Tropospheric NO2 and O3 Response to COVID‐19 Lockdown Restrictions at the National and Urban Scales in Germany

Abstract This study estimates the influence of anthropogenic emission reductions on nitrogen dioxide (NO2) and ozone (O3) concentration changes in Germany during the COVID‐19 pandemic period using in‐situ surface and Sentinel‐5 Precursor TROPOspheric Monitoring Instrument (TROPOMI) satellite column measurements and GEOS‐Chem model simulations. We show that reductions in anthropogenic emissions in eight German metropolitan areas reduced mean in‐situ (& column) NO2 concentrations by 23 % (& 16 %) between March 21 and June 30, 2020 after accounting for meteorology, whereas the corresponding mean in‐situ O3 concentration increased by 4 % between March 21 and May 31, 2020, and decreased by 3% in June 2020, compared to 2019. In the winter and spring, the degree of NOX saturation of ozone production is stronger than in the summer. This implies that future reductions in NOX emissions in these metropolitan areas are likely to increase ozone pollution during winter and spring if appropriate mitigation measures are not implemented. TROPOMI NO2 concentrations decreased nationwide during the stricter lockdown period after accounting for meteorology with the exception of North‐West Germany which can be attributed to enhanced NOX emissions from agricultural soils.

between 2020 and 2019 are solely due to meteorological influences, that is, differences in wind speed, boundary layer height, photo-chemistry etc.: In order to account for the differences resulting from meteorology and isolate the impact resulting from emission changes we subtract the difference in the simulations from the difference in the observations as follow (Qu et al., 2021), and similarly for ozone: Where, "acc" refers to meteorology accounted for, "obs" refers to in-situ or TROPOMI measured concentrations, and "GC" refers to GEOS-Chem model simulated concentrations. This approach results in values that have accounted for meteorological influence to estimate the concentration changes resulting only from COVID-19 emission reductions.

Tropospheric NO 2 and O 3 : Impact of Meteorological Conditions
Like previous studies (Çelik & İbrahim, 2007;Deroubaix et al., 2021;Ordóñez et al., 2020;Voiculescu et al., 2020), we investigated correlations between satellite and in-situ concentrations on meteorology. The correlation matrix is shown in Figure 1 for the Munich metropolitan area. We find similar correlation behavior between variables for 2019 (no lockdown) and 2020 (lockdown). Generally, satellite and in-situ 2 NO E concentrations have a negative correlation with wind speed, temperature and boundary layer height, for example, as pollutants disperse more at high wind speeds than at low wind speeds. As temperature and sunlight increases, the rate of 2 NO E photochemical loss accelerates, and the planetary boundary layer expands resulting in higher dilution. . Temperature has been shown to have a significant influence on ozone production over Europe under various NO X E conditions (Coates et al., 2016;Melkonyan & Wagner, 2013). In addition, Curci et al. (2009) show that increasing temperature causes an increase in biogenic VOC emissions, which can raise the ozone level, especially in the summer. Future climate conditions in Europe (as a result of global warming) will almost certainly have an impact on ozone pollution (Engardt et al., 2009;Forkel & Knoche, 2006;Meleux et al., 2007;Vautard et al., 2007). Europe may experience more intense and frequent heatwaves and droughts in the future, which will increase wildfire events and, as a result, background ozone levels will increase (De Sario et al., 2013; Meehl & Tebaldi, 2004

Changes in NO 2 and O 3 Concentrations in Germany Due To COVID-19 Lockdown Restrictions
In this study, we compare January through June of 2020 and 2019. This time period is divided into five sections: (a) No lockdown restrictions from January 1 to January 31, 2020. and particulate matter, implying that GC accounts for impacts of meteorology when using precise meteorological data and emission inventories. In our study, we also compare the GC and observed concentrations from 2019 to verify that the GC can reproduce the temporal variability of observed concentration changes. The 2019 (January to June) period is used to validate the GC model simulations as unlike 2020 emissions are not affected by changes resulting from COVID measures. To validate the GC model, we compared GC surface concentrations with in-situ surface concentrations, and GC column densities with TROPOMI column densities ( Figure S6 NO E changes for the study period. It is important to note that the number of TROPOMI cloud-free measurements between 2020 and 2019 may have an impact on results (for Munich, TROPOMI measurements are available for 269 days out of 363 days). In addition, the TROPOMI overpass occurs at 13.30 local time, which may make it less sensitive to traffic-related emissions (peak in the morning from 7 to 9 a.m and evening from 4 to 7 p.m). We conducted two comparisons between 2019 in-situ , which should have better agreement. We use the empirical relationship (Lorente et al., 2019) that includes boundary layer information to convert the surface concentration to column density. The TROPOMI observations agree well with the in-situ observations at the TROPOMI overpass time (mean bias (TROPOMI -in-situ) is about −13%), whereas TROPOMI underestimates 2 NO E compared to the 24-h mean in-situ value (mean bias is about −41.5%) (Figure S9, for Munich). This indicates that TROPOMI is not suitable to directly represent the 24-h mean (daily concentration), which could lead to errors in estimating lockdown effects, because the lockdown primarily reduced traffic-related emissions. Furthermore, the observed satellite column concentration is certainly influenced by the background concentration. The free tropospheric background contributes 70%-80% of background from satellite column observation is complex, because of its non-uniformity (Marais et al., 2018, thus, using column measurements is challenging for estimates of local changes in 2 NO E emissions. In contrast to satellite column measurements, background 2 NO E has little influence (5%-10%) on in-situ surface 2 NO E concentrations (R. F. Silvern et al., 2019). The discrepancies between in-situ and TROPOMI changes primarily results from unaccounted background 2 NO E influence on the satellite observation and that TROPOMI's overpass time makes it less sensitive to overall diurnal emissions. These discrepancies limit the use of satellite column measurements to infer the surface NO X E emission changes.
The 2 NO E column densities in rural locations were also investigated. During the 2020 stricter lockdown period, after accounting for meteorology, slightly increased 2 NO E vertical column densities over North-West Germany are observed compared to 2019 (Figure 4c). We hypothesize that this is due to enhanced soil NO X E emissions over North-West Germany in the 2020 stricter lockdown period (associated with increased temperature over North-West Germany ( Figure S8f); soil NO X E emissions typically increase with temperature (Oikawa et al., 2015). Soil NO X E emissions are expected to be high in the early spring (stricter lockdown period), even though the average temperature in May and June is higher than in the stricter lockdown period, because agricultural practices such as fertilizer application begin and end in the early spring (Ramanantenasoa et al., 2018;Viatte et al., 2020). Fertilized soils have high potential for NO X E emissions (Almaraz et al., 2018;S. Liu et al., 2017;Skiba et al., 2021). Figure S11 shows the monthly mean 3 NH E total column densities over Germany. High 3 NH E total column densities were observed in April as agricultural practices (fertilizer application) began in the early spring. Notably, strong enhancements were observed over North-West Germany. The total column of 3 NH E over North-West Germany in 2020 (strict lockdown period) is higher than in 2019 ( Figure S12). This supports our hypothesis that North-West Germany, which is dominated by Grass and Crop land (ESA CCI, 2020), is an agricultural region, with fertilized soil producing NO X E in elevated-temperature environments.

Ozone Sensitivity to NO X Changes
Like previous studies that reported the urban 2 NO E weekly cycle (Beirle et al., 2003;Ialongo et al., 2020), we also investigate this at the national (Germany) and urban scale (Figures S13 and S14 concentrations, because primary emission activities such as transportation typically decrease on weekends. Studies (Sicard et al., 2020;Wang et al., 2014) have demonstrated that analyzing the difference in weekday versus weekend 3 O E concentrations helps identify the ozone production regime. As NO X E emissions decrease on