Beck on CO2 and Temperature: Oceans are the "dominant CO2 store"

I’d like to comment on this, because I’ve looked at the issue in quite some detail, and even posted a paper about it on Arxiv.
In that paper, I showed that you could simulate the growth of CO2 in the atmosphere using only the temperature data, and get really good agreement with both the long term growth curve, and the short term (monthly) fluctuations in CO2 uptake (rate of change), all this assuming a simple temperature response of the carbon cycle with a given response time.
However, that result was wrong. It was wrong in that you can get the same kind of agreement if you include the “human” emissions properly, something which I had not done at the time I wrote the paper. So I am now pretty convinced that most of the increase in CO2 concentration is due, in fact to human emissions. (and yeah, if I were not so lazy, I would have removed my paper from Arxiv already…).
That being said, I learned a lot from this exercise. Making mistakes is a good way to learn.
One thing is that the carbon cycle IS dependent on temperature, something that has received surprisingly little attention in the scientific litterature. It seems to me that we could get very useful information if we studied that dependence in more detail.
But the other thing I realized is that the carbon cycle DOES react to human emissions (well, all right, nothing new here…). The carbon “sink”, whether it is forests, or phytoplankton, expands with the amount of carbon in the atmosphere. In fact, the carbon sink has expanded tremendously over the past century. The more CO2 we pump into the atmosphere, the more is absorbed. I knew it before, but playing with the data just made it more real…
But a big question is, what is the response time of that reaction? We are being told that the “residence time” of CO2 is 50, or even 100 years. In other words, it would take that long to get rid of the surplus CO2 if we stopped emitting today. I have a hard time with that figure. There just does not seem to be any empirical support for it. Sure, you can find nice simulations with such long lifetimes, superimposed on the CO2 growth curve, and there seems to be a perfect match. The problem is, you can get a similar match with about ANY lifetime value, short or long! And at some point, the uncertainty about the past history of CO2 just prevents you from discrimating between the possible values.
In the end, it is a trade-off: either the system responds strongly, but slowly, or it is less sensitive, but reacts rapidly.
A careful examination of the temperature response might help discriminating between the two options. But I must admit that I could never get a statistically significant answer. I just stopped trying…
A better knowledge of the past history of CO2 concentration would also help. I am not convinced by Beck’s data on the high CO2 concentrations during the 1930’s. Just seems too anomalous. On the other hand, I think that stomatal frequency data could provide a more precise proxy than ice cores. They do tend to show that the CO2 concentration fluctuated more than is currently admitted, and moreover that it followed the temperature.
A final comment on high CO2 concentrations in the oceans. Most people are not aware of it, but the CO2 concentration at the surface of oceans varies tremendously, depending on where you are. I mean, it varies by as much as a factor of two. Some regions of the oceans are very strong sinks, others outgas like crazy. Temperature has an effect on this, but mostly, this is the result of biological activity. Phytoplanktons are tremendous little beasts, they eat up a LOT of carbon, which they then dump at the bottom of the oceans. Their role in the carbon cycle is just beginning to be fully unraveled. There’s a lot we don’t know. Ocean biological productivity turns out to be awfully difficult to measure with any accuracy.
In the end, the role of phytoplanktons may be crucial in understanding our climate, especially the switch between glacial and interglacial, and why there are ice ages at all. They may be the reason why, as Gerald Marsh pointed out in his APS article, CO2 concentration is actually historically low.

The CO2 data can’t be any worse than the surface record of temperature as revealed by your weather station survey can it, Anthony?

I haven’t read the paper, but I’m highly dubious. What wet chemical measurements of CO2 have been made lately, with which we can compare to Mauna Loa?
Unless I’m missing something, this splice-job is just as bad as the Hockey Stick. Just like McIntyre’s call for updating the tree-ring proxies, I won’t believe a word of this until someone shows me that wet chemical measurements made today in the same way as in the past can agree with Mauna Loa.

I have often heard from the alarmists that average atmospheric CO2 concentrations are higher now (at around 360ppm) than ever before in the history of instrumental data. Beck’s work indicates concentrations of over 430 ppm a bit less than 200 years ago. Did the alarmists choose to ignore the data that didn’t fit or were they just ignorant of the larger body of climate science work by others?
I don’t quite understand the 5 yr. temperature lag. The ice cores are suggesting a much longer lag, but opposite, that is, CO2 concentrations lag temp rise. Can some one explain this difference?

Acknowledging some concern with Beck’s rigor, all of the criticism’s amount to is ankle-biting.
Mauna Loa CO2 levels are measured indirectly, via IR absorption. These spectroscopic absorption criteria were not established empirically but by means of the outdated Beer-Lambert Law that, while yielding a straightforward answer, is clearly inadequate to the task. In fact all of the spectroscopic absorption graphs commonly available were drawn from such computations, they are ‘electromagnetically naive’.
No calibration procedures are published and they persist in drying the air with H2SO4.
The most that can be said is they give the ‘shape” of an actual smoothed curve. The data deleted by the smoothing process is not ‘noise’, as though one knows the answer already and simply needs to make the data obey.
The ice-core CO2 levels aren’t ‘imprecise’ they are worthless for establishing contemporary atmospheric levels. Apart from comparing historic trends they use is only obfuscatory.
Picking at Beck’s reconstruction of historic volumetric measurements and other methods by those who have established no particular analytic insight into the methods employed is a waste of our time to read.
I do not for a moment believe Francois’ assertion that the antrhropogenic fluence can even be seen on a global scale. Deleting the ocean fluence from the seasonal sinusoid and longterm trend would still leave it indistinguishable noise on the resulting curve.
The Mauna Loa seasonal curve cannot be thought to be biogenic; the NH growth flattens by the end of August. The total biogenic fluence is on the order of 800 Gtons, that dissolved in the oceans and in intermediate bottom-dwelling carbonates is 150,000 Gtons. As the temperature rises CO2’s partial pressure rises and it enters the atmosphere, as the ocean’s temperature cools the partial pressure falls and it leaves the atmosphere.

The thing that bothers me about the ML CO2 is the actual readings declined a bit and stayed flat for a year. But humans were producing just as much CO2 as before and Pinatubo put up a bunch as well.
That tells us that there is another agent using CO2 or producing CO2.

Jim Watson:
That was a great theory when I first heard it, but global temperatures were probably higher between 1000-1100 than 1200. What this paper by Beck shows us is that this issue is very complicated, as CO2 concentration is quite variable from place to place and method to method. The oceans, if I were to guess, were probably responsible for at least 20 parts per million of this CO2 increase, and maybe higher than 30 parts per million. Unless someone can correct me….

Mark H. says: “Is Mauna Lao the only place on earth taking RDIF readings?” No, here are time series from various places around the world: http://www.esrl.noaa.gov/gmd/ccgg/carbontracker/tseries.php?type=mr
Jim Watson says: “The incredible uniformity of the recent rise of CO2 in the atmosphere has always troubled me a little…This type of smooth signature is NOT normally indicative of human activity. Shouldn’t there be fits and jabs as the CO2 rises in response to changes in human economies, the rise of China as an industrial giant, the changes in standards of living in India, etc.” Well, it hasn’t been completely smooth and there are little fits and jabs in addition to the seasonal cycle (although they probably have more to do with variations in uptake than variations in emissions). However, I don’t think that those things you listed regarding human economies have made for dramatic differences in emissions from one year to the next. Also note that the emissions is correlated with the derivative, i.e. the slope, of the CO2 concentration curve…So, even if there was a 10% change in emissions from one year to the next (which I think would be huge compared to reality, except perhaps during the Great Depression), it would not look all that dramatic on the plot of CO2 concentration vs time.
Jim Watson says: “THEORY: Could it not be that the current rise in CO2 is merely the natural result of the ‘800-year lag’ in warmer temperatures from the Medieval Warm Period?” In a word, no. For lots of different reasons, we know the current rise is due to us and that the oceans are in net absorbing…not outgassing…CO2 currently. And, it would be an awfully strange coincidence if the only time during the last 750,000 years (according to ice core measurements) that CO2 levels rose above 300ppm was the century when we started emitting significant quantities of it into the atmosphere.

I don’t think the spectroscopic measurements could be very imprecise. Spectroscopy is not rocket science, it is performed daily in thousands of labortories, and the techniques are well proven, and get better every day. And Mauna Loa is NOT the only place where CO2 is measured. There are a number of stations around the world. They don’t all agree, because CO2 does travel around. Most emissions are in the northern hemisphere, and the oceans have, as I said, local areas where they absorb and outgas, which also depends on the time of year (and the temperature). Quite frankly, I think the CO2 data are quite reliable.
If there is an area where the data is uncertain, it is that of human emissions. Before we got interested in CO2, no one really paid any attention to how much CO2 was emitted worldwide. So any data we have for worldwide emissions before, say, 20 years ago, are highly speculative. This is what, IMO, makes it hard (or easy…) to extract parameters on the lifetime of CO2. Because, in order to do that, you need to have a model linking emissions and concentration. A simple model of the form:
dC/dt = -a*(C-Ceq) + Em
where C is the CO2 concentration, Ceq is some equilibrium value, and Em are emissions. The parameter “a” is what you want to know. But the equilibrium value is not a constant. This is where you include the growing sink, and the temperature effect. In fact, you can have a set of coupled equations with CO2 in the air, in the land biomass, and in the oceans. The land biomass can also be split into the living biomass, and the CO2 in the soil. The oceans in turn have CO2 in the upper surface, and at the bottom. And so on, and so on, it can get very complicated. Various interactions can have different time constants and sensitivities.
But in trying to extract those parameters, you are stuck with the empirical data that you have. Again, IMO, the emissions data have much more uncertainty than the CO2 data post-1959, and this is what limits our ability to simulate the carbon cycle.
There IS definitely an effect of temperature, superimposed on human emissions. I found that it is almost impossible to disentangle the two. The temperature effect could be large (as much as 20% of the rise of the past century), or it could be small. In my simulations, there was no statistical difference between the two cases.

From Beck’s English links page
He links to his published paper:
180 YEARS OF ATMOSPHERIC CO2 GAS ANALYSIS BY CHEMICAL METHODS, Ernst-Georg Beck
Energy & Environment · Vol. 18, No. 2, 2007, pp 259-282
Erratum
With two published comments:
Comment on “180 years of atmospheric CO2 gas analysis by chemical methods” by ernst-georg beck pp. 635-646(12) Authors: Meijer, Harro A.J.;,
and Keeling, Ralph F.
with Beck’s reply Comments on “180 years of Atmospheric CO2 Gas Analysis by Chemical Methods” pp. 641-646(6) Author: Beck, Ernst
<a href=”http://www.biokurs.de/treibhaus/180CO2_supp.htm”Beck’s web cite links to numerous historical papers and analyses which should keep most critics happily(?!) reading for many hours.
Note Beck’s spread sheetBasic database (>90 000 series, 143 averages over 150 years, >53 locations ) rev1, 11/2007
Finally note Beck’s 113 page monograph. (Forthcoming?)
Beck E. G. (2006/2007) History of CO2 Gas Analysis of Air by Chemical Methods, pp. 1–113. “to be published.” (Can anyone find this published/posted?)
Eyeballing the chart above, note particularly the deviations in:
1822 285 ppm vs 440 ppm = 56% difference or 18 std deviations;
1856 290 ppm vs 380 ppm = 31% difference or 10 std deviations;
1942 305 ppm vs 440 ppm = 44% difference or 15 std deviations; assuming a 3% standard deviation.
Those relative deviations appear sufficiently far probability to require a serious analysis as to the causes and/or errors.