the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 29 Oct 2025
Two years of total column measurements of CO2, CH4 and CO in Paris, France
Abstract. Several cities have established atmospheric monitoring networks to track greenhouse gas emissions. In the Paris metropolitan area, continuous in-situ surface measurements have been conducted since 2015. To complement them, three ground-based solar Fourier Transform Infrared (FTIR) spectrometers provide total column concentrations of CO2, CH4, CO, and H2O (XCO2, XCH4, XCO, XH2O): Jussieu (JUS, since 2014), Saclay (SAC, since 2021), and Gonesse (GNS, since 2022) within the ICOS Cities project.
These total column estimates capture background variability and urban plumes but are influenced by calibration, measurement noise, solar geometry, a priori profiles, and surface pressure. Accounting for these factors, overall uncertainties are estimated at 0.2 ppm for XCO2, 1.2 ppb for XCH4, and 2 ppb for XCO (EM27/SUN instruments). Inter-station gradients reveal the influence of the Paris emission plume, with typical XCO2 gradients below 1 ppm (mean 0.51 ppm, 8.9 % above 1 ppm).
Observed XCO2 gradients were compared with WRF-Chem simulations driven by the dynamic Origins.earth emission inventory. Correlations are moderate – 0.47 (SAC-GNS), 0.43 (JUS-GNS), and 0.26 (SAC-JUS) – with regression slopes of 0.66, 0.72, and 0.44, respectively, suggesting a potential overestimation of emissions by about 35 %. However, the large spread between measured and modelled gradients limits the robustness of this conclusion.
The paper first describes the monitoring network and harmonized data processing, then evaluates measurement uncertainties, and finally compares observations with model simulations to assess the potential of FTIR column data for validating urban emission inventories.
- Preprint
(2154 KB) - Metadata XML
-
Supplement
(822 KB) - BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2025-4876', Anonymous Referee #1, 21 Jan 2026
-
RC2: 'Comment on egusphere-2025-4876', Anonymous Referee #2, 19 Feb 2026
General Comments:
This manuscript discusses the ground-based solar FTIR spectrometers (EM27/SUN) measurements in three stations in Paris - Jussieu, Saclay, and Gonesse. Column concentrations of CO, CO2, and H2O are measured. The measurements are compared against model simulations from WRF-Chem, suggesting that the current emission inventories might be overestimated by approximately 35%. The manuscript discusses the limitations of the role of column data in emission estimation.
This manuscript is well-structured, well-written and supported by figures adequately. This manuscript is relevant for publication in ACP after some clarifications and corrections. My main concern is the use of older PROFFAST/PROFFASTpyplot subversions in the manuscript. The authors must consider updating the analysis to the latest PROFFAST/PROFFASTpyplot subversions, if possible. Alternatively, they may choose to provide an estimate of how much the analysis would vary upon using the latest PROFFAST/PROFFASTpyplot subversions along with the changes that have gone into the latest PROFFAST/PROFFASTpyplot subversion from the one that they have used in their analysis.
Please note that a few instances in the manuscript (pages 11,12, and 22) have text in red ink. I am not sure why this is. I assume that a “track changes” version, with very minor changes (from, possibly, an internal review), was saved and uploaded. If this is the case, I request the co-authors to please check for such instances throughout the manuscript and submit the clean and unmarked version in the future.
Specific Comments:
L11: “… continuous in-situ surface measurements have been conducted since 2015” – Please state which ones.
L17: “(EM27/SUN instruments)” – Please define or state while mentioning the FTIR instruments earlier.
L19: “Observed” to “The observed”
L20: “Correlations” to “The correlations”
L31: denoted by “X”
L33: “… and industry” – Please cite
L35: “.. greenhouse gas” – Please cite.
L47: Please define ICOS before the first use.
L57: “… mitigate this issue” – How?
L 58: “…observed by satellites” – Please cite.
L68: “NE–SW” – Please define before first use.
L79: “…in spring” à “in the spring”
L87: Please check if the EM27s operate continuously or continually. The EM27s other than SN061 make sporadic measurements at Munich.
L92: “…poor agreement” – Please state the number.
L96: Please check the EM27s at Mexico City. There are six, not seven of them (see the KIT COCCON data page). SN038 is used to measure at Altzomoni and Boxo.
L97: “…high-resolution site in continuous operation” – Please state which one.
L98: “St. Petersburg and Tokyo” instead of “St. Petersburg or Tokyo”
L105: Please define “WRF” before its first use while mentioning WRF-GHG. Please define GHG as well.
L110: “Origins.earth emission” instead of “origins.earth emission”
L115: “This provides a time series of more than 10 years…” – It isn’t clear what “This” refers to here.
Table 1: Please define LSCE before first use.
L133: Please provide references for CamTracker and OPUS – either papers that discuss them, or a link to a user guide so the reader can learn more about them.
L134: MONARIS – Please expand if this is an acronym. If not, please state so.
L136: Please explain what the travel/travelling standard EM27 refers to. It would also be nice to have an image of the station/deployment set up.
L140 – PROFFAST 2.4.1 is the latest version and PROFFASTpylot is at v1.3. Why are older PROFFAST/PROFFASTpylot versions being used in this analysis? I suggest the authors update the PROFFAST/PROFFASTpylot versions for this analysis, if possible. (Or provide an estimate of how much the results would vary upon using the latest PROFFAST and PROFFASTpyplot versions.)
Please also state that PROFFAST is not an acronym but the actual name of the retrieval tool. It would also be helpful if the authors could briefly state the updates between the various PROFFAST 2 subversions.
L147: Please define CAMS first and use the acronym later.
L151: Please define PAUL.
L154: Please define and explain NUBICOS.
L156: “…while also ensuring long-term traceability” – Please provide references.
L166: “…where 𝑁𝑀𝑒𝑠𝐷𝑎𝑦𝑠 is the number of days with valid data” – Or with valid “Measurement” data?
L186: “The uncertainty is given by 𝜎/(N)0.5”- Please consider writing as an equation.
L188: Please change “after that date” to “following which”
L191: ” not “the” absolute truth
L192: lower “than” about
L196: What are the uncertainty values for XCH4 and XCO.
L199: Please start a new sentence after “average”.
Figure 3: Please explain the green strip in the figure caption.
L209-210: “For the Jussieu IFS 125 HR, days with more than 200 spectra are fewer because the site alternates usually between 210 NDACC and TCCON measurements.” – This sentence is hard to follow. Please rephrase.
L210: Please state how the 0.55 ppm estimation was made?
L213: “due to” a known issue
L220: “low SZA” – Please start a new sentence after this instead of using “;”.
L227: “noon SZA< 50o” – Why was this value used.
L227: “…assume no large geophysical trend between midday and early/late hours” – How much uncertainty would this add to the analysis?
L237: “We detect no systematic drift up to ~60°. Beyond ~65-70°, a drift emerges that grows with SZA and differs by gas: for XCO2, it becomes noticeable above ~75° (> 1 ppm) and reaches up to 3.5 ppm for 80-85°; for XCH4, it remains smaller (< 2 ppb up to 80°, up to 5 ppb between 80-85°); for XCO, it is negligible.” – It might be nice to put this in a table.
L258: “𝑋 𝑐𝑜𝑚𝑝 𝐺𝐴𝑆, the_GAS” – Please remove the underscore between “the” and “GAS”.
Figure 5: Please update the axis language to English.
L268: Please remove “however”
L271: “a_priori” – Please remove the underscore and italicize the words here and for all instances of occurrence.
L272: “(slightly)” – Please put the word in parentheses.
L278: “…alternate station 2 km away” – Please state if this introduced any uncertainty in the analysis and if so, how much.
L280: “XAIR” – Please use a subscript for “AIR” – here and everywhere else in the paper.
L292: Please define ILS.
L293: “An empirical tolerance of ±4.10-3 around 1 is typically accepted” – Please state why. Please use a multiplication sign (“×”) instead of a period (“.”) here: ±4.10-3.
L312: “Assuming independence” – How much uncertainty does this add to the analysis?
L329: “CCGvu algorithm” – Please expand if this is an acronym. If not, please state so.
L343: “It is important to note that seasonal and long-term signals are very similar at the regional scale” – Please cite relevant publications.
L347 – “that are clearly geophysical”- Please explain how.
Figure 7. – Please consider writing “Total” instead of “TOT”. It looks like an acronym.
L358 – “Saclay shows no enhancement” – Do we know why the WRF-Chem model didn’t see the enhancement, even though modest?
L266 – “Errors in transport fields can dominate apparent source mismatches. Therefore, caution is required when interpreting model–measurement differences solely as inventory biases” – Are there similar studies that draw this inference? If so, please cite them.
L402: “depending on the month.” – Please clarify which values are associated with which months.
L409: Please change “will use” to “we used”. Please remove “to finally”. Please change “compare” to “and compared”.
L444: Please change “(temperature, solar irradiation, precipitation, ...)” to “(temperature, solar irradiation, precipitation, etc.)”
L488: “Noth-East to South-West” – Please be consistent in the use of NE-SW and the full forms. Since NE-SW has been used earlier, please define it at the first instance and use the acronym form throughout the paper.
L492: “The background concentration should be very similar in the whole domain at any given moment, so we could expect a narrower distribution of background gradients between SAC and GNS.” Please change this to “The background concentration should be very similar in the whole domain at any given moment. Thus, a narrower distribution of background gradients between SAC and GNS can be expected.”
L494: “However, occasional episodes with larger differences between the background at the two stations do occur and must be taken into account when interpreting the total gradient in relation to urban emissions.” – How much will this impact the uncertainty?
L509: Please capitalize the “O” in origins.earth.
L520: “XCO2;SAC < XCO2;GNS” – Here and for all similar occurrences in the manuscript, please consider using underscore instead of semi-colon.
L554: “The best fit slope is also lower than 1 555 (0.72 for JUS-GNS and 0.44 for SAC-JUS). The best linear fit indicates a statistical error that are relatively small (0.005 for SAC-GNS, 0.015 for JUS-GNS, and 0.017 for SAC-JUS), so that a slope significantly smaller than 1 is a result that is statistically robust.” – I wonder if it would be better to put these numbers in a table.
L615: “That study, published in Lian et al., 2023 [14], concluded that emissions were underestimated by 2 to 20 %, depending on the month.” – Please state the approximate values typically associated with the months.
Citation: https://doi.org/10.5194/egusphere-2025-4876-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 346 | 145 | 33 | 524 | 65 | 25 | 26 |
- HTML: 346
- PDF: 145
- XML: 33
- Total: 524
- Supplement: 65
- BibTeX: 25
- EndNote: 26
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This manuscript describes a network of ground-based Fourier transform spectrometers for measuring greenhouse gas columns at Paris and evaluates the measurement uncertainties. The measurements are compared with model simulations based on Weather Research and Forecasting model coupled with Chemistry (WRF-Chem) to assess the potential of column measurements in verifying urban emission inventories. Correlation analysis based on observed and simulated spatial CO2 gradients suggests that the emission inventory used in the simulation (i.e., the Origins.earth emission inventory) is overestimated, while also revealing the difficulty of using column data for emission estimates.
The topic of this paper is relevant to the scope of Atmospheric Chemistry and Physics. The analysis method is appropriate and the writing structure is well organized. However, I have a number of specific comments, which need to be addressed before the publication.
-----
Specific comments:
L21: What is the basis for an overestimation in emissions of 35%? This figure is not mentioned in the main text. In addition, the Conclusion section states that emissions are overestimated by 44%.
Table 1: Please include the serial numbers (#88, #103, and #179) of the EM27/SUNs used at each station, along with the corresponding periods.
Figure 2: Regarding XCO2, instrument #88, which was deployed at SAC before 2023-05-19 and at GNS after 2023-05-19, appears to have a low bias, while instrument #103, which was deployed at SAC after 2023-05-19, has a high bias. Thus, there seems to be device-specific biases. Can the effect of the exchange of instruments be considered negligible? Additionally, Paris TCCON data are biased low by approximately 0.5 ppm, compared with other EM27/SUN data. Is it acceptable not to consider this bias?
Figure 4 and L233: Could you please clarify whether the ∆XCO2 values were calculated for each instrument, as indicated in the legend of Figure 4, or for each station, as described in the main text?
L294–295: According to Table 2, a 0.0015 change in XAIR corresponds to a 0.14 ppm change in XCO2. Which is correct?
L331: Is the observed growth rate the same at the three stations?
L465–478: Please add discussion of the differences between the simulations and observations, especially why the measured XCO2 values at Gonessa were, on average, larger than those at Saclay, contrary to the simulations.
L579: What is the emission inventory used by CAMS? This information would be helpful in understanding the differences between the WRF-Chem and CAMS simulations.
L587–588: Because the errors in the slopes for JUS-GNS and SAC-JUS are almost the same between CAMS (9 km resolution) and WRF-Chem (1 km resolution) simulations, the difference in the spatial resolution seems to be irrelevant.
L589–590: Why is underestimation in the emissions inventory considered more reasonable than overestimation?
L615–619: In L402–404, it is stated that emission estimates using in situ surface measurements are subject to significant uncertainty due to the difficulty of modeling the vertical mixing, whereas column measurements are insensitive to this modeling. Could you discuss which approach is more accurate?
-----
Technical corrections:
L64: Are the EM27/SUN data at JUS used in this study? Otherwise, it would be better to either state that fact or delete it to avoid confusion.
Table 1: What is the QUALAIR platform?
L147: “to be used” is unnecessary?
L157: the radiative transfer code GGG2020 package -> the retrieval software GGG2020
L159: TCCON framework -> TCCON archive
L225–226: we define a reference as the mean XCO2 over 10:00-14:00 local time. -> we define the mean XCO2 over 10:00-14:00 local time as a reference.
L240: modelled radiative -> modelled radiative processes
Equations (3.1)–(3.3): What do the double arrows means?
Figure 5: Please change the x-axis label to English.
L266: rSAC = 0.76 et rGNS = 0.75 -> rSAC = –0.76 and rGNS = –0.75
L312: This sentence is almost identical to L314–315, so it is redundant.
Table 3: What is the difference between “-” and “X”?
L324: According to Section 2.2, TCCON spectra were analyzed with GGG2020.
L392: CO2 plus -> CO2 plumes?
L420: Figure 10 is cited before Figure 9 in the main text. Figure 10 should be changed to Figure 9 and vice versa.
L504 and 507: et -> and
L560–564: Please add the slopes of the fits as well as the correlation coefficient.
L584: discrepancies in the simulation of emissions or atmospheric transport -> discrepancies in the atmospheric transport simulations or emissions