Fingerprints of the COVID-19 economic downturn and recovery on ozone anomalies at remote high-altitude sites in North America and Western Europe

With a few exceptions, most of the studies on tropospheric ozone (O₃) variability following the restriction measures related to the spread of COVID-19 focused on emissions-prone or urban environments. In this work, we investigated the impact of the pandemic restriction measures on surface O₃ at several high-altitude remote and rural sites across North America and Western Europe. O₃ monthly anomalies, computed with respect to the baseline period 2000–2019, were calculated for 2020 and 2021, to explore the impact of the economic downturn in 2020 and the economic recovery in 2021. A total of 41 high-altitude sites were analyzed: 5 remote stations in western Europe, 19 rural sites in the western US, 4 sites in the western US downwind of highly polluted areas, 4 sites in the eastern US, plus 9 remote sites across the globe to provide a "global" picture for comparison. In 2020, most of European high-altitude sites showed persistent negative anomalies for spring (March-May, i.e., MAM; average of -1.8 ppb) and summer (June-August, i.e., JJA; -2.7 ppb), except for April (1.8 ppb). The pattern was similar in 2021 (-1.8 ppb for both MAM and JJA), except for June (1.8 ppb). The sites in the western US showed similar behavior, with negative anomalies in 2020 (-2.1 ppb for MAM and -0.5 ppb for JJA), and 2021 (-0.5 ppb for MAM). However, the JJA seasonal average was influenced by strong positive anomalies in July, due to the large spread of wildfires in the western US. The polluted rural sites in the western US showed a negative O₃ anomaly for MAM 2020, and a slight recovery in 2021, resulting in a positive anomaly for MAM and alternating positive and negative anomalies in JJA. The eastern US sites were characterized by below baseline numbers for both MAM and JJA 2020, while in 2021 the negative values generated a "dipole" structure, with the western sites influenced by the presence of wildfires. Concerning the rest of the world, a global picture could not be drawn, as the sites, spanning a range of different environments, did not show consistent anomalies across the globe. A few sites did not experience any relevant variations (e.g., the Antarctic sites), while others showed signals of influence from the surrounding environment, or behaviors like those observed for European and US remote sites. We also attempted anomaly attribution by analyzing the behavior of several O₃ precursors (e.g., surface and columnar CO, NO, and NO₂) from the CAMS reanalysis, and the patterns of emissions reductions for 2019, 2020, and 2021.