Adjusting Temperatures for the ENSO and the AMO

NOTE: Zip file downloads of models with data have been fixed, see end of this post – Anthony

Source: Mantua, 2000

The essay below has been part of a back and forth email exchange for about a week. Bill has done some yeoman’s work here at coaxing some new information from existing data. Both HadCRUT and GISS data was used for the comparisons to a doubling of CO2, and what I find most interesting is that both Hadley and GISS data come out higher in for a doubling of CO2 than NCDC data, implying that the adjustments to data used in GISS and HadCRUT add something that really isn’t there.

The logarithmic plots of CO2 doubling help demonstrate why CO2 won’t cause a runaway greenhouse effect due to diminished IR returns as CO2 PPM’s increase. This is something many people don’t get to see visualized.

One of the other interesting items is the essay is about the El Nino event in 1878. Bill writes:

The 1877-78 El Nino was the biggest event on record.  The anomaly peaked at +3.4C in Nov, 1877 and by Feb, 1878, global temperatures had spiked to +0.364C or nearly 0.7C above the background temperature trend of the time.

Clearly the oceans ruled the climate, and it appears they still do.

Let’s all give this a good examination, point out weaknesses, and give encouragement for Bill’s work. This is a must read. – Anthony

Adjusting Temperatures for the ENSO and the AMO

A guest post by: Bill Illis

People have noted for a long time that the effect of the El Nino Southern Oscillation (ENSO) should be accounted for and adjusted for in analyzing temperature trends.  The same point has been raised for the Atlantic Multidecadal Oscillation (AMO).  Until now, there has not been a robust method of doing so.

This post will outline a simple least squares regression solution to adjusting monthly temperatures for the impact of the ENSO and the AMO.  There is no smoothing of the data, no plugging of the data; this is a simple mathematical calculation.

Some basic points before we continue.

–         The ENSO and the AMO both affect temperatures and, hence, any reconstruction needs to use both ocean temperature indices.  The AMO actually provides a greater impact on temperatures than the ENSO.

–         The ENSO and the AMO impact temperatures directly and continuously on a monthly basis.  Any smoothing of the data or even using annual temperature data just reduces the information which can be extracted.

–         The ENSO’s impact on temperatures is lagged by 3 months while the AMO seems to be more immediate.  This model uses the Nino 3.4 region anomaly since it seems to be the most indicative of the underlying El Nino and La Nina trends.

–         When the ENSO and the AMO impacts are adjusted for, all that is left is the global warming signal and a white noise error.

–         The ENSO and the AMO are capable of explaining almost all of the natural variation in the climate.

–         We can finally answer the question of how much global warming has there been to date and how much has occurred since 1979 for example.  And, yes, there has been global warming but the amount is much less than global warming models predict and the effect even seems to be slowing down since 1979.

–         Unfortunately, there is not currently a good forecast model for the ENSO or AMO so this method will have to focus on current and past temperatures versus providing forecasts for the future.

And now to the good part, here is what the reconstruction looks like for the Hadley Centre’s HadCRUT3 global monthly temperature series going back to 1871 – 1,652 data points.

Click for a full sized image

Click for a full sized image

I will walk you through how this method was developed since it will help with understanding some of its components.

Let’s first look at the Nino 3.4 region anomaly going back to 1871 as developed by Trenberth (actually this index is smoothed but it is the least smoothed data available).

–         The 1877-78 El Nino was the biggest event on record.  The anomaly peaked at +3.4C in Nov, 1877 and by Feb, 1878, global temperatures had spiked to +0.364C or nearly 0.7C above the background temperature trend of the time.

–         The 1997-98 El Nino produced similar results and still holds the record for the highest monthly temperature of +0.749C in Feb, 1998.

–         There is a lag of about 3 months in the impact of ENSO on temperatures.  Sometimes it is only 2 months, sometimes 4 months and this reconstruction uses the 3 month lag.

–         Going back to 1871, there is no real trend in the Nino 3.4 anomaly which indicates it is a natural climate cycle and is not related to global warming in the sense that more El Ninos are occurring as a result of warming.   This point becomes important because we need to separate the natural variation in the climate from the global warming influence.

Click for full sized image

Click for full sized image

The AMO anomaly has longer cycles than the ENSO.

–         While the Nino 3.4 region can spike up to +3.4C, the AMO index rarely gets above +0.6C anomaly.

–         The long cycles of the AMO matches the major climate shifts which have occurred over the last 130 years.  The downswing in temperatures from 1890 to 1915, the upswing in temps from 1915 to 1945, the decline from 1946 to 1975 and the upswing in temps from 1975 to 2005.

–         The AMO also has spikes during the major El Nino events of 1877-88 and 1997-98 and other spikes at different times.

–         It is apparent that the major increase in temperatures during the 1997-98 El Nino was also caused by the AMO anomaly.  I think this has lead some to believe the impact of ENSO is bigger than it really is and has caused people to focus too much on the ENSO.

–         There is some autocorrelation between the ENSO and the AMO given these simultaneous spikes but the longer cycles of the AMO versus the short sharp swings in the ENSO means they are relatively independent.

–         As well, the AMO appears to be a natural climate cycle unrelated to global warming.

Click for full sized image

Click for full sized image

When these two ocean indices are regressed against the monthly temperature record, we have a very good match.

–         The coefficient for the Nino 3.4 region at 0.058 means it is capable of explaining changes in temps of as much as +/- 0.2C.

–         The coefficient for the AMO index at 0.51 to 0.75 indicates it is capable of explaining changes in temps of as much as +/- 0.3C to +/- 0.4C.

–         The F-statistic for this regression at 222.5 means it passes a 99.9% confidence interval.

But there is a divergence between the actual temperature record and the regression model based solely on the Nino and the AMO.  This is the real global warming signal.

Click for full sized image

Click for full sized image

The global warming signal (which also includes an error, UHI, poor siting and adjustments in the temperature record as demonstrated by Anthony Watts) can be now be modeled against the rise in CO2 over the period.

–         Warming occurs in a logarithmic relationship to CO2 and, consequently, any model of warming should be done on the natural log of CO2.

–         CO2 in this case is just a proxy for all the GHGs but since it is the biggest one and nitrous oxide is rising at the same rate, it can be used as the basis for the warming model.

This regression produces a global warming signal which is about half of that predicted by the global warming models.  The F statistic at 4,308 passes a 99.9% confidence interval.

Click for full sized image

Click for full sized image

–         Using the HadCRUT3 temperature series, warming works out to only 1.85C per doubling of CO2.

–         The GISS reconstruction also produces 1.85C per doubling while the NCDC temperature record only produces 1.6C per doubling.

–         Global warming theorists are now explaining the lack of warming to date is due to the deep oceans absorbing some of the increase (not the surface since this is already included in the temperature data).  This means the global warming model prediction line should be pushed out 35 years, or 75 years or even 100s of years.

Here is a depiction of how logarithmic warming works.  I’ve included these log charts because it is fundamental to how to regress for CO2 and it is a view of global warming which I believe many have not seen before.

The formula for the global warming models has been constructed by myself (I’m not even sure the modelers have this perspective on the issue) but it is the only formula which goes through the temperature figures at the start of the record (285 ppm or 280 ppm) and the 3.25C increase in temperatures for a doubling of CO2.   It is curious that the global warming models are also based on CO2 or GHGs being responsible for nearly all of the 33C greenhouse effect through its impact on water vapour as well.

Click for larger image

Click for larger image

The divergence, however, is going to be harder to explain in just a few years since the ENSO and AMO-adjusted warming observations are tracking farther and farther away from the global warming model’s track.  As the RSS satellite log warming chart will show later, temperatures have in fact moved even farther away from the models since 1979.

Click for larger image

Click for larger image

The global warming models formula produces temperatures which would be +10C in geologic time periods when CO2 was 3,000 ppm, for example, while this model’s log warming would result in temperatures about +5C at 3,000 ppm.  This is much closer to the estimated temperature history of the planet.

This method is not perfect.  The overall reconstruction produces a resulting error which is higher than one would want.  The error term is roughly +/-0.2C but the it does appear to be strictly white noise.   It would be better if the resulting error was less than +/- 0.2C but it appears this is unavoidable in something as complicated as the climate and in the measurement errors which exist for temperature, the ENSO and the AMO.

This is the error for the reconstruction of GISS monthly data going back to 1880.

Click for larger image

Click for larger image

There does not appear to be a signal remaining in the errors for another natural climate variable to impact the reconstruction.  In reviewing this model, I have also reviewed the impact of the major volcanoes.  All of them appear to have been caught by the ENSO and AMO indices which I imagine are influenced by volcanoes.  There appears to be some room to look at a solar influence but this would be quite small.  Everyone is welcome to improve on this reconstruction method by examining other variables, other indices.

Overall, this reconstruction produces an r^2 of 0.783 which is pretty good for a monthly climate model based on just three simple variables.  Here is the scatterplot of the HadCRUT3 reconstruction.

Click for a larger image

Click for a larger image

This method works for all the major monthly temperature series I have tried it on.

Here is the model for the RSS satellite-based temperature series.

Click for a larger image

Click for a larger image

Since 1979, warming appears to be slowing down (after it is adjusted for the ENSO and the AMO influence.)

The model produces warming for the RSS data of just 0.046C per decade which would also imply an increase in temperature of just 0.7C for a doubling of CO2 (and there is only 0.4C more to go to that doubling level.)

Click for a full sized image

Looking at how far off this warming trend is from the models can be seen in this zoom-in of the log warming chart.  If you apply the same method to GISS data since 1979, it is in the same circle as the satellite observations so the different agencies do not produce much different results.

Click for larger image

Click for larger image

There may be some explanations for this even wider divergence since 1979.

–         The regression coefficient for the AMO increases from about 0.51 in the reconstructions starting in 1880 to about 0.75 when the reconstruction starts in 1979.  This is not an expected result in regression modelling.

–         Since the AMO was cycling upward since 1975, the increased coefficient might just be catching a ride with that increasing trend.

–         I believe a regression is a regression and we should just accept this coefficient.  The F statistic for this model is 267 which would pass a 99.9% confidence interval.

–         On the other hand, the warming for RSS is really at the very lowest possible end for temperatures which might be expected from increased GHGs.  I would not use a formula which is lower than this for example.

–         The other explanation would be that the adjustments of old temperature records by GISS and the Hadley Centre and others have artificially increased the temperature trend prior to 1979 when the satellites became available to keep them honest.  The post-1979 warming formulae (not just RSS but all of them) indicate old records might have been increased by 0.3C above where they really were.

–         I think these explanations are both partially correct.

This temperature reconstruction method works for all of the major temperature series over any time period chosen and for the smaller zonal components as well.  There is a really nice fit to the RSS Tropics zone, for example, where the Nino coefficient increases to 0.21 as would be expected.

Click for a full sized image

Unfortunately, the method does not work for smaller regional temperature series such as the US lower 48 and the Arctic where there is too much variation to produce a reasonable result.

I have included my spreadsheets which have been set up so that anyone can use them.  All of the data for HadCRUT3, GISS, UAH, RSS and NCDC is included if you want to try out other series.  All of the base data on a monthly basis including CO2 back to 1850, the AMO back to 1856 and the Nino 3.4 region going back to 1871 is included in the spreadsheet.

The model for monthly temperatures is “here” and for annual temperatures is “here” (note the annual reconstruction is a little less accurate than the monthly reconstruction but still works).

I have set-up a photobucket site where anyone can review these charts and others that I have constructed.

So, we can now adjust temperatures for the natural variation in the climate caused by the ENSO and the AMO and this has provided a better insight into global warming.  The method is not perfect, however, as the remaining error term is higher than one would want to see but it might be unavoidable in something as complicated as the climate.

I encourage everyone to try to improve on this method and/or find any errors.  I expect this will have to be taken into account from now on in global warming research.  It is a simple regression.

UPDATED: Zip files should download OK now.

SUPPLEMENTAL INFO NOTE: Bill has made the Excel spreadsheets with data and graphs used for this essay available to me, and for those interested in replication and further investigation, I’m making them available here on my office webserver as a single ZIP file


Annual Temp Anomaly Model 171K Zip file

Monthly Temp Anomaly Model 1.1M Zip file

Just click the download link above, save as zip file, then unzip to your local drive work folder.

Here is the AMO data which is updated monthly a few days after month end.

Here is the Nino 3.4 anomaly from Trenbeth from 1871 to 2007.

And here is Nino 3.4 data updated from 2007 on.

– Anthony


newest oldest most voted
Notify of
Pamela Gray

Isn’t water vapor a greenhouse gas? Is there a chart of water vapor changes at different atmospheric levels during this same time span that might coincide with increased temperatures? And if CO2’s affect on water vapor were removed, would that leave us with a correlation that more closely matches water vapor changes than CO2 changes? One more question, what does a changing ocean cycle do to water vapor? Is there a lag? Is there a point at which the climate becomes more sensitive to water vapor? In other words, is there a tipping point for water vapor? I can surmise that when ocean cycles go the other way, water vapor would start to change as well and could stall climate changes or even reverse them.


“There appears to be some room to look at a solar influence but this would be quite small.”
I am no expert, but I suspect the oceans are being driven by the solar influence. It is becoming increasingly clear that the oceans are running the climate show on this planet.


Thanks. Anthony, I am having trouble with the two spreadsheets. I have tried to download them both in open office and excel and I get a message that they are corrupted
REPLY: Try now I’ve added a ZIP file. – Anthony


There is also a chicken/egg question of warmer water outgassing or at the least not absorbing as much CO2. When oceans warm they take in less CO2 and when soils warm, they produce more CO2 due to increased rates of biological decomposition of organic matter. A recent (in the past week) paper figures that soils produce 10x more CO2 than all human activities combined. While a few degrees doesn’t make much difference in tropical regions, it can make a HUGE difference in temperate regions where a degree or two in local temperature might mean the difference between “frozen” and active biological decomposition. Also, a warmer year will result in a longer period above freezing and just a couple of weeks of additional time is again more than 10x human caused emissions over those two weeks. So an extra frost-free week could be equal to 10 weeks of additional human emissions from natural sources.
So CO2 may still not be a cause and is still quite likely to be a result of warming.

Bill: I’ll have to delve deeper into your post tomorrow, but until then, here’s something to ponder. Why does a running total of the Trenberth NINO3.4 SST anomaly data create a curve that mimics the global temperature anomaly curve?
I can scale that running total with a coefficient from a Trenberth paper on ENSO and come up with a very reasonable correlation between the running total and global temperature? WHY?

David L. Hagen

For a approach, fitting the Pacific Decadal Oscillation and CO2 to temperature trends, see: Roy W. Spencer’s model, Global Warming as a Natural Response to Cloud Changes Associated with the Pacific Decadal Oscillation (PDO)

The “PDO-only” (dashed) curve indeed mimics the main features of the behavior of global mean temperatures during the 20th Century — including two-thirds of the warming trend. If I include transient CO2 forcing with the PDO-forced cloud changes (solid line labeled PDO+CO2), then the fit to observed temperatures is even closer.
It is important to point out that, in this exercise, the PDO itself is not an index of temperature; it is an index of radiative forcing which drives the time rate of change of temperature.

Would Bill Illis’ fit above improve by using a combination of PDO, ENSO and AMO with ln(CO2)? Or are PDO, ENSO and AMO sufficiently interrelated as to not be independent?
Can the effects of the Urban Heat Island effect be quantitatively separated from CO2 forcing on temperature trends?

Anthony, only Java -type URLs come up at the links. Can you provide non-Java URLs for the data? Thanks, Steve
REPLY: Hi Steve, I’m not sure what you are seeing. I have no java involved in any of this, like you I dislike it. I noted some previous CA comment where you mentioned something similar. I think whatever version of the Java engine you have installed on your PC may be intercepting links.
My advice, uninstall java from your machine, reboot, then reinstall it.

IT would also be helpful if Bill added original sources for the data (as URLs to the ENSO version and AMO version as used, for example.)

Bill Illis

Regarding water vapour Pamela, there isn’t really good data to show changes in water vapour over time. The only non-confirmed data that there is shows there has been a very slight decline in relative humidity.
The global warming models are based on relative humidity staying more-or-less constant as temperature fluxuates up and down. Some studies show this is the case while others show there is some variation that we don’t understand right now.
The water vapour question is the big remaining question in global warming and how big the impact will be.

Bill Illis

To Steve
Here is the AMO data which is updated monthly a few days after month end.
Here is the Nino 3.4 anomaly from Trenbeth from 1871 to 2007.
And here is Nino 3.4 data updated from 2007 on.

Jeff L

This is very interesting, well thought out work on first blush. Since this is largely a statistical analysis, I would really like to see CA / Steve McIntyre take a hard look at it & render his opinion on the statistical validity (and comment here if possible). If it appears to hold up, I would encourage Bill to try to get it published, perhaps as a co-author with one or more names that could lend weight & credence to the publication.
I have taken a similar approach for seasonal forecasting of front range snowfall here in Colorado with pretty good success. Interestingly enough, I also found my best correlations with ENSO & AMO and poor correlation with solar activity & volcanic activity / optical thickness data. I think that not only do the oceans rule our long term climatic trends but also largely rule our seasonal trends.
Something to consider – we know the ocean has thermohaline circulation cycles of up to 1000+ years. If the ocean circulation has cycles of 1000 + years, could it also have heat content cycles up to 1000 + years & could that be a significant driver / component of even longer cycles of climate change we observe? Are there proxies out there that could assess this in a manor similar to what Bill just did for the last 130 years? – isotope data possibly?
I personally think that Bill is just scratching the surface of what this general multi-variate technique could bring to the table for climate data analysis.

Bill Illis

To Bob Tisdale,
There is no logical/physical reason to include a running total for the Nino 3.4 anomaly which extends over years. I could be persuaded for a running total of a few months but one just needs to examine the up and down of temperatures in the 1997-98 El Nino, for example, to see there is no accumulating impact. The direct and continuous impact appears to work better and is more logical from a physical perspective in my mind.


Great detective work Bill. Now we await the pitbull (Gavin) to see if he takes a bite!

Alan S. Blue

Good work Bill Illis.
One lingering issue I kept running into in my own (non-climate) empirical fitting work is trying to avoid adding parameters I’m not certain are needed. And at each step, trying the simplest possible influence from a factor before assuming a more complex relationship.
What that boils down to is:
What happens to your fit if you just regress on a monotonically increasing line instead of ln(co2)?
Is the fit substantially better or worse?
Because a fair number of papers split “the warming” into a slice due to humans (AGW), and a slice that isn’t necessarily. Being able to differentiate the two would be excellent.

Steve Hempell

Still can’t get spreadsheets with zip file. Lots of #VALUEs (using Excel 2007).
Only get this on the link – no way to download anything.
“ndex of
Up to higher level directory
Name Size Last Modified”
Have you read this:
On reading seems like NASA gobbly-gook and makes statements like this:
“With new observations, the scientists confirmed experimentally what existing climate models had anticipated theoretically” and
“Because the new precise observations agree with existing assessments of water vapor’s impact, researchers are more confident than ever in model predictions that Earth’s leading greenhouse gas will contribute to a temperature rise of a few degrees by the end of the century.” What’s a “few”?
What’s your take on this “News”?


Bill, I have some plots somewhere which suggest that ENSO and global temperature are correlated with about a 3 month lag and then correlated again, weakly, with about a 15 month lag.
My thinking was that the results reflected either an odd flaw in my approach or, if real, a secondary impact of ENSO on Indian Ocean SST.
I’ll see if I can find, or reconstruct, those plots during the upcoming US holiday.

Steve Hempell

Still having problems with zip file. Getting lots of #VALUE.
Also problems with
Can’t seem to download anything.
Have you read this:
Any comments?

Bill Illis

To Alan S. Blue
The theory of global warming is based on a logarithmic relationship of CO2/GHGs to temperature impact.
A linear model works fine until you move far away from the current CO2 levels of 387 ppm. In fact, right now, CO2 levels are increasing at a slightly exponential rate (0.8% acceleration) per year and the warming trend would go ballistic exponential in no time if you didn’t use the log formula.
It doesn’t make much difference for short periods of time but what would Earth’s temp be when CO2 levels were 3,000 ppm 350 million years ago – 8 times the current average of 15C or about 116C – it was only about +5C.

Bill Illis

To Steve Hempell.
Regarding the water vapour study by Dessler – I read the paper and the results are not exactly as indicated in the news releases. The study examined the change in water vapour from DJF 2007 to DJF of 2008 when the La Nina (and the AMO) reduced temperatures by 0.4C.
The study found there was a 1.5% (percentage points) decline in relative humidity in the very lower levels of the troposphere and a 1.5% increase in relative humidity in the upper layers of the troposphere. The middle layers were constant.
Given there is more water vapour in the lower levels of the atmosphere, the study really found that there was a decline in overall relative humidity when global warming theory suggests it should stay more-or-less stable.
To be fair, the models do produce results which are similar to this as temperatures go up (but not when they go down as happened between 2007 and 2008.)

Bill Illis

To davidsmith,
I think there is room for further optimization of this model, especially with the lags and trying other indices. I’ve seen your stuff before and would welcome any further thoughts.
To be honest, I built this model because I got tired of asking people to just try this or try that and then not seeing it done. I am just a layman and others need to pick this up and run with it now.

It makes me uncomfortable to look at the graph Hadley Plus Constant Versus Nino and AMO Model Only and see a hybrid of actual data and modeled data going back in time. Plotting anything that comes out of a computer on a chronologic baseline from the past is an inherently unsettling strategy, one that to my eyes appears to have been heavily influenced by AGWers’ love of GCMs.
The word “warming” on the graph, again, appears to have the ring of authority, as though there were only a single possible explanation for the divergence. That would appear to my eyes to be an argument, rather than a fact. Could the length of solar cycles have anything to do with temperatures? Could the intensity of solar cycles have anything to do with the rising temperatures? Could the PDO be at play here? Could, as someone else pointed asked, energy in the deep ocean that got there hundreds, or thousands, of years ago have raised atmospheric temperatures in the 20th century?


I’d be curious to know if the AMO has peaked on its latest cycle, and if so, what happens from here through the next 30 years? A repeat of of 1945-1975 – ala stagnant or a minor drop in temperatures – like the Western WA Professor’s paper as of recent?
.. …. …… .. 2038(?)
…… … .. …..

Mike C

Okay, I see partions of two oceans discussed here… many more to go.


Steve Hempell and Pamela
Re water vapour questions, it seems to me that with this kind of analysis, it is irrelevant. Temp increases are logarithmically tied to CO2, and H20 is essentially logarithmically tied to temperature. So a log dependence on CO2 will take care of the H2O in this type of correlation. Only the exponent of the log changes, ie the constant in front of the log term becomes a proxy for all the other parameters that change with CO2. In essence, it is perfectly acceptable to use the CO2 level as the proxy for most of the other variables that are tied to it. This is a nice piece of work.


I wish I could remember who said that the climate is the continuation of the oceans by other means.


The assumption that for time scales longer than el-nino and multidecadal oscillations the climate naturally is essentially static forms an essential, but unjustified, foundation of the AGW hypothesis. It is not possible to prove unambiguously that the warming from 1970 to 2000 was not part of some natural variation. Hence the defense of the undefensible concerning the hockey stick charade.

steven mosher

kim: Von Cloudswitz


To me, it really looks like the AMO/nino model diverges compared to observations since the 70ies, and that the difference keeps increasing with time.

Bill Ellis: You wrote, “one just needs to examine the up and down of temperatures in the 1997-98 El Nino, for example, to see there is no accumulating impact.”
Take a closer look. Note the step change in the pre-and post-1997 global temperature trends that should be attributable to the 97/98 El Nino in the:
GISS Data:
NCDC Data:
and the HADCRUT Data:
Or looking at data sets of smaller areas, notice the remarkable step changes in the Mediterranean Sea SST after the 97/98 El Nino:
and the Gulf of Mexico SST:
and the Atlantic Ocean SST:
and the western Pacific Ocean SST:
Arctic temperatures shifted, too:
Note the differences in response of the east and west Pacific Ocean SSTs (divided at 180deg longitude) to the 97/98 El Nino:
Note also the differences in the responses of the Indian Ocean to significant El Nino events:
versus La Nina events:
I discussed the above in the following posts:
Bill: I don’t have the background that would allow me to delve into the data further and pull out an accumulating impact of ENSO. But maybe someone else reading this thread does. The fact that a running total of the Trenberth NINO3.4 SST anomaly data mimics the global temperature anomaly curve hints that there is an accumulation.

Stephen Wilde

A good start in unravelling the oceanic effects on atmospheric temperatures.
Now we need to ascertain the net global effect at any particular time of ALL oceanic oscillations combined.
Sometimes they combine in the same phase to affect temperatures rapidly and at other times they offset one another.
Then tie them in with solar changes over several solar cycles and that should account for all observed temperature changes without having to involve CO2 at all.
See my various articles at :

Stephen Wilde

The oceans should be regarded as a continuation of the atmosphere as regards maintenance of global temperature and being so much more substantial the oceans are by far the greater part of the mechanism.

Bill Illis: Sorry about misspelling your last name. It’s early here. I was apparently more concerned that I had the right links.

kim (23:30:28) : I wish I could remember who said…
“Climate is the continuation of the oceans by other means”
That’s one link, kim. It seems to be all over the web.

Alan Chappell

Bill Illis says,
“I am just a layman” I ask Bill, where can i get a degree in Layman ?

Richard S Courtney

Bill Illis:
Thankyou for this cogent analysis. I have one comment on your method and its effect on your conclusion.
I understand your article to say your analytical method has the following steps.
The effect on temperature of AMO and ENSO within the time series is calculated by simple regression (this is possible because AMO and ENSO exhibit several cycles within the temporal range of the data set).
The temperature effect of AMO and ENSO is deleted from the time series to reveal a residual temperature trend in the time series.
The residual trend is assumed to be an effect of changed atmospheric carbon dioxide concentration over the temporal range of the data set.
The assumption in step 3 is used to calculate the climate sensitivity to changing atmospheric carbon dioxide concentration.
This may be correct, but the assumption in step 3 is the logical fallacy of ‘argument from ignorance’. The assumption amounts to, “The cause of the residual trend is not known so it must be changing atmospheric carbon dioxide concentration”. (If this ‘logical fallacy’ is not clear then consider, “The cause of crop failures is not known so it must be witches”.)
Of course, the residual trend may be a result of changing atmospheric carbon dioxide concentration.
However, the assumption in step 3 does not concur with the implicit assumption of steps 1 and 2 that natural cycles are affecting the temperature trend.
Other natural cycles may also be affecting the trend, and the method is not applicable to cycles with lower frequency than the time series. Such a very low frequency oscillation does seem to exist. There is an apparent ~900 year oscillation that caused the Roman Warm Period (RWP), then the Dark Age Cool Period (DACP), then the Medieval Warm Period (MWP), then the Little Ice Age (LIA), and the present warm period (PWP).
There is no known cause of this apparent low frequency oscillation: some people suggest it could be solar influence, but it could be the chaotic climate system seeking its attractor(s), and it could be … . However, there is no known cause of the AMO and ENSO, either.
Therefore, the implicit assumption of your steps 1 and 2 suggests that the residual trend determined by your steps 1 and 2 could be recovery from the LIA that is similar to the recovery from the DACP to the MWP.
Indeed, since the method adopted the implicit assumption of your steps 1 and 2, consistency suggests that all the observed rise of global temperature in the twentieth century is recovery from the LIA that is similar to the recovery from the DACP to the MWP.
Hence, the calculated climate sensitivity to changing atmospheric carbon dioxide concentration obtained by your method should be assumed to be a maximum value until this possibility of recovery from the LIA is assessed.
I hope these thoughts are helpful.
Again, thankyou for your superb work that I trust will soon be published.

Bill Illis

To Richard,
The steps you outlined are correct and there is a step 3 where there may be opportunities to find other variables/indices to explain some of the variation.
There is a fairly consistent trend going up however and the biggest explanation for that would be increasing GHGs.
But you’re right, other variables should be tested in this model.

John W

I also am having problems opening the excel files. Excel tries to repair (unsucessfully) the annual model, and I get a “cannot be accessed” error on the monthly.
Excellent analysis. I look forward to delving deeper over the holiday.


Arnd Bernaerts, yes, it was, and thank you, Roger.

Norm Kalmanovitch

No matter what scientific facts are presented to challenge the AGW ideology it is impossible for scientists to sway public opinion on this issue because the issue is political. It is very easy for high profile people who quote a scientific consensus that is supported by sophisticated computer models to convince the general public of anything that they want.
Even though the computer models have never yielded a single result that matches observations, any criticism of the models is met with some sort of complex justification that is beyond the comprehension of the general public so it is readily accepted by the masses and those questioning the validity of the models are vilified by the promoters of the AGW agenda as skeptics and deniers who are in the pockets of big oil.
The sole support for AGW is the climate models, and the sole support for the climate models with respect to CO2 is the forcing parameter. There is no actual physical rational for the forcing parameter, because it was simply contrived from the assumption that observed warming of 0.6°C was due entirely to a 100ppmv increase in atmospheric CO2 concentration. There was never any verification of this parameter either by theory or observation. There is no justification for this parameter based on the physical properties of CO2, because the molecular configuration of the CO2 molecule precludes any significant effect from CO2 beyond a concentration of 300ppmv, and the current concentration is 386ppmv.
There is no justification for this parameter based on observation because the observed notch in the spectrum created by CO2 is virtually identical for both the Earth and Mars and Mars has over 9 times the physical concentration of CO2 in its atmosphere than the Earth has in its atmosphere.
Even the reference temperature value for the parameter is faulty because the maximum temperature increase possibly attributable to human CO2 emissions is 0.1°C per century; not the 0.6°C that is used in the forcing parameter.
The climate models use a forcing parameter based on the equation:
CO2 rf = f * ln([CO2]/[CO2]prein)/ln(2)
where f= rf for CO2 doubling
In further documentation according to the IPCC, the “Radiative Forcing” ÄF, in watts per square meter, due to additional carbon dioxide in the atmosphere, can be calculated from the formula:
ÄF = 5.35 ln C/Co
The value 5.35 in this equation and the term [CO2]prein in the generalized equation demonstrate that the forcing parameter is based on the 100ppmv increase from the preindustrial value of 280ppmv and the 0.6°C of measured temperature over the time period that this 100ppmv increase occurred.
Further documentation in the IPCC reports states that the forcing of each watt/m2 raises the global temperature by 0.75°C + 0.25°C.
The Nimbus 4 satellite measured the thermal radiation spectrum of the Earth in 1970, when the CO2 concentration was 325ppmv as measured at Mauna Loa.
Mars has an atmosphere that is 95% CO2 with virtually zero water vapour and the remaining 5% of the atmosphere is comprised of O2 N2 and Ar, so CO2 is essentially the only “greenhouse gas”.
The atmosphere on Mars is so thin that the 950,000ppmv concentration of CO2 only represents about 9 times more actual CO2 than is in the Earth’s atmosphere in absolute terms.
Recent measurements of the thermal radiation spectrum from Mars should show a spectral notch from CO2 that represents an increase in forcing representing the 9 times difference in CO2 according to the equation:
ÄF = 5.35 ln C/Co
Considering that this formula gives a forcing value of 3.708watts/m2 for just a doubling of CO2, this value of 11.755watts/m2 for a 9-fold difference should be readily visible on the two measured spectra.
The spectral notch is virtually identical on both the 1970 Earth spectra with a 325ppmv and the Mars spectra from at least 9 times the concentration indicating that there is virtually no effect increases in CO2 beyond 325ppmv.
This clearly falsifies the equation and the numerical values used to determine the forcing parameter of the climate models that support the AGW hypothesis.
In addition to this physical evidence of an invalid assumption forming the basis for the forcing parameter, there is a blatantly obvious error in the actual values used in determining the magnitude of the forcing parameter. The temperature record shows that the global temperature has been increasing naturally at a rate of about 0.5°C/century since the Little Ice Age. The forcing parameter is based on the full measured 0.6°C/century without subtracting the natural warming of 0.5°C/century giving a forcing parameter that is 6 times larger than can be attributed to the measured increase in CO2.
Far less obvious, but the major fatal flaw of the forcing parameter is that it is based on an observation of temperature and CO2 concentration without taking into account the actual physical properties of CO2 and its limited effect on thermal radiation as defined by quantum physics.
As you are aware, certain gases can be caused to rotate and vibrate by thermal radiation. The rotation mode is relatively independent of wavelength but the vibration mode is limited to specific resonant wavelength bands. The rotation mode results from the interaction between the thermal energy and the dipole moment of the gas molecule. The carbon dioxide molecule is formed from two oxygen atoms equidistant from a central carbon atom and all three atoms are in a perfectly straight line. This configuration and symmetry eliminates any dipole moment, limiting the CO2 molecule to vibration modes only.
There is only a single vibration mode of CO2 that resonates within the thermal spectrum radiated by the Earth (and Mars). This bend vibration resonates with a band of energy centred on a wavelength of 14.77microns (wavenumber 677cm-1) and the width of this band is quite narrow as depicted on the spectra from Earth and Mars.
It only takes a minute amount of CO2 to fully “capture” the energy at the resonant wavelength, and additional CO2 progressively captures energy that is further and further from the peak wavelength. At the 280ppmv CO2 preindustrial level used as reference in the forcing parameter, about 95% of the energy bandwidth that could possibly be captured by CO2 has already been captured. There is only 5% of this limited energy available within the confines of this potential “capture” band left to be captured.
The greenhouse effect from CO2 is generally stated as 3°C, so an additional 100ppmv above the 280ppmv level is only capable of generating a maximum 5% increase or 0.15°C. The forcing parameter is based on a full 0.6°C which is four times the 0.15°C absolute physical limit of warming from CO2.
Furthermore if this 0.15°C increase has used up the full 5% of the remaining possible energy as the concentration reached 380ppmv, there is zero warming possible from further increases in CO2.
This is why the CO2 notch is virtually identical in the two spectra; the CO2 band was virtually saturated at the 325ppmv concentration level, so even nine times more CO2 has almost no appreciable effect.
Norm K.

John W

Richard C
You make good points. However, for your ‘logical fallacy’ example of the cause of crop failures to be analgous, there would need to be a corresponding increase in the population of witches.
And while we wouldn’t expect crop failures to “force” an increase of witches,
does the same hold true as to whether >CO2 is a cause or effect? After reading Jim Hanson’s paper where he tries to explain away the 700 or so yr. lag of CO2 following temperature changes in the Vostok data, I remain unconvinced that CO2 is, in the words of someone who’s name I can’t recall, “driving the bus or just sitting in the back”.

Bill Illis: Sometimes a change of perspective is needed. Your analysis also doesn’t account for the disparity between the magnitude and frequency of El Nino events and those of La Nina events. This can readily be seen by smoothing NINO3.4 data. I used a 7-year filter for the following graph. (Actually an 85-month filter for the monthly NINO3.4 and Global Temperature anomaly data.)
Other notes about the graph: Since the source of your data (Trenberth and Stepaniak) remark in the accompanying paper that the NINO3.4 data is questionable prior to the opening of the Panama Canal, 1914, I deleted the data before 1915. I also prepared this graph for an upcoming post that I haven’t gotten around to writing up, which is why it ends in 2005.
Note that the NINO3.4 data is predominantly positive from 1918 to 1944 and from 1977 to 2005 (periods when global temperatures rose) and that the NINO3.4 data is predominantly negative from 1943 to 1959 and from 1970 to 1977 (periods when global temperatures for the most part declined). There are some exceptions, but, in whole, it holds true. During the positive NINO3.4 period of 1959 to 1970, global temperatures started to rise but were suppressed by volcanic aerosols.


Thank you for that great post…I’ve read it 3 times, and will read it several more, I’m sure. Each time I understand a little bit more of what you’re saying.
Bill Illis,
What great work. I hope I get to see it put to good use, and provide yet another set of points that get added to the debate which we so desperately need.

Richard S Courtney

John W:
I hope this posting will clarify what I meant by my comment, and I apologise if my use of an imperfect illustration caused confusion.
To begin, I want it to be very clear that I think Illis has provided a good, useful and important analysis which warrants publication.
My witches illustration was intended to aid understanding, and it was not intended as direct analogy. However, your comment brings attention to quality of data. (The records show that the number of detected witches did increase at the time when witchfinders were appointed. But it is not clear how many witches were detected and how many witchfinders existed.)
I stress that I think Illis has provided a superb analysis, but the quality of any analysed data should always be questioned because GIGO applies to all analyses.
Also, your comment concerning “driving the bus” illustrates the importance of ascribed causality. Which was causal; did increase to the number of detected witches induce increase to the number of witchfinders, or did increase to the number of witchfinders induce increase to the number of detected witches?
In the context of the analysis Illis provides, ascribed causality has great importance. Which is causal; did increase to atmospheric carbon dioxide concentration cause increase to the residual temperature trend, or did rising temperature cause increase to the atmospheric carbon dioxide concentration, or were the temperature and carbon dioxide changes caused by some other effect(s)?
My comment was intended to explain the importance of ascribed causality on the analysis Illis provides.
I am an extreme sceptic on matters of man-made global climate change. I do not know the cause(s) of the recent rise in atmospheric carbon dioxide concentration, and I do not know what – if any – effect that rise is having on global climate. But I want to find out.
The absence of any empirical evidence for anthropogenic (i.e. man-made) global warming (AGW) leads advocates of AGW to rely on the logical fallacy of ‘argument from ignorance’ and outputs of computer models.
History is replete with examples of politicians being guided by advisors who used the ‘argument from ignorance’ fallacy to justify their advice. The advisors have always presented an appealing case based on accepted theory, and they have always ignored – or rejected – alternative possible explanations for the effects which they have asserted as justification for their advice.
In ancient times such advisors said, “We do not know what causes lightening to strike so it must be the actions of Gods and people should make sacrifices to appease those Gods.” And, as my illustration said, in the Middle Ages such advisors said, “We do not know what causes crops to fail so it must be witches and we must eliminate the witches.”
Now, advocates of AGW say, “We do not know what causes global climate change so it must be emissions from human activity and we must eliminate those emissions.” Of course, they phrase it slightly differently: they say that they cannot match historical climate change with known climate mechanisms unless an anthropogenic effect is included. But this “anthropogenic effect” is an assumption with no more empirical evidence to support its existence than the empirical evidence for ancient Gods and witches.
My comments tried to say that the final part of Illis’s analysis adopts the same ‘logical fallacy’ that is used by AGW advocates. It is an assumption – not a fact – that increased atmospheric carbon dioxide is the cause of the residual temperature trend his analysis reveals. Indeed, his demonstration that natural oscillations cause some of the temperature trend adds credence to the possibility that other observed natural oscillations may also be significantly contributing to the trend.
But, as I also said, increased atmospheric carbon dioxide may be the cause of the residual temperature trend Illis’s analysis reveals.
I hope what I intended to say is now clear.

John M

davidsmith1 (20:37:23) :
Will this or this help you track it down?


two rapid comments on the reply just above:
“the computer models have never yielded a single result that matches observations”
This looks to me like a very strong statement. Too bad it is completely false… In particular when you think many models biuld their internal set parameters based on reproducing observations over the last century.
“There is only a single vibration mode of CO2 that resonates within the thermal spectrum radiated by the Earth. This bend vibration resonates with a band of energy centred on a wavelength of 14.77microns (wavenumber 677cm-1)”
again, a very convincing argument, except that it is obviously incorrect:
There is more than just one asymetric vibration mode in CO2.

Noblesse Oblige

This analysis is very promising. It is similar to but extends the work of Douglass and Christy (Limits on CO2 Forcing From Recent Temperature Data of Earth, Energy and Environment, to be published) which considers only UAH data (1979+). Both your analysis and Douglass/Christy attempt to remove “unforced” natural effects from the climate signal and ascribe the residual to “the real global warming signal.” However, solar variability remains an untreated “forced” natural effect. You may wish to consider incorporating the work of N. Scafetta and B. West (J. Geophysical Research, Vol 112 D24S03, Nov. 2007, and their previous papers cited therein) who treat solar variability phenomenolgically via a simple thermodynamic model using various reconstructions of Total Solar Illuminance and the historical temperature over longer time periods. They find a major fraction, but not nearly all, of 20th century warminng ascribable to solar variability. It should be possible to incorporate their approach in your regressions.
The point is that no one that I know of has included both “unforced’ terrestrial cycles and solar variability in an integrated analysis.

Pierre Gosselin

Now that’s a good post. Keep ’em coming!
0.4°C more warming with a doubling of CO2 – hardly catastrophic. Certainly does not warrant expensive mitigation programs demanded by many.
“…adjustments of old temperature records by GISS and the Hadley Centre and others have artificially increased the temperature trend… ”
Now why doesn’t that surprise me? I’m curious how the other side will respond to this.

Bill Illis

To Bob Tisdale,
I tried just plugging a higher coefficient for the Nino 3.4 region because it does seem like there are periods when its impact is greater than the reconstruction allows.
However, there are many other time periods when the increased coefficient just puts the reconstruction far off the actual temperature trend. These time periods then extend out over many years, decades even, versus the very short periods when the reconstruction is off by a small amount.
So, I just decided to trust the regression and go with it. I wanted this to just be a straight-up, simple model with no plugging or smoothing in any event because there is a danger in playing around with the data too much, as Mann’s hockey sticks show.
You are the expert on analyzing ocean temps and circulation of course and I have been to your site quite often before.
In terms of the Trenberth data being questionable in earlier periods, I just decided to just go with what’s available. If we are to build a model of temperatures, we have to use what is available.
What I would like to see however, is if the raw Trenberth data would provide a better fit as this index is a five month smoothing. It is probably too variable, hence the need for the smoothing but what I’ve seen in this building this reconstruction is that information is lost as the data is smoothed or averaged. Just look at the spikes in the temperature data, the climate can move very fast.

I just returned from a climate change conference in Amsterdam organised by the Royal Geological and Mining Society of the Netherlands on the 20th of November
Two speakers: Prof Jurg Beer (physicist) of Zurich and Prof Kees de Jager of Utrecht University and founder and first director of the Utrecht Space Research Laboratory showed some remarkable correlations between temperature variations and solar activity over the last 400 years or so. They concluded that no anthropogenic signature could be detected. They purely concentrated on the statistical significance of the observations. Yet they realized that the changes in solar forcing are not enough to cause the temperature fluctuations, but they didn’t want to speculate about the possible causes. Obviously some other mechanism must explain the amplified effect on temperature. They knew about Svensmark et al but didn’t want to comment on it as they felt not qualified to do so, but they were eagerly awaiting the outcome of the forthcoming CERN experiments.
If ENSO and AMO are also correlating with the past temperature fluctuations, I can only conclude that the sun ( via clouds? ) is responsible for the frequency and strength of the various oceanic oscillations.
De Jager’s paper has just appeared in the September issue (Volume 87, no 3) of the Netherlands Journal of Geosciences.