O₂ Diving Towards Danger Point

Not only is O₂ dropping faster than CO2 rising, it is diving towards the danger point much faster than previously thought, O₂ accounting is urgently needed Dr Mae-Wan Ho

New analysis of O2 records raises alarm

In [1] O2 Dropping Faster than CO2 Rising (SiS 44), I highlighted new research showing the depletion of atmospheric oxygen accelerating since 2003, which coincided with the biofuels boom, and warned that climate policies focusing exclusively on carbon sequestration could be disastrous for all oxygen-breathing organisms including humans. I also call for O2 accounting in climate policies. The article attracted many comments not all favourable (see https://www.i-sis.org.uk/O2DroppingFasterThanCO2Rising.php).

Now, six years later, a detailed analysis of the atmospheric O₂ records in 9 stations around the world shows that O₂ is not just falling faster than CO₂ is increasing, it is actually dropping more than 10 times faster than previously thought [2]. If it continues at this rate, the danger point will be reached in thousands of years or less, instead of tens of thousands of years.

One should bear in mind that although the proportion of O₂ in the atmosphere is about 21 %, much, much higher than CO2 (currently ~0.4 %), the dangerous level of oxygen according to the US Occupational Safety and Administration and the US National Institute for Occupational Safety and Health is 19.5 %. In humans, failure of oxygen energy metabolism is the single most important risk factor for chronic diseases including cancer and death (see [1]).

The Scripps O₂ Program

The new analysis [2] has been carried out by Valeri Livinia and colleagues at National Physical Laboratory, Teddington, and John Innes Centre, Norwich in the UK. They used datasets on atmospheric O₂ levels collected by the Scripps Institution of Oceanography at La Jolla, California in 9 stations around the world (see Figure 1) [3] under the direction of Prof Ralph Keeling, who pioneered the measurement techniques in his Ph D thesis at Harvard University in 1988 [4]. Records since 1989 are available from Scripps Pier and Alert in Alaska, although these are not continuous [5]. Continuous records from 7 stations extend back to 1993, and data for the remaining two, Cold Bay in Alaska and Palmer Station in Antarctica, are available back to the mid-1990s. Oxygen levels are measured as changes in the ratio of oxygen to nitrogen O₂/N₂ of sampled air relative to a reference sample of air pumped in the mid-1980s and stored in the Scripps laboratory. The unit is ‘per meg’ such that a decline of 1 per meg is equal to 1 part per million of oxygen or 0.0001 %; or 1 molecule of oxygen per 4.8 molecules of all gases in the atmosphere, not counting water vapour [5].

The Scripps O₂ datasets also have variable timescales, with an average time interval of two weeks. Their length varies from 391 data points (Cold Bay) to 688 (Mauna Loa). For the analysis, all series were interpolated to obtain 4096 data power points separated by equal time periods.

Figure 1   Nine Scripps stations measuring changes in atmospheric O2 around the world

The down-trend is faster than linear for all stations

The data from the nine stations are displayed in Figure 2, where the downward trend is obvious in all the datasets; some looking faster than linear.

Figure 2   Downward trend of O2 concentration in all nine datasets

Livinia and colleagues applied Bayesian inference to the O₂ concentration records, and confirm that they are all dropping at least quadratically (second order polynomial) rather than linearly (first order). Bayesian inference avoids the danger of over-fitting by automatically penalising model complexity. The Bayesian Information Criterion (BIC) is a criterion for model selection, the one with the lowest BIC value being preferred.

They tested several types of polynomial models (orders from 1 to 4, including the linear Ax +B and parabolic Ax² + Bx + C), logarithmic A + Blogx and exponential –exp^Ax +B, using BIC, Akaike Information Criterion AIC and Akaike corrected AICs for model selection. The results plotted in Figure 3 show that both linear and exponential models have the highest Information Criterion values, and are discarded in favour of polynomial models, for which the quadratic is taken as the most feasible (and most conservative) selection.

Figure 3   Plots of Information Criteria for model selection in all datasets

Complete O₂ depletion in 4 500 years instead of 64 000 years

The average over the results of all nine oxygen records assuming quadratic (parabolic) decline is displayed in Figure 4.  As can be seen, 100 % depletion of atmospheric O₂ is projected to occur after 4 000 years. But the danger point, as far as humans are concerned, is a depletion of ~10 %, at <2 750 years; bearing in mind that this is a conservative estimate based on the analysis of available datasets. And it does not take into account the increase in population, fossil fuel consumption and other technological processes that create new oxygen sinks.

Figure 4   Parabolic decline to complete O2 depletion in 4 000 years

The new estimate for O₂ depletion is much earlier than the 64 000 years predicted by Ralph Keeling in 1988 [4] based on contemporary fuel consumption. Livinia and colleagues commented that the 1988 estimate may be far too optimistic [2], as technological changes and population growth would be expected to speed up consumption of fossil fuels, while the production of fertilisers and other materials would also consume atmospheric O₂. That was why they embarked on the new analysis. The problem of atmospheric oxygen deficiency should be addressed in advance, they said, before drastic changes take place, as in the recent problem of the ozone hole. “Too many environmental problems have been analysed by humanity in retrospect.” The purpose of their paper is to make an “advance warning”.

Health-threatening problems in one thousand years or sooner

Livinia and colleagues also carried out tipping point analysis on the O₂ data. Tipping point analysis is a mathematical technique for identifying bifurcation or tipping points (points of sudden changes) in time series data.

The oxygen data analysed by tipping point analysis do not indicate critical behaviour in fluctuations. The transitional behaviour is defined solely by the trend. They used the derived analytical parabolic trend to project the long-term decline of the oxygen concentration. Given the exponential growth of population and consumption, the trend is unlikely to be linear and as slow as Keeling estimated. Instead, they predict health-threatening problems for humans and other animals in one thousand years or sooner.

It is not only direct fossil fuel burning that reduces the oxygen level. There are various indirect factors that may accelerate oxygen depletion. For example, if we assume that world population stabilises and stops growing exponentially, and transport use stops growing at all, oxygen may still continue to decline due to increases in soil fertilisation and production of synthetic materials. In the past 10 years, the use of fertilisers has increased by 35 % per hectare of arable land. This one factor alone may become critical in terms of oxygen depletion within a few decades, if the dynamics remain the same.  It has been pointed out [6] that previously neglected anthropogenic disturbances to the nitrogen cycle, especially nitrate fertilizer industrial production and nitrification, have driven the cycle to a more oxidised state. Similarly, recent changes in biomes, especially the conversion of natural forests into cultivated land, have caused carbon within the terrestrial biosphere to become increasingly more oxidized over a period of decades [7].

It is also necessary to factor in population increases in population and in transport, even with new technologies such as hydrogen fuel cells [2]. World Bank statistics report global carbon emissions from transport rising from 2.5 to 6 x 10¹² kg since 1970, and increase in the number of cars per 1 000 people around the globe from 100 to 130 since 2000, while population increased from 6 to 7.1 x 10⁹ during the same period. That means exponential growth of population will lead to exponential growth not only of carbon emissions but also of air oxygen consumption. Replacing direct fossil fuel combustion by new technologies like fuel cells may restrict carbon emissions but not necessarily the use of air oxygen. For example the hydrogen fuel cell has a net reaction: 2H₂ + O₂ ⟶ 2H₂O. In addition to this use of oxygen, hydrogen production requires oxygen directly or indirectly, either in high temperature-pressure technologies using fossil fuels or in fertilisation for the production of biofuels. That means current industrial hydrogen fuel technologies maintain double sinks for atmospheric oxygen, because they use oxygen when hydrogen is synthesized, and combust oxygen when hydrogen is used as fuel.

The production of cement and fertilisers, petrochemical products of various types (oils, waxes, greases, explosives, rubbers, alcohols, asphalts), fuel cell operation and manufacturing, energy and heat generation, all require atmospheric oxygen for their industrial cycles. The Haber-Bosch process N₂ + 3H₂ ⟶ 2NH₃ for producing nitrogen fertilisers uses methane from natural gas as the source of hydrogen, at a temperature of 300 – 500 ˚C and pressure of 150-250 bars, which requires energy and oxygen consumption.

A future-technology scheme that combines green generation and storage of fuel with increased efficiency of engines is needed.

Livinia and colleagues stated at the end of their report [2]: “We suggest that all industrial processes introducing new technologies with air consumption should be assessed regarding oxygen depletion before they are deployed at full scale.”

More than that, we need to divest from fossil fuel investments and leave oil in the ground (see [8] Age of Oil Ending? SiS 65). Our goal of [9] Green Energies 100 Percent Renewables by 2050 (ISIS/TWN special report) is within sight. A low-carbon economy is not enough; even more importantly, it needs to be low-oxygen or better yet oxygen positive, in order to make up for the losses already incurred.

Article first published 18/05/15

---

References

1. Ho MW. O₂ dropping faster than CO₂ rising. Science in Society 44, 8-10, 2009.
2. Livinia VN, Martins TMV and Forbes AB. Tipping point analysis of atmospheric oxygen concentration. Chaos 2015, 25, 036403.
3. Scripps O₂ Global Oxygen Measurements. Scripps O₂ Program, accessed 14 May 2014, http://scrippso2.ucsd.edu/
4. Keeling, R. Development of an interferometric oxygen analyser for precise measurement of the atmospheric O₂ mole fraction. Ph. D. thesis, Harvard University, Cambridge, Mass. 1988.
5. Modern Records of Atmospheric Oxygen (O₂) from Scripps Institution of Oceanography. Carbon Dioxide Information Analysis Center, accessed 14 May 2015, http://cdiac.ornl.gov/trends/oxygen/modern_records.html
6. Ciais P, Manning A, Reichstein M, Zaehle S, and Bopp L. Nitrification  amplifies the decreasing trends of atmospheric oxygen and implies a larger land carbon uptake. Global Biogeochem Cycles 2007, 21, GB2030.
7. Randerson J, Masiello C, Still C, Rahn T, Poorter H, and Field C.  Is carbon within the global terrestrial biosphere becoming more oxidized? Implications for trends in atmospheric O2. Global Change Biol 2006, 12, 260-71.
8. Ho MW. Age of oil ending? Science in Society 65, 2-5, 2015.
9. Ho MW, Cherry B, Burcher S and Saunders PT. Green Energies, 100 % Renewables by 2050, ISIS/TWN, London/Penang, 2009, https://www.i-sis.org.uk/GreenEnergies.php