Alternate title: Something wonky this way comes
I try to get away to work on my paper and the climate world explodes, pulling me back in. Strange things are happening related to the BEST data and co-authors Richard Muller and Judith Curry. Implosion might be a good word.
Popcorn futures are soaring. BEST Co-author Judith Curry drops a bombshell:
Her comments, in an exclusive interview with The Mail on Sunday, seem certain to ignite a furious academic row. She said this affair had to be compared to the notorious ‘Climategate’ scandal two years ago.
Here’s the short timeline.
1. The GWPF plots a flat 10 year graph using BEST data:
2. The Mail on Sunday runs a scathing article comparing BEST’s data plotted by GWPF and the data presented in papers. They print this comparison graph:
Note: timescales don’t match on graphs above, 200 years/10 years. A bit naughty on the part of the Sunday Mail to put them together as many readers won’t notice.
3. Dr. Judith Curry, BEST co-author, turns on Muller, in the Mail on Sunday article citing “hide the decline”:
In Prof Curry’s view, two of the papers were not ready to be published, in part because they did not properly address the arguments of climate sceptics.
As for the graph disseminated to the media, she said: ‘This is “hide the decline” stuff. Our data show the pause, just as the other sets of data do. Muller is hiding the decline.
‘To say this is the end of scepticism is misleading, as is the statement that warming hasn’t paused. It is also misleading to say, as he has, that the issue of heat islands has been settled.’
Prof Muller said she was ‘out of the loop’. He added: ‘I wasn’t even sent the press release before it was issued.’
…
But although Prof Curry is the second named author of all four papers, Prof Muller failed to consult her before deciding to put them on the internet earlier this month, when the peer review process had barely started, and to issue a detailed press release at the same time.
He also briefed selected journalists individually. ‘It is not how I would have played it,’ Prof Curry said. ‘I was informed only when I got a group email. I think they have made errors and I distance myself from what they did.
‘It would have been smart to consult me.’ She said it was unfortunate that although the Journal of Geophysical Research had allowed Prof Muller to issue the papers, the reviewers were, under the journal’s policy, forbidden from public comment.
4. Ross McKittrick unloads:
Prof McKittrick added: ‘The fact is that many of the people who are in a position to provide informed criticism of this work are currently bound by confidentiality agreements.
‘For the Berkeley team to have chosen this particular moment to launch a major international publicity blitz is a highly unethical sabotage of the peer review process.’
5. According to BEST’s own data, Los Angeles is cooling, fast:
“barry says:
October 31, 2011 at 1:11 am”
Thank you for an excellent analysis!
I wish to comment on a few points you raised.
“There’s a couple of problems with the GWPF graph. It’s not 10 years (120 months), but runs from Jan 2001 to May 2010 (113 months). The time-series is already so short that interannual variability will have a big say in what the linear trend will be, but starting and ending at different times of the year could also introduce a seasonal bias.”
When I first read this, it made a lot of sense. However since we are dealing with anomalies and not absolute temperatures, does it make much difference in the end? I must confess I am not sure. Also see my earlier comment that relates to this at:
Werner Brozek says:
October 30, 2011 at 3:24 pm
“Anyway this is what I got for Jan 2001 to May 2010, with the ‘adjusted’ April 2010 – trend is 0.13C/decade.”
If it can be easily figured it out by you or someone else with better computer skills than mine, I would be curious to know what the HADCRUT3 slope is for Jan 2001 to May 2010 as well as for Jan 2001 to September 2011. I am sure the latter is lower, but I wonder if it is 0.13 lower. Since if it is, we can be reasonably confident that if BEST was right up to date, we would be able to make some reasonable conjectures.
@ur momisugly Elmer….. there are two reasons why they look different. One is the time frame, Paul goes back to 1981, the flat line graph goes back to 2001 or so, (I used 2001.75), the other is that Paul is using a 5 yr smoothing. Note, where it states “Mean (samples) and then has the value of “60”. These are monthly values, so 60 = 5 years. It should also be clarified, that to my knowledge, Dr. Curry didn’t use the flat line graph. Us skeptics did. And the Mail picked it up from somewhere.
One of the things you can take from Paul’s graph is that the BEST data is the outlier. Notice that BEST has the sharpest upward trend than anyone else. Another thing to note, is that BEST is land temps only. What it is interesting is that the land temps are warmer than the land/ocean temps.
I don’t use graphs going back that far because it distorts what is currently happening. Others will disagree with that approach, but I believe it is more valid than including events that occurred 30 years ago as somehow representative of what’s going on today.
barry:
At October 31, 2011 at 10:16 am you falsely assert:
“10 years is insufficient to determine a climate trend, which is what IPCC are projecting.”
Wrong! Absolutely not! Accept ‘A’ for trolling and ‘F’ for veracity.
If the data were showing a linear trend with variance around the trend line then the variance would determine the significance for any length of time. So, “10 years” – or any other period – is not what determines the statistical significance of a trend.
The truth is as Gail Combs explains at October 30, 2011 at 8:27 am. There are climate cycles so use of linear trends is misleading for any length of a climate time series.
Furthermore, you provide a falsehood when you write;
“The World Meteorological Organization lists a climate period as 30 years.”
No! The WMO does not! A ‘standard climate period’ is 30 years and is different from a ‘climate period’ which can be any length: e.g. the 1994 IPCC Report compares 4-year climate periods to assess a climate change.
Richard
Werner Brozek says:
October 31, 2011 at 12:22 pm
Werner, there’s no need. Just go here, http://www.woodfortrees.org/plot/hadcrut3vgl/from:2001/to/plot/hadcrut3vgl/from:2001/to:2010.42/trend/plot/hadcrut3vgl/from:2001/trend
Then, down in the lower right, there is a link labeled “raw data”. Click on it. If you scroll down, you’ll see this…… “Least squares trend line; slope = -0.00293773 per year”… that would be for the first trend line….. then scroll just a bit more, “Least squares trend line; slope = -0.00488513 per year”
Mind you, I did exactly what you asked for as far as giving people an idea about what we would expect the BEST trend to look like if the data set was complete…..http://suyts.wordpress.com/2011/10/31/more-funnies-from-the-bests-apologists/
Well, looks like Dr. Judith Curry is discovering what the rest of us have seen for a while. Political goals override good science and proper discipline on the AGW side of things. Best of luck to you, Judith. You will now likely find yourself the subject of some “kill the messenger” character assaults. Just know that there are folks of good character rooting for you.
At any rate, IMHO, it’s all just more pointless exercise in averaging the un-averagable. Where is the allowance for humidity impacts on heat? (Temperature is a lousy proxy for heat…) Where is the mass x specific heat x temperature? Where is the allowance for differences in total precipitation? Think all that 32 F snow in the North East doesn’t mean a whole lot more about HEAT loss than the same 32 F of air?
It really is a fools errand to spend so much time averaging temperatures. Yet “it’s what they do”…
(For those about to launch into a “but it’s anomalies” rant: Stow it. Look at the very first step. A daily high / low average to get a daily average. Then look at how the “monthly mean” is calculated. THAT is what is fed into GIStemp et. al. Not an “anomaly”. I won’t go into how long it takes GIStemp and others to get around to making an anomaly, but lets just say it comes long after averages of averages are calculated… The whole process is just stupid on the face of it as a method to say anything about heat flows. 32 F in the desert is treated the same as 32 F and 5 feet of snow. Just for starters…)
Septic Matthew says:
October 30, 2011 at 4:59 pm
——————————————–
“Clouds may … be … responsible for the negligible global warming of the past 12 years.”
That sentence comes from Dr. Richard Muller himself.
http://online.wsj.com/article/SB10001424052748703514404574588673072577680.html
Smokey says:
October 30, 2011 at 5:40 pm
…
I’m still waiting for you to provide any testable, empirical evidence – falsifiable, per the scientific method – directly connecting the anthropogenic rise in CO2 to any global harm. You are free to try and falsify my hypothesis:
CO2 is harmless and beneficial. More is better, at current and projected levels.
…
—————
Smokey, you know as well as I do it is not that simple. Your challenge is like a creationist saying, “prove the human eye evolved, with diagrams for every stage”. But, for the benefit of the open-minded, here is the mainstream response. AGW relies on a chain of reasoning, not just a few factoids. That chain of reasoning is this:
1. CO2 causes warming through the greenhouse effect
2. CO2 levels are increasing due to human emissions
3. Including feedback effects, the effect of doubling CO2 is thought to be between 2 and 4.5 C
4. Warming has been measured that, if viewed honestly, confirms the above
5. Other effects (Sun, ENSO, volcanoes, etc.) have been studied and found not to explain the warming, though they do make the signal noisy
6. The effect of this warming will cause undesirable effects such as rising sea levels, an increase in extreme weather events and others (MOST OF WHICH ARE YET TO HAPPEN)
Now, which of those steps do you object to?
stevo says:
October 30, 2011 at 6:20 pm
“Matt G: “Take it that you…can’t comment on whether a 10/12/17 year period is long enough or not…” don’t take any such thing. What period is long enough? It depends on the statistical test you apply, and what rate things are changing it. Only a fool would say 17 years is enough, 16 is not, or anything like that. But an example analysis which shows that 10 years is definitely not enough, while 20 years is probably OK, is here:
http://tamino.files.wordpress.com/2011/10/rates1.jpg
You didn’t understand my point about x-2001 trends being indistinguishable from x-2011 trends. Your links don’t contradict that. Look at this:
http://www.woodfortrees.org/plot/uah/from:1975/to:2001/trend/plot/uah/from:1975/to:2011/trend
Only a fool tries to draw conclusions from a decade of global temperature data. Don’t be that fool.”
REPLY
I did understand the point, firstly before I get to it, the comment you made referred to any x-2011 and any x-2001 that I showed was false.
Now with your point, it makes no difference what period I use because there seems an error in the WFT software. Whatever 30 year period I choose there is still viewing around 1980 until 2010 for the linear trend.
http://www.woodfortrees.org/plot/wti/from:1920/to:1950/trend/plot/wti/from:1940/to:1970/trend/plot/wti/from:1920/to:1950
Pick a different global data set and this does not occur.
http://www.woodfortrees.org/plot/hadcrut3gl/from:1920/to:1950/trend/plot/hadcrut3gl/from:1940/to:1970/trend/plot/hadcrut3gl/from:1920/to:1950
Neither does it show a trend warmer over a longer period up to 2011.
http://www.woodfortrees.org/plot/hadcrut3vgl/from:1975/to:2001/trend/plot/hadcrut3vgl/from:1975/to:2011/trend
Even if one does show a trend warmer over the longer period when there is no warming, it means nothing to what will happen ahead. Basically looks like typical behaviour in mathematics in any peak.
http://www.woodfortrees.org/plot/gistemp/from:1910/to:1935/trend/plot/gistemp/from:1910/to:1945/trend/plot/gistemp/from:1910/to:1965/plot/gistemp/to:1935/from:1945/trend
The graph above shows exactly the same thing as your 1975-2001 and 1975-2011 plots. Cooling followed for 30+ years, so what you show means nothing when it comes to climate.
Finally I can’t agree on how long a period is acceptable, but for example if you think 20 years is long enough, then you don’t believe the 17 year warming period which is wrongly blamed on humans is long enough either. Again just 17 years over a 77 year period, there would be naturally a warming phase expected at some time.
John B,
As usual you responded with a big ol’ strawman argument. My challenge to you was to try and falsify my hypothesis:
CO2 is harmless and beneficial. More is better, at current and projected levels.
Knowing you can’t [and it is a falsifiable hypothesis], you responded instead with various random thoughts and beliefs. Some I agree with [#1 & #2], some I disagree with [#4 & #5], and some are just evidence-free beliefs. In particular, #6 has been repeatedly falsified. If you had written, “(MOST OF WHICH
ARE YET TO HAPPENMIGHT HAPPEN IN THE DISTANT FUTURE, OR ON ANOTHER PLANET)”, then I wouldn’t object to it. But the way you wrote it demonstrates a preposterous belief system.@climatereason says:
October 31, 2011 at 2:17 am
///////////////////////////////////////////////////////////////////////////////////////////
Thanks Tony
Thank you.
CET is useful since it is the longest instrument record available, and is likely to have been compiled with care and thus as reliable as any data set available. That said, CET may of course not be typical of global temperatures as a whole perhaps because the temperatures recorded are driven by the temperature of the Atlantic. Oceans have a more tempered response but it is important to bear in mind that there can be no global warming unless the oceans are warming.
As I postulated, there is no correlation between temperature and CO2. Not even at a superficial level.
1. Superficially (whilst I dislike straight line fits), there is a gradual and steady 0.26degC per century temperature increase as from the late 1600s to date notwithstanding the exponential rise in CO2 levels running at least as from 1945.
2. As regards the pre-industrial periods, temperatures steadily rose during the period 1660 to 1840 even though there was no corresponding increase in CO2 levels.
3. As regards the post industrial period, temperatures were fairly flat between 1850 and 1875 notwithstanding rising CO2 levels.
4. Temperatures fell between 1865 and 1895 notwithstanding rising CO2 levels.
5. CO2 flat lined between 1910 and 1930 and yet temperatures continued to rise during this period.
6. The rate of temperature increase between 1925 and 1945 is broadly similar to the rate of temperature increase between 1985 to 2000 and yet the rate of increase in CO2 levels was significantly less during the former period.
7. Temperatures fell between 1948 and 1980 and yet CO2 levels was rising rapidly between this period.
8. Temperatures appear to have largely flat lined between 1995 to 2009 notwithstanding the steady and steep increase in CO2 levels.
I would defy anyone to put forward a reasoned case that CET temperatures are responding to CO2 levels. I would defy anyone to put forward a cogent and probative case establishing correlation between CO2 levels and CET temperatures.
I envisage that the same holds true to the data set compiled by BEST
MattG – had me worried there for a minute! The problem in your first graph is WTI only has data from 1979, which is when the satellite data which it is half based on starts. You’re trying to select ranges earlier than that, so it’s ignoring the ‘from’ and ‘to’, and you’re just getting repeated copies of the full WTI range 1979-present.
woodfortrees (Paul Clark) says:
October 31, 2011 at 2:43 pm
Thanks for quick responce and good to know that in a way, so it’s no good for any trend starting before 1979. I had no idea before WTI only had data from 1979 only, explains my orginal problem well.
Werner Brozek @ur momisugly here
I figured you might be more interested in the land-only data from HadCRU, to make apples to apples comparison with BEST data. The first trend (matching GWPF time period) is 0.04C/decade, and the latter (to September 2011) is -0.01C/decade.
http://www.woodfortrees.org/plot/crutem3vgl/from:2001/to:2010.42/trend/plot/crutem3vgl/from:2001/to:2011.67/trend/plot/crutem3vgl/from:2001
If you want the land+ocean analysis, just go to the link and change ‘CRUTEM3’ in each box to ‘HADCRUT3’ variance adjusted global mean.
Regarding full calendar years/seasonal cycles – I just followed the convention used by GISS and CRU. I assume they do it to avoid seasonal bias, but I’m no expert and could be wrong.
I think Huck Finn said
“There are three kinds of liars: Liars, Damn Liars, and Statisticians”
My version — there are three kinds of liars:
1. liars
2. outliers
3. and out-and-out liars
🙂
Thank you both to suyts and barry!
So let us assume: ““Anyway this is what I got for Jan 2001 to May 2010, with the ‘adjusted’ April 2010 – trend is 0.13C/decade.” So if we subtract 0.05 C/decade to get the amount BEST would probably get if we assumed Jan 2001 to September 2011, that comes to 0.08 C/decade or 0.8 C/century. It was quoted earlier we went up 0.8 C since 1750. At the rate of 0.8 C/century, it would take another 150 years to reach the 2.0 C increase. I do not believe a 2.0 C increase would be catastrophic, but that is a different discussion. But I do believe we can afford to just sit back and see what happens over the next few years before taking drastic measures to reduce CO2 should that be deemed necessary.
Steve from Rockwood says:
October 31, 2011 at 2:06 pm
Septic Matthew says:
October 30, 2011 at 4:59 pm
——————————————–
“Clouds may … be … responsible for the negligible global warming of the past 12 years.”
I mentioned this first before on WUWT many months ago, but not may, quite sure it has. But, also responsible for most of the warming too during the 17 year period.
The worst lie in statistics as well as science is extrapolation in a complex multivariable system, where some of the variables are either uncontrollable or poorly understood.
barry, make sure you use CRUTEM3 NH.
The January 2007 outlier was made up of 1484 SH temperatures and 14694 NH with 9 no latitude given according to Nick Stokes.
BEST seems to really just be Northern Hemisphere land database, which is perfect if you want claim the world is warming really fast and were not interested in reality.
Bruce, Barry.
I went through the Crutemp3 northern and southern hemisphere data – they are not that much different from eac other and they don’t match the Berkeley database with both being quite a bit lower (starting about 1990).
http://wattsupwiththat.com/2011/10/29/uh-oh-it-was-the-best-of-times-it-was-the-worst-of-times/#comment-783325
“EFS_Junior: I currently derive best-upper and best-lower as (value +/- uncertainty/2), but you think it should be just +/- uncertainty. I’m a bit surprised at this because I’m having a hard enough time matching the BEST graph as it is! Can you point me at why you came to that conclusion?”
It’s definitely the published mean +/- the published uncertainty.
I’ve plotted both the annual and decadal time series plus the uncertainty bounds, as published, about the mean, and by my eyeballs, both plots look identical to the same two plots as shown on the BEST website.
So unless something’s changed in either the BEST plots or their zipped dataset, since yesterday, it has to be the published mean +/- the published uncertainty.
If you still have any doubts, I can send you my Excel spreadsheet (BEST.xlsx).
Bruce,
What makes you think so?
Bill – yes BEST data shows greater warming in the past decades than the other sets.
>>
John B says:
October 31, 2011 at 2:07 pm
Now, which of those steps do you object to?
<<
I object to all of them.
>>
1. CO2 causes warming through the greenhouse effect
<<
The IPCC defines a value called the Global Warming Potential (GWP). The GWP of CO2 is “1”, because GWP is defined using CO2. That makes CO2 the weakest of all GHGs as the GWP of all other GHGs is greater than “1”. Notice that two of the most important GHGs are not given GWP values: water vapor and ozone. Ostensibly, it’s because they have short lifetimes in the atmosphere. Non IPCC estimates usually give these gases higher GWPs than CO2.
>>
2. CO2 levels are increasing due to human emissions
<<
As pointed out by others on this blog, ice cores show that CO2 lags temperature by about 800 years. Eight hundred years ago was the Medieval Warm Period. That would make the current rise in CO2 right on schedule.
>>
3. Including feedback effects, the effect of doubling CO2 is thought to be between 2 and 4.5 C
<<
It’s probably more like ± 5 °C. It seems unlikely that a climate system that has been around for billions of years is running with positive feedback. The long term stability of the climate system suggests negative feedbacks rule instead.
>>
4. Warming has been measured that, if viewed honestly, confirms the above
<<
Actually, the warming doesn’t confirm GHGs as the cause. The models predict a warming of the surface due to increased GHGs, but they also require a 130% to 160% warming of the atmosphere. The current warming of the atmosphere is more like 60% to 90%. That rules out GHGs as the culprit.
>>
5. Other effects (Sun, ENSO, volcanoes, etc.) have been studied and found not to explain the warming, though they do make the signal noisy
<<
Surface albedo does explain the warming. You can decrease the surface albedo (and remain well within current albedo error bars) which increases the surface temperature by the proper amount and increases the atmosphere’s temperature by the proper amount too (60% to 90%).
>>
6. The effect of this warming will cause undesirable effects such as rising sea levels, an increase in extreme weather events and others (MOST OF WHICH ARE YET TO HAPPEN)
<<
Sea levels have been rising since the start of the Holocene. But extreme weather events are driven by large temperature differences. In a warmer world caused by increased GHGs, the poles are supposed to warm faster than the tropics. That would reduce variance and by analogy reduce storm strength.
Jim
What is the accuracy of modern thermometers? Muller is claiming an accuracy 0f 0.04deg for the anomaly figures. Perhaps someone can advise – if you measure temperature to within +/- 0.5 deg. (the range for old style mercury thermomters) is that not the accuracy range of the anomaly also?
RE: Donald: (October 30, 2011 at 8:17 pm)
“We have not only “doubled” our CO2 output since then 1950; we have in fact “tripled” it.
One would have thought the ratio of temperature change due to an astronomical three fold increase in CO2 would have been “HUGE”…”
First, this is like telling a property owner on Lake Washington that he will cause a HUGE increase in lake pollution if he astronomically increases the number of his dogs from one to three. From the nominal pre-industrial level of 280 PPM (parts per million) in the atmosphere, modern industry has only increased the net CO2 concentration to about 396 PPM. This is only a 41 percent increase in atmospheric CO2 concentration.
Second, CO2 only interferes with about ten percent of the energy leaving the Earth in a narrow band around a wavelength of 15 microns. Most of the effect of any added CO2 is masked or hidden by that already there. The online MODTRAN web tool provided by the University of Chicago indicates that a 100 percent increase in CO2 will cause less than one degree C raw ground temperature increase due to the Greenhouse effect.
Below is an image that shows the minimal difference in energy leaving the earth 20 km up when the CO2 concentration is doubled from 300 PPM to 600 PPM. Note that the green 300 PPM curve can only be seen rising above the blue 600 PPM CO2 curve in a few places on the fringes of the general CO2 hole.
http://en.wikipedia.org/wiki/File:ModtranRadiativeForcingDoubleCO2.png
For Chris….