the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 10 Sep 2026
The air-sea CO2 fluxes and ship CO2 emissions in the Baltic Sea during 2019–2021, using in-situ and inversion methods
Abstract. Estimating air–sea CO2 exchange in coastal and semi-enclosed seas is challenging due to limited spatial and temporal observational coverage. Atmospheric inverse modelling offers a complementary approach, although its ability to resolve regional marine carbon fluxes is still uncertain. In the Baltic Sea, intensive shipping further complicates the regional carbon budget through substantial anthropogenic CO2 emissions. This study presents one of the first assessments combining air–sea CO2 exchange and ship-derived CO2 emissions in the Baltic Sea using both observation-based and atmospheric inversion approaches.
Air–sea CO2 fluxes for 2019 to 2021 were estimated using two independent methods. The first method combined a bulk flux calculation, which includes sea–air pCO2 differences, wind-speed-dependent gas transfer velocity, solubility, and sea-ice cover, with shipping emissions simulated by the Ship Traffic Emission Assessment Model (STEAM). The second method used atmospheric CO2 observations together with atmospheric transport modelling in the CIF-FLEXPART inversion framework to estimate surface CO2 fluxes, providing an alternative observation-based approach. The results were analyzed for the entire Baltic Sea and for 13 HELCOM sub-basins.
The STEAM-corrected bulk method indicated that the Baltic Sea acted as a net CO2 sink during 2019 to 2021, with a mean flux of -1.79 mmol m−2 d−1. Without shipping emissions, the estimated flux was -3.76 mmol m−2 d−1, showing that ship traffic increased the net flux by approximately 1.97 mmol m−2 d−1. In contrast, the CIF-FLEXPART inversion estimated a net source of 3.92 mmol m−2 d−1 for the Baltic Sea. The climatological bulk-flux estimate for 2003 to 2021, excluding shipping emissions, gave a near-neutral mean flux of 0.25 mmol m−2 d−1. The basin-scale monthly fluxes showed moderate overall agreement between the two methods (mean Pearson correlation r = 0.49), with stronger correlations (r = 0.62–0.79) in several southern basins where observational coverage was greatest.
The contrasting estimates of the Baltic Sea carbon balance show that considerable uncertainties remain in both bulk method and inversion method approaches and that shipping emissions need to be taken into account. More continuous and reliable observations, improved representation of coastal carbon processes, and further development of atmospheric inversion systems are needed to better constrain regional marine carbon budgets.
- Preprint
(7906 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2026-5442', Anonymous Referee #1, 05 Nov 2026
-
RC2: 'Comment on egusphere-2026-5442', Anonymous Referee #2, 05 Nov 2026
I think the analysis of Kettunen et al. is a potentially useful application of multiple methods to estimate basin-wide CO2 fluxes in the Baltic Sea and to better understand controls on CO2 fluxes. I think the paper would be much more impactful if it was written more clearly and discussed the results with a more mechanistic perspective.
General comments:
(1) I did not originally understand how exactly the shipping emissions factored into the analysis. It seems to me now that the methods allow for the impact of shipping on local atmospheric CO2 concentrations to be represented so that the atmospheric CO2 concentration is more representative than a large-scale aggregate. Am I correct? If so, I think this could more clearly be communicated early in the methods.
(2) I found the introduction to be overly long and I suggested specific changes below
(3) For someone unfamiliar with Bayesian approaches, this analysis would be hard to follow. "prior" and "posterior" are not simply and clearly defined in the methods and the key equations are could be better explained.
(4) I found the discussion to be unsatisfying. There are paragraphs where many different estimates of CO2 flux are stated, but there is no real interpretation of what the differences mean or why they might be different (e.g. lines 318-323). There are paragraphs that describe that some basins are different than others, and this varies by method, but no discussion of why those basins are different due to physics, biology, etc. I think the authors should decide what 4-5 key points they would like to make and organize the discussion that way.
(5) I know this might not help, but this paper reads more like a technical report one might give to a funding agency, and less like a science paper. This is fixable without a tonof work.
Specific comments:
(1) Line 38: I think this first sentence is about the Baltic specifically, but that is not stated.
(2) Lines 48-53: This paragraph interrupts the flow of the introduction with an unnecessary summary of some general features of the Baltic region. I would delete it.
(3) Lines 69-61: I don't know if a reader would understand at this point, without additional text, why one would couple the shipping estimates and the air-sea flux together. One is a controlled anthropogenic activity that impacts the atmosphere. The other is an environmental process driven by many factors that influences the atmopshere and water. If you indluce ships, why not include other coastal CO2 sources?
(4) Lines 63-66: This is methods and oddly placed here. I would remove it.
(5) Lines 67-76: The topic sentence seems ok, but the details on the inverse methods that follow in this paragraph are not really helpful here. I might move this paragraph to the methods and use it to clearly state what the various estimates are that you use in this analysis, including the terms (e.g., prior fluxes).
(6) Line 164: I think the Bayesian Cost Function is J(x) in the equation that follows. If so, state that clearly.
(7) Line 165: Define what "H" is
(8) Lines 196-202: These 2 little paragraphs belong in the methods - they are distracting as an opening to the results.
(9) Lin 214: It seems to be that "including ship emissions" could be communicated as "accounting for ship emissions". My understanding is that the emissions from the ships are not a simple addition or substraction from the air-sea fluxes. A reminder that this approach is comfusing as written.
(10) I think you need to do a better job of explaining the difference between the prior and posterior fluxes are. Not everyone is familiar with the bayesian terminology and what the values represent in this analysis. Perhaps a simple explanation of this in the methods (using exisiting text from introduction - see above - would help)
(11) Line 337: "excepted" should be "accepted"
Citation: https://doi.org/10.5194/egusphere-2026-5442-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 274 | 135 | 56 | 465 | 66 | 63 |
- HTML: 274
- PDF: 135
- XML: 56
- Total: 465
- BibTeX: 66
- EndNote: 63
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Manuscript title: The air-sea CO₂ fluxes and ship CO₂ emissions in the Baltic Sea during 2019–2021, using in-situ and inversion methods
The manuscript addresses an important scientific question by combining observation-based air–sea CO₂ flux estimates, shipping emissions from the STEAM model, and atmospheric inversion modelling using CIF-FLEXPART. The comparison provides valuable insights into uncertainties in the Baltic Sea carbon budget. However, several methodological issues and uncertainties require further clarification before the conclusions can be considered sufficiently robust.
Major Review Comments
1. Large discrepancy between the bulk and inversion-based CO₂ flux estimates
The most significant finding is the substantial difference between the STEAM-corrected bulk estimate (−1.79 mmol m⁻² d⁻¹) and the CIF-FLEXPART inversion estimate (+3.92 mmol m⁻² d⁻¹). Although the authors discuss potential uncertainties, the underlying reasons for this discrepancy are not sufficiently quantified. The authors should provide a more systematic assessment of the contributions from atmospheric transport errors, prior flux assumptions, observational constraints, and anthropogenic emissions. A quantitative uncertainty analysis or sensitivity experiments would strengthen the interpretation of these contrasting results.
2. Uncertainty associated with limited spatial and temporal coverage of seawater pCO₂ observations
The bulk flux estimates rely heavily on pCO₂ observations collected along the M/S Finnmaid shipping route, resulting in uneven spatial coverage across the Baltic Sea. This limitation is particularly important in the northern basins, where the manuscript reports unusually strong CO₂ uptake, including during winter. Furthermore, replacing January 2019 and December 2020 pCO₂ values with estimates from other months or years may introduce additional uncertainty. The authors should evaluate how these data gaps and substitutions influence the calculated annual fluxes and basin-scale carbon budgets. Sensitivity experiments excluding poorly constrained regions or using alternative gap-filling approaches would improve confidence in the results.
3. Separation of shipping emissions and natural air–sea CO₂ exchange in the inversion framework
The inclusion of shipping emissions is an important contribution of this study. However, the manuscript indicates that the anthropogenic component dominates the inversion-derived fluxes in several basins, and even produces negative anthropogenic posterior fluxes during some months. This raises concerns regarding the ability of the inversion framework to distinguish shipping-related CO₂ emissions from natural air–sea exchange. The authors should clarify how shipping emissions are represented and constrained during the inversion and assess whether uncertainties in the anthropogenic prior emissions could influence the optimized oceanic fluxes. A sensitivity experiment with fixed shipping emissions or alternative anthropogenic prior uncertainties would help establish the robustness of the inversion results.
4. Robustness of the gas transfer velocity parameterization and associated flux uncertainties
The bulk flux calculations employ the Wanninkhof (2014) wind-speed-based gas transfer velocity parameterization with an assumed uncertainty of ±20%. However, gas transfer velocities in semi-enclosed and coastal environments such as the Baltic Sea can be strongly influenced by fetch limitation, wave conditions, and surface-active substances. Although these limitations are acknowledged in the discussion, their potential influence on the estimated CO₂ fluxes is not quantitatively evaluated. The authors are encouraged to compare the results using alternative gas transfer velocity parameterizations appropriate for coastal waters and examine how these choices affect the seasonal and annual CO₂ flux estimates, particularly in regions where the two methods disagree substantially.