This interesting essay by Dr. Spencer is reposted from his blog, link here:
Global Warming Causing Carbon Dioxide Increases: A Simple Model
May 11th, 2009 by Roy W. Spencer, Ph. D.
Global warming theory assumes that the increasing carbon dioxide concentration in the atmosphere comes entirely from anthropogenic sources, and it is that CO2 increase which is causing global warming.
But it is indisputable that the amount of extra CO2 showing up at the monitoring station at Mauna Loa, Hawaii each year (first graph below) is strongly affected by sea surface temperature (SST) variations (second graph below), which are in turn mostly a function of El Nino and La Nina conditions (third graph below):
Click for larger images
Click for larger image
During a warm El Nino year, more CO2 is released by the ocean into the atmosphere (and less is taken up by the ocean from the atmosphere), while during cool La Nina years just the opposite happens. (A graph similar to the first graph also appeared in the IPCC report, so this is not new). Just how much of the Mauna Loa Variations in the first graph are due to the “Coke-fizz” effect is not clear because there is now strong evidence that biological activity also plays a major (possibly dominant) role (Behrenfeld et al., 2006).
The direction of causation is obvious since the CO2 variations lag the sea surface temperature variations by an average of six months, as shown in the following graph:
So, I keep coming back to the question: If warming of the oceans causes an increase in atmospheric CO2 on a year-to-year basis, is it possible that long-term warming of the oceans (say, due to a natural change in cloud cover) might be causing some portion of the long-term increase in atmospheric CO2?
I decided to run a simple model in which the change in atmospheric CO2 with time is a function of sea surface temperature anomaly. The model equation looks like this:
delta[CO2]/delta[t] = a*SST + b*Anthro
Which simply says that the change in atmospheric CO2 with time is proportional to some combination of the SST anomaly and the anthropogenic (manmade) CO2 source. I then ran the model in an Excel spreadsheet and adjusted an “a” and “b” coefficients until the model response looked like the observed record of yearly CO2 accumulation rate at Mauna Loa.
It didn’t take long to find a model that did a pretty good job (a = 4.6 ppm/yr per deg. C; b=0.1), as the following graph shows:
Click for larger image
The best fit (shown) assumed only 10% of the atmospheric CO2 increase is due to human emissions (b=0.1), while the other 90% is simple due to changes in sea surface temperature. The peak correlation between the modeled and observed CO2 fluctuation is now at zero month time lag, supporting the model’s realism. The model explained 50% of the variance of the Mauna Loa observations.
The best model fit assumes that the temperature anomaly at which the ocean switches between a sink and a source of CO2 for the atmosphere is -0.2 deg. C, indicated by the bold line in the SST graph, seen in the second graph in this article. In the context of longer-term changes, it would mean that the ocean became a net source of more atmospheric CO2 around 1930.
A graph of the resulting model versus observed CO2 concentration as a function of time is shown next:
If I increase the anthropogenic portion to 20%, the following graph shows somewhat less agreement:
Click for larger images
There will, of course, be vehement objections to this admittedly simple model. One will be that “we know the atmospheric CO2 increase is manmade because the C13 carbon isotope concentration in the atmosphere is decreasing, which is consistent with a fossil fuel source.” But has been discussed elsewhere, a change in ocean biological activity (or vegetation on land) has a similar signature…so the C13 change is not a unique signature of fossil fuel source.
My primary purpose in presenting all of this is simply to stimulate debate. Are we really sure that ALL of the atmospheric increase in CO2 is from humanity’s emissions? After all, the natural sources and sinks of CO2 are about 20 times the anthropogenic source, so all it would take is a small imbalance in the natural flows to rival the anthropogenic source. And it is clear that there are natural imbalances of that magnitude on a year-to-year basis, as shown in the first graph.
What could be causing long-term warming of the oceans? My first choice for a mechanism would be a slight decrease in oceanic cloud cover. There is no way to rule this out observationally because our measurements of global cloud cover over the last 50 to 100 years are nowhere near good enough.
And just how strenuous and vehement the resulting objections are to what I have presented above will be a good indication of how politicized the science of global warming has become.
REFERENCES
Michael J. Behrenfeld et al., “Climate-Driven Trends in Contemporary Ocean Productivity,” Nature 444 (2006): 752-755.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
Just back from a nice trip to Prague, Linz (Austria) and Southern Germany, saw this discussion now. Took me halve a day to wade through the comments… A lot of interest anyway!
I have made a lot of comments on the origin of the CO2 rise in the past, including Dr. Spencer’s different attempts to look at nature as cause…
Again a new round, where Dr. Spencer used this simple model:
delta[CO2]/delta[t] = a*SST + b*Anthro
There are fundamental problems with this model:
The first one is that there is no time limit in the equation for a change in sea surface temperature. That is the same problem as with the previous discussion with Frank Lansner at this blog: if the sea surface temperature goes up, the model assumes that the increase of CO2 per year remains the same over time. That is certainly not the case, as a new equilibrium between ocean temperature and vegetation at one side and CO2 levels at the other side will be reached within a limited amount of time for the fast processes (ocean solubility and vegetation growth) and after longer times for the slow processes (ice sheet/vegetation area, -deep- ocean flows). The real formula should have a term related to the change in temperature, not the relative temperature, as the overall change in CO2 is related to the change in temperature…
The second fundamental problem is that the formula shows the short term relationship between dCO2 and dT, that is the variability around the trend, but that doesn’t need to have any relationship with the trend itself, if the trend is caused by something else than temperature. Temperature then is only causing the “noise” around the trend. And eventually a small part of the trend, if there is a temperature difference between the start and the end of the period of interest. On the other side, the emissions cause only a small part of the variability around the trend, as most of the variability is temperature related.
I have used an alternative “simple model”:
dCO2 = 3*dT + 0.55*emissions
where CO2 and emissions are in ppmv (1 ppmv CO2 = 2.1 GtC emissions) and 3*dT is the factor 3 ppmv/°C which fits for short term changes, up to 8 ppmv/°C for very long term temperature influence on CO2 levels (glacials – interglacials, MWP-LIA temperature change) and dT is the temperature difference between begin and end of the period…
If we plot that formula for the short term variability (dCO2/dt), then we see the following:
http://www.ferdinand-engelbeen.be/klimaat/klim_img/egbn_trend.jpg
which is a reasonable fit to the variability around the trend.
Pieter Tans of NOAA, responsible for the CO2 data of several stations, including Mauna Loa, made a better fit by including precipitation, which influences plant growth (see the second halve of the pdf):
http://esrl.noaa.gov/gmd/co2conference/pdfs/tans.pdf
Some different view if we look at the trends themselves. Have a look at the trends of temperature and accumulated emissions vs. CO2 levels:
http://www.ferdinand-engelbeen.be/klimaat/klim_img/temp_co2_acc_1900_2004.jpg
The CO2 levels until 1960 are from different ice cores (Law Dome, best resolution 8 years), after 1959 from Mauna Loa. As can be seen, even in cooling periods (1945-1975) the CO2 levels increase at about 55% ratio with the accumulated emissions…
Even more informative, the correlation trends 1900-2004 between accumulated emissions and atmosphere:
http://www.ferdinand-engelbeen.be/klimaat/klim_img/acc_co2_1900_2004.jpg
Compare that to the temperature – CO2 correlation trend:
http://www.ferdinand-engelbeen.be/klimaat/klim_img/temp_co2_1900_2004.jpg
where a large change in temperature has little influence on CO2 increase…
Main conclusion of this all: temperature has a fast (and slow) influence on CO2 levels, but that is limited to 3ppmv/°C for short term changes, up to 8 ppmv/°C for very long term changes. Thus while temperature is mainly responsible for the variability around the trend, the emissions are responsible for the bulk of the trend itself…
Paul Vaughan 16-05-2009 (20:46:46) :
My guess is that there will be plenty of datasets & publications, including a good number focused on short timescales. Publicly-available text-format data-websites are (arguably) essential in a knowledge society. Although I’m confident shorter timescale datasets exist, I suspect you’re right that they will not be long-term and I anticipate tediously-inefficient bureaucratic access-hoops (in most/many cases).
The data of one-hour raw averages of 40 minutes of 10-second measurements (without any filtering) are available at:
ftp://ftp.cmdl.noaa.gov/ccg/co2/in-situ/
for four base stations, including Barrow, from the start year (1973 for Barrow) up to 2008…
A nice explanation of the procedures followed at Mauna Loa and other baseline stations is at:
http://www.esrl.noaa.gov/gmd/ccgg/about/co2_measurements.html
About the dip of Barrow and other far north stations: most of the time the ocean near Barrow is frozen. When the temperature increases, two main things happen: tundra grows as crazy and the ocean ice sheet disappears. The first can be seen in the increase of d13C level and both as a dip in the CO2 levels (the stil cold ocean absorbs a lot of CO2). The increase in winter is more difficult to explain, but CO2 may be transported in from other places.
Most baseline stations are in the Pacific at sea level and don’t show such a huge seasonal variability. But if you look at the variability of Schauinsland (Black Forest, southern Germany), at 1200 m when measured with sufficient wind speed and above the inversion layer, the seasonal variation is even larger than at Barrow:
http://cdiac.ornl.gov/trends/co2/graphics/schauinsland.gif
The Ferrell air moving cells may transport the CO2 levels polewards…
Hope this helped…
Paul Vaughan (02:24:30) :
Thanks for the info Looking at the “daily” data – not daily! – the minimum of the CO2 dip occurs within about +- 1 week f4rom the start of the northern records to present.
Ferdinand Engelbeen (10:47:54) :
The daily data dip in CO2 is very sharp. It goes from negative to positive slope within a couple of weeks. The slope to max co2 is very similar to to min co2 in this dip.
Is this really possible from a biological process and sea ice? It’s acting as if a CO2 sink / source switch has been thrown. No AGW agenda on this point! I just would like to know how the mass of the atmosphere can react this quickly.
bill 20-05-2009, 10:28:24 :
Both the start and end of the growing season and the ice breakup and reformation are quite rapid processes in the high north… Usually transformed from complely frozen sea to open sea (at least near the coast) and from snow covered frozen land to growing leafs and flowers within a few weeks…
The atmospheric mixing is relative fast: days to weeks at the same altitude, with some Noth-South gradient (and a steep gradient near the equator, caused by the ITCZ), weeks to months for different altitudes. But as plant growth and ocean absorption are rather continuous processes when the temperature is high enough and ice is melted, there will be a continuous difference between the places with (relative) huge vegetation growth and high absorption and the rest of the atmosphere.
The other way out is more difficult to understand, as once all is frozen, there is no real source of CO2, or there should still be a lot of bacterial life decomposing the organics on/in the (permafrost!) soils.
Alternative is that the fast changes come from elsewhere and are brought in by air circulation, where the change from polar highs in winter to more southernly air masses could be the cause…
That the dip is mainly from vegetation growth can be seen in the d13C changes: http://cdiac.ornl.gov/trends/co2/allison-csiro/graphics/alc_c13co2.jpg here for Alert (NW Territories, Canada)
Looking at the different graphs of CO2 and CO2C13 (=d13C) at the NOAA website, it seems that the behaviour of CO2 and d13C at Barrow and Hohenpeisenberg (Germany, 1,000 m) is practically the same, see: http://www.esrl.noaa.gov/gmd/ccgg/iadv/
It looks like that instream at Barrow of already mixed air from the south is the main cause of the variation…
Ferdinand Engelbeen (15:52:44) :
If the change over from -ve co2 slope to +ve slope were to take many weeks I would agree that vegetation/Ice could be the cause but using your ref for hourly data I plotted this for Barrow:
http://img132.imageshack.us/img132/7442/barrowhourlyco2.jpg
http://img190.imageshack.us/img190/1068/co2x7.jpg
Despite the change in temperature between 1975 and 2008 the minimum has barely changed. And the slope switch occurs in a metter of a couple of weeks
I still find the speed of response amazing for such a massive system!
Vegetation will affect the CO2 in a continual way for many weeks as spring approaches from the south. By the time Barrow unfreezes it will be growing vigorously a few hundred km further south.
Re: Ferdinand Engelbeen (10:47:54), (11:20:28), & (15:52:44)
Thank you for providing the very helpful comments & links.
–
Clarification:
It is not the summer dip at Alert (northern Canada), but rather the winter dip (in dCO2/dt) that is of particular interest.
I have been digging around to try to figure it out. Anything I could offer at this stage would be premature — I will refrain from speculative comment at this time.
Insights from anyone presently more-knowledgeable are certainly welcome.
– – –
Regarding “daily” CO2 time series, there is a noteworthy mystery:
The data labeled “daily” at
http://scrippsco2.ucsd.edu/data/data.html
…are NOT daily.
Fortunately, Ferdinand has pointed out truly daily CO2 time series:
ftp://ftp.cmdl.noaa.gov/ccg/co2/in-situ/
I briefly became concerned that a lot of the data had been estimated (interpolated – possibly using untenable assumptions about annual structure) after noting the long & variable-length gaps in the Scripps “daily” [but not daily] data.
I am very thankful that Ferdinand has commented. However, I will be uneasy about all CO2 data until someone sheds some light on what is up with the Scripps website’s misleading “daily” links.
–
bill (16:46:57) “Despite the change in temperature between 1975 and 2008 the minimum has barely changed. And the slope switch occurs in a metter of a couple of weeks
I still find the speed of response amazing for such a massive system!”
I imagine you are speaking of the seasonal timing of the minimum. If so, keep in mind that at the latitude of Alert, there is basically only polar-night (winter) & polar-day (summer). Think about the very low angle of the sun as it circles the sky [staying above the horizon (aside from mountain-shadows)] in the “evening” of the polar-day (i.e. approaching Sept. 21).
Just Want Truth… (16:45:32) :
CORRECTION
Ferenc Miskolczi not Miklós Zágoni.