Strong evidence for linear law removal of atmospheric carbon dioxide

Posted: September 25, 2013 by tchannon in Analysis, atmosphere, Carbon cycle, climate, cosmic rays, Geomagnetism, methodology

The IPCC base their claims about how much warming is ‘in the pipeline’ on the rate that carbon dioxide is eventually removed from the atmosphere, the ‘e-folding time’. This is different to the time it takes for any emitted joe-average co2 molecule to be re-absorbed in the carbon cycle, and reflects the assumptions made about the way the carbon cycle operates.

The IPCC relies on the ‘Bern model, which was cooked up many years ago by Bert Bolin and other atmospheric scientists of the warmist persuasion. The Bern model makes assumptions which lead to a very long e-folding time of hundreds of years, a figure which has been disputed by several able researchers, and discussed here at the talkshop in previous posts.

Now talkshop co-blogger Tim Channon has made novel use of data which shows what has happened to the radioactive carbon 14 isotope carrying co2 levels since the atmospheric nuclear bomb tests of the early 1960’s. The results look like another hole below the waterline for the IPCC and the climate alarmists. – tb


Figure 1


Figure 2

Figures 1 and 2 are demonstrating both northern[1] and southern[2] hemisphere decay from a Dirac injection[3] of a test signal. The consequent effect is very close to perfectly linear, proportionality between pressure and effect of pressure over more than an order of magnitude of data variation (hence linear law). This seems to destroy IPCC claim of a non-simple law. Deviation is <1%

In addition the effect is a simple low pass filter on all kinds of atmospheric carbon dioxide. A later article might cover this in detail.

Note: this article is cross posted from the authors blog, discussion is probably more appropriate on the Talkshop. Some of the content has been the subject of wide discussion around the ‘net but not so far as I know here.  — Tim

There are few readily available 14CO2 datasets.

Figure 1 is as labelled, from inside the Arctic circle near to the Soviet test site at Novaya Zemlya. Figure 2 from New Zealand, nearer to western test sites and differing in largely ocean environs. Both data are irregularly sampled, both have non-trivial breaks, both have changes in the method of collection and sample measurement. For example, the Norwegian data used the older decay event counting where later on the count integration period was increased to 4 days. This may account for the slight change in the apparent noise level.

Normal circumstances

The earth is irradiated by cosmic rays, a few of these impact atmospheric atoms, in these case the important one is nitrogen, breaking out a proton and then it is literally transmogrified into an isotope of carbon, carbon 14. (12, 13 and 14 exist)

Free carbon and free oxygen will combine in the atmosphere forming a molecule of radioactive carbon dioxide.

Carbon dioxide is removed from the atmosphere as well as interchange mechanisms being present.

The data has been compensated for the slight removal by radioactive decay (half-life ~6ky, back to nitrogen), diffusion present in the collection mechanism (chemical capture) and reference change. (data [1] presents both raw and compensated)

As a consequence of this, air contains a very small proportion of 14CO2, the amount set by a balance between formation and removal.

There are three removal mechanisms

  1. radioactive decay , this is negligible because 2 and 3 dominate for the atmosphere
  2. update by photosynthesis where the 14C is combined into new molecules which are part of dead plant, or shell or whatever
  3. dissolves in water, such as the ocean

2 and 3 are interchanges with CO2 also being emitted

At this point I need to mention the matter of fossil fuel usage. Because of it’s age it contains near zero 14C and therefore usage dilutes atmospheric 14CO2, leading to a reduction in the isotope ratio in the atmosphere. The effect is small and contentious.

The bomb spike

Decode of the exponential decay is simply done as a linear regression on the natural logarithm of the Y data. I chose 1966 as the start to avoid the excess annual variation. r2 of fit is 0.99 in spite of the additional excess injection by later bomb tests (which will elevate levels slightly)


Figure 3

A simple graphic method of determining time constant is take the tangent of the curve and measure the time to origin. In this case is broadly based on the Norwegian data.

Roughly, northern ~15 years, southern ~17 years, near enough the same. Both are R2 > 0.99 of exact law.

The bombtest curve and its implications

IPCC etc. insist CO2 remains in the atmosphere for a very long time yet that is in bald conflict with the curves shown, particularly an assertion of an elevated oceanic impedance to flow.

There is no sign of further time constants in these results, which would result in curvature.

A typical example of establishment thinking is here at Yale

It turns out that while much of the “pulse” of extra CO2 accumulating in the atmosphere would be absorbed over the next century if emissions miraculously were to end today, about 20 percent of that CO2 would remain for at least tens of thousands of years.

State lifetime 5 to 200 years and note “c”

c No single lifetime can be defined for CO2 because of the different rates of uptake by different removal processes.

So how does bomb CO2 distinguish?

This brief PDF says 400 years

If these plots are taken at face value?

There are two recent collections of PDF produced by specialists in slightly different fields, both strongly disputing the establishment opinion.

Readers may have seen both.

The first a trio of on-line papers by “Gösta Pettersson, Chemical Center, University of Lund, Sweden”, where someone has published on a blog about page

Gösta Pettersson, born Nov 24 1937

Professional carrier at the University of Lund, Sweden

Ph D in biochemistry 1966
Professor assistant in biochemistry 1968–77
Professor in biochemistry 1978–2001

So we have a veteran biochemist. Web site is here and has been done specifically to publish these papers. I’ll put safety copies on the Talkshop.

Pettersson uses the bomb test data as part of criticism of the IPCC view on carbon dioxide in the atmosphere.

I pick up specifically on page 5 of paper one, highlighting mine.

This means that the Bern model overestimates the final equilibrium amount of airborn carbon dioxide by a factor of about 15. Since the
hydrosphere is the predominant sink for an excess of atmospheric carbon dioxide, one could say that the Bern model makes carbon dioxide about 15 times less water soluble than it has been empirically established to be according to the carbon cycle data reported by the IPCC [9].

Gösta Pettersson papers
Blog site, HTML content and PDF papers.
Local copies [4]

Secondly a short while ago I highlighted a triplet of papers by David Coe [5] who has a slightly different specialism in atmospheric gases.

This work is complementary to Pettersson’s where at the end of part 3 Coe writes

In stark contrast this fingerprint evidence, along with seasonal variations of atmospheric CO 2 and O 2 and interannual variations of CO 2is explained in a rational and consistent manner by a theory based solely on the precept that transport of gases is governed simply by variations
in partial pressure between the deep ocean, the sea surface and the atmosphere.

Model used here

As with an innovative work on lunar temperature I have used SPICE to model a process where there is an exact dual in a different field, merely needing unit of measure translation.

An involved model could be derived, excellent for what-if quick investigations.

For example for illustrative purposes


Figure 4

A parametric run produces a family of curves plus one near straight line. (parametric, software repeats analysis using a range of parameter values)


Figure 5

As I mentioned, any deviation from a pure law will show curvature on a log plot. Far more complex schemes could be produced.

Post by Tim Channon

1. Norwegian data is part of the CDIAC archive

14C archived “bomb” data

14C archived complementary data for ocean content

2. Baring Head 14CO2 dataset (ascii file)

NIWA website, same data available.

WDCGG (World Datacentre for Greenhouse Gases), part of WMO (World Meteorological Organisation) is carried by JMA (Japan Meteorological Agency)[1]

“CREDIT FOR USE: This is a formal notification for data users. “For scientific purposes, access to these data is unlimited and provided without charge. By their use you accept that an offer of co-authorship will be made through personal contact with the data providers or owners whenever substantial use is made of their data. In all cases, an acknowledgement must be made to the data providers or owners and the data centre when these data are used within a publication.”

3. Dirac pulse (has other names too) is a theoretic concept, an infinitely large and short pulse which acts to stimulate an associated system, responding in some dynamic manner from which information can be deduced about the system. Classic example is the man who taps railway wheels and listens to the result, looking for cracks in the steel.

4. Copies of Gösta Pettersson papers

paper1 paper2 paper3

5. Coe papers
Original article here

  1. hunter says:

    This is an intriguing paper. Its implications are important: yet another IPCC/AGW climatocracy assumption bites the dust.

  2. tchannon says:

    I hope so, assuming I haven’t make mistakes, yet the whole thing is contentious.

  3. michael hart says:

    One of the arguments I have seen, is that the observed rate constant for atmospheric 14C removal (as 14CO2) is much greater than that for 12C because while the 14C bomb spike is diluted into a rapidly-responding reservoir(s) containing far less pre-existing 14CO2, the same reservoir(s) is/are already much closer to equilibrium as far as 12CO2 is concerned.

    This causes me some difficulties:
    1) The atmospheric 14CO2 appears to fall back to the original pre-bomb levels, implying that the rapid reservoir(s) is/are large compared to the size of the atmospheric reservoir.

    2) The atmospheric 14C concentration didn’t display any seasonal effects in the data I have seen, implying something other than a small, rapidly-exchanging reservoir(s).

    3) Sequestration by photosynthetic fixation is a process that is NOT close to equilibrium (in the way that thermal dissolution/out-gassing might be said to be close to equilibrium). Therefore I would not expect to see much (land based?) photosynthetic isotope discrimination: The plant/tree will grab the CO2 molecules as fast as they arrive during active photosynthesis (the absolute biochemical isotope discrimination being very small). So this form of photosynthetic carbon removal would not produce a markedly different atmospheric CO2 residence time.

    4) As I’ve mentioned before, the chemical speciation resulting from aqueous hydration of CO2 to carbonic acid/bicarbonate is much slower than a diffusion-controlled reaction. The biosphere makes ubiquitous use of carbonic anhydrase to speed up the reaction by up to six or seven orders of magnitude. (And there is a biofilm covering most of the worlds oceans.)
    If this catalysed reaction is responsible for any significant amount of the performance of carbon sinks then it may well torpedo one of the unstated assumptions in the ‘consensus’ mechanism above:-That the rate of conversion of 14CO2 to 14C-bicarbonate is independent of the rate of conversion of 12CO2 to 12C-bicarbonate. If described by an enzymatic-biochemical mechanism (to any ‘significant’ extent), then both 14CO2 and 12CO2 will be competing for (the relatively tiny amount of) the same active site in the carbonic-anhydrase molecule, meaning that the rate constants for conversion of both isotopic forms of CO2 would be approximately similar.

  4. tchannon says:

    I suspect the 14C data is differential with 12C+13C atmospheric, proportional and therefore the annual cycle is subtracted out, hence we do not see it unless there is a distinction made by a consuming process.

    Getting a clear answer out of the data authors is in my experience rather difficult.

  5. Roger Andrews says:

    Numerous estimates of residence time based on bomb test 14C decay have been made since 1957. Thirty of them are listed on the table below (number two is from Bert Bolin, incidentally):

    C3: The Liberal Attack On Science: The IPCC Fabrication of Atmospheric CO2 'Residence Time'

    It’s not clear what the table means by “Atmospheric Residence Durations”, but if it’s residence time, i.e. when the 50th in a sequence of 100 molecules gets reabsorbed, then residence time = e-folding time times 0.6933.

    A question here is whether 14C decays more quickly than 12C or 13C. I don’t know the answer to that one but maybe someone else does.

    I don’t see any “linear law removal”. The decays on the two graphs are beautifully lognormal.

  6. tchannon says:

    ” The consequent effect is very close to perfectly linear, proportionality between pressure and effect of pressure over more than an order of magnitude of data variation (hence linear law).”

    If a constant excess was maintained the flow would be constant. It is self decaying so it feeds itself and results in a exponential decay, the effect from a one shot.

    The proportionality shows no evidence of non-linearity and that is why I question additional time constants being present.

  7. R J Salvador says:

    This reminds me of an undergraduate experiment with radioactive isotopes many eons ago. A chlorine containing organic compound, highly insoluble in water, is placed in contact with an aqueous solution of a metallic chloride salt solution where some of the chloride ions are radio active. The solubility of each compound in the other is already known by previous chemical analysis. In a short period of time we found that the concentration of radioactive chlorine in the organic phase was orders of magnitude higher than what solubility would predict. Conclusion the chloride ions and the chlorine atoms in the organic compound could replace each other across the phase boundary layer. The world of chemistry never seemed stable after that.

    I can see this how this a contentious issue. TC the 15 years maybe only one half the kinetics. There is also the kinetics of CO2 coming out of the ocean. If so then the controlling step is the transport of CO2 probably in the form of HCO3- away from the boundary layer at the surface of the ocean.
    Another source of chaos.

  8. michael hart says:

    Thanks, Tim. The second of my four points requires more thought.

  9. kuhnkat says:

    Dr. Salby will be pleased I am sure!! 8>)

  10. wayne says:

    I’m intrigued by your very first graph.

    Looks like the smoothed green curve crosses 800 in late 1963.
    Crosses the 400 level in 1974, eleven years to halve.
    Crosses the 200 level in 1985, eleven years to halve again.

    So I assume the half-life is very close to 11 years. But much longer values are given further deeper in the article and I am wondering why. Might be you are speaking of e-folding which would be 0.367 (1/e) and not half, is that right? So the eleven could be stated 15 years to e-fold (11/0.367/2) and I’m just checking if that is correct (little rusty on e-folding). maybe you could just give the equation of the relation you are using for clarification to some here. Sometimes articles speak as if everyone is thoroughly embedded in that branch of science and the units that branch uses and I bet there are others, like myself, who are not. You can forget a lot of picky details like that since college, like calculating e-folds. Also does ‘time constant’ on the further graphs imply an e-fold as the units? They do not specify so the graphs don’t mean much at a glance.

    As far as IPCC speaking of hundreds or a thousand years… hogwash, propaganda.

    Pretty sure biological systems or sea-water give no huge preference in the uptake of co2.14 over the two other main isotopes twelve an thirteen. Use to work with co2.14 in the later years college (primary degree in physiology) and wear the little radiation badges. 😉

    Tim, love the way you use the circuits as parallel models, neat and correct in the symmetries of physics and chemistry.

  11. tjfolkerts says:

    I think you are confusing two different concepts.

    One concept, often called “turnover time”, is how long it takes for one particular CO2 molecule to cycle in and out of the atmosphere. This is usually considered to be a short time — 3-5 years from the estimates I have seen. One molecule gets absorbed by a plant for example, and another molecule gets exhaled by an animal. This simply cycles CO2 molecules, but does not remove them from the atmosphere.

    The other concept, often called “residence time”, is the time to remove a CO2 molecule for good. So for example, the ocean might absorb 101 molecules and only emit 100 molecules, for a net loss of CO2 from the atmosphere as a result of the system being out of equilibrium. This time is usually considered to be much longer — decades or centuries. This seeks to answer the question “if the equilibrium CO2 concentration is ~ 280 ppm and we raise it to 400 ppm, how long will it take to return to 280 ppm?” Simply cycling the CO2 molecules in and out every 3-5 years doesn’t do this.

    The strange thing is, I would have expected the radioactive CO2 to get removed on time scales related to the turnover time — ie a time constant of ~ 4 year. You get a time sort of half-way in between these two time scales.

  12. Roger Andrews says:

    The curious thing about the atmospheric CO2 residence time dispute is that there actually isn’t one. Even Skeptical Science acknowledges that CO2 residence time is short. (“It is true that an individual molecule of CO2 has a short residence time in the atmosphere.”)

    The dispute arises over what residence time means in the practical sense. The skeptics assume that a molecule of atmospheric CO2 disappears into the ocean and stays there when its residence time is up. However, Skeptical Science claims that; “in most cases when a molecule of CO2 leaves the atmosphere it is simply swapping places with one in the ocean. Thus, the warming potential of CO2 has very little to do with the residence time of CO2.”

    So if it has very little to do with residence time, what does it have to do with? Back to the carbon cycle 🙂

  13. tallbloke says:

    Tim F: In previous discussions here, where Richard Courtney has joined in, residence time is the ‘turnover time’ you refer to, and the e-folding time is the long term removal time. These are the generally accepted definitions so far as I’m aware.

    ” I would have expected the radioactive CO2 to get removed on time scales related to the turnover time — ie a time constant of ~ 4 year.”

    This depends on whether the radioactive co2 molecule gets absorbed by something which removes it semi-permanently, like a growing phytoplankton shell which then dies and takes the shell to the abyss, or a short term sink like the winter-time upper ocean, where it will more likely be re-emitted the following summer.

    “You get a time sort of half-way in between these two time scales.”

    Errm, what makes you think the absorbers prefer 14C? If the e-folding time for 14C carrying co2 molecules is 15 years, its fifteen years for all co2 molecules.

  14. tallbloke says:

    Roger A: Remind me again how much of the extra co2 in the atmosphere is of human origin according to your model, if the e-folding time is 15 years. 😉

  15. Roger Andrews says:


    Trying to avoid potential terminological confusion:

    A 15-year residence time is roughly equivalent to a 21 year e-folding time and gives 60% anthropogenic CO2.

    An 8-year residence time is roughly equivalent to an 11 year e-folding time and gives 40% anthropogenic CO2

    A 5-year residence time is roughly equivalent to a 7 year e-folding time and gives 25% anthropogenic CO2

    Assuming, of course, that the CO2 molecules are absorbed for good when their residence time is up and don’t get replaced by sneaky CO2 molecules going the other way.

  16. tallbloke says:

    So if we have a residence time around 5 years and an e-folding time around 15 years, what parameters in your model would have to change to accommodate the ratio? If co2 molecules were absorbed for good after residence time is up, why wouldn’t the e-folding time would be the same as the residence time?

  17. Roger Andrews says:

    If we accept that residence time is when the 50th of 100 molecules gets absorbed then residence time is 0.6933 times e-folding time. So to get a 5 year residence time and a 15 year e-folding time we would either have to change the 50th molecule to some other molecule or repeal the laws of mathematics. What’s your preference? 😉

  18. Roger Andrews says:


    Here’s the math:

    I make two assumptions. First that CO2 in the atmosphere decays exponentially, and second that the “residence time” is when the 50th molecule in a chain of 100 emitted CO2 molecules (or the 5 quintillionth in a chain of 10 quintillion) gets reabsorbed. This occurs when:

    e^-(t/T) = 0.5

    Where t is the residence time and T (tau) is the e-folding time.

    Solving for t/T gives:

    ln(0.5) = -t/T

    t/T = 0.69315

    So the residence time is always 69.315% of the e-folding time (I was going from memory above and got it slightly off).

  19. TB, there is a huge difference in “excess CO2 decay” time of 14CO2 and 12CO2. That is not because of the fast responses with vegetation (seasonal) and the ocean surface (1-2 years). The bomb spike 14CO2 was rapidely distributed between the atmosphere and these reservoirs and in 1960 certainly completed.

    The difference is in the deep ocean circulation: what goes into the deep oceans is the isotopic composition of today. What comes out of the oceans is the composition of some 1,000 years ago…

    Here a graph of the (estimated) CO2 fluxes and the 14CO2 concentration in the different fluxes around 1960, the peak bomb spike year where the 14CO2 concentration is set to 100%. The pre-bomb equilibrium was around 50% of the bomb spike:

    What one can see is that while the ocean surfaces and vegetation are in rapid equilibrium with the atmosphere, the deep oceans return about 98% of all 12CO2, but less than 45% of 14CO2 which was going into the deep oceans. Thus the decay of a 14CO2 excess peak is much faster than of a 12CO2 excess peak. About three times faster.
    Here the same graph for the year 2000:

    Something similar happens to the 13C depleted human emissions: the drop of d13C in the atmosphere is about 1/3rd of what may be expected if there was no “thinning” by the deep ocean exchanges. This can be used to estimate the deep ocean exchanges:

    Thus the excess 13CO2 decay is three times faster than the excess 12CO2 decay. The 14CO2 decay is similar faster than the 12CO2 decay.

    The real decay time of excess 12CO2 is a measured drop of ~2 ppmv/year (~4 GtC/year) for an excess CO2 pressure in the atmosphere of ~100 ppmv (~210 GtC) above equilibrium. That gives an e-fold time of 210/4=52.5 years. Or a half-life time of ~40 years.
    Far above the ~5 years residence time, but far below the centuries of the Bern model, which are based on the saturation of the deep oceans (and land vegetation sinks), for which there is no sign in the foreseeable future.

  20. tchannon says:

    Sorry about the outage.

    I’ve shown what I found. Nothing new. been known for a long time, showing the duality is perhaps novel The meaning is a different matter.

    Anyone know of more 14CO2 datasets, preferably high res?

    Yesterday I reverse engineered part of the data. Result is as expected not interesting, is noise, data is too poor.

    One of the effects of the natural low pass filter is removing variation from atmospheric CO2 and also, Tallbloke will like this, part explains why 14C proxy data does not show solar. This begs some questions.

  21. Roger Andrews says:
    September 26, 2013 at 7:09 pm

    I see that the confusion again is complete… Even the IPCC uses all definitions in complete puzzlement…

    One definition is clear: residence time. That is the average time that a CO2 molecule resides in the atmosphere before being exchanged with a molecule from another reservoir. That is around 5 years. Nobody disputes that, not even the IPCC.

    The residence time can be calculated by a simple formula:
    amount in the reservoir / exchange rate (that is influx or outflux).
    In the case of CO2: 800 GtC / 150 GtC/yr = ~5 years.
    If there is no difference between influx and outflux, then nothing happens with the amounts in the reservoirs, even if about 20% of all CO2 is replaced each year.
    If there is a one-shot influx of some “strange” CO2 (human or red-colored or 14CO2) one can calculate the decay rate of that different CO2 molecule if nothing is returned or if some is returned and one knows the size of the other reservoirs. But that is not the relevant decay rate (e-fold time) we want to know.

    What we want to know is how fast and excess in mass of CO2 above the (temperature dictated) equilibrium is going into the other reservoirs. That has nothing to do with the residence time, as that is independent of the size of the fluxes but only depends of the difference between these fluxes. The difference is influenced by the exerted CO2 pressure in the atmosphere above equilibrium. The decay rate in this case is given by the ratio disturbance / effect:
    210 GtC / 4 GtC/yr = 52.5 years.

    Quite different decay rates for different types of decay…

  22. DocMartyn says:

    This is actually more tricky than it appears on first glance.
    14C is measured as a ratio of total CO2. As there is addition of, and loss of CO2, into the atmosphere during the 14C plot you have to work out the degree of dilution of 14C due o an increase in the size of the atmospheric carbon pool.
    Much more interesting than the rate is the estimate of relative reservoir ratio. If the size or the atmospheric carbon reservoir was the same size as the interrogated ocean reservoir, then starting from a delta of 800, the end point would be 400, if the ocean reservoir was three times bigger than the atmospheric reservoir the end-point would be 200, and if it the ocean reservoir was 7 times bigger than the atmospheric carbon reservoir then he end-point would be 100.
    If you do the plot, you get an endpoint near 20-50, due to noise. This suggests that the size of the carbon reservoir the atmosphere is talking to is at least >20 times the size of the amount of carbon in the atmosphere.
    This position of this end-point is actually the most important piece of information that we can get from the 14C plot; what it says is that there is no fat tail, the ocean doesn’t have layers, top to bottom, were carbon moves very slowly. The carbon in the atmosphere is sequestered quickly and mixed with the whole of the ocean, very quickly. I believe that it is due to the speed that organic ‘marine snow’ falls from the surface to the depths, being converted to DIC, all the way down.

  23. Peter Shaw says:

    tchannon, tallbloke –

    Dr Pettersson argues that the Bern model has been “tuned” to the Keeling curve, and that this isn’t the best approach – a simple decay model suffices (as you show).
    However, your Fig 1 also discloses information which I hope is not off-topic for this blog, as it may be of importance. It will complicate any model, but provide insights and constraints.

    The Nordkap data shows:
    > An overall simple exponential decay (your fit);
    > A seasonal variation peaking in June/July (?);
    > This itself decays rapidly from ~20% (range) with TC ~1½-2 years.

    Add to this the Mauna Loa data:
    > Slow secular rise of CO2;
    > Seasonal variation peaking in May (ie not in phase with Nordkap?);
    > Seasonal range 2%, but larger at high N latitude (see GLOBALVIEW from NOAA).

    Very tentatively, I offer the following deductions for discussion:
    1 A single large “irreversible” CO2 sink (the ocean);
    2 An annual partially reversible sink (temperate/boreal continental vegetation);
    3 A longer-term (decadal?) reversible sink (degree of reversibility not determinable);
    4 Seasonal NH continental growth annually reduces CO2 by ~20%;
    5 A combination of rapid (3-5 yr) primary ocean CO2 exchange and slower 3 gives the aggregate ~15-year decay.

    If these stand up:
    > Net Primary Production is non-zero (short-term, and probably longer);
    > The form of the Bern model is unsuited for this system;
    > Any CO2 model that excludes life is incomplete.

    For readers in the NH, Earth has two summers per year. A single seasonal peak indicates a dominant hemisphere.
    Seawater contains ~130x atmospheric CO2 (v/v as bicarbonate), so 14C is immediately considerably diluted on solution.
    [Ta for the links to 14C data]

  24. DocMartyn,

    The difference in (carbon) size between the (deep) oceans and the atmosphere is about 40,000/800 or 50:1. Further complicated by the fact that the output to the deep oceans and the input from the same is effectively decoupled for about 1000 years. Thus the first 1000 years virtually no bomb spike 14CO2 returns from the deep. After 1000 years, the spike probably is already mixed in with the rest of the deep ocean CO2 and hardly measurable…

    Further there are two distinct layers in the oceans: the deep oceans and the surface (“mixed”) layer of 100-200 meter which is rapidely exchanging CO2 and other stuff with the atmosphere. Carbon content ~1000 GtC, comparable with the atmospheric content. Despite the rapid exchanges, the 14CO2 is not evenly distributed between these two: the distribution is about 10:1, as result of the”Revelle” (buffer) factor. Thus a 100% increase in 14CO2 (or 12CO2) in the atmosphere only gives a 10% increase in the mixed layer.

    Most of the exchange between the atmosphere and the deep oceans bypasses the mixed layer: a huge sink in the polar waters of the NE Atlantic and upwelling places in the equatorial Pacific. But these exchanges are rather restricted: about 40 GtC/year or 5% of the atmospheric content.

  25. Roger Andrews says:

    Ferdinand Englebeen says:
    September 26, 2013 at 9:37 pm

    “I see that the confusion again is complete… Even the IPCC uses all definitions in complete puzzlement…
    “One definition is clear: residence time. That is the average time that a CO2 molecule resides in the atmosphere before being exchanged with a molecule from another reservoir. That is around 5 years. Nobody disputes that, not even the IPCC.
    “The residence time can be calculated by a simple formula:
    amount in the reservoir / exchange rate (that is influx or outflux).
    In the case of CO2: 800 GtC / 150 GtC/yr = ~5 years.”

    When I started looking into this question I found two measures of the length of time that CO2 endured in the atmosphere that were robust and mathematically definable – the half life and the e-folding time. The e-folding time seemed to be in more common use, so I chose that.

    But the numerous previous researchers in the field had mostly estimated “residence times” (of between 5 and 25 years, incidentally). What did they mean by residence times? According to Wikipedia residence time is the average amount of time that a particle spends in a particular system – basically the same definition as yours – and I interpreted this, rightly or wrongly, to be when the 50th molecule in an emitted chain of 100 gets absorbed. This I could relate mathematically (see above) to the e-folding time, whereupon I had two robust and mathematically definable metrics I could use.

    Now we come to your estimate of 800/150 = ~5 years. The expression residence time = reservoir size/exchange rate would indeed be a good choice IF we had reliable numbers to plug in. But we don’t. We can be reasonably confident that the atmospheric reservoir contains about 800 Gt of CO2 (762 according to AR4) but we can’t be in the least certain that CO2 interchange between the atmosphere and terrestrial/ocean sinks is 150 Gt/year (213 according to AR4). This number is basically a guess, and it will remain such until we have a much better understanding of how the carbon cycle works than we do now.

    So no. I’m not in the least puzzled by the conflicting definitions of atmospheric CO2 residence time. But evidently others are.

  26. DocMartyn says:

    Ferdinand Engelbeen; you either believe the data or you don’t. The sequestration of 14C is essentially first order and the atmospheric 14C is being exchanged with and diluted by a reservoir a least 20 times larger.

    The plot shows, the half life is unchanged across the 6 decades.

    Now you are free to believe whatever mechanism you like, you can invoke magic or authority or the ‘Revelle buffering’; however, you are quite simply wrong, and you are wrong because your description does not match reality.

    You are incapable of looking at the data and accepting what is shows in terms of kinetics and mass action.

    The truth is far simpler than you biased judgments.

  27. tchannon says:

    A very common law is rate = driving force / resistance
    Diffusion is like that.

    Looks to me that here there is a near perfect sink.

  28. tchannon says:

    Peter Shaw,

    If you use NIWA data beware it is not a proper time series. Either corruptions or something else. (check the sample points)

    I was ignoring the the burst of annual, lots of possibilities to do with global spread and irradiation in more than one location. A lot was written, too much.

  29. DocMartyn says:
    September 27, 2013 at 1:14 am

    I don’t think we differ so much in the basics: 14CO2 is removed in a much larger reservoir. We differ in opinion about the mechanism involved:

    Bomb spike 14CO2 in 1960 was already distributed over the fast reacting reservoirs: ocean surface and vegetation. These doen’t play much role in the sink rate of 14CO2 after 1960.

    Remains the deep oceans. Your idea is that these have a relative huge exchange with the atmosphere and redistribute the 14CO2 over the two reservoirs. My idea is that the exchange rate is rather modest, but that the return rate of 14CO2 is less than halve the peak bomb spike for the next ~1000 years. The net effect in both cases is exactly the same.
    My idea is based on the observed decay of 13CO2 in the atmosphere, which is caused by burning 13C depleted fossil fuel. But the observation shows only 1/3rd of the drop in δ13C as can be expected if there was no dilution by the deep ocean exchanges:

    Something similar happens with 14CO2. That makes that the decay rate of a 13CO2 spike or a 14CO2 spike doesn’t tell us what the decay rate is of a 12CO2 spike…

  30. Roger Andrews says:
    September 26, 2013 at 11:33 pm

    “I interpreted this, rightly or wrongly, to be when the 50th molecule in an emitted chain of 100 gets absorbed. This I could relate mathematically (see above) to the e-folding time, whereupon I had two robust and mathematically definable metrics I could use.”.

    Your e-fold time is right for the wrong decay…

    Let us try an example:

    You have a local bank in France with a lot of money going in and out during the day, mostly French euro’s. The cash flow is about 20% of the deposit of the bank per day, thus a residence time of ~5 days. The bank makes a small gain per year. Any extra gain is used for investments, in ratio with the gain.

    One day someone comes in and deposits a lot of German euro’s, about 20% above the total deposit.
    What you have calculated is the e-fold time of these German euro’s midst the bulk of the French euro’s. Because of the huge turnover, the e-fold time for the German euro’s in the bank in the bulk French euro’s is quite short: a few days..

    Because the total deposit increased with 20%, the bank start to invest in shares in a ratio of half the extra income per year. The first year that means 10% above the total deposit, the second year 5%, etc…
    The e-fold time of investments above the normal gain in this case is 20%/10%/year = 2 years. Even so, the German euro’s within the bulk of French euro’s disappeared with the short e-fold time of a few days…

    Thus two quite different e-fold times in play which have no connection with each other…

  31. DocMartyn says:

    “My idea is based on the observed decay of 13CO2 in the atmosphere, which is caused by burning 13C depleted fossil fuel”

    Did you by any chance examine the 13/12C ratio of limestones and coals, by global source, and then examine the shift in production of cement, iron, coal, oil and methane since the 1960’s?

    You see I had a look at limestone and coals, I found that there are huge differences in the 13/12C ratio in the limestones used to manufacture both cement and Pig-Iron, and also in coals.
    We don’t know which coals and which limestones have contributed to the 13C in the atmosphere, so we cannot estimate the size of the atmospheric reservoirs dilution, based on a fixed 13/12 ratio of the addition of fossil fuel/limestone based CO2. Remember that limestone quarries have different 13/12 ratios at different bed depths, to early layers can have quite different rations than higher ones; so the same quarry can have a different annual out of stone, depending where they are blasting.

    You may have much better sources for the annual addition of different 13/12 ratios of carbon since the 60’s. If you have please do share them.

  32. Roger Andrews says:


    “Thus two quite different e-fold times in play which have no connection with each other…”

    Well, that may be true of French and German Euros, but all I used e-folding times for was to find out which e-folding time for the emitted CO2 gave me the best fit to observed CO2 in a simple mass-balance model. (It turned out to be about 50 years, but as others have pointed out this result isn’t statistically diagnostic.) And the model considered only CO2. It didn’t distinguish between French CO2, German CO2, anthropogenic CO2 or natural CO2, or between 12C, 13C or 14C.

    Now there could indeed be some differences between the e-folding times for different varieties of CO2, but I submit that a much larger uncertainty is the IPCC’s ~5 year estimate of CO2 residence time in the atmosphere, which according to you “nobody disputes”. Well, I’m going to dispute it. I say that the IPCC’s estimate of annual carbon exchange between the atmosphere and land/ocean sinks is subject to so much uncertainty as to render its residence time estimate meaningless. If you can present any hard numbers to prove that this isn’t the case and that the IPCC estimate is robust after all I would be delighted to see them, because I haven’t been able to find any.

  33. Roger:

    The residence time of any CO2 molecule in the atmosphere is ~5 years, or somewhere between 2 and 10 years, based on different estimates. The isotopic composition of the atmosphere shows ~5.3 years:
    That means that about 20% of all CO2 is exchanged with CO2 from other reservoirs (the German euro’s) within a year. But also that any human contribution is rapidely exchanged for “natural” CO2 from other reservoirs.

    Nobody disputes the shortness of the residence time of CO2, whatever it is between 2 years and 10 years. Because the residence time of an individual molecule is not of the slightest interest for what happens with an injection of extra CO2 above equilibrium (the additional amount of German or other euro’s above normal deposit triggering more investments). That has no connection with the residence time above and has its own decay rate, which indeed is over 50 years e-fold time and practically independent of the isotopic (euro origin) composition.

    That was already discussed many years ago by Peter Dietze at the blog of the late John Daly:

  34. DocMartyn says:
    September 27, 2013 at 7:20 pm

    I am well aware of the differences in isotopic composition for different fossil fuels and carbonates. The average d13C level of the combination of fossil fuels use and cement factories is what is calculated by different organisations and may be interpreted with large margins of error.

    But that doesn’t change the fact that as long as the combined addition has a different δ13C than the (pre-industrial) atmosphere, the ocean circulation will change the difference between what goes into the oceans as 13CO2 (and 14CO2) and what returns out of the deep oceans more than for 12CO2, thus shortening a 13CO2 (or 14CO2) peak decay compared to a 12CO2 peak decay.

    That is the imporant consequence of the disconnection between the deep ocean inputs and outputs.

  35. DocMartyn says:

    “Because the residence time of an individual molecule is not of the slightest interest for what happens with an injection of extra CO2 above equilibrium”


    Take a look at the profiles of O2/DOC and DIC. Note that it is quite clear that life plus sunlight converts CO2 into organic carbon.
    Now the steady state level of living carbon in the oceans is about 2.5 GtC, but the throughput of carbon through the marine biosphere is 90 GtC.
    The 90 GtC ends up as respired CO2, and also as ‘marine snow’, excreata and dead bodies. Marine snow begins its fall to the bottom, being oxidized on its way down. Aerobes consume the O2, which is why oxygen is not in EQUILIBRIUM from the surface to the bottom. The oxygen is consumed, converting DOC into DIC. The marine snow falls below the hypoxic zone and is converted to CO2 and also to CH4.


    Methane is a lovely food and as soon as you have the combination of methane and oxygen, it is consumed. However, you can seen that, from 4 km up, DOC is converted to CH4, and this very light gas makes its way to the surface, but is intercepted by methane oxidizing organisms.

    Now, neither O2, CH4, CO2 or any-other biotic gas is in equilibrium in the ocean, w.r.t. depth. The biotic activity completely swamps chemical processes. Note also that the surface of the ocean, especially the top 5m, is denuded of DIC/CO2. The CO2 is being fixed and replenished from DIC/CO2 fluxes from the lower depths and from the atmosphere.

    Why do you think that atmospheric CO2 is in ‘equilibrium’ with the top of the ocean, when no other biotic metabolite is?

  36. DocMartyn says:

    Ferdinand, I must admit that I do not understand your 13/12C point at all. I am not trying to cause an argument here, but I was trying to model changes in the 13/12C in the atmospheric/aquatic bulk phases. I assumed, a prior, that the residency time of CO2 was at least a decade and using 1960 as a start date, my simulations showed I could get any line-shape I wanted if just limestone quarried in the US and China had different, extreme, 13/12C ratios.
    The shift in cement/pig-iron production is so huge that I decided it couldn’t be done, given the range of 13/12 ratios in the literature.

  37. Roger Andrews says:


    Thank you for the links. I’m encouraged to see that Dietze – in 1997 I believe? – came up with a 55-year e-folding time. This compares very closely with my best-fit 2012 estimate, which I stated above was “about 50 years” but which was actually 54 years.

    Anyway, the Dietze paper gives me the opportunity to put my problem with the data in a specific context. The arrows in Dietze’s Figure 2 show a net exchange of about 220 GtC/yr between the atmosphere and land/ocean sinks. This number is very similar to the IPCC’s number, which first appeared in the 2001 TAR and which doesn’t seem to have changed since.

    My simple question is, where did the number come from? And how good is it?

  38. DocMartyn,

    I have not the slightest problem with huge differences in local CO2, including the upper few meters of the oceans, thanks to the huge production rate of the biological world. That makes that local CO2 levels may change a lot diurnal and seasonal and with more or less upwelling from the deep oceans or deposit to the deep oceans,

    But I don’t see the relevance for the difference between residence time and excess mass decay time.

    About the 14CO2 bomb spike: more is known of the distribution, because it is a nice tracer to follow the ocean currents and other stuf:
    from slide 13 on.

  39. Roger Andrews says:
    September 27, 2013 at 10:17 pm

    “My simple question is, where did the number come from? And how good is it?”

    The number is based on seasonal changes in temperature, CO2, δ13C and δO2.

    The latter two give how much the total fluxes in and out vegetation are over the seasons. The difference between the calculated (~30 ppmv) and what is measured (~5 ppmv) is what the oceans absorb with lower seasonal temperatures and release with higher seasonal temperatures, countercurrent with vegetation. In addition, about halve of the CO2 release from vegetation (60 GtC – 30 ppmv) seems not to reach the bulk of the atmosphere but is readily absorbed by local vegetation. The same for the oceans, where the opposite happens: there is a continuous bulk transfer of CO2 between the oceans in the tropics and near the poles, which returns in upwellings near the equator some 1000 years later. The atmospheric transfer is calculated from the pCO2 differences ocean-atmosphere and wind speed (which increases the CO2 transfer rate ocean-atmosphere).

    Error bars are huge in all counts (from halve to double if you want), but an exchange rate of ~150 GtC/year or 5.3 years residence time seems to fit in the long range of estimates…

    But all by all, that doesn’t influence the removal rate of some extra CO2 in the atmosphere, as that is based on the (measured) increase in pressure in the atmosphere above the equilibrium setpoint (derived from ice cores) for the current temperature and the (measured) sink rate of CO2 under that pressure difference.

  40. DocMartyn,

    I suppose that most of the variability of the 13C/12C ratio is leveled off on global scale, but on local scale that may give huge differences. And there may be huge shifts over longer time scales: exhausting sources, new sources, the shift from coal to natural gas, etc…

    Here some interesting report about local/regional shifts in 13C/12C ratio as result of the change in fuels used over the seasons:

    Click to access bush07geo.pdf

    “Carbon dioxide emissions from fossil fuel combustion have been estimated with energy use statistics and combined with average global values of the isotopic composition of fossil fuels to estimate the isotopic composition of emissions (Tans, 1981; Andres et al., 2000).”
    But it seems that this need to get more detailed for the exact sources and their isotope ratio’s.

  41. Roger Andrews says:


    Thanks for your response. I was hoping that there might have been a “landmark” paper that explained how the numbers were derived hidden away somewhere but I guess there isn’t one.

    You mention error bars of from half to double. I did some work on this and found that that’s about what you get when you use the optimistic one-sigma errors the IPCC lists in the AR4 carbon budget (emissions = 7.2 +/- 0.3 GtC/yr, absorbed by ocean sinks 2.2 +/- 0.5, absorbed by land sinks 0.9 +/- 0.6, remaining in atmosphere 4.1 +/- 0.1). If you use two sigma-errors you can make ocean absorption range anywhere from 3.4 down to 1.0 Gt/yr and land absorption anywhere from 2.1 down to minus 0.3 Gt/yr.

    And if that sounds extreme, consider this. According to the IPCC there’s a net carbon flux of 0.9 GtC/year from the atmosphere to land sinks. According to the Integrated Carbon Observing System the number is 2.8 GtC/year. Why the difference? Because ICOS identifies 1.6 Gt/yr of emissions from “land use changes” that the IPCC says doesn’t exist, and the ICOS carbon cycle models say that this extra 1.6 Gt/y all goes into land sinks. So right there we have a difference of a factor of three and we haven’t even begun to talk about error bars.

  42. Roger

    Agreed, error bars are huge… In my calculations I never use land use changes, as that is a very triggy business. At one side that adds to the human contribution, but it gives an enormous increase in the extra uptake of land vegetation (sea vegetation is hardly influenced, as CO2 is not the limiting factor there). All we know for sure is the overall net sink rate of the extra CO2 in the atmosphere, hardly where it sinks… That is about 55% of the human emissions (as mass, not origin) if one excludes land use changes or 45% if you include them…

  43. Roger Andrews says:


    So where does all this leave us? I think here:

    1. The only carbon cycle variables we can measure with reasonable accuracy are a) the increase in atmospheric carbon content (from CO2 records), b) the size of the atmospheric “reservoir” (ditto) and c) carbon emissions from fossil-fuel burning (indirectly from fuel consumption etc. data).

    2. We have no definitive measurements of any of the other variables we need to balance the carbon cycle, including how much CO2 is absorbed by land and ocean sinks, how much has been emitted by land-use changes (mostly deforestation) and also how much has been emitted by biomass burning (cooking fires etc. – another potentially large contributor that generally gets ignored). It seems that we’re still not even sure we’ve identified all the main carbon sinks (what about soils?) The capacities of the oceanic and terrestrial biosphere “reservoirs” also seem to be based largely on assumptions and/or guesswork, although feel free to correct me if I’m wrong.

    3. Under these circumstances any estimate of atmospheric CO2 “residence time” based on reservoir size/exchange rate is, as I said earlier, clearly meaningless. (And even if the numbers were right it would still probably be meaningless because this isn’t the way CO2 gets exchanged in the real world, but I’m not going to get into that here.)

    4. E-folding times can be estimated from observational data without making any assumptions regarding the carbon cycle. Exponential decay is also probably the way CO2 gets removed from the atmosphere in the real world. The only source of uncertainty is emissions – we will get shorter e-folding times if we include deforestation and biomass burning. We will also find that carbon sinks absorb maybe two-thirds of total emissions rather than ~55 percent if we include them.

    5. The problem with e-folding times, however, is that many people aren’t familiar with the term. For this reason I would suggest we use “half life” instead.

    6. The original purpose of this post, which seems to have gotten lost in the shuffle, was to see how the 10-15 year e-folding times measured from bomb-test 14C decay figured into the equation. Given that 14C accounts for only one in one trillion carbon atoms and that the decay rates of 12CO2, 13CO2 and 14CO2 may be different anyway it’s difficult to see how we can deduce very much from this result. But if we estimate decay rates from all of the CO2 in the atmosphere, 99% of which is 12CO2, the problem goes away.

  44. Roger,

    Agreed with all points… Only an addition to point 6:

    The IPCC’s long decay times, where a part of CO2 resides (near) forever in the atmosphere is based on the Bern model, which is certainly wrong for the total emissions up to now. While the decay time of 14C is too short, there are no signs that the 50-55 years calculated by Peter Dietze in 1997 is increasing, to the contrary. Thus as long as the deep oceans and vegetation (the more permanent storage in peat, browncoal,…) goes on unlimited, there is no reason why that mechanism would be limited in time. Thus Tim Channon is completely right that the Bern models fails to represent current reality.

    The only way the Bern model may have merit is if we burn all available oil and gas and lots of coal, 10-20 times more than we have done up to today. Then the deep oceans may become saturated and only the slow(er) reactions of carbonate deposits may go on with longer decay times… But that is not the case in the foreseeable future.

  45. Roger Andrews says:

    Ferdinand: Thanks for you comments on all this. They’ve helped me get my ideas together. I’m now in the process of reviewing observational data (CO2, O2, NDVI, SST, emissions etc.) to see if there’s anything they can tell us about how the carbon cycle really works. Initial results are intriguing but it’s too early to draw any firm conclusions. Maybe more later.

    Stay tuned 🙂

  46. Roger Andrews says:

    Ferdinand: One more question if you’re still around. Fig. 7.3 of the IPCC AR4 shows natural exchanges of 70 GtC/yr between the atmosphere and ocean and 120 GtC/yr between the atmosphere and land sinks and sources. Do you know where these two numbers came from and what errors they might be subject to?

  47. tchannon says:

    I’ve been keeping out of this, different subject.

    As time permits some tricky detail investigation is going on here where there are novel technical problems to resolve. Fun part is I have no idea what is going to emerge.

    Incidentally, the time constant value means the cutoff is about 110 years if that is what is going but like the other kind of irradiance there are mind twisting oddities which make me reluctant to say much until I know more clearly.

  48. Roger,

    Their own estimate is +/- 20%, but -10 to- 50% would be more appropriate I suppose. I use the NASA figures, these are without land use changes:
    In both cases halve of the respiration CO2 from vegetation and soils doesn’t even reach the bulk atmosphere, thus doesn’t influence the residence time. If one includes the full movement of vegetation CO2, that would shorten the residence time to ~3 years, which seems too short.

    One point where both differ is the human contribution in the sea surface exchanges. The IPCC graph shows an extra CO2 sink of 22.2 GtC/yr and an extra CO2 source of 20 GtC/yr from the ocean surface caused by humans, but that can’t be true: any increase in the atmosphere causes an increase in flux into the cold polar oceans and a DEcrease in CO2 release from the equatorial oceans. Thus with a net sink of 2.2 GtC/yr, about halve is from more sink and halve is from less release. There are no increases in oceanic CO2 releases due to human intervention.

  49. Roger Andrews says:

    Hi Ferdinand

    The NASA cartoon shows 210 GtC getting exchanged between the atmosphere and land/ocean sinks/sources every year, but like everyone else NASA doesn’t say where these numbers come from. And a little farther on NASA contradicts itself by stating that only 10-100 GtC in fact gets exchanged:

    “the fast carbon cycle moves 10^16 to 10^17 grams of carbon per year”.

    And below that it says sorry, the number is actually 1-100 GtC:

    “Between 10^15 and 10^17 grams (1,000 to 100,000 million metric tons) of carbon move through the fast carbon cycle every year.”

    Saints preserve us.

    We can in fact make a reasonably good estimate of how much carbon gets exchanged from the seasonal CO2 cycle, which is dominated by the growth and decay of NH vegetation. Globally this cycle has a mean amplitude of about 6 ppm CO2, representing an exchange of only 13 GtC/yr.

  50. Roger,

    Be careful, the global cycle indeed is about 6 ppmv, but that is the result of two opposite fluxes: oceans and vegetation. NH vegetation wins the battle, mainly because there is abundant land and vegetation in the NH mid to higher latitudes which restart life in spring and ends life in fall. The NH doesn’t show much amplitude over the seasons…

    As far as I remember, they have based their figures on oxygen en d13C balances, which are mainly from vegetation, while oceans give a little difference in oxygen with temperature and some change in d13C in opposite direction. DIC measurements in the seawater surface also show a huge change over the seasons.

    There are some hints of the carbon cycle here:

    Click to access BenderGBC2005.pdf

    but they don’t think that there is a huge seasonal source/sink of CO2 from the ocean (surface)…

  51. Of course the
    “The NH doesn’t show much amplitude over the seasons”
    must be the SH…

  52. Roger Andrews says:

    “Be careful, the global cycle indeed is about 6 ppmv, but that is the result of two opposite fluxes: oceans and vegetation.”

    There seem to be a question as to exactly which way CO2 moves across the atmosphere/ocean interface. There are, for example, two ways of interpreting the HOT Aloha results (data at link below); either increased atmospheric CO2 is causing a transfer of CO2 from the atmosphere to the ocean or it’s causing less CO2 to be transferred from the ocean to the atmosphere. But the question is largely academic because observations strongly suggest that changes in the ocean don’t even generate a significant seasonal atmospheric CO2 cycle. The cycle is dominated by NH vegetation, as you note.

    The one thing that might be wrong with my 13 GtC estimate is that it’s derived from fluxes averaged over large areas and month-long time periods. We would certainly get much larger fluxes if we went down to the microscale and counted molecules exchanged per nanosecond, which can presumably go in both directions at the same time (a CO2 molecule about to get photosynthetically absorbed by a tiny new grass shoot passes a CO2 molecule that’s just been released by last year’s decaying grass, and they wave to each other as they pass and say “see you next year in the troposphere”.) But how do you measure something like that? And even if you could, what would it mean?

    I looked at the Bowdoin paper you linked to and it was the same-old-same-old. Annual CO2 flux balances similar to those published by IPCC, NASA etc. and no mention of the ~200 Gt that supposedly gets exchanged each year during the carbon cycle. I’m beginning to think this number was put together by people spinning roulette wheels in smoke-filled rooms, and the smoke wasn’t tobacco smoke. 😉

  53. tchannon says:

    Monday, September 30, 2013
    Mathematical & observational proof that CO2 has no significant effect on climate
    A recent comment at WUWT sets out the mathematical and observational proof that the effect of CO2 on climate is effectively nil.

    The comment is a succinct summary of many recent posts here on the Hockey Schtick.

    Bart says:
    September 28, 2013

  54. Roger Andrews says:


    Getting way O/T here, but your comment gives me an opportunity to ask a question that’s been bothering me for some time.

    There are some scientists who have forgotten more about atmospheric physics than I’m ever going to know and whose judgement I respect, for example Richard Lindzen. He agrees that CO2 does have an impact on temperature although not necessarily a large one. Now along comes this guy who claims to have proved that it has basically no effect at all. Why should I believe him and not Lindzen?

  55. kuhnkat,

    I have replied at JoNova’s site that the first 1000 meters over land are worthless for “background” CO2 levels, historical as well as modern. 95% of the atmosphere can be represented with one station, no matter if that is at the South Pole, Mauna Loa or near the North Pole (Barrow, AK, USA)…

  56. kuhnkat says:


    yes I have read much of what you have written on the popular sceptic sites. Unfortunately you are less than persuasive with your shaping of the data and interactions.

    CO2, as a Green House Gas, plant fertilizer, or just mass, is not limited as to where it functions.

  57. tchannon,

    I have had many discussions with Bart on the topic of the cause of the increase of CO2. He is a brilliant theoritical guy, but has no idea how nature works. As many before him, he thinks that curve fitting equals cause and effect and then ignores all what proves that he is wrong.

    In this case he is already wrong from the beginning:

    “If we now presume that there is a positive feedback between CO2 and temperature, we get a positive feedback loop, which would be unstable.”

    If the positive feedback is modest, then the system is not unstable, only gives an extra increase of temperature and CO2 levels with (fb) vs. without feedback (nofb) of CO2 on temperature:

    The rest can be read at WUWT:

  58. kuhnkat says:


    Who matches the observations better??

    Oh wait, it is really difficult to figure out what is really happening with CO2 due to the multiple cross effects from so many directions.

    My ignorance has me looking at the current lack of warming for 17 years, and probably longer if we had an honest temperature record instead of the mess we have. Based on no warming, the fact that the best effort models (are they really best effort for truth or agenda??) simply are not worth what has been spent on them. The fact that Lindzen was indoctrinated with so many others in his education, I see no reason to believe anything other than he was trying to keep from being excommunicated with his acceptance of GHG warming!!

    In fact, we can accept a reasonable amount of GHG warming and still have no issue with pumping gigatons more into the atmosphere every year, INCLUDING Fingelbeans beliefs on the sources, sinks, and lifetime. There is no evidence the SYSTEM cannot deal with even more energy than what is hypothesized and some evidence that it can.

    For instance, the IPCC types claim that they have viewed localized hot spots. Their theory requires a positive feedback for the hotspot generation and later problems. For the Hot Spots and Strat cooling to happen and then go away would be a strong indicator to me that the system adjusts to a higher efficiency of moving the heat off the planet preventing the positive feedbacks from continuing.

    Without a new hypothesis of positive feedbacks fitting the physics I see no reason to continue this scam. Learning the intricacies of the system is useful, but, not an emergency.

  59. Roger Andrews says:


    You ask “Who matches the observations better??”

    Well, the best match to observations I’ve seen was produced last year by Tallbloke, but he needed a significant contribution from CO2 to get it:

    “Lindzen was indoctrinated with so many others in his education …. he was trying to keep from being excommunicated with his acceptance of GHG warming!!”

    Well, let’s see how many others were indoctrinated during their education to accept at least some GHG warming, apart from Tallbloke, that is. We have Curry, We have Spencer. We have Christy, Pielke and Scafetta. We have Watts. We have McIntyre and McKittrick. We have Singer, Michaels, even the late John Daly. And we have – I can hardly believe this myself …… we have …… Christopher Monckton!! Whoda thunk the warmists could ever have indoctrinated him?

  60. Roger Andrews says:


    You say: “95% of the atmosphere can be represented with one (CO2) station, no matter if that is at the South Pole, Mauna Loa or near the North Pole.”

    I’m not sure that we can put absolute percentages on it, but the CO2 records are indeed remarkable in their consistency from place to place. Overall I would rate them as among the most reliable in all of climate science.

    And if, as some seem to believe, this is because the records have been fudged then we’re looking at the most successful international conspiracy in history, one that’s been going on for over 50 years and which has involved uninterrupted plotting by scientists and technicians in the US, the UK, Russia and the old USSR, Iceland, France, Germany, Italy, Hungary, China, Japan, Australia, New Zealand, South Africa and a couple of dozen smaller countries too.

  61. tchannon says:

    Looks like I have cracked directly filtering irregularly sampled data. How far this can be pushed remains to be seen. Initial result is surprisingly good.

  62. tchannon says:


    Lets see your stable positive feedback.

  63. tchannon,

    Simple control theory: as long as the total gain from the primary driver (temperature) and the secondary feedback (CO2) is lower than unity, there is no runaway of the total process. In practice, there is no runaway if the total gain is less than ~0.8.

    From Wiki (

    “If the functions A and B are linear and AB is smaller than unity, then the overall system gain from the input to output is finite, but can be very large as AB approaches unity. In that case, it can be shown that the overall or “closed loop” gain from input to output is:
    G = A/(1-AB)
    When AB > 1, the system is unstable, so does not have a well-defined gain; the gain may be called infinite.”

  64. tchannon,

    That doesn’t imply that the response of temperature to increased CO2 levels is huge (which I don’t think at all), but one can’t exclude that there is a positive feedback, because CO2 lags temperature on near all scales.

  65. kuhnkat says:


    thanks for making assumptions as to what I meant by GHG’s causing warming. I would mention that I am not a denier in the respect that I deny radiative transfer. I deny that it makes a significant difference in weather or climate at this time. If we accept the belief that it is an exponential curve then a doubling of CO2 will only make about 1C difference. In other words from 400 to 800ppm, however long that actually takes will be about 1c due to the GHG. That is miniscule and easily lost in our turbulent climate. The doubling from 800-1600ppm is a little more mythical.

    So, yes, I tend to be a little more extreme than most of the people who have had more conventional education and often wonder why they waste their time. Of course, if one wishes to stay within polite society one MUST believe in the warming of GHG’s!!


  66. kuhnkat says:


    the positive feedback could definitely be there as I am assured that it is. In fact, a warmist over at Curry’s presented me with a wonderful “proof” of how small the water vapor feedback is!! My suggestion had originally been that if CO2 could cause problems due to a water vapor feedback then water vapor would already be causing the problem.

    Maybe you can explain to me in a way I will understand how CO2 can cause a more dangerous feedback than water vapor itself. Yes I understand that water vapor precipitates out. It doesn’t matter whether CO2 or Water Vapor provides the energy that causes it to evaporate in the first place though, water vapor precipitates. As I reminded that same genius, the amount of GHG in the atmosphere is what does or does not cause a warming, not some future possible level due to lifetime.

    My arm waving self tends to think that the reason Trenberth and others have seen the Hotspot, but, that it was temporary, is due to the fact that the positive feedbacks are small and induces changes in the atmosphere which allows the effects to dissipate. It may be that their issue is simply how rough the models are compared to analog reality.

    I say they should believe their observations when the Hotspot dissipates. A warming atmosphere increases the space between molecules meaning that the radiation from lower altitudes will NOT be restricted as much as they think due to the increased amount of GHG’s and that the temps at the same altitudes will be warmer so the average radiation levels will NOT be as cold as computed.

    Yup, a tiny error in physics emulation algorithms could be the whole issue here, yet, what is being done to open the modeling so it can be verified that it is not??? Then again it could be a tiny error in the math that has been accepted as correctly representing the physics. So many possible errors that could be causing this issue and no one willing to work themselves out of a job!!


  67. kuhnkat says:


    “You ask “Who matches the observations better??”

    Well, the best match to observations I’ve seen was produced last year by Tallbloke, but he needed a significant contribution from CO2 to get it:

    No, that is what you consider to be the best match to observations. Matching ONE item in the array of components of the climate is a game that the IPCC has been fobbing us off with. They could probably match the atmospheric temps almost exactly if they weren’t trying to push an agenda and if the temps weren’t being adjusted to make them easier to match!!

    Oh, and using CO2 as a knob to MATCH something is funny!! Reminds me of the IPCC type morons who told us that they couldn’t match the temp rise without CO2 in their models. Their “proof” was that their models will not warm enough with their CO2 knob turned down. Does anyone else see the circularity in that argument???

  68. kuhnkat says:


    “I’m not sure that we can put absolute percentages on it, but the CO2 records are indeed remarkable in their consistency from place to place.”

    If the temperature stations were sited and maintained with the effort that the CO2 stations are we would have a much more believable record.

    CO2 is measured where there is little anthropogenic disturbance as opposed to temps that are measured with a LARGE anthro disturbance!!

    The accepted standards for CO2 also reject readings outside of acceptable ranges. It helps in making that beautifully consistent record.

  69. kuhnkat:

    “The accepted standards for CO2 also reject readings outside of acceptable ranges. It helps in making that beautifully consistent record.”

    Hardly makes any difference if you include or exclude the outliers. Here a plot of Mauna Loa and the South Pole showing raw hourly data, “cleaned” daily and monthly data:

    Look at the scale if you want to see how “huge” the outliers at Mauna Loa are (volcanic vents and vegetation in the valleys, depending of wind direction).

    And I agree that any effect of more CO2 at maximum is modest and even beneficial…

  70. kuhnkat says:


    “Hardly makes any difference if you include or exclude the outliers. Here a plot of Mauna Loa and the South Pole showing raw hourly data, “cleaned” daily and monthly data:”

    A few years ago there was an uproar when the monthly ML came out with a SIGNIFICANT DROP in CO2. After WUWT contacted them they pulled the data point. Eventually they explained that it was caused by throwing out so much data for the month that the few lower readings that were left was all that got plotted. Oh, and they then started infilling also.

    Excuse me if I don’t accept your glib assurances about CO2 levels from a remote area in the Pacific high on the side of an active volcano where actual CO2 in the air causes spikes that are rejected.

    Like the temp record, averaging distorts the record. A couple of years ago the Japanese improved their algorithms and put out a great video of the biosphere breathing. Recently NASA put out a similar video. It showed CO2 levels varying up to 20ppm over short distances. Arboreal Canada was the strongest I believe. This is not a local Urban issue and does not show on the ML or other records. I won’t say the official record is useless, but, it is limited by the desire to show this mythical “background” level.

  71. kuhnkat,

    If you have read the follow-up story at WUWT about the missing data: that was caused by a hard drive crash. The drop in the monthly data was because the average was of only 10 days at the end of the month in a seasonal decreasing trend.

    No data are thrown out from the raw (hourly averaged) database. That is what I have plotted and that data can be found at:

    Some of the data are not used for daily, monthly and yearly averages. The rules for excluding such outliers in the averages are very strict:
    But even if you include all outliers, that doesn’t change the yearly average or trend with more than 0.1 ppmv.

    The largest differences are from seasonal changes (+/- 8 ppmv at Barrow, +/- 1 ppmv in the SH), which exchange somewhat 20% of all CO2 in the atmosphere. That gives a global variation of only +/- 1% of the full scale of CO2 in the atmosphere and there is a 1-2 years lag between the NH and the SH, which proves that the main source is in the NH, despite a larger sink capacity by vegetation:

    The Japanese satelliet data only show the CO2 fluxes, unfortunately not absolute values. Near ground over land in forests and other places with lots of sources and sinks you can find any level of CO2 you want. Dig a hole in the ground and it is filled with over 1000 ppmv in short time on a wind silent day…

    But that doesn’t matter for the radiation balance: Even if the first 1000 meter is at 1000 ppmv, that has little effect on temperature (without feedbacks either way). It is the full 70 km column that matters, as far that there is an effect (with feedbacks).

    All I can say is that we only can hope that one day the temperature record would be as rigourisly quality controlled as the CO2 record…

  72. Roger Andrews says:

    Tim, if you’re still around:

    I was going to ask Ferdinand this question – and if you want to respond Ferdi, please feel free to do so – but then it occurred to me that it might be more up your street.

    I’ve been using ppm changes in atmospheric CO2 to estimate seasonal carbon exchange between the atmosphere and global carbon sinks/sources and have come up with hugely different numbers depending on the measurement interval:

    One month CO2 measurement interval: ~25 GtC/yr (I forgot to count the carbon coming back out again in the 13 GtC/yr estimate I posted earlier).

    One-day CO2 measurement interval: ~150 GtC/yr

    One-hour CO2 measurement interval: ~1,500 GtC/yr.

    The reason for these differences is that going to a one-day time interval adds a lot of day-to-day carbon flux exchanges that get cancelled out in the monthly averages while going to a one-hour time interval adds a lot of hourly carbon flux exchanges that get cancelled out in the daily averages.

    The question is, is there any analytically objective way of determining the time interval to use in a case like this? Or am I stuck with a range of uncertainty of a factor of sixty?

  73. Roger,

    I am afraid that looking at too short time intervals gives you the answer for the short time intervals: mostly noise. It is like estimating the sea level changes by extrapolating the movements of waves (seconds) via tides (hours) to sea level (at least 25 years of average tide movements needed).

    Even the monthly intervals may give you a false impression of the total CO2 exchange to the underestimating side: the permanent flow of CO2 from the equatorial deep ocean upwelling waters to the polar sink places doesn’t change the amounts in the atmosphere, but does add to the exchange rates…

    Thus there is no simple way to know the real exchanges…

  74. Roger Andrews says:


    In cases like this I normally make my points by attaching graphs, but my computer is still in hospital and the standby machine I’m using doesn’t do graphics, so I’m going to have to ask you to do it for me.

    Go to the ESRL website at the link below, download the Point Barrow daily CO2 readings for 2012 (line 30) and plot them up:

    Now compare the plot with the Point Barrow snow cover data for 2012:

    You will notice that the daily CO2 record is a lot more erratic when there’s no snow cover. Day-to-day CO2 changes in fact show about four times as much variance when there’s no snow on the ground as when there is. Clearly this isn’t a result of “noise”.

    A specific example is the ~4 ppm CO2 spike in mid-May, which coincides to the day with the beginning of the spring thaw. What caused it? My best guess is a pulse of CO2 released by just-uncovered decaying vegetation left over from 2011, but whatever the cause it wasn’t noise. If we assume that this pulse was representative of the Alaskan North Slope – an area of about 250,000 sq km – then all by itself it caused the release of about 16 million tonnes of CO2, and there are a lot of other cases like it. And all of them get cancelled out at the monthly time scale.

    And when you plot hourly CO2 data for May 2012 (data at line 30 on ESRL) you find that the spike amplitude doubles to 8ppm, showing that we get strong cancellation effects even at the daily time scale.

    So I’m afraid I have to disagree with you when you say; “looking at too short time intervals gives you the answer for the short time intervals: mostly noise.” By not looking at shorter time intervals you are in fact throwing out most of your signal.

  75. Roger, I have plotted the 2012 daily averages of Barrow here:

    You need to know the circumstances at Barrow: it is a coastal station, where only measurements are taken into account that are deemed “background”, that is when the wind is blowing from the ocean side. Only these are used for (daily) averages. The number of hours taken into account for the daily averages are given in the 7th column.

    All hourly data are retained, thus that includes data with wind from land side with a lot of snow for half a year and tundra the other halve…

    The problem is in the definition of “noise”: for global averages one want the averages of “background’ CO2, which is over the oceans and above 1000 meters over land. Even the tundra – and in general all extratropical vegetation uptake/release of CO2 – is readily mixed in within a few days within one latitude/altitude band. But one should not use data directly influenced by the same tundra, as that shows much more variation that isn’t mixed in yet in the bulk of the atmosphere…

    Thus the CO2 movements over the tundra near Barrow is “noise” for the (sealevel latitudonal) measurements there. Even with careful looking at wind direction some influence of the tundra around Barrow may get into the measurements, but that is not representative for most of the atmosphere in that band.

  76. Roger Andrews says:


    “You need to know the circumstances at Barrow: it is a coastal station, where only measurements are taken into account that are deemed “background”, that is when the wind is blowing from the ocean side.”

    And “for global averages one wants the averages of “background’ CO2, which is over the oceans and above 1000 meters over land”.

    We’re coming at this from completely different viewpoints. You see the CO2 records as a vehicle for estimating absolute “background” CO2 concentrations, i,e those in well-mixed air remote from terrestrial sources of contamination, and to make these estimates you have to remove the impacts of these sources of contamination, such as the CO2 blowing in from the tundra south of Barrow and the CO2 blowing upslope into Mauna Loa.

    I, on the other hand, see them as a vehicle for estimating the size of the seasonal carbon cycle, and to do this I don’t even have to know what the absolute CO2 values are. But I do need all of the available information on the seasonal cycle, which includes the CO2 blowing into Barrow from the south and the CO2 being carried into Mauna Loa by the upslope winds.

    As they say, one man’s noise is another’s signal.

    Hope that clarifies things.

  77. Roger,

    It is impossible to estimate the seasonal swings from local disturbances. One can only do that if “local” is from a (very) large area. That is the case for what is measured by tall towers (200 m height), which measure CO2 at different heights and may be representative for an area up to 1000 km in diameter, including the upgoing and downgoing fluxes of CO2. For any local point, that is impossible (including Mauna Loa, Barrow and other “baseline” stations). See the difference in diurnal CO2 levels between a modern station mid-land in Germany at Giessen, a semi-rural town with agriculture and stations like at Barrow, Mauna Loa and the South Pole (all raw data):

    The tall tower in Cabauw (The Netherlands) gives a nice overview of the change with height:

    The US maintains a large network of small and tall towers over most of the US and beyond to measure CO2 fluxes over different types of vegetation and over larger areas:

    But even so, not an easy task to know the real fluxes over the seasons…

  78. Roger Andrews says:


    I went through the old Giessen CO2 data a year or so ago when I was reviewing the “Beck curve” (which is a figment of Beck’s vivid imagination, by the way) but I didn’t know they were still taking measurements there. It’s interesting to see that current CO2 values are pretty much the same as what Kreutz measured there back in the early 1940s – that is about 50 ppm higher than everywhere else.

    But here’s the problem. If we assume the diurnal CO2 cycle at Giessen applies to an area of only one square kilometer it still represents the exchange of up to 3,000 tonnes of CO2 a day between the atmosphere and the land, and it’s reasonable to suppose that the much of the CO2 released by the land during the upslope of the diurnal cycle will get blown away and that much of the CO2 that gets reabsorbed during the downslope will not be the same CO2. So just in this tiny area the diurnal cycle could be causing an atmosphere-land exchange of hundreds of thousands of tonnes of CO2 a year, but we won’t detect it in monthly averages nor even in the daily averages.

    However, few records are as user-friendly as Giessen, and when you say “It is impossible to estimate the seasonal swings from local disturbances” I find it hard to disagree with you. Bottom line? We have no idea what the amplitude of the seasonal carbon cycle is, except that it’s probably very much larger than the 150-200 gigatonne estimates published by the IPCC and others.

    And how much discussion of this issue do we find in Chapter 6 of the IPCC AR5; “Carbon and Other Biogeochemical Cycles”, which is 169 pages long? None. Evidently it just isn’t considered important.

  79. Giessen indeed started again near 2 decades ago with modern measurements of CO2 and a lot of other stuff:
    Interesting that it is only a few km from the place where the historical measurements were taken, again at the edge of the village. The only difference is the number of cars since the 1940’s…
    The graph I sent was on a few summer days with inversion at night, that gives a huge diurnal difference with more wind, the difference is a lot smaller. But even so, there still is a local bias visible in the monthly averages (some months even extreme):

    And I agree with you about the seasonal swings. The only possibility is frequent scans with satellites with a high accuracy (both are lacking until now)…

    On the other hand, I suppose that the 150 GtC seasonal exchange is not far off, as much higher fluxes would lead to a much shorter residence time, for which is no empirical proof.

  80. Roger Andrews says:


    This discussion started with me questioning where the ~150 GtC seasonal exchange numbers we see on all the carbon cycle cartoons came from. I’m still trying to find out where but seem to have reached a dead end. Here’s a summary of progress, or lack of it:

    Figure 6.1 of the IPCC AR5 shows a seasonal flux exchange of 170GtC (60 ocean, 110 land) but doesn’t say where these numbers come from.

    Figure 7.3 of the AR4 shows higher seasonal flux exchanges of 190 GtC (70 ocean, 120 land) and claims that these numbers were “modified from Sarmiento and Gruber 2006”.

    S&G 2006 say that their fluxes (which were much lower than the IPCC’s at 130 GtC, 60 ocean, 70 land) came from Sarmiento & Gruber 2002.

    S&G 2002 say their numbers were “slight modifications of those published by the Intergovernmental Panel on Climate Change”, i.e the 2001 TAR numbers.

    The IPCC says that the TAR numbers (which were 60% higher than S&G 2002 at 210 GtC, 90 ocean, 120 land) came from Schimel, 1995.

    I’ve found one of Schimel’s figures and it shows 150 GtC total (60 land, 90 ocean), but the paper is behind a paywall and to judge from the abstract it wouldn’t tell me where these numbers came from anyway.

    Anything you can add?

  81. Roger,

    Oh, help! I give up…

  82. wayne says:

    Priceless! Sorry, just can’t help but laugh. 😆

    More seriously, we can finally see the gist of how IPCC comes up with some of its numbers. How many like that? Thanks Roger. Good investigating. You would never undercover that reading any single IPCC document.

  83. tchannon says:

    The stuff of sketches. Need a reference, look up what your Schimel gives as a figure, who actually copied his Schimel and what with typos, the fish that got away and Indian railway miscarriages, poor dear truth is toast.

  84. Roger Andrews says:

    Success! Well, sort of.

    I stumbled across a comparison of output from the “Carbones” and some other carbon cycle models at:

    The Carbones model shows a global carbon exchange ranging from ~10 gigatonnes per month at the trough of the seasonal cycle to ~17 gigatonnes/month at the peak. This represents an annual exchange of ~160 gigatonnes, which is ~ the number I’m looking for. So I’m going to assume that it was from a carbon cycle model like this, at some time in the distant and unrecorded past, that the chain of estimates that led to AR5 Figure 6.1 emerged.

    And is the Carbones model realistic? It shows a constant exchange of about 6.5 GtC/month, or 80 GtC/year, between the atmosphere and land/ocean sinks/sources in the tropics, and I don’t know of any observational data that we can check this number against. It also shows an 80 GtC/yr seasonal cycle, which converts into a ~40 ppm seasonal swing in atmospheric CO2, about seven times higher than what observations show. Could it be that the model is simulating some of the pesky diurnal and other short-term fluxes that get cancelled out in the monthly CO2 averages? No way of checking that either that I can see.

    Once more back to square one, dear friends. 😉

    I think I’m going to give the carbon cycle a break and move on to something which is a little less like trying to shovel a cloud into a bucket.

  85. tchannon says:

    RA. A climatic model is exactly what I expected as the “data” source.

    I also expect a large annual cycle.

    What might be going on is as follows.

    There is a long period low pass effect en-mass which computes to an attenuation of annual by 100x, an order larger than your 7x

    However, the larger attenuation requires an homogeneous atmosphere which I doubt is present, too many point effects causing localism.

    Whether this hangs together is I think open. It might be possible to extract more information from what data is available.

  86. Roger Andrews says:


    If you want to try extracting some information from a “global” data set the hourly records for Barrow, Mauna Loa, Tutuila and South Pole are at:

    These four records contain about 1.4 million hourly readings. Daily averages for these stations are also available, but these smooth out diurnal variations and include only hourly readings taken when the wind was blowing in the “right” direction, which limits their usefulness.

    The specific problem will be deciding which of the wiggles on the hourly records represent legitimate CO2 exchanges and which are “noise”. Maybe some spectral analysis might help.

  87. tchannon says:

    I was thinking of a different tack aimed at laws.

    Hourly MLO is well known here but only the original data because the messed with later stuff is broken.

    There is diurnal in MLO but it is very weak. Looked hard at this probably 2008 concluding it didn’t tell anything. As I recall all the other datasets are fragmented, not good. Almost as though nothing is ever let out which can be cross checked.

    I’ll put that one on look when a suitable moment appears.

  88. Roger,

    The problem with the Carbones and other models is what you count as “exchange”:
    The diurnal cycle may be huge, but if the CO2 released from the debirs on/in the soils is immediately absorbed by the canopy of the trees above it, it doesn’t reach the bulk of the atmosphere.

    There is also the continuous exchange between the equatorial upwelling and the polar sink places, which doesn’t reflect in any diurnal or seasonal measurements. My own estimate is that this deep ocean exchange is about 40 GtC/yr, based on the d13C decrease rate:

    I wonder what the seasonal swings in the tropics are? Temperatures in the tropics shift around the equator, but do they increase/decrease that much over the whole 30N-30S area?

    Takahashi did compile a lot of real world data (pCO2 difference air-ocean and wind speed), thus may be more realistic than models, but as “net” changes are given, the ocean exchanges are much larger than presented in the overview…

  89. Roger Andrews says:


    Now that my computer is finally fixed I can present graphs again, and here’s an interesting one:

    The American Samoa station is located on a hill at the extreme east end of the island and normally enjoys strong southeasterly sea breezes. When these breezes blow there’s no impact from the tropical vegetation that covers most of the island and the daily CO2 record is substantially flat. But every so often, such as between January 9th and 17th, something happens (probably a wind shift; daytime high temperatures were elevated over this period) and we see a diurnal variation that presumably originates from the vegetation.

    If this is what is happening – and there are no guarantees that it is – we can take a shot at estimating carbon exchange from the tropical forests of American Samoa during January 2012. The average amplitude of the diurnal cycle between January 9th and 17th was about 4.5 ppm, and if we assume that this is representative of the whole month the carbon exchange becomes 4.5 (ppm) times 31 (days) times 2.12 (GtC per global ppm CO2) times 100 sq km (area of forest on American Samoa) divided by 510 million sq km (area of Earth), which works out to 58,000 tonnes.

    And if we assume that January is representative of the whole of 2012 (I haven’t plotted the data so don’t know whether it is or not) multiplying 58,000 tonnes by 12 gives us 700,000 tonnes of total carbon exchanged in the year.

    And if we want to get really brave we can prorate American Samoa over the entire 15.5 million sq km of Earth covered by tropical forest. This gives us 700,000 tonnes divided by 100 multiplied by 15,500,000,which comes out to 109 gigatonnes. Considering the highly speculative assumptions and massive extrapolations used to obtain this number it’s remarkable that it should be so close to the 80 gigatonnes that gets exchanged in the tropics each year according to the Carbones model. But most likely it’s a fluke.

  90. KuhnKat says:


    I beg to differ with your story. Yes they very well may have made up a hard drive crash story later, BUT, why did they tell us they were going to start infilling instead of just throwing out data?? Also, are you telling me they never throw out data outside their standards??

    See, this is why I think you are all liars.

    Another reason is our previous discussions of glaciers.

    Hypsithermal Finglebean. The Antarctica may not have melted completely, but, there simply is NOT a continuous, readable record in the ice like y’all claim. The previous interglacial was allegedly even warmer than the current UNPRECEDENTED warm one. Again, how did the glaciers survive Finglebean!!! Remember that pole amplification?? 2 degrees at the equator means how many at the poles??? We allegedly have most of the world glaciers melting at current temps!!! What would happen if we had 500 years at 2c warmer than now???

    Ouch, those lies come back don’t they???

  91. KuhnKat says:


    the Japanese flick did show just fluxes. It was the first we saw how strong they were over short periods and regionalized. The Nasa data showed absolute values. By the way, I am suggesting the Average Trend would be HIGHER not lower and I need the definition of well mixed. Er, why would you consider regional seasonal variation outliers?!?!?!

    Actually where in the column the higher density occurs DOES matter as the bottom part warms and cools the Nobel Gases and the upper section does the radiating to space. If there could be a low density of GHG’s at the bottom it would slow the heating and cooling of the lower atmosphere slowing the convective cooling.

    Just reread you comment about no data being thrown out of the raw hourly averaged DB. Either you are trying to finesse it or you ARE lying. The raw data takes spikes from the volcano that are ridiculously large. Are you telling us that they do not throw that out?? I have read specific statements where they say they do.

  92. KK,

    If you have any real proof that people at NOAA are faking CO2 data, just to pleasure themselves or the IPCC, then please show that. The only possibility to influence the data over tens of stations in tens of countries with hundreds of people involved is by manipulating the calibration gases (calibration gases are made by NOAA for most of the world) with some 0.02 ppmv/day. Even that would be seen by Scripps, which still uses its own calibrations and takes samples at Mauna Loa, independent of NOAA. And they are still mad on NOAA, as Scripps delivered the calibration gases for a very long period before NOAA did take over, thus they would have a lot of fun taking NOAA to shame…

    BTW, here the raw data (where nothing is thrown out) and the “manipulated” data averages for Mauna Loa and the South Pole:

    As you can see, there are no raw data at MLO during near a month, simply because of equipment failure. In that case, the monthly average of the missing data is calculated as the average curve of previous years + the increase measured in the other months with sufficient data.
    If you think that “missing data” are thrown away, because of extra CO2 from the volcano, then simply compare these data with the trend at the South Pole or any of the other stations.

    Again, no data are thrown out from the raw hourly average (calculated from 2×20 minutes of 10-second voltage snapshots and 3×2 minutes of calibration gases snapshots fo the NDIR instrument). All the hourly averages are available, including the variability of the snapshots. If the variability is large, then that datapoint is not used for daily, monthly or yearly averages. That is all. If you think that is a kind of manipulation, then calculate the averages and trend of the raw and “manipulated” data (I did…) and see what difference that gives (less than 0.1 ppmv). Or even ask for the raw voltage data (I did…) if you want to check their calculations. Pieter Tans of NOAA will readily send you a few million 10-second voltage readings if you like that…
    Here the calculation for one hour of voltage snapshots:

    If Greenland didn’t melt completely during the previous interglacial, then Antarctica certainly didn’t, simply because the Greenland summit is average around -30°C, while the South Pole station is in average at -50°C. The Vostok ice core is in average -40°C… Thus you need a lot of polar amplification to melt all that ice…

    BTW, if some 20% of all CO2 is replaced in the atmosphere within a year over the seasons and you don’t see more than 2% of full scale global variability in the bulk of the atmosphere, then I call that well mixed. If you have a better definition, I like to hear that…

  93. The .xls sheet with the one-hour calculation of CO2 from the 10-second voltage readings is here: