See attached pdf

Review on Weimer et al., 2022: Effects of denitrification …

In this manuscript, the effects of denitrification on the distribution of HNO3, ClONO2, and O3 in the Arctic and Antarctic polar vortex during spring are investigated. Essentially, simulations performed with the WACCM model are compared to observations with satellite-based MIPAS and ozonesondes. The main focus is on sensitivity simulations where the NAT number density is varied to improve the agreement with the observational data. Based on the current work, the authors recommend a new value for the NAT density for future simulations with the WACCM model.

The manuscript presents an interesting study of the capabilities and limitations of the WACCM model and how the agreement with trace gas observations can be improved.

I recommend the publication of this manuscript. However, some points should be investigated and discussed in more detail beforehand.



General Comments

The most important point I would like to address is the lack of comparison of the simulated NAT density with observations. Although this is the most important set screw for this study, no attempt is made to constrain the NAT density used in this modeling study with observational data. Satellite-based observations of NAT density are mentioned in the introduction but are not discussed further. The model results are compared to more than 10 years of trace gas observations. I assume that observations of NAT density are also available during this period, at least temporarily.

The NAT density now recommended as an input parameter differs by a factor of 20 from the previous one. Although the simplifications of the model might not allow direct integration of the measured quantities, it is necessary to relate the model simulations to observations of the NAT density.

The manuscript states that reduced NAT density in the model setup leads to better agreement between model simulations and observations. And this seems to be true for HNO3, ClONO2 and ozone. But why is this so? What are the underlying physical and chemical processes that are responsible? At least a rudimentary attempt should be made to explain this plausibly.

In general, a more detailed discussion of the simplifications of the model would be helpful to better classify the quality of the presented comparisons between model and observations in this study. For example: the statement that "... so observed NAT particle abundances may therefore not be the best guide for this parameter choice ..." (page 3, line 68) should be discussed in a bit more detail.

Specific comments

Page 4/line 105: Although the term "maximum difference" might be common in the modeling community, it would be useful for a broader readership to see this concept presented and explained in a bit more detail.

Page 5/line 121: "... Turning off all heterogeneous chemistry except for N2O5 + H2O ..." Why is this reaction not turned off?

Figure 1: The labels in the partial figures take up too much space (HetAll ....). This means that the actual content of the figure is not displayed optimally. Since the color coding of the simulations is the same in all subfigures, it could be placed separately next to the figures. The indication "max(d)" should of course remain in the figure.

Figure 1: What does the indication "nprof:...." at the top of the figure mean?

Figure 1 to 4: The volume mixing ration of nitric acid in the stratosphere varies with altitude, similar as ozone (Figure 5). In addition, sedimentation of PSC particles leads to a change in the height distribution of HNO3. In Figures 1 through 4, only altitude-independent concentration is given. Please explain briefly how this value is obtained. Would it not be possible, at least for one case of HNO3, to choose a similar representation as for ozone in Figure 5?

Page 5/line 124: "... In HetAll.1e-2, larger HNO3 values are more common in all panels compared to the other simulations..." Can you explain this result? Why does a higher density in the particle phase lead to a higher gas phase concentration of HNO3 and vice versa? What processes underlie this behavior?

Page 7, line 148: In the manuscript, the discrepancy between model and measurement is attributed to the lower accuracy of the MIPAS instrument. By how much was the accuracy of the measurement reduced compared to the measurements included in Figure 1? Can model processes be excluded for this discrepancy?

Figure 5: The "interquantile range" is very difficult to see in the figure.