# Review of "Atmospheric Carbon Dioxide Measurement from Aircraft and Comparison with OCO-2 and CarbonTracker Model Data" by Wang et al.

The authors describe their approach on how to measure CO_2 mixing ratios and

columns with an airborne LIDAR system. The study is motivated by addressing

the limitations of fixed ground-based measurements and the limited

spatio-temporal resolution of satellite measurements. Airborne measurements

can have similar spatial resolution as ground-based measurements,

and at the same time cover a larger area.

The study will be a valuable contribution to the community,

and motivate others to use airborne LIDAR systems to infer

CO_2 and probably other trace gases with high resolution on a regional scale.

However, the manuscript needs a major revision because of some

omissions that should be addressed in a scientific publication.

## General comments and major points

1. The authors should give error ranges to all their average estimates,

from the abstract to the conclusions, the values given in the text should

have the form xxx.xx ± x.xx ppm, with an appropriate number of significant decimal places.

I'd suggest to round to 1 significant decimal place (xxx.x ± x.x)

or at most 2 (xxx.xx ± x.xx),

as too many can give a false impression of the precision.

2. The authors need to add units to all quantities, in particular the

ones used in the equations.

3. My major issue is the description of their "pulse integration method" with

a reference to an earlier paper using the same instrument. To my understanding,

what is described is not a pulse integration method, which uses the sum or average of

multiple pulses to increase the SNR.

Here the authors use "points on the pulse", which is something different.

Thus, the first paragraph in Sect. 3.2 needs to be re-written, it is not clear what the authors did.

It would be helpful if the authors could provide an independent reference

that is not a self-citation.

Then, depending on the interpretation, Eqs. (8) and (9)

might have to be adjusted;

to my understanding in the case of un-weighted averaging,

the average variance is calculated by dividing by N, not by N^2.

As given, Eq. (8) resembles the standard error of the mean,

not the standard deviation as described in the text.

On the other hand, if the variance at each point is known,

then that is the variance of the signal, which gives the SNR.

Averaging points sounds as if the authors are binning in time.

But that has to be done for all points on the pulse, including the peak.

It does not make sense to average points before and after the peak,

and then to exclude the peak itself.

In addition, if the points before and after the pulse peak are averaged,

this average should be zero, as this amounts to averaging noise

as presented in Figs. 4--6.

If it is not zero, there might be additional terms from the variance

of the alpha values themselves.

In the case the authors use these points to get an estimate of the noise

in their signals to calculate the SNR in the first place, then this should

be described, and Eq. (8) should be replaced by the equation for the usual variance.

As a third option, the noise could be estimated from the residuals

after subtracting the matched filter signal.

Then, it is also not clear what quantities are averaged, the voltage?

How are alpha and sigma estimated?

## Minor points and detailed comments

### Abstract

- l 19: As mentioned in general comment 1, ± error ranges should be given.

- l 24: As mentioned in general comment 1, ± error ranges should be given.

- ll 26--27: Measurements can "indicate" but not "prove" something, which is a very strong statement.

  I'd suggest to remove "trend" and replace "which proved the existence of" by "showing". 

### Introduction

- Please re-introduce the acronyms again in the main text at their first appearance.

- l 42: Remove "the": "the methane" -> "methane"

- l 44: Change "vulnerable to the" to "affected by"

- l 80: Continue with "discussed in Sect. 3, and our conclusions are presented in Sect. 4."

### Section 2

- Please add  a few (1--3) sentences to motivate the measurements and data,

  i.e. why you need CarbonTracker when you have OCO-2 data?

#### Section 2.1

- l 84: "-1" should be a superscript in "h^-1"

- l 86: Please add the spectral resolution of the channels, also to table 1.

- l 87: Introduce "the": "the 532 and 1064 nm channels",

  remove "the": "detect the aerosols" -> "detect aerosols"

- l 90: Remove "our previous article": "described in (Zhu et al., 2020)."

- Table 1 should be on one page. Add a column with the spectral resolution.

#### Section 2.2

- Table 2 should be on one page.

#### Section 2.3

- l 126: typo: "UUGA" -> "UGGA"

- l 131: Insert "the": "the various types of surfaces."

- l 138: Remove "the": "at regional and global scales."

- l 139: Change "making coincident measurements" to "simultaneously measuring"

- l 142: Section 2.3.3 The authors should motivate their use of the

  CarbonTracker data, in particular since they also have OCO-2 data available.

#### Section 2.4

- As mentioned in general comment 2, all the quantities used in the equations

  should be given with proper units, SI or SI-derived.

- l 153: Replace "except" by "than"

- l 159: Use lower case "where" without a comma after it.

- l 162: Refer to Eq. (2) after "caused by CO2": "caused by CO2 (given by Eq. (2) below)"

- l 163: "detection signals" do you mean "monitor signals"?

- l 164: Insert "tau_CO2": "...of the CO2, tau_CO2, can..."

- l 167: Use lower case "where" without a comma after it.

- l 168: The sentence "R_G is the height ... aircraft platform." is the same as

  in line 160 and can be removed.

- l 169: Change "using" to "according to"

- l 172: Use lower case "where" without a comma after it.

- ll 172--173: What is R_v exactly? If I got it right, then it has the unit

  1 / A, (A for Ampere), which does not look like a rate (~ 1 / s),

  and not like a power unit such as W or V·A (Volt·Ampere), as mentioned at the

  end of the sentence. Thus this is also related to my general comment 2 to add

  proper units to all quantities.

- l 173: Change "Therefore," to "Using Eq. (3),"

- l 175: Change to "detection signal voltages",

  or probably "monitor signal voltages", see comment on line 163.

- l 176: Change to "echo signal voltages".

- l 177: Change to "calculated using the following equations:"

- l 179, Eq. (6) upper and lower case "P" are used for pressure, please use

  the same consistently throughout the manuscript (see also the comment below)

- l 180: Use lower case "where" without a comma after it.

- l 180: Upper vs lower case "P" for pressure, in line 168 lower case "p" is

  used, please rectify to be consistent.

### Section 3

#### Section 3.1

- l 194: Remove "Moreover": "No significant difference..."

- l 195: Change to: "over the residential and mountain areas."

#### Section 3.2

- The first paragraph "data processing" could be in the "methods" section.

- l 199: Refer to Eq. (2) after "DAOD".

- ll 199--211: As mentioned in general comment 3, this part needs to be

  rewritten and better explained what was done. As I read it, the authors bin

  the signal of a single pulse, but do so differently for the background noise

  ("before and after the peak") and the actual peak signal, where no averaging

  is indicated. In addition, Eqs. (8) and (9) might need to be adjusted.

  What quantities are alpha and sigma?

  What are N and k exactly, and what is denoted by the superscript l?

- ll 214, 216, 222--223: As mentioned in general comment 1, ± error ranges should be given.

  Please use the same order of ocean, residential, and mountainous areas for

  tau, IWF, and XCO2 consistently.

  If I use Eq. (5) with the values of tau and IWF as given, I get different results

  for XCO2, could the authors clarify?

#### Section 3.3

- l 226: The first sentence should be removed.

- l 229: Replace "because it produced errors and sudden spikes" by "because of sudden spikes"

- l 229: Insert "associated": "the associated sudden pressure changes."

- l 230: Consider replacing "concentration" by "volume mixing ratio",

  concentration can be confused with the more typically used molar concentration

  with units of mol / m3.

  It is also possible to introduce the acronym "vmr" at the same place and use that

  instead of "concentration" later on.

- l 230: Remove "the": "is largest".

- l 232: Remove "the": "in northeast China", and "Moreover, northeast China..."

- Table 3: Consider adding a column for the measured XCO2 values for an easier overview

- l 239: Remove "to be": "was likely caused by"

- l 243: What is shown in Fig. 16 exactly? Consider indicating the boundary layer height

  as mentioned in the text.

- l 246: The reason needs to be strengthened, a similar variation or trend

  alone does not imply a good agreement. There good be a large bias between the two.

- l 247: Consider replacing "concentration" by "volume mixing ratio", see comment on line 230.

- l 247: Remove "the": "is highest"

- ll 248--249: As mentioned in general comment 1, ± error ranges should be given.

- l 250: Remove "relatively"

- ll 251--252: This is a general statement, what exactly does it imply for the UGGA vs ACDL comparison?

  Is it feasible to compare them or not?

#### Section 3.4

- l 257: typo: "were removed."

- ll 257--258: As mentioned in general comment 1, ± error ranges should be given.

#### Section 3.5

- l 266: Replace "more" by "higher".

- ll 266--267: Be careful with the term "upper atmosphere", which usually

  refers to altitudes above 50 km, or sometimes 100 km.

  I suggest to use "atmosphere above" instead.

### Conclusions

- l 274: Remove "relatively"

- ll 275--279: Related to general comment 3, this part should be improved to

  describe the method better. Please consider the following question:

  Is this part of the method essential to the results of the paper?

  In my opinion these details could be removed from the conclusions

  as they are very specific to the instrument used.

- ll 281--282: As mentioned in general comment 1, ± error ranges should be given.

- l 283: Remove "the": "was largest"

- ll 283--289: Consider replacing "concentration" by "volume mixing ratio", see comment on line 230.

- l 286: Is it variation or trend? As commented on line 246, neither a similar

  variation nor a similar trend imply a good agreement. The authors should

  discuss their definition of agreement, in particular with respect to

  a potential bias.

- l 288: Remove "relatively"

- ll 289--290: This is a general statement, is that a conclusion of the paper

  connected to the particular measurements? This sentence should be better

  connected to the method and the rest of the paper.

### Figures

- Fig. 3: Add a label to the x-axis, probably "time".

  "distance measured by the lidar", distance between what?

  If it is the distance between the plane and ground, that should be described.

  Last line: "dividing line for different".

- Figs. 7--9: Include axis labels with units for all panels (a).

  The panels (b) need to be reconsidered after addressing general comment 3.

  What units are used for SNR?

- Figs. 10--12: Replace "in" by "over": "XCO2 results over ocean..."

  Add a label to the x-axis, probably "time".

  By referring to Fig. 10, the last two sentences in the captions of Figs. 11 and 12

  can be removed.

- Fig. 14: Legend: There is no need to list the values with 4 decimal places

  if the error range has only 2. Please round to 2 significant digits on both,

  too many will give a false impression of the precision.

- Fig. 15: Designate the lines as "lines", i.e. "blue line", "black line", and "red line".

- Fig. 16: Add a label to the x-axis, probably "time".

- Fig. 17: Add a label to the x-axis, probably "time".

  Use the same scaling for IPDA XCO2 and UGGA CO2 to make them better comparable

  in absolute terms.

- Fig. 18: Put (a), (b), and (c) on one page, include panel indicators (b) and (c)

  in the second row.

- Fig. 19: Consider replacing "concentration" by "volume mixing ratio", see comment to line 230.

  Include the dates in the panels or as column headers to make it easier to

  find which panel shows the measurements from a certain date.

  "Red and blue scatter" is better referred to as "red and blue errorbars".

General

The paper introduces aircraft based measurements of column weighted CO₂ mixing ratio using a lidar at the Chinese coast including comparison with satellite data. The paper is within the scope of the journal but especially section 3.2 requires revision because of misleading sentences. Here the authors should use the papers of the other groups applying similar methods (Refaat, Amediek in introduction). Sections 3.4 and 3.5 are too short.

Specific

Line 143ff: "two-way nested chemistry-transport model Tracer Model 5’’ (see also Peters et al, 2004). Improve sentence, it is not consistent with the provided references. The reference to CarbonTracker (Babenhauserheide et al, 2015) should be earlier. Unfortunately the references use different full names for TM5 but not 'transfer model'.

Line 151: I suppose the trace gas is CO₂ here, i.e. online means on a CO₂ line. If yes please say so.

Line 160: "hard target": is that the surface or the cloudtop? Please be more precise here.

Line 170ff: Please define all quantities in equations in the text. What is for example P_p? Use lower case for atmospheric pressure (p).

Section 3.1: Please improve text for the non expert reader. 

Section 3.2: I suppose Eq. 7 describes the signal and Eq.8 the noise, if yes please write that (see also reviewer #1). Please correct misleading sentences. What is on the abscissa of panel a of Figs. 7-9 (with units)?

I would suggest to rearrange section 3.3 to have every results for the 14 March flight together, i.e. exchange the paragraph beginning with line 245 with the part from line 226 to line 244.

Sections 3.4 and 3.5: Please say more to Fig. 18, including the shown standard deviations. Please include the flight data in panel 18b in the same color scale, or maybe a slightly shifted scale, to consider that the satellite data must have a systematic low bias because of the influence of the altitude region with lower CO₂ above the flight track (upper troposphere and the stratosphere). This bias should be mentioned in the text, here and also in section 3.5 as justification of the use of CarbonTracker (please spell out in caption of Fig.19). The last sentence of section 3.5 has to be replaced, I don't think you refer to the stratosphere here when in the figure is only the troposphere. It might be useful to indicate the flight altitude in Fig.19.

Line 376: Please replace the preprint by: Krol, M. C., S. Houweling, B. Bregman, M. van den Broek, A. Segers, P. van Velthoven, W. Peters, F. J. Dentener, and P. Bergamaschi (2005), The two-way nested global chemistry-transport zoom model TM5: Algorithm and applications, Atmos. Chem. Phys., 5, 417– 432.

Technical corrections

Additional to the remarks of reviewer #1 there are the following issues:

Line 64: Don't create fantasy names for existing institutes. The correct name is 'German Aerospace Center (DLR)'.

Line 143: Typo in citation.

Table 3: Is 'wind scale' 'wind strength in Beaufort'? 

Line 266: Typo

References:  Please remove control sequences (e.g. line 312) or blanks (e.g. line 449) and use subscripts instead.

Several times the name of the journal and the volume are missing, indicated by ',,', please insert it. In case of Yokota also the DOI is missing, meaning that it is impossible to find the paper. For books please provide publisher and city.

Use µ instead of mu, and CO₂.

References

W. Peters, M. C. Krol, E. J. Dlugokencky, F. J. Dentener, P. Bergamaschi, G. Dutton, P. v. Velthoven, J. B. Miller, L. Bruhwiler, and P. P. Tan (2004), Toward regional-scale modeling using the two-way nested global model TM5: Characterization of transport using SF₆, J. Geophys. Res., 109, D19314, doi:10.1029/2004JD005020.