There is a WordPress blog called Http://klimakatastrophe.wordpress.com that recently did an analysis of some changes in the way Hadley Climate Research unit presents its HadCRUT data. They suggest that HadCRUT has been doing some adjusting, and gotten closer to GISS. It has been translated into English with the help of WUWT reader Pierre Gosselin and shown below, but you can read the original in German here:
Lets all do some investigation into this to see what is actually going on and if the claim holds up.
UPDATE: There is something wrong with this analysis, possibly a comparison to two data sets that are land and SST rarher than the combined HadCRUT index. Please don’t reference this story until I can hear from the original writer.
UPDATE2: This analyis has confirmed errors, disregard it. I’m not sure if this was accidental or intentional, but the HadCRUT data has not been modified as the German blog author claims.
UPDATE3: Thanks to readers Nick Stokes and Pieree Gosselin, a corrected analysis has been posted, see it here:
The Met Office adjusts its HadCRUT3-database to better match the GISS NASA database.
Written on Sunday, 13 April, 2008
GISS Global temperature of NASA has shown substantial deviations over the last years when compared to the other three data sets of global temperature (HadCRUT3, RSS and UAH), particularly for March 2008. Now the Met Office has itself apparently decided to adjust HadCRUT3 databases to better match GISS values. See the following diagrams:
The database is shown by the black points with red line (moving annual average), how it appeared last week by the Met Office. The circles with blue line (moving annual average) represent the new, highly corrected database. So far I haven’t been able to find any reason for this, or why this correction took place.
As the lower part of the illustration shows, on a long-term basis, the annual average (the red line), was corrected approx. 0.1°C. The month of March was adjusted from 0.166 to 0.43°C. Thus around 0.264°C. Recently the Met Office has again changed its presentation of the annual average temperatures. The start of 2008 was particularly cold, and led to a temperature dive in the yearly series. For this reason the temperature was removed from the early 2008 series.
Now we have a massive problem. Because of the corrections, the surface measured temperatures (GISS, HadCRUT3) show a trend which no longer agrees with the satellite measurements (RSS, UAH).
Two groups seem to have formed, using different calculation and presentation methods for global temperature. This circumstance is obviously due to missing climate warming. Such adjustment attempts are not only dubious, but also cast great doubt on these (Hadley, GISS) climate scientists, and risk losing credibility.
Update: As I have just determined, the monthly average values are constantly changing, and that going all the way back to 1850. It is constantly being readjusted…to be more and more in line with GISS database.
Conditions 11.01.2008
2007/01 0.923
2007/02 0.680
2007/03 0.595
2007/04 0.634
2007/05 0.522
2007/06 0.531
2007/07 0.545
2007/08 0.589
2007/09 0.520
2007/10 0.525
2007/11 0.432
Conditions 18.03.2008
2007/01 0.632
2007/02 0.520
2007/03 0.441
2007/04 0.473
2007/05 0.374
2007/06 0.377
2007/07 0.403
2007/08 0.370
2007/09 0.409
2007/10 0.360
2007/11 0.266
2007/12 0.202
2008/01 0.056
2008/02 0.194
Conditions 11.04.2008
2007/01 0.366
2007/02 0.361
2007/03 0.310
2007/04 0.286
2007/05 0.265
2007/06 0.332
2007/07 0.354
2007/08 0.282
2007/09 0.294
2007/10 0.228
2007/11 0.149
2007/12 0.112
2008/01 0.116
2008/02 0.154
2008/03 0.166
Conditions 13.04.2008
2007/01 0.632
2007/02 0.520
2007/03 0.441
2007/04 0.472
2007/05 0.375
2007/06 0.376
2007/07 0.403
2007/08 0.370
2007/09 0.414
2007/10 0.356
2007/11 0.265
2007/12 0.201
2008/01 0.056
2008/02 0.187
2008/03 0.430
In my Plots of 11.04.2008 I used the values of 11.04.2008.
The “correction” looks to be about 0.1 K on average. This is a major change when differences of 0.3 – 0.5 K over the past 30-40 years are being used to justify either alarm or complacency over global temperature trends. They surely have to explain and justify the change.
With all the various doctoring of data — the heat island business, for example, the corrections to the satellite data — it becomes increasingly difficult to take the data too seriously. At the very least, error bars, whatever they are, should be put on the various graphs.
I have to laugh when people claim that global temperature is as well documented as the law of gravity! We’re talking, what, 14 orders of magnitide difference in the relative errors?
Okay, then. If the satellite measures are not matching the ground measures, then it begins to add up. Maybe. Begins.
Depending on how accurate the satellite measures are and how well lower troposphere corresponds to actual surface temps, the miccrosite bias is probably found in the difference between the satellite and ground measures.
One wonders what that difference is, both in terms of trends and absolute. One would prefer surface air temperatures over land and sea, not a land/ocean (3m under surface) measure for the comparison. And over-land measurements for both ground and satellite. Then maybe we will be getting somewhere.
It’s hard for me to believe that the Germans, the world standard bearer of precision, would put up with this laughable farce.
There’s something odd here. The data for 18.03 and 13.04 are the same, and are those that have been used elsewhere, eg by lucia. So it doesn’t seem to be a permanent change. Do we know where the aberrant sets labelled 11.01 and 11.04 appeared?
Oh man!
These translation softwares are a total diasater.
Anthony, please allow me to clean this up:
—————————————–
The Met Office adjusts its HadCRUT3-database to better match GISS NASA database.
GISS Global temperature of NASA has shown substantial deviations over the last years when compared to the other three data sets of global temperature (HadCRUT3, RSS and UAH), particularly for March 2008. Now the Met Office has itself apparently decided to adjust HadCRUT3 databases to better match GISS values. See the following diagrams:
The database is shown by the black points with red line (moving annual average), how it appeared last week by the Met Office. The circles with blue line (moving annual average) represent the new, highly corrected database. So far I haven’t been able to find any reason for this, or why this correction took place.
As the lower part of the illustration shows, on a long-term basis, the annual average (the red line), was corrected approx. 0.1°C. The month of March was adjusted from 0.166 to 0.43°C. Thus around 0.264°C. Recently the Met Office has again changed its presentation of the annual average temperatures. The start of 2008 was particularly cold, and led to a temperature dive in the yearly series. For this reason the temperature was removed from the early 2008 series.
Now we have a massive problem. Because of the corrections, the surface measured temperatures (GISS, HadCRUT3) show a trend which no longer agrees with the satellite measurements (RSS, UAH).
Two groups seem to have formed, using different calculation and presentation methods for global temperature. This circumstance is obviously due to missing climate warming. Such adjustment attempts are not only dubious, but also cast great doubt on these (Hadley, GISS) climate scientists, and risk losing credibility.
Update: As I have just determined, the monthly average values are constantly changing, and that going all the way back to 1850. It is constantly being readjusted…to be more and more in line with GISS database.
——————–
Anthony,
To sum it up. They’re cooking the books – big time!
REPLY: That remains to be seen.
It’s bed time for me.
Gute nacht!
They’re revising history in England!
This site archived values, including Hadcrut3, in February here. The latest Hadcrut3 is . here There’s virtually no change.
Adding to Mike’s thoughts, an individual datum here is taken to have an error of +/- 0.5 degrees. How are adjustments of < 1 degree to a datum justifiable on any grounds? Anthony’s audit would seem to indicate measurements to greater precision are an empty conceit.
I expect that Hadley Centre will be giving us some explanation in due course. Whether the explanation will mollify critics is unlikely, but it will help calm the faithful in these troubling cool times.
“Now the Met Office has itself apparently decided to adjust HadCRUT3 databases to better match GISS values.”
But wait! We’re being told that the GISS is more accurate because of the Arctic extrapolation. Does this mean that HadCRU is being adjusted to account for this “Arctic estimate”? Or is it just because Jones et al are tired of Hansen et al getting all the attention because the GISS charts are more widely used to support AGW?
Also, there will STILL be some offset due to the difference in reporting periods, won’t there?
I suspected as much from this graph (2008 = average of Jan + Feb + March). I calculated about 0.2 from this graph which is updated weekly
http://www.cru.uea.ac.uk/cru/climon/data/themi/g17.htm just watch for a sudden change. Is this correct Anthony? It is quite interesting to see quite a few people beginning to save before and after scenarios. Don’t be surprised if soon data will not made available to public
Misreadings. The Germans seem to be making elementary errors here. I don’t know where the first set of figures come from. The second and fourth sets are correct Hadcrut3. And the third set (11.04) are hadsst2 sea surface temperatures.
Explanation of plot. In the first plot, the red curve is combined land/sea temperature, Hadcrut3. The lower, blue curve is sea surface temperature, Hadsst2. It isn’t “before and after” Hadcrut3.
REPLY: Hmm, I wonder why they’d do that? I’ll change the post. Thanks for pointing it out.
I’m also going to contact the blog owner and ask them why they’d make a comparison this way, it makes no sense.
The climate “scientist” mafia, in the pay of Big Government, want to get their stories straight.
Evan Jones, you raise an old chestnut of a question which still hasn’t been answered by anyone:
What, precisely, is the temperature of the the planet?
The land surface.
The sea surface.
The surface atmospheric temperature/pressure product
The ionospheric temperature
The temperature at the tropopause.
The sum total of incoming against outgoing energy – of all forms.
OK In writing my rhetorical question list, I finally managed to formulate a definition of global warming/cooling.
I can’t sleep, and so here I am.
Please note that I only did a rough correction of the text to make it a bit more understandable, and not more. I know the English is bad.
Also the German who posted this is one of the good guys.
He’s only doing what McIntyre etc are doing – trying to expose the shady work of a few Brits. Errors are possible.
VG
It damn well better stay available. Our taxes are paying for it!
Sorry, the mouse jumped my keyboard before I could finish:
OK In writing my rhetorical question list, I finally managed to formulate a definition of global warming/cooling.
If the TOTAL amount of energy dumped into the Earth, including the atmosphere and magnetosphere and orbital energy, increases, then the “temperature” of the Earth must increase until the radiated energy is equal to the increased received energy. This is global qwarming.
The whole point about this definition is that it is unambiguous and concise and quantifiable.
Global warming isn’t about local land temperatures. This defintion of global warming is precise.
Maybe his English isn’t so good, and perhaps he misread some headings.
But you’re right, if he made a mistake, it only hurts us. His conclusion may have unjustly put HadCrut in a bad light.
I’m curious how this German will react to being corrected by some cowboys 😉
REPLY: Pierre, since you know some of the German language, can you make an inquiry at that blog?
His overall conclusion is correct though: Temperature averages from GISS, and possibly HadCrut, are moving away from satellite averages. This has to be explained.
Well, I for one am glad the error was caught. I added it to this morning’s article regarding what we are seeing when we are shown a global temperature chart. Much of the material was from here.
CoRev, Editor
globalwarmingclearinghouse.blogspot.com
Anthony,
Sure will…
I already sent him an e-mail.
Right now it’s 2 a.m., and so expect another half day or so to find out what’s up.
Robert Wood, I like the definition. Hope you don’t mind if I use it over at my house?
OK What are the ramifications of my definition of global warming?
1. AGW cannot be true. Why? Because we do not control the energy input or output of this planet. We may, indeed, affect the way it is redistributed around the planet, but no more than that.
2. It is not important whether there are heat sinks, or iris effects, or any such details. Whatever climatic mechanisms lead to heat transfer between different layers and parts of this planet, the energy balance, therefore temperature, is decided by external factors.
3. Even IF AGW were true, temperatures will be regulated by Stephan-Bultzmann, NOT Al Gore.
His overall conclusion is correct though: Temperature averages from GISS, and possibly HadCrut, are moving away from satellite averages. This has to be explained.
Ummmm . . . Let me think . . . MICROSITE VIOLATIONS, MAYBE?
RW: Surface temperatures are important. We have huge amounts of energy stored in the earth’s core. So long as it stays there we don’t have a problem. Overall input and output are important, yes, but they are only part of the story.
That “distribution” thingie is an all-important consideration becuase we live on the surface and if that changes a whole lot of other things change too.
No, I don’t think surface temps are changing outside the natural ups and downs, and while I think CO2 has a slight effect, it is dwarfed by other considerations. But “where” the heat is does matter.
BTW, I agree with your splits. They are all important. What matters is keeping them all tallied and not confusing them.
Everybody needs to use the same scepticism WRT this as everything else.
I want to see the data. I want to see the methods. I want to see the code.
Then double check. then have someone else double check.
I’ve seen small changes in Hadcru, but nothing of this order. So, I am suspect.
Still. check. double check. you engineers know the drill.
Hold your horses everybody. There does not appear to be major tinkering going on. Being a bit sloppy, and having a large harddrive, I have the hadcrut3 data to 2007/11 and 2007/12 and 2008/02 and 2008/03 lying around. No major changes. No changes period, before 2007. I’ve uploaded the datasets and “diffs” (unix/linux users will know what these are) as http://clients.teksavvy.com/~walterdnes/changes.zip alongside my own analysis http://clients.teksavvy.com/~walterdnes/temperatures.zip (before anyone asks, my date scheme is 12 months “different” from what you would intuitively expect).
It’s not new that HADCRU changes its temperatures to “make up” the warming trend. Just compare their graphs for global temp in TAR and 4AR : http://skyfal.free.fr/?p=122
Evan,
You are right, the internal mechanisms do matter, I was attempting to clarify the whole concept in my head. Now, just where do we stick that thermometer:-)
So far he hasn’t replied to my e-mail. So I posted a message at his blog (along with reader here Nick Stokes). He’s probably off to work.
I had a look through his website and it looks legit.
“Update: As I have just determined, the monthly average values are constantly changing, and that going all the way back to 1850. It is constantly being readjusted…to be more and more in line with GISS database.”
At what point does this become a complete and total fraud?
Herr Mueller has updated the above posting
in English:
—————————
Important Advisory: Thanks for the advisory from Nick Stokes. As I have just noticed, the databasis I used on 11 April 2008 are not HadCRUT3 values(global surface temperatures), but rather they are the HadSST2-values (global sea surface temperatures). In retrospect there is no way for me to know if this was a mix-up on my part, or a mix-up on the part of the Met Office. Usually I always download the HadCRUT3 data from the same Met Office link.
I’ve removed the plot and will update it. I view this correction as necessary, as my post was taken over 1:1 by other blogs. That should not be done without first checking it over carefully. Therefore in the future I kindly request that it be carefully checked over first, or that I be consulted about it.
The following remains as fact: The HadCRUT3 databasis average monthly values are undergoing an adjustnment process, as the values from 18 March 2008 and 13 April 2008 show. I will continue to watch these values and save the link to their source just to be on the safe side.
GISS Global temperature of NASA has shown substantial deviations over the last years when compared to the other three data sets of global temperature (HadCRUT3, RSS and UAH), particularly for March 2008.
But the Met Office also had problems, and had to make corrections. Just recently the Met Office changed its diagram of the annual mean temperature. The start of 2008 was particularly cold, and led to a temperature dive in the yearly series. For this reason the temperature was removed from the start of the 2008 year series.
Now we have a massive problem. The surface measured temperatures (GISS) show a trend which no longer agrees with the satellite measurements (RSS, UAH). Two groups seem to have formed, each using different calculation and presentation methods for global temperature. This circumstance is apparently due to the missing climate warming of the last years. Such adjustment attempts are not only dubious, but make the climate scientists (Hadley, GISS) lose their credibility.
Update: As I have just determined, the monthly average values of the HadCrut3 data basis (global temperature) are constantly changing. That is, corrections are constantly being made.
Date of data: 18 March 2008
2007/01 0.632
2007/02 0.520
2007/03 0.441
2007/04 0.473
2007/05 0.374
2007/06 0.377
2007/07 0.403
2007/08 0.370
2007/09 0.409
2007/10 0.360
2007/11 0.266
2007/12 0.202
2008/01 0.056
2008/02 0.194
Stand 13.04.2008
2007/01 0.632
2007/02 0.520
2007/03 0.441
2007/04 0.472
2007/05 0.375
2007/06 0.376
2007/07 0.403
2007/08 0.370
2007/09 0.414
2007/10 0.356
2007/11 0.265
2007/12 0.201
2008/01 0.056
2008/02 0.187
2008/03 0.430
(See his new plot:)
http://klimakatastrophe.wordpress.com/2008/04/13/das-met-office-hat-seine-hardcrut3-datenbasis-der-giss-nasa-datenbasis-angeglichen/
———————————
Hope this clears things up!
PG
REPLY: Thanks to Pierre and to everyone that contributed to getting this corrected.
Ah, disagreement between the surface and aerial measurements. I wonder what the major media will make of this, and the climate scientists.
There’s a real chance that this global warming business is going to crash, especially if the data sets turn out to be spurious, and take a lot of the credibility of science with it. Not that the non-climate science types will have anyone much to blame but themselves.
Earth has tremendous capabilities to affect long term climate change. Land over the poles makes a VERY big difference in how the oceans distribute heat. However, plate movements influence long term trends, not the kind I would use to determine whether or not I should take land out of conservation programs and start planting wheat again. I look at short term influences to help me think into the near future (50 years ahead). Given that, I hope CO2 is the factor in our warming trend. If it is, I am going to plant like crazy, flood irrigate, and get three hay cuttings while using dry land for grazing only. If it isn’t, I will still use dry land ground for grazing only but plant lower valley ground with wheat, since food shortages will result in higher wheat prices. Tinkering with surface averages means little to me. Understanding short term forcing means everything.