the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 10 Apr 2025
Spatiotemporal properties of intrinsic sea level variability along the Southeast United States coastline
Abstract. The influence of intrinsic ocean variability on coastal sea level remains largely unexplored but is of potential importance for emerging forecasting efforts. As in weather forecasts, intrinsic variability will amplify uncertainty in initial conditions. However, variability originating from intrinsic processes may be predictable in a forecast system with sufficient resolution and accurate initialization. Here, we examine the spatiotemporal properties of intrinsic sea level variability along the Southeast United States coast using a suite of global ocean/sea-ice simulations at 0.1° horizontal resolution. In model simulations, intrinsic variability is a dominant component of the monthly de-seasonalized and detrended sea level variability in deep waters, but it is damped along continental shelves, where it comprises ~10–30 % of the sea level standard deviation. Our analyses demonstrate that US East Coast and Gulf of Mexico shelves exhibit a common intrinsic mode of sea level variability, with maximal amplitude in the South Atlantic Bight and almost no expression north of Cape Hatteras. This coastal mode is coherent with sea level along the Gulf Stream axis after detachment from Cape Hatteras. Intrinsic sea level variability in the detached Gulf Stream leads the coastal mode by 2–3 months, suggesting that intrinsic coastal sea level variability may exhibit predictability.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(2822 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(2822 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-1571', Marcello Passaro, 14 May 2025
I appreciated the opportunity to review this manuscript. The study presents an engaging analysis and contributes to the ongoing discussion on intrinsic sea level variability. The identification of a connection between along-shore variability and a region across the shelf-break is a noteworthy and potentially valuable insight. I have assigned minor review since further work is needed for the following points:
- The current structure places essential methodological details in the Appendix. I recommend integrating the Data and Methods section into the main body of the paper to enhance transparency and reproducibility. If this is not possible due to the “Letter” format, I request at least to answer to the points that I raise concerning the Appendix
- The discussion would be more impactful if it were more explicitly connected to existing findings in the literature.
Please note that the questions raised below are not intended merely for a response to the reviewer, but rather highlight areas where the manuscript lacks sufficient explanation. These should be addressed directly in the main text.
Both major and minor points are described in the line-by-line review below. Note that the comments about the Appendix are anticipated, since I considered that section as a “Data&Method”, which anticipate the results
Line 24:
Could you please clarify what type of predictions are being referred to here? Additionally, over what time scales are these predictions intended to apply?Lines 25–26:
The phrase “assessment of the capabilities” is somewhat vague in scientific terms. Are you referring to a validation process? Furthermore, when you mention “ocean forecasting,” could you specify the time scales involved?Line 52:
Earlier, you defined “intrinsic” variability as being generated by the ocean itself, independent of atmospheric forcing. In this context, what exactly do you mean by “sources” of intrinsic sea level variability?------
“APPENDICES”
Why are the Data and Methods sections placed in an appendix?
Given their importance for evaluating the results, I recommend restructuring the manuscript to integrate these sections into the main body of the text.Appendix A appears to be written for a highly specialized audience, which may limit accessibility. It would be helpful to include a brief explanation of the HR FOSI and HR RYF simulations. Specifically:
What processes do these simulations represent?
What type of ocean models are they based on?
Why were these particular simulations selected over others?
Additionally, please clarify the rationale for applying 32 years of HR RYF data, forced with boundary conditions from a single year (2003–2004). What is the scientific justification for this approach?
Is monthly resolution sufficient to capture the temporal scales of intrinsic sea level variability relevant for forecasting? This question ties back to the earlier point regarding the nature of the forecasts being discussed. Why was the analysis limited to monthly resolution, especially considering the availability of high-frequency observational data such as tide gauge records (e.g., GESLA), daily altimetry grids (noting their limitations in effective temporal resolution)
Why was the Measure dataset selected for gridded sea level data? Measure uses a maximum of two altimeters. How does its spatial resolution compare to more comprehensive products, such as the one provided by Copernicus (SEALEVEL_GLO_PHY_L4_MY_008_047), which incorporates all available altimeters?
Line 219: What is a ball tree algorithm?
A brief explanation would be helpful for readers unfamiliar with this method.------
Figure 1:
The tide gauge (TG) points are currently not visible. Could you clarify why only three tide gauges are shown in this region, which is among the most extensively monitored globally?
Additionally, I suggest adding a panel that shows the difference between panels a and b. For clarity in the subsequent discussion, it would also be helpful to indicate the blue area from Figure 3 within Figure 1.Line 77:
The statement “the model generally underestimating...” may be influenced by the temporal resolution of the data. This result could differ significantly if daily rather than monthly data were used. It is worth noting that satellite observations typically underestimate high-frequency variance in shelf circulation.Line 88:
Please clarify that the results discussed here pertain specifically to a region dominated by a western boundary current. This is not a general characteristic of the deep ocean, as illustrated globally in Close et al. (2020), Figure 1c.Section 2.3:
The propagation described in this section is intriguing and might be better resolved using higher temporal resolution data. The cited study by Close et al. (2020) used “successive 5-day averages.” Could you explain why higher temporal resolution was not used in your analysis, and why it was feasible in other studies?Line 158:
The finding regarding the along-coast coherence of PC1 is particularly interesting. I recommend expanding the discussion in light of existing observational studies. For example, Oelsmann et al. (2024) (https://doi.org/10.1029/2024jc021120 ) analyzed the along-shore coherence of monthly sea level variability using tide gauges and coastal altimetry. Their Figure 7 shows that, along the U.S. East Coast (a western boundary), the observed clusters do not strongly correlate with interannual variability from typical climate indices, unlike what is observed along eastern boundaries. Your results, consistent with previous work, suggest that these clusters are linked to intrinsic dynamics. Notably, both your model and the observations show a separation at Cape Hatteras.Line 158 (continued):
Please specify that the “robust 2–3-month lag” refers to the offshore region highlighted in Figure 3.Lines 160–166:
These statements are somewhat unclear. Figure 3c shows that the lag-correlation of PC1 peaks at around 3 months along the entire shelf. How does this reconcile with the claim that sea level anomalies (SLAs) travel much faster, at sub-monthly scales, once they pass Cape Hatteras, which are “unseen” in your experiments?Best regards,
Marcello Passaro
Citation: https://doi.org/10.5194/egusphere-2025-1571-RC1 -
AC2: 'Reply on RC1', Carmine Donatelli, 13 Aug 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-1571/egusphere-2025-1571-AC2-supplement.pdf
-
RC2: 'Comment on egusphere-2025-1571', Anonymous Referee #2, 09 Jun 2025
This paper presents a novel and valuable contribution to our understanding of intrinsic sea level variability along the U.S. East Coast. The authors use high-resolution ocean model simulations to demonstrate for the first time that 10–30% of the shelf sea level standard deviation may originate from ocean-internal processes alone. They also examine lead-lag relationships to assess predictability and explore possible mechanisms underlying a major mode of variability that is dominant south of Cape Hatteras. Overall, this study represents an important step toward identifying the full range of processes contributing to monthly-to-interannual sea level variability along the Northeast U.S. coastline.
The overall presentation of the results, methods, data, and conclusions—along with the manuscript's structure, language, and conciseness—is very strong and concise. I have only a few questions that should potentially be addressed. Note that most of my questions aim to clarify which processes could be relevant in the setup where only intrinsic variability is allowed, compared to the setup that includes external variability. Some of these comments may be somewhat subjective, as I occasionally found it difficult to determine which previous studies should be referenced here, or not.
Major comments:
Previous studies on mechanisms linking open-ocean and coastal sea level changes:
There is a substantial body of literature investigating the drivers of sea level variability along the U.S. East Coast (e.g., Frederikse et al., 2017; Calafat et al., 2018; Piecuch et al., 2018; Dangendorf et al., 2021, 2023; Steinberg et al., 2024; Wang et al., 2024). However, these studies are not cited or discussed in the introduction. I think this is likely because some of the previously discussed mechanisms are not relevant (with regards to intrinsic variability), or because these works primarily analyze historical observations or model simulations that include externally forced variability, rather than isolating intrinsic (unforced) variability, as is done in this study.That said, I wonder whether some of the physical mechanisms proposed in those earlier studies—such as the link between offshore steric height anomalies and bottom pressure fluctuations on the shelf in modulating coastal sea level (including the communication along the coast via coastally trapped waves)—might also be relevant in the context of intrinsic variability. If such mechanisms are indeed applicable regardless of the forcing source, then it would be appropriate to acknowledge relevant studies and situate your findings within the broader framework of previously proposed mechanisms.
In the conclusions, you note: “the ocean mechanisms involved in the communication of off-shelf anomalies to the coast (and from Cape Hatteras to the Gulf of Mexico) are likely the same as those regulating the transfer of other sea-level signals, regardless of their forced or intrinsic nature.” This statement suggests that some mechanisms identified in earlier studies (e.g., Calafat et al., 2018; Dangendorf et al., 2021; 2023; Steinberg et al., 2024) could in fact be relevant to your results, particularly given your finding of significant coherence between the leading coastal principal component (PC1) and off-shelf sea level, especially at periods longer than one year. Overall, it may be helpful for the reader to explain (maybe in introduction), which of the previously discussed forcings (e.g., wind forcing) are not relevant for your study focusing on intrinsic variability (just to provide a little bit more context).
I think that, if some of the previously investigated mechanisms are relevant, the authors should try to emphasize a bit better why and how their findings, e.g., the detection of the dominating and coherent first mode south of Cape Hatteras, is different or similar to previous findings, or novel, considering the study on coherent sea level in that region by Calafat et al., 2018. That ‘This coastal mode is coherent with sea level along the Gulf Stream axis after detachment from Cape Hatteras’ is a central point of this paper (in the abstract), and therefore it’s novelty needs to better clarified in light of previous research. You address several of these points in the discussion section (citing Wise et al., 2018; Hughes et al., 2019; Wise et al., 2020), but it may strengthen the manuscript to introduce some of these mechanisms earlier, in the introduction.
Another key finding of the study is that “Intrinsic sea level variability in the detached Gulf Stream leads the coastal mode by 2–3 months, suggesting that intrinsic coastal sea level variability may exhibit predictability.” I agree that this is an intriguing result from a mechanistic standpoint. However, I wonder how relevant this is in practical terms—specifically, what implications it holds for the predictability of sea level in the real world.
If I understand the results correctly, intrinsic variability accounts for roughly 20–30% of the total standard deviation in sea level (based on monthly data), relative to the forced simulations. Assuming a maximum coherence of ~0.6 between offshore sea level and the leading coastal mode (for periods >1 year), this corresponds to roughly 36% of the variance in the coastal mode being explained by offshore signals. Even if we assume that the first mode captures all of the intrinsic sea level variability, the offshore signal would then explain about 7–11% of the total sea level variability (i.e., 0.36 × 0.2–0.3) in a realistically forced setup.
If this reasoning is approximately correct, then a key question arises: To what extent does this lead–lag relationship actually translate into meaningful predictability of sea level in the real world, where forced variability dominates? Clarifying this point would strengthen the practical interpretation of the findings and help position the study within the broader context of coastal sea level forecasting.
Minor comments (some points are related to the major comments):
Abstract L14-17: I think it would be helpful to more clearly articulate what distinguishes your study from existing research. The separation of sea level correlation patterns around Cape Hatteras has already been documented in observational studies. Therefore, I would suggest emphasizing what makes your work novel. Is this the first study to shows that this separation also exists in intrinsic SSH variability, or is this the first study that shows this phenomenon using climate or ocean models? And/or is your key contribution the finding that this separation may be predictable? Clarifying this would strengthen the positioning of your study.
L52-55: Related to the 1. major comment: Are the results of previous studies investigating the causes of sea-level variability along the U.S. East Coast (e.g., Frederikse et al., 2017; Calafat et al., 2018; Piecuch et al., 2018; Dangendorf et al., 2021; 2023; Steinberg et al., 2024; Wang et al., 2024) relevant here, or not?
L63: Minor comment: Shouldn’t the methods not be in the main text in this journal?
Methods/Appendix:
L194: Please explain what ‘four consecutive cycles’ means here already.
L202: ‘The HR RYF simulation was carried out by applying a single year of JRA55 boundary conditions (May 2003 to the end of April 2004).‘ What if that year has a substantially different atmospheric variability (or annual cycle amplitude) than the average of 1958-2018? Wouldn’t that not bias your estimates? Would it make sense to use the 1958-2018 average as a forcing?
Fig1. b) It may be helpful to improve visibility of TGs (add white marker edge color?). It would also be interesting to see some more available TGs in this region (showing STD). C-e) What are the numbers at the bottom in the time series plots? g) as you write in the text, this fraction can be as large as 1 (100%). Could it actually be larger than visible by the color scale (that is cut off at 1, or 100%)?
Fig1. What about having a look at the differences in the power spectrum between the two experiments. At what frequencies is the variance on the shelf most strongly reduced?
L75: I assume that ’Total sea level standard deviation (mean across the FOSI members; Fig. 1a)’ is the temporal STD of the ensemble mean, correct?
L81: It may be helpful to quantify this agreement (correlation, rms), even if that’s from Little et al., 2024.
L89-90. I think it would be helpful to include some interpretation of the patterns observed in the intrinsic variability from an ocean dynamics perspective. For example, what causes intrinsic variability to be smaller or larger in certain regions?
L96: It may be nice for the reader if you’d better explain why you apply the EOF analysis overall.
L105: ‘Specifically, the PCs exhibit fluctuations at sub-annual to interannual time scales, including significant multi-year sea level trends (for example PC1, over a 5-year period beginning around month 90) (Penduff et al., 2019).‘ Is the significance based on visual inspection? A spectral analysis incl. significance test could be useful here.
Section 2.2.: The intrinsic variability north of Cape Hatteras appears to be much weaker compared to south of it, causing the EOF’s to mostly pickup regions in the south. It’s probably difficult to answer, but do you have any idea why the intrinsic variability North of Cape Hatteras is weaker (from an ocean dynamics perspective)?
L120. What is the 0.3 threshold based on? Did you take into account autocorrelation in the time series?
L123-125: Why do you compare these correlation pattern with the minimum offshore intrinsic fraction?
2.3. Considering previous work, what about having a look at offshore steric and coastal SL correlations (maybe also at different frequencies). Or what about adding the steric averages in Fig. 4. computed over the same offshore-region?
Fig 3. If the geographical extent shown in the figure were expanded, would similar lagged correlations emerge in other regions—for example, in the Caribbean? This could be worth exploring, particularly in light of previous findings (e.g., Dangendorf et al., 2024, Supplementary Fig. 6; Calafat et al., 2018).
Discussion
L157-167 and 180 -189: You also mention that the processes communicating signals to the coast are “likely the same as those regulating the transfer of other sea-level signals, regardless of their forced or intrinsic nature.” It would be useful to clearly rule out previously discussed processes (e.g., Dangendorf et al., 2021, Calafat et al., 2018) that are not possible within the framework of a setup that includes only intrinsic variability.
Citation: https://doi.org/10.5194/egusphere-2025-1571-RC2 -
AC1: 'Reply on RC2', Carmine Donatelli, 13 Aug 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-1571/egusphere-2025-1571-AC1-supplement.pdf
-
AC1: 'Reply on RC2', Carmine Donatelli, 13 Aug 2025
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-1571', Marcello Passaro, 14 May 2025
I appreciated the opportunity to review this manuscript. The study presents an engaging analysis and contributes to the ongoing discussion on intrinsic sea level variability. The identification of a connection between along-shore variability and a region across the shelf-break is a noteworthy and potentially valuable insight. I have assigned minor review since further work is needed for the following points:
- The current structure places essential methodological details in the Appendix. I recommend integrating the Data and Methods section into the main body of the paper to enhance transparency and reproducibility. If this is not possible due to the “Letter” format, I request at least to answer to the points that I raise concerning the Appendix
- The discussion would be more impactful if it were more explicitly connected to existing findings in the literature.
Please note that the questions raised below are not intended merely for a response to the reviewer, but rather highlight areas where the manuscript lacks sufficient explanation. These should be addressed directly in the main text.
Both major and minor points are described in the line-by-line review below. Note that the comments about the Appendix are anticipated, since I considered that section as a “Data&Method”, which anticipate the results
Line 24:
Could you please clarify what type of predictions are being referred to here? Additionally, over what time scales are these predictions intended to apply?Lines 25–26:
The phrase “assessment of the capabilities” is somewhat vague in scientific terms. Are you referring to a validation process? Furthermore, when you mention “ocean forecasting,” could you specify the time scales involved?Line 52:
Earlier, you defined “intrinsic” variability as being generated by the ocean itself, independent of atmospheric forcing. In this context, what exactly do you mean by “sources” of intrinsic sea level variability?------
“APPENDICES”
Why are the Data and Methods sections placed in an appendix?
Given their importance for evaluating the results, I recommend restructuring the manuscript to integrate these sections into the main body of the text.Appendix A appears to be written for a highly specialized audience, which may limit accessibility. It would be helpful to include a brief explanation of the HR FOSI and HR RYF simulations. Specifically:
What processes do these simulations represent?
What type of ocean models are they based on?
Why were these particular simulations selected over others?
Additionally, please clarify the rationale for applying 32 years of HR RYF data, forced with boundary conditions from a single year (2003–2004). What is the scientific justification for this approach?
Is monthly resolution sufficient to capture the temporal scales of intrinsic sea level variability relevant for forecasting? This question ties back to the earlier point regarding the nature of the forecasts being discussed. Why was the analysis limited to monthly resolution, especially considering the availability of high-frequency observational data such as tide gauge records (e.g., GESLA), daily altimetry grids (noting their limitations in effective temporal resolution)
Why was the Measure dataset selected for gridded sea level data? Measure uses a maximum of two altimeters. How does its spatial resolution compare to more comprehensive products, such as the one provided by Copernicus (SEALEVEL_GLO_PHY_L4_MY_008_047), which incorporates all available altimeters?
Line 219: What is a ball tree algorithm?
A brief explanation would be helpful for readers unfamiliar with this method.------
Figure 1:
The tide gauge (TG) points are currently not visible. Could you clarify why only three tide gauges are shown in this region, which is among the most extensively monitored globally?
Additionally, I suggest adding a panel that shows the difference between panels a and b. For clarity in the subsequent discussion, it would also be helpful to indicate the blue area from Figure 3 within Figure 1.Line 77:
The statement “the model generally underestimating...” may be influenced by the temporal resolution of the data. This result could differ significantly if daily rather than monthly data were used. It is worth noting that satellite observations typically underestimate high-frequency variance in shelf circulation.Line 88:
Please clarify that the results discussed here pertain specifically to a region dominated by a western boundary current. This is not a general characteristic of the deep ocean, as illustrated globally in Close et al. (2020), Figure 1c.Section 2.3:
The propagation described in this section is intriguing and might be better resolved using higher temporal resolution data. The cited study by Close et al. (2020) used “successive 5-day averages.” Could you explain why higher temporal resolution was not used in your analysis, and why it was feasible in other studies?Line 158:
The finding regarding the along-coast coherence of PC1 is particularly interesting. I recommend expanding the discussion in light of existing observational studies. For example, Oelsmann et al. (2024) (https://doi.org/10.1029/2024jc021120 ) analyzed the along-shore coherence of monthly sea level variability using tide gauges and coastal altimetry. Their Figure 7 shows that, along the U.S. East Coast (a western boundary), the observed clusters do not strongly correlate with interannual variability from typical climate indices, unlike what is observed along eastern boundaries. Your results, consistent with previous work, suggest that these clusters are linked to intrinsic dynamics. Notably, both your model and the observations show a separation at Cape Hatteras.Line 158 (continued):
Please specify that the “robust 2–3-month lag” refers to the offshore region highlighted in Figure 3.Lines 160–166:
These statements are somewhat unclear. Figure 3c shows that the lag-correlation of PC1 peaks at around 3 months along the entire shelf. How does this reconcile with the claim that sea level anomalies (SLAs) travel much faster, at sub-monthly scales, once they pass Cape Hatteras, which are “unseen” in your experiments?Best regards,
Marcello Passaro
Citation: https://doi.org/10.5194/egusphere-2025-1571-RC1 -
AC2: 'Reply on RC1', Carmine Donatelli, 13 Aug 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-1571/egusphere-2025-1571-AC2-supplement.pdf
-
RC2: 'Comment on egusphere-2025-1571', Anonymous Referee #2, 09 Jun 2025
This paper presents a novel and valuable contribution to our understanding of intrinsic sea level variability along the U.S. East Coast. The authors use high-resolution ocean model simulations to demonstrate for the first time that 10–30% of the shelf sea level standard deviation may originate from ocean-internal processes alone. They also examine lead-lag relationships to assess predictability and explore possible mechanisms underlying a major mode of variability that is dominant south of Cape Hatteras. Overall, this study represents an important step toward identifying the full range of processes contributing to monthly-to-interannual sea level variability along the Northeast U.S. coastline.
The overall presentation of the results, methods, data, and conclusions—along with the manuscript's structure, language, and conciseness—is very strong and concise. I have only a few questions that should potentially be addressed. Note that most of my questions aim to clarify which processes could be relevant in the setup where only intrinsic variability is allowed, compared to the setup that includes external variability. Some of these comments may be somewhat subjective, as I occasionally found it difficult to determine which previous studies should be referenced here, or not.
Major comments:
Previous studies on mechanisms linking open-ocean and coastal sea level changes:
There is a substantial body of literature investigating the drivers of sea level variability along the U.S. East Coast (e.g., Frederikse et al., 2017; Calafat et al., 2018; Piecuch et al., 2018; Dangendorf et al., 2021, 2023; Steinberg et al., 2024; Wang et al., 2024). However, these studies are not cited or discussed in the introduction. I think this is likely because some of the previously discussed mechanisms are not relevant (with regards to intrinsic variability), or because these works primarily analyze historical observations or model simulations that include externally forced variability, rather than isolating intrinsic (unforced) variability, as is done in this study.That said, I wonder whether some of the physical mechanisms proposed in those earlier studies—such as the link between offshore steric height anomalies and bottom pressure fluctuations on the shelf in modulating coastal sea level (including the communication along the coast via coastally trapped waves)—might also be relevant in the context of intrinsic variability. If such mechanisms are indeed applicable regardless of the forcing source, then it would be appropriate to acknowledge relevant studies and situate your findings within the broader framework of previously proposed mechanisms.
In the conclusions, you note: “the ocean mechanisms involved in the communication of off-shelf anomalies to the coast (and from Cape Hatteras to the Gulf of Mexico) are likely the same as those regulating the transfer of other sea-level signals, regardless of their forced or intrinsic nature.” This statement suggests that some mechanisms identified in earlier studies (e.g., Calafat et al., 2018; Dangendorf et al., 2021; 2023; Steinberg et al., 2024) could in fact be relevant to your results, particularly given your finding of significant coherence between the leading coastal principal component (PC1) and off-shelf sea level, especially at periods longer than one year. Overall, it may be helpful for the reader to explain (maybe in introduction), which of the previously discussed forcings (e.g., wind forcing) are not relevant for your study focusing on intrinsic variability (just to provide a little bit more context).
I think that, if some of the previously investigated mechanisms are relevant, the authors should try to emphasize a bit better why and how their findings, e.g., the detection of the dominating and coherent first mode south of Cape Hatteras, is different or similar to previous findings, or novel, considering the study on coherent sea level in that region by Calafat et al., 2018. That ‘This coastal mode is coherent with sea level along the Gulf Stream axis after detachment from Cape Hatteras’ is a central point of this paper (in the abstract), and therefore it’s novelty needs to better clarified in light of previous research. You address several of these points in the discussion section (citing Wise et al., 2018; Hughes et al., 2019; Wise et al., 2020), but it may strengthen the manuscript to introduce some of these mechanisms earlier, in the introduction.
Another key finding of the study is that “Intrinsic sea level variability in the detached Gulf Stream leads the coastal mode by 2–3 months, suggesting that intrinsic coastal sea level variability may exhibit predictability.” I agree that this is an intriguing result from a mechanistic standpoint. However, I wonder how relevant this is in practical terms—specifically, what implications it holds for the predictability of sea level in the real world.
If I understand the results correctly, intrinsic variability accounts for roughly 20–30% of the total standard deviation in sea level (based on monthly data), relative to the forced simulations. Assuming a maximum coherence of ~0.6 between offshore sea level and the leading coastal mode (for periods >1 year), this corresponds to roughly 36% of the variance in the coastal mode being explained by offshore signals. Even if we assume that the first mode captures all of the intrinsic sea level variability, the offshore signal would then explain about 7–11% of the total sea level variability (i.e., 0.36 × 0.2–0.3) in a realistically forced setup.
If this reasoning is approximately correct, then a key question arises: To what extent does this lead–lag relationship actually translate into meaningful predictability of sea level in the real world, where forced variability dominates? Clarifying this point would strengthen the practical interpretation of the findings and help position the study within the broader context of coastal sea level forecasting.
Minor comments (some points are related to the major comments):
Abstract L14-17: I think it would be helpful to more clearly articulate what distinguishes your study from existing research. The separation of sea level correlation patterns around Cape Hatteras has already been documented in observational studies. Therefore, I would suggest emphasizing what makes your work novel. Is this the first study to shows that this separation also exists in intrinsic SSH variability, or is this the first study that shows this phenomenon using climate or ocean models? And/or is your key contribution the finding that this separation may be predictable? Clarifying this would strengthen the positioning of your study.
L52-55: Related to the 1. major comment: Are the results of previous studies investigating the causes of sea-level variability along the U.S. East Coast (e.g., Frederikse et al., 2017; Calafat et al., 2018; Piecuch et al., 2018; Dangendorf et al., 2021; 2023; Steinberg et al., 2024; Wang et al., 2024) relevant here, or not?
L63: Minor comment: Shouldn’t the methods not be in the main text in this journal?
Methods/Appendix:
L194: Please explain what ‘four consecutive cycles’ means here already.
L202: ‘The HR RYF simulation was carried out by applying a single year of JRA55 boundary conditions (May 2003 to the end of April 2004).‘ What if that year has a substantially different atmospheric variability (or annual cycle amplitude) than the average of 1958-2018? Wouldn’t that not bias your estimates? Would it make sense to use the 1958-2018 average as a forcing?
Fig1. b) It may be helpful to improve visibility of TGs (add white marker edge color?). It would also be interesting to see some more available TGs in this region (showing STD). C-e) What are the numbers at the bottom in the time series plots? g) as you write in the text, this fraction can be as large as 1 (100%). Could it actually be larger than visible by the color scale (that is cut off at 1, or 100%)?
Fig1. What about having a look at the differences in the power spectrum between the two experiments. At what frequencies is the variance on the shelf most strongly reduced?
L75: I assume that ’Total sea level standard deviation (mean across the FOSI members; Fig. 1a)’ is the temporal STD of the ensemble mean, correct?
L81: It may be helpful to quantify this agreement (correlation, rms), even if that’s from Little et al., 2024.
L89-90. I think it would be helpful to include some interpretation of the patterns observed in the intrinsic variability from an ocean dynamics perspective. For example, what causes intrinsic variability to be smaller or larger in certain regions?
L96: It may be nice for the reader if you’d better explain why you apply the EOF analysis overall.
L105: ‘Specifically, the PCs exhibit fluctuations at sub-annual to interannual time scales, including significant multi-year sea level trends (for example PC1, over a 5-year period beginning around month 90) (Penduff et al., 2019).‘ Is the significance based on visual inspection? A spectral analysis incl. significance test could be useful here.
Section 2.2.: The intrinsic variability north of Cape Hatteras appears to be much weaker compared to south of it, causing the EOF’s to mostly pickup regions in the south. It’s probably difficult to answer, but do you have any idea why the intrinsic variability North of Cape Hatteras is weaker (from an ocean dynamics perspective)?
L120. What is the 0.3 threshold based on? Did you take into account autocorrelation in the time series?
L123-125: Why do you compare these correlation pattern with the minimum offshore intrinsic fraction?
2.3. Considering previous work, what about having a look at offshore steric and coastal SL correlations (maybe also at different frequencies). Or what about adding the steric averages in Fig. 4. computed over the same offshore-region?
Fig 3. If the geographical extent shown in the figure were expanded, would similar lagged correlations emerge in other regions—for example, in the Caribbean? This could be worth exploring, particularly in light of previous findings (e.g., Dangendorf et al., 2024, Supplementary Fig. 6; Calafat et al., 2018).
Discussion
L157-167 and 180 -189: You also mention that the processes communicating signals to the coast are “likely the same as those regulating the transfer of other sea-level signals, regardless of their forced or intrinsic nature.” It would be useful to clearly rule out previously discussed processes (e.g., Dangendorf et al., 2021, Calafat et al., 2018) that are not possible within the framework of a setup that includes only intrinsic variability.
Citation: https://doi.org/10.5194/egusphere-2025-1571-RC2 -
AC1: 'Reply on RC2', Carmine Donatelli, 13 Aug 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-1571/egusphere-2025-1571-AC1-supplement.pdf
-
AC1: 'Reply on RC2', Carmine Donatelli, 13 Aug 2025
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 653 | 57 | 19 | 729 | 15 | 30 |
- HTML: 653
- PDF: 57
- XML: 19
- Total: 729
- BibTeX: 15
- EndNote: 30
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Carmine Donatelli
Christopher M. Little
Rui M. Ponte
Stephen G. Yeager
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(2822 KB) - Metadata XML