UPDATED: Roy Spencer on how Oceans are Driving CO2

NOTE: Earlier today I posted a paper from Joe D’Aleo on how he has found strong correlations between the oceans multidecadal oscillations, PDO and AMO, and surface temperature, followed by finding no strong correlation between CO2 and surface temperatures. See that article here:

Warming Trend: PDO And Solar Correlate Better Than CO2

Now within hours of that, Roy Spencer of the National Space Science and Technology Center at University of Alabama, Huntsville,  sends me and others this paper where he postulates that the ocean may be the main driver of CO2.

In the flurry of emails that followed, Joe D’Aleo provided this graph of CO2 variations correlated by El Nino/La Nina /Volcanic event years which is relevant to the discussion. Additionally for my laymen readers, a graph of CO2 solubility in water versus temperature is also relevant and both are shown below:

  

Click for full size images

Additionally, I’d like to point out that former California State Climatologist Jim Goodridge posted a short essay on this blog, Atmospheric Carbon Dioxide Variation, that postulated something similar.

UPDATE: This from Roy on Monday 1/28/08 see new post on C12 to C13 ratio here

I want to (1) clarify the major point of my post, and (2) report some new (C13/C12 isotope) results:

1.  The interannual relationship between SST and dCO2/dt is more than enough to explain the long term increase in CO2 since 1958.  I’m not claiming that ALL of the Mauna Loa increase is all natural…some of it HAS to be anthropogenic…. but this evidence suggests that SST-related effects could be a big part of the CO2 increase.

2.  NEW RESULTS: I’ve been analyzing the C13/C12 ratio data from Mauna Loa.  Just as others have found, the decrease in that ratio with time (over the 1990-2005 period anyway) is almost exactly what is expected from the depleted C13 source of fossil fuels.  But guess what? If you detrend the data, then the annual cycle and interannual variability shows the EXACT SAME SIGNATURE.  So, how can decreasing C13/C12 ratio be the signal of HUMAN emissions, when the NATURAL emissions have the same signal???

-Roy

Here is Roy Spencer’s essay, without any editing or commentary:

Atmospheric CO2 Increases:

Could the Ocean, Rather Than Mankind, Be the Reason?

by

Roy W. Spencer

1/25/2008

            This is probably the most provocative hypothesis I have ever (and will ever) advance:  The long-term increases in carbon dioxide concentration that have been observed at Mauna Loa since 1958 could be driven more than by the ocean than by mankind’s burning of fossil fuels.

            Most, if not all, experts in the global carbon cycle will at this point think I am totally off my rocker.  Not being an expert in the global carbon cycle, I am admittedly sticking my neck out here.  But, at a minimum, the results I will show make for a fascinating story – even if my hypothesis is wrong.  While the evidence I will show is admittedly empirical, I believe that a physically based case can be made to support it.

            But first, some acknowledgements. Even though I have been playing with the CO2 and global temperature data for about a year, it was the persistent queries from a Canadian engineer, Allan MacRae, who made me recently revisit this issue in more detail.  Also, the writings of Tom V. Segalstad, a Norwegian geochemist, were also a source of information and ideas about the carbon cycle.

            First, let’s start with what everyone knows: that atmospheric carbon dioxide concentrations, and global-averaged surface temperature, have risen since the Mauna Loa CO2 record began.  These are illustrated in the next two figures.

Both are on the increase, an empirical observation that is qualitatively consistent with the “consensus” view that increasing anthropogenic CO2 emissions are causing the warming.  Note also that they both have a “bend” in them that looks similar, which might also lead one to speculate that there is a physical connection between them.

Now, let’s ask: “What is the empirical evidence that CO2 is driving surface temperature, and not the other way around?”  If we ask that question, then we are no longer trying to explain the change in temperature with time (a heat budget issue), but instead we are dealing with what is causing the change in CO2 concentration with time (a carbon budget issue).  The distinction is important.  In mathematical terms, we need to analyze the sources and sinks contributing to dCO2/dt, not dT/dt.

So, let us look at the yearly CO2 input into the atmosphere based upon the Mauna Loa record, that is, the change in CO2 concentration with time (Fig. 3).

Here I have expressed the Mauna Loa CO2 concentration changes in million metric tons of carbon (mmtC) per year so that they can be compared to the human emissions, also shown in the graph.

Now, compare the surface temperature variations in Fig. 2 with the Mauna Loa-derived carbon emissions in Fig. 3.  They look pretty similar, don’t they?  In fact, the CO2 changes look a lot more like the temperature changes than the human emissions do.  The large interannual fluctuations in Mauna Loa-derived CO2 “emissions” roughly coincide with El Nino and La Nina events, which are also periods of globally-averaged warmth and coolness, respectively.  I’ll address the lag between them soon. 

Of some additional interest is the 1992 event.  In that case, cooling from Mt. Pinatubo has caused the surface cooling, and it coincides in a dip in the CO2 change rate at Mauna Loa.

These results beg the question: are surface temperature variations a surrogate for changes in CO2 sources and/or sinks?

First, let’s look at the strength of the trends in temperature and CO2-inferred “emissions”.  If we compare the slopes of the regression lines in Figs. 2 and 3, we get an increase of about 4300 mmt of carbon at Mauna Loa for every degree C. of surface warming.  Please remember that ratio (4,300 mmtC/deg. C), because we are now going to look at the same relationship for the interannual variability seen in Figs. 2 and 3.

In Fig. 4 I have detrended the time series in Figs. 2 and 3, and plotted the residuals against each other.  We see that the interannual temperature-versus-Mauna Loa-inferred emissions relationship has a regression slope of about 5,100 mmtC/deg. C. 

There is little evidence of any time lag between the two time series, give or take a couple of months.

So, what does this all show?  A comparison of the two slope relationships (5100 mmtC/yr for interannual variability, versus 4,700 mmtC/yr for the trends) shows, at least empirically, that whatever mechanism is causing El Nino and La Nina to modulate CO2 concentrations in the atmosphere is more than strong enough to explain the long-term increase in CO2 concentration at Mauna Loa.  So, at least based upon this empirical evidence, invoking mankind’s CO2 emissions is not even necessary. (I will address how this might happen physically, below).

In fact, if we look at several different temperature averaging areas (global, N. H. land, N.H. ocean, N.H. land + ocean, and S.H. ocean), the highest correlation occurs for the Southern Hemisphere ocean , and with a larger regression slope of 7,100 mmtC/deg. C.  This suggests that the oceans, rather than land, could be the main driver of the interannual fluctuations in CO2 emissions that are being picked up at Mauna Loa — especially the Southern Ocean.

Now, here’s where I’m really going to stick my neck out — into the mysterious discipline of the global carbon cycle.  My postulated physical explanation will involve both fast and slow processes of exchange of CO2 between the atmosphere and the surface. 

The evidence for rapid exchange of CO2 between the ocean and atmosphere comes from the fact that current carbon cycle flux estimates show that the annual CO2 exchange between surface and atmosphere amounts to 20% to 30% of the total amount in the atmosphere.  This means that most of the carbon in the atmosphere is recycled through the surface every five years or so.  From Segalstad’s writings, the rate of exchange could even be faster than this.  For instance, how do we know what the turbulent fluxes in and out of the wind-driven ocean are?  How would one measure such a thing locally, let alone globally?

Now, this globally averaged situation is made up of some regions emitting more CO2 than they absorb, and some regions absorbing more than they emit.  What if there is a region where there has been a long-term change in the net carbon flux that is at least as big as the human source? 

After all, the human source represents only 3% (or less) the size of the natural fluxes in and out of the surface.  This means that we would need to know the natural upward and downward fluxes to much better than 3% to say that humans are responsible for the current upward trend in atmospheric CO2.  Are measurements of the global carbon fluxes much better than 3% in accuracy??  I doubt it.

So, one possibility would be a long-term change in the El Nino / La Nina cycle, which would include fluctuations in the ocean upwelling areas off the west coasts of the continents.  Since these areas represent semi-direct connections to deep-ocean carbon storage, this could be one possible source of the extra carbon (or, maybe I should say a decreasing sink for atmospheric carbon?).   

Let’s say the oceans are producing an extra 1 unit of CO2, mankind is producing 1 unit, and nature is absorbing an extra 1.5 units.  Then we get the situation we have today, with CO2 rising at about 50% the rate of human emissions.

If nothing else, Fig. 3 illustrates how large the natural interannual changes in CO2 are compared to the human emissions.  In Fig. 5 we see that the yearly-average CO2 increase at Mauna Loa ends up being anywhere from 0% of the human source, to 130%.  

It seems to me that this is proof that natural net flux imbalances are at least as big as the human source.

Could the long-term increase in El Nino conditions observed in recent decades (and whatever change in the carbon budget of the ocean that entails) be more responsible for increasing CO2 concentrations than mankind?  At this point, I think that question is a valid one.