Guest post by Willis Eschenbach
Inspired by this thread on the lack of data in the Arctic Ocean, I looked into how GISS creates data when there is no data.
GISS is the Goddard Institute for Space Studies, a part of NASA. The Director of GISS is Dr. James Hansen. Dr. Hansen is an impartial scientist who thinks people who don’t believe in his apocalyptic visions of the future should be put on trial for “high crimes against humanity”. GISS produces a surface temperature record called GISTEMP. Here is their record of the temperature anomaly for Dec-Jan-Feb 2010 :
Figure 1. GISS temperature anomalies DJF 2010. Grey areas are where there is no temperature data.
Now, what’s wrong with this picture?
The oddity about the picture is that we are given temperature data where none exists. We have very little temperature data for the Arctic Ocean, for example. Yet the GISS map shows radical heating in the Arctic Ocean. How do they do that?
The procedure is one that is laid out in a 1987 paper by Hansen and Lebedeff In that paper, they note that annual temperature changes are well correlated over a large distance, out to 1200 kilometres (~750 miles).
(“Correlation” is a mathematical measure of the similarity of two datasets. It’s value ranges from zero, meaning not similar at all, to plus or minus one, indicating totally similar. A negative value means they are similar, but when one goes up the other goes down.)
Based on Hansen and Lebedeff’s finding of a good correlation (+0.5 or greater) out to 1200 km from a given temperature station, GISS show us the presumed temperature trends within 1200 km of the coastline stations and 1200 km of the island stations. Areas outside of this are shown in gray. This 1200 km. radius allows them to show the “temperature trend” of the entire Arctic Ocean, as shown in Figure 1. This gets around the problem of the very poor coverage in the Arctic Ocean. Here is a small part of the problem, the coverage of the section of the Arctic Ocean north of 80° North:
Figure 2. Temperature stations around 80° north. Circles around the stations are 250 km (~ 150 miles) in diameter. Note that the circle at 80°N is about 1200 km in radius, the size out to which Hansen says we can extrapolate temperature trends.
Can we really assume that a single station could be representative of such a large area? Look at Fig.1, despite the lack of data, trends are given for all of the Arctic Ocean. Here is a bigger view, showing the entire Arctic Ocean.
Figure 3. Temperature stations around the Arctic Ocean. Circles around the stations are 250 km (~ 150 miles) in diameter. Note that the area north of 80°N (yellow circle) is about three times the land area of the state of Alaska.
What Drs. Hansen and Lebedeff didn’t notice in 1987, and no one seems to have noticed since then, is that there is a big problem with their finding about the correlation of widely separated stations. This is shown by the following graph:
Figure 4. Five pseudo temperature records. Note the differences in the shapes of the records, and the differences in the trends of the records.
Curiously, these pseudo temperature records, despite their obvious differences, are all very similar in one way — correlation. The correlation between each pseudo temperature record and every other pseudo temperature records is above 90%.
Figure 5. Correlation between the pseudo temperature datasets shown in Fig. 3
The inescapable conclusion from this is that high correlations between datasets do not mean that their trends are similar.
OK, I can hear you thinking, “Yea, right, for some imaginary short 20 year pseudo temperature datasets you can find some wild data that will have different trends. But what about real 50-year long temperature datasets like Hansen and Lebedeff used?”
Glad you asked … here are nineteen fifty-year long temperature datasets from Alaska. All of them have a correlation with Anchorage greater than 0.5 (max 0.94, min 0.51, avg 0.75). All are within about 500 miles of Anchorage. Figure 6 shows their trends:
Figure 6. Temperature trends of Alaskan stations. Photo is of Pioneer Park, Fairbanks.
As you can see, the trends range from about one degree in fifty years to nearly three degrees in fifty years. Despite this huge ~ 300% range in trends, all of them have a good correlation (greater than +0.5) with Anchorage. This clearly shows that good correlation between temperature datasets means nothing about their corresponding trends.
Finally, as far as I know, this extrapolation procedure is unique to James Hansen and GISTEMP. It is not used by the other creators of global or regional datasets, such as CRU, NCDC, or USHCN. As Kevin Trenberth stated in the CRU emails regarding the discrepancy between GISTEMP and the other datasets (emphasis mine):
My understanding is that the biggest source of this discrepancy [between global temperature datasets] is the way the Arctic is analyzed. We know that the sea ice was at record low values, 22% lower than the previous low in 2005. Some sea temperatures and air temperatures were as much as 7C above normal. But most places there is no conventional data. In NASA [GISTEMP] they extrapolate and build in the high temperatures in the Arctic. In the other records they do not. They use only the data available and the rest is missing.
No data available? No problem, just build in some high temperatures …
Conclusion?
Hansen and Lebedeff were correct that the annual temperature datasets of widely separated temperature stations tend to be well correlated. However, they were incorrect in thinking that this applies to the trends of the well correlated temperature datasets. Their trends may not be similar at all. As a result, extrapolating trends out to 1200 km from a given temperature station is an invalid procedure which does not have any mathematical foundation.
[Update 1] Fred N. pointed out below that GISS shows a polar view of the same data. Note the claimed coverage of the entirety of the Arctic Ocean. Thanks.
[Update 2] JAE pointed out below that Figure 1 did not show trends, but anomalies. boballab pointed me to the map of the actual trends. My thanks to both. Here’s the relevant map:
Why not infill using the satellite data?
Satellites only look “across” (they are in pole-to-pole orbit). Therefore there is no direct lookdown capability (for an unknown reason) and they cannot observe either north or south pole.
I also understand there are some issues with MW reflection off ice that makes life difficult.
The sort of procedure used by GIStemp, using the correlation structure in the data to fill in the gaps, is not dissimilar to the geostatistical tools used by mining companies estimate how much reserves there are from scattered data.
Well, that explains why mining companies are so abominably bad at estimating reserves. My own method turns out to be far more accurate:
1.) Take the given estimate. Multiply it by ten.
2.) Be sure that step 1 is an underestimate. Probably a gross underestimate.
Works for everything from tin to oil. (But for projected warming, divide by 10 rather than multiply and assume an overestimate.)
@ur momisugly Andrew P.
The issue is not over measurement of todays temperatures. Todays temperatures are easy – not only do we have the buoys you mention, but we also have satellite data. We know the absolute temperature in the polar region quite accurately.
The issue is over the measurement of the baseline which is subtracted from todays accurate temperatures to determine the `anomaly’ (difference from normal). The problem is that the baseline is set using historical temperature data dating back to the period before we had accurate data from satellites or buoys. So that baseline data is in fact highly innaccurate and constructed using a lot of questionable assumptions, particularly in a region like the arctic where there isn’t a lot of historical data to go by.
We simply don’t know what the `normal’ temperature (if such a thing exists) for parts of the world should be. At best we have a `guesstimate’. And many here feel that every effort has been made in the construction of such guesstimates to come up with a number as cold as possible to make it look like the world is currently unusually hot.
Anomalies are differences. The accuracy of a difference is no greater than that of its least accurate component. Subtracting inaccurate historical guesstimates from highly accurate modern satellite data makes the resulting anomalies inaccurate.
Keith G (18:32:14) :
Frankly, the conclusion seems obvious: given the exceedingly sparse temperature data sets for the Arctic, the construction of forecasts based on this data can only be regarded (at best) as being ‘unreliable’.
The more reasonable position to take would be: “on the basis of data that we currently have, we do not know what the recent temperature trends in the Arctic have been. Nor can we say much about what they will be.”
In that case where do you suppose the data for the DMI Arctic temperature linked to on this site come from? Willis shows 3 stations on the edge of the 80ºN parallel.
[snip – we won’t be turning this thread into a discussion of tobacco company issues and arguments – A]
This whole issue raises several points, not the least of which is again, the near uselessness of a global average temperature and the problems with expressing it as an anomaly number. Grossly dependent on and subject to manipulation of the base period, and not really a useful metric anyway. There are to many widely different temp distributions that can come up with the same average, even when done in smaller spacial cells.
Which brings up another point. Even without any manipulation to support a desired outcome (higher temps), either deliberate or inadvertent (selection bias) this is a prime example of what happens when mathematics is divorced from the underlying physical reality at hand. If you draw a line between two points separated by a distance, and say one end of this line has temp A, and the other temp B, what is the temp halfway between? Why, just do a simple interpolation! Mathematically that is valid, but not when the numbers are associated with underlying physical reality and not just isolated scalar metrics. If A is on the top of a high mountain, B is on the ocean, and the point in between is in the middle of a steamy jungle, obviously the situation is a lot more complex and doesn’t yield to simple analysis or interpolation.
It is a bad practice just on the basis of rigor, let alone when it can be manipulated in ways to support a predetermined outcome. Too much possibility of gross error, but I get the impression too many scientists don’t understand the difference between pure mathematics and math constrained by real world processes and such.
Given that climate science is so politicized, should the guys who politicized it (for example, Hansen and Jones) be in charge of collecting the basic data? Is there anybody less biased who could take over the job of data collection? There has to be an independent organization collecting the data. This org should be staffed by people who are not climate scientists. Maybe the solution is to take climate science out of NASA. It is a shame how Hansen has sullied NASA’s reputation.
My numerical analysis professor told me not extrapolate.
Joe Bastardi is a clever fellow. He has opened another can of worms for the alarmists without being the guy who actually presses charges. Well done Joe and well done Willis.
24 March: UK Daily Mail: Will Stewart: Russia’s top weatherman’s blow to climate change lobby as he says winter in Siberia may be COLDEST on record
‘The winter of 2009-10 was one of the most severe in European part of Russia for more than 30 years, and in Siberia it was perhaps the record breaking coldest ever,’ said Dr Alexander Frolov, head of state meteorological service Rosgidromet. ..
Climate change adherents say the planet is warming due to man-made factors but Russian expert Professor Arkady Tishkov said yesterday that Siberia and the world are in fact getting colder.
‘From a scientific point of view, talk about increasing average temperatures on earth of several degrees are absurd,’ he said.
‘Of course we can’t say that global warming is a myth and falsification. In many regions of planet the temperature is higher than expected because of human impact.
‘But the climate system of the planet is changing according to different cycles – from several years to thousand of years.
‘From the scientific point of view, in terms of large scale climate cycles, we are in a period of cooling.
‘The last three years of low temperatures in Siberia, the Arctic and number of Russia mountainous regions prove that, as does the recovery of ice in the Arctic Ocean and the absence of warming signs in Siberia.’
Mr Tishkov, deputy head of the Geography Institute at Russian Academy of Science, said: ‘What we have been watching recently is comparatively fast changes of climate to warming, but within the framework of an overall long-term period of cooling. This is a proven scientific fact.
‘The recent warming – and we are talking tenths of a degree at most – is caused by human activity, like forest elimination, the changing of landscapes.
‘The greenhouse gases so much discussed now do not in fact play big role. We have to remember that all the impact of industrial enterprises in Russia cannot be compared with one volcano eruption on our planet.’….
http://www.dailymail.co.uk/news/worldnews/article-1260132/Russian-weatherman-strikes-blow-climate-change-lobby-announcing-winter-Siberia-coldest-record.html
Willis:
You might want to make notice of this curiosity for future investigation.
Notice how your graph here:
http://wattsupwiththat.files.wordpress.com/2010/03/north_pole_giss_polar_2.jpg
is almost a dead match to the Arctic Oscillator Loading Pattern advanced eastward 90 degrees here:
http://www.cpc.ncep.noaa.gov/products/precip/CWlink/daily_ao_index/loading.html
That is a very curious coincidence, if it is a coincidence. If not, it seems to lead you to assume that one is made from the other even though I don’t see how that could be possible in nature. One is temperatures and the other is pressures if I am correct.
Drifting buoys usually freeze up in the winter, however they do send good Arctic info when not frozen. Does GISS use this stuff?
http://www.sailwx.info/shiptrack/shiplocations.phtml?lat=88.62838836211&lon=39.796694558758&radius=500
To claim that there are not any thermometers- well at least some of the time.
read comments too:
22 March: BBC Paul Hudson Blog: A breakthrough in long range forecasting?
I am indebted to Dr Jarl Ahlbeck, from Abo Akademi University, Finland, who contacted me about his fascinating new piece of research relating to this winters severe cold across much of Europe, and a possible link to the very low solar activity we have been experiencing….
In essence what this research shows is that there is a link between the level of solar activity, the stratosphere, and the weather patterns that we experience, and gives more weight to the idea that solar effects may influence our weather (and hence climate) more than is currently accepted or understood. ..
You can read Dr Ahlbeck’s research paper in PDF format by clicking here [214KB PDF]
http://www.bbc.co.uk/blogs/paulhudson/2010/03/-i-am-indebted-to.shtml
Just did a quick check on the numbers for Eureka (where i used to work many years ago). Average for 51-80 as defined by GISS normal for winter months -38C
winter 2009-10 was -36. so it was 2 degrees above normal, whop-de-du! Love the color scheme, it pushes the melt going on up there at that temp!!!
‘The greenhouse gases so much discussed now do not in fact play big role.’
As the Russian software engineer who sat next to me said one day, while I was having an intense losing battle with Unix Script & EBSIDIC readouts, “Is not end of world. End of world is big mushroom cloud”.
GISS is not end of the world.
@Willis Eschenbach (17:25:18) :
We’ve been over this several times. GISS does not do this now for the Arctic Ocean, for two reasons. The satellites don’t go that far north. The satellites don’t measure ice temperatures.
The satellites cover up to 82.5 °N.
http://www.discoveringthearctic.org.uk/images/8b_ahdr.jpg
That’s the halfway point between the very inner circle, and the 80 °N latitude circle.
The satellites Microwave Sounding Units measure the intensity of upwelling microwave radiation from atmospheric oxygen. They do not measure “temperatures” directly – the data must be mathematically inverted to deduce the temperature of the lower troposphere. This math is rather detailed and difficult – UAH (University of Alabama in Huntsville) got it wrong for decades.
I gave you links to detailed analysis of NOAA satellite data in the Arctic – I have no desire to become an expert on the intricacies of how summer melt ponds on top of sea ice affects satellite SST measurements by MSU’s, for instance, but perhaps you are as interested as you seem. Good luck.
Additionally, the GISS temperature trends are given from a base period of 1951-1980. If you have satellite temperature measurements of the arctic ocean for that period, please cite them …
They have satellite measurements from 1979. But even those have that 82.5 °N limitation that bugs purists. I myself wish they had 90 °N inclination satellites measuring the entire planet, for the last 5 decades. But 20 decades ago they were still working out the basics of geology, for comparison…
I see you are not happy with the various methods for guesstimating Arctic temperatures in areas that are not directly measured – I guess you’ll have to throw out that entire dataset you showed me from 1875 to 1979.
pat (21:05:55) :
‘However, during low solar activity the easterly QBO causes a considerably stronger negative AO than the westerly QBO is able to cause a positive AO. Furthermore, easterly QBO is more common than westerly QBO during the Nordic Hemisphere winter.’
That’s an interesting concept.
A couple of strongly negative QBO’s could knock the Northern temps down like a pile driver.
Cement a friend (18:51:18) :
However, I thought that a correlation was only significant if greater than 0.9 (or less than -0.9)
Not true. A correlation could be significant with a very small R value. You can pick any significance (confidence) level you would like to consider whether such a small R is relevant to you or not.
For example, imagine several thousand pairs of data points where the measured data has a huge spread around an upward sloping trend line, like measuring a very noisy signal over time. Suppose it has an R^2 value of of 0.04. The R then is 0.2. It could still be quite apparent that this signal is increasing over time, and with increasing N, it will become more and more significant. At some point it will be determined that it is unlikely such a trend could develop in the data by chance alone, and once this threshold is crossed (your confidence interval), you can be confident that it is significant (not by chance at your chosen confidence).
The correlation R or determination R^2 describes how well it is possible to predict the next data point based on the model (in this case with increasing time). You will not be able to predict with any accuracy where the next few data points will lie due to the fact of so much unexplained variance. Your model only explains 4% of the variation. So a low R means an inability to predict the next outcome with confidence based on a given input value. But this does not mean your model is wrong, it just means you don’t have enough variables included in your model to explain the variation you see. If it turns out that your model is correct, and that it holds true beyond the measured range (dangerous territory to predict here), you will be able to predict the general direction of a collection of many outcomes based on the independent value, but not of a single value. The noise will remain until you improve your understanding / measurements / model to reduce unexplained variation.
Where so many climate studies fall apart is that very noisy data is taken over a very short time frame, these linear trends are then projected forward to predict a future outcome (generally with the independent variable being the all-evil CO2 molecule). But the underlying data is in-fact cyclical when a longer time frame is considered, superimposed on an even warming trend since the LIA. Meaning the proposed model only worked over the time frame when the slope of natural and reasonably predictable temperature cycles happened to coincide with sharply increased CO2 emissions.
This is no mistake. The only way to build an alarming story is to purposely ignore previous cyclical events, or “off them” as the hockey stick attempted to do. Once we have chosen to ignore obvious natural cycles, this opens the door for a higher correlation with high linear slope during the time since WWII. The really nice part is, if we only use this time, the better statistics also indicate higher predictive ability, meaning our model is much, much better (and we’re smarter too since we’ve greatly reduced unexplained variation – a real win-win situation). Well, since CO2 alone cannot be responsible for that much warming over such a short period using even the most aggressive greenhouse gas calculations, the only way to explain this is to over inflate effect of CO2 by surmising positive feedback effects that amplify the small warming effect that CO2 is presumed to have. Effects which have never been demonstrated. Now you have a model you can project into the future that can really scare people. YAAAY!
All of this, of course, develops while researchers also fail to study natural negative feedbacks due to clouds and thunderstorms which strongly counteract warming, by whatever cause. Ignoring the past in this case is much like predicting it will be 3600° outside by September because the temperature increased 1°F from 10:00 to 11:00 this morning.
So we’ve developed insane models based on the coincidence of a few sine waves added together plus a new variable which has little practical effect on temperature, we focused on the highest slope area we could find to improve our predictive ability and reduce noise, amplified it, then projected it as far as the eye can see. Which is why it’s falling apart before our eyes. Such a model has no chance of success in practical terms since it fails to consider the cyclical and chaotic nature of climate. In political terms, it just needs to hang on by a thread for a little longer for the anthrophobes to succeed.
I still think this is the best demo I’ve seen yet of how dangerous the last sharp warming trend was (that ended in 1995)…
http://wattsupwiththat.com/2009/12/12/historical-video-perspective-our-current-unprecedented-global-warming-in-the-context-of-scale/#more-14034
DMI has very accurate measurements north of 80N, and has for many decades.
http://ocean.dmi.dk/arctic/meant80n.uk.php
Ah, another carefully crafted comment moderated away.
That means I can’t comment here again for 24 hours – too bad, Willis had a good topic and some interesting points.
Cheers.
REPLY: you can comment on topic, just leave tobacco out of the discussion – Anthony
Anu (20:46:33) :
[snip – we won’t be turning this thread into a discussion of tobacco company issues and arguments – A]
Thank you Anthony. 🙂
Are you sure of this? I visited the Heartland site about four months ago and got the impressions that they were only involved in the fight over cancer and secondhand smoke, not cancer and smoking itself, and that they were not a PR organization but a thinktank.
I find statistics rather fascinating and often confusing, even though I have to use them on a regular basis. Like many other implements, it is just a tool that only works if applied properly. I keep going back to the following site on Inferential Statistics when I get confused:
http://faculty.vassar.edu/lowry/webtext.html
It gives a very understandable explanation of the concepts. Pertinent to this discussion are chapter 3 & 4 on correlation, regression and significance.
Steve Koch (20:52:38) :
Maybe the solution is to take climate science out of NASA.
That would require taking Washington out of NASA.
In the 60’s there was a glorious push by Washington to put a man on the Moon. NASA was an arm of Washington to accomplish that. Now there is an inglorious push by Washington for Cap N Trade from ‘global warming’. Again NASA is an arm of Washington to accomplish that. NASA goes where Washington wants it to go.
I hate it because I we’re supposed to be seeing accomplishments as awesome as landing on the Moon coming from NASA. But if you have inglorious leaders you have the inglorious goals they set.
All we have now is James Hansen and Gavin Schmidt. 🙁
We are like Charlie Brown on Halloween always getting a rock!
pat (20:56:06) :
24 March: UK Daily Mail: Will Stewart: Russia’s top weatherman’s blow to climate change lobby as he says winter in Siberia may be COLDEST on record
Those poor prisoners in Siberia!! Yes, there are still prison camps there.
St Paul’s dictum comes to mind:
“As in a mirror , darkly” anomalies say something about temperature which temperature is a proxy for heat and heat content is what signifies cooling or heating.
Anomalies are a third level proxy, and as I have said often enough, on the earth, due to the several heat transport mechanisms existing in the atmosphere and the oceans, anomalies may have as much connection to reality as a mirage in the desert to the oasis.
Mods / Anthony,
In case someone hasn’t pointed out to you about the newspaper story on the VS comment stream over on Bart’s blog.
http://www.examiner.com/x-9111-Environmental-Policy-Examiner~y2010m3d24-Global-warming-Bigger-than-Climategate-more-important-than-Copenhagenits-statistical-analysis
Based on my many hours spent looking at the VS stream, the article is amazingly accurate.
John