Tropospheric ozone (O3) is a key component of air pollution and an important anthropogenic greenhouse gas1. During the twentieth century, the proliferation of the internal combustion engine, rapid industrialization and land-use change led to a global-scale increase in O3 concentrations2,3; however, the magnitude of this increase is uncertain. Atmospheric chemistry models typically predict4,5,6,7 an increase in the tropospheric O3 burden of between 25 and 50 per cent since 1900, whereas direct measurements made in the late nineteenth century indicate that surface O3 mixing ratios increased by up to 300 per cent8,9,10 over that time period. However, the accuracy and diagnostic power of these measurements remains controversial2. Here we use a record of the clumped-isotope composition of molecular oxygen (18O18O in O2) trapped in polar firn and ice from 1590 to 2016 ad, as well as atmospheric chemistry model simulations, to constrain changes in tropospheric O3 concentrations. We find that during the second half of the twentieth century, the proportion of 18O18O in O2 decreased by 0.03 ± 0.02 parts per thousand (95 per cent confidence interval) below its 1590–1958 ad mean, which implies that tropospheric O3 increased by less than 40 per cent during that time. These results corroborate model predictions of global-scale increases in surface pollution and vegetative stress caused by increasing anthropogenic emissions of O3 precursors4,5,11. We also estimate that the radiative forcing of tropospheric O3 since 1850 ad is probably less than +0.4 watts per square metre, consistent with results from recent climate modelling studies12.
Access optionsAccess options
Subscribe to Journal
Get full journal access for 1 year
only $3.90 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Rent or Buy article
Get time limited or full article access on ReadCube.
All prices are NET prices.
The isotopic data and main LOCK-IN firn data that support the findings of this study are available from the PANGAEA database (https://doi.pangaea.de/10.1594/PANGAEA.901154). The LOCK-IN firn analysis is ongoing, so additional firn data underlying sensitivity tests in Extended Data Fig. 6 will be published elsewhere and made available freely and immediately upon request.
The computer codes used to support the findings of this study are available from the authors upon reasonable request.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
National Research Council. Rethinking the Ozone Problem in Urban and Regional Air Pollution (National Academies Press, 1991).
Cooper, O. R. et al. Global distribution and trends of atmospheric ozone: an observation-based review. Elementa 2, 000029 (2014).
Lamarque, J. F. et al. Historical (1850–2000) gridded anthropogenic and biomass burning emissions of reactive gases and aerosols: methodology and application. Atmos. Chem. Phys. 10, 7017–7039 (2010).
Murray, L. T. et al. Factors controlling variability in the oxidative capacity of the troposphere since the Last Glacial Maximum. Atmos. Chem. Phys. 14, 3589–3622 (2014).
Young, P. J. et al. Pre-industrial to end 21st century projections of tropospheric ozone from the Atmospheric Chemistry and Climate Model Intercomparison Project (ACCMIP). Atmos. Chem. Phys. 13, 2063–2090 (2013).
Parrella, J. P. et al. Tropospheric bromine chemistry: implications for present and pre-industrial ozone and mercury. Atmos. Chem. Phys. 12, 6723–6740 (2012).
Archibald, A. T. et al. Impacts of HOx regeneration and recycling in the oxidation of isoprene: Consequences for the composition of past, present, and future atmospheres. Geophys. Res. Lett. 38, L05804 (2011).
Volz, A. & Kley, D. Evaluation of the Montsouris series of ozone measurements made in the nineteenth century. Nature 332, 240–242 (1988).
Marenco, A., Gouget, H., Nédélec, P. & Pagés, J.-P. Evidence of a long-term increase in tropospheric ozone from Pic du Midi data series: consequences: positive radiative forcing. J. Geophys. Res. 99, 16617–16632 (1994).
Pavelin, E. G., Johnson, C. E., Rughoopth, S. & Toumi, R. Evaluation of pre-industrial surface ozone measurements made using Schönbein’s method. Atmos. Environ. 33, 919–929 (1999).
Worton, D. R. et al. Evidence from firn air for recent decreases in non-methane hydrocarbons and a 20th century increase in nitrogen oxides in the northern hemisphere. Atmos. Environ. 54, 592–602 (2012).
Checa-Garcia, R., Hegglin, M. I., Kinnison, D., Plummer, D. A. & Shine, K. P. Historical tropospheric and stratospheric ozone radiative forcing using the CMIP6 database. Geophys. Res. Lett. 45, 3264–3273 (2018).
Zhang, Y. et al. Tropospheric ozone change from 1980 to 2010 dominated by equatorward redistribution of emissions. Nat. Geosci. 9, 875–879 (2016).
Newland, M. J. et al. Changes to the chemical state of the Northern Hemisphere atmosphere during the second half of the twentieth century. Atmos. Chem. Phys. 17, 8269–8283 (2017).
Alexander, B. et al. Quantifying atmospheric nitrate formation pathways based on a global model of the oxygen isotope composition (Δ17O) of atmospheric nitrate. Atmos. Chem. Phys. 9, 5043–5056 (2009).
Erbland, J. et al. Air–snow transfer of nitrate on the East Antarctic Plateau – part 1: isotopic evidence for a photolytically driven dynamic equilibrium in summer. Atmos. Chem. Phys. 13, 6403–6419 (2013).
Yeung, L. Y., Ash, J. L. & Young, E. D. Rapid photochemical equilibration of isotope bond ordering in O2. J. Geophys. Res. 119, 10552–10566 (2014).
Yeung, L. Y. et al. Isotopic ordering in atmospheric O2 as a tracer of ozone photochemistry and the tropical atmosphere. J. Geophys. Res. Atmos. 121, 12541–12559 (2016).
Taylor, K. C. et al. Dating the Siple Dome (Antarctica) ice core by manual and computer interpretation of annual layering. J. Glaciol. 50, 453–461 (2004).
Vinther, B. M. et al. A synchronized dating of three Greenland ice cores throughout the Holocene. J. Geophys. Res. Atmos. 111, (2006).
Buizert, C. et al. The WAIS Divide deep ice core WD2014 chronology – part 1: methane synchronization (68–31 ka BP) and the gas age–ice age difference. Clim. Past 11, 153–173 (2015).
Legrand, M. et al. Inter-annual variability of surface ozone at coastal (Dumont d’Urville, 2004–2014) and inland (Concordia, 2007–2014) sites in East Antarctica. Atmos. Chem. Phys. 16, 8053–8069 (2016).
Witrant, E. & Martinerie, P. Input estimation from sparse measurements in LPV systems and isotopic ratios in polar firns. IFAC Proc. 46, 659–664 (2013).
Witrant, E. et al. A new multi-gas constrained model of trace gas non-homogeneous transport in firn: evaluation and behaviour at eleven polar sites. Atmos. Chem. Phys. 12, 11465–11483 (2012).
Santer, B. D. et al. Human and natural influences on the changing thermal structure of the atmosphere. Proc. Natl Acad. Sci. USA 110, 17235–17240 (2013).
Boothe, A. C. & Homeyer, C. R. Global large-scale stratosphere–troposphere exchange in modern reanalyses. Atmos. Chem. Phys. 17, 5537–5559 (2017).
Butchart, N. et al. Chemistry–climate model simulations of twenty-first century stratospheric climate and circulation changes. J. Clim. 23, 5349–5374 (2010).
Lin, P. & Fu, Q. Changes in various branches of the Brewer–Dobson circulation from an ensemble of chemistry climate models. J. Geophys. Res. Atmos. 118, 73–84 (2013).
Gelaro, R. et al. The Modern-Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2). J. Clim. 30, 5419–5454 (2017).
Schmidt, G. A. et al. Configuration and assessment of the GISS ModelE2 contributions to the CMIP5 archive. J. Adv. Model. Earth Syst. 6, 141–184 (2014).
Sherwen, T., Evans, M. J., Carpenter, L. J., Schmidt, J. A. & Mickley, L. J. Halogen chemistry reduces tropospheric O3 radiative forcing. Atmos. Chem. Phys. 17, 1557–1569 (2017).
Legrand, M. et al. Alpine ice evidence of a three-fold increase in atmospheric iodine deposition since 1950 in Europe due to increasing oceanic emissions. Proc. Natl Acad. Sci. USA 115, 12136–12141 (2018).
Eiler, J. M. “Clumped-isotope” geochemistry—the study of naturally-occurring, multiply-substituted isotopologues. Earth Planet. Sci. Lett. 262, 309–327 (2007).
Eiler, J. M. & Schauble, E. 18O13C16O in the Earth’s atmosphere. Geochim. Cosmochim. Acta 68, 4767–4777 (2004).
Wang, Z., Schauble, E. A. & Eiler, J. M. Equilibrium thermodynamics of multiply substituted isotopologues of molecular gases. Geochim. Cosmochim. Acta 68, 4779–4797 (2004).
Yeung, L. Y., Hayles, J. A., Hu, H., Ash, J. L. & Sun, T. Scale distortion from pressure baselines as a source of inaccuracy in triple-isotope measurements. Rapid Commun. Mass Spectrom. 32:1811-1821 (2018).
Severinghaus, J. P., Grachev, A., Luz, B. & Caillon, N. A method for precise measurement of argon 40/36 and krypton/argon ratios in trapped air in polar ice with applications to past firn thickness and abrupt climate change in Greenland and at Siple Dome, Antarctica. Geochim. Cosmochim. Acta 67, 325–343 (2003).
Capron, E. et al. Synchronising EDML and NorthGRIP ice cores using δ18O of atmospheric oxygen (δ18Oatm) and CH4 measurements over MIS5 (80–123 kyr). Quat. Sci. Rev. 29, 222–234 (2010).
Schwander, J. et al. The age of the air in the firn and the ice at Summit, Greenland. J. Geophys. Res. Atmos. 98, 2831–2838 (1993).
Buizert, C. et al. Gas transport in firn: multiple-tracer characterisation and model intercomparison for NEEM, Northern Greenland. Atmos. Chem. Phys. 12, 4259–4277 (2012); corrigendum 14, 3571–3572 (2014).
Fourteau, K. et al. Analytical constraints on layered gas trapping and smoothing of atmospheric variability in ice under low-accumulation conditions. Clim. Past 13, 1815–1830 (2017).
Huber, C. et al. Evidence for molecular size dependent gas fractionation in firn air derived from noble gases, oxygen, and nitrogen measurements. Earth Planet. Sci. Lett. 243, 61–73 (2006).
Severinghaus, J. P. & Battle, M. O. Fractionation of gases in polar ice during bubble close-off: new constraints from firn air Ne, Kr and Xe observations. Earth Planet. Sci. Lett. 244, 474–500 (2006).
Battle, M. O. et al. Controls on the movement and composition of firn air at the West Antarctic Ice Sheet Divide. Atmos. Chem. Phys. 11, 11007–11021 (2011); corrigendum 14, 9511 (2014).
Wang, Z. et al. The isotopic record of Northern Hemisphere atmospheric carbon monoxide since 1950: implications for the CO budget. Atmos. Chem. Phys. 12, 4365–4377 (2012).
Trudinger, C. M. et al. Modeling air movement and bubble trapping in firn. J. Geophys. Res. Atmos. 102, 6747–6763 (1997).
Burr, A. et al. Pore morphology of polar firn around closure revealed by X-ray tomography. Cryosphere 12, 2481–2500 (2018).
Yeung, L. Y., Young, E. D. & Schauble, E. A. Measurements of 18O18O and 17O18O in the atmosphere and the influence of isotope-exchange reactions. J. Geophys. Res. 117, D18306 (2012).
Butler, J. H. et al. A record of atmospheric halocarbons during the twentieth century from polar firn air. Nature 399, 749–755 (1999).
Kreutz, K. J., Mayewski, P. A., Whitlow, S. I. & Twickler, M. S. Limited migration of soluble ionic species in a Siple Dome, Antarctica, ice core. Ann. Glaciol. 27, 371–377 (1998).
Mitchell, L. E. et al. Observing and modeling the influence of layering on bubble trapping in polar firn. J. Geophys. Res. Atmos. 120, 2558–2574 (2015).
Ahn, J., Brook, E. J. & Buizert, C. Response of atmospheric CO2 to the abrupt cooling event 8200 years ago. Geophys. Res. Lett. 41, 604–609 (2014).
Rommelaere, V., Arnaud, L. & Barnola, J.-M. Reconstructing recent atmospheric trace gas concentrations from polar firn and bubbly ice data by inverse methods. J. Geophys. Res. Atmos. 102, 30069–30083 (1997).
Extier, T. et al. On the use of δ18Oatm for ice core dating. Quat. Sci. Rev. 185, 244–257 (2018).
Sapart, C. J. et al. Can the carbon isotopic composition of methane be reconstructed from multi-site firn air measurements? Atmos. Chem. Phys. 13, 6993–7005 (2013).
Eastham, S. D., Weisenstein, D. K. & Barrett, S. R. H. Development and evaluation of the unified tropospheric–stratospheric chemistry extension (UCX) for the global chemistry-transport model GEOS-Chem. Atmos. Environ. 89, 52–63 (2014).
Usoskin, I. G. & Kovaltsov, G. A. Production of cosmogenic 7Be isotope in the atmosphere: full 3-D modeling. J. Geophys. Res. Atmos. 113, D12107 (2008).
Yeung, L. Y., Ash, J. L. & Young, E. D. Biological signatures in clumped isotopes of O2. Science 348, 431–434 (2015).
Koch, D. M. & Rind, D. H. Beryllium 10/beryllium 7 as a tracer of stratospheric transport. J. Geophys. Res. 103, 3907–3917 (1998).
Field, C. V., Schmidt, G. A., Koch, D. M. & Salyk, C. Modeling production and climate-related impacts on 10Be concentration in ice cores. J. Geophys. Res. 111, D15107 (2006).
Usoskin, I. G. et al. Short-term production and synoptic influences on atmospheric 7Be concentrations. J. Geophys. Res. 114, D06108 (2009).
Usoskin, I. G., Bazilevskaya, G. A. & Kovaltsov, G. A. Solar modulation parameter for cosmic rays since 1936 reconstructed from ground-based neutron monitors and ionization chambers. J. Geophys. Res. 116, A02104 (2011).
Clette, F. & Lefèvre, L. The new sunspot number: assembling all corrections. Sol. Phys. 291, 2629–2651 (2016).
Fraser-Smith, A. C. Centered and eccentric geomagnetic dipoles and their poles, 1600–1985. Rev. Geophys. 25, 1–16 (1987).
Hansen, J. et al. Climate impact of increasing atmospheric carbon dioxide. Science 213, 957–966 (1981).
Miller, R. L. et al. CMIP5 historical simulations (1850–2012) with GISS ModelE2. J. Adv. Model. Earth Syst. 6, 441–478 (2014).
The NCAR command language, version 6.6.2 (UCAR/NCAR/CISL/TDD, 2019); https://doi.org/10.5065/D6WD3XH5.
This work was supported by the David and Lucile Packard Foundation Science & Engineering Fellowship and by the European Commission’s Seventh Framework Programme ERC2011 under grant agreement number 291062 (ERC ICE&LASERS). We thank M. Twickler, G. Hargreaves and R. Nunn at the National Science Foundation Ice Core Facility for curating and providing ice-core samples for this study. We also thank X. Faïn, A. Lemoine and G. Teste for CO2 and CH4 laboratory measurements at IGE; G. P. Lee and W. Sturges and his team at the University of East Anglia for halocarbon measurements on the LOCK-IN canisters; J. Freitag and C. Florian Schaller at AWI and K. Fourteau at IGE for providing a high-resolution LOCK-IN density profile; E. Le Meur for evaluating the site elevation; K. Fourteau, C. F. Schaller and J. Savarino for discussions; and A. Landais for comments on the manuscript. The LOCK-IN field and scientific programme was funded by Institut Polaire Français Paul-Emile Victor programme number 1153 and Centre National de la Recherche Scientifique INSU/LEFE programme NEVE-CLIMAT. We thank the field personnel at the LOCK-IN site: D. Colin, P. Dordhain, P. Possenti, as well as P. Godon for setting up the field logistics.
Extended data figures and tables
In green: gas age distributions in LOCK-IN firn at depths of 84.2 m (short-dashed line), 98.6 m (long-dashed line), 104.8 m depth (solid line) and 107.65 m (short–long dashed line). In red: gas age distribution of GISP2 ice samples. In blue: gas age distributions of WAIS Divide ice samples, estimated using a diffusivity based on NEEM-EU data (solid line) and NEEM-US data (dashed line). In purple: gas age distribution of Siple Dome ice samples. Source Data
Extended Data Fig. 2 Comparison of the most precise model (in isotopic δ notation23) and the simplified model used to include ice data45.
Left, black stars show ∆36 (‘D-36’, in parts per thousand) data in LOCK-IN firn plotted against mean gas ages with uncertainties (±2 s.e.m., calculated using the pooled standard deviation) shown as vertical bars. Lines represent reconstructed atmospheric trends. The preferred scenario is obtained using a SCRIPPS-based O2 trend (see Methods, ‘Δ36 for firn modelling’) and is constrained by LOCK-IN firn data excluding the deepest value. The black and green solid lines show the best-guess trend obtained with the most precise and simplified models, respectively. Long-dashed lines show the uncertainty envelope. Short-dashed lines show the results of the two models when including the probably contaminated deepest data point (most precise model in green, simplified model in black). Right, δ18O18Ocor data (‘d18O18Ocor’, in parts per thousand; see Methods, ‘Δ36 for firn modelling’) in LOCK-IN firn against depth (symbols), compared with model results (lines). The four model simulations only differ in the very deep firn. The deepest data point at 107.65 m, which is probably contaminated (see text), is not shown in the left panel. Its mean gas age is 110 yr, corresponding to 1906 ad. Source Data
Shown are δ15N values of N2 measured at LSCE (black stars); δ18O values of O2, divided by 2, measured at LSCE (green crosses) and at Rice University (red circles); and δ18O18O values of O2, divided by 4, measured at Rice University (blue circles). The black line shows the barometric slope. Source Data
Shown are the corrections obtained using the δ15N value of N2 measured at LSCE (black stars), the δ18O value of O2 measured at LSCE (green crosses), the δ18O value of O2 measured at Rice University (red circles) and the ∆36 value of O2 measured at Rice University (empty blue circles). Source Data
Shown are results obtained with or without the deepest LOCK-IN data point, and with constant or SCRIPPS-based O2 trend estimates, as well as forward firn model tests of atmospheric model scenarios. Top left, ∆36 data in firn and ice (LOCK-IN in green, GISP2 in red, WAIS Divide in blue, Siple Dome in purple) plotted against mean 18O18O age, compared with atmospheric trends obtained by inverse firn/ice modelling. Shown also are the ±2σ-equivalent uncertainty envelope for the inverse model (long-dashed black lines) and the best-guess trends obtained using: the SCRIPPS-based O2 scenario and excluding the deepest LOCK-IN data point (short-dashed black line); a constant-O2 scenario and excluding the deepest LOCK-IN data point (red line); the SCRIPPS-based O2 scenario and all LOCK-IN data points (dashed grey line); and a constant O2 scenario and all LOCK-IN data points (blue line). Top right, δ18O18Ocor in firn and ice (defined in Methods, ‘Δ36 for firn modelling’; LOCK-IN in green, GISP2 in red, WAIS Divide in blue, Siple Dome in purple) plotted against depth, compared with model results in firn and ice using the SCRIPPS-based O2 scenario. The solid lines show the simulation excluding the deepest LOCK-IN data point and the dashed lines correspond to the simulation with all data points. Bottom left, ∆36 data in firn and ice (same colours as in upper panels) compared with simulated profiles using the forward firn model (LOCK-IN, solid lines; WAIS Divide, long dashed lines; GISP2, short dashed lines; Siple Dome, short–long-dashed lines). Outputs shown correspond to SCRIPPS-based atmospheric concentration trends for O2 and constant values for δ18O and ∆36 (black lines); constant values for O2, δ18O and ∆36 (grey lines; results are nearly the same as the black lines); SCRIPPS-based atmospheric concentration trends for O2 and constant values for δ18O, with the +25% box model scenario for ∆36 (orange lines); the +200% box model scenario for ∆36 (blue lines); and the +300% box model scenario for ∆36 (red lines). Source Data
Extended Data Fig. 6 Results of sensitivity tests on atmospheric trend reconstructions from the inverse firn model.
∆36 data in firn and ice (stars with ±2 s.e.m. uncertainties shown as vertical bars; LOCK-IN in green, GISP2 in red, WAIS Divide in blue, Siple Dome in purple) plotted against mean 18O18O age and compared with modelled atmospheric trends (lines). The solid black line is the preferred scenario, obtained using a SCRIPPS-based O2 concentration trend and excluding the deepest LOCK-IN data point, with its uncertainty envelope shown alongside (dashed black lines). The left panel shows the simulation that includes the deepest LOCK-IN data point (red line), a simulation with the deepest LOCK-IN data point corrected (grey star) from a maximum estimate of 10% surface air contamination (purple line), tests of the sensitivity to the optimal solution (grey lines; see Methods, ‘Sensitivity tests on atmospheric trend reconstructions’), the simulation excluding WAIS Divide data (blue line), and simulations excluding the Siple Dome data (green solid line) or excluding the Siple Dome data and using NEEM-US-data-based diffusivity to simulate WAIS Divide firn (green dashed line). The dashed grey line shows that a straight trend with a weak slope can remain in the uncertainty envelope. The right panel shows tests of LOCK-IN firn physics parameters (green; see Methods, ‘Sensitivity tests on atmospheric trend reconstructions’) and tests of LOCK-IN diffusivity constrained with field data only (blue), all nearly superimposed to the preferred trend. Source Data
a, Firn and ice-core Δ36 values (means of replicates) plotted against mean gas age. b, Kernel-smoothed probability density distributions of bootstrap-resampled mean values of each dataset, showing a significant (P < 0.002) difference between the means of the ice-core and firn (above 105 m) datasets. Uncertainties are omitted for clarity. Pooled standard deviations for each sample type are 0.03‰–0.04‰ (see Methods). C.E., common era. Source Data
Shown are 12-month grid-scale moving averages at the surface near each of the four polar sampling sites between 1850 ad and 2015 ad. Source Data