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:
“This appears to be a case of “Willis doesn’t believe it, therefore its not true”. Hardly an adaquate basis for evaluating the method.” – Richard Telford (15:54:25).
What? Nonsense.
Doing mining with “guesses”, yes educated guesses, is one thing. Doing Climate Science using the same “guessing” techniques is simply frabricating data that just was not there. This is especially the case when you go on to make extreme and wild claims of doom and gloom and cause out of your wild claims massive amounts of public monies to be spent supporting the wild claims. It’s fraud when you do this knowing that you’re doing it. It’s fraud when you attempt to block other scientists work that refutes yours.
The problem with the climate data bases is that they have so many fudge factors and so much missing data that it’s not possible to use them for much of anything with any accuracy. You simply can’t use one temperature monitoring station for a 1200 km radius. To do so is like saying that you can take the temperature of the entire planet with just one thermometer. It’s junk science to even attempt to do so.
Dr. James Hansen, Junk Scientist who Fabricates Data out of thin air.
Science needs higher standards than this type of slop that Dr. James Hansen peddles. Much higher standards.
Doug Badgero (15:29:06)
“Are those Alaska temps NASA GISS “value added” trends or are they raw data trends? If they are raw data trends, has Alaska really been on such a continuous warming trend?”
In short, no.
there was a step change between 1974 and 1979 of 5 degrees warming. The rest of the world at the time was worried about the ‘coming ice age’. Strangely enough it peaked in 2004 and has dropped backed to the 60 year mean.
The Analysis done by the Alaska Climate Research Center
http://climate.gi.alaska.edu/ClimTrends/Change/TempChange.html
“It can be seen that there are large variations from year to year and the 5-year moving average demonstrates large increase in 1976. The period 1949 to 1975 was substantially colder than the period from 1977 to 2009, however since 1977 little additional warming has occurred in Alaska with the exception of Barrow and a few other locations. The stepwise shift appearing in the temperature data in 1976 corresponds to a phase shift of the Pacific Decadal Oscillation from a negative phase to a positive phase.”
Satellite data goes to 85 degrees or so. Misses most of the permanent arctic sea ice. It’s the surface temp over the permanent ice cover that’s of interest. I wouldn’t think it matters how well the average correlation or trend is between stations 1200 kilometers apart when you have such radical surface changes across the divide such as going from land to open water to seasonal ice to permanent ice. Plus there’s an arctic ozone hole to consider as well. It’s not as big as the antarctic hole but it’s there and the existing stations aren’t under it. That’s another variable that isn’t reflected in correlations between distant stations.
Hansen must be aware of the above. Of course he’ll just say it doesn’t matter because we know what the trend is because arctic sea ice extent is on a downward trend. But that doesn’t indict CO2. I think Pielke is right on the money in this article:
http://wattsupwiththat.com/2009/08/21/soot-and-the-arctic-ice-%E2%80%93-a-win-win-policy-based-on-chinese-coal-fired-power-plants%E2%80%9D/
No source of black soot can make it to the south pole which handily explains why the antarctic interior isn’t warming.
The good part of soot for the anti-human crowd who thinks people are destroying the planet is that most of the soot is anthropogenic. The bad part for them is that the United States doesn’t produce much of it since the Clean Air Act of 1963 required drastic reductions through particulate filters on industrial smokestacks and clean burning diesel and not using slash & burn agriculture and so forth. So someone else has to get the blame and in all fairness that means someone else is responsible for the cost of amelioration.
Of course I’m still of the opinion that no one has a cost benefit analysis that clearly shows we should spend one thin dime on amelioration. Polar bear population is actually increasing. A northwest passage is something to be desired. What exactly is the downside of less arctic sea ice?
At this point assuming Dr. James Hansen were to be a reputable scientist he would say, oh, that’s interesting… you are right, I bow down and salute you for refuting a key method that underpins the entire GISS dataset. We will immediately forthwith cease using this technique and we will review all past data sets fabricated by Nasa GISS or that are derived thereof from our work and alert all scientists who have written papers based upon our work and warn the to make corrections to their science accordingly.
I’m listening… James?
Giss has the center of their most poleward sub-boxes at 84.3n and they are 9 degrees wide. The sub-box bottoms are at 81.9n. Since any station within 1200 kilometers of 84.3n is used in calculating the value for those sub-boxes, stations as far south as 73.6n can have an effect on the polar sub-boxes. This results in the effect being propagated 1800 kilometers to the pole. The effect is minimal at that distance if other closer stations are available, but there are occasions when a polar box will have only one station which may be at extreme distance which qualifies. In that case its full value is used. Isachsen NW 78.78n early in the century comes to mind as an example.
Side note:
Giss also has a rule which states that if a station is within a sub-box it it used to calculate that sub-box anomaly. This is targeted at one station. Amundsen at the south pole. That is why you always see total coverage from the pole to 82s when the smoothing radius is set to 250 kilometers. Amundsen is actually 635 kilometers to the sub-box centers and otherwise wouldn’t be shown at all.
JAE (16:06:04)
Jae, you are correct. However, at the moment, the anomaly and the trends are quite similar, and I couldn’t find a GISS map of the trends. Here’s the closest I could find, zonal trends by month. I didn’t know if most folks would be able to interpret it, so I showed the anomaly map.
As you can see, they suffer from the same problem, which as you point out is the incorrect application of the 1200 km extrapolation.
Just as I suspected.
Global Warming a Boon for Greenland’s Farmers
Same thing my buddies up north have been telling me about global warming -the more the better.
Quite frankly the fact that the high northern latitudes are warming faster than anywhere else is I believe more like a godsend coming at just the right time in just the right place to boost agricultural output in order to feed a growing population. Maybe I should invest in land in Siberia in anticipation of my granchildren being able to lease it out for agriculture.
Huh. The post is about GISS, GISTemp is used, yet E.M. Smith has not yet expertly commented, nor has carrot eater proclaimed the divine infallibility of Hansen’s work as confirmed by the genius of St. Tamino which incontrovertibly shows that E.M.S. is an incompetent ignorant buffoon (to use the language I’ve seen used on Tamino’s site).
Thus this thread is too young. I’ll check back later.
I have a question about the goal of temperature trend interpolation. Interpolation and homogenization make for pretty pictures, but do they tell us anything we need to know? Instead, it seems to me that any significant “global” temperature trends would readily appear in statistical correlations of thermometers as long as the data of the two thermometers can be compared over similar time spans. This should have the added benefit of not throwing away useful discrete environmental information for individual thermometers that reveals correlations to local changes such as urbanization, agricultural changes, etc.
Regardless of the validity of their extrapolations, GISS more or less correctly shows much of the Arctic and most of the Antarctic cooling over the last 80 years in their trend maps. The red area over Russia is probably incorrect.
http://data.giss.nasa.gov/work/gistemp/NMAPS/tmp_GHCN_GISS_1200km_Trnd02_1930_2010/GHCN_GISS_1200km_Trnd02_1930_2010.gif
The areas around the Pacific warmed a lot during the PDO shift in 1977, and this is reflected by red colors in Canada and Alaska.
The big problem is with the GISS baseline period, which was largely a period of unusual cold, so the anomaly maps always look warm.
[snip – sorry Willis, I’m not going to let this turn into a tobacco war started by Anu, his comments have been snipped as well – Anthony]
Anu (16:32:13)
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.
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 …
Willis,
Thanks. I see now that the linear trend lines were just to make a point. The underlying data is not linear.
In response James Hansen says,”what you talkin bout Willis?”
Anu (16:32:13)
How about we just leave it empty, along with all the other parts of the planet that don’t have temperature stations, since we don’t know what the temperature is?
How about you notice that since much of the earth has only occasional widely scattered temperature stations, we don’t know the “planetary average”?
How about you deal with the fact that despite high correlations, nearby stations can have very different trends, so using one station to extrapolate out 1200 km can give a very wrong answer?
These questions and more …
Look at all the grey in the maps over ocean areas. GREY???? Grey isn’t on the scale. What does grey mean? I can only suppose that it is supposed to represent areas of insufficient data.
But … but …. you see the issue don’t you.
Why has the Arctic been filled in while these other ocean areas of insufficient data have not. Wouldn’t it have been more honest to leave the arctic grey as well?
Anu (16:32:13)
I don’t recall saying that they did. Take a look at Figure 1. Do you notice the areas in gray? Those are areas of missing data. They are not filled with anything, not the planetary average, nothing. CRU does the same, as far as I know, when data is missing.
Or as I quoted Trenberth saying above,
Missing. Not “average global anomaly”. Missing.
subtlety.leads.to.confusion (16:44:35) :
[quote]But wait a second, aren’t all the global temperature analyses done based on 5×5 grid cells?
And isn’t it true that the farther from the equator one goes, the smaller the physical area of each grid cell becomes?
And then, if you fill in some high temperature numbers in some high latitude cells, those numbers will be over-represented in the subsequent summation process?
Shouldn’t each grid cell be weighted by latitude[/quote]
Not to mention the huge scaling bias using such a map to present such ‘results.’
subtlety.leads.to.confusion (16:44:35)
Good questions. Different teams use different gridcells. GISS uses equal-area gridcells, GHCN and CRU use 5×5 gridcells. The latter are area-averaged as you say, using the cosine of the mid-latititude of the cell.
Scott (15:20:17)
R^2 is the coefficient of determination, and roughly translates as the amount of variation explained by the model (0 to 1). So using data with 0.5 R results in using data where one independent variable (the temperature at one station) could explain at most 0.25 or 25% of the other variable’s (other station’s) variation. While this may be useful somehow, this, as Willis explains, says nothing of the trend differences, magnitude of correlation slope, intercept, etc.
The best way to eliminate this in an experiment is to MEASURE more stuff!
Hmmm…
Series 0,1,2,3 and 0,5,10,15 are very well correlated indeed, with correlation coefficient r=1. Yet, second the trend is five times larger than the first one, isn’t it? That’s pretty basic stuff… and pretty basic flaw in their reasoning.
So what happened to the peer review process, how come no-one noticed it before?
Willis: The following is a comparison graph of the GISTEMP (LST+SST) annual zonal mean trends from 1880 to 2009 with 250km and 1200km smoothing. (For some reason, EXCEL didn’t like the data and refused to plot both datasets, so I had to rely on another spreadsheet, one that I’m not familiar with.) The GISTEMP zonal mean trend data shows that the trends correlate well between 55S and 55N. Poleward of those latitudes, the 250km and 1200km smoothed data diverge. In fact, at high latitudes, the zonal mean data with the 1200km smoothing can rise with latitude while the zonal mean data with the 250km smoothing drops. Interesting effect.
http://i42.tinypic.com/21abja0.png
And here’s a gif animation of the two corresponding maps:
http://i41.tinypic.com/2d2j1if.gif
Very informative, Willis – loved your gentle instruction on the meaning of correlation and love your posts as always. Just one thing: I got hung up on the caption for figure 3 that describes the Arctic Ocean as being roughly three times the size of Alaska so I checked out the actual areas: the Arctic Ocean is 14.056 Million square km vs Alaska’s area of 1.718 square km so the Arctic Ocean is more than 7 times the size of Alaska. Of course this strengthens your argument regarding the serious problem with the extrapolation of a few data points.
From the arctic research center at the University of Alaska. No ‘slow steady warming’…a huge step in 1975.
[image]http://climate.gi.alaska.edu/ClimTrends/Change/graphics/temp_dep49-09_F_sm.jpg[/image]
Hmm, the thing that gets me with this is that you might be able to now show strong correlation and trending between two points on the globe 1200kms apart – but this is not guaranteed to be true into the future.. It could just be a pure artifact of sampling and measurement errors as well.
I also had a look through the GISS code around this 1200km circle stuff and found that they also pull in stations within a 1200km ‘square’ lat/long bounding box containing the 1200km circle as projected onto the globe – stations in this in-box/ex-circle space then get given the same weighting as the furthest station within the circle (if memory serves)… sounds like a fudge to make up for a lack of stations in certain areas..
Someone should do a recode to show on the globe confidence in terms of mean true distance of contributing stations per cell. that might shed some useful light on where things are going right off the rails and achieving low earth orbit..