Ozone mystery laid to rest

Measurements of atmospheric ozone levels taken during the nineteenth century cast doubt on the computational models used today to simulate the atmosphere. An independent proxy of past ozone levels offers reassurance.
Mathew Evans is at the Wolfson Atmospheric Chemistry Laboratories, Department of Chemistry, University of York, York YO10 5DD, UK.

Search for this author in:

Before the 1950s, direct observations of the composition of the atmosphere were extremely limited. Air trapped in pockets of snow and ice has allowed the observational record for some gases to be extended back hundreds of years. But certain gases that affect climate and air quality, such as ozone (O3), are not stable in ice or snow, limiting their records to the past few decades. Writing in Nature, Yeung et al.1 report that isotopic observations of oxygen (O2) molecules trapped in polar snow and ice can provide a new constraint on estimates of ozone levels in the troposphere (the lowest 12 kilometres of the atmosphere) over the past 150 years. This greatly extends our knowledge of the concentration of this key atmospheric gas, and might finally address a problem that has worried atmospheric chemists for decades.

Advances in atmospheric science are often made by taking advantage of the ‘experiments of opportunity’ that occur as a result of natural and human-driven changes to the atmosphere. It is therefore crucial to make long-term measurements of the atmosphere. The modern era for such measurements began in the late 1950s, with observations of carbon dioxide levels2 from the Mauna Loa Observatory in Hawaii. Long-term measurements of other atmospheric components started soon after that. It was only in the 1990s that observational networks expanded sufficiently to provide a global perspective of a wide range of the air’s components3.

However, the impact of the atmosphere’s changing composition depends on both the present-day concentrations of pollutants and their concentrations before the change started. Many atmospheric changes (climate change being the most notable example) are associated with the Industrial Revolution, and so, to understand the magnitude of those effects, we need to understand the composition of the atmosphere over the past 100–200 years.

The changes in concentrations of some components of the air, such as CO2 and methane, are known quite well from the ice record and from modern observations. However, because the ice record provides no constraints on estimates of changes in ozone levels, computer models of atmospheric chemistry and physics are used instead. These models calculate a roughly 40% increase in tropospheric ozone levels between 1850 and the present4. The Intergovernmental Panel on Climate Change uses these estimates to evaluate the relative importance of ozone as a greenhouse gas, compared with CO2 and methane, and thereby to develop climate policy5.

The use of computational estimates of ozone wouldn’t be a concern if it weren’t for an unlikely set of ozone measurements made in the nineteenth century by scientists at several sites around the world (Fig. 1). These measurements were made using various techniques that have since been assessed by present-day scientists — who calibrated the data to place the results on modern scales and worked out that other pollutants, such as sulfur dioxide, could have interfered with the measurements6,7. It was difficult, however, to assess the magnitude of such interference. Surprisingly, the data suggested that nineteenth-century ozone concentrations were very low, and increased by about 300% during the industrial period — a much larger change than was calculated by models8.

Pic du Midi Observatory, France

Figure 1 | The Pic du Midi Observatory, France. Measurements of atmospheric ozone levels were made during the nineteenth century, including some at the Pic du Midi Observatory. Yeung et al.1 report an independent constraint on atmospheric ozone levels that suggests the concentrations recorded in the nineteenth century were artificially lowered by other pollutants.Credit: Hemis/Alamy

On the face of it, these results have enormous implications. Given that present-day concentrations of ozone calculated by models are roughly correct9, a failure to simulate concentrations before the industrial period would suggest that there is a fundamental problem with our understanding of atmospheric chemistry. In that case, how could we trust the models’ predictions of future atmospheric composition, and therefore formulate climate policy? If ozone concentrations had increased much more than was previously thought, then the role of the gas in climate change would be larger than had been assumed, and so efforts to lower ozone levels would have greater potential to reduce global warming.

It seemed implausible to most atmospheric chemists that the computer models could be so wrong. The most likely explanation was that sulfur dioxide and other chemical reductants produced by coal burning during the Industrial Revolution interfered with the observational techniques used at that time. But the absence of an independent observational constraint on historical ozone levels has meant that the nineteenth-century observations have vexed atmospheric chemists for the past few decades.

Although ozone molecules do not remain trapped in ice and snow, oxygen molecules do. Yeung et al. therefore measured the amounts of the common oxygen-16 isotope and of the much less common oxygen-18 isotope in oxygen molecules trapped in polar ice and snow. The production of ozone in the atmosphere changes the proportion of these isotopes in atmospheric oxygen molecules. The isotopic record of trapped oxygen molecules therefore contains a history of ozone concentrations over the past 150 years.

Yeung and colleagues’ analysis shows that the increase in tropospheric ozone over that period was around 40%, much smaller than the increases indicated by the nineteenth-century observations, and consistent with the numbers predicted by models. It therefore seems likely that interference from sulfur dioxide and other gases had indeed artificially lowered the ozone concentrations recorded in the historical measurements.

Those nagging doubts of atmospheric chemists can probably now be laid to rest. Perhaps disappointingly for some, there doesn’t seem to be a fundamental problem with our understanding of atmospheric chemistry, which means that the scope of ozone management in helping to reduce climate change is limited: the radiative forcing (greenhouse warming) caused by tropospheric ozone is only around 22% of that caused by CO210. However, efforts to reduce tropospheric ozone concentrations shouldn’t stop. Every little helps in the fight against climate change, and reductions would help to prevent some of the approximately one million deaths estimated to be caused by tropospheric ozone each year11.

Other mysteries remain in atmospheric chemistry, including the sources of atmospheric organic aerosols, the chemistry of halogen-containing molecules, and how naturally occurring emissions interact with those associated with human activities. But Victorian scientists don’t have much to say about any of those.

Nature 570, 167-168 (2019)

doi: 10.1038/d41586-019-01752-z


  1. 1.

    Yeung, L. Y. et al. Nature 570, 224–227 (2019).

  2. 2.

    Keeling, C. D. Tellus 12, 200–203 (1960).

  3. 3.

    Schultz, M. G. et al. Elementa 3, 000067 (2015).

  4. 4.

    Young, P. J. et al. Atmos. Chem. Phys. 13, 2063–2090 (2013).

  5. 5.

    Intergovernmental Panel on Climate Change. Climate Change 2014: Synthesis Report (IPCC, 2014).

  6. 6.

    Volz, A. & Kley, D. Nature 332, 240–242 (1988).

  7. 7.

    Marenco, A., Gouget, H., Nédélec, P., Pagés, J.-P. & Karcher, F. J. Geophys. Res. Atmos. 99, 16617–16632 (1994).

  8. 8.

    Mickley, L. J., Jacob, D. J. & Rind, D. J. Geophys. Res. Atmos. 106, 3389–3399 (2001).

  9. 9.

    Young, P. J. et al. Elementa 6, 10 (2018).

  10. 10.

    Myhre, G. et al. in Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (eds Stocker, T. F. et al.) Ch. 8, 659–740 (Cambridge Univ. Press, 2013).

  11. 11.

    Malley, C. S. et al. Environ. Health Perspect. 125, 087021 (2017).

Download references

Paid content

Nature Briefing

An essential round-up of science news, opinion and analysis, delivered to your inbox every weekday.