You can thank Chris Horner of CEI for making this happen.
In August 2007, I submitted two Freedom of Information Act requests to NASA and its Goddard Institute for Space Studies (GISS), headed by long-time Gore advisor James Hansen and his right-hand man Gavin Schmidt (and RealClimate.org co-founder).
I did this because Canadian businessman Steve McIntyre — a man with professional experience investigating suspect statistical claims in the mining industry and elsewhere, including his exposure of the now-infamous “hockey stick” graph — noticed something unusual with NASA’s claims of an ever-warming first decade of this century. NASA appeared to have inflated its U.S. temperatures beginning in the year 2000. My FOIA request asked NASA about their internal discussions regarding whether and how to correct the temperature error caught by McIntyre.
NASA stonewalled my request for more than two years, until Climategate prompted me to offer notice of intent to sue if NASA did not comply immediately.
On New Year’s Eve, NASA finally provided the Competitive Enterprise Institute (CEI) with the documents I requested in August 2007.
…
Regarding U.S. temperatures, Ruedy confessed to Hansen on August 23, 2007 to say:
I got a copy from a journalist in Brazil, we don’t save the data.
——————————————–
The Ruedy relationship with a Brazilian journalist raises the matter of the incestuous relationship between NASA’s GISS and like-minded environmental reporters. One can’t help but recall how, recently, the UN Intergovernmental Panel on Climate Change (IPCC) claim of glacier shrinkage in the Himalayas was discredited when found to be the work of a single speculative journalist at a popular magazine, and not strict peer-reviewed scientific data. The emails we obtained include several instances of very close ties and sympathetic relationships with journalists covering them.
…
In an August 15, 2007, email from Ruedy to Brazilian journalist Leticia Francisco Sorg, responding to Sorg’s request for Ruedy to say if warming is accelerating, Ruedy replied:
“To observe that the warming accelerates would take even longer observation times” than the past 25 years. In fact, it would take “another 50-100 years.”
This is a damning admission that NASA has been complicit in UN alarmism. This is not science. It is debunked advocacy. The impropriety of such policy advocacy, let alone allowing unsubstantial scientific claims to become part of a media campaign, is self-evident.
More here.
================
Climategate 2.0: The NASA Files
In August 2007, Christopher Horner of the Competitive Enterprise Institute (CEI) submitted two Freedom of Information Act requests to NASA and its Goddard Institute for Space Studies (GISS), headed by long-time Gore advisor James Hansen and his right-hand man Gavin Schmidt (and RealClimate.org co-founder).
Canadian businessman Steve McIntyre — a man with professional experience investigating suspect statistical claims in the mining industry and elsewhere, including his exposure of the now-infamous “hockey stick” graph — noticed something unusual with NASA’s claims of an ever-warming first decade of this century. NASA appeared to have inflated its U.S. temperatures beginning in the year 2000.
The FOIA request asked NASA about their internal discussions regarding whether and how to correct the temperature error caught by McIntyre.
NASA stonewalled for more than two years.
Quietly — on New Year’s Eve 2009 — NASA finally provided the documents:
Read Christopher Horner’s analysis of the documents here.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
To those who might not be familar with Gilestro’s work, he claimed to show that adjustments are symmetric around zero, and his published histogram was very tightly squeezed so the distribution looked like a needle. Blowing up the figure shows the skewing.
He didn’t realize that even his analysis showed the adjustements accounted for 15% of the reported warming in the 20th century.
That last sentence in John M (18:27:00) right above the link is mine, not Gilestro’s.
carrot eater (22:11:52) : f they just saved the GHCN and USHCN source files they used for each month, they could just recreate whichever calculation they ever did. Perhaps they should do that.
Nope. The programs and hardware (and compilers) too mutate over time. Such as the 15 Nov 2009 date when they swapped from USHCN to USHCN.v2 input (with USHCN.v2 being adjusted “warmer” than USHCN and with 1/10 F instead of the silly 1/100 F precision and with a boat load of individual data items changed) that required a change to the GHCN / USHCN merger code. Oh, and a “STEP6” was added. I’ve also demonstrated a compiler dependent bit of code that warms 1/10 of the records by 1/10 C in the F to C conversion. So depending on what compiler you use, you get different answers…
And if they FIXED that code, then you have a code change…
So it’s not enough just to save the input. You also need the code. And the compilers. And the hardware…
Or, of course, they could just have saved 45 MB of data (roughly) from each run… just about what would fit a few times over on a single “round tape” of the era. So they could have saved a bunch of the log files and intermediary files too. Cost was about $9 each IIRC.
Kate (00:15:29) : The latest independent analysis of world climate data by acclaimed skeptic blogger ‘Chiefio’ (aka E. M. Smith) and his blog contributors confirm that the Global Historical Climatology Network (GHCN) has cynically dumped the world’s second oldest and reliable climate record at Prague in the Czech Republic for no scientific reason.
It is important to point out that while the thermometer was dropped as of January 2010 it can come back later, maybe. That is, it could be a Zombie Thermometer:
http://chiefio.wordpress.com/2010/02/15/thermometer-zombie-walk/
And we get to wait a few weeks, or months, or maybe even a few years, to see the ultimate fate. (One thermometer in Canada came back after several years of being ‘walkabout’ so who knows how many “dead” thermometers are really just Zombies…)
This has the interesting side effect that some thermometers are constantly returning, but just a few months late. This might allow for the current month to always maintain a “delta” to prior months… Such an interesting tool the Zombie could be. (So much to analyze, so little time…)
So it’s possible that the oldest Czech thermometer might be dead, or it might “just be resting” for a while… A Zombie to walk the earth again, in just a few weeks… or months…
Andrew30 (00:32:55) :
When Gavin Schmidt say something like this:
“The other factor might be that lampasas is overall cooling, if we use climatology to infill in recent years, that might give a warm bias. ”
Is this type of statement ’significant’?
Oh Yes!
One of the frequent spears thrown at me is the assertion that “the anomaly will save us” and that there is no way changing the composition of the thermometer inventory could possibly cause the anomaly map to change. yet here we have Gavin himself saying just that.
“Climatology” is the “Jargon” used to differentiate the Climatology Anomaly Method of “homgenizing” and “in fill” from the other methods. So you have First Difference as one approach “Climatology” as the other. (I have a reference to a paper somewhere…)
Ah, here it is:
http://homepage.mac.com/williseschenbach/.Public/peterson_first_difference_method.pdf
It also tossed a couple of rocks at The Reference Station Method used in GIStemp ( GIStemp does some RSM and some CAM when it does offsets and anomalies so it gets the worst of both worlds, IMHO).
Worth a read to figure out the jargon of things like “climatology” if nothing else…
This statement amounts to an admission that GIStemp certainly IS sensitive to using ‘Fill In’ for missing stations via the “Climatology” (i.e. Anomaly and Reference Station Method) technique and means that “thermometer change matters”.
You have no idea how happy that one email makes me 😉
PS. It makes me think that ‘climatology’ as some kind of spice to warm up a meal, or a trick for adding a warm bias to data?
Am I mistaken?
Nope.
The only question is this: Given that they clearly KNEW it could be done, did they do it as a voluntary act or as an accident (via ignoring it)? Sin of omission or commission? Well; and maybe the follow on question of “If voluntary, was it done with intent?”.
But that they knew a bias could be introduced is clear. Only in doubt is what was done with this knowing.
John M (18:27:00) :
First off, the contention here was that adjustments only ever go one way. This is clearly not the case; the distribution is very nearly symmetric. Hard to imagine that homogenisation is some sort of intentional manipulation, with a distribution like that.
Yes, the mean is not exactly zero. This was noted and discussed, despite your assertion. The mean is small enough that one would not expect it to make much difference. And indeed, it does not.
See here. Look at the two graphs under Q4. Global average using raw and adjusted.
http://www.ncdc.noaa.gov/cmb-faq/temperature-monitoring.html
Of course, even if it did show some difference, that’d be fine if the adjustments were justifiable. But as it is, the difference due to adjustment is negligible.
Note that these are global data.
John M (18:30:33) :
“He didn’t realize that even his analysis showed the adjustements accounted for 15% of the reported warming in the 20th century.”
You don’t have sufficient information there to determine that.
But once you make the global mean plot, you can see the difference. No such 15%. See the plot at NOAA.
E.M.Smith (19:37:55) :
Yes, yes. They’d have to archive the code, as well. That’s something they should do anyway; I don’t know if they do. Or they could have archived all the results. But they didn’t.
E.M.Smith (20:16:07) :
“This might allow for the current month to always maintain a “delta” to prior months… Such an interesting tool the Zombie could be.”
That really doesn’t make much sense. Whatever change there is when those stations get added, it catches up to you when they are added. So the trend map at any given point in time would be unaffected, so long as you ended your trend about 1 year before the current time. It’s the trends that matter.
Adding stations will have some impact, yes. If you add them to an undersampled part of the world, it will have more impact than adding them to an oversampled part. But you can sit there and watch how much each month’s anomaly (and more importantly, the trend) changes as those stations get updated over time.
“One of the frequent spears thrown at me is the assertion that “the anomaly will save us” and that there is no way changing the composition of the thermometer inventory could possibly cause the anomaly map to change.”
That is untrue, or at least, if anybody throws that spear at you, it is an untrue spear. The point everybody is making is that simply removing a station in a cold place will not make the record warmer. That much is very true, and is directly aimed at points you have made. On the other hand, removing a station that had different trends from its near neighbors will make a difference. To the extent that no two stations are 100% perfectly correlated, removing or adding a station will always have at least a tiny effect at that grid point.
“This statement amounts to an admission that GIStemp certainly IS sensitive to using ‘Fill In’ for missing stations via the “Climatology””
FILNET had a very slight impact for the US. That was never hidden. But for certain individual stations, it could make a big impact. Perhaps at this station, it was doing something unreasonable (if in fact it’s FILNET they’re talking about, I can’t tell)
“(i.e. Anomaly and Reference Station Method) technique and means that “thermometer change matters”.”
You’re mixing things up here. Of course things like FILNET and homogenisation matter; otherwise you wouldn’t bother doing them. Though in the end, the net impact is small compared to the trends.
And yes, thermometer change matters for the reasons I outlined above. That was never under dispute. It was the idea that simply removing high latitude/high altitude stations causes warming because they are cold – it’s that idea that’s heavily disputed, because it doesn’t make any sense when you use anomalies.
RockyRoad (10:53:28) :
You might be onto something. I was hoping you’d say something like “analysis is flawed” because that opens up the whole CO2-absorption mythology.
Unfortunately your source for this is flawed, and is wrong on the physics of emission of radiation by gases.
RichieP (16:33:49) :,
I was blogging this and wanted to give the author credit. Instead of just saying Author. The reason for bylines is to make that easy. And Anthony a while back said he was going to do that.
And exactly how was I to know that a link at the bottom of the piece was to the original? Well you got the link love anyway. I credited the piece to Author.
Long gone is the day when we could trust governmnet for even the simple things
carrot eater (20:52:56) : And yes, thermometer change matters for the reasons I outlined above. That was never under dispute. It was the idea that simply removing high latitude/high altitude stations causes warming because they are cold – it’s that idea that’s heavily disputed, because it doesn’t make any sense when you use anomalies.
Such magnificent singing and dancing. Yet another example of “once shown to be” a roundly “disputed” problem becomes “always known”… Priceless.
Jones must be giving “of course the MWP existed” and “We knew the hockey stick was wrong” lessons 😉
BTW, I can explain EXACTLY why and how “high altitude” and “high latitude” station deletions / drops do, in fact, cause a warming influence. It is only your understanding of it that “doesn’t make any sense”. It all comes down to the way anomalies are used and a neat little overlooked temperature artifact / behaviour. Once you allow anomalies to be “basket of thermometers A” vs “basket of thermometers B” at a different time, all sorts of “unexpected side effects” can leak through.
If you actually think about it for a few months, you too can figure it out. If you chose to assume “it can’t be” you never will.
I’m ‘kicking the idea’ around with some (few, limited) folks with an eye to what to do with it… so no more will be posted on that point. For now it must stay a “teaser”.
Why, you have to wonder, are NASA, GISS, CRU et al all SOOOOO frightened of historic temperature data (see Kate’s blog above)..?
You don’t think it could be because it shows zero to negligible warming over two or three centuries, do you..?
Nah – that’s just a cynical skeptics view…
E.M.Smith (00:14:58) :
“Yet another example of “once shown to be” a roundly “disputed” problem becomes “always known”… Priceless. ”
____
Ever since 1987, at least, GISS has been calculating error bars due to incomplete spatial coverage. So yes, I do think the issue of incomplete spatial coverage has been known for a while. I don’t know how much more clearly to put that.
They also, in that 1987 paper, examined other ways of combining the stations, as well as decided to drop stations with less than 20 years of overlap with the neighbors. They admitted that in their method, the results could vary slightly if they used a different station as the first one into the basket. So they’ve known the method of calculation has some impact on the results.
___
“BTW, I can explain EXACTLY why and how “high altitude” and “high latitude” station deletions / drops do, in fact, cause a warming influence. It is only your understanding of it that “doesn’t make any sense”. It all comes down to the way anomalies are used and a neat little overlooked temperature artifact / behaviour. Once you allow anomalies to be “basket of thermometers A” vs “basket of thermometers B” at a different time, all sorts of “unexpected side effects” can leak through.”
___
Each station is added to the ‘basket’ after adding or subtracting an offset. Once you do that, there is no memory of whether it was a hot or a cold station. Only whether it was warming or cooling. Now, this could get dicey if the periods of overlap are not long enough to calculate the proper offset (especially if the stations don’t correlate well with each other), which is why they exclude those with too little overlap.
carrot eater (10:37:27) :
Alan the Brit (10:13:25) :
How can you say all that, when the example at hand is inconsistent with all that?
A journalist wanted to know if it’d be correct to say global warming is now accelerating, and GISS told them, no, you can’t say that because it isn’t.
This is an example of the person at GISS doing the right thing. So why is it being highlighted as a bad thing?
My final word on the subject in this post. Telling the truth is ALWAYS the right policy in the end, although I would concede that the truth is open to interpretation in many cases. Yes he said the right thing, but if he or his colleagues said something different in public then that is politics, pure & simple, & politicians are always looking for a cause célèbre & a soap box from which to shout it.
HAGWE carrot eater, & the same to everyone else in Blogistan! :-))
Alan the Brit (02:15:07) :
Have your last word, but it seems to me that you allow there is nothing wrong with this email.
“but if he or his colleagues said something different in public”
IF they ever said that the temperature trend was currently observed to be accelerating, then that would be wrong. Have they ever done so? Seems a little unfair to just leave that hypothetical out there, as if they might have done. People have maybe talked about possible acceleration in other things, but I’m not aware of GISS people specifically claiming a current acceleration in temperature.
carrot eater (20:28:30) : and carrot eater (20:31:30) :
Point taken. But I merely pointed out what gg claimed and asked what you thought of it. I also pointed out that the distribution was not symmetrical and that stretching the graph out helped to visualize that point. Do you dispute that?
What is there not enough information for?
That gg says the average adjustment average is 0.017deg/decade?
That he erroniously says that warming from 1900-2000 was 0.2 deg/decade?
That the actual warming from 1900-2000 was .065 deg/decade?
http://www.woodfortrees.org/plot/hadcrut3vgl/from:1900/to:2000/plot/hadcrut3vgl/from:1900/to:2000/trend:2000/plot/none
That 0.017/.065 x 100% is 15%? (well, actually, you’ve got me there, it’s 26%)
That he didn’t understand his error wrt to 0.2 deg/decade, even after having it pointed out to him at least twice?
What information do we lack?
If the NOAA analysis shows the adjustments are much smaller than that, take it up with gg, not me.
But as long as you mention the NOAA analysis, where are their results for 1900-2000?