the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 06 Oct 2025
Response of a semi-enclosed sea to perturbed freshwater and open ocean salinity forcing
Abstract. The sensitivity of Baltic Sea salinities to changed fresh water forcing and other forcing factors have been debated during the last decades, since changed salinities would have large impacts on the marine ecosystems, and since this parameter still shows a high degree of uncertainty in regional climate projections. In this study we performed a sensitivity study where fresh water forcing and salinities at the outer boundaries of the North Sea were perturbed in a systematic way in order to obtain a second-order Taylor polynomial of the statistical steady state mean salinity. The polynomial was constructed based on perturbations of a 57-year long hindcast run for the period 1961–2017 with a regional ocean model covering the North Sea and the Baltic Sea. The results show that the Baltic sea is highly sensitive to fresh water forcing and that only about one third of the boundary salinity change propagates into the Baltic Sea. The results are also analysed in terms of a total exchange flow analysis in the entrance region, and it is found that the Baltic Sea salinity sensitivity to a large degree can be explained by increased freshwater input causing (1) dilution inside the Baltic Sea, (2) decreased inflows caused by changes to the mean sea level gradient in the entrance region, and (3) reduced inflow salinities due to recirculation of outflowing Baltic water in the entrance region where the inflow water consists of about two parts outflowing Baltic water and one part North Sea water.
- Preprint
(1443 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-4735', Anonymous Referee #1, 13 Oct 2025
-
AC1: 'Reply on RC1', Lars Arneborg, 15 Dec 2025
Referee comments in italics and our answers in normal text.
Review of manuscript “Response of a semi-enclosed sea to perturbed freshwater and open ocean salinity forcing” by Arneborg et al. 2025
This study uses high-resolution model sensitivity experiments to examine how the Baltic Sea’s steady-state salinity responds to variations in freshwater forcing and salinity at the boundary of North Sea. From these experiments, the authors constructed a second-order polynomial that relates the basin-mean steady-state salinity to changes in freshwater forcing and boundary salinity. Results show that the Baltic Sea’s response to freshwater forcing is large and non-linear, whereas its response to boundary salinity is more linear but less significant. The authors also analyze the impact of freshwater forcing and boundary salinity changes on the freshwater volume fluxes in and out of Baltic Sea driven by circulation changes.
Overall, the manuscript presents some important results, but it needs some revising and reorganizing before it can be considered for publication. The detailed comments are provided below:
Thanks for a detailed and constructive review. We have answered the comments below, and described how we will improve the manuscript according to the comments where relevant.
1. The study examines the sensitivity of Baltic Sea salinity to net freshwater input into both the Baltic and the North Sea, as well as to variations in salinity at the North Sea boundary. However, it is not clear which specific boundary of the North Sea is used for prescribing or evaluating the boundary salinity. Could you clarify this, and indicate the location on Figure 1? Additionally, it is very likely that changes in the net freshwater forcing would influence the boundary salinity in the North Sea itself. If so, how do you separate the individual impacts of the freshwater forcing and boundary salinity within your analysis?
We mean the open boundary salinities and will make this clearer in the text. It will also be made clearer that the boundary salinity change is intended to represent changes in Nort-East Atlantic salinities rather than salinities at the boundary of the Baltic Sea.
The separation between boundary salinity and fresh water forcing is done by construction in the experiments. When we raise the runoff in the North Sea we freshen it a little bit. Whatever, presumably small, effect that has on our experiments counts as a runoff effect, not as an open boundary salinity effect. So they are naturally separated by the experiment
The influence of fresh-water forcing on salinities in the North Sea are seen in Fig. 5 for the RP+ and RP- cases. At the boundaries the influence is almost zero. This may partly be because the boundaries are specified so, and in reality there may be an influence caused by recirculation of outflowing water into the domain, which we do not capture with a one-way coupled model. The focus in this study is on the changes in the Baltic Sea, and we have placed the open boundary further away from the Baltic Sea than most previous studies. The inflows to the Baltic Sea do include influences of outflowing fresh water to the North Sea, so we believe that we do have a good separation between impacts of freshwater forcing and boundary salinities at the boundary of the model.
2. The figure captions are vaguely written. For example, in Fig.2, it is not clear during which period the observed salinity profiles are compared with the hindcast profiles? Are these averaged over a period?
The time period will be included in the captions. The solid line is the temporal mean and the dashed lines are the 5th and 95th percentiles as described in the caption.
3. Figure 2: I suggest plotting the observed salinity profiles in thin lines rather than dots. The multiple grey dots at each depth level
We tend to think that is nice to see the observational points as in the present plot.
4. Figure 3: Please report quantitative statistics for these comparisons. How do the mean and standard deviation of salinity differ between the observations and the hindcast experiments? It would be helpful to list metrics such as bias, root-mean-square error (RMSE), and correlation coefficients.
From the figure, it appears that the bias in the hindcast salinity increases with depth. To better illustrate this, you could include a vertical profile of correlation coefficients (with statistical significance indicated) to show how model–observation agreement varies with depth.
We could add these metrics, but we do not see the value of doing so for this paper. If doing an intercomparison study or optimizing a model, such metrics would make sense, but here we do not see how the metrics would help the reader to decide whether the model describes the necessary processes or not. We don’t know about any relation between, e.g., RMSE for salinity and the model sensitivity to boundary salinity changes. To help the reader, we will instead refer to the Baltic Sea Model Intercomparison project, adding a sentence like: “Generally, the model salinities compare as well with observations as those of the better models in the Baltic Sea Model Intercomparison Project (Gröger et al., 2022) and the model is therefore well suited for the sensitivity study which is the focus of this work.”
5. Figure 4 needs more explanation. Note that the salinity trends are different RP- and RP-SB- runs as compared to the CTL and all other runs. Can this be explained? When you say yearly salinities, do you mean the mean salinity values averaged over each year? The caption is not clear.
We will add more explanation to describe what the figure shows, including “when the freshwater forcing decreases (blue lines), the salinities increase”. We write “yearly mean salinities averaged over the whole Baltic Sea volume inside the sill transect (Fig. 1)” in the text and “Yearly mean salinities for the Baltic Sea averaged over the volume inside the sill transect” in the captions.
6. Figure 5: How do these salinity anomaly maps for different runs compare with the CTL run?
The hindcast run (CTL) would only show zeros. We have added “(CTL)” to make it clearer that CTL is the hincast run.
7. Figure 6 caption is too vague. Specify the units.
Caption will be improved.
8. Figures 8 and 9: Over what period have these volume and salt fluxes been integrated? Is there no inflow of salt into the Baltic sea, given all positive values of F in Fig. 9?
The period is 1990 – 2017 which will now be mentioned also in the caption. There are inflows of about 250000 kg/s (maximum value of F) on average for the CTL case. Have we written somewhere that there are no inflows?
9. Figure 11: In the caption, you need to mention that these are the differences in fluxes between the RP+ and RP- experiments.
We will add the following to the caption: “The fluxes are based on the differences between runs RP+ and RP- but normalized to represent an increase in Qf of 100 m3/s.”
10. I recommend moving the equations and the calculations estimating percentage changes in salinity due to variations in inflows, outflows, and related terms from the Summary section to the Results section. The final section should focus primarily on synthesizing and highlighting the main findings.
We prefer not to move these calculations and discussion. The main results of the study are the polynomial and the total exchange flow analysis results based on the 3D model calculations. The calculations in the discussion part of Section 4 are more speculative with a number of underlying assumptions, and are meant more as a suggested simplified interpretation of the Baltic Sea sensitivity to the investigated forcing factors than a firm result of the simulations.
11. The manuscript would be strengthened if the authors placed their results in better context with previous studies. In the Introduction, the authors state that their work differs from earlier studies because the model includes the North Sea domain. Could the authors elaborate on what difference this inclusion makes for the analysis and interpretation of salinity variability in the Baltic Sea? How do the results compare with those from previous studies that did not include the North Sea
We will extend the text comparing our results with earlier studies in Section 4. We will not focus too much on the importance of including the North Sea domain, but we will write in the introduction that this is important in order to study the influence of North Atlantic salinities on the Baltic Sea, which has been investigated to a very limited degree in the past.
Citation: https://doi.org/10.5194/egusphere-2025-4735-AC1 -
AC2: 'Reply on AC1', Lars Arneborg, 15 Dec 2025
Sorry, but this seems to be the reply to anonymous referee # 2. We will upload the correct reply below.
Citation: https://doi.org/10.5194/egusphere-2025-4735-AC2
-
AC2: 'Reply on AC1', Lars Arneborg, 15 Dec 2025
- AC3: 'Reply on RC1', Lars Arneborg, 15 Dec 2025
-
AC1: 'Reply on RC1', Lars Arneborg, 15 Dec 2025
-
RC2: 'Comment on egusphere-2025-4735', Anonymous Referee #2, 18 Nov 2025
Please find the comments in the attached file.
-
AC4: 'Reply on RC2', Lars Arneborg, 15 Dec 2025
Referee comments in italics and our answers in normal text.
Review of manuscript “Response of a semi-enclosed sea to perturbed freshwater and open ocean salinity forcing” by Arneborg et al. 2025
This study uses high-resolution model sensitivity experiments to examine how the Baltic Sea’s steady-state salinity responds to variations in freshwater forcing and salinity at the boundary of North Sea. From these experiments, the authors constructed a second-order polynomial that relates the basin-mean steady-state salinity to changes in freshwater forcing and boundary salinity. Results show that the Baltic Sea’s response to freshwater forcing is large and non-linear, whereas its response to boundary salinity is more linear but less significant. The authors also analyze the impact of freshwater forcing and boundary salinity changes on the freshwater volume fluxes in and out of Baltic Sea driven by circulation changes.
Overall, the manuscript presents some important results, but it needs some revising and reorganizing before it can be considered for publication. The detailed comments are provided below:
Thanks for a detailed and constructive review. We have answered the comments below, and described how we will improve the manuscript according to the comments where relevant.
1. The study examines the sensitivity of Baltic Sea salinity to net freshwater input into both the Baltic and the North Sea, as well as to variations in salinity at the North Sea boundary. However, it is not clear which specific boundary of the North Sea is used for prescribing or evaluating the boundary salinity. Could you clarify this, and indicate the location on Figure 1? Additionally, it is very likely that changes in the net freshwater forcing would influence the boundary salinity in the North Sea itself. If so, how do you separate the individual impacts of the freshwater forcing and boundary salinity within your analysis?
We mean the open boundary salinities and will make this clearer in the text. It will also be made clearer that the boundary salinity change is intended to represent changes in Nort-East Atlantic salinities rather than salinities at the boundary of the Baltic Sea.
The separation between boundary salinity and fresh water forcing is done by construction in the experiments. When we raise the runoff in the North Sea we freshen it a little bit. Whatever, presumably small, effect that has on our experiments counts as a runoff effect, not as an open boundary salinity effect. So they are naturally separated by the experiment
The influence of fresh-water forcing on salinities in the North Sea are seen in Fig. 5 for the RP+ and RP- cases. At the boundaries the influence is almost zero. This may partly be because the boundaries are specified so, and in reality there may be an influence caused by recirculation of outflowing water into the domain, which we do not capture with a one-way coupled model. The focus in this study is on the changes in the Baltic Sea, and we have placed the open boundary further away from the Baltic Sea than most previous studies. The inflows to the Baltic Sea do include influences of outflowing fresh water to the North Sea, so we believe that we do have a good separation between impacts of freshwater forcing and boundary salinities at the boundary of the model.
2. The figure captions are vaguely written. For example, in Fig.2, it is not clear during which period the observed salinity profiles are compared with the hindcast profiles? Are these averaged over a period?
The time period will be included in the captions. The solid line is the temporal mean and the dashed lines are the 5th and 95th percentiles as described in the caption.
3. Figure 2: I suggest plotting the observed salinity profiles in thin lines rather than dots. The multiple grey dots at each depth level
We tend to think that it is nice to see the observational points as in the present plot.
4. Figure 3: Please report quantitative statistics for these comparisons. How do the mean and standard deviation of salinity differ between the observations and the hindcast experiments? It would be helpful to list metrics such as bias, root-mean-square error (RMSE), and correlation coefficients.
From the figure, it appears that the bias in the hindcast salinity increases with depth. To better illustrate this, you could include a vertical profile of correlation coefficients (with statistical significance indicated) to show how model–observation agreement varies with depth.
We could add these metrics, but we do not see the value of doing so for this paper. If doing an intercomparison study or optimizing a model, such metrics would make sense, but here we do not see how the metrics would help the reader to decide whether the model describes the necessary processes or not. We don’t know about any relation between, e.g., RMSE for salinity and the model sensitivity to boundary salinity changes. To help the reader, we will instead refer to the Baltic Sea Model Intercomparison project, adding a sentence like: “Generally, the model salinities compare as well with observations as those of the better models in the Baltic Sea Model Intercomparison Project (Gröger et al., 2022) and the model is therefore well suited for the sensitivity study which is the focus of this work.”
5. Figure 4 needs more explanation. Note that the salinity trends are different RP- and RP-SB- runs as compared to the CTL and all other runs. Can this be explained? When you say yearly salinities, do you mean the mean salinity values averaged over each year? The caption is not clear.
We will add more explanation to describe what the figure shows, including “when the freshwater forcing decreases (blue lines), the salinities increase”. We write “yearly mean salinities averaged over the whole Baltic Sea volume inside the sill transect (Fig. 1)” in the text and “Yearly mean salinities for the Baltic Sea averaged over the volume inside the sill transect” in the captions.
6. Figure 5: How do these salinity anomaly maps for different runs compare with the CTL run?
The hindcast run (CTL) would only show zeros. We have added “(CTL)” to make it clearer that CTL is the hincast run.
7. Figure 6 caption is too vague. Specify the units.
Caption will be improved.
8. Figures 8 and 9: Over what period have these volume and salt fluxes been integrated? Is there no inflow of salt into the Baltic sea, given all positive values of F in Fig. 9?
The period is 1990 – 2017 which will now be mentioned also in the caption. There are inflows of about 250000 kg/s (maximum value of F) on average for the CTL case. Have we written somewhere that there are no inflows?
9. Figure 11: In the caption, you need to mention that these are the differences in fluxes between the RP+ and RP- experiments.
We will add the following to the caption: “The fluxes are based on the differences between runs RP+ and RP- but normalized to represent an increase in Qf of 100 m3/s.”
10. I recommend moving the equations and the calculations estimating percentage changes in salinity due to variations in inflows, outflows, and related terms from the Summary section to the Results section. The final section should focus primarily on synthesizing and highlighting the main findings.
We prefer not to move these calculations and discussion. The main results of the study are the polynomial and the total exchange flow analysis results based on the 3D model calculations. The calculations in the discussion part of Section 4 are more speculative with a number of underlying assumptions, and are meant more as a suggested simplified interpretation of the Baltic Sea sensitivity to the investigated forcing factors than a firm result of the simulations.
11. The manuscript would be strengthened if the authors placed their results in better context with previous studies. In the Introduction, the authors state that their work differs from earlier studies because the model includes the North Sea domain. Could the authors elaborate on what difference this inclusion makes for the analysis and interpretation of salinity variability in the Baltic Sea? How do the results compare with those from previous studies that did not include the North Sea
We will extend the text comparing our results with earlier studies in Section 4. We will not focus too much on the importance of including the North Sea domain, but we will write in the introduction that this is important in order to study the influence of North Atlantic salinities on the Baltic Sea, which has been investigated to a very limited degree in the past.
Citation: https://doi.org/10.5194/egusphere-2025-4735-AC4
-
AC4: 'Reply on RC2', Lars Arneborg, 15 Dec 2025
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 237 | 85 | 37 | 359 | 21 | 16 |
- HTML: 237
- PDF: 85
- XML: 37
- Total: 359
- BibTeX: 21
- EndNote: 16
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The manuscript investigates the sensitivity of the salinity of the Baltic Sea to changes in the freshwater forcing and the salinity at the open boundaries in a model setup comprising the Baltic Sea and the North Sea. As mentioned by the authors, several studies have already been performed on the sensitivity to freshwater forcing. However, considering the salinity at the open boundary is a novel approach and also the idea of approximating the salinity of the Baltic Sea by a Taylor polynomial depending on the effects of the two forcings under consideration is new. Finally, the study adds novel insights to the water exchange between the North Sea and the Baltic Sea. All in all, the study meets the quality and scope of Ocean Science.
General comments
1. The study could provide a bit more background / context information, e.g., about typical variations of freshwater forcing and boundary salinities, about how they are expected to evolve in the future, whether there is any kind of interdependence etc… . Why did you choose specifically those two parameters? Aren’t other parameters like wind patterns / sea level rise more important for the salinity in the Baltic Sea than the boundary salinity in the North Sea (the importance of sea level rise is at least mentioned at the end of the text)? What exactly is the use of the Taylor polynomial? To explore the parameter space without having to run simulations for every combination of parameters? Are there alternative approaches and if so, why did you choose this particular approach? A final assessment of how this study advances the existing knowledge at the end of the manuscript would also be great.
2. There could be a few more references. Others have already looked at, for instance, the impact of runoff on inflows via changed sea level gradients; maybe, there are also more studies on the water exchange between the North Sea and the Baltic Sea or about the salinity at the North Sea boundary. See suggestions in the attached pdf.
3. The model validation could be a bit more comprehensive. For instance, a validation of transports in the entrance area of the Baltic Sea would be great as they are important for the study. The salinities in the central Baltic Sea look quite good (with some exceptions mentioned in the detailed comments) but they might be “right for the wrong reason”. In addition, possible inaccuracies introduced by the Taylor polynomial approach, namely by the short averaging period / nonstationarity of the time series in this period, the discrete differences, and the truncation of the polynomial, are only discussed very briefly (in Figure 7) and might deserve a few more sentences.
In case the suggested modifications lead to a too long manuscript, less important parts might be moved to a supplementary file.
Detailed comments can be found in the attached pdf.