Barmy bloopers from John Christy and co at WUWT seven years later

Another Oh My! article has hit the denier-waves. This time from John Christy of UAH infamy, and friends (WUWT article with link is archived here, update here). It purported to be about green houses that are gassy, or something like that. The first sentence in the preface is:

On December 15, 2009, EPA issued its Green House Gas (GHG) Endangerment Finding, which has driven very significant and costly regulations beginning with CO2.  

Seriously. These guys claim to be writing about greenhouse gases and they don't even know how to spell greenhouse?

Warning: this is another long article.

It's not warming? Huh?

I've noticed lately at WUWT and at Judith Curry's place, the latest trend is to claim it's not warming. This seems particularly foolish right now, after the four hottest decades on record, the two hottest years on record (and heading for a third) and eleven consecutive hottest months on record.

So I was only mildly surprised to find that John Christy has co-authored an article for some disinformation purpose. His co-authors are, wait for it, Joe D'Aleo (would you believe) and James P Wallace III. Joe is in cahoots with Anthony Watts on his dormant or defunct Open Atmospheric Society. This trio wrote an article protesting the EPA's seven year old Endangerment and Cause or Contribute Findings for Greenhouse Gases under the Clean Air Act. Yep. Seven years it's been out, and it's taken this mob all that time to come up with an angle they think might fly. Or might fly with the scientific illiterati at any rate.

Pardon, your age is showing...

The first thing you'll notice is the quaint use of capital letters spattered throughout the document. Take the second paragraph of the preface. I'm surprised they didn't capitalise "dangerously":

Focusing primarily on the time period since 1950, EPA’s Endangerment Finding predicated on Three Lines of Evidence, claims that Higher CO2 Emissions have led to dangerously higher Global Average Surface Temperatures (GAST).

A précis - it's not warming, it's the sun and it's El Niño

What they claim is:

• it's not really warmed, or not significantly.
• it's the sun (based on cumulative TSI - what the heck did they do there?)
• it's El Nino (based on cumulative MEI - what the heck did they do there?).

I mean, seriously? Is John Christy about to retire? Is he looking to let go of his contracts with NOAA/NASA? Why didn't Roy Spencer co-author this? Was it too much to stomach even for him?

"Cumulative Total Solar Intensity Anomaly" and more gems

Some denier gems = I'll let you try to translate the gobbledegook:

• periods of increasing Cumulative Total Solar Intensity Anomaly (Cum TSI Anomaly) would lead to time periods involving more intense and more frequent El Ninos and vice versa
• The results, which are obviously quite robust, make it very clear that the Cum MEI variable will capture much of the solar trend cycle influence on temperature trend slopes.
• the statistically significant linear trend component of the Cum MEI Trend Cycle over this period is due to the cumulative impact of the 1977 Pacific Shift and it is totally appropriate for ENSO Adjustment to take out its linear trend impact.
• The 1977 Shift variable is also always used when the data begins prior to that date. 
• the Cumulative ENSO activity, has been quite similar to that of Cum Solar Activity
• the positive linear trend component in the Solar Activity’s trend cycle over this period

It hasn't warmed since 1950 - ahem (if you subtract the warming)

Looking at their argument that it hasn't warmed, or not really, the way they figured this is as follows. (They sure took their sweet time about all this by the way. The document runs to 68 pages of nonsense.)

Games with the Multivariate ENSO Index (MEI)

For ENSO they did something quite unique. The developed what they called the "cumulative MEI". Now the Multivariate ENSO Index (MEI) is an index that's used to indicate ENSO events. From NOAA: the MEI is calculated as the first unrotated Principal Component (PC) of six observed fields (combined) in the tropical Pacific. These are: sea-level pressure (P), zonal (U) and meridional (V) components of the surface wind, sea surface temperature (S), surface air temperature (A), and total cloudiness fraction of the sky (C). There's more.

All values are normalized for each bimonthly season so that the 44 values from 1950 to 1993 have an average of zero and a standard deviation of "1".

How or why anyone would distort the MEI into a cumulative chart is anyone's guess. But these chaps did. Not only did they do that, but the MEI is provided as a bi-monthly number. For each year it goes: Jan-Feb, Feb-Mar, Mar-April and so on to Nov-Dec. It took me a while to figure out what these authors did. I tried accumulating the values but ended up with values ranging from 64 to -110. That was an order of magnitude different to what these authors came up with. Turns out it was simple in the end.

1. Take an annual average (why on earth?)
2. Add the second year's data to the first
3. For each subsequent year, add the current year to the sum of all prior years.

This is the actual MEI going from 1950 to the present, using data from NOAA. Positive values are El Niños and negative values are La Niñas.:

Figure 1 | Multivariate ENSO Index. Data source: NOAA

Now compare that to what happens when you try to average the MEI by year. Click the chart to enlarge it:

Figure 2 | MEI and MEI averaged annually. Data source: NOAA

• The highs and lows disappear into a smudgy soup. In the averaged plot, the peaks and troughs are pretty well the same height. 
• Some peaks and troughs all but disappear altogether.

Part of the reason is that by averaging over a year, you're losing the peaks and troughs. I've kept the scale the same, and you can see how different the plots are. The annual average never reaches the highs and lows of the monthly.

Another problem is that some years there is both a La Nina and an El Nino, so they cancel each other out. For example, 1997-98 was a massive El Nino followed immediately by a La Nina from 1998 to 2001. In the averaged chart, the 1998 El Nino looks no different to the ones before it.

Similarly, El Ninos typically span two years, from April-May in the first year to April-May in the second year. Therefore taking a calendar year average mushes the El Nino months with the ENSO neutral months (or La Nina months).

Now look what happens when you accumulate the annual averages:

Figure 3 | The Christy and co "cumulative MEI". This seems to be calculated by first averaging all the MEI values over each calendar year, then accumulating the annual numbers.

To my way of thinking the above chart is somewhere between useless and very useless. However John Christy and his co-authors found a use for it. Guess what they did next. They appear to have deducted some calculation based on the above from the average global surface temperature. It's not at all clear to me just what they subtracted, except that they did remove the increase in temperature, using their scrambled, weird abomination that they called the "cumulative MEI".

What? I hear you gasp. Yep. This is what they wrote:

There is one additional very important point that needs to be made here. From Figure VII-2 above, it is clear that the Cum MEI variable has a Trend Cycle pattern the Trend component of which will have a statistically significant positive slope if the linear regression on a time variable is run over the entire data window 1959 -2015.

By the way, there wasn't any Figure VII-2 above.  This was from their section VI, so I figure they meant Figure VI-2, which showed their annually averaged then accumulated MEI. They continue, explaining why  they have to remove the rise in temperatures in the 1970s. It was because the temperature went up. (Shades of John McLean, eh?)

The fundamental question addressed now is whether or not it is appropriate that the methodology, used to remove ENSO impacts in doing so, removes this linear Trend impact. The answer is that it is totally appropriate, in fact a must, because the linear trend impact in Cum MEI results from the Natural ENSO related 1977 Pacific Shift.

As shown above, the statistically significant linear trend component of the Cum MEI Trend Cycle over this period is due to the cumulative impact of the 1977 Pacific Shift and it is totally appropriate for ENSO Adjustment to take out its linear trend impact. In fact, it would be totally inappropriate to not take its impact out. However, the CO2 explanatory variable is statistically indistinguishable from a (straight line) linear trend. So, a key question addressed in this research remains whether or not MEI Adjusted Temperatures have a statistically significant positive Trend slope that might be attributed to CO2.

Games with solar radiation

As you probably know, solar radiation has been declining since the middle of last century. The current solar cycle is the weakest it's been for ages. Despite this, this pack of ratbags did the same thing with TSI that they did with MEI. They developed what they called the "cumulative TSI".  They put up the chart below:

Figure 4 | Abomination of a chart from Christy and co. Source: Their abomination of a document.

Here is the estimated and observed TSI according to resident WUWT solar expert, Leif Svalgaard. I've added a blue box to mark the period since the 1950, and an arrow to show that it's been declining:

Figure 5 | TSI estimated and observed. Source: Leif Svalgaard

Is there any way that this recent decline in TSI could cause the world to heat up? Well, according to Christy and co, the solar radiation accumulates. (Why it didn't accumulate or cause global warming from the 1700s to the early 20th century, they don't explain.) What they did say was this:

It is not surprising that periods of increasing Cumulative Total Solar Intensity Anomaly (Cum TSI Anomaly) would lead to time periods involving more intense and more frequent El Ninos and vice versa. Thus, inclusion of the natural Cum MEI variable in the ENSO adjustment modeling process can be expected to capture such cumulative solar impacts on ENSO behavioral patterns.

You scratch your head and wonder how a declining TSI can cause warming. And I say - umm, good question.

Honestly, these chaps are nuts. They wrote how TSI went up and down with MEI, and then made pronouncements about the sun causing the warming. They'll convince no-one except the denialati.

I haven't even got to their "Tropical Hot Spot". Read on if you dare.

Random number cumulative

Here's something that John Christy and Joe D'Aleo and James P Wallace III could try. Instead of messing about with cumulative MEIs and TSIs, why not just use a random number generator? Below are ten plots that I generated successively (no cheating) with numbers that ended up between +3.3 and -2.3. I then performed the same calculations as they did to get the cumulative distribution.

If you've got a charting package and a random number generator you can try this for yourself. (I just used MS Excel.)

Added by Sou at 5:15 pm Fri 23 September 2016 AEST

On the warming rate in the upper troposphere - denier's so-called "tropical hot spot"

The article or whatever you want to call it, started out badly. The authors wrote how their "evidence" invalidates the EPA's 2009 Greenhouse Gases Endangerment Finding:

These analysis results would appear to leave very, very little doubt but that EPA’s claim of a Tropical Hot Spot (THS), caused by rising atmospheric CO2 levels, simply does not exist in the real world. 

But where was the EPA's claim? I looked and couldn't find it.

Much of their article was about what they called a "tropical hot spot", which is not hot and not a spot. In the EPA Endangerment finding itself, the word "spot" appears not at all. The word "hot" appears five times, associated with days, nights and temperatures, but no "hot spots". The word "tropical" appears only once, in association with tropical cyclones, not hot and not spots. 

This isn't surprising because the term "tropical hot spot" is not a term that climate scientists typically use. They tend to refer to amplification of the rate of warming in the upper troposphere. In the EPA Finding, the words troposphere or tropospheric appeared ten times, but not in relation to any amplified rate of warming aloft.

In other words, the EPA didn't base their finding on any misnamed "tropical hot spot". If they had they would have said so in the finding.

I then went to the technical support document, and I could find no mention that could be related to amplified warming rate in the upper troposphere. You can look for yourself and tell me where or if I missed it. (It wasn't in the Finding document or the tech support document - Michael S found it in their response to public comments. So not any basis for the EPA finding, just a response to some denier or other.)

Remove the warming trend and say there's no warming

I won't go into all their tropical troposphere charts. They had lots and lots of them. They also had some surface temperature charts. These so-called scientists did a John McLean again. They took a whole lot of data sets from the troposphere and the surface, removed the warming and then said there wasn't any warming. How clever is that?

They justified it by claiming they were removing the effects of ENSO events and solar radiation. They weren't. They used their "cumulative MEI" and "cumulative TSI" and took out the "1978 Pacific Shift". Those tortuous meaningless fudges were apparently the only way they could get rid of this warming. Here is the chart of the average of 12 months to August each year. 

Figure 6 | Global mean surface temperature anomaly for the 12 months to August each year. The base period is 1951-1980. Data source: GISS NASA

You can see why they must have puzzled for seven whole years to try to come up with a way to say there's no warming.  In the end their effort was a pathetic failure.

By the way, if you want to see the real impact of removing ENSO from the temperature chart, Tamino will show you - here and here. Below is his chart for GISTemp, with changes attributed to ENSO and solar and volcanic eruptions removed:

Figure 7 | Global mean surface temperature minus effects of changes in solar, ENSO and volcanic eruptions. Source: Tamino

From the WUWT comments

I don't need to tell you that most of Anthony's readers will lap up anything they are fed, provided it's nonsense. Here's a sample:

George Tetley sets the ball rolling, and behaves like an idiot in and out who can't add 2+2.

September 22, 2016 at 10:56 am
Now it is to late to teach Obama to read, but there must be someone in the global political world who can add 2+2 without coming up with 8.
idiots in idiots out !

Tom Halla likes the nonsense and doesn't realise that Christy and co didn't adjust for ENSO. They adjusted to remove the warming using denier mathturbation:

September 22, 2016 at 11:13 am
So once ENSO is compensated for, there is no tropical hotspot. I am fairly sure that the true believers will argue the compensation is flawed, as the hotspot must be there according to the models, and the models are never wrong.

 Anthony Watts, who wouldn't know pseudo-science from the real thing, according to his good friend Willis Eschenbach (and all the evidence on his blog), must have guessed that the analysis was shonky because he said he was anticipating a comment from Nick Stokes.

September 22, 2016 at 11:25 am
“…true believers will argue the compensation is flawed…” Cue Nick Stokes in 3, 2, 1…

Greg seems to regard this woeful document as "serious analysis".

September 22, 2016 at 11:43 am
Isn’t this what Bob Tisdale suggested about 5 years ago?
Maybe now someone has done some more serious analysis, it may get looked at more seriously.
Lot’s like all these natural “oscillations” are not necessarily net zero effects after all. Maybe we’ll have to stop automatically ‘detrending’ any and all climate variables, attributing the trend to AGW and calling the rest an “oscillation”.

Tom Dayton points out that the so-called "tropical hotspot" isn't a fingerprint of greenhouse warming. It happens when there's any surface warming.

September 22, 2016 at 11:57 am
Michael Bastasch, please provide a quote or at least an actual, specific place in the EPA’s documentation to support your contention “When EPA released its CO2 endangerment finding in 2009, it used three lines of evidence to bolster its argument that greenhouse gases threatened human health through global warming.The crux of EPA’s argument rested on the existence of a ‘tropical hotspot’ where global warming would be most apparent. That is, there should be enhanced warming in the tropical troposphere — the “fingerprint” of global warming.”
In reality, neither climatologists in general nor the EPA have ever claimed that the tropospheric hotspot is a fingerprint of CO2-caused warming, which is what you strongly imply in your two sentences. Instead, a tropospheric hotspot is a consequence of the lapse rate, and is expected to increase if global warming occurs, regardless of the cause of that warming.

ristvan is another hard core denier who applauds almost every bit of nonsense at WUWT. Not only that but he defames quality scientists whenever the urge takes him. He's wrong in every respect of course:

September 22, 2016 at 12:59 pm
TD, you are diesrincorrect in several ways. Read the very general ‘three lines of EPA evidence’ posted in the replies to your comment. 1. General understandings. CO2 is a GHG, yes. But that says nothing about sensitivity or attributiin. Natural variation is overlooked, specifically the period 1920-1945 that is essentially indistinguishable from 1975-2000. The former cannot be caused by delta CO2, which means attribution of the latter to CO2 cannot be proven. 2. By comparison to paleo. That is provably false, just as Mann’s hockey stick and Marcotts mess are. 3. By models. Those models all produce a tropical troposphere hotspot that does not exist in reality, thereby showing the models cannot be relied upon.
This new research directly attacks 1 by showing there is no warming 1975-2000 after correcting for ENSO using NOAA’s official MEI. And it directly attacks 3 by showing there is no observable tropical troposphere hotspot. It does not need to attack 2 because McIntyre already has.
And all that is admissable evidence in a court proceeding attacking the EPA endangerment finding.

Michael S has found the missing "hotspot" reference. It wasn't in the Finding document, nor was it in the Technical Support document. It was in a "response to public comments". So, like I said up top, John Christy and Co were lying when they claimed it was one of "Three Lines of Evidence in its CO2 Endangerment Finding". (Note the capitalisation!) It wasn't, and nor should it have been. (Read the responses and guess what types submitted them. They are a hoot. The EPA was very polite, but must have been amazed and appalled at the level of ignorance of Americans who pretended to know something of climate.)

September 22, 2016 at 1:56 pm (excerpt)
Mr. Dayton,
The EPA specifically commented on questions about he “fingerprint” in justifying their decision in the endangerment finding. They used the flawed studies such as Allen/Sherwood (2008) using wind date to help the EPA ignore the inconvenient facts of the temperature data. They also used Santer (2008) to try to justify the differences between modeled and observational trends. To argue that climate alarmists and the EPA have not tried to save the “fingerprint” argument as a key element in “proving” AGW is either naive or a prevarication.
From the EPA own website:
https://www3.epa.gov/climatechange/endangerment/comments/volume3.html

The fake sceptic ristvan again, who'll regurgitate any lie he's fed about climate, if it happens to suit his agenda:

September 22, 2016 at 12:19 pm
Correct scientifically, but not legally. EPA endangerment finding is based on their conclusion that the tropical troposphere hot spot exists. To attack that legally required fact finding to declare CO2 a regulanle pollutant, you show it to be factually wrong. As this unassailable research does.

And this one's priceless. ristvan cites Jo "Force X and the Notch" Nova as "tearing to shreds" the paper by Steve Sherwood. She's torn it oodles of times, but she's incapable of tearing a scientific paper to shreds. (Neither was any Sherwood paper "shredded" by his peers. He is a much respected scientist who does terrific work.)

September 22, 2016 at 1:03 pm
Toneb, the second Sherwood paper you cite is pal review garbage. Go read it. Was Sherwood’s second peer reviewed effort to find the mythical hotspot. His first was shredded by his peers. His second is so methodologically bad that nobody has even bothered. And you don’t need an advanced degree in statistics to spot its flaws. Jo Nova did all by herself.

Tom Dayton is the first to touch on the issues I raised:

September 22, 2016 at 1:41 pm
What is their rationale for using *cumulative* TSI and *cumulative* MEI? 

I had to save the WUWT page again to get the contributions that Anthony Watts was aching for. Nick Stokes gives a lesson in basic statistics about the so-called "hotspot":

September 22, 2016 at 3:29 pm
That’s not what they say. They say positivity of trends are not statistically significant. That means roughly that there is a 5% chance (or more) they were not positive, consistent with the null hypothesis. But that still leaves a 95% chance that they were.

And a follow-on from Nick Stokes, pointing out any error could come from observations as much as anywhere else:

September 22, 2016 at 6:56 pm
“If we can’t get to a overwhelmingly high level of confidence, then all bets are essentially off.”
But that’s not the claim here. They claim “And, therefore, the THS does not exist.”. And they go on to say, therefore the GCMs are wrong, and so the EPA is FUBAR.
In fact the big doubt here is the measurement methods. The question of homogeneity of radiosondes is very real. And trying to sort out levels in satellites is very uncertain – just see the change in UAH from V5.6 to v6 (still in beta).

Nick Stokes elaborates further here.

September 22, 2016 at 2:38 pm
“Isn’t the proper statistical approach to this : what are the sufficient conditions for rejecting the hypothesis that there is no THS ?”

Indeed so. The paper has that elementary logic wrong. The muddle starts in the first sentence of the abstract:
“These analysis results would appear to leave very, very little doubt but that EPA’s claim of a Tropical Hot Spot (THS), caused by rising atmospheric CO2 levels, simply does not exist in the
real world.”

Firstly there’s the usual problem with sceptic “refutations” that they don’t quote what they are actually refuting. Not once in all those words do we learn what the EPA actually said about the hotspot.

But then, their basic claim is this:
“Section V. Sufficient Conditions for Rejection of
the THS Hypothesis
If, after adjusting for the natural ENSO impacts, all relevant temperature time series have linear trend slopes that are not positive and statistically significant; then rising CO2 emissions in combination with Non ENSO related solar and volcanic activity, cannot have had a statistically significant impact on the tropical temperature data trend slopes in the real world over the past 50 plus years. And, therefore, the THS does not exist.

And as a matter of elementary logiuc, that is just wrong. Statistical significance never proves anything. It may fail to reject the null hypothesis. That actually relies on a theoretical uncertainty model, which may be inappropriate. But at best, all you can deduce is that you didn’t have enough data. It’s a failure, not a proof.

The thing that Nick didn't comment on was the method used by Christy and co to adjust for ENSO and changes in solar forcing. It was warped and wrong and weird IMO.

References and further reading

Endangerment and Cause or Contribute Findings for Greenhouse Gases under Section 202(a) of the Clean Air Act - EPA website

Correcting for more than just el Niño - article by Tamino about removing ENSO, solar and volcanic etc


From the HotWhopper archives

• About that tropical "hot spot" - May 2015, with references
• How Roy Spencer and John Christy trick Anthony Watts and his deniers once again - October 2013
• Roy Spencer's latest deceit and deception - February 2014 with John Christy's deception
• Roy Spencer grows even wearier... - May 2014 - more about John Christy's deception
• Joseph D’Aleo Fails Meteorology 101 on WUWT - June 2013.