Green World Trust

[Derek]

Part 1


“We” do NOT measure CO2.


This is a plot that will be familiar to almost all people, it is the Mauna Loa record of CO2 levels,
as recorded on a volcanic island in the Pacific (very near to the world's most active and another very active volcanoe).
The constant, unerring rise due to (purportedly) man’s activities and our emitted CO2.



So, instantly it is fairly reasonable to assume that the Mauna Loa Observatory (MLO) measures the CO2 level in the atmosphere.
Here is a link to a page describing, in some detail, how and what the MLO does.
http://esrl.noaa.gov/gmd/ccgg/trends/
Trends in Atmospheric Carbon Dioxide - Mauna Loa
@ US Department of Commerce,
National Oceanic and Atmospheric Administration.

Here is a page linked from the above page, describing in more detail the method used.
http://esrl.noaa.gov/gmd/ccgg/about/co2_measurements.html
How we measure background CO2 levels on Mauna Loa.
Pieter Tans and Kirk Thoning, NOAA Earth System Researh Laboratory,
Boulder, Colorado
September, 2008

The Scripps Institution of Oceanography has something very similar as well.
http://cdiac.ornl.gov/trends/co2/sio-mlo.html
Atmospheric Carbon Dioxide Record from Mauna Loa
R.F. Keeling, S.C. Piper, A.F. Bollenbacher and J.S. Walker
Carbon Dioxide Research Group
Scripps Institution of Oceanography
University of California
La Jolla, California 92093-0444, U.S.A.

Here is a link to the Scripps Institution’s monthly mean record for MLO from 1958 to 2007
http://cdiac.ornl.gov/ftp/trends/co2/maunaloa.co2
Atmospheric CO2 values (ppmv) derived from in situ air
samples collected at Mauna Loa, Hawaii, USA
Carbon Dioxide Research Group
Scripps Institution of Oceanography (SIO)

I have plotted this monthly data on an excel sheet, and it shows a level of uniformity that is surprising.
There seems to be a very constant rate of change in just about every way, from month to month, year to year, and season to season.
Virtually whichever way these monthly means are plotted you get virtually straight lines.
Below I plotted the month to month rate of change between each month for each year from 1960 to 2007.
A single year due to seasonal influences forms an S shape, this plot has all 47 years plotted.



Forty seven years, in that time there are virtually no outliers to within 1 part per million of CO2.
Furthermore the most any month can seem to vary + or – is less than 0.75 of a part per million.
I find it hard to comprehend nature and man’s influence combined is, or could be that regular a rate of change.

This leaves one at a cross roads, do you accept the given monthly means “as is” or
do you delve a little deeper to check the data and the figures.
Well, I’ll try to save you a bit of time here, if you look further you can find daily averages, and
maybe also hourly averages for MLO by NOAA, and possibly SIO as well.
Sometimes these figures can be referred to as “raw hourly averages”, but
being averages they must have been processed, so they by their name can not be raw.
This is important, because only the raw data has not been altered according to
various assumptions and omissions used in processing the raw data.
Processing data, assumptions, and omissions are all perfectly normal procedure, as long as,
a) the raw data is still available, and
b) the assumptions / omissions are clearly stated and known.

This is so that other people at a later date can go back through the data to check
the data, the assumptions, the omissions,
and any other taken onto account factors are as they are purported to be.
Measuring a trace gas to such a high level of accuracy, that so much is built upon in recent times,
that has such political importance, and could be said to be ruling our lives in many respects,
you would of thought the MLO is both open and verifiable.

Asking further questions about raw data usually leads to, well, less than polite, or
just plainly (and seemingly deliberately) confusing answers.
Getting past the “raw hourly averages” is nigh on impossible, but sometimes you can.
I have seen two days of what are called “voltage” figures taken at 10 second intervals at MLO.
From a supposedly continuous 50 year record - that is not very open.

So, this raises a question,
1) What is the voltage figure produced by the Infra Red Gas Analysis (IRGA) method a measure of ?
http://esrl.noaa.gov/gmd/ccgg/about/co2_measurements.html
Excerpt,
“Infrared absorption.
How does the CO2 analyzer work? Air is slowly pumped through a small cylindrical cell with flat windows on both ends.
Infrared light is transmitted through one window, through the cell, through the second window, and is measured by
a detector that is sensitive to infrared radiation. In the atmosphere carbon dioxide absorbs infrared radiation,
contributing to warming of the earth surface. Also in the cell CO2 absorbs infrared light.
More CO2 in the cell causes more absorption, leaving less light to hit the detector.
We turn the detector signal, which is registered in volts, into a measure of the amount of
CO2 in the cell through extensive and automated (always ongoing) calibration procedures.”

In other words the IRGA method used measures the amount of absorbed infra red radiation by the sample.
ie, the overall effect of the greenhouse gases (GHGs) present in the sample.
Later in the description it is explained that water vapour has been removed from the sample before it is tested
by a cold trap. This is because water vapour is THE most important greenhouse gas, so
to measure CO2 levels H2O is first removed, by cooling or rather by freezing.
The then dry air is passed into the test chamber, from which the voltage reading is taken.

At this point it might be useful to look at a plot of the IR spectrums of several well known GHGs.
I have put the below plot together from spectral plots as published in a climate science journal.

From these I put together this “plot”.


From this plot the strong and wide spectral absorbtion by H2O is clearly visible.
So how does the “plot” look with water vapour removed, as the IRGA method does.



Immediately a problem is evident, depending upon amounts present “other” GHGs could be (ARE)
contributing to the measured IR absorbtion of the measured sample.
So, the IRGA method DOES NOT measure CO2.
This raises a second question,
2) Of the amount of IR absorbtion measured by the IRGA method, how much is due to CO2 ?
Obviously this question depends upon spectral response and amount of each GHG gas present in the (dry) sample,
and applies to Ozone, Methane, and Nitrous oxide.
Without knowing (or measuring) any or all of the actual amounts present of the four GHGs in the sample,
how can the amount of any one of the GHGs be determined in the first place.

It would be like giving a blind man a bucket full of coloured ping pong balls and asking him
how many red ping pong balls are in the bucket.
He would not have a clue.

Being the helpful type I am I have done another plot to illustrate this point.



Is this why there is no released raw CO2 data for MLO.
There is no raw data, because CO2 is not directly measured.

So how is there a 50 year record of CO2 measurements from MLO ?
I would guess assumptions of the amounts of the other GHGs present in the samples are
subtracted from the voltage measured in the processing that produces the “raw hourly averages”.
I hope I’m wrong, because if I am correct then there is no (directly measured) CO2 record.

It may well have been that back in the late 1950s and early 1960s that this new method was thought to be
a great improvement upon the wet chemical analysis methods then,
and still used now (but not allowed to verify MLO et al...).
If that was the case, then why was the method, “invented” by the Scripps institute made and
then quickly shipped to the middle of the Pacific ocean (MLO) whilst not allowing anyone else to verify it.
MLO and the other 19 stations (at least, it might be nearer 40 stations) in the official CO2 monitoring / measuring network
all use the same machinery and / or technique, to this day.
To the best of my knowledge the IRGA method has never been independently verified,
possibly the above plots might illustrate why.


To be continued.


Part 2
Continued..

So far I have tried to illustrate one or two concerns regarding the IRGA method used to
“measure” CO2 levels at MLO et al.
Let us not forget these official CO2 figures

are used to convince us that the global level is the same as these figures,
yet wet chemical analysis shows a completely different CO2 level and history by direct measurements.
http://www.biokurs.de/treibhaus/180CO2_supp.htm

So, on to more detailed concerns, or rather “smaller” concerns that may add up..

System status information, or quality control checks for MLO are not available on the internet.
From, http://esrl.noaa.gov/gmd/ccgg/about/co2_measurements.html
Excerpt,
“Auxiliary measured variables
System status variables measured and continuously recorded are:
analyzer temperature, cold trap temperature, room temperature, sample flow rate, and
both pressure and flow rate through both air intake lines from the tower. “

The cold water trap used to dry the air sample before testing is surely a critical measure or parameter.
This is because the solubility of CO2 in water is raised as water is cooled, so
the cooling water vapour WILL absorb CO2 out of the sample whilst it is being cooled.
The amount of water vapour in the sample prior to cooling has to be known so that
the amount of CO2 removed from the sample by “x” amount of cooling water vapour can be
added back on the voltage reading taken after drying the sample. As the amount of water vapour
in the fresh samples varies so does the correction factor required,
I have read elsewhere that this can easily amount to a 20ppm correction being required.
How water vapour varies over time has to be taken into account.
If the cold water trap alters temperature (over any time scale) then this would also need to be accounted for.
Is this type of solubility correction also required for O3, N2O, and CH4 ?
If so, how large are these corrections.


This type of system status measure is crucial if the claimed accuracy of measurement is creditable, or even possible.

How many machines are measuring at any one time.
Pardon, yes that had not occurred to me either.
In an email from Dr. Pieter Tans who is the head of the
Carbon Cycle Greenhouse Gases Group (CCGG), at NOAA he mentioned that the MLO has 4 machines.
So does that mean 4 machines are constantly recording. ?
Are all four data sets used to calculate hourly averages. ?
Does each machine have it’s own “system status” data ?
If less than 4 machines are used at any one time, is or are the machine/s identified. ?
There are many other factors recorded at MLO that need to be referred to specifically
to ambient and machine conditions so that the raw data quality is known.
Apparently there are frequent calibrations done at MLO, of all 4 machines, or just one,
and if so which machine/s have been calibrated. ?

This raises the thorny question of calibration gases, and standards.
At first MLO used an air and CO2 mixture reference gas.
Given the air in the “standard” contained O3, N20, and CH4, this does not sound a reliable reference.
Unsurprisingly the reference or calibration gas has changed twice since.
Yet there are no steps in the data……

I have given the above as limited examples of how quickly what should be easily understandable,
is soon so confusing as to be unfathomable.
1 or 4 machines, how many are calibrated, how often, with what, what correction/s to be applied, etc, etc.
There are so many examples it is difficult to know what to say or do next,
which maybe the idea after all…..

There is however a jump from the voltage figures every ten seconds to hourly averages
expressed as CO2 in parts per million. It would seem there is no intervening step of
raw CO2 ppm (or ppmv) figures, the voltage being calculated into an hourly CO2 ppm average figure
using an as yet unseen algorithm..

Corrections used.
It is hopefully clear now that to get from the voltage figure as measured to
a CO2 ppm or ppmv hourly average figure “corrections” must be being applied
for the other GHGs present in the dry air sample.
With this in mind,
On this page,
http://esrl.noaa.gov/gmd/obop/mlo/
Ozone hardly gets a mention. Is any variation in the levels of
methane, ozone and, N2O allowed for, or are constants assumed. ?
http://www.newscientist.com/article/dn14211
Tropical ocean sucks up vast amounts of ozone
New Scientist
26 June 2008
Excerpt,
“ Ozone is known to be largely broken down by sunlight and water vapour. T
his produces hydroxyl radicals, which in turn remove methane from the atmosphere.
When they measured how the concentrations of ozone in the troposphere varied over time,
they discovered that 13% of ozone is lost each day in this area, not 8% as predicted by models. “

As a slight aside, is this last article completely wrongly titled.
The oceans do not absorb ozone, to be released later, as the word “sucks” in the title implies.
The oceans release chemicals as the article describes in sea spray that promote breakdown of ozone,
which in turn releases other chemicals that promote the breakdown of methane.
“Tropical oceans breakdown ozone by about 50% more than realised previously
and this accelerates methane breakdown as well” would have been a more accurate title.

The below figures show how the modelled levels of ozone vary to the measured levels of ozone.
What is interesting is how different the curves are for all the locations plotted.
This suggests “correction factors” used for IRGA must be both local and varying.


As further evidence of the overall variation in the combined GHGs variations between locations
this is an interesting paper.
http://ams.confex.com/ams/pdfpapers/24874.pdf

Given recent discoveries and the resulting advances in our understanding of both ozone and methane
and their relationships with each other, as well as how the oceans may be effecting them.
It would seem reasonable to say that how ozone and methane levels vary is surely a significant factor
in the corrections applied to the voltage figure generated by measuring the dry air samples.

The basic point is that the other GHGs do not appear to be measured before or after sampling
so there must be assumed correction factors applied to the voltage figure to calculate
the CO2 ppm or ppmv hourly average.
Are these corrections constants ?
Do these corrections vary for each location and at each location ?
If the corrections do vary on what basis are they calculated from ?
I hope these corrections are NOT “modelled”….

Why is it that MLO (and all the other up to 40 sites) are not measuring more variability than they are ?
Is this lack of variability just in the processing, or have they allowed correctly for
what they knew of,
what they knew of incorrectly,
and what they did not know of.

It is amazing if they have done so, and the MLO record is correct.
It is a possibility after all, however remote a possibility it may seem.

As far as I can see or find the raw data, the other factors measured,
the system status measurements / calibrations for MLO and the correction factors applied
are not open or available, so I can not say if it (the MLO et al CO2 data sets) is correct,
and neither can anyone else.

The system status measures,
the correction factors applied, and
the algorithms used, are not available on the internet.

However if “you ask nicely”………..Yeah right.



Part3

Elsewhere I have condensed the above into a series of questions and or statements
I'll let you know of any meaningful responses....

" In short,
To my knowledge the IRGA method has never been independently verified.

IRGA does NOT measure CO2 directly.

The corrections used to get from a voltage figure to a hourly average
ppm or ppmv figure are unknown. These corrections are unknown in
value or how they are derived as they are not stated. It is even
unknown if they vary or have any localized differences.

Are solubility corrections applied for all the GHGs (if applicable)
present in the sample after drying.

Why has the changes of calibration gas make up (three I believe)
never shown up in the data sets.

Could the slow and gradual rise be due to warming cold traps..

Recent discoveries have altered relevant knowledge tremendously
recently, ozone and methane and oceans, but this has not effected the
released data..

How do you separate four GHGs in a sample that only measures the
overall reading,
without measuring several of the gases directly.... .
In which case there would be no need for the indirect method that
IRGA is.

When and how was IRGA proved to be as accurate as claimed, or
demonstrably better than wet chemical methods.

None of the algorithms, or formulas for calculating the CO2 figures
from the 10 second voltage measurements have been released so we do
not know what is calculated. We only know one of the (how many ?)
variables or constants used.

We have only seen 2 days of voltage figures of a 50 year record that
MUST include many, many other variables.

Maybe back in the late 1950s and early 1960s this new method of
measuring CO2 looked very promising but now,
I am not so sure it is even vaguely accurate.

The measurement is indirect, with unknown variables used to calulate
a figure that in reality there is no actual raw data for.

At least I understand a bit better why the raw CO2 data has never
been released.
There is no raw data, not for any part of the supposed 50 year record.
As far as direct measurements are concerned that is absolutely
correct because,
IRGA DOES NOT MEASURE CO2.

yours Derek.
_________________

An induced feeling of guilt is not sufficient reason
to convict, or punish the human race.

[Richard111]

A very impressive post Derek. I haven't read it all but recoginise a lot of the graphs. My personal bias is that most anthropogenic CO2 is generated in the northern hemisphere and then from the major cities. How long does it take for this CO2 to be come globally "well mixed"? How much gets washed out by rain? Been a very wet year in the UK this year. Bound to have reduced our contribution.

So MLO is definitely a NO. We need many more world wide stations with standard procedures, this includes over ocean readings. Going to be a while before this happens, if ever. IMO.

[Derek]

Thanks Richard,
I think the point I'm really trying to illustrate is the difference between IRGA CO2 levels and history, compared to wet chemical analysis.
They are SO different, one must be wrong, but which one...

There is little point in theorising from what MLO says is the CO2 level and history,
if it is completely wrong in the first place.
Hence Ferdinand has been so defensive of MLO.

I hope Lucy takes some solice from these posts.
_________________

An induced feeling of guilt is not sufficient reason
to convict, or punish the human race.

[ian]

Ta' so much for all your work Derek. Will take me some time to wade through.
best wishes, ian

[Derek]

Thank you ian, and happy voting..
_________________

An induced feeling of guilt is not sufficient reason
to convict, or punish the human race.

[Mike Davis]

Derek:
You made your point very well. When our favorite bell ringers come to tear it apart I would like to join the food fight.
Thank You for your time and effort.

[Derek]

[Mike Davis]

Montreal was a mistake that should never have been made. when I heard they were patting each others back about the success of the MP 10 years later! I was already aware of the cycles in the ozone hole and that they had reversed prior to the PACT actuallyy starting to reduce CFC's. Also when they were announcing to the world what they had acomplished in 10 years the hole was the same size as when the pact was signed. I enjoy watching the ocillations that naturally occure in nature.
I am on sattlite so do not have the speed for searches. Sorry but that plus other issues I can only offer to read what you have written and comment how it relates to my memory.

[Lucy Skywalker]

Very interesting Derek. But I don't know which way to poll, I feel I still don't know enough.

There was a really interesting piece at WUWT recently The CO2-Temperature Link by Frank Lansner. What interested me was his suggestion that the CO2 levels may show a cycle that crashes periodically rather than showing the regular up-down-up-down sine curve. He believes the biosphere grows steadily to accommodate increasing CO2 levels with increasing temp (as we've seen recently with global satellite sensors); then with decreasing ocean temp the CO2 levels go down but the biosphere is still "hungry" and drains the CO2 before readjusting with dieback.

This might at last give room to accommodate Beck alongside MLO. I'm not sure if there is hanky-panky at MLO. I don't like Keeling junior's attitude to Beck, but that is no proof of data massage either. I've just emailed Beck, from Derek's leads, to see if we can get his paper on the MLO in English without paying - it's free in German! Surely we want Beck's stuff in the open too, like MLO?

Now the MLO curve: why does it rise so steadily and at an increasing rate, if we are right? This is what I think (assuming no deliberate fraud):

(1) earlier on, it was still rising slowly from the LIA. The marine environment does a pretty good smoothing-out job, volcanoes notwithstanding.
(2) then the solar activity of recent decades kicked in to speed it up. The TSI increased but not quite enough to explain such an increase; however the Svensmark cloud-clearing effect, following a great increase in solar magnetic activity, gave the push.
(3) it is still increasing longterm from the LIA as well as showing thermal inertia effects from recent solar activity. Downturn is slow - in the same way, the Artcic ice does not start to increase until September or so.
(4) the ice core record splice with the MLO record is fraudulent, as Jaworowski showed, and the ice record is untrustworthy and should be abandoned pro tem until we can get a more believable correlation between Beck's records and the ice records.
(5) Is the Mauna Loa volcano getting more active? Yet I would not expect it to get active in incredibly smooth progression.
(6) the interesting plateau-tending-to-dip time between November and April, in the multiple annual anomalies chart. Doesn't this just seem to show remarkably clearly the effects of the huge NH biosphere and the much smaller SH biosphere, as they go through their respective growing seasons?

[Derek]

I dunno Lucy, I was just looking at the method used at MLO to measure the CO2 level.
Which IRGA does not do directly..That's the point.

NB - Anyway seasons greetings to everyone, and I do mean everyone,
I'm only here to question reasonably and discuss.
I'll see you all in January,
buns sharpened...
_________________

An induced feeling of guilt is not sufficient reason
to convict, or punish the human race.

[Derek]

Hi All,
I'm looking back and forth at the moment.
Back - There were plenty of techniques to measure CO2 in the 1940s and 1950s, but Keeling made a new one.
(with help from people with an unproven hypothesis that needed data to "prove" it, ie Callender / Revelle)
He (in very, very short) took a closed area measuring method and applied it outdoors,
on a known volcanic island, in the middle of a de-gassing ocean ....

Going forwards - The "next generation"...
http://www.licor.com/env/Products/GasAnalyzers/7000/7000.jsp -
LI-7000: The Next Generation CO2/H2O Gas Analyzer
manufactured by Li-Cor

Hmmm, note manufacturers name...
Here is their presentation regarding the machine / instrument, and method.
http://www.licor.com/env/Products/GasAnalyzers/eddyPresentation/EC_master_Presentation_files/_frame.htm

Some excerpts..
Notes slide 20.
" None of these errors is trivial. Combined, they may sum to over one hundred percent of the initial measured flux value. To minimize such errors, a number of procedures exist within the Eddy Covariance technique. Here we show the relative size of errors on a typical summer day over a green vegetative canopy, and then we provide a brief overview of the remedies. Step-by-step instructions on how to apply these corrections are given in the Data Processing Section of this presentation. "
and,
Notes slide 112
" Overall, perhaps the most important point about Eddy Covariance workflow is the necessity for an individualized customized approach to each experiment, because Eddy Covariance, to a large extent, is a site-specific method.

Each study site is different and requires unique treatment in terms of experimental design, tower placement, instrumental set up, data collection, and processing and analysis.

Built-in flexibility of the Eddy Covariance method, in conjunction with user knowledge and understanding of the method and the study site, will allow for successful implementation of site-specific arrangements that are tailor-made for a specific scientific purpose under specific ambient conditions. "

end of excerpts.

That's the next generation.........

I think the phrase
" for successful implementation of site-specific arrangements that are tailor-made
for a specific scientific purpose under specific ambient conditions. "
says it all.

Results taylor made to your requirements - anywhere, and,
incomprehensibly hidden in the corrections / assumptions applied to the measured data only you know.

Incidentally it is obvious that temperature and pressure records are absolutely vital to the IRGA (or NDIR if you prefer) method employed at MLO.
Anyone ever seen a record of such.


I thought not.

NB - Lucy is it possible to edit the poll question on this thread to include the 2 I missed out in "global CO levels",
I errr typed it wrongly. -
I have tried editing the first post, but the poll does not appear to be editable from there.
Thanks in advance Derek.
_________________

An induced feeling of guilt is not sufficient reason
to convict, or punish the human race.

[Richard111]

Thanks for the above info Derek. More questionable "science" in the market place. (sigh)

[anna v]

I would like to remind about the AIRS CO2 measurements

[FerdiEgb]

Dear Derek and all,

You know, I had comments in the past about your investigation of the MLO data, after I had helped you as much as possible, But I did give up as I had the feeling that whatever was said by the researchers at MLO, you never would believe the data...

But I will try to give some more comments, where you may be right and where you surely are wrong...

To begin with: the spectral analyses. The spectra you did plot are right for the same level (moles) of the different GHGs. But the GHGs are not at the same level at all:

H2O: 1,000-10,000 ppmv
CO2: ~385 ppmv
CH4: ~2 ppmv
N2O: ~0.3 ppmv
O3: ~0.1 ppmv
CFC's: sum of several: less than 0.001 ppmv

As you can see, water has the largest influence, and must be removed (although there are alternatives, see http://www.ejournal.unam.mx/atm/Vol16-1/ATM16101.pdf ) if one need an accurate measurement of CO2.
CH4 and CO2 don't overlap for several bands, thus no problem.
And the rest of the GHGs is too small to be problematic at all.
In fact as the reference before says, CO2 is a problem for N2O measurements, not the reverse...

Thus a CO2 measurement only measures CO2 and water at the bands where is measured. Water is removed by freezing and ice doesn't contain measurable amounts of gases (contrary to ice formed from compressed snow where air remains in enclosed bubbles). Thus no CO2 is lost, except at the surface of all ice, lines, valves, where some CO2 is retained. But that is solved by allowing several minutes of flushing without measurement, every time that a flow is switched between ambient air and reference gases and in between reference gas switches.

I am very grateful for the plot of the monthly means over the 47 years at MLO. It proves that little year by year variation happens in nature, including the human emissions, which effect is only 2 ppmv/yr nowadays or 0.08 ppmv/month or 0.003 ppmv/day or 0.0001 ppmv/hour. The latter two not detectable even with the best available techniques.

But be aware that the monthly avarages are "cleaned" from outliers. Not that it makes much difference, as most outliers are not more than +4 ppmv from volcanic outgassing of -4 ppmv from upwind depleted CO2 by vegetation at MLO. Including or excluding the outliers doesn't change the average or trend with more than a tenth of a ppmv.
Other stations like the south pole have not such problems, except mechanical ones, and show the same trends:


If you need the voltages from a given year for research, simply ask them (I have asked for a few days and did receive them, as 8 million datapoints for a single year is a little overdone for a check of the methodology).

From these two days of raw voltages, the calculated CO2 level is the (quadratic) interpolation between three voltages measured over several minutes from three reference gases, each with well known CO2 content, for every one of the 40 minutes of 10 second ambient air measurements expressed in voltage and calculated as ppmv. The reported result is one average value, and the standard deviation of the 240 measurements caculated in ppmv. These values (hourly averages) are available on line, the individual values not, but can be (re)calculated at any moment in the present or the future if one requests for the raw voltage data and/or if one of the reference gases need recalibration. There is no substraction for other GHGs, as these are not interfering with more than the accuracy of the measurements.

As the supposed variability through the seasons and the human addition is so small, it can be sufficient to take a sample once a week (in duplo/triplo) to be sure. This is effectively done at MLO by different laboratories and different organisations and different methods, as a check for the continuous measurements. Average deviations: less than 0.2 ppmv for the different series.

The NDIR method was far more accurate (+/- 0.1 ppmv) than most of the old wet methods (+/- 10 ppmv) until that day of its introduction by Keeling. He tested the new method with a manometric method up to 1:40,000 units accurate. The method was and is robust, accurate enough to see the seasonal and continuous variations and asks for a minimum of maintenance and oversight. There are a few alternatives which are more accurate: gaschromatography, which is used in several stations (which give similar results), but that is prone to more (mechanical) failure and needs more maintenance, and mass spectrometry, very expensive and a lot of maintenance, but used e.g. for isotope fraction determination, mostly off-line on flask samples.

As said before, even the MLO data are checked by different organisations, different laboratories, different persons and different methods. That is described in
http://esrl.noaa.gov/gmd/ccgg/about/co2_measurements.html#replication

As far as I know, there are not four machines at MLO, only one. I suppose that there is some misunderstanding here...

The first calibration gas standard was CO2 in nitrogen. Only years later it was discovered that calibration of CO2 in N2 gives different results than CO2 in air. Therefore all NDIR instruments over the world were recalibrated with the new CO2-air mixtures (which don't have trouble with O3, N2O or CH4 at ambient levels). This caused a jump of about 12 ppmv in most data, but as the recalculation was easy, as based on raw voltages, all older data were recalculated with the new reference values, thus no jump, only a more accurate result for the first years.

I suppose that most of the questions posed by you are answered here. With one exception:

None of the algorithms, or formulas for calculating the CO2 figures
from the 10 second voltage measurements have been released so we do
not know what is calculated. We only know one of the (how many ?)
variables or constants used.

This is clearly wrong: the full method, algorithms and formulas for calculating the CO2 levels from the voltages are described in
http://esrl.noaa.gov/gmd/ccgg/about/co2_measurements.html#instrument

I have recalculated the CO2 values of the two days of raw voltages via an excel sheet, took me a few hours to write the program that does the job, only based on the above description. And the recalculated values are exactly the archived ones, with a little deviation in the second decimal (as I, a little lazy, used a linear interpolation, not a quadratic curve interpolation). The results were as published, without any correction or other pre-processing.

Your last sentence is quite strange to say the least:

IRGA DOES NOT MEASURE CO2

Do you mean that thermocouples (mV) or mercury (expansion) don't measure temperature, neither do gaschromatographs measure anything (mass/ionic retention+flame ionisation), mass spectrometers (idem),... odometers (magnetic pulses or friction), don't measure speed and distance, because that are not direct measurements of the real variable you want to know with high accuracy, and these don't give the raw data (mV, mm Hg, ionisation V measurements, pulses, rpm of the weels,...)?

[FerdiEgb]