The paper describes experiments to validate the hypothesis that NOx and HOx are generated spatially separated in lightning. This is an important aspect when trying to consider the impact of lightning induced HOx generation on the global oxidizing capacity. The experiments support the hypothesis, but I find that many details are missing or poorly described. Maybe it is necessary to read the earlier papers to understand, but I feel more information in the manuscript would make the paper much better.   

Adding a schematic view of the experiment would be helpful (even if there is a figure in Jenkins 2021 it would be good to have one in this manuscript as well). And even after reading to the end of the paper, I am still wondering what you consider is the size of the core, ie. in your experiments what volume do you consider to contain NOx, and from which volume samples NOx and GTHOS? For what reason did you do pressure dependent experiments? Initially I was thinking that you would look for increased diffusion of NOx out of the core volume with decreasing pressure, but you never draw any conclusions from the results obtained in pressure dependent experiments. Is the only difference the change in rate constants for pressure dependant reaction in MCM with decreasing pressure?

Many details are missing: what is the time-gap between the different sparks, ie. what is the time window for the 10 sparks compared to the total measurement time such as shown in Figure 1? And at what repetition rate you generate spark packets?  It would be interesting to see the raw data of the time evolution of OH and HO2 measurements for a spark packet. How the GTHOS measurements have been synchronized with the discharge? Over what time window do you integrate the peaks? Over what distance do you move the discharge?  What is the time resolution of the NOx analyser and how well resolved are the NOx measurements for one spark packet? You say: “ so any change in the NOx mixing ratio across the different positions was assumed to come from diffusion and not chemical loss. “: can you give an order of magnitude of the NOx loss over the different positions, maybe show a figure? At different pressures? How do you calibrate your HOx measurements? What is the HO2 conversion factor within GTHOS? Line 142: what do you means with “The model experiments ran for 0.5 seconds of experiment time“? How long was the experiment time?

I understand that the conclusion is, that NOx is generated spatially different from HOx. My question is how do you take this into account when you measure HOx and NOx? Because it should be very sensitive to the position of the sampling points for GTHOS and the NOx analyser? Maybe add a Figure 4 type where you indicate the (estimated) size of the core as well as the precise position of the NOx and HOx sampling. Would it be possible to move the sampling points? Maybe at least for the NOx analyser? You correct both species for 15% due to not perfect sampling and diffusion, but I do not clearly understand how you can deduce from the NOx measurement that the correction factor should be the same for the HOx measurements.

When you say “ The laboratory air was found to contain ~20 ppbv of CO “ I guess you mean the purified air that entered the reactor? How did you measure and can you safely assume the absence of other trace gases in the purified air? As any remaining VOCs could generate many different species in the discharge, that could react with OH, it seems to me important to verify the quality of the purified air.

Figure 1: why are there no measurements with higher NOx at the higher pressure? Any good reason? Because you say “OH and HO2 decays became progressively steeper, as shown Figure 1 (970 hPa and 360 hPa)“ and to be really convincing it would have been interesting to see the trace at 970 hPa for 100 ppb NO at least. Did you not do the measurements at higher NO or were there not enough data points available at higher NO (even though the 50ppb data look good enough to still expect good data also at 100ppb)? Even if it seems clear, maybe explain in the legend that HOx is the sum of OH and HOx, ie. the total signal together with (in the main text) some information on the calibration procedure of the HOx signal.

Line 161 and 168: I guess you talk about the Figure 2 and not Figure 3? It would be good to give the absolute concentration of spark NO and not just giving the % so the reader can have an idea of how much this is compared to the added NO.

Line 248: The sentence “All the core LOH is also titrated to <1 pptv (our limit of detection in these experiments) over the same time frame the HONO is generated, so it would not be detected by GTHOS in the laboratory experiments, consistent with our observations.“ is not clear to me: I understand that you cannot make a difference between core and outer LOH in the measurement, so what is consistent with your observations? 

Line 81 : Should probably be 5 kHz.

The scaling in Figure 3c is missing.

In “Spatially separate production of hydrogen oxides and nitric oxide in lightning”, Jenkins et al. report the generation of OH from lightning (LOH) based on laboratory studies and F0AM box model simulations. The authors suggest that OH can persist in the atmosphere as it is generated in the corona sheath, which is spatially separated from the lightning core, where large amounts of NO titrate OH immediately. They further propose that large amounts of HONO are generated in lightning strokes from the reaction of NO with OH.

This research is highly important given the limited knowledge of lightning and particularly the role of OH and suits the scope of ACP. However, I have several important questions which need to be addressed before I can recommend this manuscript for publication. While the introduction is well written and can be followed easily, the methods and results parts need some restructuring and additional information, as it is sometimes difficult to follow the reasoning. The authors often refer to their previous studies - I recommend adding a paragraph on these results as they are important for this manuscript, but not all readers might be familiar with them. Further, a schematic of the experimental set up could help to understand the laboratory methods better. I additionally wonder if the authors could carry out experiments under upper tropospheric conditions, where temperature, pressure and water vapor concentrations are different than those pursued by the authors, but lightning is most relevant. How relevant are the experiments to the actual conditions of lightning in the atmosphere (mostly in the UT)? Please find my specific comments and questions in the following.             

Specific comments:

Line 33 & 46: 100s and 10s could be mistaken for 100 and 10 seconds. Maybe this could be spelled out.

Lines 53 ff: “UV radiation can also make extreme OH…” As far as I understood the authors earlier, the UV radiation is created from the corona. However, here it sounds like corona and UV radiation represent to different mechanism for creating LOH / LHOx. Could you please clarify this? And what would be the mechanism of HOx formation from corona? How much HOx is formed in the corona versus from UV radiation?

Line 61: “spatially separation LHOx and LNO production is possible” Is there a word missing? Or “spatially separate”?

Line 64: A previous reviewer of this manuscript or of another paper? If it refers to a different study, I recommend removing “as suggested by a previous reviewer” and replacing it by a reference to the study it follows up on.

Lines 73 ff.: It could be helpful for the understanding to add a schematic of the experimental set up, including the location of the sampling and the positions of the discharge generation.

Line 76: Is the flow tube “wide enough” to capture the center of the spark and the corona individually?

Line 96: What’s the mixing ratio of NO in the flow tube generated from the spark?

Line 99: Have you tried to carry out an experiment under upper tropospheric conditions, i.e. lower temperature (e.g. 220K) and lower pressures (e.g. 200hPa), where lightning is most frequent? If that’s not possible with your set-up I recommend discussing the implications of different conditions. Reaction (R2) seems to be highly dependent on temperature and pressure

Line 108: Could your detection limit (20ppbv?) for O3 be too low to capture it? If you generate short wave UV radiation from the sparks, O3 should probably be formed both from O2 and from NO2. The amount of O3 generated would be dependent on the amount of NO added and can also react with OH and HO2. Is the timescale of these reactions not relevant to the decay of HOx or can you exclude the impact of O3 via the model simulations?

Lines 114 f.: Could the losses for HOx and NO be different, e.g. through wall effects?

Line 119: Did you also change the location of the sampling or are you referring to the different locations of the discharge generation?

Line 137: Do these cases also include 0ppbv of added NO, only considering spark NOx?

Line 149: Which reactions do you expect to occur? NO + HO2 -> NO2 + OH, NO + OH -> HONO, NO2 + OH -> HNO3? Which one is dominant? How important is recycling of OH through NO + HO2? I recommend adding the reactions already in the introduction to explain the effects of NOx on HOx.

Figure 1: Are the differences between 0ppbv and 50ppbv of added NO really significant for 970hPa?

Line 155: How many measurements are included in each data point?

Line 166 ff.: Do I understand correctly that in your experiment, HOx and NOx is both generated from the spark? So, for the 0% spark NOx case in your model, do you assume an initial HOx mixing ratio based on the experiments? Please clarify.

Figure 2: Is the scaling of the y axis for panels C-F the same as for A-B? The minimum is not visible. And does the OH axis somehow relate to the HO2 axis? - The outline of the figure is a bit irritating (2 black outlines for the left and 3 for the right panels).

Line 186: Does this mean that the default model run only includes HOx from the spark? Or NOx from the spark is increased by x% because the amount of LNOx is uncertain? Does this mean you were previously able to reproduce the HOx decay in the model without assuming spatial separation?

Lines 249 ff.: How about upper tropospheric pressures? Would we expect a lower agreement given that the pressure is even lower?

Table 1: What are the units of these values?

Line 258: Are you able to measure HONO and confirm the model results?