Measurements from the Scripps O2 Program show that from January 1992 to January 2009, the O2 concentration under unpolluted conditions at La Jolla decreased by 320 per meg. This corresponds to a loss of 320 O2 molecules out of each one million O2 molecules in the atmosphere, or a loss of 0.032%. The decrease measured at other stations in our network is very similar, showing that the trend is global.
The annual global O2 consumption from fossil-fuel burning in year 2005 was 917 Tmol O2 or 9.17×1014 moles of O2. The corresponding figure for year 1995 was 741 Tmol. The cumulative fossil-fuel O2 consumption from the dawn of the industrial era to 2005 was 36001 Tmol O2 or 3.6×1016 moles of O2. Moles of O2 can be converted to grams of O2 by multiplying by the molecular weight of 32.00 g/mol.
These estimates are based on compilations of the CO2 emissions from fossil-fuel usage from the Carbon Dioxide Information and Analysis Center (CDIAC) combined with O2 consumption factors from the Ph.D. Thesis of R. Keeling. The consumption factors are as follows: Gas (1.95), Liquids (1.44), Solids (1.17) , Cement Manufacture (0.0), Gas Flaring (1.98), where the factors are in units of moles O2 consumed per mole of CO2 released.
The answer can be estimated from the formula:
δ(O2/N2) decrease in per meg units= (FF/M)×106
where FF is the number of moles of O2 consumed by fossil-fuel burning, and M= 3.706×1019 moles is a reference value for the total number of O2 molecules in the atmosphere. (It suffices here to use a constant value for M because the change in M is so small). The extra factor of 106 is needed to convert to per meg units.
Annual decrease from fossil-fuel burning in 2005: (9.17×1014)/(3.706×1019)×106 = 24.8 per meg. Cumulative decrease from the dawn of the industrial era to 2005: (3.6×1016)/(3.706×1019)×106= 972 per meg. The decrease of 972 per meg corresponds to losing 972 O2 molecules per million O2 molecules in the atmosphere, or a loss of roughly 0.1% of the O2 content.
Our program only started making measurements in 1989, so we don’t have direct measurements of the earlier changes. The overall decrease must be dominated, however, by the O2 loss from fossil-fuel burning, which is estimated to have been around 0.1% of the O2 content of the atmosphere through year 2005.
This point was addressed in an article by Broecker (Science, Vol. 168, 1537-1538, 1970). Although written many years ago, the basic point made by the article remains valid. The maximum potential loss of O2 from fuel burning, when fossil fuel reserves (mostly coal) are exhausted is only a few percent of the atmospheric burden. Since even this loss will take many centuries to materialize, it's hard to see this as high on the list of possible environmental concerns. Fossil-fuel burning causes much larger relative changes in atmospheric CO2, which is much less abundant in air than O2.
At higher elevations, the O2 partial pressure is lower, just as the total air pressure is lower. The O2 concentration, expressed in terms of the O2/N2 ratio, stays very constant with elevation, however, because the N2 partial pressure decreases by the same percent with height. The homogenization results because the atmosphere is turbulent.
One process which drives non-negligible changes in N2 is warming and cooling of the ocean surface and the associated changes in the N2 solubility in the seawater. N2 is more soluble in cold than warm water. As seawater warms, N2 is released from the water into the air. The opposite happens if seawater cools. This process has a small but significant effect on the seasonal cycles in O2/N2 ratio. Other processes causing changes in atmospheric N2, such as nitrogen fixation or dentrification don’t appear large enough to drive significant O2/N2 changes.
The ppm unit for CO2 is a measure of the mole (or molecular) fraction. One ppm corresponds to one CO2 molecule per million air molecules. Although we could easily report O2 changes also in mole fraction units, this would create complications. Suppose you have an air sample containing one million molecules, of which 350 are CO2 and 210000 are O2, yielding CO2 and O2 concentrations of 350ppm and 210000ppm, respectively. Adding an additional molecule of CO2 molecule to the sample increases the CO2 concentration to 351/1000001 = 0.00035099965 or 350.99965 ppm, whereas adding an additional O2 molecule to the sample increases the O2 concentration to 210001/1000001 = 0.21000079 or 210000.79 ppm. CO2 increases by almost exactly 1 ppm, while O2 increases by 0.79 ppm. The O2 change is smaller than 1 ppm because the total number of molecules increases, partly offsetting the increase in O2. The same effect for CO2 is negligible. If this isn’t confusing enough, it is easy to show that the O2 mole fraction changes even for the CO2 addition alone (dropping by 0.21 ppm). Clearly, using mole fraction units is a confusing basis for tracking flows of both O2 and CO2. Expressing O2 concentrations in terms of the O2/N2 ratio avoids these complications.
These units refer to different types of quantities, so the question needs to be sharpened before it can be clearly answered. Suppose a tree consumes exactly one molecule of CO2 for each O2 molecule produced by photosynthesis. The changes in atmospheric O2 and CO2 near the tree will then be inversely proportional. What is the proportionality factor in per meg/ppm? The answer is 1/.2095 = 4.8 per meg/ppm, where 0.2095 is the O2 mole fraction of air. This can be derived realizing that, because N2 is constant, the relative change in the O2/N2 ratio is the same as the relative change in O2 and calculating the relative change requires dividing by its abundance.
APO is useful for extracting signals of oceanic origin from combined O2/N2 and CO2 data. The exchanges of O2 and CO2 with land biota are very tightly coupled by the chemistry of photosynthesis and respiration. APO is defined in such a way that the effect of land plants on the δ(O2/N2) term largely cancels the effect on the XCO2 term, leaving APO unchanged. APO is strongly influenced by air-sea exchanges of O2 and CO2, which are not typically coupled in tight proportions, and by burning of fossil-fuels, particularly petroleum and natural gas, due their higher O2:C combustion ratio.