the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 19 Feb 2026
Mobile Zenith DOAS Measurements of NO2 and HCHO, and their Relationship to Collocated In-Situ measurements
Abstract. The TRACER-AQ (TRacking Aerosol Convection interactions ExpeRiment- Air Quality) field campaign, performed in Houston, Texas in the summer of 2021, presented opportunities to compare Mobile DOAS (Differential Optical Absorption Spectroscopy) and in-situ measurements of nitrogen dioxide (NO2) and formaldehyde (HCHO), which can be used to obtain additional information about the vertical distribution of trace gas, without the need of a full sky scan.
During TRACER-AQ1, a Zenith DOAS instrument retrieved Tropospheric Vertical Columns of NO2, with typical ranging from 1–3 × 1016 molecules cm−2 as well as HCHO columns, which typically ranged from 2.5–3.5 × 1016 molecules cm−2. The retrieved columns were then compared to in-situ measurements made on the same mobile platform in order to develop a column-to-surface ratio, which describes the vertical heterogeneity of the measured columns. The average NO2 ratio ranged from 100 m to 3 km, and HCHO from 200 m to 5 km. In addition, the Column to Surface Ratio was used on two case days to determine the origin of a local ozone episodes observed by the mobile lab.
- Preprint
(3021 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-5179', Anonymous Referee #1, 20 Mar 2026
-
AC1: 'Reply to RC1', Davis Earley, 21 May 2026
This manuscript by Earley et al. reports on mobile zenith DOAS measurements of
formaldehyde and nitrogen dioxide performed in the Houston area in the summer of 2021.
In addition to most existing studies using mobile DOAS measurements, the DOAS data are
complemented by in-situ data yielding surface concentrations of both trace gases. By
calculating their ratio (CSR), information not only on the horizontal but also on the vertical
distribution can be inferred. Since many (mobile) DOAS applications - to some extend -
assume horizontal and vertical homogeneity which is clearly not always the case, this
study offers an interesting approach. Several relevant aspects, such as a sensitivity study
aiming to address uncertainties in the assumed AMFs used for the analysis of the DOAS
data, are addressed. Lastly, two selected days are discussed in more detail as case
studies. Given the overall content of the manuscript it is within the scope of AMT.
However, despite several improvements compared to the initial submission, it still seems
at several instances that the manuscript is not in its final shape yet. Several major and
minor including technical/editorial aspects should be considered and the manuscript
adjusted before publication in AMT.
General comments:
The consistency of several aspects should be improved throughout the manuscript and the
presentation style further harmonised. Firstly, this refers to a consistent naming, e.g.,
Zenith DOAS vs. zenith DOAS vs. ZSL-DOAS, or the use of plural forms of AMF vs. AMFs as
well as AOS vs. AODs. There should also be a consistent use of abbreviations. These
should be clearly defined at their first appearance and after that they should be
consistently used. Lastly, also the way information is provided might be addressed.
Examples here are the representations of several words (SZAs vs. SZA’s, SSAs vs. SSA’s,
cross sections vs. cross-sections, boxAMF vs. box AMF, etc.). See also the specific
comments below for other examples.
Although there will be a proper language editing after acceptance of the manuscript, there
are several obvious grammatical as well as spelling mistakes, e.g., the use of articles as
well as plural vs. singular forms. Further, several punctuation marks or spaces (especially
before/after references and brackets) are missing. These issues disrupt the flow of reading
in many places and make it difficult for the reader to follow the content. See also detailed
comments below for further examples. Note that such mistakes were only marked in the
first half of the manuscript to give an overall impression. Corrections should be done
throughout the manuscript.
The figures and the presentation style might be further improved. This includes, besides
others, the use of font sizes, the colours involved and the formatting of chemical
structures, such as O2O2 and NO2 in legends, titles and axislabels. Often different panels
are specified as (a), (b), … in the captions but none of them is then used in the actual
figures. Captions of figures showing the same but for different trace gases should be
harmonised. Further details are given in the specific comments.
Main/major comments:
Structure of the paper:
In several places, it seems that paragraphs or even entire sections are misplaced or a shift
to other parts of the manuscript would be beneficial for the presentation quality and to
make the manuscript clearer. In the current shape, several pieces of information are
diffusely spread throughout the manuscript. A non-exhaustive list of suggestions is given
below:
Current Sect. 1.1 could be shifted to Sect. 2 as it mainly introduces the TRACER-AQ
campaign being the basis of the measurements described in Sect. 2. Here, the pieces of
information could be merged with the ones from page 5 (lines 114 - 131). For details, refer
to the specific comments below.
• Section 1.1 was shifted into section 2, and the information about the TRACER-AQ
campaign and instrumentation was moved into this section from other parts of the
paper.
Further, a short paragraph on how the manuscript is organized could be added at the end of
Sect. 1 right after the overview on previous studies on the relationship between VCDs and
surface concentrations (lines 61 – 66).
• The organizational paragraph was already included but was merged with the old
section 1.1, now this paragraph should be more clearly visible now that Section 1.1
has been moved.
The paragraphs on pages 3 and 4 (lines 87 – 101) describe the concept of CSR. However, it
would be more logic if this was an own paragraph/subsection in Sect. 2 which deals with
the measurements and methods used in the manuscript.
• This paragraph was moved to Section 3 to introduce CSR and lead into presentation
of the data.
In the current version, the sensitivity study on the AMFs, including all its subsections (2.2,
2.2.1, 2.3, 2.4, 2.5), is embedded into Sect. 2 which describes the measurements and their
analysis. Since this study is a major part of this work, I suggest to give it an own section
number with the content of 2.2. being the introduction to this new section and then all
other subsections can be just one level below (i.e., X.Y).
• All sensitivity studies subsections were shifted into their own Section, now Section
3, with the section hierarchy updated accordingly.
Information on AERONET data could be merged into one paragraph. In the current version,
this information is spread over several parts of the manuscript and is hard to follow for the
reader.
• Information about AERONET was condensed into the discussion on the TRACER-AQ
field campaign, to explain the relevant instruments, and Sensitivity Study data
preparation section, to explain how the instrument data was used.
A more general recommendation: The manuscript should to some extent be cleaned up
and information should be merged at one location per topic.
More detailed suggestions are given below in the specific comments.
Sensitivity study and result part:
The manuscript includes a good sensitivity study on uncertainties introduced by the use of
fixed AMFs instead of AMFS simulated for each individual measurement. However, it is
recommended to stress the context and the aim of the sensitivity in more detail. Further, it
remains unclear throughout the manuscript which AMFs are actually used to convert the
measured SCDs into VCDs? In my understanding, mean aerosol properties are used to
simulate AMFs to be used in the analysis (page 5, lines 127 – 131). However, I strongly
recommend to clearly state/explain this in more detail. Probably, in Sect. 2.1 after Eq. 2.
• The AMFs used for the SCD conversion (mean aerosol conditions), is more explicitly
stated with further detail. Also, the goals and context of the study are described
more explicitly.
A procedure to filter for cloudy scenes is described in Sect. 3.1. However, it remains
unclear where this filtering is actually applied. Are all data discussed in the result part
(Sect. 3) filtered? As the aim of the filtering is to remove cloudy scenes, it would be good to
state whether it successfully removes cloud contaminated data when O2O2 SCDs are high.
These aspects should be considered in a clear and concise way.
• Information on what data is cloud filtered is now more explicitly stated. All data
discussed in the results section is now stated to have been filtered for clouds.
While the VCDs are nicely discussed in Sect. 3.1, the manuscript lacks a short discussion
of the surface concentrations. This is especially the case for HCHO and should be added.
• More discussion about the behaviors of the in situ data for both trace gases was
added, with an emphasis on the behavior of surface concentrations in relation to
the VCDs.
In Sect. 3.2, it should become clearer where CSR is discussed in general and where the two
case days are discussed. One could also think about adding subsections for both case
days. Further, a clearer motivation why these two days were selected could be added. In
my understanding this is related to the ozone levels. I suggest to stress this more. Further,
O3 episode might be replaced by something like “high O3 episodes”. Also adding some plot
showing the O3 levels in some way would be beneficial to stress the special character of
these days. Lastly, a lot of times and time periods are mentioned in this section. To guide
the reader, I suggest to mark them by lines in the respective plots? Some more details are
mention below in the specific comments.
• Each case day now has its own subsection, and moved the initial discussion on the
CSR to Section 3.1. The goal of the case days is stated more explicitly now, which is
to further understand potential causes of high ozone.
At several instances, more in-depth discussions are missing. For instance, the topic how
the observations relate to ozone chemistry is missing some depth. Further, the relationship
between CRS and boundary layer height was highlighted in the motivation part but lacks a
more detailed discussion in the result part.
• Further comments on the relationship between the behavior of ozone and
observations were added to the text.
Lastly, Sect. 3.3 reads like a motivation rather than a discussion. Parts of the information
provided here might be merged with the introduction part of why CSR is important. Apart
from that, most might be skipped. Probably a more in-depth discussion of the results and
findings of this work could be performed here (compare previous comment) in a similar
way as the last two paragraphs of the section.
• Adjusted Section 3.3 (now Section 4.4) to provide a more in depth discussion on the
results and their implications for future studies.
• Line Item comments can be assumed to have been changed as stated unless
specific responses were given, and can be seen in the tracked changed version of
the manuscript.
Minor/specific and technical comments:
Page 1, line 4: I guess it should read “… vertical distribution of trace gases, …”.
Page 1, line 9: Better something like “…., and for HCHO from…”.
Page 1, line 10: Consider to write something like “… the origin of episodes of enhanced
ozone…”.
• The phrase “ozone episode” was used to satisfy requirements set by the
grant issuer, and cannot be changed
Page 1, lines 17 + 22: Missing spaces.
Page 2, lines 24 + 25: Like for TEMPO, you might add the long version of TROPOMI and the
primary reference of the instrument, namely Veefkind et al., 2012.
Page 2, line 28: Better “… are sensitive to the total column abundance …”?
Page 2, line 31: I suggest to add “… usually smaller than 75°.” since in theory retrievals are
also possible at larger SZAs and this is a more practical aspect.
Page 2, line 34: Consider skipping the word “two” in this sentence as also other nitrogen
oxides exist.
Page 2, lines 38 – 41: The paragraph first is about NOx and later only NO2 is mentioned.
Consider to mention that (passive) DOAS measurements cannot detect NO and that the
study here is therefore limited to NO2. Then the considerations about NO2 can follow. Since
the lifetime can vary depending on the atmospheric conditions, a range of how far
emissions are concentrated around their sources might be added.
Page 2, lines 43 – 49: This paragraph about HCHO reads rather diffuse. I suggest to
restructure it in a more concise way. In particular, the fact that HCHO is an intermediate
product of VOC oxidation is mentioned twice.
Page 2, line 46: Better “VOCs” rather than “VOC’s”?
Page 3, line 69: Year of publication of “Judd et al.” is missing.
Page 3, line 80: Probably skip the word “retrieved” in this sentence.
Page 3, line 81: Check formatting of O3.
Page 3, line 82: Missing space and “was operated” should be skipped.
Table 1: The abbreviation VCD was so far not introduce. I suggest to do so in the
introduction at the first occurrence of “vertical column density”. Further, pressure should
be abbreviated by p rather than P. Also, AERONET and its relevant products (SSA, ASY, AOD)
should be mentioned in the text, since the abbreviations might not be familiar to everyone
and were not yet introduced.
Page 4, lines 92 + 93: I guess it should read something like “… to the total columns of trace
gases, and in situ measurements sample only surface air.”
Page 4, lines 96 + 97: A word seem to be missing.
Page 4, line 99: One “layer” too much.
Page 4, line 105: Should it read “… trace gas profiles.”?
Page 4, lines 103 – 112: This paragraph could be the last part of section 1 as mentioned in
the main comments.
Page 4, line 106: Full stop (.) is missing. This sentence might be the one regarding the
AERONET properties mentioned above and could go there.
Page 4, line 110: Should it be “… the AMF error…”?
Page 5, line 123: Please check formatting of 1.5°. In my understanding, this refers to the
FOV of the sky mode. If so, please add this information to the text.
Page 5, lines 114 – 131: Many of this information describes aspects summarised in Table 1
and is related to the description of the TRACER-AQ campaign. Also, some pieces of
information are doubled. Please consider merging these pieces into one subsection
summarising the campaign. My suggestion would be to do this in Sect. 2 as indicated
above (main comments).
Page 5, line 138: Better write “accounted for” instead of “estimated and removed”.
Pages 5 + 6, lines 139 – 146: First, I suggest to write something like “Equation 1 is the so
called DOAS equation …”. Next, in the description of Eq. 1 it should be a Δτ being the
differential optical depth, I guess. Further, I wonder if σi is really a differential cross-section
or a full one? In my understanding differential here refers to the usage of a reference
spectrum. Consider to revisit the DOAS description a little bit taking these suggestions into
account.
• The isolation of the differential absorption cross-section is achieved by
taking the difference between measurement and reference.
Page 6, line 145: Consider adding “(ΔSCD)”, to stress that this is the actual output of a
DOAS fit.
Page 6, line 150: Please be consistent in writing cross-section vs. cross section throughout
the manuscript.
Page 6, line 151: “…measured…” instead of “… measures …”.
Page 6, line 154: “Table 2 lists the corresponding…” as well as “… corrections were
performed …”.
Page 6, line 157: Please add a reference for the CINDI-2 campaign, e.g., Kreher et al., 2020.
Regarding the I0-correction: not I0 equals some value but a typical SCD is assumed for
performing this correction as for instance described by Aliwell et al., 2002. Further, the unit
for these SCDs is missing here (molec cm-2).
Table 2: In the caption you might use the abbreviations for UV and VIS as done above.
Please cross-check whether the temperatures for the BrO and HCHO cross-sections are
correct or if they should be 223 K and 298 K (or the cross-section from 1993), respectively.
Page 6, line 158: Probably better write “The retrievals of …”.
Page 6, lines 160 – 162: It would be more precise to write something like “The SCD is a trace
gas concentration integrated along the average photon path and … wavelength accounting
for…”.
Page 6, line 163: Are you referring to the calculation of VCDs here? If so, I suggest to write
“… derive the tropospheric VCD from the zenith measurements …”.
Equation 2: I strongly recommend to harmonise the quantities and their representations in
equations (compare Eq. 1 and 2) as well as the text. For example, so far ΔSCD was used
and now this is referred to as ΔS. Also, their dependencies (T, λ, etc.) should be consistently
indicated in both equations. Lastly, it is not necessary to provide the unit here in brackets.
Page 7, line 168: “… at a fixed location, …”
Page 7, line 170: “… 8 km from the coast.”
Page 7, lines 170 – 174: Was the cloud screening described below already applied for the
creation of the universal reference spectrum? This information might be added to the
description.
Page 7, lines 177 – 180: I find these two sentences contradicting. Therefore, I suggest to
write in the first sentence “…., most previous Mobile….” and in the second sentence “Other
mobile DOAS studies, however, …”.
Page 7, 1st bullet point: As suggested above (Table 1), please introduce these abbreviations
already before. Also, the listing of the AERONET sites might be included in the section
about TRACER-AQ when the instrumentation is introduced. Here, it could then be stated
that “..aerosol properties were from these AERONET stations as described above.”.
Page 7, 3rd bullet point: Better “representing” rather than “estimating”?
Page 8, line 195: “… using the LibRadtran/MYSTIC …”
Page 8, line 196: Better “… in terms of the so-called box AMF (bAMF).”?
Page 8, line 197: “…, the trace gas bAMF profile …“ instead of “… (as function of height) ”
Page 8, lines 195 – 199: Probably, add a reference for bAMFs.
Page 8, line 199: Briefly mention the limitations of the HG phase function. Further, I would
argue that also the albedo of the surrounding matters. Could you comment on that?
Page 8, Sect. 2.2.1.: As indicated in the main comments, this subsection could go one level
up in a potential new section on the sensitivity study.
Page 8, line 207: “… the aerosol type in the RT simulations.”
Page 8, lines 208 + 209: I suggest to mention to which wavelengths the properties were
interpolated to (the ones mentioned above). Also mention how this was done, since I guess
also here an Angstrom parameter was used?
Page 8, line 211: Please remove the extra “The”
Page 8, line 211: Better “For that, the ceilometer signal …”?
Equation 3: Please mention in the text that N is the normalised profile.
Equation 5: Please cross-check this equation. If I am not wrong, it should be -α in the
exponent.
Page 9, lines 226 - 228: I cannot fully understand how these values are obtained. Shouldn’t
it be 1120 profiles? Please cross-check. Probably also comment how these three scenarios
are selected/defined.
• The number of scenarios was incorrect, corrected both total scenario
number and total profile number, which also included header lines
Figure 1: Please cross-check the indicated colour: 460 nm should be black in my
understanding.
Page 9, line 231: Please chose a consistent way of writing LibRadtran (compare page 8, line
195).
Page 9, line 232: “… and RH data …”. Please abbreviate pressure as p instead of P.
Page 9, line 233: Above (Sect. 2.2) eight profiles were mentioned.
Table 3, caption: “… in the the sensitivity study.”. It seems that the table is too wide too fit on
the page.
Figure 2: A space is missing before the bracket of the reference. Consider changing the x
label to Trace Gas Concentration. Brown line seems to be no box, is this a simple sampling
artefact?
• Sampling artefact causes brown line to not be a perfect box when plotted
Page 9, lines 236 + 237: Could you comment in how far this assumption is justified?
Page 10, lines 241+ 242: Please provide the values for both concentrations.
Section 2.3: It becomes not clear which AOD scenario was used for the investigations in
this section. This information should probably be added.
Figure 3: Consider using single scattering albedo or SSA as y-label in the left panel.
Page 11, line 245: “sites” instead of “sited”
Page 11, line 253: Better use “change” or so rather than “shift”?
Page 11, lines 253 – 255: Please consider re-phrasing this long sentence to improve clarity.
For instance, what does it mean that SSA deviates from the minimum AMF? Please
consistently uses VIS or visible throughout the manuscript (compare main comments).
Page 11, line 259: Could you briefly comment how these values are obtained? Looking on
the grey line, I have the impression that the change of the AMFs from min. to max. SSA is
almost 10 %. Or is this just a misunderstanding?
Page 12, lines 262 – 264: Please check this sentence.
Figure 4: Consider using single scattering albedo or SSA as respective x-axis labels. If
possible, adjust formatting of O2O2.
Figure 5: I suggest to write “hourly mean L2 AOD values …” instead of “averaged every
hour”. I wonder why this figure is not Fig. 1, as it only referred to on page 8 (line 204)?
Page 12, lines 269 – 271: Where is the information on the SZA in Fig. 6? Please add this
information in the text.
Page 13, line 283: It might help to mention how the difference is calculated to get a more
intuitive idea of Fig 7. because difference between O2O2 and trace gas (line 281) sounds
contra intuitive to Fig. 7 (negative values). I assume its AMFtrace gas – AMFO2O2 shown in Fig. 7?
Page 13, line 285: “… high solar zenith angles (morning and evening) …”. Please skip the
word “other”.
Figure 7: No panels are marked but mentioned in the caption. Is there a blue line (profile 1)
in fig. 7? If so, please mention in the text that is covered by another (orange?) line.
Figure 8: Consider using “Trace gas AMF” as y-label. Differences are mentioned in the text
but absolute AMFs are shown this should be harmonised. No panels marked as mentioned
in the caption. Please check which column is which wavelength since titles and caption are
contradicting each other.
Page 14, lines 290 + 291: Better “trace gas profiles” instead of “other profiles”. In general,
clearly use “trace gas profiles” if they are referred to in this paragraph.
Page 14, lines 293 – 295: This sentence reads rather complicated. Probably just state that
the differences between the trace gas profiles are smaller than the differences to the O2O2
profile which can also be seen in Fig. 2. since the profiles deviate less. Better “changes in
the vertical profile” rather than “changing of the vertical profile”, since you are referring to
the final measurements?
Page 14, line 300 onwards: What is meant with constant profile?
Page 15, line 303: Please consider using ZSL or Zenith/zenith DOAS (one of both)
throughout the manuscript.
Page 15 lines 307 – 309: This is unclear. Didn’t you state above that the choice of the trace
gas profile is not crucial (page 14, lines 294 +295)? Probably comment on that.
Page 15, line 310 onwards: Which AMFs are actually used for the analyses of the
measurement data in this work? This question arises several times when reading the
manuscript (compare also main comments).
Consider adding a short paragraph where the main findings and the uncertainties related to
the AMFs are briefly summarised. This would also help to follow the discussions in Sect.
2.6.
Section 2.5.1: I wonder if this is the best place for this section. I have no clear idea where to
put it but it is not directly related to the sensitivity study. When restructuring the sections
this section might go to the measurement and analysis description section (Sect. 2). Add a
reference for the Langley plot method.
Page 17, line 321: What does 2.2-4.2 SCDref refer to? Is a full stop missing?
Page 17, line 322: “… for the UV retrieval …”. How where the errors estimated?
Page 17, lines 326 – 328: What is the value of the stratospheric VCD? Only the uncertainty
is mentioned.
Page 18, line 341 onwards: In the current version, Sect. 3 is the results section. Please use
the correct numbering (after restructuring). How is this AMF uncertainty exactly
calculated? Are different values for different spectral ranges from above considered?
Estimation of the individual terms remains vague. Moreover, the values for the error of the
reference SCD differs from the ones listed in Sect. 2.5.1.. I suggest to provide more
information on how these values were obtained.
Page 18, line 351 – 353: Check if words are missing this sentence as it is difficult to read.
Table 4: Also values for NO2 UV are shown please modify caption.
Page 18, line 355: Consider removing “… of the month.”.
Page 18, line 362 onwards: I suggest to introduce this abbreviation and its long form already
above in the measurement section. “… surface concentration …”.
Page 19, line 367: Consider to use the abbreviation VCD here. “… are also plotted in black.”
Figures 9 + 10: Probably use VCD as y-label? The legend is very small, probably enlarge it.
UV NO2 VCDs seem to be consistently smaller. Can you comment on that since for VCDs
radiative effects are already accounted for? Again, which AMFs are used (compare main
comment above)? Do they not account for that?
Page 19; line 386: The abbreviation dAMF can be introduced above in line 378 where it is
first used. Where is this filter actually applied? This is not mentioned (compare main
comment above).
Page 19, lines 388 – 391: Consider rearranging this sentence to improve clarity. Fig. 10 is
not referred to, which should certainly be done.
Page 20, line 399: I would argue on 25 September the correlation is not good for HCHO,
especially the large peak at 9 a.m. is only seen by the in-situ instrument.
• Added condition “afternoon of the 25th”
Page 20, line 401: “… in the trace gas abundance …”
Pahe 20, line 403 onwards: This part is not completely clear. I would rather write “…
depends on the specific vertical distribution of the trace gas …” or so. Also, the “while
sentence” is not clear. I recommend to say that the combination of both allows to infer
information on the vertical distribution and to mention that this is actually the CSR which is
then described in the next sentences.
Page 21, line 417: “… averaged measurements”, Probably better use “tens of metres” since
“10 s” looks like seconds. Also values above 2 km occur frequently. Better write “… up to a
few kilometres …”?
Page 21, line 420: Probably again mention on which days the instrument was the stationary.
“… by consistent values …”, and I suggest to add example days for this.
Page 21, line 428 onwards: This is confusing. The paragraph before deals with HCHO and
then there is a jump to NO2 again. I strongly recommend to rearrange these paragraphs to
make the content more fluent and logical. Further, I wonder if Fig. 11 could already be
mentioned here, since the described aspects are clearly visible there rather than in Fig. 8.
In fact, Fig .11 is not referred to at all in the current version.
Figure 11: Again, no panels marked in the figure but mentioned in the caption.
Page 22, line 436: At 14:30 it seems that a local minimum in CSR is visible. I think it’s more
14:15.
Page 23, line 443: In Fig. A7, a clear increase of HCHO is visible. I guess it’s the earlier peak
you are referring to. I suggest to mention Fig. 10 or even A7 here.
Page 23, lines 449: Check the mentioned times as they are not consistent to the caption of
Fig. A8. “… the air masses originated largely from …”
Page 23, line 454: “… NO2 and HCHO …”
Page 24, line 460: This information is already given before.
Page 25, line 475: I think it should be VCDs since they show the spikes in Fig. 13.
Page 25, line 477: “… in both HCHO and NO2 VCDs, the NO2 concentration …”
Figure 13: Caption of this figure and Fig. 11 should be the same with the difference only in
trace gas shown. Consider using the abbreviation VCDs.
Page 25, line 479 onwards: This paragraph is unclear. Please elaborate more what it means
that the transported trace gases interact with the HCHO present? If not, this part of the
sentence might be skipped.
Page 25, line 480: Regarding Fig. A11 please mention first that the plot shows data from a
different day to guide the reader.
Page 26, lines 484 onwards: This paragraph has no real link to before as no wind arrows are
shown in Fig. A11 the way it is the case in Fig. 12. I suggest to put this information to a
different part of the manuscript where it is more relevant. This could be either the part
where the measurements are described or where the wind arrows are shown.
• Added more info to make paragraph more relevant to September 9
discussion
Page 26, line 491: Consider skipping “more than”.
Page 26, line 496: “… trace gas distribution”
Page 26, line 497: It is unclear what exactly can be validated. Please comment on that.
Page 26, line 500: Isn’t that what you do? I am little bit confused at this stage.
Page 26, line 517: Please skip one “both”.
Section 4: I really like the way the conclusions are written. More of this very precise and
clear sentences should be written during the whole result and sensitivity study parts
(compare comments given so far).
Figures A1, A2 and A3: In the caption better use “…retrieved in the VIS/UV” and “AMFs at”. Is
22 September really the reference day? Before it was mentioned that 21 September is the
day on which the reference spectrum was determined. I suggest to skip “reference day”
here to avoid confusion. There are no red dots and lines in the figure. Please check that.
Further, the second sentence in A1 and A2 is unclear. What exactly are these bins, probably
elaborate this in the main text.
Figure A4: There are O2O2 data for days were no NO2 is available, why?
• The O2O2 data shown is before both cloud filtering and data quality flagging,
for instance Sept 12-17 was a period of heavy clouds due to an incoming
hurricane into the Houston area. The unfiltered O2O2 data was used
extensively for cloud filtering, and is left unfiltered to provide a comparison of
before and after quality checking of the data.
Figure A9: Better “… on that day.” rather than “… during this time frame.”?
Figure A10: “… were taken.”
Figure A11: Is the sentence about reference spectrum really relevant here?Citation: https://doi.org/10.5194/egusphere-2025-5179-AC1
-
AC1: 'Reply to RC1', Davis Earley, 21 May 2026
-
RC2: 'Comment on egusphere-2025-5179', Anonymous Referee #2, 23 Apr 2026
General Comments:
The paper by Earley et al. presents a study from a mobile-DOAS observation made in Houston in 2021. Tropospheric vertical columns of NO2 and HCHO were successfully derived from the mobile-DOAS measurements, with some detailed error estimations. The author also tried to study the column-to-surface ratio (CSR) of NO2 and HCHO, utilizing the remote sensing observed column values and in situ surface results. The author claims CSR can be a good indicator for trace gas vertical distribution. However, I think the CSR results here will not be meaningful before removing their correlation with meteorological factors, such as boundary layer height. Some discussion and explanation are needed. Nevertheless, this work is a good match with AMT and within the scope of the journal. I recommend publishing this work after addressing the following issues
Specific Comments:
Line 223: the Angstrom number that was used should be provided here, with some minimal justification.
Line 230: The approximate time of ozonesonde lunches for each day should be provided.
Line 268: the discussion of Figure 6 started here, just after the discussion of Figure 4. The order of the figures or the narrative of the results should be modified to enable the reader to follow your thought.
Figure 6: The figure is informative, but it’s more useful if we have a % view (i.e., the difference divided by the mean condition value). By the end of the day, the same absolute difference of AMF has a very different impact on the results.
Line 290-297: should provide clear arguments why this correction is needed for O2O2 AMF. Also, how has this correction been made? Is it just by using the linear relation found in Figure 7? I see the author try to justify the error for AMF, but I am not sure I understand the logic. Better wording is needed.
Line 310-314: the whole section needs some rewording. If the author is to provide a method, the logic is to state it first, and then perform the justification, and lastly discuss the benefits. It reads like the author proposes to estimate the O2O2 AMF from sondes and perform a “correction” to get the NO2 AMF. Anyway, I am not sure we should call such a process as a “correction”.
Line 326-328: Stratospheric column of NO2? Please make this clear. Also, is this a monthly climatology with interpolated hourly resolution? Such basic information should be presented.
Line 321-325: How did you estimate the error for SCDref? Is this just a fitting error? If not, please provide some information. HCHO Langley fit looks more scattered, but only has a ~10% error, while NO2 UV has ~30% error?
Figure A1 to A3: The captions of the figures do not match the plots. I can not see any red dot or red line. I could not see any binning. From the captions, it sounds like you are using the modified Langley method. Also, as stated, the Langley method relies on the assumption that the total column is stable for the entire fitting period (half-day). The figures A1 to A3 look wrong. It is important to verify this for both NO2 and HCHO. i.e., random noise in the total column is not the worst, but linear changes in a half-day period will induce a large error in your SCDref derived. Such work can be done using Pandora direct-sun observations. Also, the HCHO fitting (A3) looks much noisier than NO2. Any comments?
Line 345-346: What do you mean by real data? This is not clear to me. Also, except for HCHO, the numbers here don’t seem to match the errors for SCDref stated in Line 323. Please double-check those numbers.
Table 4 and Line 351-353: This table is good and informative. Maybe state this is on a 1-sigma level? However, the large difference between NO2 vis and uv error estimates is from the median VCD of NO2 retrieved in vis and uv. I.e., as the median UV NO2 VCD is much lower, it translates to a higher % error. However, we all know that there is only one real NO2 column value, no matter which wavelength you select in measurement. Can you plot the histogram of both vis and uv NO2 columns (entire records, and/or only coincident records)? Why do we have a >50% offsets between them (on their median). Do they still have a similar distribution? Theoretically, UV and VIS results are basically sampling different airmasses. However, a 50% offsets still large and should be explained. Can you do some estimation on the horizontal sampling areas for these two settings? Anyway, if the author believes NO2 UV truly have doubled the error than NO2 VIS, please provide some discussions and suggestions.
Line 515-522: More quantitative analysis is needed to support the claim that CSR can be used for meaningful information on trace gas vertical distributions. As discussed here, the method will be limited by the sensitivity of in situ and mobile-DOAS. In which range, do you think CSR provide useful knowledge on mixing height? CSR vs. MAX-DOAS profiling results might provide more information. Also, CSR will be strongly affected by factors such as BLH, which is measured by your ceilometer. Have your test their relationship? I think if you can “remove” the impact of daily BLH changes, the modified CSR might truly give you more information on vertical distribution.
Technical Comments:
Line 208: The first time the discussion of a figure is Figure 5, which is very strange. If Figure 5 has to be mentioned here, it should be renamed and relocated as Figure 1.
Line 211: backscatter “of “? Rewrite the sentence to make it clearer.
Citation: https://doi.org/10.5194/egusphere-2025-5179-RC2 -
AC2: 'Reply to RC2', Davis Earley, 21 May 2026
The paper by Earley et al. presents a study from a mobile-DOAS observation made in Houston in
2021. Tropospheric vertical columns of NO2 and HCHO were successfully derived from the
mobile-DOAS measurements, with some detailed error estimations. The author also tried to
study the column-to-surface ratio (CSR) of NO2 and HCHO, utilizing the remote sensing
observed column values and in situ surface results. The author claims CSR can be a good
indicator for trace gas vertical distribution. However, I think the CSR results here will not be
meaningful before removing their correlation with meteorological factors, such as boundary
layer height. Some discussion and explanation are needed. Nevertheless, this work is a good
match with AMT and within the scope of the journal. I recommend publishing this work after
addressing the following issues
Specific Comments:
Line 223: the Angstrom number that was used should be provided here, with some minimal
justification.
• The exponent was calculated from the AERONET data for each aerosol scenario and
therefore is unique to each scenario. This explanation has been made more clear in the
text.
Line 230: The approximate time of ozonesonde lunches for each day should be provided.
• Ozonesonde launches were not performed at set times, but did occur routinely between
12:00 and 20:00 UTC depending on the number of launches that day. This is now stated
more clearly in the text to improve clarity.
Line 268: the discussion of Figure 6 started here, just after the discussion of Figure 4. The order
of the figures or the narrative of the results should be modified to enable the reader to follow
your thought.
• Adjusted Figure 6 to Figure 1, and moved figure location closer to where it is mentioned
in the text
Figure 6: The figure is informative, but it’s more useful if we have a % view (i.e., the difference
divided by the mean condition value). By the end of the day, the same absolute difference of
AMF has a very different impact on the results.
• Changed figure to percent difference, as well as description of figure in the text.
Line 290-297: should provide clear arguments why this correction is needed for O2O2 AMF.
Also, how has this correction been made? Is it just by using the linear relation found in Figure 7?
I see the author try to justify the error for AMF, but I am not sure I understand the logic. Better
wording is needed.
Line 310-314: the whole section needs some rewording. If the author is to provide a method,
the logic is to state it first, and then perform the justification, and lastly discuss the benefits. It
reads like the author proposes to estimate the O2O2 AMF from sondes and perform a
“correction” to get the NO2 AMF. Anyway, I am not sure we should call such a process as a
“correction”.
• Lines 290-314 response: The discussion on the modification of O2O2 AMFs to match the
AMFs calculated from the seven profiles used for the sensitivity study was edited to
improve clarity on the goal, methodology, benefits, and limitations.
Line 326-328: Stratospheric column of NO2? Please make this clear. Also, is this a monthly
climatology with interpolated hourly resolution? Such basic information should be presented.
• Further detailes on the stratospheric no2 column are now included, to fully explain the
stratospheric NO2 column and how it was removed from the VCD.
Line 321-325: How did you estimate the error for SCDref? Is this just a fitting error? If not,
please provide some information. HCHO Langley fit looks more scattered, but only has a ~10%
error, while NO2 UV has ~30% error?
• Reference error now more explicitly defined as the fitting error for Langley extrapolation.
Also, the errors in Section 2.2 have been updated to match the correct errors given in
Section 2.3, where both HCHO and NO2 UV errors are ~30%
Figure A1 to A3: The captions of the figures do not match the plots. I can not see any red dot or
red line. I could not see any binning. From the captions, it sounds like you are using the
modified Langley method. Also, as stated, the Langley method relies on the assumption that the
total column is stable for the entire fitting period (half-day). The figures A1 to A3 look wrong. It
is important to verify this for both NO2 and HCHO. i.e., random noise in the total column is not
the worst, but linear changes in a half-day period will induce a large error in your SCDref
derived. Such work can be done using Pandora direct-sun observations. Also, the HCHO fitting
(A3) looks much noisier than NO2. Any comments?
• Corrected captions of A1-3 to match each other, removed mentions of red colors and
bins. The data was chosen to minimize the linear changes in dSCD during the day, and
the true reference day was not used for this reason.
Line 345-346: What do you mean by real data? This is not clear to me. Also, except for HCHO,
the numbers here don’t seem to match the errors for SCDref stated in Line 323. Please double
check those numbers.
• Corrected the errors to match Section 2.21 with 2.3
Table 4 and Line 351-353: This table is good and informative. Maybe state this is on a 1-sigma
level? However, the large difference between NO2 vis and uv error estimates is from the median
VCD of NO2 retrieved in vis and uv. I.e., as the median UV NO2 VCD is much lower, it translates
to a higher % error. However, we all know that there is only one real NO2 column value, no
matter which wavelength you select in measurement. Can you plot the histogram of both vis
and uv NO2 columns (entire records, and/or only coincident records)? Why do we have a >50%
offsets between them (on their median). Do they still have a similar distribution? Theoretically,
UV and VIS results are basically sampling different airmasses. However, a 50% offsets still large
and should be explained. Can you do some estimation on the horizontal sampling areas for
these two settings? Anyway, if the author believes NO2 UV truly have doubled the error than
NO2 VIS, please provide some discussions and suggestions.
• The difference is due to a lower signal to noise ratio in the UV NO2 dSCDs, leading to an
increase in error propagating through all points of analysis, especially the reference
estimation. This is supported by the HCHO errors, which are similar to the NO2 UV
errors.
Line 515-522: More quantitative analysis is needed to support the claim that CSR can be used
for meaningful information on trace gas vertical distributions. As discussed here, the method
will be limited by the sensitivity of in situ and mobile-DOAS. In which range, do you think CSR
provide useful knowledge on mixing height? CSR vs. MAX-DOAS profiling results might provide
more information. Also, CSR will be strongly affected by factors such as BLH, which is measured
by your ceilometer. Have your test their relationship? I think if you can “remove” the impact of
daily BLH changes, the modified CSR might truly give you more information on vertical
distribution.
• The limitations of the CSR, as well as the ability of this study to provide more
quantitative analysis of the CSR and boundary layer height is discussed further, with
recommendations on potential future studies involving the CSR
Technical Comments:
Line 208: The first time the discussion of a figure is Figure 5, which is very strange. If Figure 5
has to be mentioned here, it should be renamed and relocated as Figure 1.
• Moved Figure 5 up to where it is first discussed, now it is Figure 1.
Line 211: backscatter “of “? Rewrite the sentence to make it clearer.
• More clearly explained how aerosol profiles are calculated based on ceilometer
backscatter profiles.Citation: https://doi.org/10.5194/egusphere-2025-5179-AC2
-
AC2: 'Reply to RC2', Davis Earley, 21 May 2026
Data sets
TRACER_AQ2021 LARC data archive Laura Judd https://www-air.larc.nasa.gov/cgi-bin/ArcView/traceraq.2021
AERONET data archivr Pawan Gupta https://aeronet.gsfc.nasa.gov/
Model code and software
BlickP Alexander Cede https://www.pandonia-global-network.org/home/documents/software/
QDOAS Thomas Danckaert https://uv-vis.aeronomie.be/software/
Libradtran Bernhard Mayer https://www.libradtran.org/doku.php?id=download
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 774 | 411 | 77 | 1,262 | 107 | 84 |
- HTML: 774
- PDF: 411
- XML: 77
- Total: 1,262
- BibTeX: 107
- EndNote: 84
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This manuscript by Earley et al. reports on mobile zenith DOAS measurements of formaldehyde and nitrogen dioxide performed in the Houston area in the summer of 2021. In addition to most existing studies using mobile DOAS measurements, the DOAS data are complemented by in-situ data yielding surface concentrations of both trace gases. By calculating their ratio (CSR), information not only on the horizontal but also on the vertical distribution can be inferred. Since many (mobile) DOAS applications - to some extend - assume horizontal and vertical homogeneity which is clearly not always the case, this study offers an interesting approach. Several relevant aspects, such as a sensitivity study aiming to address uncertainties in the assumed AMFs used for the analysis of the DOAS data, are addressed. Lastly, two selected days are discussed in more detail as case studies. Given the overall content of the manuscript it is within the scope of AMT.
However, despite several improvements compared to the initial submission, it still seems at several instances that the manuscript is not in its final shape yet. Several major and minor including technical/editorial aspects should be considered and the manuscript adjusted before publication in AMT.
General comments:
The consistency of several aspects should be improved throughout the manuscript and the presentation style further harmonised. Firstly, this refers to a consistent naming, e.g., Zenith DOAS vs. zenith DOAS vs. ZSL-DOAS, or the use of plural forms of AMF vs. AMFs as well as AOS vs. AODs. There should also be a consistent use of abbreviations. These should be clearly defined at their first appearance and after that they should be consistently used. Lastly, also the way information is provided might be addressed. Examples here are the representations of several words (SZAs vs. SZA’s, SSAs vs. SSA’s, cross sections vs. cross-sections, boxAMF vs. box AMF, etc.). See also the specific comments below for other examples.
Although there will be a proper language editing after acceptance of the manuscript, there are several obvious grammatical as well as spelling mistakes, e.g., the use of articles as well as plural vs. singular forms. Further, several punctuation marks or spaces (especially before/after references and brackets) are missing. These issues disrupt the flow of reading in many places and make it difficult for the reader to follow the content. See also detailed comments below for further examples. Note that such mistakes were only marked in the first half of the manuscript to give an overall impression. Corrections should be done throughout the manuscript.
The figures and the presentation style might be further improved. This includes, besides others, the use of font sizes, the colours involved and the formatting of chemical structures, such as O2O2 and NO2 in legends, titles and axislabels. Often different panels are specified as (a), (b), … in the captions but none of them is then used in the actual figures. Captions of figures showing the same but for different trace gases should be harmonised. Further details are given in the specific comments.
Main/major comments:
Structure of the paper:
In several places, it seems that paragraphs or even entire sections are misplaced or a shift to other parts of the manuscript would be beneficial for the presentation quality and to make the manuscript clearer. In the current shape, several pieces of information are diffusely spread throughout the manuscript. A non-exhaustive list of suggestions is given below:
Current Sect. 1.1 could be shifted to Sect. 2 as it mainly introduces the TRACER-AQ campaign being the basis of the measurements described in Sect. 2. Here, the pieces of information could be merged with the ones from page 5 (lines 114 - 131). For details, refer to the specific comments below.
Further, a short paragraph on how the manuscript is organized could be added at the end of Sect. 1 right after the overview on previous studies on the relationship between VCDs and surface concentrations (lines 61 – 66).
The paragraphs on pages 3 and 4 (lines 87 – 101) describe the concept of CSR. However, it would be more logic if this was an own paragraph/subsection in Sect. 2 which deals with the measurements and methods used in the manuscript.
In the current version, the sensitivity study on the AMFs, including all its subsections (2.2, 2.2.1, 2.3, 2.4, 2.5), is embedded into Sect. 2 which describes the measurements and their analysis. Since this study is a major part of this work, I suggest to give it an own section number with the content of 2.2. being the introduction to this new section and then all other subsections can be just one level below (i.e., X.Y).
Information on AERONET data could be merged into one paragraph. In the current version, this information is spread over several parts of the manuscript and is hard to follow for the reader.
A more general recommendation: The manuscript should to some extent be cleaned up and information should be merged at one location per topic.
More detailed suggestions are given below in the specific comments.
Sensitivity study and result part:
The manuscript includes a good sensitivity study on uncertainties introduced by the use of fixed AMFs instead of AMFS simulated for each individual measurement. However, it is recommended to stress the context and the aim of the sensitivity in more detail. Further, it remains unclear throughout the manuscript which AMFs are actually used to convert the measured SCDs into VCDs? In my understanding, mean aerosol properties are used to simulate AMFs to be used in the analysis (page 5, lines 127 – 131). However, I strongly recommend to clearly state/explain this in more detail. Probably, in Sect. 2.1 after Eq. 2.
A procedure to filter for cloudy scenes is described in Sect. 3.1. However, it remains unclear where this filtering is actually applied. Are all data discussed in the result part (Sect. 3) filtered? As the aim of the filtering is to remove cloudy scenes, it would be good to state whether it successfully removes cloud contaminated data when O2O2 SCDs are high. These aspects should be considered in a clear and concise way.
While the VCDs are nicely discussed in Sect. 3.1, the manuscript lacks a short discussion of the surface concentrations. This is especially the case for HCHO and should be added.
In Sect. 3.2, it should become clearer where CSR is discussed in general and where the two case days are discussed. One could also think about adding subsections for both case days. Further, a clearer motivation why these two days were selected could be added. In my understanding this is related to the ozone levels. I suggest to stress this more. Further, O3 episode might be replaced by something like “high O3 episodes”. Also adding some plot showing the O3 levels in some way would be beneficial to stress the special character of these days. Lastly, a lot of times and time periods are mentioned in this section. To guide the reader, I suggest to mark them by lines in the respective plots? Some more details are mention below in the specific comments.
At several instances, more in-depth discussions are missing. For instance, the topic how the observations relate to ozone chemistry is missing some depth. Further, the relationship between CRS and boundary layer height was highlighted in the motivation part but lacks a more detailed discussion in the result part.
Lastly, Sect. 3.3 reads like a motivation rather than a discussion. Parts of the information provided here might be merged with the introduction part of why CSR is important. Apart from that, most might be skipped. Probably a more in-depth discussion of the results and findings of this work could be performed here (compare previous comment) in a similar way as the last two paragraphs of the section.
Minor/specific and technical comments:
Page 1, line 4: I guess it should read “… vertical distribution of trace gases, …”.
Page 1, line 9: Better something like “…., and for HCHO from…”.
Page 1, line 10: Consider to write something like “… the origin of episodes of enhanced ozone…”.
Page 1, lines 17 + 22: Missing spaces.
Page 2, lines 24 + 25: Like for TEMPO, you might add the long version of TROPOMI and the primary reference of the instrument, namely Veefkind et al., 2012.
Page 2, line 28: Better “… are sensitive to the total column abundance …”?
Page 2, line 31: I suggest to add “… usually smaller than 75°.” since in theory retrievals are also possible at larger SZAs and this is a more practical aspect.
Page 2, line 34: Consider skipping the word “two” in this sentence as also other nitrogen oxides exist.
Page 2, lines 38 – 41: The paragraph first is about NOx and later only NO2 is mentioned. Consider to mention that (passive) DOAS measurements cannot detect NO and that the study here is therefore limited to NO2. Then the considerations about NO2 can follow. Since the lifetime can vary depending on the atmospheric conditions, a range of how far emissions are concentrated around their sources might be added.
Page 2, lines 43 – 49: This paragraph about HCHO reads rather diffuse. I suggest to restructure it in a more concise way. In particular, the fact that HCHO is an intermediate product of VOC oxidation is mentioned twice.
Page 2, line 46: Better “VOCs” rather than “VOC’s”?
Page 3, line 69: Year of publication of “Judd et al.” is missing.
Page 3, line 80: Probably skip the word “retrieved” in this sentence.
Page 3, line 81: Check formatting of O3.
Page 3, line 82: Missing space and “was operated” should be skipped.
Table 1: The abbreviation VCD was so far not introduce. I suggest to do so in the introduction at the first occurrence of “vertical column density”. Further, pressure should be abbreviated by p rather than P. Also, AERONET and its relevant products (SSA, ASY, AOD) should be mentioned in the text, since the abbreviations might not be familiar to everyone and were not yet introduced.
Page 4, lines 92 + 93: I guess it should read something like “… to the total columns of trace gases, and in situ measurements sample only surface air.”
Page 4, lines 96 + 97: A word seem to be missing.
Page 4, line 99: One “layer” too much.
Page 4, line 105: Should it read “… trace gas profiles.”?
Page 4, lines 103 – 112: This paragraph could be the last part of section 1 as mentioned in the main comments.
Page 4, line 106: Full stop (.) is missing. This sentence might be the one regarding the AERONET properties mentioned above and could go there.
Page 4, line 110: Should it be “… the AMF error…”?
Page 5, line 123: Please check formatting of 1.5°. In my understanding, this refers to the FOV of the sky mode. If so, please add this information to the text.
Page 5, lines 114 – 131: Many of this information describes aspects summarised in Table 1 and is related to the description of the TRACER-AQ campaign. Also, some pieces of information are doubled. Please consider merging these pieces into one subsection summarising the campaign. My suggestion would be to do this in Sect. 2 as indicated above (main comments).
Page 5, line 138: Better write “accounted for” instead of “estimated and removed”.
Pages 5 + 6, lines 139 – 146: First, I suggest to write something like “Equation 1 is the so-called DOAS equation …”. Next, in the description of Eq. 1 it should be a Δτ being the differential optical depth, I guess. Further, I wonder if σi is really a differential cross-section or a full one? In my understanding differential here refers to the usage of a reference spectrum. Consider to revisit the DOAS description a little bit taking these suggestions into account.
Page 6, line 145: Consider adding “(ΔSCD)”, to stress that this is the actual output of a DOAS fit.
Page 6, line 150: Please be consistent in writing cross-section vs. cross section throughout the manuscript.
Page 6, line 151: “…measured…” instead of “… measures …”.
Page 6, line 154: “Table 2 lists the corresponding…” as well as “… corrections were performed …”.
Page 6, line 157: Please add a reference for the CINDI-2 campaign, e.g., Kreher et al., 2020. Regarding the I0-correction: not I0 equals some value but a typical SCD is assumed for performing this correction as for instance described by Aliwell et al., 2002. Further, the unit for these SCDs is missing here (molec cm-2).
Table 2: In the caption you might use the abbreviations for UV and VIS as done above. Please cross-check whether the temperatures for the BrO and HCHO cross-sections are correct or if they should be 223 K and 298 K (or the cross-section from 1993), respectively.
Page 6, line 158: Probably better write “The retrievals of …”.
Page 6, lines 160 – 162: It would be more precise to write something like “The SCD is a trace gas concentration integrated along the average photon path and … wavelength accounting for…”.
Page 6, line 163: Are you referring to the calculation of VCDs here? If so, I suggest to write “… derive the tropospheric VCD from the zenith measurements …”.
Equation 2: I strongly recommend to harmonise the quantities and their representations in equations (compare Eq. 1 and 2) as well as the text. For example, so far ΔSCD was used and now this is referred to as ΔS. Also, their dependencies (T, λ, etc.) should be consistently indicated in both equations. Lastly, it is not necessary to provide the unit here in brackets.
Page 7, line 168: “… at a fixed location, …”
Page 7, line 170: “… 8 km from the coast.”
Page 7, lines 170 – 174: Was the cloud screening described below already applied for the creation of the universal reference spectrum? This information might be added to the description.
Page 7, lines 177 – 180: I find these two sentences contradicting. Therefore, I suggest to write in the first sentence “…., most previous Mobile….” and in the second sentence “Other mobile DOAS studies, however, …”.
Page 7, 1st bullet point: As suggested above (Table 1), please introduce these abbreviations already before. Also, the listing of the AERONET sites might be included in the section about TRACER-AQ when the instrumentation is introduced. Here, it could then be stated that “..aerosol properties were from these AERONET stations as described above.”.
Page 7, 3rd bullet point: Better “representing” rather than “estimating”?
Page 8, line 195: “… using the LibRadtran/MYSTIC …”
Page 8, line 196: Better “… in terms of the so-called box AMF (bAMF).”?
Page 8, line 197: “…, the trace gas bAMF profile …“ instead of “… (as function of height) ”
Page 8, lines 195 – 199: Probably, add a reference for bAMFs.
Page 8, line 199: Briefly mention the limitations of the HG phase function. Further, I would argue that also the albedo of the surrounding matters. Could you comment on that?
Page 8, Sect. 2.2.1.: As indicated in the main comments, this subsection could go one level up in a potential new section on the sensitivity study.
Page 8, line 207: “… the aerosol type in the RT simulations.”
Page 8, lines 208 + 209: I suggest to mention to which wavelengths the properties were interpolated to (the ones mentioned above). Also mention how this was done, since I guess also here an Angstrom parameter was used?
Page 8, line 211: Please remove the extra “The”
Page 8, line 211: Better “For that, the ceilometer signal …”?
Equation 3: Please mention in the text that N is the normalised profile.
Equation 5: Please cross-check this equation. If I am not wrong, it should be -α in the exponent.
Page 9, lines 226 - 228: I cannot fully understand how these values are obtained. Shouldn’t it be 1120 profiles? Please cross-check. Probably also comment how these three scenarios are selected/defined.
Figure 1: Please cross-check the indicated colour: 460 nm should be black in my understanding.
Page 9, line 231: Please chose a consistent way of writing LibRadtran (compare page 8, line 195).
Page 9, line 232: “… and RH data …”. Please abbreviate pressure as p instead of P.
Page 9, line 233: Above (Sect. 2.2) eight profiles were mentioned.
Table 3, caption: “… in the the sensitivity study.”. It seems that the table is too wide too fit on the page.
Figure 2: A space is missing before the bracket of the reference. Consider changing the x-label to Trace Gas Concentration. Brown line seems to be no box, is this a simple sampling artefact?
Page 9, lines 236 + 237: Could you comment in how far this assumption is justified?
Page 10, lines 241+ 242: Please provide the values for both concentrations.
Section 2.3: It becomes not clear which AOD scenario was used for the investigations in this section. This information should probably be added.
Figure 3: Consider using single scattering albedo or SSA as y-label in the left panel.
Page 11, line 245: “sites” instead of “sited”
Page 11, line 253: Better use “change” or so rather than “shift”?
Page 11, lines 253 – 255: Please consider re-phrasing this long sentence to improve clarity. For instance, what does it mean that SSA deviates from the minimum AMF? Please consistently uses VIS or visible throughout the manuscript (compare main comments).
Page 11, line 259: Could you briefly comment how these values are obtained? Looking on the grey line, I have the impression that the change of the AMFs from min. to max. SSA is almost 10 %. Or is this just a misunderstanding?
Page 12, lines 262 – 264: Please check this sentence.
Figure 4: Consider using single scattering albedo or SSA as respective x-axis labels. If possible, adjust formatting of O2O2.
Figure 5: I suggest to write “hourly mean L2 AOD values …” instead of “averaged every hour”. I wonder why this figure is not Fig. 1, as it only referred to on page 8 (line 204)?
Page 12, lines 269 – 271: Where is the information on the SZA in Fig. 6? Please add this information in the text.
Page 13, line 283: It might help to mention how the difference is calculated to get a more intuitive idea of Fig 7. because difference between O2O2 and trace gas (line 281) sounds contra intuitive to Fig. 7 (negative values). I assume its AMFtrace gas – AMFO2O2 shown in Fig. 7?
Page 13, line 285: “… high solar zenith angles (morning and evening) …”. Please skip the word “other”.
Figure 7: No panels are marked but mentioned in the caption. Is there a blue line (profile 1) in fig. 7? If so, please mention in the text that is covered by another (orange?) line.
Figure 8: Consider using “Trace gas AMF” as y-label. Differences are mentioned in the text but absolute AMFs are shown this should be harmonised. No panels marked as mentioned in the caption. Please check which column is which wavelength since titles and caption are contradicting each other.
Page 14, lines 290 + 291: Better “trace gas profiles” instead of “other profiles”. In general, clearly use “trace gas profiles” if they are referred to in this paragraph.
Page 14, lines 293 – 295: This sentence reads rather complicated. Probably just state that the differences between the trace gas profiles are smaller than the differences to the O2O2 profile which can also be seen in Fig. 2. since the profiles deviate less. Better “changes in the vertical profile” rather than “changing of the vertical profile”, since you are referring to the final measurements?
Page 14, line 300 onwards: What is meant with constant profile?
Page 15, line 303: Please consider using ZSL or Zenith/zenith DOAS (one of both) throughout the manuscript.
Page 15 lines 307 – 309: This is unclear. Didn’t you state above that the choice of the trace gas profile is not crucial (page 14, lines 294 +295)? Probably comment on that.
Page 15, line 310 onwards: Which AMFs are actually used for the analyses of the measurement data in this work? This question arises several times when reading the manuscript (compare also main comments).
Consider adding a short paragraph where the main findings and the uncertainties related to the AMFs are briefly summarised. This would also help to follow the discussions in Sect. 2.6.
Section 2.5.1: I wonder if this is the best place for this section. I have no clear idea where to put it but it is not directly related to the sensitivity study. When restructuring the sections this section might go to the measurement and analysis description section (Sect. 2). Add a reference for the Langley plot method.
Page 17, line 321: What does 2.2-4.2 SCDref refer to? Is a full stop missing?
Page 17, line 322: “… for the UV retrieval …”. How where the errors estimated?
Page 17, lines 326 – 328: What is the value of the stratospheric VCD? Only the uncertainty is mentioned.
Page 18, line 341 onwards: In the current version, Sect. 3 is the results section. Please use the correct numbering (after restructuring). How is this AMF uncertainty exactly calculated? Are different values for different spectral ranges from above considered? Estimation of the individual terms remains vague. Moreover, the values for the error of the reference SCD differs from the ones listed in Sect. 2.5.1.. I suggest to provide more information on how these values were obtained.
Page 18, line 351 – 353: Check if words are missing this sentence as it is difficult to read.
Table 4: Also values for NO2 UV are shown please modify caption.
Page 18, line 355: Consider removing “… of the month.”.
Page 18, line 362 onwards: I suggest to introduce this abbreviation and its long form already above in the measurement section. “… surface concentration …”.
Page 19, line 367: Consider to use the abbreviation VCD here. “… are also plotted in black.”
Figures 9 + 10: Probably use VCD as y-label? The legend is very small, probably enlarge it. UV NO2 VCDs seem to be consistently smaller. Can you comment on that since for VCDs radiative effects are already accounted for? Again, which AMFs are used (compare main comment above)? Do they not account for that?
Page 19; line 386: The abbreviation dAMF can be introduced above in line 378 where it is first used. Where is this filter actually applied? This is not mentioned (compare main comment above).
Page 19, lines 388 – 391: Consider rearranging this sentence to improve clarity. Fig. 10 is not referred to, which should certainly be done.
Page 20, line 399: I would argue on 25 September the correlation is not good for HCHO, especially the large peak at 9 a.m. is only seen by the in-situ instrument.
Page 20, line 401: “… in the trace gas abundance …”
Pahe 20, line 403 onwards: This part is not completely clear. I would rather write “… depends on the specific vertical distribution of the trace gas …” or so. Also, the “while sentence” is not clear. I recommend to say that the combination of both allows to infer information on the vertical distribution and to mention that this is actually the CSR which is then described in the next sentences.
Page 21, line 417: “… averaged measurements”, Probably better use “tens of metres” since “10 s” looks like seconds. Also values above 2 km occur frequently. Better write “… up to a few kilometres …”?
Page 21, line 420: Probably again mention on which days the instrument was the stationary. “… by consistent values …”, and I suggest to add example days for this.
Page 21, line 428 onwards: This is confusing. The paragraph before deals with HCHO and then there is a jump to NO2 again. I strongly recommend to rearrange these paragraphs to make the content more fluent and logical. Further, I wonder if Fig. 11 could already be mentioned here, since the described aspects are clearly visible there rather than in Fig. 8. In fact, Fig .11 is not referred to at all in the current version.
Figure 11: Again, no panels marked in the figure but mentioned in the caption.
Page 22, line 436: At 14:30 it seems that a local minimum in CSR is visible. I think it’s more 14:15.
Page 23, line 443: In Fig. A7, a clear increase of HCHO is visible. I guess it’s the earlier peak you are referring to. I suggest to mention Fig. 10 or even A7 here.
Page 23, lines 449: Check the mentioned times as they are not consistent to the caption of Fig. A8. “… the air masses originated largely from …”
Page 23, line 454: “… NO2 and HCHO …”
Page 24, line 460: This information is already given before.
Page 25, line 475: I think it should be VCDs since they show the spikes in Fig. 13.
Page 25, line 477: “… in both HCHO and NO2 VCDs, the NO2 concentration …”
Figure 13: Caption of this figure and Fig. 11 should be the same with the difference only in trace gas shown. Consider using the abbreviation VCDs.
Page 25, line 479 onwards: This paragraph is unclear. Please elaborate more what it means that the transported trace gases interact with the HCHO present? If not, this part of the sentence might be skipped.
Page 25, line 480: Regarding Fig. A11 please mention first that the plot shows data from a different day to guide the reader.
Page 26, lines 484 onwards: This paragraph has no real link to before as no wind arrows are shown in Fig. A11 the way it is the case in Fig. 12. I suggest to put this information to a different part of the manuscript where it is more relevant. This could be either the part where the measurements are described or where the wind arrows are shown.
Page 26, line 491: Consider skipping “more than”.
Page 26, line 496: “… trace gas distribution”
Page 26, line 497: It is unclear what exactly can be validated. Please comment on that.
Page 26, line 500: Isn’t that what you do? I am little bit confused at this stage.
Page 26, line 517: Please skip one “both”.
Section 4: I really like the way the conclusions are written. More of this very precise and clear sentences should be written during the whole result and sensitivity study parts (compare comments given so far).
Figures A1, A2 and A3: In the caption better use “…retrieved in the VIS/UV” and “AMFs at”. Is 22 September really the reference day? Before it was mentioned that 21 September is the day on which the reference spectrum was determined. I suggest to skip “reference day” here to avoid confusion. There are no red dots and lines in the figure. Please check that. Further, the second sentence in A1 and A2 is unclear. What exactly are these bins, probably elaborate this in the main text.
Figure A4: There are O2O2 data for days were no NO2 is available, why?
Figure A9: Better “… on that day.” rather than “… during this time frame.”?
Figure A10: “… were taken.”
Figure A11: Is the sentence about reference spectrum really relevant here?