The results are finally in. Year-end numbers have been posted by the three organizations which, as far as I know, publish online the most relied-upon global average temperature estimates. NASA GISS posted their year-end results first, around Jan. 9, NCDC updated about Jan. 15, and HadCRU has finally updated their global-average files to reflect results from 2007. Let’s take a look.
Each organization reports a global (land+sea) average temperature anomaly for each month. HadCRU data begin in 1850, GISS and NCDC don’t begin until 1880. Let’c first take a look at annual averages during the period of overlap, from 1880 to the present:
It’s obvious that the HadCRU numbers are consistently lower than GISS and NCDC. If fact they’re all three on different scales. This is because they all use a different reference period. GISS anomaly is the difference between a given month’s average and the average for that month during the reference period 1951 to 1980; NCDC anomaly uses the reference period 1901 to 2000; HadCRU uses the reference period 1961 to 1990.
It turns out that global average temperature during the GISS reference period (1951 to 1980) was about the same as the average during the NCDC reference period (1901 to 2000), so those two series are approximately on the same scale. But during the HadCRU reference period (1961 to 1990) the planet was about 0.1 deg.C warmer. Hence all the HadCRU anomaly numbers are about 0.1 deg.C lower, because they’re being compared to a “higher standard.” This simplified picture — GISS and NCDC on the same scale with HadCRU about 0.1 deg.C lower — is only approximate. If we want to be precise we need to choose a common reference for all three series (and so we shall).
But as it is, let’s use our approximate alignment, and show the same graph but with the HadCRU numbers shifted upward by 0.1 deg.C. We can accomplish this by plotting them on a different axis. We’ll use the left-hand axis for GISS and NCDC, the right-hand axis for HadCRU, and align that axis so that it’s 0.1 deg.C higher:
This shows that our approximate shift, adding 0.1 deg.C to HadCRU, brings all three temperature estimates closely into alignment. We also see that they all show roughly the same long-term trend, together with similar noise (the short-term ups and downs above and beyond the trend). I’ve also outlined the four “episodes” of temperature evolution during the period of common coverage: 1880-1910, 1910-1946, 1946-1975, and 1975-present.
But it really would be better to bring them all into alignment using a common reference period. For this purpose I chose the period 1950 to 1980, which is not quite the same as the 1951-1980 GISS reference period. But it’s a reasonably modern 30-year period, so I computed the average anomaly of each time series during this reference period and subtracted that from each time series, so that the remainder, the common-anomaly series, has an average anomaly exactly zero during the reference period. This enables us to compare the time evolution of each series directly:
This graph really isn’t very different from the one before it, which emphasizes that the earlier approximation (HadCRU numbers about 0.1 deg.C low relative to GISS and NCDC) is actually a pretty good approximation.
Let’s also look at each of the “episodes” I’ve delineated in the data. To make comparison between them a little easier, I’ll plot each one on the same y-axis scale: the y-axis height will be 1 deg.C. The most recent episode is what I like to call the “modern global warming era,” from 1975 to the present:
Clearly the three records are in very good agreement. But there are differences, notably the extreme high temperature reported by HadCRU in 1998 (but not quite so high in GISS or NCDC). We also see that for last year (2007), GISS was highest, NCDC in the middle, and HadCRU lowest. The GISS/HadCRU difference seems to be because HadCRU omits the arctic while GISS estimates it by interpolation. In that case, the noted result is what we would expect: if it’s true that the arctic is the fastest-warming part of the planet, and HadCRU omits that fastest-warming region, we would expect their estimate to be low. If anyone knows the details of how NCDC estimates (or doesn’t) data for regions like the arctic, feel free to volunteer that information.
The next most-recent interval is the post-war era, from 1946 to 1975. During this period, all three series show no real long-term trend in temperature at all:
In fact the numbers determined by fitting a regression line to the data are very slightly positive for all three data sets, but in no case is the indicated warming statistically significant.
Next comes the early-century warming period:
Again all three series show notable warming, but we can also see that for the early years of of this period the HadCRU estimates are actually a little lower than GISS or NCDC, so HadCRU actually shows the fastest warming.
Finally we have the early period of overlap, from 1880 to 1910:
Again the data are similar, but the differences are greater than for later years. This indicates that the uncertainty in estimated global temperature is greater 100 years ago than it is today. Surprise! Even so, the general agreement of the three series is encouraging.
We can compute the warming rates for each time period, for each data series, and here they are:
| Time Span | GISS | NCDC | HadCRU |
| 1975-2008 | 0.0181 | 0.0178 | 0.0183 |
| 1946-1975 | 0.0013 | 0.0022 | 0.0045 |
| 1910-1946 | 0.0129 | 0.0132 | 0.016 |
| 1880-1910 | -0.0008 | -0.0047 | -0.0068 |
Finally, for those who want to place bets on whether or not global warming has stopped I’ll offer a suggestion how such a wager should be settled. For each data series, use only the data from 1975 to 2000 to determine a regression-line fit. Then superimpose the regression-line fit on a plot of the actual data, and extend the regression line beyond 2000 into the future. So far, we have this (dotted lines are regression-line fits):
It certainly seems as though all three data series have evolved in a way which is consistent with a continuation of the 1975-2000 trend; global warming has *not* stopped yet (at least, there’s no evidence that it has). We can extend the regression lines even further into the future, and add some grid lines, to get this:
For each data series, for each year, we can compute the residual, which we’ll define as the difference between the estimated temperature and what it would have been if it exactly followed the regression-line fit. Here are the residuals so far:
Part of the reason HadCRU shows a more negative residual for 2007 than the other series is that its temperature estimate is lower. But another part of the reason is that its regression line is (slightly) steeper; it has more warming to “live up to” than the other series.
If all three series show a statistically significant (at the 95% confidence level) departure from the 1975-2000 trend, we can reject the hypothesis that global temperature is continuing its present trend. We can then conclude that global warming has either slowed, or stopped, or reversed itself, or accelerated, depending on the direction of the departure from the present trend.
If all three series show continued warming with a statistically significant departure from being flat, then we can reject the hypothesis that global warming has stopped. If the departure is downward, we can conclude that warming has reversed; if it’s upward, then we can conclude that it never really stopped.
By 2015, the expected temperature from the regression-line fit and that expected from the “no change” hypothesis will be far enough apart that we’ll probably be able to distinguish between them with statistical significance. In other words, by 2015 either we’ll know that global warming has changed (possibly stopping, possibly reversing), or there’ll be no more of this “global warming stopped in 1998″ malarkey.
It’s entirely possible that the numbers may give us statistically significant evidence even before 2015. If so, I’ll report the result. If it turns out that global warming is not continuing (which I seriously doubt), then I’ll readily admit that I was wrong. In fact, I’ll be keeping a close eye on the future evolution of global temperature and actively looking for such results, so if we do get valid evidence that global warming has stopped, I just might be the *first* one to say so.
If 2015 rolls around, and temperature have risen above present-day levels by enough to be demonstrably significant, I’ll announce that too. Will those who have so often chanted the “no more global warming” mantra admit that they were wrong? Somehow, I doubt it. I suspect that instead, they’ll be flooding blogs, newspapers, magazines, and Faux News reports with claims that “global warming stopped in 2013.”
150 responses so far ↓
Eli Rabett // January 24, 2008 at 9:59 pm
I have a slightly different take on this. If you compare the interannual variability of GISS, HadCRU UAH and RSS since 1979, HadCRU appears not to vary as much on a year to year basis esp in recent years although 1998 is a interesting outlier to this pattern.
Dodo // January 24, 2008 at 10:57 pm
Nice analysis, thank you very much. But why do you make those smug remarks in the last paragraph? I mean, your 2015 commitment is a laudable one, from a skeptic’s perspective, and I think many would be willing to join your challenge and admit that the debate must be settled by then. Surely, if rapid warming returns, it falsifies much of what many skeptics are claiming. And if it cools…
Unfortunately you will not get much resonance with that attitude of yours. Why not try some politeness towards the bad guys? You do know that, most of them have never denied the existence of a warming trend since 1975. But if you make it since 2001 and look at the satellite measurements, you have a different (short) story, so far.
[Response: Why not just say that "if you make it since November 2007 then you have a different story, so far"? Would that be a valid argument? Would that even constitute evidence -- let alone proof -- that global warming stopped in December 2007? Why not?
What I said in the last paragraph is what I believe to be true. My guess is that denialists will refuse to admit the truth then, just as they do now. This is evidenced by your repeating what they've made their latest mantra: that global warming stopped in 1998 (or 2001, or whatever). Look at the final graph, of residuals for each data set from the trend determined by its 1975-2000 data: is there ANY evidence whatever -- even a *hint* -- of a departure from the existing trend? No. Not in GISS data, not in NCDC data, not even in HadCRU data. But denialists have taken advantage of the statistical naivete of the general public to create the *impression* that global warming has stopped. Enough of them have statistical training that they *know* this is a faulty conclusion, which makes it deliberate deception rather than simple mistakenness. But they still tell this story far and wide. You appear to have swallowed it.
My original question is a valid one, which you should answer if only for yourself: is declining temperature from November to December 2007 evidence that global warming has stopped? If not, then why not?]
David B. Benson // January 24, 2008 at 11:31 pm
From about 2002 until just now solar irradiance was decreasing. I have no idea how much (short term) difference this made.
Martin // January 24, 2008 at 11:36 pm
Thank you for a fascinating article. There is nothing like real facts and careful analysis to clear the mind after a lot of nonsense on other boards and threads.
One small question: the Arctic above 80N is only a pretty small fraction of the earth’s surface (less than 1% I think). I wonder how GISS figures for such a small proportion of the planet could cause the discrepancy with the HadCRUT figures. Am I right in saying that the satellite measurements don’t cover the poles either?
[Response: The arctic above 80N is indeed less that 1% of the earth's surface. But I'm not sure that that's the only region ignored by HadCRU (and estimated by GISS), even in the far north. In fact, the HadCRUT3gl.txt file (HadCRUT3 for the globe) lists the fraction of the earth's surface which is covered by their analysis, and for 2007 it was 82%, so that fully 18% of the globe isn't estimated by HadCRU.
Satellite measurements also don't cover the poles, extending (for the lower troposphere) from about latitude 70S to 82.5N.]
Lab Lemming // January 25, 2008 at 12:15 am
How can a satellite orbit cover different latitudes in different hemispheres?
[Response: Good question, I don't know. Might it have to do with the ice cover in Antarctica making observations difficult or inaccurate below 70S?
Response #2: Checking again, I find that RSS reports 70S to 82.5N for the lower troposphere (where surface terrain may have its biggest affect), but for the mid-troposphere and stratosphere they report 82.5S to 82.5N.]
henry // January 25, 2008 at 4:08 am
[But it really would be better to bring them all into alignment using a common reference period. For this purpose I chose the period 1950 to 1980, which is not quite the same as the 1951-1980 GISS reference period.]
Thank you, thank you, thank you.
Finally, a real look at what the various data sets look like when they all use the same reporting period.
From this perspective, we are now able to see exactly what the difference is between the HadCRU and GISS, and just how much “residual” there is when the extrapolation for arctic temps is used (by GISS).
It still appears close enough, and some groups can now look at a particular “outlier” year to check their data (comparing their raw data with others to see if the difference is there, etc).
Also, number of reporting stations could be looked at (more or fewer than those in another series).
Still, good to finally see a “common ground” being used.
henry // January 25, 2008 at 4:11 am
[In fact, the HadCRUT3gl.txt file (HadCRUT3 for the globe) lists the fraction of the earth’s surface which is covered by their analysis, and for 2007 it was 82%, so that fully 18% of the globe isn’t estimated by HadCRU.]
Yet the difference in temps between HadCRU and GISS is in the hundereths or thousandths of a degree.
cce // January 25, 2008 at 5:14 am
The difference in coverage between the Hadley/CRU (Jones) and GISS(+Reynolds) analysis is shown in Plate A1 of Hansen’s 1999 paper:
http://pubs.giss.nasa.gov/docs/1999/1999_Hansen_etal.pdf
Effect of removing portions of the Arctic (but not Antarctic) by latitude:
http://data.giss.nasa.gov/gistemp/graphs/ArcticEffect.pdf
My attempt at comparing different analyses,and explaining several skeptical attacks:
http://cce.890m.com/part06
(about 18 minutes)
I need to update to RSS 3.1, and when describing the interpolation of the Arctic I refer to trends instead of anomalies. In the interest of eliminating any other errors, I’d appreciate any input that anyone has.
fred // January 25, 2008 at 8:47 am
Nice post. Thanks. One can see how much work goes into this, and its appreciated. Informative stuff.
Though to me the question has never really been is it warming, but rather why and through what mechanism and to what extent unprecedentedly.
Dodo // January 25, 2008 at 9:54 am
Can’t you read? As a moderate skeptic, I was willing to accept your 2015 challenge, and I think you should offer it (politely) to others as well. But of course, it is much more fun to push the controversy, isn’t it?
JCH // January 25, 2008 at 3:04 pm
Fred, why is it warming? What mechanism do you think is behind it? Why do you think it matters at all whether it is unprecedented or not?
Hank Roberts // January 25, 2008 at 4:03 pm
Slightest of typos in first post:
“If fact …” should read “In fact …”
> politeness
Tamino isn’t merely irate with people who should know better. He’s irate with people who know better, and who lie convincingly to the public, and who treat scientific work as an opponent to be beaten rather than a shared source of enlightenment.
When you’re in the dark and there are grues, beware those who attack your sources of light.
“See an amazing and cogent review of our energy and climate situation, by Dr. Steven Koonin, chief scientist of BP and my former Caltech classmate. I cannot over-emphasize how useful and important Koonin is!
http://www.youtube.com/watch?v=wt_mluFK7xk
Here you have one of the smartest guys anywhwere, chief scientist for a major oil company, laying it all out and demolishing the “let’s ignore climate change” mantra of the right. Perfect “ostrich ammo.” He’s done an even better presentation that I’ll link-to, when it’s available online.”
— David Brin http://davidbrin.blogspot.com/
Hank Roberts // January 25, 2008 at 4:15 pm
Please take the time — 80 minutes, can you afford it? — to watch that movie.
Then check David Brin’s blog — his site’s now blocked from access by US government employees at work. Guess why.
“I have called upon the professionals of the civil service, the intelligence services, the many agencies of law and accountability, the scientific community and the U.S. military officer corps, to remember their oaths — to protect the people from all enemies, foreign and domestic.
That’s it, really. My ‘inappropriate’ sin. Pointing out that our professionals - both in and out of government - have been the top victims of the neoconservative putsch. I have praised the skilled men and women who dedicate their lives to public service, and made it clear - in bold but meticulously legal terms - that they do not have to put up with being bullied by incompetent, dogmatic hacks, who have oppressed them in a concerted campaign of intimidation, ever since the Bush Administration began.
All right, I did more than that. I also called upon Democrats - and decent conservatives who remember true patriotism - to make this a centerpiece issue, not only because it is just and right, but because no other accusation (backed by proof) could damage the current GOP ruling cabal more than this one.”
— David Brin http://davidbrin.blogspot.com/2007/12/this-blog-blacklisted-for-demanding-we.html
Barton Paul Levenson // January 25, 2008 at 4:29 pm
Dodo posts:
[[But of course, it is much more fun to push the controversy, isn’t it?]]
What controversy?
Jim Arndt // January 25, 2008 at 5:39 pm
Hi,
If you take the current trend of no or little significant warming since 1998 you could conclude that the warming stopped in 1998. If the cooling continues til 2015 will the AGW movement accept that? I think your right in that the skeptic side will have its “denial” if the temperatures start to rise again. Considering the sun is in a low activity cycle ( and if it continues) and the PDO and AMO seem to be going toward the negative and then have warming then the argument of AGW is confirmed. But what if the skeptic side is right? I worry about political advocacy in science when its has a darker purpose. Its like the DDT scare and now millions have died from a mis-guided movement. How many theories that hey had it right and then got blown out of the water. Be careful standing under trees cause an apple may hit you.
[Response: Believe it or not, I understand your viewpoint.
That said, I think you've got some of your facts wrong. I lay the blame at the feet of the denialists who have mislead you. You speak of the "current trend of no or little significant warming since 1998," but using GISS data the trend since 1998 is 0.017 +/- 0.016 (2-sigma) deg.C/yr, which is *not* little but *is* significant (even compensating for red noise).
You then go on to ask what may happen "If the cooling continues..." But there isn't any cooling since 1998, not according to GISS, HadCRU, or NCDC -- so the question assumes as fact that which is false. I'm guessing you got that, too, from denialists -- who love to use misleading phrases which give the impression that something (in this case, cooling) is established fact, when it's actually established falsehood.
I too "worry about political advocacy in science when its has a darker purpose." But I wonder, what darker purpose are you referring to? You give as an example the DDT scare, which may or may not have been misguided -- I really don't know. But even if so, I don't see what the "darker purpose" is; it seems the *purpose* of DDT regulation was to protect living things, and that doesn't fit my definition of "darker." What's the "darker purpose" of advocating reduction of greenhouse-gas emissions?
You wonder "How many theories that hey had it right and then got blown out of the water." Plenty. But AGW isn't some newfangled theory that hasn't stood the tests, both of time and of intense scrutiny. The basic physics was worked out over a century ago, the details over the next half century or more, and over the last several decades the impact of human activity on climate has been the focus of some of the most intense scrutiny in the history of science. As the scrutiny has become more intense, confidence in the basic hypothesis (the climate is warming, and human activity is the cause) has gotten ever stronger. In my opinion, the confidence which should rightly be assigned to this conclusion is far greater than is required to justify not only action, but extreme action.
What if the skeptic side is right? Then we'll have devoted considerable resources to a cleaner environment and a sane, sustainable energy policy.
What if the skeptic side is wrong?]
dhogaza // January 25, 2008 at 6:36 pm
You could pick a starting point of January 2009 and, at the moment, conclude there is no climate at all.
I can’t believe that after all the discussion here, and all the information available on the web and elsewhere, that you simply don’t get the fact that even IF there was no warming trend since 1998 (which is false), it is FUNDAMENTALLY DISHONEST and INHERENTLY UNSCIENTIFIC to cherry-pick a starting year that’s known to have been an exceptionally strong El Niño year.
I’ve got an idea. Let’s take the coldest week in December, 2007 as our starting point. I’ll bet you any sum you please that we’ll see significant northern hemispheric warming over the several month period ending August 1, 2008.
Since you think that cherry-picking’s totally cool, honest, and ethical ‘n all that.
Ah, you’ve covered two of the bogus science denialist bases. You really gotta stop reading Junkscience and crap sites like that, dude. It’s bad for your mental health.
You could start by telling us what “darker purpose” was involved in showing that indiscriminate use of DDT
1. harms a variety of bird species that are high on the food chain
2. causes DDT resistance to rapidly arise (in a matter of years, in a variety of locales on more than one continent) in the very species of mosquito being targeted
Well, our host probably doesn’t want the diversion, but it’s interesting to know that you’ve swallowed the ouput not one, but two, right-wing, industry-driven denialist campaigns against legitimate science.
Zeke // January 25, 2008 at 7:08 pm
Tamino,
I wrote a post on this subject a week or so ago over at the Media Forum. However, Nov/Dec HadCRU data wasn’t available yet, so I had to plug in GISS numbers for those months (yielding a mean 2007 temperature that put it at the 6th warmest rather than the 7th warmest year in HadCRU, but not a huge distortion).
http://www.yaleclimatemediaforum.org/dept/0108_globaltemp.htm
I also had a chance to compare the difference between standardized GISS and HadCRU datasets, though it didn’t make it into the post. It seems that there is no consistent bias of GISS relative to HadCRU in the last 30 years, and the lower HadCRU temps for the past few years are well within past variation between the sets. You can see the graph here:
http://i81.photobucket.com/albums/j237/hausfath/GISSHadCRUdifferences.jpg
ChrisC // January 25, 2008 at 10:11 pm
A simple request.
BEFORE you make a post or a comment along the lines of “global warming stopped in 19998″, please, please pretty please do some reading on the subject. Start with this very blog (for example http://tamino.wordpress.com/2007/08/31/garbage-is-forever/). Have a look at some of the data sets available. Just do some reading!
The reason I ask is that the constant repetition of this claim is really getting very boring. For one, it is easily proved false. Secondly, it has been proved false in so many different venues, so many friggin times that to have to constantly thrash it out on every friggin thread iss tiresome nad distracting.
This claim is one made by three kinds of people. the innocently ignorant, the innumerate and the dishonest. The good news is that the first two can be overcome by reading and learning. The third cannot.
So from now on, if I read that global warming stopped in 1998 (or 2001, or 2005…) again (unless it is accompanied by a VERY convincing and novel argument), I’m simply going to assume that you are being dishonest, as you’ve had plenty of time to do the reading.
Once again Tamino, keep up the good work.
dhogaza // January 25, 2008 at 10:32 pm
Actually I missed something earlier…
So, clearly Jim Arndt thinks there’s a “darker purpose” to climate science than just figuring out … climate science.
Lay it on us, Jim. What is this “darker purpose” we’re guilty of?
I’ve always wanted to be guilty of a “darker purpose”, and it’s simply HORRIBLE to be guilty of such without knowing what that “darker purpose” is!
What master am I serving? Satan? Stalin?
Hank Roberts // January 25, 2008 at 11:45 pm
Soros.
null{} // January 26, 2008 at 1:43 am
tamino, why do you not display the error ranges for any of the plots?
thanks
[Response: Error ranges for individual years, and due to a number of different sources, are readily available from HadCRU but not, as far as I'm aware, from GISS or NCDC. However, the ranges for GISS/NCDC are bound to be very nearly the same as those for HadCRU. But even if they were readily available for all three series, I'd still have omitted them.
The reason is that I made a choice: to keep the graphs simple. The first graph, for instance, has three lines color-coded to indicate source. Now add 6 more -- an upper and lower error limit for each series -- and with 9 different colored lines the graph gets crowded beyond the point of clarity. Or go to the extreme and add 18 more lines, an upper and lower limit for the error in each series due to each of three different sources; then we have 24 lines all very close to each other and it's really hard to tell what's what. Adding shade rather than lines would be a little clearer perhaps but still overcrowded in my opinion. And for this blog I make graphs with ExCel, which is a nice graphing tool but doesn't have nearly the capability of a real scientific graphing package, so it would be a lot of work to make a graph which I think is actually less informative.
I've always favored clarity and visual impact in graphs over trying to include every bit of information possible. In fact one of my least favorite graphs in the peer-reviewed climate literature is the *original* hockey stick; enough lines crowded together (even though there are only four) to reduce visual impact significantly.]
steven mosher // January 26, 2008 at 2:14 am
1976 1909 1945 are better cuttoffs.
Or you could do a regime change analysis to figure the periods
cce // January 26, 2008 at 3:36 am
Recently seen on CA
http://wattsupwiththat.wordpress.com/2008/01/25/double-whammy-friday-roy-spencer-on-how-oceans-are-driving-co2/
Now, I try to understand all sides of an issue, and give people the benefit of the doubt, but this is . . . stupid. I can’t think of any other word to describe it. Layman-style skeptics make this argument, but here’s Spencer doing it.
Mauna Loa links up with the ice records reasonably well. Why would anyone with any kind of creditials suggest that changes to the El Nino/La Nina cycle be responsible for “most” of the CO2 increase, when we’ve never seen anything like this over the entire 850,000 year ice core record? Even if we were responsible for only half of the CO2, that would still put us beyond the natural range. Whatever, dude.
Heretic // January 26, 2008 at 4:03 am
Jim Arndt, you’ve been exposed to too much of the rethoric of so called free-market advocates. A closer look at the history of DDT will show you what really happened.
The industry producing it pushed the use of completely outrageous quantities, far beyond what was necessary to achieve the goal of mosquito control. This was done to sell more of the stuff and make more money out of it than what the real need would have allowed, a perfectly normal business practice for other products. The problem, of course, is that a toxic substance become a lot more toxic when a lot more of it is used (profound thought that may not be part of business classes curricula).
The result was widespread toxic effects. The result of that result was widespread public perception that the stuff was evil. That perception was both right and wrong. It is evil when used at the ridiculous concentrations that its producers pushed. It is not if used properly at the exact doses and only where it is needed.
The backlash against DDT was the fault of the irresponsible and greedy individuals who tried to shove it down the throat of every piece of land where a mosquito had been seen, not to mention saturating beaches with it. The stupidity of DDT impementation leads to the question: what the hell were they thinking? Well, that’s easy: $, and only $.
The “darker purpose” remark seems to suggest that the people who advocated against DDT use because they observed and quantified its very real toxic effects had an anterior motive of depriving others of its use in the future, so as to cause their death. This qualifies as a conspiracy theory and is so grotesque that it does not deserve further discussion.
To all here, I recently ran into this while reading through some Deltoid posts and it is so refreshing that I am compelled to pass it along:
http://www.uwgb.edu/dutchs/PSEUDOSC/SelfApptdExp.htm
Dutch also has a nice take on pseudoscience, including considerations on conspiracy theories, which you should read, Jim:
http://www.uwgb.edu/dutchs/pscindx.htm
Tilo Reber // January 26, 2008 at 6:42 am
I can’t believe the three instrument records that you choose for your analysis. Your choices would seem to indicate that you are more interested in selling AGW than in getting at the truth.
First, NCDC is uncorrected for any heat island effect. And if you look at the records of where some of these measurement stations are, there is a lot of that.
Second, you use the GISS temperature record. That is James Hansen’s temp record. Here we have a man who is trying to prove his own model’s predictions using his own temperature record. This after years of trumpeting global disaster (25 meters sea rise). GISS is hardly a source that can be considered to come from a disinterested party.
And of course you have ignored the two satellite records, RSS and UAH. Since they show much lower temperature anomolies, we can, of course, understand why.
[Response: I was wondering when a denialist (that's you) would show up to say that the surface temperature record is useless. You've lived up to the reprehensible reputation of denialists by attempting to discredit the GISS temperature record with nothing more than a personal smear of Jim Hansen.
As for RSS and UAH, satellites don't measure surface temperature. They measure temperature in very broad layers of the atmosphere. That information has been used to estimate lower-troposphere temperature, but that's still not the same as surface temperature. You also reveal real ignorance by suggesting that their "lower temperature anomalies" are somehow indicative of less global warming, when the real reason is that they too use a different "reference period" for computing anomalies. And then of course there are the satellite-derived series from U.Washington (Fu et al.) and U.Md. (Vinnikov & Grody) which show far *more* warming than the surface record; you don't mention those (we do, of course, understand why). Those interested in a glimpse of comparison of GISS to RSS(TLT) when plotted with roughly the same zero point can look here.]
dhogaza // January 26, 2008 at 9:08 am
CCE, that’s not CA, but rather Watts blog.
But, still, it’s fascinating. Roy Spencer is putting forth the “warming increases CO2 so it is current warming that is causing CO2 to increase, not CO2 causing warming”.
I just skimmed it but the first thing I noticed is that he doesn’t seem to mention the CO2 isotope ratio fingerprint fossil fuels is leaving on the atmosphere.
He raises up the old “humans are only responsible for 3% of CO2 production therefore can’t really effect concentrations” canard.
He, like the solar influence crowd, invokes an unknown mechanism (some unknown long-term change in El Niño/La Niña) to account for data, discounting the well-understood physics of OC2 IR absorbtion, etc.
He at least starts with an honest statement:
Except that those experts will likely substitute “know” or “think”.
Stepping back to take a wider view … isn’t it pathetic that a working scientist like Spencer will allow his own political and religious beliefs drive him to the extent that he’ll spend time advancing total bogosity of this sort?
It’s sad. He mentions he’s only spent about a year studying the carbon cycle and here he is, turning an entire field of science on its head while conveniently trumpeting a hypothesis that shouts “see, we can burn as much fossil fuel as we want, because doing so doesn’t really significantly increase CO2 in the atmosphere after all!”
Sad, pathetic, not surprising.
dhogaza // January 26, 2008 at 9:17 am
That thread over on watts blog contained a pointer to a recent (Jan 24)Spencer editorial in which he says
It’s amazing that a scientist working in climate science would blatantly lie about the extent of the climate science consensus this way.
The rest of the editorial is just pseudo-scientific crackpottery, which given the fact that Spencer’s a Creationist doesn’t surprise me.
In the past, while Spencer’s personal biases were never hidden, he at least was doing real science (MSU satellite data analysis). Now that even the UAH measurements show that the earth is warming, he seems to have dropped any pretense of doing science at all, and has switched to full-blown pseudo-scientific denialism.
Barton Paul Levenson // January 26, 2008 at 1:15 pm
Is it possible Spencer hasn’t read the AR4? I mean, they have this nice diagram of the carbon cycle, with the oceans emitting 90 gigatons of carbon a year and taking in 92. Net sink, not net source. And he could probably get it from a lot more places than that, if he just cracked a book or did an internet search. It’s hard to believe he doesn’t know this.
Deech56 // January 26, 2008 at 2:00 pm
Tilo Reber, if you are interested in looking at an overlay containing the satellite data you have only to hop on over to Eli’s place: http://rabett.blogspot.com/2008/01/food-for-tamino-while-looking-at.html
cce // January 26, 2008 at 7:21 pm
dhogaza,
I found it on CA.
Re: satellites measurements
The RSS analysis of the lower troposphere shows more warming than all of these instrumental analyses. So, if tamino wanted to emphasize warming, he would have used RSS.
And thanks to the good work of Anthony Watts, we know that the GISS methods do the job of thousands of photographs of surface stations.
cce // January 26, 2008 at 7:51 pm
Also, Spencer’s “analysis” ignores the decrease in Oxygen.
dhogaza // January 26, 2008 at 9:23 pm
It is possible. But not likely. In the editorial I referenced above, Spencer pretty much puts it on record that he won’t accept any result that results in “a loss of his freedom”. I’m distilling a lengthy editorial into a soundbite, but, honestly, I don’t think I’m mispresenting his editorial. Politics/religion first, science spun to support those, rather than to seek truth.
I have to wonder if Spencer’s been such a fruitcake all along. The original MSU stuff was “Christy and Spencer”, and Christy got most of the ink that I saw, at least (which doesn’t mean much, but does anyone recall reading much of what Spencer had to say 10 years ago?).
Now Spencer’s gone public, with one result being he’s declared himself to be an intelligent design creationist who disbelieves modern biology, something that surely doesn’t help his credibility elsewhere.
And in the climate science arena, he’s gone totally nutty.
Now it looks like he wants to be the Michael Behe of Climate Science.
Hmmm, I wonder what the result would be if people started referring to him that way.
“Roy Spencer, the Michael Behe of Climate Science, recently …”
Hmmm …
Tilo Reber // January 26, 2008 at 10:18 pm
Sorry, don’t buy your alarmist explanation. First of all, when you plot your graph against HadCrut you show GISS as being warmer in 2005 than in 1998. When you plot GISS against RSS you show GISS being warmer in 1998 than 2005. Can’t make up your mind how you want it to look?
[Response: You're just plain wrong. Perhaps you THINK the graph shows GISS as warmer in 1998 than 2005 because you've failed to notice that for 2005, the GISS and RSS values are so close that the black RSS dot covers the red GISS dot and makes it nearly impossible to see. Your snotty tone when making an unfounded and blatantly false accusation is offensive -- and not just to me.]
Then there is the question of why you stopped in 2005 for the RSS graph.
[Response: Again, you're just plain wrong. Both data series go through 2006, which at the time this graph was posted (mid-2007) was the last complete year available for either data set. But for the years 2005 and 2006, the GISS and RSS values are so close that the black RSS dot obscures the red GISS dot. For a graph of these data which are *not* set to the same zero point but offset (because the two series use different reference periods to define "anomaly"), look here.]
Also, looking at the data from Junk science, there is a growing divergence between GISS and both RSS and UAH. They start fairly close together and diverge to about .25C.
[Response: Sorry, don't buy your denialist diversions.]
Tilo Reber // January 26, 2008 at 10:23 pm
Concerning Hansen, is or is he not validating his own predictions with his own temp record. And as far as his predictions go, you are desperately trying to stay alive in those error bands, aren’t you?
Tilo Reber // January 26, 2008 at 10:33 pm
Thanks Deech56. You can actually see the divergence that has grown between GISS and the satellites in that overlay.
Deech56 // January 26, 2008 at 10:56 pm
RE: Tilo Reber // January 26, 2008 at 10:18 pm: I think you are reading the GISS/RSS chart incorrectly. Start from 1998 and count out the years. And as far as Dr. Hansen’s validation, what’s wrong with using the GISS temps? The other records generally follow the same pattern as the GISS records. Tamino just showed this, and Eli did the comparison for the satellite records as well.
Deech56 // January 26, 2008 at 11:08 pm
Tilo Reber, regression analyses are better than eyeballing. Please look back at the comparison that Tamino posted for your benefit.
henry // January 26, 2008 at 11:11 pm
Tamino:
Sorry to see your work on the charts get caught up in the AGW vs denialist, CA is bad, jump on Watts bandwagon.
To get back on topic:
If the only difference between GISS and HadCRU is the way the Arctic is handled, wouldn’t you expect to see a reletively small, constant variation between them? It appears they trade places on some years.
Also, it appears at first, that a “range” of trends is developing (if that makes sense). On the last charts shown, it looks like HadCRU and GISS are “bracketing” the center player, NCDC.
Further data handling could compare to see if that “bracketing” existed over the entire range.
conard // January 26, 2008 at 11:18 pm
dhogaza,
“The rest of the editorial is just pseudo-scientific crackpottery, which given the fact that Spencer’s a Creationist doesn’t surprise me.”
The good news is that I am now bored by your absurd statements.
Your understanding of scientific history seems thin, maybe deficient. Next time you are near the library take some time and read through the works of a prominent scientist or thinker from history. I am quite sure you will be amazed at the number of theories, ideas, and behaviors that are wrong, naive, and boorish. Yet today we admire them, and rightly so, for the things they did get right.
If you only accept truth from sources without blemish you have no choice but to turn to religion, putting you in with company you disparage. I fear that your dismissal of people for having bad or wrong ideas will leave you intellectually sterile, immune to change, angry, and lonely.
Hank Roberts // January 26, 2008 at 11:57 pm
http://scienceblogs.com/stoat/2008/01/spencer_is_totally_off_his_roc.php
Tilo Reber // January 27, 2008 at 12:35 am
“I think you are reading the GISS/RSS chart incorrectly.”
Could be. Looks like I lost the red dot behind the black one.
“And as far as Dr. Hansen’s validation, what’s wrong with using the GISS temps? ”
It’s a conflict of interest. And as I said before, you cannot assume that a guy who is out talking to politicians and trumpeting 25 meters of sea level rise is without an agenda.
“Please look back at the comparison that Tamino posted for your benefit.”
It’s Tamino’s chart that I’m worried about. Lay a ruler across the starting end of Rabbets graph and then lay it across the end. There is a definite divergence.
JCH // January 27, 2008 at 1:05 am
25 meters of SLR by when, and where does he say that?
Deech56 // January 27, 2008 at 2:07 am
RE: Tilo Reber // January 27, 2008 at 12:35 am
“It’s Tamino’s chart that I’m worried about. Lay a ruler across the starting end of Rabbets graph and then lay it across the end. There is a definite divergence.”
You are falling into the trap of defining a linear regression by the first and last points, and that ain’t the way its done. To make a regression, one must consider all the data in between as well. Also take a look at the residuals to look for increased deviation from the calculated values. Tamino’s chart shows the calculated slopes of the two data series, and they are remarkably close.
Thus is all very elementary statistics, and it would behoove you to go back to Tamino’s analyses (like any of those dealing with the temperature records) and explanations; he really does make these points clear.
Deech56 // January 27, 2008 at 2:11 am
More reply to Tilo: “It’s a conflict of interest. And as I said before, you cannot assume that a guy who is out talking to politicians and trumpeting 25 meters of sea level rise is without an agenda.”
I swear that some day the word “agenda” will make my head explode. Look at the title of this post and the various charts posted and tell me: Would using HadCRU or NCDC make a difference in his conclusions?
Tilo Reber // January 27, 2008 at 3:06 am
“You are falling into the trap of defining a linear regression by the first and last points, and that ain’t the way its done.”
I understand. And the trend lines actually merge slightly instead of diverging with the linear regression. But it looks like almost all of that comes from 1998. The rest of the trend looks more like it’s diverging.
Tilo Reber // January 27, 2008 at 3:17 am
“I swear that some day the word “agenda” will make my head explode. Look at the title of this post and the various charts posted and tell me: Would using HadCRU or NCDC make a difference in his conclusions?”
HadCRU would make him a little more wrong about his projections than he already is - that is all. But that’s not the point. He should not be the validator of his own projections. I’m sorry about your head, but, again, do you really want to say that a man who talks to our legislators about 25meter sea level rise and death trains doesn’t have an agenda? When you have that much bluster and ego invested in your conclusion, impartiality in your work cannot just be assumed.
EliRabett // January 27, 2008 at 4:36 am
Tilo Reber says:
“Also, looking at the data from Junk science, there is a growing divergence between GISS and both RSS and UAH. They start fairly close together and diverge to about .25C.”
Which appears to be very wrong.
cce // January 27, 2008 at 5:07 am
I think Hansen’s agenda is to stop a huge problem.
BTW,
GISS vs Hadley/CRU vs UAH vs RSS
http://cce.890m.com/temp-compare.jpg
(each monthly anomaly offset to create a Y intercept of zero.)
dhogaza // January 27, 2008 at 8:31 am
Dodo // January 27, 2008 at 9:10 am
No matter how much you keep repeating that global warming has continued (unabated) during the 2000’s, I beg to differ.
Let’s get back to Tamino’s challenge: allow empirical data to end this controversy (yes, there is one, if not several). If, by 2015, the global mean temperature has returned to the rising trend between 1975 and 1998, we should all accept that the increase in atmospheric CO2 is the main driver of global mean temperature.
If, on the other hand, the temperature trend remains flat or even changes sign, we all accept that other forcings are more important.
It is just seven years from now. Meanwhile, we can all do our best to save the planet in the way we think is most appropriate. As an environmentally conscious citizen of the EU, I challenge Tamino and his government to respond to our efforts without delay.
null{} // January 27, 2008 at 11:56 am
thanks for the info at null{} // January 26, 2008 at 1:43 am, tamino.
However the plots are incomplete, and some would say that the most important information has been omitted, without the range information.
Additionally, without the range information one cannot actually say that the data series are different, or the same.
Deech56 // January 27, 2008 at 1:06 pm
RE: Tilo Reber // January 27, 2008 at 3:17 am
“HadCRU would make him a little more wrong about his projections than he already is - that is all. But that’s not the point. He should not be the validator of his own projections.”
Tilo, I am assuming that you are writing about Hanson, et al. 2006 (PNAS 103: 14288), in which he showed good agreement between the model predictions and the actual temperature record. I really do not see your point. Scientists make predictions and gather data to test their predictions all the time. They do not make predictions and wait for others to test their predictions.
The predictions have been out in the literature for 20 years, so anyone could have done what he did. Also, by publishing his results he presented his case to the scientific community. Others can build on his results, run their own favorite temperature records and publish a comparison. To state that a researcher cannot comment on his own predictions is absurd.
george // January 27, 2008 at 1:48 pm
“If all three series show continued warming with a statistically significant departure from being flat, then we can reject the hypothesis that global warming has stopped. If the departure is downward, we can conclude that warming has reversed; if it’s upward, then we can conclude that it never really stopped… If it turns out that global warming is not continuing (which I seriously doubt), then I’ll readily admit that I was wrong. ”
If one means by “warming” the increase in temp in a literal sense, I’d agree, but if one means “influence of greenhouse gases on the climate”, wouldn’t one have to be careful?
If there is decade-long cooling, isn’t it important to first understand it’s source(s), just as climate scientists are now trying to thoroughly understand all the possible source(s) of recent warming?
Suppose, for example, that Chinese sulfur dioxide emissions (or some other aerosols) caused a decade long cooling — not unlike what has been proposed as a reason for the downturn at mid-20th century.
If the statistical analysis shows a downturn in the temp graph over the period that you refer to, in a literal sense, we can say with fairly high confidence from the data that “global increase in temp has ceased”, but if CO2 and other GHG emissions had continued over the same period, the cooling temp might not really indicate that “global warming has stopped”, at least not in the broader sense of “influence of human emitted GHG’s on climate”, only that it has been temporarily counteracted or masked.
If the counteracting effect does not last, the warming WILL return eventually if GHG concentrations keep going up — and the temp may even return to the value it would otherwise have had without the temporary masking effects of the aerosols.
To me, this seems to be the biggest problem about the “global warming stopped in {insert year}” claims. Without considering the why, such claims mean very little in the event of continued increase GHG emissions.
Of course, if one does not accept that human GHG emissions cause warming to begin with — or that the warming they do cause is insignificant compared to “natural” effects — , then it is a different story — and I am sure the climate change skeptics would simply call the above argument a “lame attempt to evade the obvious”.
Eli Rabett // January 27, 2008 at 3:36 pm
Ooooo ……we have a concern Dodo
Climate change is a problem with HUGE procrastination penalties which is why waiting is not a smart thing.
JCH // January 27, 2008 at 5:10 pm
If it cools slightly in the next seven years, we just keep pumping CO2 into the atmosphere until all the fossil fuel is burned? As we say in the barn, udder nonsense.
The only thing that should cart AGW to the scrap yard is new science, not weather variability.
null{} // January 27, 2008 at 5:15 pm
cce // January 27, 2008 at 5:07 am
Same question as I’ve asked tamino. Why are there no range bars on your plot?
Some would argue that the most unbiased way to present raw data is solely as points without any scale indications on the axes.
George // January 27, 2008 at 6:11 pm
Tilo Reber said:
But isn’t much original research just that? — ie, a validation by an individual scientist of that scientist’s own predictions/hypotheses?
The logic behind “He should not be the validator of his own projections” assumes that all scientists are so self-invested that they simply can not be objective when it comes to testing their own hypotheses and theories..
By that logic, a scientist could never do his/her own experiments to test his/her own ideas!
Hank Roberts // January 27, 2008 at 6:16 pm
On error bars, here’s why they muddy up a page visually, compare the examples in this how-to page:
http://peltiertech.com/Excel/ChartsHowTo/ErrorBars.html
Tamino, anyone reading here who wants to see error bars is most likely to have access to the Excel software on their own; perhaps you can just point to the raw numbers and they can do it themselves, because with the freedom to zoom in on their own PC they can deal with the visual mess.
Dodo // January 27, 2008 at 6:26 pm
Eliiiiii… Listen up. As a European, I am not waiting, but taking part in a massive attempt to reduce carbon emissions. What are you Americans doing, besides preaching?
Hank Roberts // January 27, 2008 at 6:48 pm
“Nous sommes desolÈs que notre prÈsident soit un idiot. Nous n’avons pas votÈ pour lui.”
http://www.ufppc.org/content/view/487/37/
JCH // January 27, 2008 at 7:20 pm
11 months to go!
How does Europe fare when misguided individuals hold power?
cce // January 27, 2008 at 8:31 pm
The range is shown here:
http://cce.890m.com/range.jpg
The difference between the max and min anomalies has been decreasing for the last 10 years, but that may be an artifact of more “boring” global temperature anomalies. i.e. no strong el ninos or volcanic eruptions in the last few years that tend to bring out the differences between each analysis.
dhogaza // January 27, 2008 at 8:32 pm
On the one side, we have statistically valid trend analysis.
On Dodo’s side, we have … “I beg to differ”.
Fair summary, no?
jl // January 27, 2008 at 10:10 pm
since there is some discussion on GISS vs HadCRUT
I took the time to chart the differences at
http://www.geocities.com/jacob_inmt47/giss/
both should be January of 2007 with a base period of 1961-1990
both came from netcdf files and where rendered with Panoply
[Response: Very nice charts, well worth study.]
Dodo // January 27, 2008 at 10:33 pm
Hank, the first time you didn’t vote for him, but in 2004 you did. I wonder what part of the democrats’ message was difficult to understand? Maybe the one about climate change - people from coal producing states like Illinois may get nervous.
JCH, your question about European politicians in impossible to answer, because nobody really holds power. Seldom in history have so many spoken so much about a subject they know very little about, in order to make people pay so much for something they do not understand.
I mean, if the EU’s “202020″ climate policy fails, who is going to get fired or voted out of office? Nobody - instead, member states will start blaming each other.
Tilo Reber // January 28, 2008 at 12:11 am
Deech56:
“Tilo Reber, regression analyses are better than eyeballing. Please look back at the comparison that Tamino posted for your benefit.”
Okay Deech, I want to play with this myself. I was able to find the GISS data and import it to Excel. Does anyone have links to the RSS, UAH, and HadCRU data?
[Response: The first paragraph of this post includes links to global land+sea temperature data (and other files as well) from GISS, NCDC, and HadCRU. Links to satellite data sets are here. Be careful that you retrieve the correct files.]
“Scientists make predictions and gather data to test their predictions all the time.”
And that is a good thing. If Hansen were making predictions and testing it against other people’s data sets, that would be fine. But making predictions and then coming out with statements of the type, “2007 was the second hottest year on record”. Which is only according to his own data set, and which serves to reinforce the dire predictions that he had made in the past - that is a conflict of interest.
[Response: Further insults against James Hansen, with ZERO evidence, will be summarily deleted. Keep a civil tongue or take it elsewhere.]
Hank Roberts // January 28, 2008 at 12:19 am
> 2004
Wish us luck for 2008.
http://dekerivers.wordpress.com/2008/01/26/president-bushs-favorite-painting-not-what-it-seemsor-is-it/
and
http://www.ias.edu/SpFeatures/kurt_godel/godel-2.html
The examiner …:
Mr. Gödel, where do you come from?
Gödel: Where I come from? Austria.
The examiner: What kind of government did you have in Austria?
Gödel: It was a republic, but the constitution was such that it finally was changed into a dictatorship.
The examiner: Oh! This is very bad. This could not happen in this country.
Gödel: Oh, yes, I can prove it.
conard // January 28, 2008 at 12:36 am
jl,
thanks for the post.
chriscolose // January 28, 2008 at 1:29 am
Regarding this “2015 challenge,” I think that if the warming “trend” magically stops out to 2015, or declines, then we need to accept that our understanding is off. That is not to say that we could have a party with CO2 emissions, because I’d be more worried “something weird” is happening, then the idea that nature if completely offsetting us, or that CO2 radiative physics is wrong.
If 2015 by itself is anomalously cold, it is probably due to some internal variability– did global warming “stop” after a volcanic eruption? This by itself is not statistically significant. It would go the same way if the trend declined since 1998, but 2005 stuck out as anomalously warm. That is not what is happening. 1998 was one of the biggest EL Nino’s, so its net effect (~0.1-0.2 K) is easily identifiable visually on a plot of global T vs. time, but it is superimposed on the much larger (~0.8 K) long-term trend.
If you are too concerned with which years are the nth warmest on record, we’re getting a bit away from the climate discussion at hand, and only encouraging skeptics to cherry pick individual years, since all of a sudden it is becoming a “that was warmer than that year” game. Silliness.
P.S. Tamino, as a nitpick you might want to fix this piece of verbiage
“If fact they’re all three on different scales.”
EliRabett // January 28, 2008 at 3:09 am
Dodo. . . well at least the EU is moving in the right direction as for me, there is Eli’s simple plan for saving the world. The bunny notes that Sarkozy is adopting our no free riders plan to the usual stuck pigs squeals.
Lee // January 28, 2008 at 3:19 am
Tilo Reber:
So let me get this straight.
A scientist has a model, which he wants to validate against the real world.
So he collects real-world data that has been generated by others, and is readily available for those willing to do the work, and analyzes that data.
H e updates that analysis on an ongoing basis, so that one can see the evolution of the system in near-real time - and so that both expected and unexpected results are available to everyone nearly immediately.
He publishes his analytical methods, and indeed has released the computer code that does the analysis, so that anyone can see how he did the analysis.
He publishes his results, puts them out there for comparison against similar analyses by others who use that same real-world data, and for comparison to satellite analyses done with entirely independent data and analytic methods, so that anyone can see the similarities (huge) and differences (small) in both the surface analyses, and the satellite analyses.
Hell, that’s the entire point of this thread.
And somehow this Tilo guy considers that to be disqualifying - the fact that Hansen is doing open and transparent science somehow, in Tilo’s mind, disqualifies him as a scientist, and therefore Hansen can not be trusted. How does this leav e Tilo any room to even believe that science has any utility? The POINT of science is that when one develops a model or prediction, that one gathers data and does analyses to test it. He wants all scientists to be running around asking others to test their ideas for them?
Its hard to imagine that someone actually believes that, but I’ll give Tilo the benefit of doubt, and attribute the argument to a dumbfounding lack of logical clarity on Tilo’s part, rather than to dishonesty.
Heretic // January 28, 2008 at 3:57 am
And Eli, France has experienced a decrease in CO2 emissions last year, a result in which the (almost) all nuclear electricity production plays a major part.
Barton Paul Levenson // January 28, 2008 at 1:24 pm
Tilo Reber writes:
[[Concerning Hansen, is or is he not validating his own predictions with his own temp record.]]
Where did you get the idea that Hansen owns the GISS temperature record? Do you think he bribed the thousands of people who maintained those stations to submit false readings? What in the world are you thinking here?
Barton Paul Levenson // January 28, 2008 at 1:28 pm
Tilo Reber writes:
[[“And as far as Dr. Hansen’s validation, what’s wrong with using the GISS temps? ”
It’s a conflict of interest.]]
Do you have any idea what the phrase “conflict of interest” actually means? You’re saying that a scientific organization which makes empirical observations has a conflict of interest when it tries to use or interpret their own findings! If that were taken seriously it would pretty much shut down all science, anywhere, ever.
Hank Roberts // January 28, 2008 at 4:38 pm
Tilo, is this you? Or someone else using the same name?
http://dotearth.blogs.nytimes.com/2008/01/24/earth-scientists-express-rising-concern-over-warming/#comment-9238
Dodo // January 28, 2008 at 4:45 pm
Chris. Thanks for elaborating on the problems with Tamino’s 2015 challenge. Of course, we should have a commitment that eliminates both cherry- and nitpicking.
Tamino, how would you define the trend lines that would decide in favor or either “camp”?Surely we must reduce the weight of outlier years like 1998 somehow.
Barton Paul Levenson // January 28, 2008 at 5:49 pm
Dodo writes:
[[Tamino, how would you define the trend lines that would decide in favor or either “camp”?Surely we must reduce the weight of outlier years like 1998 somehow.]]
Two words: Linear regression.
Tilo Reber // January 28, 2008 at 6:07 pm
So I’m still gathering data sets in order to do my own comparisons. I have GISS and HadCrut3v.
Looking at the UAH data I assume that I should use the column that is marked “Globe” and that begins on 12, 78 with -0.13 and ends 12, 07 with -0.06.
For RSS data I’m assuming that I need the one marked -82.5/82.5 and beginning on 1, 79 with -0.136 and ending on 12, 07 with -0.023.
The GISS and HadCru data that I have is yearly. The UAH and RSS data is monthly. To get comperable data sets can I simply take the twelve months average for every year, or do I need something else?
[Response: Be sure to use the satellite files for "TLT" (lower-troposphere temperature); it's not surface temperature but it's the closest to it from satellite measurements. It looks like you have identified the correct columns with which to work.
You can also get monthly data from GISS and HadCRU (at the links in the 1st paragraph, and on the new "Climate Data Links" page). Last I checked, the HadCRUT3v file wasn't updated through December 2007, but the HadCRUT3 file was; you may have more current information.
You can estimate trends from annual or monthly averages; they'll be about the same but a little more precise (only a little) if based on monthly data (if you use monthly data, you'll want to convert year/month to decimal years).]
elspi // January 28, 2008 at 6:21 pm
Dodo said
“how would you define the trend lines that would decide in favor or either “camp”? Surely we must reduce the weight of outlier years like 1998 somehow.”
Calculus it is then.
Take a data set {(x_i, y_i)} (in this case x_i is the date and y_i is the temp.)
Consider any line y =ax +b
Define f(a,b) = sum of the square of distances form y_i to the point on the line over x_i (that is ax_i +b )
Thus f(a,b) = (ax_1 +b -y_1)^2 + (ax_2 +b -y_2)^2 + … +(ax_n+b -y_n)^2.
This function is nice and differentiable and so we can find its minimum by taking the derivative and setting it equal to zero.
This give you THE numbers a and b which minimized the square of the vertical distances from the line to the data. This is called the least squared fit. It is the standard way of fitting a line to data. The process of find this line is called linear regression.
It can of course be done for other curves than lines, but there should justification for using something more complicated.
As to why the derivative of a function must be zero (or undefined) if the function is to have a min (or max), well the function “looks” like the tangent line at a point (to first order anyway). So if the tangent line is sloping up, then when you move to the right the function immediately increases and similarly when you go to the left it decreases. The argument for downward sloping is analogous.
Julian Flood // January 28, 2008 at 7:37 pm
Interesting to see the three marching together, particularly through the ‘35 to ‘50 period. Am I correct in assuming that all three series have the Folland and Parker SST correction incorporated? If so, I’d really like to see the plots with raw data — I am not enamoured of that particular data manipulation. Give or take a bit of smoothing, obviously, to damp down ludicrous volatility.
Are corrections applied to all the data and, if so, do they differ between the sets?
JF
Dodo // January 28, 2008 at 9:33 pm
Ok, elspi, thanks for the math. I haven’t done calculus for 20 years, and am not going to do any now, so let’s go for simple linear regression. The last data point shall be global mean temperature in 2015. Shall we start at 2007 then, or 1998? And which dataset should we use? GISS would be nice (high 2007), but shouldn’t we take averages of several temp series?
What about the slope? What if it is just slightly positive or negative, but statistically insignificant? How would we define a draw?
The 2015 challenge is not worth much if it is not formulated exactly by statistically skilled people form both “camps”. I rest my case, and my brain.
[Response: The method outlined by elspi is the "method of least squares", so-called because it seeks the paramters which minimize the sum of squares of the deviations between the data and the model. You may be pleased to learn that "linear regression" actually *is* least-squares regression, when the "model" is a straight line.
If you're working in ExCel, then by choosing "add a trend line" and ensuring the trend line is linear (I believe that's the default), it will compute it according to least-squares regression. You can also choose the option "display equation on chart" to get the numerical values of the slope and intercept.]
Tilo Reber // January 29, 2008 at 1:11 am
“Do you think he bribed the thousands of people who maintained those stations to submit false readings? What in the world are you thinking here?”
Who decides what corrections are made to the data from each site for things like heat island effect? Who decides when sites are swapped in and out? Who decides how data is processed and weighed?
“Do you have any idea what the phrase “conflict of interest” actually means?”
Hope so! For example, if I’m a politician and I’m sitting on a committee that is debating about awarding a contract to a company in which I own shares, then I have a conflict of interest and I should recuse myself from the committee. The basic idea is that I should not be a beneficiary as a result of making decisions on an issue where I should be impartial. Do I get an A or an F?
So, let’s take Hansen’s case. He makes predictions about future climate change. He is very vocal, not only about his science, but also about what he thinks politicians should do about it. If his predictions are correct, he stands to gain a great deal of prestige. If they are wrong, his prestige suffers. I don’t know if financial incentives are involved, so I won’t go into that. In the area of climate reporting with GISS, he makes decisions every day about the issues that I mentioned above. So, are we to assume that he makes these decisions with no regard about how they effect his predictions. You can make that choice if you want to. I believe that Hansen has a conflict of interest. For example:
(Since I’m just getting the hang of this thing and since I only looked at these numbers once, I need to go back and recheck but, here’s the info).
I tried to follow up on the convergence, divergence checking of data sets that Eli started. Over the long time span GISS, HadCrut3, RSS, and UAH have no significant divergence that I can see. But in the short term (last 6 years) HadCrut3, RSS and UAH all trend down. Only GISS trends up. Of course I realize that people here will say that it’s too short a time period to be significant, but at this point we can’t know if it’s signal or noise. We do, however, know that Hansen’s predicted 3C of forcing per CO2 doubling begins to look doubtful if the temp is not rising and if we don’t have an explanation for why it’s not rising. Well, we didn’t have any volcanoes over that time period. We’ve had a solar minimum only for 2007 so far. I don’t think that there was a dominance of La Nina activity over that period of time. So if CO2 forcing is such a dominant climate factor, shouldn’t the temp just keep trucking up the scale? It would seem that Hansen would have a strong motive for showing that it does just keeps trucking up the scale.
The question about scientists checking their own theories seems like a red herring to me. I write software for launch vehicles. When I do, I try to produce data sets and simulations that will test that software as well as possible. But at some point I turn it over to a system test group that will use their own test scenarios to try to break my software. I don’t get to make the final decision on the quality and status of my software. In fact, my company often hires an outside company to do Verification and Validation, just to insure that we are not mislead by misplaced confidence in our own work or by shortcuts taken in order to meet a schedule. So, yes, Hansen should check his predictions. But he should not be the final arbiter of the quality of his predictions by using his own data work to judge them.
“Tilo, is this you?”
That be me Hank. I will sometimes use a handle when posting, but I don’t think that the AGW crowd has gotten extreme enough yet to behead me if I disagree with them. LOL.
[Response: I believe you have a false picture of what's going on at GISS (and HadCRU and NCDC etc.). For example, nobody decides that "such-and-such" UHI correction will be applied to any particular location. Instead, a method of applying UHI correction is chosen based on analysis of comparison between nearby rural and urban stations. Then for a given location, the amount of the correction is determined by the behavior of nearby rural stations, and the stations to which it is applied are selected on the basis of objective criteria. You need to invest some serious effort learning how the corrections are applied; the NASA GISS website has links to the scientific papers which document it. It appears that you have a lot to learn before you even know how it's done, let alone can offer any valid criticism of the process.
As for the recent data being too brief to offer any real evidence that global warming has stopped or even slowed down, that's been established in a number of posts. So answer (at least for yourself) this question: is the lower temperature of December 2007 compared to November 2007 valid evidence that global warming has stopped or reversed? If not, why not? When you truly understand *that* issue, you'll have begun to understand the issue here.
All your accusations are nothing but speculation, based on zero evidence, and you sling the mud despite the fact that it seems clear you don't even know how the data are processed. Any further implications of misconduct on the part of James Hansen must either be accompanied by real evidence (including links to documentation of that evidence) or will be summarily deleted; there will be no further speculative accusations or implications allowed. If you believe Hansen has defrauded us, either offer hard evidence or take it elsewhere.]
Deech56 // January 29, 2008 at 2:37 am
RE: Tilo Reber // January 29, 2008 at 1:11 am
“The question about scientists checking their own theories seems like a red herring to me. I write software for launch vehicles. When I do, I try to produce data sets and simulations that will test that software as well as possible. But at some point I turn it over to a system test group that will use their own test scenarios to try to break my software. I don’t get to make the final decision on the quality and status of my software. In fact, my company often hires an outside company to do Verification and Validation, just to insure that we are not mislead by misplaced confidence in our own work or by shortcuts taken in order to meet a schedule. So, yes, Hansen should check his predictions. But he should not be the final arbiter of the quality of his predictions by using his own data work to judge them.”
Go up to Lee’s post and read again. Dr. Hansen does science. He does science the way just about every other scientist does science (just better than most of us). The final arbiter of the quality of his predictions (assuming his manuscripts pass peer review) is the scientific community as a whole. Just like for every other scientist.
Please educate yourself on how scientists works and science operates before making specious arguments. I’ve published a number of scientific papers and had a number of scientific abstracts accepted for meeting presentation and publication. None of my data were given to outside auditors for validation and verification. Unless one is working under GLP conditions, that is just not done. Peer reviewers have the opportunity to examine my claims and the scientific community has an opportunity to replicate my studies, if they choose.
thomas lavin // January 29, 2008 at 5:36 am
i posted this at dot.erth.com but still havent had much of a response.
if you plot the giss data using 1881 to 1950 as the baseline and compare to 1950 to 2007 as the comparison set, you just dont get much global warming and do not reproduce the graph above. Even better try it by season, and most of the anomaly shows up only in the far north and only in the winter. use 1200 smoothing and the haydl reyn ocean data. see what you come up with
http://data.giss.nasa.gov/gistemp/maps/
[Response: Using a different baseline only shifts the entire plot up or down, without affecting the changes from one time to another. So it's hard to believe your result. I suspect you've made an error somewhere.]
thomas lavin // January 29, 2008 at 5:46 am
and by the way, the giss maps show mostly localized “global warming”, which of course is not reflected in temperature trends. the americas have much less or no “global” warming compared to europe and russia in all of these maps., including
http://www.geocities.com/jacob_inmt47/giss/
so is it even fair to say this is “global warming” or should we say localized warming of sufficient extent that when averaged over all latitudes and all longitudes it provides a global effect on the averaged temperature
cce // January 29, 2008 at 7:37 am
Comparing 1951-2007 to 1881-1950 gives me a 0.31 degree global temperature anomaly. Not sure what you mean by “not producing the graph above,” since it doesn’t create graphs like that (it does create a graph based on warming by latitude, however).
But, if you look at tamino’s graph and break it into two parts, which (roughly) correspond to the break at 1950. Just eyeballing it, the first half looks like it has an average of about -0.1 degrees, and the second half looks lik about +0.2 degrees, or an increase of 0.3 degrees.
The world is made up of regions. When everything is in equilibrium, you will have regional climate changes, as the energy is essentially pushed from one area to another. If you have warming averaged over the whole world, that energy has to be coming from somewhere, and the number of candidates for that is short.
dhogaza // January 29, 2008 at 8:58 am
“global” impl