Aircraft data reveal rising peak CO over the mid-latitude North Pacific

Airborne measurements from 1991 to 2018 indicate that the most polluted air over the mid-latitude North Pacific upper troposphere became more carbon-monoxide-rich, even as background CO showed a weak downward tendency. In the mid-latitude region, monthly maximum CO increased by 0.31 ppb per month, equivalent to 37.2 ppb per decade, with a 95% confidence interval of 0.06 to 0.57 ppb per month. The interval excludes zero, making this the clearest statistically significant regional trend in the observational analysis.

The finding comes from a multi-decadal synthesis of in-situ aircraft measurements, satellite observations and global chemical transport modeling. The primary aircraft record consists of routine measurements made by the Pacific Greenhouse Gases Measurement and In-service Aircraft for a Global Observing System program from July 2012 through December 2018. Earlier observations came from the PEM-West A and B campaigns in 1991 and 1994, TRACE-P in 2001, HIPPO in 2009–2011 and ATom in 2016–2018. MOPITT Version 8 satellite retrievals, Mauna Loa flask measurements and the Integrated Modelling System supplied spatial, long-term and process context.

The investigators aggregated observations into monthly bins and calculated the mean, median, 25th and 75th percentiles, minimum and maximum. They fitted linear trends separately for each statistic and then repeated the analysis through 10,000 randomly sampled regressions in which monthly measurements were perturbed within a stated ±7% measurement uncertainty. A trend was treated as statistically significant only when its 95% confidence interval excluded zero. This percentile-based approach distinguished the behavior of polluted air masses from background air rather than relying on a mean that could combine the two.

Mean CO had positive but statistically non-significant point estimates in both the higher-latitude and lower-latitude North Pacific domains. The higher-latitude estimate was 0.15 ppb per month, or 18.0 ppb per decade, with a 95% confidence interval of −0.02 to 0.32 ppb per month. The lower-latitude estimate was 0.06 ppb per month, or 7.2 ppb per decade, with a 95% confidence interval of −0.04 to 0.16. Although both intervals included zero, they were predominantly positive and therefore were described as suggestive rather than significant increases.

The maximum behaved differently from the average. Its trend was not significant in the higher-latitude domain, at 0.32 ppb per month with a 95% confidence interval of −0.50 to 1.14. In the mid-latitude North Pacific, however, the maximum rose by 37.2 ppb per decade and the confidence interval excluded zero. Monthly minimum CO, used as a measure of background conditions, had negative point estimates in both domains: −0.01 ppb per month in the higher latitudes and −0.07 ppb per month in the lower latitudes, with respective 95% confidence intervals of −0.07 to 0.05 and −0.15 to 0.01. Neither minimum trend was statistically significant.

Seasonal observations provided evidence for vertical redistribution of the pollutant over the North Pacific. From 2000 to 2018, upper-tropospheric CO was higher during June, July and August than during the preceding winter months, while lower-tropospheric CO showed the opposite seasonal pattern. Upper-tropospheric CO was also negatively correlated with upper-tropospheric ozone and positively correlated with water vapour, relationships consistent with convective lifting of CO-rich lower-tropospheric air during the more active summer circulation. By contrast, the North Atlantic upper troposphere showed statistically significant CO decreases in both the upper and lower troposphere over the 2000–2018 comparison.

The regional contrast followed different upwind emission trajectories. EDGAR version 6.1 estimates used in the analysis indicate that Asian anthropogenic CO emissions rose through much of the study period, while emissions from North America and western Europe generally declined. The authors also identified differing transport and mixing conditions: summertime convection and warm conveyor belts efficiently lift Asian emissions into the North Pacific upper troposphere, whereas the North Atlantic receives a larger fractional contribution of CO-poor stratospheric air. The observations therefore do not support a uniform basin-wide decline, and the paper does not treat the geographic contrast alone as proof of a cause.

The Integrated Modelling System was used as a supporting attribution tool, not as a substitute for the aircraft observations. Its historical simulations used constant anthropogenic emissions after 1990 and therefore could not reproduce post-2003 emission changes. Comparisons with Mauna Loa, IAGOS and MOPITT showed that the model reproduced broad variability and relationships but tended to overestimate low CO and underestimate elevated CO in some aircraft comparisons. MOPITT correlated with IAGOS but generally underestimated aircraft-measured concentrations. These discrepancies limit precision in attribution and explain why the authors emphasized direct upper-tropospheric measurements.

The main limitations are the uneven temporal and spatial sampling of the aircraft record, the inclusion of short campaign datasets that were used primarily for distribution comparisons and baselines, and the limited vertical resolution of satellite retrievals. The analysis window ended in 2018 to preserve continuity with the selected MOPITT product, avoid the altered commercial-flight sampling associated with the COVID-19 period and match the EDGAR inventory through 2018. Consequently, the results characterize trends through 2018 and should not be read as a continuous post-pandemic estimate. The study nevertheless indicates that rising peak pollution can coexist with declining or weakly changing background CO, making the full distribution of upper-tropospheric CO important for evaluating long-term change.

Subject of Research: Trends in upper-tropospheric carbon monoxide over the North Pacific using aircraft observations, satellite data and chemical transport modeling

Article Title: Long-term trends of carbon monoxide over the North Pacific upper troposphere from in-situ airplane measurements 1991–2018

Article References: Wang, K.-Y., Nedelec, P., Thouret, V., Clark, H., Wahner, A., & Petzold, A. (2026). Long-term trends of carbon monoxide over the North Pacific upper troposphere from in-situ airplane measurements 1991–2018. PLOS One, 21(10), e0359855. https://doi.org/10.1371/journal.pone.0359855

Image Credits: AI Generated

DOI: 10.1371/journal.pone.0359855

Keywords: carbon monoxide, North Pacific, upper troposphere, aircraft observations, IAGOS, MOPITT, atmospheric transport, chemical transport model, air pollution trends, Asian emissions, Long-term, trends