Constraints on methane emissions in North America from future geostationary remote-sensing measurements

Bousserez, Nicolas; Henze, Daven K.; Rooney, Brigitte; Perkins, Andre; Wecht, Kevin J.; Turner, Alexander J.; Natraj, Vijay; Worden, John R.

doi:https://doi.org/10.5194/acp-16-6175-2016

Articles | Volume 16, issue 10

https://doi.org/10.5194/acp-16-6175-2016

© Author(s) 2016. This work is distributed under
the Creative Commons Attribution 3.0 License.

https://doi.org/10.5194/acp-16-6175-2016

© Author(s) 2016. This work is distributed under
the Creative Commons Attribution 3.0 License.

Articles | Volume 16, issue 10

Research article

|

20 May 2016

Research article |

| 20 May 2016

Constraints on methane emissions in North America from future geostationary remote-sensing measurements

Nicolas Bousserez, Daven K. Henze, Brigitte Rooney, Andre Perkins, Kevin J. Wecht, Alexander J. Turner, Vijay Natraj, and John R. Worden

Download

Final revised paper (published on 20 May 2016)
Preprint (discussion started on 10 Jul 2015)

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC C5412: 'Review', Anonymous Referee #2, 29 Jul 2015
- AC C12412: 'Responses to anonymous referee #2', Nicolas Bousserez, 15 Feb 2016
RC C5475: 'Review', Anonymous Referee #1, 30 Jul 2015
- AC C12410: 'Responses to anonymous referee #1', Nicolas Bousserez, 15 Feb 2016

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

AR by Nicolas Bousserez on behalf of the Authors (15 Feb 2016) Author's response Manuscript

ED: Referee Nomination & Report Request started (08 Mar 2016) by Mathias Palm

RR by Anonymous Referee #1 (14 Mar 2016)

Suggestions for revision or reasons for rejection

I am impressed by the seriousness the authors’ revision of their paper. There are still a few pending issues that I am listing hereafter.

Major issues:
- My question about retrieval error as a function of satellite altitude: a same instrument receives less photons on average when placed at a higher altitude, and the uncertainty of the column retrieval consequently (and significantly) increases. This effect does not seem to be accounted for here.
- My question about the absence of prior error correlations: the authors justify their diagonal matrix by the impossibility to compute “accurate” or “rigorous” prior error correlations (l. 188, l. 439). However, in the absence of perfect knowledge, the authors do not justify why zero is better than, e.g., an e-folding length of 300 km. Taking the extreme case may not be a fair choice, which the authors implicitly acknowledge l. 308 by speculating on a resulting systematic bias in their results (actually more details on this speculation would be needed because, if the existence of a bias seems obvious, its “sign” is not to me).
- L. 251: am I missing something or have the authors forgotten about model temporal error correlations? There are likely very large from one hour to the next (see, e.g., Lauvaux et al., doi:10.5194/bg-6-1089-2009).

Minor issues:
- What is the link between l. 89 “their computational cost can be prohibitive, since many perturbed inversions (typically about 50) are needed” and l. 129 “Here an ensemble of 500 random gradients of the cost function are used”. The word “prohibitive” seems to be author-centric.
- At several places (at least l. 223 and 255), the authors use the generic word “error” in place of “error standard deviation”. The authors should use the exact expression.
- L. 228: some details are needed about the vertically resolved covariance matrix.
- L. 429: “nation” likely means the US, but it should be explicited.

Hide

ED: Reconsider after minor revisions (Editor review) (15 Mar 2016) by Mathias Palm

Dear authors, after consideration of your revised manuscript, there are still some questions open, but they seem comparatively minor as compared to the first reviews.

Please consider the question raised the the third review:

Major issues:
- My question about retrieval error as a function of satellite altitude: a same instrument receives less photons on average when placed at a higher altitude, and the uncertainty of the column retrieval consequently (and significantly) increases. This effect does not seem to be accounted for here.
- My question about the absence of prior error correlations: the authors justify their diagonal matrix by the impossibility to compute “accurate” or “rigorous” prior error correlations (l. 188, l. 439). However, in the absence of perfect knowledge, the authors do not justify why zero is better than, e.g., an e-folding length of 300 km. Taking the extreme case may not be a fair choice, which the authors implicitly acknowledge l. 308 by speculating on a resulting systematic bias in their results (actually more details on this speculation would be needed because, if the existence of a bias seems obvious, its “sign” is not to me).
- L. 251: am I missing something or have the authors forgotten about model temporal error correlations? There are likely very large from one hour to the next (see, e.g., Lauvaux et al., doi:10.5194/bg-6-1089-2009).

Minor issues:
- What is the link between l. 89 “their computational cost can be prohibitive, since many perturbed inversions (typically about 50) are needed” and l. 129 “Here an ensemble of 500 random gradients of the cost function are used”. The word “prohibitive” seems to be author-centric.
- At several places (at least l. 223 and 255), the authors use the generic word “error” in place of “error standard deviation”. The authors should use the exact expression.
- L. 228: some details are needed about the vertically resolved covariance matrix.
- L. 429: “nation” likely means the US, but it should be explicited.

Kind regards
Mathias Palm

Hide

AR by Nicolas Bousserez on behalf of the Authors (25 Mar 2016) Author's response Manuscript

ED: Publish as is (15 Apr 2016) by Mathias Palm

Short summary

This work provides new insight into the observational constraints provided by current low-Earth orbit (LEO) and future potential geostationary (GEO) satellite missions on methane emissions in North America. Using efficient numerical tools, the information content (error reductions, spatial resolution of the constraints) of methane inversions using different instrument configurations (TIR, SWIR and multi-spectral) was estimated at model grid-scale resolution (0.5° × 0.7°).