General comments:

Tropospheric chemistry often begins with a trace gas molecule reacting with one of the tropospheric oxidants: OH radicals, NO₃ radicals, or O₃ molecules. (Reactions initiated by Cl atoms, generated from ClNO₂ photolysis, also contribute to the atmosphere’s oxidising capacity, although Cl chemistry was not considered in this paper.)  Such reactions determine the lifetime of traces gases in the atmosphere; this chemistry also leads to the production of secondary pollutants such as tropospheric ozone, organic nitrates (PAN etc) and secondary aerosol particles. Under most conditions, OH radicals make the largest contribution to the atmosphere’s oxidising capacity. However there are gaps in our detailed understanding of OH sources, notably concerning the importance of the photolysis of nitrous acid (HONO).

This paper reports observations of HONO made in an urban location (Tai’an city) in the central North China plain over a period of several weeks in summer. The authors find that simplest “default” HONO source from the OH + NO reaction (which is not a net source of OH) only accounts for 13% of the observed HONO, averaged over the measurement campaign. Thus a large additional “unknown” HONO source must exist, and the authors explore 6 possibilities. The dominant HONO source at the measurement site is identified as the heterogeneous conversion of NO₂ into HONO on ground surfaces, either photo-enhanced by sunlight during the day (approx. 80% of the source strength), or without light in the dark at night (approx. 70%). Small contributions (of up to around 10% each) variously come from the conversion of NO₂ into HONO on aerosol during the day or at night, the photolysis of particulate nitrate, or direct emissions of HONO (e.g. from traffic or soils). The authors have done a good job to pick apart and characterize the strengths of these different possible HONO sources. This information is valuable to atmospheric chemists and modellers.

In the second half of the paper, the authors employ a zero dimensional box model to investigate what contribution OH derived from HONO makes to the radical chemistry at their measurement site. The model is used to model OH from HONO and from other sources (ozone photolysis, unsaturated VOCs reacting with ozone, etc), sources & sinks of HO₂ and RO₂ radicals, sources & sinks of NO₃, and the changing contributions of OH- and NO₃-initiated oxidation chemistry over the 24 hour cycle. The model’s many outputs are nicely illustrated on well-constructed and informative plots. But this part of the work was less convincing due, primarily, to the simplicity of the 0-D model which cannot consider the vertical gradients in HONO concentrations that are known to exist from ground-based HONO sources. The authors are honest about the limitation of their modelling approach and say that a 1-D box model would provide better conclusions. The nighttime oxidant NO₃ and its reservoir N₂O₅ were not measured in this study, but modeled instead, and this leads to further uncertainties in assessing the relative contributions of OH and NO3 reactivity (as acknowledged by the authors).

Specific comments:

[Comment 1] The success (or otherwise) of the 0-D box model depends on making the right choice for HONO’s mixing layer height (section 3.2.2.4, line 343). The authors choose 50 m for HONO’s MLH because that generates the best agreement between modeled and measured HONO production rates, P(HONO), in Fig 6(b). But Figure 6(b) shows that setting MLH to 35 m instead of 50 m increases P(HONO) by roughly a factor of 1.5; alternatively, setting MLH to 100m decreases P(HONO) by a factor of 2. Thus relatively small changes in the MLH generate big changes in modeling the ground HONO source (i.e. the dominant HONO source), with implications for understanding the OH source strength and radical chemistry. This raises questions that require answers:

* How much do this paper’s conclusions depend on the precise choice of the MLH? – how would the conclusions change if the actual MLH were 35 m or 100 m or 200 m?

* Is there a way to independently verify that the optimum choice of HONO’s MLH is 50 m? – can this MLH value be proved by comparing HONO observations made at the foot site in Tai’an city with HONO measurements at the nearby summit of Mount Tai (using data from the companion paper)? Is the vertical gradient in HONO at this site consistent with other studies e.g. Brown et al (J GeoPhys Res, 118, 8067, 2013; doi:10.1002/jgrd.50537) who also observed vertical gradients in NO₃, N₂O₅ and ClNO₂?

* How does HONO’s MLH change with e.g. wind speed, time of day, daytime versus night? Was a constant MLH = 50 m assumed to model the HONO time series (Fig 7a) and HONO’s average diurnal profile (Fig 7b)? How would variability in MLH affect this work’s conclusions?

[Comment 2] Figures 1 and 2 show detailed time series of the observations, and Table 2 provides statistics (all good). The information that is missing is average diurnal profiles of NO, NO₂, O₃, J(HONO) etc – please add these plots either in section 3.1.1 or the supporting information. Please add the diurnal profile of J(HONO) to one or all of Fig 4, Fig 6 and Fig 7(b) [and Fig S7(b) and Fig S8(b)], so that reader can easily distinguish day vs night, sunrise and sunset, photolytic vs dark HONO sources.

Another reason why I request the diurnal plots of NO, NO₂, O₃ and J(HONO) etc is because the diurnal plot of P(HONO) is not symmetric around 12:00 noon (Fig 4 and Fig 6). The unknown HONO source is faster in the morning than the afternoon, and peaks at 11am. The authors note this asymmetry on line 259 but they don’t explain it. What is the cause of this asymmetry in HONO? And is P(HONO) expected to also be asymmetric at other locations? [I’m guessing it’s not due asymmetry in J(HONO) because P(OH) from ozone photolysis *is* symmetric around 12:00 noon in Fig 14c.]

[Comment 3] How do the conclusions of this paper translate to other locations in China and other countries? In terms of air quality, this is a polluted site, so are conditions here somehow special? For example, because high [NO₂] produces an especially large ground-based HONO source at this site? And/or because high ozone, [O₃] > 100 ppbv, is reacting rapidly with unsaturated VOCs to produce an especially active HO₂ and RO₂ chemistry at this site? For example, NO₃ is traditionally seen as the dominant nighttime oxidant, yet the modeling here in this study shows the sink for isoprene (C₅H₈) at night is approximately 50% to OH and 50% to NO₃ – it is unusual to see such active OH reactions at night. [BTW: I agree with the authors that it is better to measure NO₃ than to model it.]

Technical corrections:

Generally, the paper is written well. However, some of the sentence structure is long and complex, especially those sentences that list multiple measurements, ideas or findings. Consider rewording and/or breaking into separate sentences [e.g. lines 40-43; line 132 (see below)]. Some figure captions also have long and complex sentence structure; they likewise could be improved by breaking them into several shorter sentences, e.g. Fig 9 “Simulated concentrations of (A) OH… and (D) NO3. [New sentence] The different coloured lines show…”. Fig 13: the 3 notes in the figure caption would work better as 3 separate sentences. Fig 4: caption needs separate sentences to explain the main panel and the pie charts.

 Abstract line 32: “Our model’s default HONO source from the OH + NO reaction could only reproduce 13% of the observed HONO…”

Abstract line 37: “A HONO/NOx ratio of 0.7% from direct emissions [plural] was inferred…”. Also please state what are the direct sources of HONO at this measurement site?

Abstract line 43: State clearly that NO₃ concentrations come from the 0-D model. Otherwise, without reading the main body of the paper, readers might assume NO₃ was measured.

Abstract lines 46-48: the sentence repeats itself.

Line 79: NO3 radicals [plural]

Line 99: However, very few observations have been reported of OH and NO3 concentrations…

Line 115: Mt Tai is located in…

Line 117 / end of section 2.1.1: Add a sentence to explain the scientific motivation for making HONO observations at this ground site and at the summit of Mt Tai, and briefly explain what the authors found. Readers should be able to know this without having to find and read the companion paper.

Line 120-121: Here it states the LOPAP measurements took place in 2017. But line 109 and the time axis labels in Fig 1 and 2 show 2018. Which is correct?

Line 132: Reword: 10 minute data were used in X. [New sentence] Hourly data were used in Y.

Line 133: PM2.5 measurements were obtained… [plural]

Line 136: SZA functions [plural]

Line 144: interferences in [not “of”] the NO2 measurements [plural]

Line 169-171: Air masses [plural] observed at this site [delete “was”] originated from… which are shown in the wind rose plot in Figure S2.

Line 256: HONO mixing ratio differences… were calculated… and they are compared with P_unknown in Figure S5. [plurals]

Line 257: What do the authors mean by the word “typical”? – the average diurnal profile of P_unknown inferred from their HONO observations? I doubt P_unknown shown in Fig 4 can typically be applied to all other measurement sites.

Line 278: Add a reference to the recent study of HONO emissions from traffic in a road tunnel by Kramer et al, who found a similar HONO/NOx ratio of 0.85%. https://doi.org/10.5194/acp-20-5231-2020.

Line 291 / Table 4: Why are there two values listed for HONO (and two values for NO and two values for NOx)? Are they the HONO concentrations at the start and end of each plume of fresh emissions? Also units = ppbv?

Line 302: What is meant by “overestimated one” and “popularly used one”?

Line 340: “reflected by” is not the right verb. [I’m not sure what the correct verb should be because I didn’t understand this sentence.]

Line 355: Therefore, the MLH for HONO was set at 50 m, with sensitivity tests performed with the MLH set at 35 and 100 m.

Line 371: the model with the present HONO source parameterizations performed well in predicting… [the original adjective “magnificent” is too strong].

Line 378: reinforcing our conclusion that aerosol-derived sources played only a minor role in daytime HONO formation.

Line 400: rather than “The former one..,”, it is clearer simply to begin the sentence with “HONO emissions from soils may occur…”

Line 447 / Fig 10: I found it difficult (& sometime impossible) to distinguish the different OH sources and sinks, where two shades of the same colour are plotted next to each other. The rainbow of colours is visually pretty, but the plot is impossible to interpret. Likewise P(NO3) and L(NO3).

Line 447 / Fig 10: Why is NO₃ photolysis (during the day) not plotted in the NO₃ reactivity plot, panel (D)? Or is photolysis included in the “inorganic” reactions (orange colour)?

Line 476: hemiterpene emitted by very many species of vegetation…

Line 483 / section 3.3.3.1: In addition to forming particulate nitrate, it should also be noted that HNO3 formation from OH+NO2 and from N2O5 hydrolysis on aqueous aerosol are also the major daytime and nighttime sinks for removing NOx from the atmosphere. Without these NOx sinks, NOx photochemistry would produce tropospheric ozone even more rapidly.

Line 487: the dominant one

Line 556: The default HONO source, NO + OH, significantly underestimated the observed HONO concentrations. [New sentence] This reaction could only account for 13% of the observed HONO, revealing a strong…”

General comments:

HONO and related parameters were measured at the foot of Mt. Tai in the summer 2018. 0-D box model coupled with the MCM were used to explore the budget of HONO, OH, ROx and NO3 radical chemistry. The homogeneous reaction of NO and OH has been adopted as the default HONO source in the box model and account for 12%-15%. The family constraint was used in this Model scenario to correct for interferences of NO2 measurements. Large amount of unknown source of HONO appeared especially on the noontime. Then many sources of HONO were discussed and added in the models. Corrected NO2, direct emission, heterogeneous reactions of NO2, and photolysis reactions were considered in the model. Another part of the manuscript studied the Radical chemistry. The authors gave very detailed consideration on the sources of HONO, and some corrected methods were suggested. These results is meaningful for the development of HONO investigation.

There also existed some problems the authors need to improve the manuscript. The manuscript had two parts, one was about the sources of HONO, the other was about the radical chemistry. The connection between these two parts was not very tightly. The first part, more focused on the sources of HONO which had some relationship with OH radical, but how about NO3？ I suggested the authors gave some descriptions on the connection between these two parts. For example, the significant of first part was that model was corrected more accuracy and could give more accurate results of radicals, such as ROx, NO3? Some relationship of HONO in NO3 chemistry?

Specific comments:

1. The logical of Introduction was not very well. The authors should give more discussions between the relationship of the investigation of HONO sources and radical (ROx + NO3) chemistry.
2. More detailed information of foot site should be presented especially the real environment around the site, which were very useful for the analysis of HONO sources.
3. In 3.1.2. Since the NO2 concentration is not credible by using thermo 42i, how did authors prove that the model results of NO2 correction were reliable. Additionally, the interference could be as high as +75% after adding HNO3 in model simulation, which corrected NO2 was used, consider HNO3 or not? If not, please give the explanation on why not considered HNO3 interference? By the model results, PAN had most impact on the NO2 concentration, how accuracy about the model results of PAN, have compared with observation PAN?
4. What’s the meaning of Fig S3? NOz*??? Line 223, also NOz*?
5. In 3.2.2.1. What was the correlation between ∆HONO/∆NOx data in table 4 and HONO/NOx data in Fig 5. In Fig 5, the phenomenon of “the observed HONO/NOx is convergentas NO/NO2 increases” was unclear, this was not convincing for the further correction on ∆HONO/∆ There were definitely different meanings for ∆HONO/∆NOx and HONO/NOx, why authors choose NOx concentrations?
6. Give the explanation of why HONO from direct emission (HONOemi) is likely significantly overestimated with a constant ∆HONO/∆NOx because of different lifetimes of HONO (τ(HONO)) and NOx (τ(NOx)) in the daytime. Please give the more reasonable explanation of the modified factor of 𝝉(𝑯𝑶𝑵𝑶)/𝝉(𝑵𝑶𝒙) in equation 3, and detail information on the calculation of 𝝉(𝑯𝑶𝑵𝑶)/𝝉(𝑵𝑶𝒙).
7. The observation site is special, how to choose NO2 uptake coefficient on aerosol surfaces and ground surfaces? what’s the reasonable? Why the γa was larger than γa_dark? As shown in Eq-5, photo-enhanced effects had been considered. Similar question also appeared on the γg and γg_dark. Please give the explanation for the higher value of γa and γg.
8. MLH values have great impact on the simulation results, so the reasonable MLH value was very important. Why 50 m was good? please combined the real environment and give the reasonable discussions.
9. Line 380: how HONOemi was included in the model? The HONOemi was not the production rate data by Eq-2 and Eq-3.
10. Line 534: what’s the percentage of HONO contribution to OH radical not considering only HONO and O3 photolysis? From Fig 10, there were many sources in production of OH, and HONO not the most important sources.
11. I can’t understand why put the foot and summit of Mt. Tai together in the title, through the two part manuscripts, the Part I was the results on the summit of Mt. Tai, while Part II was the results on the foot of Mt. Tai. the comparation of these two sites was only given some discussions in this manuscript 3.3.5, but these discussions had no new sights and meaning. Furthermore, the analysis methods were different in these two parts. I suggested the authors revised the title, this manuscript was “Atmospheric Measurements at Mt. Tai-Part II: HONO Budget and Radical (ROx + NO3) Chemistry in the Lower Boundary Layer”.

General comments:

Firstly, the English used in the manuscript is, at times, quite poor and it is recommended that the authors make use of an English-language proofreading service to tidy up the manuscript. However, having said that, the scientific content of the manuscript is generally fine and, after following the specific recommendations below, I would fully support its publication in ACP. One other general comment relates to the consistency of units: please standardise reaction rates with the widely-used ppbv h-1 (sometimes ppbv s-1 is used instead).

Specific comments:

Figure 2 – it would be good to see diurnal profiles of the same key species, especially considering almost all of the other results are presented as diurnal profiles rather than full time series.

Figure 10 – in A and B, the gradient-style colouring makes it difficult to distinguish the categories, please use more contrasting colours as in C and D. Also, for C and D, the caption states that unmeasured species are summarised in the “other” category. Does this refer to intermediates (e.g., OVOCs) generated in the model? i.e., model-generated species that were not measured but do contribute to the model reactivity? Please clarify.

Figure 13 – in the caption, what is meant by the “integration problem”? I cannot find it mentioned in the main text. In general, the caption could do with rewording as it is not the most clear. Also, the caption states that equilibrium reactions were not considered, yet (net) PAN formation is shown in C?

Line 65 – please cite Slater, ACP, 2020 (https://doi.org/10.5194/acp-20-14847-2020) and Whalley, ACP, 2021 (https://doi.org/10.5194/acp-21-2125-2021) as examples of the importance of HONO photolysis in OH formation.

Section 2.2.2 could do with more explanation/clarity, e.g. line 160, “significantly enhanced” by how much?

Line 170 – was any back-trajectory analysis performed?

Line 194 and Table 3 – please cite Crilley, AMT, 2019 (https://doi.org/10.5194/amt-12-6449-2019) as a more recent example of HONO levels in Beijing.

Section 3.1.2/line 208 – misleading – the chemiluminescence method is fine (for the measurement of NO), it is the NO2 (to NO) converter chemistry that gives rise to interferences from other NOy species.

Lines 211, 214, and 215 – what is organic nitrates*? This is defined as RONO2 and ROONO2 in the Figure 3 caption, but please include the same definition in the main text.

Line 224 – the time is given as 11:00, should it really be 13:00?

Line 326 – the EF value of 15.6 appears to have come from your companion paper, therefore please reference this.

Line 433 – please give the average percentage contribution of HO2+NO – same for HONO photolysis on line 434.

Line 446 – delete “and OH” – OH reactivity is not affected by [OH].

Line 501 – give reference for the summertime O3 increase in the NCP.

Line 508 – ozonolysis chemistry also produces RO2