy!
Bob @ #7, the shortest period Pachauri refers to, as supported by the graph, is the last 25 years..as you are aware. He does not have a trend plotted for the last ten years; annual means are represented by black dots. But Duffy claims in black and white: "For the past decade it represented temperatures climbing sharply." It simply does not. Duffy cannot read a graph; this impedement to his credibility is not at all overcome by the rest of his labour!
The Global Mean Temperature graph includes four time series (periods): 25yrs (yellow), 50yrs (orange), 100yrs (blue) & 150yrs (red).
Michael Duffy asked why the last decade was not included as a time series - a reasonable question.
And being so profoundly intelligent your rebuttal graph cherry picks the last 7 months...
I think your response is emotional, petty and immature. Asking for a time series of seven to ten years is not an unreasonable request. You did not need to do a petulant dummy spit and childishly show a graph for the last seven months.
Arrogance and emotional immaturity are an ugly combination (unfortunately shared by too many scientists).
Via Denialism Blog, here's what Ioannidis thinks of denialists using his work:
Fair enough. A decade is not enough time for internal variability (unforced variability--things like el nino/la nina) to even out. 25 years is enough time.
Of course, I'm sure Duffy has been enlightened a hundred times and is unable or unwilling to learn.
"But to claim that this data shows the increase is accelerating seems to be relying on cherry-picking the analysis"
Bob O'H, do you even know what cherry-picking means?
To answer the question of if the 25 year span was "cherry-picked" you look at the trends you get by using an interval of about 25 years ending about now, and look at the average of these trends. If (as in this case) the original trend is not significantly different than the average of these trends, THEN THERE WAS NO CHERRRY-PICKING BY DEFINITION.
Cherry picking is by definition choosing an interval specifically because it shows a trend which most other intervals close to it don't show.
If you don't know what a word means then don't use it.
Andrewr @ 10,
My understanding is that the last decade was not given as a separate series because 30 years or longer is the standard time period in which meaningful long-term trends (climate), as opposed to short term fluctuations (weather) can be seen. I am a bit suprised Pachauri used 25 years rather than 30 years, actually. Starting a series 10 years ago would mean starting at the peak of the 1998 El Nino (the black dot high in the top right corner) which would obviously distort the long term trend. The statement from Hadley is quite clear that the reason the last decade has been relatively flat is that we have gone from an El Nino peak to a La Nina trough (Actually, can someone enlighten me on this - how does El Nino/La Nina warm/cool the planetary average? is the heat being transferred from/to the deep ocean where we don't sense it?).
Duffy also states that "for the past year there has been a sharp cooling". By showing a graph of the past year, Tim is directly rebutting this point, not "dummy spitting" (Although that 19 looks like a typo to me).
If you want to complain about emotional, petty and immature people, I suggest you hop over to Bolt's blog.
You forgot one thing radicals ,it is impossible for any model or anyone to predict the future climate or anything else in 100 yrs ,if you dont agree then simply produce a quantified solution to collect your nobel prize for doing the impossible ,if not then stop the lies and scam to get our taxes ,its worth noting the radicals have not been right in 50 yrs and they were started by the nazis in 1934 as a properganda group against the west,c02 has never been proven to control the weather either .its all lies without quantified proof .
'melanie' : Trolling works better if you wait until the second or third response to break out Hitler and/or Nazis. Also, I like my trolling to have sentences please! Thanks a bunch.
Yeah, and are we radicals or did we forget radicals? Thanks for the clarification, melanie.
Gee, melanie, what was that about?
Melanie, have you considered drinking less coffee?
@Bob O'H
I completely agree with you.
Tim Lambert writes quoting Dr. :
re: #14 James Houghton
Good comments. Appreciate your point about 30yr time series for meaningful long term climatic trends...
It has been a while since I had to draw a line of best fit on a graph, but i am sure that a line that intersects only 2 points, with 2 points above it and 5 points below it is not classed as a good example. My year 7 maths teacher would be shaking her head. If we are only going to hit two points, i could easily draw one that goes down from the from month 3 to month 6 and have 3 points on one side and 4 points, showing a sharp cooling trend.
Timedw, you're not making much sense. Goodness of fit has little or nothing to do with how many points intersect the fit (heck, you can have an excellent fit with zero points intersecting the line) nor with how many points lie above or below. The fit is created in a precise mathematical computation that is straightforward. Are you saying that the data doesn't represent a linear trend?
"Pachauri's statement is correct and it is Duffy who is completely wrong."
I don't think so. If you have a short period of accelerated warming, then the more of the past that you throw away when you cart it, the more accelerated your trend line looks. But by throwing away more of the past, you are showing nothing except that you had 20 years of rapid warming. It's simply not very meaningful because it does nothing to tell you that the 20 years will continue. Especially since that 20 years was followed by 11 years of no warming. There is no evidence for anything coming from the period of 1977 to 1997 except that we had a lot of El Ninos over that time period.
Timedw:
[Curve fitting](http://en.wikipedia.org/wiki/Curve_fitting)
[Least Squares](http://en.wikipedia.org/wiki/Total_least_squares)
Think before you spout.
I just went to this site and ran the correlation:
http://faculty.vassar.edu/lowry/corr_big.html
The rate of warming given by the slope is
0.029233*1200 = 35.08 degrees per century. With only 9 points, and the given variability, the 99% confidence interval is somewhere between
-.012* 1200 = -14.4 C per century and
z=.0704*1200= +84.5 C per century.
Tmedw may not know the theoretical background for computing slopes, but he's right that the given 9 months don't have enough data points to give meaningful results, given the variability in the monthly data.
With only 9 points, it's virtually impossible to exclude
data as biased, but there was a grain of truth in Timedw's objection regarding the number of runs. If the number of runs is significantly different than the median, there may be a strong non-linear relationship involved.
Check out the
"Swed-Eisenhart" runs test. Here's one link:
http://books.google.com/books?id=brQ9AAAAIAAJ&pg=PA122&lpg=PA122&dq=swe…
If the number of runs is significantly less than the median, say at the 5% or 1% level, you may have a parabolic or exponential trend rather than a linear trend.
Likewise, if the number of runs is significantly more than the median, you may have some sort of "sine" or negative feedback function involved.
Sorry, part of my post was deleted.
The median slope was 35.08 C per century with a
99% confidence interval between - 14.4 C and +84.5 C per century
Yes I do.
And then you ask "why 25 years"? Why not, say, 10 years? Or why not look at the smoothed graph? Does anyone really think that temperature change is linear over time? If you don't then I hope you see that drawing straight lines through 25 years of data is going to be problematic: in this case it covers up (literally - see below) the slowing-down that is seen recently.
Nick - The plot does suggest a rapid rise over the last 10 years: that's what the yellow line is showing. Because it's so thick, it obscures the shorter trend, over the last 10 years or so. When presented with that graph in a talk, most people simply won't see the last few years of data, because it's covered by teh thick yellow line.
Any idea how tim gets the 19 deg per decade? Even taking that he started his series at the coolest month in 10 years, I still can't see how roughly 0.25 degrees over 9 months gets to 19 degrees a decade?
19 deg per decade? Nice one...
Oops, copied the intercept instead of the slope from the regression results. The correct slope is 3.5K per decade.
@Bob, Nick and elspy:
Maybe it is helpful for the discussion to bear in mind what the IPCC actually said in the
--> "FAQ 3.1" footnote:
"Figure 1. (Top) Annual global mean observed temperatures1 (black dots) along with simple fits to the data. The left hand axis shows anomalies relative to the 1961 to 1990 average and the right hand axis shows the estimated actual temperature (°C). Linear trend fits to the last 25 (yellow), 50 (orange), 100 (purple) and 150 years (red) are shown, and correspond to 1981 to 2005, 1956 to 2005, 1906 to 2005, and 1856 to 2005, respectively. Note that for shorter recent periods, the slope is greater, indicating accelerated warming. The blue curve is a smoothed depiction to capture the decadal variations. To give an idea of whether the fluctuations are meaningful, decadal 5 to 95% (light blue) error ranges about that line are given (accordingly, annual values do exceed those limits). Results from climate models driven by estimated radiative forcings for the 20th century (Chapter 9) suggest that there was little change prior to about 1915, and that a substantial fraction of the early 20th-century change was contributed by naturally occurring influences including solar radiation changes, volcanism and natural variability. From about 1940 to 1970 the increasing industrialisation following World War II increased pollution in the Northern Hemisphere, contributing to cooling, and increases in carbon dioxide and other greenhouse gases dominate the observed warming after the mid-1970s. (Bottom) Patterns of linear global temperature trends from 1979 to 2005 estimated at the surface (left), and for the troposphere (right) from the surface to about 10 km altitude, from satellite records. Grey areas indicate incomplete data. Note the more spatially uniform warming in the satellite tropospheric record while the surface temperature changes more clearly relate to land and ocean. 1 From the HadCRUT3 data set".
1) First of all: The periods chosen were not the idea of Pachauri, but confirming IPCC AR4. So the data were indeed not cherry picked by Pachauri.
2) However the way I see it, the IPCC claim when comparing the 150, 100, 50, 25 years should be balanced out by adding that the 150 and 100 periods both started with a cooling (thus resulting in less warming) and the past 50 and 25 years started with a warming. There was accelerated warming from 1981-2005 (or 1978-2005) alright, but a similar accelerated warming occured 1910-1940. The latter was disguised because of the period chosen (not picked, smile).
3)From the above IPCC-link:
Bob O'H:
The yellow line doesn't obscure anything, nor does it show rapid warming over the last 10 years.
what it does show is a warming trend averaged over the past 30 years, but if you look at the handful of dots for the year 2000 onwards, then you can see that they are all around the same level and no 10 year warming trend from 1998 onwarnds can be seen.
However, what the graph as a whole - from the 1800s to present - shows, is how silly it is to take 10 years of data and say that there is an *important* cooling trend. You can see lots of 10 year cooling trends in this data e.g. 1850-1860, 1940-1950.
But that that time scale looks kind of noisy to be picking out a long term trend doesn't it?
Longer term graphs might say something more again, but that is a whole other argument, best left for another thread.
I love how Bolt likes to paste up the satellite data as some kind of gotcha to back his case, even though it shows a warming trend over its duration. He has a habit of posting graphs that show the exact opposite of what he is trying to argue.
(a) the last 10 years is part of the last 30 years. So what it shows for the last 30 years it also shows for the last 10 years. It's a straight line, so what it shows is the same.
(b) How carefully do you have to look at the graph to see what happened over the last few years? You have to look carefully: it's squashed up at one end with a big thick yellow line on top. The line does obscure the pattern, and in a talk (which is where Pachauri presented this) few people will look that carefully.
OK, so it's someone else's error, or his for not looking at this critically. He said:
which may have been true a few years ago, but as you wrote:
and I fully concur with this. We're at a stage now where there is little or no warming - the data shows this. I assume there are good explanations for this (I'm not a denialist, and I've seen more than enough data to be happy that this is just a short term fluctuation), but Pachauri's comment was, at the very least, ill-advised. It's based on an awful approach to model fitting. I think the longer-term trends speak for themselves, and there's enough evidence to suggest we're not starting a period of stasis (or that we're about to enter a period of lowering of temperatures). Why leave yourself open to attack, when you don't have to?
Two points missed in all of this: 1. In much of Europe and in many higher latitudes, the trend is linear and sharply upawards (meaning that regional increases in temperature are highly significant over the past 30 years); 2. According to NASA, October was the warmest October since records began and easily ranks in the top five for all monthy records.
Bob writes: "We're at a stage now where there is little or no warming - the data shows this".
Incorrect. It does not show this. As you are a biometrician and biologist, Bob, I'd expect you would know better than to make such a statement. Your statement appears to reflect a poor understanding of temporal stochasicity versus determinism. For a hugely deterministic system, like climate control at the global scale, one cannot extrapolate over ten years and draw conclusions. As a fellow scientist, I certainly would have trouble doing this in a mesochosm experiment, let along an experiment incorporating every system across the biosphere. One would have to break the system down into regional systems at this scale, and even this would be risky. Regional systems do, however, show a significant upward warming trend (as, for example, here in western Europe). Thirty years would have to be a minimuim baseline and even then we are pushing things. I'd never boldly state that we could fully understand biogeochemical and hydrological cycles across the biosphere in time scales that the sceptics are now using to downplay warming.
I'd like you to define what you mean by 'now'. This tends to have an anthropocentric ring to it.
Pachauri is a railway engineer turned economist, not a scientist. NGO's are heavily involved in the IPCC process, Greenpeace's International Campaign Director for many years, Bill Hare, is a "guest scientist" at the Potsdam Institute in Germany and also a lead author for the 4th assessment. Al Gore's favourite scientist after James Hansen and a former IPCC chairman, Bob Watson, is now in charge of Britain's climate agenda. He recently came out with the 4 degree scare which has zoomed around the world.
The real science deniers are the one's who still push the agenda in spite of current lack of warming against still rising CO2, historical knowledge of warmer climate in the MWP and denial of the facts that all that has really happened is a recovery from the terrible times of the Little Ice Age.
Bob, I think you're blaming the victim here. Pachauri may have left himself "open to attack" because he has better things to do than to anticipate the ill-intentioned pseudo-scrutiny of Duffy. Life's too short.
Who is Michael Duffy?!!?
BTW i'm a Brit and i'm not really looking for an answer, heh, heh...
Maybe Jeff Harvey should look again at those NASAGISS results for October.
"How does El Nino/La Nina warm/cool the planetary average"
Ignoring the oxymoron, a paper by White et al addresses this very issue;
http://meteora.ucsd.edu/papers/auad/Global_Warm_ENSO.pdf
The abstract states "Thus global warming and cooling during Earth's internal mode of interannual climate variability (ENSO) arise from fluctuations in the global hydrological balance, not the global radiation balance."
The Tsonis paper looks specifically at one such transition period in 1976;
http://www.agu.org/pubs/crossref/2007/2007GL030288.shtml
This 1976 event, which effectively saw off the period of cooling over the preceding 30 years, has been noted by various other papers by Guildersen and Schrag, McPhaden and Zhang, Wijffels and Myers and McLean and Quirk who associate the partial cessation of deep water upwelling in the Pacific and Indian oceans with a marked and abrupt increase, or 'step-up' in temperature. Others such as Cai et al have associated the variation of the IPO with this event, while McPhaden and Zhang have noted that the partial cessation which began in 1976 continued until the end of 1998 when the trend reversed. Bob Tisdale has done a series of graphs and analysis on the step effect of El Nino and notes another one in 1998 during the super El Nino; there was also another one in the early '80's. Arguably, therefore, it is possible to claim that temperature movements during the 20thC have been a product of ENSO phase shifts which either cause or are caused by shoaling of the ocean thermocline; this is not an energy build-up but rather a periodic repositioning of that energy whereby ocean zonal retention and realease has profound effects on SST and atmospheric temp; a useful study would be of deep ocean temp during these upwelling variations; presumably during an upwelling cessation/-ve PDO period the thermocline should steepen as the difference between the SST and deep water temperature increases.
Jeff, #36. I notice from other sites GISTEMP is using Sept figures for Oct in Siberian towns such as Olenek. I think GISTEMP is wrong for October. It seems there was a lot of cold weather events in October.
Olenek was 3.1 deg according to GISTEMP, so was Sept and this is above the Arctic circle where October is generally a lot colder as the sun hours decrease at about 10 mins per day.
The White paper won't link; it's citation is;
White and Cayan, Dettinger, Auad; "Sources of global warming in upper ocean temperature during El Nino"; Journal Of Geophysical Research, VOL. 106, NO C3, PP4349-4367, March 15, 2001
Bob O'H -- the World Meteorological Organization (WMO) defines climate as mean regional or global weather over a period of 30 years or more. Ten years mean nothing; it is too small a sample size to be statistically significant. Pachauri was right and Duffy was wrong.
harbinger writes:
Wow, three errors in one run-on sentence.
1. We are still warming. See:
http://www.geocities.com/bpl1960/Ball.html
http://www.geocities.com/bpl1960/Reber.html
2. Climate in the Medieval Warm Period was not warmer than now.
3. What is the mechanism behind "recovery from... the Little Ice Age?" Do you think of the climate system as a sort of spring which, if pulled or pushed out of shape, automatically tends to move in the other direction when released? It doesn't really work that way.
Sorry, that should have been 1, 2 and 3, of course, not 1, 1 and 2.
#38
Amusing how the Bush administration chose to back a railway engineer turned economist over a real scientist.
Jeff Harvey -
Well "now" normally means now, not 5 years ago. I agree that we should be careful with short-term trends, but that means we should be careful about fitting straight lines to wiggle data, which is what is shown in the graph. For the last 5 years or so the 30-year trend has been decreasing, not increasing as suggested.
I think there's a lot of truth in this. I guess Pachauri was not as careful as he might have been (this was a talk, after all). But on this specific point I think Duffy is right - over the last few years there has been little or no increase. Now, one could argue that Pachauri and Duffy are talking about slightly different things, and that Pachauri's statement should be clarified, but I don't think it's right to argue that Duffy was "completely wrong" when he wrote
Or if you are, then show he's wrong by showing us the data. The first sentence certainly looks to be true (take a look). The second seems to be from the graph, but I'm not sure what that data covers, and one year is definitely in the "weather not climate" category (so even if it's right, it's not very relevant).
I am curious. Of the people who are now insisting that, based on the previous 5 (or 10) years of data, temperatures are steady or decreasing, how many were writing on blogs or elsewhere in 1998 claiming that global temperatures are increasing at an impressive rate, based on the previous 5 years of data?
If I recall correctly, those people were saying that the satellites showed that temperatures were cooling, and you can't trust the surface record.
According to that graph if you draw a 25 year trend from 1915 - 1940 you get a similar increase in temp as the past 25 yet it was before we increased CO2 output.
According to that graph if you draw a 25 year trend from 1915 - 1940 you get a similar increase in temp as the past 25 yet it was before we increased CO2 output.
you are right.
if you cherry-pick both end points and the length of the time span, you can generate a SINGLE period that is somewhat similar to what we re seeing over the last couple of decades.
now please start looking for 7 year flats or 1 year downs (lol!)
and then start wondering, why these clowns are allowed to publish stuff on those newspaper sites....
"...journals will be more likely to publish research that overturns the scientific consensus, since that is what would make the biggest splash..."
The latest RealClimate post affirms this, discussing the paper from three years ago that claimed oceans were cooling, only to find an error with instrumentation:
http://www.realclimate.org/index.php/archives/2008/11/mountains-and-mol…
News for all:
Giss Nasa has just removed the "hot" October temperature with the "heatwave" in Russia, and the map on the gistemp site shows the September data again.
Bob #48,..."for the past 5 years ,the 30 year trend is decreasing, not increasing as suggested." No, the 30 year trend is_the_30_year_trend, not 5 years of a 30 year data set.
In this case,Pachauri has compared the 25 year trend-through 25 annual data points- with the 50 year trend, and observed, correctly, a difference in rate of increase. Duffy's mendacity is clear in his carefully worded opening paragraph:
"For the past decade, it[the graphic] represented temperatures climbing sharply."
It "represented" no such thing: it's not about the last decade, though the annual means are part of the total graphic, it's about comparing 25,50 75 and 100 year trends over 158 years of records!!! Duffy is talking about a different graphic...the one he wants to see.
"looking for 7 year flats or 1 year downs"
1978-1994
http://woodfortrees.org/plot/uah/from:1978/to:1994/trend/plot/uah/from:…
Here's GISS from 1940-1977
http://woodfortrees.org/plot/gistemp/from:1940/to:1977/trend/plot/giste…
The period from 1995-2008 needs to be looked at as well.
7 year flats or 1 year downs part 2;
1995-2000
http://woodfortrees.org/plot/uah/from:1995/to:2000/trend/plot/uah/from:…
2001-current
http://woodfortrees.org/plot/uah/from:2001/to:2009/trend/plot/uah/from:…
AGW is problematic until it can explain these fluctuations in temperature which have occurred during a period of sustained increase in CO2. I might add that Douglass and Christy and Lucia have removed ENSO and volcanoes from the mix and obtained very interesting results for temperature movements, presumably due to GHG's, for the period 1979-2008.
"Amusing how the Bush administration chose to back a railway engineer turned economist over a real scientist."
--------------------
#47
According to can-explain-all Al Gore, in NYT, "the [Bush] administration threw its weight behind the ''let's drag our feet'' candidate, Dr. Rajendra Pachauri of New Delhi, who is known for his virulent anti-American statements", Source
Yeah, no kidding, that's what fat Al thought of his co-nobelist back in 2002.
Now that Prophet Al can't play the blame game anymore with Bush ousted by Messiah Obama, we're living in interesting times to see how the green Khmers will walk the talk.
Hey, cohenite's got wood (fortrees)...but he still can't get it up. Cancel the shoot...
Khmer is an ethnicity you dolt. Try Bolshevik. Or hey, you could even try cracking a book, provided any survived the pyre that is.
Melanie's right. You free radicals are a cancer on society and aging our culture before its time. Fortunately, her sort of people found the solution - oxidizing reactants - around the time of the glorious Middle Ages. When the world was warmer than it is now, and there were unicorns.
> scientists are conducting a scare campaign ... because they are environmental activists
Well, obviously: anyone arguing that society ought to do X about Y is *by definition* an "Y activist".
Duh.
Oh dear, Nick seems to have succumbed to this recently diagnosed AGW syndrome;
http://www.ecoenquirer.com/global-warming-porn.htm
Nick @55 (responding to me at 48) -
Sorry, I should have been clearer. If we take the 30 year trend ending in (say) 2000, and compare it to the 30 year trend ending in 2007 (for example), then the latter has a lower slope. This suggests that the trend is decreasing (i.e. decelerating), so it is not faster than ever. Clear, I hope, and also clear that this is so for the last 5 years or so. What this means is that
isn't true for now - it was true a few years ago.
Huh? This makes no sense. If the graphic represents the trends, and the trends are straight lines and are increasing, and the trends include the last 10 years, how can you deny that they represent increasing temperatures over the last 10 years?
Bob. This is about Duffy's take on Pachauri. The shortest period Pachauri analyses is the 25 years up to and including 2007. The 25 year trend is established from 25 points,not the last ten ,or the last five ,or the first eight, or the middle thirteen ,or any other combination,but 25 consecutive years. Therefore, the yellow bar represents increasing temperatures over the sampled twenty five years, not "the last ten years". Pachauri is not commenting on the last ten years (as this is conventionally regarded as too short a sample)whether Duffy likes it or not.
Khmer is an ethnicity you dolt. Try Bolshevik.
#60
The Khmers Rouges are NOT an ethnicity, neither are the Green Khmers. It's a bunch of sick minds who dream of an egalitarian technology-free society. You ignorant or/and denier.
@Nick #55
No, Duffy is talking about the shaded area of 1996-2005 of the same graphic. You obviously did not read the footnote:
"To give an idea of whether the fluctuations are meaningful, decadal 5 to 95% (light blue) error ranges about that line are given (accordingly, annual values do exceed those limits)."
@65
There is a qualifier for it: lying by omission.
Nick - the last 10 years are part of the last 25 years, aren't they? So comments about the last 25 years include the last 10 years. No?
The trend line over the whole 25 years simply doesn't fit the last 5 years, so using the full trend to say something about now (where "now" covers the last few years) is nonsense, because the data itself is showing something different. basically, there's more going on than straight lines, so using straight lines to describe what's going on is misleading.
the last 10 years are part of the last 25 year
Your buttocks are also part of your body, Bob. I'm sure you don't consider yourself a complete arse though.
Hey demesure, Khmer
.
.
[listens to canned blather]
.
.
Um ok, now go away.
The trend line over the whole 25 years simply doesn't fit the last 5 years, so using the full trend to say something about now (where "now" covers the last few years) is nonsense, because the data itself is showing something different. basically, there's more going on than straight lines, so using straight lines to describe what's going on is misleading.
the most misleading thing in climate science are "5 year trend lines". simple fact.
Bob O'H writes:
That's a meaningless sentence. You can't find a 30-year trend with 5 years of data.
cohenite writes:
How many times have you been told, on other blogs, that CO2 is not the only thing that affects temperature? Yet you keep on bringing up the same straw man argument over and over and over again. Don't you have any shame?
Bob O'H writes:
Bob, you need to study an area of statistics called "linear regression." Then you'll understand why your sentence above doesn't make any sense.
Barton - sorry, but I have studied linear regression, I work as a statistician in a department of Mathematics and Statistics.
What we teach is that you should check the fit of a regression. An easy way of doing this is to plot the residuals. If the model is fitting well, then the residuals should be a random cloud. In this case they're not - the plot will look wiggly (curses, not even a simple frown or smile!), and will be decreasing at the end of the series. This is indicative of a poor fit of the model: there is obvious structure that it's not picking up. The advice in this case is don't trust the model: it's not fitting to the data well, so any inferences should be treated with care.
sod - I'd agree that a 5 year trend line isn't good, but at least it fits the data! My point (one of them, anyway) is that trend lines aren't very useful here, because the relationship in the data isn't a straight line.
(I'll ignore anthony - I think that falls under the category of "Prsnl tttcks [sc] n thr cmmntrs")
@Bob O'H
The Pachauri statement re accelerating warming over 150/100/50/25 years back is correct. Even if we take September or October 2008 instead of 2005 (IPCC) as the "now", the "25 year back" slope doesn't change significantly. You might like to play with the interactive graph as provided by cohenite under #56, just as I did.
But I agree with you if you mean the last five years (2003-2008) are not included in the last 10 or 25 years of the above graph (1981-2005). Smile.
.
This is a valid argument. But I think the answer is "not true". I don't have the data at hand right now, but by eyeball measure, using cohenite link, the slope 1978-2007 is steeper than the slope 1971-2000. If I recall correctly, the 30-year trend started to decelerate in 2006. It stands now at approx. +0.16°C per decade (HadCRUT3, September 2008).
Climatepatrol has a synoptical view of the unsmoothed temperature history 1979 - September 2008 representing five different datasets, if you are interested.
Yet you keep on bringing up the same straw man argument over and over and over again. Don't you have any shame?
They don't have any arguments. Or testable science.
Shame has nothing to do with it.
Best,
D
"What we teach is that you should check the fit of a regression. An easy way of doing this is to plot the residuals. If the model is fitting well, then the residuals should be a random cloud. "
YOU TEACH THAT ALL NOISE IS WHITE??????????
Do your students a favor and go read the posts on noise over at open mind. http://tamino.wordpress.com/
elspi - no I don't, don't put words into my mouth. If the noise is demonstrably coloured, then the model isn't fitting well.
Well, Barton Paul, I happen to agree with you, CO2 and temperature increase is a straw man; but it is not my straw man; AR4, Executive Summary, pp131-132 states that the combined radiative forcing of the GHG's is 2.63Wm-2; of that the RF for CO2 is 1.66Wm-2, or 63%; as well, the doubling of CO2 is supposed to translate into a temperature increase of approximately 4.5C; it is, therefore, legitimate to look at how temperature and CO2 have tracked during the 20thC; as my erotic graphs show, regardless of whether you have 40 year periods, 20 year, 10 year or 5 year sampling, there is no consistent correlation between CO2 and temperature. Your remarks about shame are disingenuous and relevant only if you regard AGW as a religion.
Dano; Miskolczi has had empirical verification; actually even if Dessler is correct about SH increase in the atmosphere this will not impact on the correlation between Miskolczi and HARTCODE; and if Dessler is correct, and that is a big if, then as far as I can tell that will be the first verification of AGW theory.
"If the noise is demonstrably coloured, then the model isn't fitting well."
I would loan you my backhoe but you seem to be digging the hole very nicely all by your self.
So if the noise is red then the model is wrong. WOW.
Once again, do your students a favor and meander over to open mind. He knows what he is talking about and maybe some of it will rub off on you.
Marion Delgado: I think you've hit on it; the reason for the medieval warm period was that people back then believed in oxidising free radicals (such as heretics, pagans and witches) in large numbers, thus increasing CO2 levels. I'm sure Energy and Environment would publish a paper to that effect if we wrote one.
I see. Would you care to explain that, or are you just trolling?
Bob O'H writes:
You're a STATISTICIAN and you were still capable of writing a sentence like "The trend line over the whole 25 years simply doesn't fit the last 5 years???" I find that very hard to believe. The trend in a 25-year sample is fitted to all the points, including the last 5. To take 5 years out of your sample and complain that the linear regression on 25 points doesn't fit the last 5 is just statistically illiterate. It's classic cherry-picking, something that will get you a flunking grade in any introductory data analysis class.
What university are you supposed to work at?
cohenite writes:
Oh, really? Someone empirically verified that water vapor goes down with increasing carbon dioxide? Not that I ever heard. Someone verified that the Earth's gray infrared optical thickness is fixed at 1.87? I haven't heard that, either -- except from the science illiterates who venerate Miskolczi.
If you use Miskolczi's model, you get results fairly close to reality for present-day Earth and Mars. It's spectacularly wrong for Venus and Titan, for early Earth and Mars (the Faint Young Sun problem), and if it's true we should not have had ice ages. The empirical evidence is all against Miskolczi's crackpot theory.
Barton - just because you can fit a straight line to a data set doesn't mean it's a good fit. Go here for some examples.
And where dis I say anything about taking out the last 5 years' data? Nowhere - that's your fabrication.
Bob O'H, let me see if I can't end the suffering here. The point is, I think and elspi can correct me if I'm wrong, that if you're interested in a long term trend (climate) and there appear to be factors in short term noise (weather), that this is not necessarily a problem. In fact, that could be anticipated and decidedly not a problem... (remembering that we come into this with a priori knowledge of both the relative magnitude of weather noise, the fact that this can extend to factors that persist over a few years- ENSO for example- and, relatedly, that we're probably not dealing with a system where the climate turns on a dime, short of ice sheet collapse or other cataclysm). In econometric terms, the quality of a model is not given by its ability to explain everything, only that which you seek to know.
It furthermore bears pointing out that we're not really talking about a model here. It's a trend. If I understand you right, you're essentially saying that a 30 year trend is not meaningful because of residual clustering at the end. Given that we care about 30 year trends because of what we know of the climate system in the first place, that is not convincing.
Majorajam - I'm saying that the 30 year trend is not meaningful because it assumes a linear change, and it's obvious from the data that the changes aren't linear.
We know changes in climate aren't linear, so why describe them as such? Why fit straight lines on top of wiggly graphs? It simply mis-represents the complexity that's there.
Bob O'H writes:
I take it you've done partial F-tests to see whether it's statistically justifiable to add nonlinear terms to the straight-line fit?
Bob, we know changes in climate are determined by forcings and feedbacks whose interaction can't be properly described by a regression equation full stop, let alone a linear one (although, the log of ghg emissions growth is vaguely linear). We fit long term trends mostly because if the climate were not warming it would mean that clearly scientists have missed something big in the physics of what accounts for the climate, and perhaps also that greenhouse gas forced warming may not be a problem. That would be great news of course, and no one would be happier to discover that kind of ignorance in climate science than I. Unfortunatley, since the earth is warming and considerably, despite the fact that their understanding of the climate is still far from complete, it is high time we started taking seriously what it is telling us.
Well done, Tim! Great critique!
Majorajam (#91 Nov 14) says:
"Bob, we know changes in climate are determined by
forcings and feedbacks whose interaction can't be
properly described by a regression equation full
stop, let alone a linear one.."
Quite. Many of the discussions of "trends" are really no more than disucssion of whether the global temperature is a linear function of the date.
This strikes me as a bit silly, given that the real question is whether the temperature is, and has consistently been, close enough to what climate models say it ought to be for us to have confidence in those models.
I'm pretty sure the models don't say we ought to be seeing a nice straight linear time-trend.
True, but I think scientists are just using linear trends as a quick way to summarize what the climate models are saying.