Review ‘Background nitrogen dioxide (NO₂) over the United States and its implications for satellite observations and trends: effects of nitrate photolysis, aircraft, and open fires’ by Dang et al.

The study by Dang et al follows up on the ongoing conversation on the importance of free tropospheric NO₂ for satellite retrievals and their interpretation. The paper puts forward interesting hypotheses and sensitivity tests for tropospheric NO₂ that can be seen as guidelines for future research. The prediction of how satellite UV/Vis air mass factors may change in response to increasing aircraft NO_x emissions is exciting, and provides a clear message to the retrieval community to stay on their toes in accounting for possible changes in atmospheric composition.

The paper is very well written and clear in reporting its model-based findings and providing recommendations. The one difficulty I have with judging the relevance of the findings is that many of the statements rely on the confidence the authors have in the accuracy of their baseline simulations of FT NO2. The paper would be much stronger if the authors would evaluate their simulated FT NO2 profiles with aircraft NO2 profiles that have been measured over the contiguous US over the last two decades via campaigns, or with simulations from other chemistry transport models (e.g. GMI). Such a comparison could help to substantiate the claim that including an NO₂ source from nitrate photolysis in the GEOS-Chem simulation can explain ‘a large and increasing contribution from background NO₂ in the free troposphere’.

Another question is why the effect of enhanced NO₂ from nitrate photolysis is present so ubiquitously throughout the free troposphere. I would think that the nitrate aerosol vertical distribution has -on average- lowest mass density in the upper troposphere and highest mass density in the lower troposphere (above the boundary layer), and therefore constitutes a stronger source of NO₂ in the mid-troposphere than in the upper troposphere. But the NO₂ increase from nitrate photolysis shows up throughout the entire free troposphere. Could this be related to stronger radiation levels in the UT? Do the authors also see a spatial gradient in the NO₂ addition with stronger NO₂ additions in the warm seasons, and close to the coasts?

Specific comments

L169-171: please indicate for which wavelengths the photolysis of particulate nitrate occurs.

L296: I propose to (also) cite the study by Castellanos et al. [2015] here, who showed that AMF calculations that explicitly account for aerosol absorption and scattering are on average 10% and up to a factor 2 higher than AMF calculations with implicit aerosol corrections.

References

Castellanos, P., Boersma, K. F., Torres, O., and de Haan, J. F.: OMI tropospheric NO₂ air mass factors over South America: effects of biomass burning aerosols, Atmos. Meas. Tech., 8, 3831–3849, https://doi.org/10.5194/amt-8-3831-2015, 2015.

In order to clarify the causes of the flattening of OMI NO₂ columns over the US after 2009, despite the decreased anthropogenic NOx fluxes, Ruijun Dang et al. investigate the sources and trends of background NO₂ levels over the US. In particular, the paper explores the impact of aerosol nitrate photolysis, aircraft emission trends, and wildfires. The major finding is that the inclusion of aerosol nitrate photolysis in the simulations leads to an annual increase of 13% of the model NO₂ column, and to a 7% decrease of the OMI NO₂ column when using the shape factor (AMF) re-calculated based on the model profiles. Both changes contribute to improve the model agreement with the data. The concomitant increasing trend of aircraft emissions and decline of surface anthropogenic NOx emissions leads to an increase of the AMF over the years (by about 10% between 2009 and 2017). 

However, it is not clear from the paper whether the proposed effects do quantitatively explain the observed near-absence of NO₂ column trend seen in the OMI dataset. The paper builds upon recent papers by the same group: in particular, Shah et al. (2022) had already included particulate nitrate 

photolysis in the GEOS-Chem model. The innovative aspects of the present study work are (i) the use of the model shape factors (accounting for aerosol nitrate photolysis) to re-calculate OMI NO2 columns for 2017; (ii) the estimated effect of the increasing aircraft emissions on the AMFs. This study underscores the importance of the free tropospheric background when interpreting satellite NO2 data. The article could be published if the following concerns are adequately addressed in a revised version. 

Major comments: 

1) A careful model evaluation against particulate nitrate observations is missing but is indispensable. We lack information on uncertainties regarding this important parameter. Does the model skill improve when accounting for the photolysis reaction? Could you provide maps 

of the distributions of aerosol nitrate from GEOS-Chem? 

2) The enhancement factor (EF) which mulitplies the HNO₃ photolysis frequency is only crudely dependent on aerosol composition. Obviously this factor is very uncertain, and furthermore, the enhancement is very likely wavelength-dependent. The shape of the absorption cross section is likely to change considerably in the condensed phase. This should be acknowledged and discussed.  

3) In Table 1, the background sources of lightning, soils and fires are higher in 2017 than in 2009. These sources are very variable though. Could you provide information on the variability of these sources between 2009 and 2017? I wonder whether the results of Figure 4 would change if another year were chosen (e.g. 2016).

4) Please include the results for 2009 in Fig.4 (or in a Supplement). It is important to estimate the difference between the two years. Does the temporal correlation between model and observations improve when including nitrate photolysis?  

5) The initial goal was to explain the OMI NO₂ flattening after 2009. It is not at all clear that you reached your goal. The model should be run for both years with the appropriate emissions, and the resulting NO₂ columns should be compared with the AMF-corrected OMI data. Furthrmore, you need to convince the reader that comparing the years 2009 and 2017 is sufficient to conclude on the trend, given the interannual variability.