While Alan mulls over big red dots in NOAA’s graph where no data exists, Frank Lanser finds that GISS global temperature trend is warmed up by weighting land data more.
Guest post by Frank Lansner http://hidethedecline.eu/
The simple task of combining Land surface temperature with Sea surface temperatures has become an odd complex algorithm for GISS. It seems that they weight land data more and more during the 20th century leading to extra heat added to the GISS global temperatures.
Thus in 1900-1920 the GISS LST+SST graph is mostly spot on the used SST graph (the HADISST/Reynoldsv2). This means that the GISS global temperature around 1900-1920 appears to weight land data zero %
The land fraction is increasing during the 20th century, especially after 1980:
Fig2
The real land fraction of the Earth is of course 30%, but around 1980 GISS uses 40%, in 1988 55% – and in 1995 no less than 73%. (The high land % weighting around 1995 leads to a reduction in temperature decline due to Pinatubo volcanic cooling.)
GISS ends up in 2007 using a land weighting of 67%.
In general GISS defends use of larger land fraction due to their 1200km zones around land stations reaching some Ocean areas. But this does obviously not explain a land fraction that appears to go from near zero to around 70% globally during the 20th century.
Besides, the land temperatures for all stations including the coastal stations appear to show significantly different behaviour compared to the SST´s:
(Taken from PART3 of my new article, http://hidethedecline.eu/pages/posts/the-perplexing-temperature-data-published-1974-84-and-recent-temperature-data-180.php)
And therefore, just shifting SST´s out with land temperatures are questionable.
The topic has been discussed somewhat at Joanne Nova’s site:
http://joannenova.com.au/2010/07/did-giss-discover-30-more-land-in-the-northern-hemisphere/
What is the impact of the still greater land fraction in GISS data? If the 30% land fraction from the real world was used, GISS 2007 would be 0,55K warmer than GISS 1900.
With the still increasing GISS land fraction actually used, we have GISS 2007 0,72K warmer than GISS 1900.
The difference is 0,17K added by not using 30% land constantly. But this calculation could be done in many ways.
We know for a fact that the oceans cover 70% of the planet. So why not use 70% of data from SST?
data sources:
GISS global Land + SST
http://data.giss.nasa.gov/gistemp/tabledata/GLB.Ts+dSST.txt
GISS land temp
http://data.giss.nasa.gov/gistemp/tabledata/GLB.Ts.txt
Both HADISST and Reynolds can be seen using:
http://climexp.knmi.nl/selectfield_obs.cgi?someone@somewhere
We have to admit, that HadCRUT is not THAT bad compared to GISTEMP or NDCD datasets. It fares poorly in tropics, central Africa and Central Asia compared to MSU since 1979, but reasonably well in polar regions and northern/southern extratropics.
Geo, you make my day!!
Your “logic” is that if some conslusion is in the beginning of an article, its probably false 🙂
Well, the presen little writing is around 2% of the real article i 4 parts here:
http://hidethedecline.eu/pages/posts/the-perplexing-temperature-data-published-1974-84-and-recent-temperature-data-180.php
The subject we discuss here is in fact placed in the END OF PART 4 !
So wow! Then it must be true??
I think you ough me to read my article that I just linked to, and then come back and tell your opinion.
K.R. Frank
If this keeps up, Earth officially won’t have any oceans in a few years, with 100% land weighting, and we can then petition the US Government to disband NOAA.
It is being growing boresome to see so many repetitions of lies than nobody any longer believes in but only the bedwetting fools, who, by the way, will always exist.
Frank Lansner: You wrote, “1) How do you justify around ZERO % land data used 1900-20 and then 70% land data used in 2007?”
Not sure where you get these numbers. As I wrote to you over on the next thread, You wrote to Steven Mosher, “In the present case for example we have a GISS LST+SST that is the same as the SST (HADISST) 1900-1920.”
What’s the source of your data? A quick trip to the KNMI Climate Explorer shows that to be incorrect. There are significant differences between Global GISTEMP LST + SST (1200km radius smoothing) and Global HADISST from 1900 to 1920:
http://i25.tinypic.com/2mi14bm.jpg
And here’s the difference (GISTEMP MINUS HADISST):
http://i25.tinypic.com/2rwrbpg.jpg
Frank Lansner: You wrote in the post, “The real land fraction of the Earth is of course 30%, but around 1980 GISS uses 40%, in 1988 55% – and in 1995 no less than 73%. (The high land % weighting around 1995 leads to a reduction in temperature decline due to Pinatubo volcanic cooling.)”
How are you determining these percentages?
Rattus Norvegicus said on July 18, 2010 at 4:31 am:
Good, you have confirmed that you attempted to disprove the existence of the adjustment in the GISS code with something other than the GISS code. While the ccc coding effort does try to follow the original Hansen algorithms and work off the publicly-released code, it of course cannot be used to disprove the existence of such an adjustment in the actual code NASA is using.
There is something troubling I have noticed on the GISS Updates page, which roughly serves as a change log. There are several examples where the programs actually used are modified then the “publicly-available source codes” are subsequently modified. If the code released was the actual code used by NASA, then why would the released code also need modification? Wouldn’t it be a simple matter of placing NASA’s own current code to where it is publicly available? This hints that the released code is the “consumer” version while the “in-house” version may have differences, which has actually been true in certain instances noted on that page.
I’ve also noticed the increasing frequency of notes on that Update page. Perhaps the increased public scrutiny has brought this about, as they document how they are trying to keep GISS as accurate as possible.
Speaking of the publicly-released codes, it’s good to read of the problems the Chiefio (E.M. Smith) has had getting that steaming pile up and running on his own computer. And what he has already found about how GIStemp actually works, as opposed to what the supporters say it does… It’s not good, and such can really leave one wondering why anyone wants to trust it for much of anything.
Neil Robertson says:
July 18, 2010 at 12:46 am
I have a question. The expenditure of trillions of dollars are being contemplated on the basis of what everyone (on both/all sides of this debate) recognises is an incomplete data set….
In fact, wouldn’t automated systems of this sort be greatly preferable even in more urbanised areas, thus removing many of the practical issues that lead to the massive prevalence of siting problems recorded on WUWT and elsewhere, and missing data from many stations?
____________________________________________________________________
You are making the assumption that those advocating “The expenditure of trillions of dollars” are really interested in whether there is any “manmade” global warming. They are not. CAGW is just a convenient excuse for crippling western civilization and imposing “Global Governance” (Nations especially the USA must be weakened and taken by stealth since a take over by direct means is too risky.) That intent has been stated very bluntly several times but no one seems to believe it even when it stares them in the face.
In 1973, just after the first Earth Summit chaired by Oil Mogul, Maurice Strong.
“President Obama’s top science adviser, John P. Holdren, advocated the “de-development” of the United States in books he published in the 1970s.
“A massive campaign must be launched to restore a high-quality environment in North America and to de-develop the United States,” Holdren wrote in a 1973 book he co-authored with Paul R. Ehrlch and Anne H. Ehrlich. “De-development means bringing our economic system (especially patterns of consumption) into line with the realities of ecology and the global resource situation.”“ click
More recently:
“”In searching for a new enemy to unite us, we came up with the idea that pollution, the threat of global warming, water shortages, famine and the like would fit the bill … All these dangers are caused by human intervention and it is only through changed attitudes and behaviour that they can be overcome. The real enemy, then, is humanity itself.”
— in The First Global Revolution, pp.104-105 by Alexander King, founder of the Club of Rome and Bertrand Schneider, secretary of the Club of Rome
“Isn’t the only hope for the planet that the industrialized civilizations collapse? Isn’t it our responsibility to bring that about?” – Maurice Strong, May, 1990
Check out Maurice Strong, Bill Clinton and the UN REFORM – Restructuring for Global Governance
David Rockefeller, the man behind the World Bank, states very bluntly in his 2002 Autobiography 2 “Memoirs” on page 405: “For more than a century ideological extremists at either end of the political spectrum have seized upon well-publicized incidents… to attack the Rockefeller family for the inordinate influence they claim we wield over American political and economic institutions. Some even believe we are part of a secret cabal working against the best interests of the United States, characterizing my family and me as “internationalists and of conspiring with others around the world … If that’s the charge, I stand guilty, and I am proud of it.”
I swear people waking up in slave chains will still be saying it is all a “right wing conspiracy theory” rather than face the fact they have been bamboozled, fleeced and made slaves.
Great analysis. I said on a GISS jump thread last year that I thought Jimmy James had started bumping up the lower toposphere ratio. Looks like that hunch was right, but this properly done study carries far more weight.
Well done!
cce says: “The GISTEMP “Land” series is not an index representing changing land temperature. It is an index that attempts to represent global temperature using only land based measurements.”
That makes about as much sense as looking for your keys where the light is better, instead of where you dropped them. Almost as crazy as trying to measure historic temperatures with a few carefully selected tree rings. They’ve all gone mad.
Great post! it is mind boggling.
In summery, am I right that Gisstemp basically uses about HadiSST/Reynolds from 1900-1935, then as the globe starts to cool, it takes about 30% land data, thus it cools more until about 1959. Then after 1975, land portion goes up and up including UHI effects and polar warming (temperature above the Arctic Ocean is of course smeared over by coastal land temperature data) comes in throughout the two last decades? No polar and no UHI warming at GISS until 1935, but all of it since 1975 and Gisstemp keeps shooting up unlike the other data sets!!
climatepatrol says:
July 18, 2010 at 6:41 pm
Great post! it is mind boggling.
In summery, am I right that Gisstemp basically uses about HadiSST/Reynolds from 1900-1935>>
It is almost impossible to say since GISS doesn’t actually publish the SST data they use. They publish Land and Combined (land/ocean). To try and figure out what they were using I compared Land to Combined and isolated all the grid points that had data in the Combined set, but not in the Land set. By using the 250 km smoothing data to reduce the amount of overlap, this gives you most of the SST data that they are using. There is of course additional SST data merged with Land data at grid points where both Land and Ocean exist, but this method still captures about 75% of the data, and it is reasonably well distributed. We can then trend the SST data that we’ve isolated and compare to both Hadl/Reyn and NOAA. NOAA and GISS use a 1951 to 1980 mean as reference while Hadl/Reyn uses 1961 to 1990. To compare, I averaged Hadl/Reyn for 1951 to 1980 and adjusted it accordingly. Put the three trends on a graph and you get this:
http://knowledgedrift.files.wordpress.com/2010/07/giss-1.png
Looks like Hadl/Reyn at the cold end and NOAA at the warm end with some flip flopping around in the middle.
That said, I came to the same conclusion as Frank a while back, my estimate being that Land is weighted about 6% (based on the 250 km smoothing) in 1880 and based on the increased number of land stations reporting becomes a much larger percentage over time.
kadaka,
A development line and a release line? Oh, you don’t seem to be a software developer. It seems that the code is updated as bugs are found or changes in the procedures used are incorporated.
You might try corresponding with Nick Barnes at CCC, he has the skinny on a lot of this. Ignored stations are AFAIK contained in a list that can be downloaded from GISS. I know that at one point Rito Reudy had expressed to CCC an interest in using the reimplementation from CCC as the actual code. CCC and Peter O’Neill seem to have been the only outside parties providing fixes to GISS. I don’t see E.M. Smith providing any input, but at least two other people have been able to get it to run so I think that E.M. Smith’s problems lie with E.M. Smith.
Most of the rest of the changes have to do with problems with the input data, some of which have been highlighted by the contrarian community. This is a good thing, but has had little impact on the analysis. Many more seem to have been due to problems with the data sets which were discovered through internal investigation. This is also a good thing, but has had little impact on the analysis.
I do have one question to ask: if GISS is changing the weighting of land vs. ocean in the final product, when does the trend agree so well with CRU, NCDC, JMA, UAH and RSS? Are they all colluding to jigger their respective analysis in the same manner? And what about the work which was done by the Muir Russell committee, which wrote their own code and got pretty much the same results from the publically available data? Or how about the various implementations done by bloggers over the last few months, all of which get about the same results, using methods that differ from GISS? Are they all doing this?
Frank Lansner says:
July 18, 2010 at 7:41 am
Geo, you make my day!!
++++
Frank, I’m always happy to make someone’s day. Pay it forward, ‘mmkay?
The context that is presented me is the context presented me. I shall be happy to read the rest of your work, which wasn’t obvious from what I read above, when I get a chance.
Excerpts from: Rattus Norvegicus on July 18, 2010 at 9:05 pm
Warning! Warning! Misleading diversionary tactic detected!
I did not mention development and release lines. I mentioned the possibility of in-house and consumer versions, i.e. what they release is not exactly what they themselves use, which is a different thing.
He has a whole section for GIStemp at his site, has sometimes posted here on WUWT some of what his in-depth analysis has revealed as well. I’ve seen the “buzz” the defenders of GIStemp and Hansen generate when they see his work, and the insulting personal attacks against him as well. Someone at NASA must be noticing his fine work.
Publicly released code, with all those researchers of the temperature record out there who are interested in using GIStemp and seeing how it works, and you can only find two people on your own who got it to run?
Go to the updates page and see the June 9, 2008 entry. They moved their “analysis” off a 15-year old machine onto a new one, and it did affect results. They modified some routines to minimize machine dependence. Since then, I see no note about making the code machine independent. E.M. Smith noted some stuff was for big-endian while PC’s are little-endian. Plus there are uses of non-standard Fortran. Etc. Add everything together and it becomes clear how getting GIStemp up and running can be a real, long, and excruciating pain.
At this point I count five questions, and decide I’ve spent enough time on this. I got you on your “adjustment” comment, now it just looks like you’re diverting attention from what you did. For four of the five, as has been discussed here before, you start with the same buggered-up temperature records then you get similar results, there are differences between satellite and temp record trends, and in any case while warming has been noted the link to CO2 has not been shown. The PDO and other factors can account for the warming.
For the one that remains, regarding your claim within your question about Muir Russell, provide links describing this code they themselves wrote and preferably a link to their source code. Given the short time span of the inquiry, it seems unlikely on the face they, more likely others under their direction, wrote such code from scratch to process the temperature records. Besides, Muir Russell was charged with
absolvinginvestigating CRU, not GISS, and it has been acknowledged CRU is doing a better job than GISS (far as that goes). If you wish to say Muir Russell shows GISS is pretty good, you better have some good references.Jose Suro
It’s the same number. Each degree Celsius is the same “size” as a degree Kelvin. The only difference is that Celsius has its zero at the freezing point of water and Kelvin at absolute zero (where all molecular and atomic motion ceases). Zero K is at -273.15C.
climatepatrol says: “In summery, am I right that Gisstemp basically uses about HadiSST/Reynolds from 1900-1935, then as the globe starts to cool, it takes about 30% land data, thus it cools more until about 1959. Then after 1975, land portion goes up and up including UHI effects and polar warming (temperature above the Arctic Ocean is of course smeared over by coastal land temperature data) comes in throughout the two last decades? No polar and no UHI warming at GISS until 1935, but all of it since 1975 and Gisstemp keeps shooting up unlike the other data sets!!”
No. The method Frank uses to determine the percentage of land to sea surface data is faulty. He has been informed of this. In order to use the methods he’s using, he would have to mask the land surface data where it extends out over the oceans. This should be clear in my rebuttal that Anthony posted immediately after this one:
http://wattsupwiththat.com/2010/07/17/giss-land-and-sea-ratios-revisited/
I created the following maps of GISTEMP combined land and sea surface temperature data (1200km radius smoothing) with the KNMI Climate Explorer to help illustrate this. And I used KNMI as a third party source of GISS data. Here’s a map of the GISTEMP data in 1915:
http://i27.tinypic.com/20kpq1v.jpg
Areas in white have no data. The only major areas with data missing are the Antarctic and Southern Oceans and a part of the Arctic, which GISS treats as land, and some of Africa and South America. So GISS is presenting land surface data for a significant portion of the global land mass, not 0%.
And here’s a map for 1945:
http://i26.tinypic.com/2e49f6e.jpg
And one for 1975:
http://i29.tinypic.com/2ym90fd.jpg
The major difference between 1945 and 1975 is the addition of Antarctic data and that has no impact on global temperature since the linear trends of the GISTEMP data with and without the Antarctic and Southern Ocean are the same:
http://i30.tinypic.com/11awlqb.jpg
And one more map, 2005:
http://i30.tinypic.com/yj5ue.jpg
Can you see the significant changes in the percentage of land surface temperature data that you described above and that Frank Lansner presents in the post?
tallbloke says: “Great analysis. I said on a GISS jump thread last year that I thought Jimmy James had started bumping up the lower toposphere ratio. Looks like that hunch was right, but this properly done study carries far more weight.”
Unfortuately, the method used to calculate percentage of land to ocean in this post is wrong, hence my rebuttal post which follows this one:
http://wattsupwiththat.com/2010/07/17/giss-land-and-sea-ratios-revisited/
Listen! Please! You just don’t install multi-million dollar weather stations anywhere. There has to be a couple three good reasons. The major factors in location are: Population Centers >1M, International Airports, Amenities for NOAA Personel. We’re in the middle of the Great Recession, we can’t afford what most of you’re suggesting. The US is becoming a Third World nation and the Chinese won’t lend us the money to put one of their thermometers just anywhere.
I get really annoyed at people who never look at NASA code to figure out what NASA is doing. back in 2007 a bunch of us hounded NASA to get them to release the code.
it was a success.
Several of us went over that code with a fine toothed comb. There is nothing much of note there, except some clarifications that the papers were too vague on. Getting the code running was a bitch ( the guys I was working with gave up after a couple weeks, it was basically compiler issues and OS issues) Nick barnes and company had more skill and patience: They succeeded. If you havent read their code ( I have) then you should probably keep your mouth closed. As a data and code libertarian, nothing is more annoying that people who talk about code and data that is available without ACTUALLY reading it or using it.
If you read the code and ran the code you would understand the mistakes in this post. For the short version, see Zeke’s post at Lucia’s
I will say it again. The key issue is metadata and adjustments. put your brain cells on that.
Bob Tisdale:
You write: “The method Frank uses to determine the percentage of land to sea surface data is faulty. He has been informed of this. In order to use the methods he’s using, he would have to mask the land surface data where it extends out over the oceans. ”
You are in title to your opinion.
The official GISS LAND data is weighted around zero in 1900-20.
This you see because the GISS Land+SST lies on top of the HADISST graph:
http://hidethedecline.eu/media/GISSglobal/fig1b.jpg
I havent seen any of your writing anywhere near explaining this. Weight zero means all ocean areas + all land areas from the GISS LST is hardly weighted at all. So when you just show the minor variance in giss LST ocean cover you are 10.000 miles from explaing things.
Then i pointed out that in 1980 the GISS land weight was around 40% while in 1995 it was 73%.
So in 15 years, the lad date use are allmost doubbled.
So now, Bob, will you please show me the maps where GISS LST(land+ocean area) is alomost the doubble in 1995 than 1980?
I havent seen anything just remotely exolaing this grotesque increase of GISS LST (land0ocean) in any of your maps or writings.
So please (!) come up with the actual explanations for the grotesk increase of GISS land before you conclude anything.
K.R. Frank
The IPCC uses one tidal gauge mounted on sinking geology in Hong Kong Harbor as the bellweather of the world’s ocean level changes, and Michael Mann uses the tree rings from one pine tree in the Yamal Pennisula to construct his FrankenGraph Hockey Stick. So, why are we to be surprised that NASA GISS is increasing the weighting of land based measurements when the land area remains constant? Who will be around 30 years from now – when the Thames and Hudson Rivers are frozen over in the middle of the Grand Solar Minimum – to remind us of these charlatains that twisted the climate science to promote their social agendas? We will be too cold and too hungry from lack of fuel and food to care.
Frank Lansner says: July 19, 2010 at 2:54 pm: This appears to be the same comment you left for me a Jo Nova’s thread here:
http://joannenova.com.au/2010/07/did-giss-discover-30-more-land-in-the-northern-hemisphere/
So I’ll repeat my reply to you from there, just to make sure you receive it. Note that the comment numbers refer to the Jo Nova post, not the thread here at WUWT.
##############
Frank Lansner # 104: You replied, “You are in title to your opinion.”
Actually, Frank, you agreed with me in your comment # 85, when you replied, “On this basis, then you are correct, it would make sence to use time and energy to go further like you advise, i really totally agree with you on this (mask ocean area year fo year and other studies).”
You replied, “The official GISS LAND data is weighted around zero in 1900-20.
This you see because the GISS Land+SST lies on top of the HADISST graph:
http://hidethedecline.eu/media/GISSglobal/fig1b.jpg”
Frank, you can’t simply eyeball these things. Subtract the HADISST from the GISTEMP Data. The difference is the impact of the GISS Land Surface Temperature anomalies. I did it for you in my response to your reply to Steven Mosher. See:
http://wattsupwiththat.com/2010/07/17/giss-land-and-sea-ratios-revisited/#comment-433545
*****
You wrote in that thread, “In the present case for example we have a GISS LST+SST that is the same as the SST (HADISST) 1900-1920.”
And I replied, What’s the source of your data? A quick trip to the KNMI Climate Explorer shows that to be incorrect. There are significant differences between Global GISTEMP LST + SST (1200km radius smoothing) and Global HADISST from 1900 to 1920:
http://i25.tinypic.com/2mi14bm.jpg
And here’s the difference (GISTEMP MINUS HADISST):
http://i25.tinypic.com/2rwrbpg.jpg
******
You wrote, “I havent seen any of your writing anywhere near explaining this.”
Actually, you stopped in and left a message at the post at WUWT that was a rebuttal to yours, which Anthony posted immediately after yours. Here’s a link to your comment on that thread:
http://wattsupwiththat.com/2010/07/17/giss-land-and-sea-ratios-revisited/#comment-433362
You wrote, “So now, Bob, will you please show me the maps where GISS LST(land+ocean area) is alomost the doubble in 1995 than 1980?”
I can’t. Why? Because it hasn’t doubled. Only your erroneous calculations make you believe it has doubled. Here’s a gif animation of GISTEMP Combined Land And Sea Surface Temperature anomalies with 1200 km radius smoothing for January 1980 and January 1995. White represents areas with no data, and the only land surface area without data is the small area along the shore of Antarctica, south of the South Pacific.
http://i31.tinypic.com/33zfo5f.jpg
Other than that, the land surface area coverage is complete and it hasn’t changed.
I’ve asked you for the method you use to calculate this growing difference in land surface data percentage, and as of now, I have not found your reply. So I’ll ask, are you assuming that the growing difference between Global Temperature Anomalies and Sea Surface Temperature Anomalies is caused by an increase in the amount of land data? If so, it’s not. Land surface temperature anomalies rise significantly faster than sea surface temperature anomalies during warming periods and cool faster than SST anomalies during cooling epochs. If you’re not making that assumption, how do you account for the naturally growing difference between land and sea surface temperature anomalies with your calculations?
Oops. The opening of the above reply to Frank Lansner should have been presented differently. Let’s try the opening again:
Frank Lansner: With respect to your July 19, 2010 at 2:54 pm comment, this appears to be the same comment you left for me a Jo Nova’s thread here…
Yup, that’s a much clearer opening.
Hi Bob!
I wrote yo you guys at
http://rankexploits.com/musings/2010/the-gistemp-land-fraction/comment-page-1
My comment 49368 time 6.04 i cant see, await moderator
my comment 49371 time 6.22 i cant see, await moderator
For some reason my coments are not shown yet – unlike yours – but i would very much take the debate there since it seems.
I hope that my coments will come up soon.
K.R. Frank