Guest post by Ira Glickstein
Thanks to WUWT readers who posted estimates of how much of the supposed 0.8ºC Global Warming since 1880 was due to Data Bias, Natural Cycles, and AGW (human-caused warming). I am happy with the results even though the average for AGW came out higher than my original estimate.
This is the fifth of my Tale of the Global Warming Tiger series where I allocated the supposed 0.8ºC warming since 1880 to: (1) Data Bias (0.3ºC) , (2) Natural Cycles (0.4ºC) , and (3) Human-caused global warming – AGW (0.1ºC). Click Tiger’s Tale and Tail :^) to read the original story.
WUWT COMMENTERS SAY
As the above graphic indicates, WUWT Commenters who provided their own estimates generally agreed with my allocation, with the interesting exception of AGW, where the average is 0.18ºC, nearly double my original allocation of 0.1ºC. Natural Cycles averaged out at 0.33ºC, a bit lower than my original 0.4ºC. Data Bias averaged out at 0.28ºC, a bit lower than my original 0.3ºC. While this is not a scientific poll, it certainly shows a wide variety of Climate Science opinion is alive and well here at WUWT.
Far from being a Global Warming Tiger mostly due to atmospheric CO2 from human burning of fossil fuels and land use, and on its way to 2ºC to 5ºC or more, according to the IPCC, it appears we are actually dealing with a Global Warming Pussy Cat, with warming since 1880 around 0.5ºC to 0.6ºC,and stabilizing despite continued rise in CO2, much of it human-caused.
Some who responded put AGW as low as ZERO (while others put it as high as 0.7ºC), some put Natural Cycles as low as ZERO (while others put it as high as 0.55ºC), and some put Data Bias as low as ZERO (while others put it as high as 0.65ºC). At the end of this posting, I’ve tabulated your estimates, along with the names of those kind enough to provide them. THANKS!
When everything settles out over the coming decades, which I believe will be marked by stabilization of Global temperatures, and perhaps a bit of Global Cooling, I think your estimates will turn out to be more prescient than that of the official climate Team! One Commenter humorously posted “Jim Hansen’s” estimates as: AGW = +3.3ºC, Natural Cycles = – 2.5ºC, and, of course, Data Bias = 0.0ºC.
IS ALL THE TEMPERATURE DATA USELESS – OR IS IT THE ANALYSIS?
When I discussed the controversy about the temperature data collected since 1880 with my PhD advisor (with whom I am still in regular contact) he reminded me that, given a large number of measurements by different observers, using a variety of thermometers, and taken at a variety of locations and times, the random errors would largely cancel each other out. Even systematic errors in given thermometers, which might be calibrated a bit high or low, and given observers, who might tend to round the numbers up or down, would largely cancel out. Indeed, he said, even long-term systematic bias would hardly show up in the temperature trends. Thus, he assured me, while any individual reading may or may not be accurate, the overall temperature trend would be quite robust, to a high level of precision.
Of course, he is correct from an academic point of view. As a brilliant analyst once humorously explained to me, once we ASSSUME a perfectly smooth elephant with negligible mass, all sorts of wonderful circus tricks become possible!
Yes, errors may be categorized as:
- Perfectly Random (due to “noise” in the measurement process, and equally likely to be higher or lower than the truth) or,
- Perfectly Systematic (due to miscalibration of the measuring instrument, off by a constant amount, equally likely to be higher or lower than the truth), and assumed to be
- Perfectly Independent (not affected by any other measurement).
In the real world, however, these conditions seldom obtain, but they are necessary assumptions for statistical analysis to operate correctly. When a scientific study concludes that the results are correct, plus or minus a given amount (say +/- 0.05ºC), to a given statistical certainty (say 95%), they are implicitly assuming the three items above are satisfied.
In many cases, even if those assumptions are not perfectly true, they are close enough for the statistical results to be valid. How can we tell if Global Warming is one of those cases? Well, for a start, we can ask how ROBUST are the results. In other words, when they are analyzed by different people at different times, do they all come up with close to the same results? In the case of Global Warming data, as I have shown, even when the same exact data is analyzed by the same exact members of the official climate Team, the results vary by +/-0.2ºC or more, indicating that something is wrong with their basic assumptions.
Case #1
According to my posting, a graph of the US Annual Mean Temperature record from 1880 to 1998, published by NASA GISS in 1999, differs substantially for the record for the same years, published by them in 2011, see blink graphic below:
A commenter suggested that the 1999 chart did not look like what had been published by GISS in that year. Well, the 1999 chart I used came from a posting by Anthony who credited Zapruder.nl. An almost identical chart appeared at Climate Audit in 2007, linking to a Hansen 1999 News Release but that link now brings up a damaged image. However, I found an almost identical chart at GISS in a Hansen 1999 paper. The 2011 graphic I used was downloaded from GISS last month. The GISS re-analysis makes data after about 1960 warmer by up to 0.3ºC, while that prior to 1950 gets cooler by 0.1ºC.
Case #2
According to a GISS email, released under the Freedom of Information Act, records for US Annual Mean Temperature for 1934 and 1998 were re-analyzed seven times, and that resulted in a reduction of 1934’s lead of 0.5ºC warmer to a virtual tie. [The email is embedded in the graphic below.] In the latest GISS accounting, done after the date of the email, 1998 pulled ahead by a bit. (Our tax dollars at work.)
There is a need to analyze and adjust the raw temperature data when stations move or are encroached by development or when other changes are made to the equipment and enclosures or in the times of observation, etc. It seems that most of those changes would tend to exaggerate the amount of warming, yet those charged with analyzing the data seem to think otherwise. The reported temperatures always seem to increase with each re-analysis. That suggests an agenda on the part of those entrusted with the analysis.
DOES SATELLITE TEMPERATURE DATA SOLVE THE PROBLEM?
Satellite temperature measurements have been available starting in the late 1960’s, with good surface and tropospheric data available since late 1978. So, it would appear that, at least from 1979 on, given a uniform Global source set of data, global temperature trends have been accurately reported. However, according to Wikipedia
Satellites do not measure temperature. They measure radiances in various wavelength bands, which must then be mathematically inverted to obtain indirect inferences of temperature. The resulting temperature profiles depend on details of the methods that are used to obtain temperatures from radiances. As a result, different groups that have analyzed the satellite data have obtained different temperature trends. Among these groups are Remote Sensing Systems (RSS) and the University of Alabama in Huntsville (UAH). Furthermore the satellite series is not fully homogeneous – it is constructed from a series of satellites with similar but not identical instrumentation. The sensors deteriorate over time, and corrections are necessary for satellite drift in orbit. Particularly large differences between reconstructed temperature series occur at the few times when there is little temporal overlap between successive satellites, making intercalibration difficult. …
They go on to say “Satellites may also be used to retrieve surface temperatures in cloud-free conditions, generally via measurement of thermal infrared …”[Emphasis added] so it would appear that this type of instrumentation cannot reliably measure surface temperatures below clouds. That is problematic, since anyone who has been to a beach knows how cold it gets when a cloud happens to pass overhead and block the Sun!
Roy Spencer, PhD updates the UAH Global temperature datasets based on satellite data. He writes:
Since 1979, NOAA satellites have been carrying instruments which measure the natural microwave thermal emissions from oxygen in the atmosphere. The signals that these microwave radiometers measure at different microwave frequencies are directly proportional to the temperature of different, deep layers of the atmosphere. Every month, John Christy and I update global temperature datasets … that represent the piecing together of the temperature data from a total of eleven instruments flying on eleven different satellites over the years. As of early 2011, our most stable instrument for this monitoring is the Advanced Microwave Sounding Unit (AMSU-A) flying on NASA’s Aqua satellite and providing data since late 2002.
… Contrary to some reports, the satellite measurements are not calibrated in any way with the global surface-based thermometer record of temperature. They instead use their own on-board precision redundant platinum resistance thermometers calibrated to a laboratory reference standard before launch.[Emphasis added]
The last sentence is somewhat reassuring, but it does not resolve my questions about how they compensate for cloud cover. It appears highly likely that Global temperatures have increased since 1880 by around 0.5ºC, which would most likely increase the water vapor content of the atmosphere and, over time, result in more clouds, on average. Thus, depending upon how the satellite temperature data analysis corrects for cloudiness, that data might report more warming than actually occurs. In any case, it appears that the satellite data will help improve the general reliability of global tempeature data, assuming that the analysis is done properly, by experts who do not have any political agenda to “prove” or “disprove” Catastrophic AGW. Spencer appears to be a solid citizen in that respect.
CONCLUSIONS
In my postings (A-, B-, C-, D-) in this Tale of the Global Warming Tiger series, I asked for comments on my allocations: to: (1) Data Bias 0.3ºC, (2) Natural Cycles 0.4ºC, and (3) AGW 0.1ºC. Quite a few readers were kind enough to comment, either expressing general agreement or offering their own estimates. Here is a tabulation of their interesting inputs. THANKS!
| Anomaly due to — | Human (AGW) | Natural Cycles | Data Bias |
|---|---|---|---|
| A- | ºC | ºC | ºC |
| Bill Illis | 0.225 | 0.275 | 0.300 |
| Brian H | 0.450 | ||
| Edmh | 0.100 | ||
| Ágúst Bjarnason | 0.250 | 0.250 | 0.100 |
| B- | |||
| Ed Caryl | 0.000 | 0.300 | 0.500 |
| James Barker | 0.000 | 0.480 | 0.320 |
| JimF | 0.100 | 0.500 | 0.200 |
| richard verney | 0.000 | 0.550 | 0.250 |
| Scarface | 0.000 | 0.150 | 0.650 |
| Dave Springer | 0.500 | 0.000 | 0.300 |
| Mike Haseler | 0.100 | 0.300 | 0.200 |
| C- | |||
| Leonard Weinstein | 0.300 | 0.400 | 0.100 |
| TimC | 0.100 | 0.400 | 0.300 |
| Steve Reynolds | 0.400 | 0.250 | 0.150 |
| Eric Barnes | 0.150 | 0.450 | 0.200 |
| Lucy Skywalker | 0.000 | 0.300 | 0.500 |
| D- | |||
| Wayne | 0.100 | 0.300 | 0.400 |
| Eadler | 0.700 | 0.200 | 0.000 |
| Nylo | 0.200 | 0.400 | 0.200 |
| Minimum | 0.000 | 0.000 | 0.000 |
| Maximum | 0.700 | 0.550 | 0.650 |
| AVERAGE | 0.179 | 0.331 | 0.275 |
| Ira’s Estimates | 0.100 | 0.400 | 0.300 |
| Anomaly due to — | Human (AGW) | Natural Cycles | Data Bias |
Can anyone really measure a 1/2 of a degree….
…no
You can only arrive at that through math.
When you consider rounding, thermometers shrinking, stations getting dropped, the past getting colder, whole years missing, no statistical warming in a decade…………
…when that entire 1/2 of a degree can be accounted for in one blink chart
Snow a thing of the past, snowrain, warmcold, droughtflood…..
….more severe weather events, when those events are on a major decline
I don’t believe one word of any of it any more….
….and all of this fuss is about 1/2 of a degree
There are or were only about 6000 land stations in the GISSTemp records. Various investigators have analysed in good detail New Zealand, Australia and parts of Europe. WUWT readers have done grand work on the American mainland records. So the UHIE problem seems to have been nailed down in many areas. Also, the great dying of records in 1990 is well connected to the sudden global temperature anomaly rise in 1990. The work seems to have been done.
What is left is for someone to pull the “counter-fixed” station data into a group file and repeat the process that gives us a global history. Your work is interesting in that it shows that some have done this, at least with part of the data. When do we see a “data revealed” Skeptic graph? When do the Watts et al put their/our cards on the table and say “This, is what is really going on!”?
If the global temperature has truly increased by about 0.5C in the past 130 years, this is less than 0.05C per decade. Whilst I do not consider that one can measure temperatures to such accuracy, it is hardly anthing to get alarmed about.
It is interesting that Dave Springer does not consider that any of the ‘assessed’ temperature increase is due to natural causes. This is surprising since even the ‘Team’ generally accept that natural causes drove temperature increases through to the 1940s. It would be very odd if all those natural causes suddenly switched themselves off post 1940.
Ira, an interesting set of articles even if there is no real place for consensus views in Climate Science.
It should be easy to see how the natural cycles correlate to recent warming… Anthony, can you put a post together on that? It would be nice to see the data collected and well presented on this, and a great follow up to this post…
I don’t know why climate scientists bother doing scientific research when all they have to do to figure out how much global warming is anthropogenic is poll WUWT readers. This is how real science is done, alarmists! Stop wasting our tax dollars!!
Aw… Dana, no comment on what matters (O’Donnell et al 2010) vs (Steig 2009). PC’s must mean “Personal Computers” to some folks 😉
Ira, love it! And I haven’t even read a word yet!
I no longer see that tiger, just Sambo and all of that
oilbutter.Yes but according European Union of Climate science or something, pre industrial era T was normal, so everything since then is abnormal. This is clearly stated in their policy to keep T below 2 degrees above pre industrial era T. Somebody tell these monkeys that pre industrial era T was a little ice age whuih even the most fanatical warmists agrees was a natural cold period.
Ira, how in the heck do you have me under AGW, yuck!
Have you never read what I have been saying?
Ira,
The measurement bias is obviously “underestimated”. As your advisor pointed out, in any “untampered” distributions, one can convey that the “bias” can be sorted out. This is not the case, since all measurements have been “corrected” for various external “influences”. The data is just reduced to a series of numbers, +/- X degrees. Without the measurement in control (especially in a multi-variant), the rest is sci-fi or fantasy.
I don’t understand how it’s been acceptable for the two most important surface temperature archives to be left in the custody of Wigley/Jones, both attendees at the proto-IPCC Villach meeting in 1985 (the objective of which was to prove that human CO2 drove the climate — there was no evidence it had or would) and Hansen who by 1979 had the same conviction (without evidence) and who has recently stated that “…..The trains carrying coal to power plants are death trains. Coal-fired power plants are factories of death…”.
Truly bizarre — a situation which wouldn’t have been be tolerated in any other discipline I assume (as a layman).
I am not really surprised that the AGW estimate pans out at 0.179C. That is more or less Dr Roy Spencer’s estimate (around 20% of the 0.8C increase) and I believe that Professor Lindzen is in that ballpark as well. Having read Dr Spencer’s books and his arguments I am inclined to believe (as a layman) that this is probably right. That would give a decadel increase of around 0.02C rather than the AGW 0.2C per decade.
Strictly speaking you’ve missed out one of the figures from my own which I think was reduction in global dimming. But as it wouldn’t significantly change the result I’m not going to complain.
Obviously it’s totally unscientific, and speaking personally I’ve only got professional experience with temperature measurement and the rest is just the best information I have gleamed from others – but very insightful!
Before the exercise, I used a rough guess of 1/3:1/3:1/3. But having been forced to think through the exercise and reading the comments of others, I think the final averages are better guesses.
Ira:
My estimate of 0.250° 0.250° 0.100° (sum 0.6°C) is from the year 1998
( http://www.agust.net/sol/ ). As most of the other estimates add up to 0.8°C we can scale this old estimate up to approximately 0.33° 0.33° 0.13° (sum 0.8°C).
(Sorry for the precision in the numbers. I would have liked one significant number, maybe something like 0.3° 0.3° 0.2°).
When asked, I often answer that about half of the observed warming could be caused by humans, about half caused by natural variations, plus some measurement error. Then I add that uncertainty is so great that the word “half” can mean anything in the range 20% to 80%. Probably in favor of natural variations…
This does not affect your result. The result 0.179C (~ 0.2°C) for AGW is well acceptable in my mind.
Graphics: Less is more, Ira!
John Peter says: February 8, 2011 at 12:38 am
“I am not really surprised that the AGW estimate pans out at 0.179C. … That would give a decadel increase of around 0.02C rather than the AGW 0.2C per decade.”
I’m not sure of this 0.2C. I thought the IPCC had a range 0.13 – 0.52C/decade!
If, as I recall, the typical noise figure in the temperature record is around 0.1C/decade, then it’s already unlikely to have had this masking as to have had a decade without warming there would need to have been at least -0.13C/decade of climate noise.
If there is no evidence for further warming this century**, as the minimum required noise to hide this “warming” increases, we should get more and more confidence that the predicted warming is not happening. (aka BS)
Just for fun, I’ve tried to estimate when this theory is busted!
I hope you forgive my statistics (I’ve not found a single article on the probability statistics of 1/f type noise as the climate has) but using a rule of thumb, I’d guess that the equivalent of 95% confidence is a difference between predicted and actual of:
0.2 C in any one decade (2x typical noise of 0.1C = 95% confidence Gaussian)
0.17C in two decades (90% confidence)
0.13C in three decades (80% confidence)
I’ve scaled these figures down based these of it being unlikely to have two successive years of cooling to “hide” the warming. Having a run of one decade of cooling is twice as likely has two decades of cooling, which is again twice as likely as three decades of cooling. I suppose I ought to use single sided probability distribution figures, but as it’s only a rule of thumb (and unlike climate “scientists” I’m not paid to be right) I’m not being fussy.
This means the prediction of a minimum of 0.13C/decade warming is busted if by:
2011 it had cooled 0.07C
2021 it has warmed by 0.09C
2031 it has warmed by 0.26C
So it looks to me that if proper statisticians looked at it, the GW scam is busted if:
a) The trend since 2001** at any time shows net cooling below -0.07C/decade
b) or by 20165 there’s no net warming.
** (since 2001 – the year of the IPCC prediction of 1.4-5.8C by 2100)
Hadley and CRU give their surface measured data error bands as +/-1.0C. So how can any average be below this range. It is not statistically correct. It is pure alarmism and a figure plucked from a computer readout.
Satellites have errors but at least they get the overall picture of temperature which surface measurements do not.
According to some physicists a global average temperature is meaningless. To get an accurate temperature of anything is must be at equilibrium. This planet’s climate is never at equilibrium.
they are implicitly assuming the three items above are satisfied.
————
This is an assumption on your part.
It depends on the kinds of measurements and statistical analysis required.
1. there may be good and well understood reasons for making certain statistical assumptions.
2. In areas where it is important the statistics will actually be used to determine if the analysis assumptions are valid.
“As the above graphic indicates” …
Actually, Ira, the above graphic indicates nothing to me, as it’s so prettified that any embedded data there might be is completely lost.
I see I’m not the only person who thought so, (Peter Czerna, 1:47 am)
That suggests an agenda on the part of those entrusted with the analysis.
————
We’ll this is a big fat assumption in the face of ignorance of:
1. the analysis method
2. Who did the analysis
3. The agenda of the person doing the analysis
4. The honesty of the person doing the analysis
5. What checks and balances are in place.
In any case, it appears that the satellite data will help improve the general reliability of global tempeature data, assuming that the analysis is done properly, by experts who do not have any political agenda to “prove” or “disprove” Catastrophic AGW. Spencer appears to be a solid citizen in that respect.
—————
This is seriously weird. Roy Spencer is a very well known and committed climate skeptic. You could just as easily accuse Roy Spencer of having an agenda and thereby fudging the data as any other climate scientist.
On top of that Roy Spencer has been accused of cocking up several years of the satellite data analysis. I sort of assume this is true, though I would like to hear Roy’s version of events.
Digging into the attribution using the 1880- temperature and CO2 records I cannot get close to 18%. Seeking attribution statistically is often a mug’s game although a decent R2 should give an indication of the level of variability in annual temperature than can be explained by variability in the independent variable. Putting trust in Ockham and going for the simplest statistical analysis first: variation in CO2 does not appear to explain any of the observed variation in temperature. Regressing annual % change in temperature v annual % change in CO2 both contemporaneously and lagged the R2 using OLS and MLR does not suggest any causation. At this level, in my field, I wouldn’t bother going any further. Both data series look to me like random walks with some upward drift but that’s where the similarity ends. Statistically they appear quite unconnected. Maybe I am missing something. I had better find out before I try anything more sophisticated!
Ira,
I don’t see why the reader inputs are apparently grouped by A, B, C, and D when their answers don’t support that grouping. For example, Ed Caryl is under B-Data Bias but Lucy Skywalker who gives the same breakdown of 0.000, 0.300, and 0.500 is under C-Natural Cycles. Also, James Barker, JimF, richard verney, and Mike Haseler all give bigger weight to Natural Cycles but are apparently part of the B-Data Bias group. The presentation is misleading.
“Contrary to some reports, the satellite measurements are not calibrated in any way with the global surface-based thermometer record of temperature. They instead use their own on-board precision redundant platinum resistance thermometers calibrated to a laboratory reference standard before launch.”
In the “www.burnsengineering.com” is interesting papers about “Error Sources
That Effect Platinum Resistance Thermometer Accuracy” and they show, that platinum resistance thermometers resistance is changing due to prolonged exposure to high temperatures/radiation and industry standarts alows change by approximately 0.1°C after 1000 hours for a 350°C exposure. I checked publically available papers of precision redundant platinum resistance thermometers which were used in NASA’s Aqua satellite and producers of these construction elements in their own WWW page FAQ recomends to check accuracy of thermometers yearly… 🙂
Hi Wayne, my spreadsheet table is organized according to under which of my topic postings commenters provided their estimates.
You gave your estimate under my “Some People Claim There’s a Human to Blame” topic:
That is why your estimates appear in that section of my table. I did not mean to imply that you think AGW is a major component. Indeed, your entry in my table reads “Wayne 0.100 0.300 0.400” with only 0.1ºC for Human-AGW, 0.3ºC for Natural Cycles, and 0.4ºC for Data Bias, exactly as you specified in your comment reproduced above.
NOTE: I have updated the table to remove the text and left only the “A-“, “B-“, “C-“, and “D-” to indicate to which of my postings the commenter was replying.