the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
The Space CARBon Observatory (SCARBO) concept: Assessment of XCO2 and XCH4 retrieval performance
- 1Laboratoire de Météorologie Dynamique/IPSL, CNRS, École polytechnique, Institut Polytechnique de Paris, Sorbonne Université, École Normale Supérieure, PSL Research University, 91120 Palaiseau, France
- 2Institut de Planétologie et d’Astrophysique de Grenoble, Université Grenoble-Alpes, 38058 Grenoble, France
- 3ONERA/DOTA, BP 80100, chemin de la Hunière, 91123 Palaiseau, France
- 4SRON Netherlands Institute for Space Research, Utrecht, the Netherlands
- 5NOVELTIS, 31670 Labège, France
- 6Airbus Defence and Space, 31 rue des Cosmonautes, 31402 Toulouse, France
- 1Laboratoire de Météorologie Dynamique/IPSL, CNRS, École polytechnique, Institut Polytechnique de Paris, Sorbonne Université, École Normale Supérieure, PSL Research University, 91120 Palaiseau, France
- 2Institut de Planétologie et d’Astrophysique de Grenoble, Université Grenoble-Alpes, 38058 Grenoble, France
- 3ONERA/DOTA, BP 80100, chemin de la Hunière, 91123 Palaiseau, France
- 4SRON Netherlands Institute for Space Research, Utrecht, the Netherlands
- 5NOVELTIS, 31670 Labège, France
- 6Airbus Defence and Space, 31 rue des Cosmonautes, 31402 Toulouse, France
Abstract. Several single-platform satellite missions have been designed during the past decades in order to retrieve the atmospheric concentrations of anthropogenic greenhouse gases (GHG), initiating worldwide efforts towards better monitoring of their sources and sinks. To set up a future operational system for anthropogenic GHG emission monitoring, both revisit frequency and spatial resolution need to be improved. The Space CARBon Observatory (SCARBO) project aims at significantly increasing the revisit frequency of spaceborne GHG measurements, while reaching state-of-the-art precision requirements, by implementing a concept of small satellite constellation. It would accommodate a miniaturized GHG sensor named NanoCarb coupled with an aerosol instrument, the multi-angle polarimeter SPEXone. More specifically, the NanoCarb sensor is a static Fabry-Perot imaging interferometer with a 2.3 × 2.3 km2 spatial resolution and 200 km swath. It samples a truncated interferogram at optical path differences (OPDs) optimally sensitive to all the geophysical parameters necessary to retrieve column-averaged dry-air mole fractions of CO2 and CH4 (hereafter XCO2 and XCH4). In this work, we present the Level 2 performance assessment of the concept proposed in the SCARBO project. We perform inverse radiative transfer to retrieve XCO2 and XCH4 directly from synthetic NanoCarb truncated interferograms, and provide their systematic and random errors, column vertical sensitivities and degrees of freedom as a function of five scattering error-critical atmospheric and observational parameters. We show that NanoCarb XCO2 and XCH4 systematic retrieval errors can be greatly reduced with SPEXone posterior outputs used as improved prior aerosol constraints. For two thirds of the soundings, located at the centre of the 200 km NanoCarb swath, XCO2 and XCH4 random errors span 0.5–1 ppm and 4–6 ppb, respectively, compliant with their respective 1-ppm and 6-ppb precision objectives. Finally, these Level 2 performance results are parameterized as a function of the explored scattering error-critical atmospheric and observational parameters in order to time-efficiently compute extensive L2 error maps for future CO2 and CH4 flux estimation performance studies.
- Preprint
(12918 KB) -
Supplement
(3839 KB) - BibTeX
- EndNote
Matthieu Dogniaux et al.
Status: final response (author comments only)
-
RC1: 'Comment on amt-2021-224', Anonymous Referee #2, 01 Nov 2021
The authors introduce a new concept called the Space CARBon Observatory (SCARBO), which aims to measure CO2 and CH4 from a constellation of ~20 satellites in sun-synchronous low Earth orbit, with a multi-angle polarimetric aerosol instrument to account for scattering effects. SCARBO will have higher spatial coverage and revisit frequency compared to existing greenhouse gas missions. The authors assess the performance of SCARBO for a variety of scenarios, both with and without the aerosol instrument. They find that systematic errors in column-averaged CO2 and CH4 (XCO2 and XCH4) retrievals can be greatly reduced by using aerosol information from the polarimeter. The authors also parameterize results as a function of relevant parameters in order to facilitate efficient computation of error maps for CO2 and CH4 flux estimation.
The manuscript is well written and the topic extremely relevant to the greenhouse gas remote sensing community. However, a few issues need to be addressed before it is ready for publication.
Lines 148-149: “Entanglements between CO2, CH4, O2, H2O and aerosols signals have been considered, with the assumption that albedo models are constant over all four spectral bands.” What is the impact of varying albedo on the results? Also, only soil, vegetation and desert types are considered. What about water? Many emission sources (e.g., power plants) are near the ocean, so coastal areas would need to be considered.
Lines 235-236: “The interfering impact of temperature has not been taken into account for the latest optimized OPD selection used in this work, and is not considered in the state vector.” What is the impact of this assumption on the retrievals?
What is the impact of retrieving profile scaling factors for CO2 and CH4 as opposed to retrieving the vertical profile (that is traditionally done by OCO-2, for example)? Have the authors assessed the impacts on accuracy and on downstream flux estimation?
Lines 300-301: “For this synthetic performance study, constant trace gas concentration profiles have been used: 394.85 ppm for CO2 and 1855.3 ppb for CH4.” This seems (unnecessarily) restrictive (see also previous comment). There needs to be an assessment of how results change for realistic CO2 and CH4 profiles.
Aerosols: the authors might want to say that the fine mode particles are assumed to be spherical. It would also be useful to have a sentence describing how the aerosol single scattering properties were calculated (e.g., Mie for spherical, T-Matrix for spheroidal?).
Is SCARBO only going to make measurements in the nadir mode? If not, the viewing zenith angle needs to be a parameter that is considered in the evaluation of the scattering error.
Grammatical Errors / Typos:
Line 118: Acronym OPD already defined
Line 151: FOV (2) an analytical approximation -> FOV, and (2) an analytical approximation
Lines 152, 258: “line-by-line” would be more appropriate than “pseudo-infinite”
Lines 171-172: “The constellation sizing aims at ensuring intra-daily revisit of the largest possible amount of anthropogenic CO2 emission hotspots which emission rate is compatible with the 1 ppm SCARBO ð!!! precision objective.” Awkwardly phrased
Line 173: “performed” -> “compiled”?
Line 178: remove “a” before “global coverage” and “daily revisit”
Line 179: compromise well -> provides an optimal compromise
Line 191: measures -> measurements
Lines 197-198: what is meant by “without artificial noise”? The text indicates that instrument noise is considered in the retrievals. I would recommend removing this phrase to avoid confusion.
Line 260: measure -> measurement
Line 268: Acronym FOV already defined
Line 276: fasten -> speed up
Line 368: “more disadvantageous” is too vague. Please use a more descriptive term.
Lines 369-360: “more favourable” please use a more quantitative term (more forward scattering?)
Line 448: of all atmospheric layers -> in all atmospheric layers
Line 498: on the optical path -> in the optical path
- AC1: 'Reply on RC1', Dogniaux Matthieu, 30 Jun 2022
-
RC2: 'Comment on amt-2021-224', Anonymous Referee #3, 03 Apr 2022
General comments
The authors present a simulation-based assessment of CO2 and CH4 column retrievals for a novel nanosatellite constellation concept called SCARBO. The constellation would involve 20+ small satellites each carrying a Fabry-Perot interferometer (NanoCarb) for CH4 and CO2 and a multi-angle polarimeter (SPEXone) for aerosols. The authors present extensive error analysis with a focus on aerosol-related retrieval errors and the ability of the SPEXone auxiliary instrument to mitigate those errors. They use for this purpose standard OSSE methods including Rodgers (2000) optimal estimation techniques, and show that their NanoCarb-SPEXone instrument design should be capable of delivering high-precision XCO2 and XCH4 retrievals. The paper is well written and a good fit for AMT. I recommend acceptance for publication subject to the following comments and questions.
Specific comments
- L. 36: Please clarify what is meant by “on small areas”.
- L. 40: Can you provide a reference for 2x2 km2 resolution being fine enough to resolve point sources? Of what magnitude? TROPOMI can resolve only extreme methane point sources at similar (5.5x7 km2) resolution (e.g., Pandey et al., 2019).
- L. 41: Extensive recent work has shown that plumes observed by imaging spectrometers do not look Gaussian. For example Cusworth et al. (2021) use an integrated mass enhancement method to quantify CO2 emissions from individual power plants observed by the PRISMA satellite instrument, and the TROPOMI team and others have used a variety of methods to quantify CH4 plume emission rates (eg, Pandey et al., 2019) at km-scale resolution, but Gaussian plume modeling seems poorly suited to the problem.
- L. 66: “requirements for operational top-down monitoring of anthropogenic GHG emissions” – What are these requirements and what does “operational” CO2/CH4 monitoring mean?
- L. 79: “geophysical parameters necessary to retrieve XCO2 and XCH4” – can you say what these parameters are or point to them in the text?
- L. 80-81: “2.3 x 2.3 km2 spatial resolution, enabling to detect emission plumes from megacities and hotspots (e.g. > 10 Mt CO2 yr-1 power plants)” – You seem to use "hotspot" and "point source" interchangeably. Megacities are examples of hotspots and power plants examples of point sources. For point sources, where does the 10 Mt/y threshold come from?
- L. 86: Please describe CO2M.
- L. 95: “scattering error-critical atmospheric and observational parameters” – Are these the “geophysical parameters” you mentioned before? It would be helpful to list these out somewhere in the introduction if not too lengthy or point to them in the text.
- Ray Nassar and Dan Cusworth’s works about satellite monitoring of CO2 emissions from power plants should be cited somewhere. Same for TROPOMI methane plume papers (Pandey et al. and others) since there has been a lot of recent work on these topics.
- Can you explain why SCARBO uses an FP rather than grating? Is it about financial cost, instrument size/weight, something else? Also please cite other instruments/concepts that use FP - eg Jervis et al. (2021).
- L. 165: Not clear what “0.003” means, is it an error (1 or 2 sigma)?
- L. 173: What is an “emission clump”? This terminology is non-standard.
- L. 234: Are you referring to the instrument temperature?
- Table 2: How conservative is the 4 hPa (0.4%) error for surface pressure? It seems quite small.
- L. 271: I would suggest pointing to the appendix here because I initially wondered if the “combination” of single-pixel measurements was through averaging or something more.
- L. 289-290: “Errors arising from the interpolation have been assessed and are negligible (not shown)” – What is the magnitude of the error?
- L. 297-300: CH4 falls off rapidly above the troposphere, so why use a uniform vertical profile? Is the impact of this unrealistic profile on the retrieval small enough to be neglected?
- Fig. 7: Is the much lower DOFs for FOD in the with-SPEX scenario merely due to using a much lower prior error for FOD compared to no-SPEX?
- L. 424-425: Do the albedo DOFs actually increase? They seem to be equal to 1 in both scenarios.
- L. 437-439: This seems odd since the albedo DOFs look to be almost or exactly 1.0.
- Fig. 8: It’s not clear how you compute column averaging kernels when your state vector doesn’t include a vertical column but rather a single scaling factor for each gas. If the state vector included CH4 and CO2 at different vertical levels then you would obtain A = dxhat/dx giving the AK for each vertical layer. How do you get column averaging kernels when optimizing just a scaling factor?
- Fig. 8: Also the column averaging kernels look quite smooth, what is the vertical resolution here?
- Several plots show regions where retrievals did not converge satisfactorily but I cannot find in the manuscript what method you use for the optimization. Is it Newton, Levenberg-Marquardt, something else?
- Fig. 13: Why is there striping in this figure? Because of the loss of precision with increasing transversal position?
Technical corrections
- L. 39: “large-swath”
- L. 51: “best fits”
- L. 151: “radiance spectrum”
- L. 164: “at 50 spectral band” seems like a typo?
- L. 178: “compromises”
- L. 276: “fasten” doesn’t seem like the right word here. Do you mean “speed up” or something similar?
- L. 422: “in these situations” typo
- L. 612: “then” typo
- L. 637: “mentioned” typo
- Consider changing “scattering error-critical” to “scattering-error-critical” everywhere. If I understand correctly it’s meant to be a compound adjective and might be clearer with two –'s.
References
Pandey, S., Gautam, R., Houweling, S., van der Gon, H. D., Sadavarte, P., Borsdorff, T., Hasekamp, O., Landgraf, J., Tol, P., van Kempen, T., Hoogeveen, R., van Hees, R., Hamburg, S. P., Maasakkers, J. D., and Aben, I.: Satellite observations reveal extreme methane leakage from a natural gas well blowout, P. Natl. Acad. Sci. USA, 116, 26376–26381, https://doi.org/10.1073/pnas.1908712116, 2019.
D.H. Cusworth, R.M. Duren, A.K. Thorpe, M.L. Eastwood, R.O. Green, P.E. Dennison, C. Frankenberg, J.W. Heckler, G.P. Asner, C.E. Miller: Quantifying global power plant carbon dioxide emissions with imaging spectroscopy, AGU Adv., 2 (2) (2021), https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1029/2020AV000350
Jervis, D., McKeever, J., Durak, B. O. A., Sloan, J. J., Gains, D., Varon, D. J., Ramier, A., Strupler, M., and Tarrant, E.: The GHGSat-D imaging spectrometer, Atmos. Meas. Tech., 14, 2127–2140, https://doi.org/10.5194/amt-14-2127-2021, 2021.
- AC2: 'Reply on RC2', Dogniaux Matthieu, 30 Jun 2022
Matthieu Dogniaux et al.
Matthieu Dogniaux et al.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 586 | 184 | 17 | 787 | 38 | 10 | 8 |
- HTML: 586
- PDF: 184
- XML: 17
- Total: 787
- Supplement: 38
- BibTeX: 10
- EndNote: 8
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1