Review of “Measurement  report:  Vertical  distribution  of  biogenic  and anthropogenic  secondary  organic  aerosols  in  the  urban  boundary layer over Beijing during late summer” by Ren et al.

The manuscript describes observational results of SOA tracers from a tall tower located in Beijing at the end of summer 2015 for about 5 weeks, which encompassed a period of tighter emission control in the end of august. Daily PM2.5 sampling was conducted at three different heights, allowing in turn to study the vertical profiles of biogenic and anthropogenic tracers. This is a quite interesting dataset, particularly showing how different heights ranging from 8m up to 260m at an urban site can lead to quite distinctive conclusions regarding the contribution of SOA precursors. I do identify though some major issues that need to be addressed prior acceptance.

General comments:

#1 – Interpreting changes in SOA (or their tracers), is highly complex because it depends on meteorology (particularly for BSOA), air mass transport, pre-existing aerosol population and so on. The manuscript generally assumes that if concentration at 260m is higher than at 8m, then it’s regional, otherwise local, but this is oversimplified (a local VOC source could also produce maximum SOA at 260m high, depending on vertical mixing, oxidation time, etc.). Here are some suggestions to help data interpretation: i) Provide a significanly thorough site description. As most of those tracers can be formed within hours (or less), there is a high potential of a source being local. For example, what is the vegetation cover and its type surrounding the site? ii) prior performing back-trajectories, analyze polar plots of tracer concentration vs wind speed/direction to help identify local vs regional formation; iii) add information of meteorology (particularly including solar radiation) on interpreting SOA tracers temporal variability, which is particularly important on assessing the effects of strict emission controls, but also the pollution events. iv) add any possible ancillary measurements (CO, NOx, O3, VOCs) that could help better interpret the observations. For example, if CO and DHOPA is higher at 260m than 8m, than its regional contribution is obvious. Eventually EC could also be used as normalizing parameter; v) add information on PM2.5 levels, and if possible its composition, especially during pollution or parade period to link SOA tracers with PM composition.

#2 - I suggest to change the order section 3 is presented. As it stands it starts highly descriptive and offers only generic interpretations (as P6L7-L12, for example) to explain the dataset. Then, some possible impacts of BVOCs (3.1.2) is given, and then finally the actual tracers are used to interpret the data, considering its oxidation steps and different branching, which is the main advantage of such methods compare to bulk analysis such as WSOC or AMS-like source-apportionment. I suggest beginning this section with a discussion on VOC sources, then, as the tracer profiles are presented, interpret them using first and later stage oxidation products, as well as different branchings.

#3 - I invite the authors to give it a careful and complete read to insure high quality text. I found several typos and reported on technical comments, but it's likely that I missed some.

#4 – Lastly, I find that the number of references can be significantly reduced, by at least a factor 3. Reducing the number of references will improve the readability with a clearer information tracing. For broad claims such as P.2L.2, all those 8 references could be replaced by to the latest IPCC report, for example.

Minor comments:

Could you please use colors instead of circle, stars and triangles for the three heights into all plots? It would significantly improve readability.

P3L19-22: As curiosity, is there simulation results that could complement the results presented here?

P4L24: I’m not an expert on this type of analysis, but I understand that recovery rate is an important part of the quantification process. Why recoveries were not used for correction here?

P4L29-P5L5: several minor issues and confusing sentences, please rewrite them in a clearer manner.

P5L7: please define which additional information.

P5L9: How close where the buildings surrounding the sampling site?

P5L12-14: It’s difficult to see from the plot, but it seems that at times (e.g. end of E1) there are at most 1-2 degrees difference between lowest and highest level, but >10% RH difference between 120m and 260m. The same is not observed during E3, for example. Why is that? It could be interesting to add solar radiation on this plot, for example.

P5L15: Please rewrite.

P5L15: How were defined the pollution episodes?

P5L20: as it stands, it’s difficult to compare OC and WSOC between heights and with variability (std, I assume?), perhaps target only a few values, for the rest it’s listed on Table 1. 

Fig. S1-S2: Have you performed a polar plot analysis of SOA concentration considering wind direction and intensity? This would help identify the role of local vs regional sources before assuming all is long-range transport and could be explained by back-trajectories.

Fig. S4: What are the values showed to the right on the vertical profile? Average and std?

Table S1: Is it for the whole period or just during the period impacted by restrictions linked to the parade. Please correct the caption if that’s not the case.

Table S2: This table is not very clear, with the a’s, b’s and b^b’s. If the objective is identify statistically meaningful difference those can be indicated in bold, for example. Also, please rewrite the caption (perhaps “difference” was meant?).

P6L7-11: Globally I agree with the three points indicated by the authors, but I do not classify them equally to explain the differences on tracer levels among the three heights. I believe that it’s a local vs regional impact (argument #2) that explains such variability. This is a strong result presented by the paper, raising a caveat on observations conducted at 8m (which is already quite high for typical urban sites, ranging usually to 3 or 4 meters) as representative of regional chemistry to be compared with meso-scale 3d models, for example.

Figure2: As suggestion, the caption could be “SOA tracers of (a) isoprene, (b) monoterpenes and (c) sesquiterpenes. Measurement heights are 8m (triangles), 120m (circles) and 260m (star) in PM2.5. Relative mass fractions are shown in (d).” I remind also the authors that monoterpenes and sesquiterpenes make a group of several species (unlike isoprene, which is a single compound), so they should be referred in plural. I suggest modifying other captions as well to reduce repetitions and make easier to understand.

Figure3: Figure difficult to read.

P7L2-3: Be careful not to mix tracer concentration with SOA concentration (as later discussed in section 3.3).

Figure 7: perhaps would be more interesting to compare sum of SOC to WSOC (as a proxy for total SOC), it would probably correspond to about 50% of total SOC.

Section 3.1.3: It could be interesting to calculate enrichment factors during the pollution events (perhaps normalized by deltaEC, or deltaCO, if available, from non-pollution periods).

P10L10-L15: Be mindful that correlation and causality are not the same thing. The fact that there is correlation between isoprene tracers and DHOPA, or that traffic can emit some VOC is not itself an indicative of biogenic-anthropogenic interaction.

P10L25: Difficult to read when so many values are listed with their standard deviation. 

Section 3.3: Could you add a discussion on the SOC mass ranges using the defined uncertainties for the ratios? Is there perhaps more up-to-date values to be used?

P11L12-L14: To improve readability, could you compare Parade with average before and after?

Technical comments:

Please check section numbering, 3.1.2 is repeated, and the reason to change from 3.1.3 to 3.2 is unclear to me.

Fig. 1: Unclear what the authors meant by “Obvious meteorological conditions were found during the sampling period.”

P2L1: I think the authors mean “can impact radiative forcing”.

P2L6: remove “a” between “photooxidation” and “of”.

P2L15: This sentence could be review – changing CCN size also affects the radiative forcing. I suggest “…influencing the climate negatively impacting human health” given that those aspects were already described earlier.

P2L22: “…events in China highlights the urgent...” & “…processes of SOA formation in the atmosphere”

P2L26: “…urban boundary layer are lacking, …”

P2L30-34: Unclear the objective of this sentence, please rewrite to make it clearer.

P3L3: “severe” instead of “serve”.

P3L3-6: Please rewrite this sentence. It feels like it’s repeating several times the same phrase “understand SOA formation mechanisms to improve air quality”.

P3L8: “emission control” & “improve” instead of “guarantee the”.

P3L9: remove “the chemical behaviours and regional  transport  of”

P3L16-17: this sentence is unclear

P3L17: “To THE best of our knowledge”.

P3L19: “megacity in China” or “chinese megacity”.

P3L24: I missed here a more detailed description of the site location itself, such as lat/long for example.

P4L2: Do not skip line here.

P4L3: Change Tem for T, also in the figure.

P4L24: “blank”

P5L8: Do you mean something like: ”Whereas the prevailing winds at 8m were either easterly or westerly, at 260 m the wind direction was dominated by northerlies.”?

P5L22: Table S2

P6L31: “…, while tracers of monoterpenes and sesquiterpenes SOA did not show a marked increase with height.”

P7L1-2: repetition of information.

P8L15: Which other pollution events? I thought they were only three.

P11L12: “Before”.

P11L16: remove “obviously”

P1L27: please rephrase.

The authors present measurements of aerosol mass and composition (i.e., tracer concentrations) at a tall tower in Beijing. Measurements of vertical distributions, as presented here, are valuable and generally fairly scarce, so presenting these measurements is itself of value to the community. I believe the work is useful and worth publishing in this journal, but suffers from some scientific overreach that needs to be addressed first. Specifically, as described below, the authors need to temper the strength of some of their statements to more accurately reflect the strength of their evidence, and the authors need to re-evaluate some of their interpretation of tracer ratios by either providing support from the literature or correcting their claims. 

General comments:

1. It is a little confusing that the methods discuss Parade-based periods, but most of the paper actually is more about the pollution episodes.

2. There is a fair amount of English-language issues, mostly odd phrasing and the like, that should be cleaned up.

3. Interpretation of tracer data is somewhat confusing and does not seem accurate to me, particularly in the case of the isoprene tracers. In particular, the authors' interpretation of the 2-MT/2-MET and the 2-MTs/C5-ATs ratio are not, to the best of my knowledge, grounded in recent literature on the sources of these tracers, and the citations provided by to the authors do not support their interpretations as far as I can tell. 

4. The authors seem to draw fairly broad conclusions from somewhat limited evidence. While the vertical distributions are certainly interesting, some of the assumptions regarding regional vs. local transport, changes in partiotioning, impacts of primarily particles, etc., are not very strongly supported. In some cases, these claims seem to based off of previous work, but that is not always clear. In other cases, these claims are based on tracer ratios, but as described above, it's not clear these claims are always based on a current understanding of the current tracer literature.

Specific comments:

P1L29: This is a very broad statement, which is fine as an opening sentence, but why are these citations specifically selected? Some of the early ones makes sense, but, for instance, what is the information being conveyed by the Huang et al. reference?

P2L6: extra "a"

P2L9-10: This distinction between ASOA and BSOA is a bit out of date, in the years after these citations the idea of ABSOA being dominant became somewhat more accepted. The next sentence clarifies this a bit, but the statement that 90% of SOA is BSOA is more a historical perspective than actually informative so should be re-framed as such or removed.

P3L6: should say decision "makers"

P3L12: What is the basis for the claim that 120 m and 260 m are regional? There are two citations - do they measure boundary layer height? Or model it? Or measure tracer compounds in some way?

P3L27: I'm not sure "the typical urban site" really means anything. The following description of the site is more useful.

P4L6: "OC and EC in an aliquot filter" is phrased oddly and should be re-stated

P4L17: Are the author's sure it is a Hewlett-Packard? The last HP GC I was aware of, at least in the United States, was the 5890, and I thought subsequent GCs and MSs were all sold under the "Agilent" branding (i.e., Agilent 7890 GC and Agilent 5975 MS). But perhaps it is different in other countries?

P4L24: Were they not corrected for recoveries because recovery was near 100%? That should be stated if so.

P5L20-23: These sentences seem to contradict, claiming both increase with height and no significant differences with height.

P527-30: Sometimes it is not clear to me when the authors are making a new claim, vs. stating a previously published result. This statement is one of those examples - is the claim that the lower WSOC:OC ratio at ground level is due to biological aerosols a claim made (and presumably supported) by Wang et al., or is that a new claim here?

P6L4 and P6L29: This sentence is a bit misleading - some monoterpene tracers decrease, but others (MBTCA, HDCCA) increase. Also, there is only one sesquiterpene tracer, so it is a bit tough to make generally claims like this

P7L4: The claim that isoprene is regional and MTs/SQTs are more local is not necessarily true. As the authors note, the vertical distribution could be due to regional transport, but conversely could be due to vertical differences in chemistry and/or partitioning.

P7L26: Is there a citation for the claim that sesquiterpenes are mainly emitted by crops and herbs? I'm not sure that is true, they are released from many plants, particulary for reasons related to chemical signalling and plant protection (e.g., increase SQTs with herbivory: Faiola et al, 10.1021/acsearthspacechem.9b00118)

P8L4: The phase" methacryloyl peroxynitrate (MPAN, e.g. methacrolein, methyl vinyl ketone and methyl butanediols)" is odd, as those latter species are not a subset of MPAN but rather separate compounds

P8L7: I find the use of 2-MT to mean 2-methylthreitol while 2-MTs means the sum of both isomers to be confusing. I would recommend calling the sum 2-MTs (which is fairly standard) and maybe calling the isomers 2-MT_eryth and 2-MT_threi (where "_X" denotes a subscript).

P8L16: I am not aware of work showing that the 2-MT/2-MET ratio is indicative of anything in particular. The two citations in this sentence do not seem to include such claims either. As someone who has thought a fair amount about isoprene and monoterpene tracers, It's not clear to me what this ratio is telling me, or why the authors include it.

P8L20-23: The interpretation of C5-alkene triols as precursors in the oxidation of 2-MTs is confusing to me, to the point of making me feel the authors are interpreting their tracer data through an outdated lens. Since the Wang et al., 2005 paper, lots of work has been done on IEPOX oxidation pathways, and I'm not aware that any of it has made the claim the authors are making here. Even in the Wang et al., 2005 paper, Scheme 1 shows both C5-ATs and 2-MTs to be products of IEPOX (one through addition and one through rearrangement). Since then, there has been a fair amount of work to understand what C5-alkene triols are actually "telling us", in particular from the Surratt group and Goldstein group, and I think both groups would agree it's still not quite clear. See for example: Cui et al. doi.org/10.1039/C8EM00308D and Yee et al. 10.1021/acs.est.0c00805.

P8L23: Typo: "vitations"

P10L28: It would be helpful in the figures and pie charts about source apportionment if they also included what fraction of OC and/or WSOC was not captured by the source approtionment. I think this sentence here is telling me that only 8-13%% of SOC is accounted for in their source apportionment, but it's not totally clear to me.

P11, Sect. 3.4: Are these reductions in total WSOC, or just the fraction of SOC that is captured in the source apportionment?

Figures:

HDCCA is not defined anywhere in the main text

Figure 5. The caption is in the wrong order, and the description of panel (d) is tough to understand.