By Steve Goddard
As Bob Tisdale pointed out, Tom Karl’s NCDC trend claims don’t match his graph. The trend line is less than either of the claimed V2 or V3 trends in the graph below.
But beyond this blatant error, there are other problems with his graph. Why did it start in 1900? NCDC has data going back to 1880. As you can see below, temperatures dropped from 1880 for about 30 years, which reduces the long term slope considerably.
http://www.ncdc.noaa.gov/img/climate/research/global-jan-dec-error-bar-pg.gif
Now let’s compare his graph vs. Had-Crut, which goes back to 1850. Had-Crut is shown in green, and NCDC V3 is shown in red, at the same scale. Note that there is a huge divergence over the last ten years vs. Had-Crut. The Had-Crut long term trend is 0.45C/century, about half of what Tom Karl is claiming.
It appears that someone is hiding the decline at both ends of the graph.
UPDATE from Steve Goddard:
When I wrote the article I did not recognize that the “global” data presented to the Senate in slide 21 was land only. Thanks to Bob Tisdale for pointing it out. The land only data has diverged from global data over the last decade and explains the discrepancy at the right side of the graph. It does not explain why the slide is marked as “global” or why the land only data set was presented to Congress as “unequivocal” evidence of “global” warming. My apologies to Tom Karl for not recognizing which data set he was presenting to Congress in that particular slide.
latitude says:
May 20, 2010 at 3:22 pm
“NCDC has data going back to 1880″
So how far would we have to go back to see the even bigger trend.
It’s hell coming out of a LIA.
______________________________________________________________
If you want to know take a look at Lucy’s super graph Click
@Nigel Leck
Extend that line back to the MWP or the RWP. You sound like you bought into Mann’s straight line going back forever. That has been throughly debunked here and elsewhere. You must be new here. Relax and just sip the cool aid instead of gulping it.
the thing that is missing from these graphs are ERROR bars.
A.J. Strata dug through the whistleblower released CRU information and found the error bars.
A.J. States:
“…Before we dive into CRU data we need to step back and understand the concept of ‘accuracy’ in scientific measurements. One rule of reality is you cannot process data (run statistics) to create more accuracy than originallly captured in the raw data. If you measure something in meters or yards, no amount of statistical analysis can increase your accuracy to inches or centimeters….
…I also want to note the size of the so-called temperature anomaly depicted on this graph. It shows a -0.3°C (circa 1910) to +0.5°C (circa 2000) change over the period. Therefore the error in temp data cannot exceed 0.5°C or else all these global temp indexes are statistically equal (i.e., no warming)….
The title of this graph indicates this is the CRU computed sampling (measurement) error in C for 1969. Note how large these sampling errors are. They start at 0.5°C, which is the mark where any indication of global warming is just statistical noise and not reality. Most of the data is in the +/- 1°C range, which means any attempt to claim a global increase below this threshold is mathematically false. Imagine the noise in the 1880 data! You cannot create detail (resolution) below what your sensor system can measure. CRU has proven my point already – they do not have the temperature data to detect a 0.8°C global warming trend since 1960, let alone 1880…..”
The graph is here: http://strata-sphere.com/blog/wp-content/uploads/CRU%20Sampling%20Error.gif
geo says: “Re Bob Tisdale’s question –but if that’s the wrong dataset, then why is the agreement so close prior to the last 10 years?”
Look again. Is it so close prior to the last 10 years, or is it the way Steve overlaid them?
geo says: May 20, 2010 at 3:57 pm . . .
Totally off topic, but great work on the Iowa station survey, geo. Much appreciated.
Now who’s gonna get Indianola?
Mr Mike McMillan;
I’ve wanted to ask some questions for awhile now ;
1)_What subset of the Illinois stations did you plot for those blinkers?
2) Was there any indication or statement that the version had changed from V1 to V2, if indeed that is the difference?
3) What rationale was offered for ‘RAW’ data changing so dramatically?
4) What is your opinion of the source of the dramatic shifts in individual trends?
When I show that page to AGW believers, they reject it as being cherry-picked or not representative.
Beyond the download dates, what else can you say about the provenance of the underlying data sets?
Big thanx
RR
Mike G: Take the line to when the asteroid killed the dinosaurs it may even show a cooling as I’m fairly sure that it was quite warm on that day.
Maybe if you want to be serious you should look at this instead :_
http://climateprogress.org/2010/05/16/nasa-easily-the-hottest-january-and-hottest-jan-april-in-temperature-record/
jb says: “It seems to me that you could perhaps ignore older temperature variations as natural, and claim that only the changes that coincide with increases in CO2 should be considered as part of the trend line.”
Circular reasoning lies at the base of your suggestion. Correlation is not causation.
Nigel Leck,
OK, I read your article and the links. Did you read the comments? I didn’t count, but it appears there are more [true] skeptics than those supporting the article’s contention. [BTW, I dislike blogs where the owner repeatedly jumps into the thread to argue with people who have a different point of view than he does. Almost all alarmist blogs do that, and it tends to drive away the folks who get put in their place by the presumed “authority,” even though they often turn out to be right after all.]
But anyway, let’s assume you’re right, and there can be no other explanation for the isotope disparity, even though the authors conclude that “the (isotopic) fractionation between compartments is not established well enough to allow conclusions to be drawn unequivocally.”
So please explain how an isotopic change from 0.0111073 of CO2 to 0.0110906 is going to radically alter the global climate. This recent peer reviewed paper shows that there has been no net increase in atmospheric CO2 over the past 150 years, or even over the past 5 decades. It appears that the biosphere is responding to increased airborne fertilizer.
The central question is whether CO2 will cause runaway global warming, isn’t it? Because if CO2 has only a slight to moderate effect on temperature, or its effect is naturally countered by other feedbacks, then there is no reason to continue spending the exorbitant amounts of tax dollars on beating the dead carbon horse.
Finally, CO2 has no measurable effect on temperature. Rising CO2 is not the cause, it is the effect of past rising temperatures. Also, what we see now has happened regularly in the past.
If you accept the wisdom of Occam’s Razor, you know there is no reason to add an extraneous entity like CO2 to the explanation of rising and falling global temperatures. From what is actually observed, that tiny trace gas has less influence over the planet than U.S. postal rates.
@Nigel
Maybe if you want to be serious you should look at this instead :_
That’s covered pretty well of various current threads. One, recent El Nino, now deceased. Two, GISS extrapolates METAR corrupted temperature data to encompass the entire arctic, convieniently inflating their numbers.
For Nigel: http://i42.tinypic.com/vpx303.jpg
So, that’s how ya’ll do it. Cook the books.
Reminds me of a certain chemist we used to call to get boron concentration. She’d ask what we expected it to be and then go grab the sample and analyze it. Pretty soon she’d call back and tell us it was what we’d thought it was going to be. Of course, we did know what it was going to be because we knew how much we’d diluted the RCS since the last sample.
RuhRoh,
I’ve wondered about some of those things, too. Notice that the later charts always show faster warming, whether it’s done by lowering past temperatures, or by raising current temperatures.
But the most glaring issue is that both charts in V1 and V2 have the notation “July 09 Raw” and “November 09 Raw”. GISS could not have made the same error inadvertently on every station. One or both of the data months for each station must have been “adjusted.”
It is hard to avoid the conclusion that James Hansen is deliberately manipulating the climate data in his custody to show alarming warming.
Smokey:
“no net increase in atmospheric CO2” We can directly measure the levels of CO2 in the atmosphere, so sorry I do not agree with you on that point.
“CO2 will cause runaway global warming” A doubling of the CO2 levels by itself is predicted to produce 1.1c increase in temperatures BUT it’s the feedbacks both positive and negative that the debate is over. One of these feedback is as ice caps melt the dark oceans will adsorb more heat. The sum total of all these feedbacks is predicted to increase the average temperatures by ~3c for a doubling of the CO2 levels.
http://www.ccrc.unsw.edu.au/Copenhagen/Copenhagen_Diagnosis_LOW.pdf
Dan Inesanto
They are at the same scale. If you look closer you will see that they are normalised with NCDC shifted down by 0.1 degrees, due to a different baseline.
Mike G: You know that we have been in a period of a solar minimum yet record high temperatures. A sinlge El Nino doesn’t explain multi decade trends.
http://science.nasa.gov/science-news/science-at-nasa/2009/03sep_sunspots/
Bob,
The NCDC presentation has disappeared off their web site as Anthony predicted. From looking at the copies of his slides on WUWT there is not much indication that they are land only, but after graphing the NCDC data from the web site I think you are correct.
Why would he be using land-only data???? Hansen tells us that global warming is about the oceans. Sounds like yet another deception.
Nigel Leck,
There is no valid reason to exclude any Had-Crut data.
https://spreadsheets.google.com/oimg?key=0AnKz9p_7fMvBdDFaMF92Y19odXZoLUhZMHJBUVk1LWc&oid=1&zx=qd0s83-pcpv8g
The slope is 0.45C/century. Now justify your cherry picking.
Yes Nigel,
But that level of feedback is derived by tweaking gains in questionable climate models to get the warming that CRU said took place and was claimed to be entirely man-made (tweaking the models to match the thoroughly debunked hockey stick, ignoring UHI, adjusting past temperatures downswards, etc.). They are deniers of natural variability.
Bob,
I calculated the NCDC land trends from their web site. 0.76C/century since 1900 and 0.72C/century since 1880.
But their land-ocean trend since 1880 is only 0.57C/century!
Given that the ocean is 2/3 of the planet, I find it mind-boggling that he is using a shortened trend from 1900, using only land data and excluding the oceans.
Nigel Leck,
See Mike G’s link above. Please explain.
And when you say, “so sorry I do not agree with you on that point,” without giving any basis for your feelings, you do realize that you are not arguing with me, you are arguing with the conclusions of a peer reviewed paper.
I have lots of charts, and graphs, and articles from knowledgeable climate scientists, if you’d like to keep going.
I know your mind is made up, but lots of other folks read these comments. They are the ones for whom I’m really providing this information. But maybe you will learn something in the process.
RuhRoh says: May 20, 2010 at 5:55 pm
I’ve wanted to ask some questions for awhile now ;
1)_What subset of the Illinois stations did you plot for those blinkers?
That is the entire Illinois USHCN set, not a subset. The stations are listed on surfacestations.org
http://www.surfacestations.org/USHCN_stationlist.htm
2) Was there any indication or statement that the version had changed from V1 to V2, if indeed that is the difference?
All charts were downloaded from
http://data.giss.nasa.gov/gistemp/station_data/
with “raw GHCN data+USHCN corrections” selected in the “Data Set:” box.
Prior to V2, there was no other set. The current page has a “Note to prior users,” but it is at best ambiguous, and doesn’t mention anything about a Version 2. That the dataset changed between the dates is demonstrated by the charts.
3) What rationale was offered for ‘RAW’ data changing so dramatically?
http://www.ncdc.noaa.gov/oa/climate/research/ushcn/
4) What is your opinion of the source of the dramatic shifts in individual trends?
UHI is not included among the adjustments. I don’t doubt that other adjustments were fine tuned within reasonable limits to show an unambiguous warming trend, but that’s only opinion. The new raw data is much more closely aligned with GISS homogenized data, fact, suggesting an effort to fit the data to the theory, opinion.
When I show that page to AGW believers, they reject it as being cherry-picked or not representative.
That’s every Illinois station. Here’s Iowa and Wisconsin, every station –
http://www.rockyhigh66.org/stuff/USHCN_revisions_iowa.htm
http://www.rockyhigh66.org/stuff/USHCN_revisions_wisconsin.htm
Beyond the download dates, what else can you say about the provenance of the underlying data sets?
Charts are direct from the govt web site, overlaid to set the scales in register, then .gif’d.
Richard Hill says:
May 20, 2010 at 4:29 pm
Where are the statisticians? Could Bob T. or WIllis E. kindly explain why people are still allowed to draw straight lines through the these data series when the issue of “stationarity” means that it is statistically invalid to do so. Has “stationarity” gone away as an issue? I would appreciate an explanation.
Richard, I would also like a second opinion on that aspect.
However, I came away from that tremendously long interplay with VS a few months ago that a linear regression has two completely different uses. One is to see backwards at what data has done in the past while minimizing variances, nothing more. Two is to attach a confidence level to that data and regression and try to infer some inkling of it’s probability to continue along that regression line into the future, a true trend that is (and I know the term “trend’ gets thrown around with dual meaning applying to both cases). To me, that is where “stationary” come blazing in to play and clearly indicates when that is not only wrong but fictional, a figment so to speak. If I’m incorrect there in my thoughts, I would like to hear from someone savvy in statistics because I really want to understand and know I am thinking correct when in that realm.
geo: I just checked, and, as I suspected, the similarities between the land surface temperature and combined land plus sea surface temperature are really a matter of appearance. (Steve actually did a great job of overlaying the two graphs.) Here’s a comparison of annual land surface temperature anomalies (CRUTEM) and combined land plus sea surface temperature anomalies (HADCRUT) from 1850 to 2009:
http://i50.tinypic.com/rscs45.png
The two datasets appear similar for the mid to latter part of the 20th century. But if we subtract the HADCRUT data from the CRUTEM data we can see the actual differences:
http://i50.tinypic.com/icidmu.png
When global temperatures are rising, land surface temperatures rise more than the combined, and when global temperatures are dropping, land surface temperatures drop more.
Smokey says: May 20, 2010 at 6:40 pm
But the most glaring issue is that both charts in V1 and V2 have the notation “July 09 Raw” and “November 09 Raw”.
I put that label on the charts, which indicates the file date when I downloaded the two charts. I don’t recall knowing at the time that there was a version 2, or when the changeover was.
Wisconsin blinks were downloaded later, but there were still no indications on the download pages, so I continued to label them with just the date.
I have indicated v1 and v2 on the Iowa charts.
Hmm.
Somebody FOIA noaa for the correspondence regarding the disappearing of presentations that have previously been public. And maybe check the DQA.