the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 11 May 2023
Next generation atmosphere-ocean climate modelling for storm surge hazard projections
Abstract. Due to on-going global warming, extreme storm surges are expected to threaten a greater number of coastal communities worldwide. However, global and regional climate simulations of extreme events are still not accurate enough to respond to the growing needs of the local decision makers to prepare for these rising hazards. We present a new approach using (sub-)kilometre-scale coupled atmosphere-ocean-wave models and demonstrate the feasibility to provide meter-scale assessments of the impact of climate change on storm surge hazards. As a proof of concept, we focus in the Adriatic Sea and analyse the sea levels of two kilometre-scale 31-year long simulations used in evaluation and extreme warming modes. First, we demonstrate that, at 1-km resolution, the model errors are reduced by up to a third compare to state-of-the-art regional and global models. Second, we show that meter-scale storm surge results – obtained by further downscaling extreme events extracted from the kilometre-scale simulations – contrast with the previously published literature. In particular, we found that some understudied regions of the Adriatic coast might be more vulnerable to sea level rise and atmospherically driven storm surges induced by extreme climate warming than the well-researched Venice Lagoon. Following these preliminary results, we present a newly developed methodology directly downscaling extreme events from global climate models. Within this framework, the numerical resources, previously spent to produce long-term simulations, are used efficiently to quantify the climate change uncertainty and to properly assess the meter-scale storm surge hazards.
- Preprint
(2821 KB) - Metadata XML
- BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on egusphere-2023-913', Anonymous Referee #1, 21 Jul 2023
The topic of the manuscript and the modelling effort put on this research are of great relevance. However, I recommend against publication of the manuscript in its current state. I appreciate the effort put by the two authors, and the potential of their analysis, so I would like to challenge authors to resubmit a new version including a specific research aim, sound scientific description of the hypothesis, experimental design, and findings, and how these represent a substantial contribution to understanding future surge hazards in the Adriatic Sea or for modelling of extreme sea levels.
Please find more detailed comments in the attachment, i wish authors understand them to be constructive.
-
AC1: 'Reply on RC1', Clea Denamiel, 03 Oct 2023
The authors truly appreciate the careful review and detailed comments provided by the reviewer and would like to thank them for their time and consideration.
Hereafter the comments of the reviewer are highlighted in italic while the responses of the authors are preceded with "R:".
The topic of the manuscript and the modelling effort put in this research are of great relevance. However, I recommend against publication of the manuscript in its current state. I appreciate the effort put by the two authors, and the potential of their analysis, so I would like to challenge them to resubmit a new version including a specific research aim, sound scientific description of the hypothesis, experimental design, and findings, and how these represent a substantial contribution to understanding future surge hazards in the Adriatic Sea or for modelling of extreme sea levels. I admit that with such large and detailed analysis it can get very complicated, but the current presentation quality is rather low. Additionally, although the scientific significance and quality of this manuscript for the understanding of future surge hazards in the Adriatic sea has the potential to be high, in the current stage is uncertain due to missing details or discussion. Although I wrote several major and minor comments on the current manuscript, I would like to first share some general insights below to support authors with the resubmission of their manuscript. Please understand them to be constructive.
R: The authors acknowledge their failure to provide and present a general modelling framework for meter-scale storm-surge projections. Consequently, they will reframe the aim of their study to the specific studied area: the Adriatic basin.
- The introduction is wide and general. There is little focus on the case study. I also miss insights on what the current scientific gaps are and what are those taken in this manuscript, what is new? is that increase on resolution? The extreme events simulations? why is the PGW used over other methods?
- There is little reflection on what others have done regarding waves and surge projections in the introduction.
R: The introduction will be rewritten in order to highlight several aspects missing in the present version. The overly extensive historical literature review that evidences the research and knowledge gaps will be shorten and a transition towards the motivations to target the Adriatic basin will be added. The objective of the article will be reframed to demonstrate the added value of the modelling framework for the projection of storm-surges in the complex Adriatic environment (e.g., topography and bathymetry, basin wide seiches, etc.) compared to more classical, simple approaches using lower resolutions and less coupling. This will include a thorough review of the available storm surge research in the Adriatic Sea. Finally, the justifications of using the presented baroclinic framework instead of simpler barotropic models (as used in the final step of the nesting chain) will be added to the article. This will also include a clear justification for the usage of the PGW methodology.
- The manuscript is difficult to read, in my opinion using too much decoration, I advise to adopt a more scientific approach on statements. In fact, the positioning is not clear, is a methodological finding or a case study presented? For example, Line 62- Finally, the local application of the (sub-)kilometre-scale methodology to the Adriatic Sea is only used as a proof of concept and all the presented approaches and results can be replicated and/or adapted at any location in the world where extreme sea level hazard assessments fully depend on the accurate representation of the complex geomorphology of the coastal areas I advise in such a case to present the experiment that validates the hypothesis that upscaling of the methodology is possible to any location in the world. More precisely, I would encourage the authors to be cautious with limitations, once their methodology might struggle to simulate future meteo-tsunamis, that are of great relevance for extreme sea level hazard assessment in the Adriatic sea.
R: As stated above the reframed aim of the study will be to demonstrate the added value of the modelling framework for the projection of storm-surges in the complex Adriatic environment (e.g., topography and bathymetry, basin wide seiches, etc.) compared to more classical, simple approaches using lower resolutions and less coupling.
As for modelling the frequency and intensity of past and future meteotsunami events, the authors recently published a methodology for this specific purpose (Denamiel et al., 2023). This is indeed another important contribution to extreme sea levels in the Adriatic Sea but it would require either to use sub-kilometer-scale atmospheric model at the climate scale (unrealistic at this stage due to the numerical cost) or the use of the methodology presented in Denamiel et al. (2023) and at the end of this article. To this date, such an approach is still under development (i.e., synoptic index performances should be increased, method should be tested in different places, etc.) and, as suggested by the other reviewer, is not necessarily relevant to this particular article.
Denamiel C, Belušić D, Zemunik P and Vilibić I. Climate projections of meteotsunami hazards. Front. Mar. Sci., 10,1167863 (2023). https://doi.org/10.3389/fmars.2023.1167863
- The manuscript requires that methods are reproducible, important elements are missing in the methodology. For example, several models such as ROMS, ADCIRC, and SWAN are used, where parameterizations, spatial, and temporal discretization are really impacting results. Details on how the pseudo-global warming method is applied are neither provided.
R: The modelling framework of the AdriSC modelling suite has been presented in many other articles and the authors thought that citing this extensive literature would be enough. However, as both reviewers specifically ask for more details on the modelling suite, they will be added to the new version of the article.
- Regarding the above, if the manuscript aims at developing a new modelling approach for investigating future extreme surges, I encourage to present the experimental design elaborating on sensitivities or uncertainties arising from model choices.
- Then, how do sensitivities from modelling choice compare with the changes seen in the future scenario simulated? If authors have addressed this in other research publications it is worth to incorporate explicitly in this manuscript.
R: Sensitivities and uncertainties introduced by the choice of the models used to downscale the regional (or global) climate model results can be important. However, in this article, we have demonstrated that the kilometer-scale modelling suite is performing well during the 1987-2017 period (comparison of the extreme sea level distributions at different tide gauges). The authors thus believe that representing the climate uncertainty (i.e., having different scenarios of climate warming from different global climate models) is much more critical than representing the uncertainty linked to the choice of the models used for the downscaling. This is something that cannot be done with the presented methodology due to its numerical cost. This will be discussed in the “discussion and conclusions” section.
- The manuscript structure is ok regarding sections used; however I found significant information out of their corresponding section; a few examples are: Line 134-138 In methods but it is part of the results section. Line 138-143 In methods but it is part of the discussion section. Line 188-196 In methods but it is part of discussion In general, in the results section new data is described.
R: The authors will make sure to move the results from the method section.
- In my opinion, there are choices that are inconsistent, and that are not discussed appropriately. For example: If interaction between mean sea level rise and other sea level components are of interest in this research, why the regional model does not account for it, but local ones do? I consider that superimposing mean sea level rise in the boundary conditions of the local models could lead into numerical artifacts in the local results, to accommodate for inconsistent boundary conditions, I would say that, if it is one of the research objectives of this manuscript, at least understanding the effect of this superposition deserves investigation.
R: The authors are not responsible for the choices made by the Med-CORDEX community (regional climate simulations of the Mediterranean Sea) that did not include sea level rise in their simulations and decided to only add the tides (as sea-level and barotropic currents) in their baroclinic simulations to not create too much imbalances between 3D current forcing and sea-levels (continuity equation) and to not count twice the effects of the thermal expansion (included in the sea level rise datasets provided by the IPCC-AR5 2015 ensemble). Further, adding sea level rise to the initial/boundary conditions of models like ADCIRC is a common practice in the extreme sea level community using barotropic models which is not known to create any instability as it just raise the initial/boundary sea levels of the domain and, in contrast with baroclinic models, the balance between sea-levels and depth-averaged currents is not critical as depth averaged currents are generally not well represented by barotropic models (i.e., often not even analyzed). Finally, during the development and implementation of the AdriSC modelling suite, the choice of the locations of the boundaries of the different atmosphere-ocean models has been made with extreme care in order to be able to represent all important processes driving the Adriatic Sea circulation. For example, to this date, the AdriSC climate model is the only model capable to represent the impact of the North Ionian Gyre (or BioS) within the Adriatic Sea (Denamiel et al., 2022). The authors thus believe that, despite all the modelling errors (see evaluation of the AdriSC climate model; Denamiel et al., 2021 and Pranić et al., 2021), the AdriSC climate modelling suite has already been proven to perform well within the Adriatic basin.
Denamiel, C., Pranić, P., Ivanković, D., Tojčić, I., and Vilibić, I. Performance of the Adriatic Sea and Coast (AdriSC) climate component—a COAWST V3.3-based coupled atmosphere–ocean modelling suite: atmospheric dataset, Geosci. Model Dev., 14, 3995–4017 (2021). https://doi.org/10.5194/gmd-14-3995-2021
Denamiel, C., Tojčić, I., Pranić, P. et al. Modes of the BiOS-driven Adriatic Sea thermohaline variability. Clim Dyn 59, 1097–1113 (2022). https://doi.org/10.1007/s00382-022-06178-4
Pranić, P., Denamiel, C., and Vilibić, I. Performance of the Adriatic Sea and Coast (AdriSC) climate component—a COAWST V3.3-based one-way coupled atmosphere–ocean modelling suite: ocean results, Geosci. Model Dev., 14, 5927–5955 (2021). https://doi.org/10.5194/gmd-14-5927-2021
- I find figure 1 too general, a proper schematic of the manuscript workflow would help very much on understanding the methodological approach and how it benefits the assessment of future extreme surge.
R: Figure 1 will be redrawn to better present the needs and added value of the sub-kilometer-scale approach in storm surge projections.
- In general, consider the use of appendixes when needed.
R: As the article will be fully restructured in order to prove the added value of the presented approach, material will definitely be moved to either appendixes or supplementary material.
Citation: https://doi.org/10.5194/egusphere-2023-913-AC1
-
AC1: 'Reply on RC1', Clea Denamiel, 03 Oct 2023
-
RC2: 'Comment on egusphere-2023-913', Anonymous Referee #2, 25 Aug 2023
General comments
The study proposes a novel, high-resolution ocean-atmosphere coupled modelling framework to increase the accuracy and therefore local relevance of storm-surge projections under climate change. Their modelling framework benefits from kilometer-scale atmospheric and ocean resolutions to reproduce sea-levels and their extremes and includes a final downscaling module to simulate selected extreme events at sub-kilometer scale with further coupling possibilities (waves, SLR). The framework is developed for the Adriatic Sea and it is validated against tide-gauges. Impact of climate change on storm-surges is assessed by comparing simulations of historical and a future climate change scenario (RCP8.5).
The complexity of the modelling framework is impressive, and together with the targeted resolutions it definitely brings added value at local scales compared to the existing coarse projections in CMIP6, which lack the relevant processes to represent extreme sea-levels (tides, storm-surges). However, the manuscript lacks a clear objective and a clear thread, and statements are made that should be either clarified or carefully rephrased. Additionally, I feel that the main claim of the paper – that such complex, computationally expensive framework is needed to resolve storm-surges/sea-levels at a local scale – is not thoroughly proved. There is no quantitative assessment of the improvement of these variables under the additional resolution or coupling processes included, while the opportunity to demonstrate it (at least partially) is there given the availability of multiple nested models within the framework. Additionally, most conclusions are likely to be specific to the region at hand, and the design of the framework itself elsewhere in the world might not be straightforward (for example, there is little point on targeting decameter scale resolutions – as done in the last module of the framework- if high-resolution bathymetric data is poor or not available). Furthermore, the framework relies on reanalysis data that, at least for the ocean component, comes from a dedicated, regional dataset, which is not available all around the world. Because of these conditions, and the complexity of the modelling framework which hampers an easy relocation to other places in the world, I propose to reframe the study as a dedicated study for the Adriatic Sea (and reflected in the title). The objective should be clearly addressed at the end of the introduction, following a literature review that evidences the research and knowledge gap (which is currently done from a historical point of view – overly extensive, in my opinion -, and at a global/continental scale, but a transition towards the motivation to target the Adriatic Sea is currently lacking). Since this is not a model validation paper, and since it is not suited to be the presentation of a general framework as it is currently suggested, a clear objective should be stated highlighting the expected outcomes (in my view, the demonstration of the added value of such an (expensive) modelling framework compared to more classical, simple approaches – lower resolution, less coupling) and clearly stating the focus of the assessment (in this case storm-surges, which is not clearly stated from the beginning – while it is in the title, and not properly argued in the methods and results) .The manuscript should be restructured and rephrased following the objective at hand, and the target variables and metrics. In terms of target variable, I wonder about the suitability to focus on storm-surges to highlight the added value of the framework (as opposed to total sea-levels), when (extreme) storm-surges are often modelled using comparatively much cheaper barotropic modelling chains. It would be good if, in the introduction, the author could justify the decision to use such a complex system to target storm-surges (providing literature that has, to some extent, evaluated the sensitivity of extreme surges to certain of the elements in the current modelling framework – coupling (in this case between ocean-atmosphere, to justify the use of a 3D ocean model), resolution, non-linear interactions with non-barotropic components, etc. Otherwise, the results could focus on total sea-level extremes, which are more influenced by baroclinic processes.
Finally, the authors propose an approach to exploit the current framework in an efficient manner, by finding a way to identify/isolate the occurrence of extreme events directly from the CMIP6 models and then downscaling those selected events with the framework. I have serious doubts that this is a simple as it is suggested, given considerable snip-up periods associated with 3D regional ocean models (and their sensitivity to initial conditions). This idea is poorly elaborated, and it is highly speculative. I propose it is somewhat left out, or simply included as a potential prospect at the end of section 4.
Because of the raised concerns, I suggest a major revision.
Other specific comments per section are:
- Introduction: I propose to shorten the description of the history of climate projections and focus on current gaps, gearing up towards the motivation behind this study, and finishing with a clear statement on the objectives and approach. Additionally, include other references to regional modelling exercises other than the previous publications of the author.
- Models and Method:
- 1 Here I am missing a plot showing the domains and subdomains involved in the modelling chain, plus some description of the processes included (tides – are tides forced at the boundaries? plus coupling atmosphere-ocean) and background datasets (bathymetry). What about vertical resolution?
- 2/2.3 How are storm-surges isolated from the total sea-level series?
- Results:
- I really miss some plots demonstrating the (presumably positive) impact of the added complexity of the framework to represent (extreme) storm-surges(or sea-levels, if the focus switches to sea-levels)
- It would be nice to have, if metrics on the improvement due to resolution are provided, a map showing the gain in the resolved spatial-scales of the maximum sea-level map between the two resolutions. In order to understand if we really resolve new scales with the additional resolution (or if the extremes – and their changes – are rather smooth instead).
- Here, what is most confusing to me is that you are using a criterion to classify your extreme events based on TWL extremes, but then applying the classification to the storm-surge component alone, therefore losing the correspondence between the category (=return period range) and the targeted events/values. Please explain this clearly in the methods (2.2).
- Discussion and conclusions:
- The discussion section has to be much more elaborated (at the expense of other sections that are too long, you need to be more concise in the description of the results, and refer to tables/figures for a detailed overview) and improved considerably. A proper benchmarking is missing (using regional/Mediterranean assessments, e.g. Toomey et al. 2022, something more comparable to the current analysis). The limitations of the current analysis are not even mentioned. For example, the current assessment uses a single GCM model as forcing, and the method employed to introduce climate change in the model - PGW - has also it's limitation and might lead to climate change impacts completely different to other methods, so the currently presented changes in storm-surges are to be considered with care and highly uncertain. Furthermore, this would be the place to discuss the quantified effects of resolution (and why not, coupling if possible, as this is potentially one of the aspects that makes this setup so heavy). Here, the justification of the added value of this framework should be highlighted, with proper literature review – e.g. why not simply use a modelling framework with barotropic models, without coupling?)
- Section 4.2 is interesting but very speculative. I propose to reduce it and shouldn’t represent a section in itself at the current state of development of the idea. Instead, mention it as a prospect (but also describing the questionable aspects about it – for example the spin-up, or the fact that GCMs will represent the Mediterranean poorly so intermediate modelling activities will be needed, such as Cordex, to avoid large propagating biases along your boundaries ) and it should not be in the abstract, in my opinion. Here you should also mention potential future directions of research given the limitations of the current study (which need to be discussed!) or further exploitation of the dataset generated in this study.
- In general: Avoid using the numbering or items in phrases ( (1)…(2)), most of the times these can be replaces by a simple ‘and’ (especially in cases when just 2 items are described)
I have included detailed comments on the manuscript text in the attached pdf.
-
AC2: 'Reply on RC2', Clea Denamiel, 03 Oct 2023
The authors truly appreciate the careful review and detailed comments provided by the reviewer and would like to thank them for their time and consideration.
Hereafter the comments of the reviewer are highlighted in italic while the responses of the authors are preceded with "R:".
General comments
The study proposes a novel, high-resolution ocean-atmosphere coupled modelling framework to increase the accuracy and therefore local relevance of storm-surge projections under climate change. Their modelling framework benefits from kilometer-scale atmospheric and ocean resolutions to reproduce sea-levels and their extremes and includes a final downscaling module to simulate selected extreme events at sub-kilometer scale with further coupling possibilities (waves, SLR). The framework is developed for the Adriatic Sea and it is validated against tide-gauges. Impact of climate change on storm-surges is assessed by comparing simulations of historical and a future climate change scenario (RCP8.5). The complexity of the modelling framework is impressive, and together with the targeted resolutions it definitely brings added value at local scales compared to the existing coarse projections in CMIP6, which lack the relevant processes to represent extreme sea-levels (tides, storm-surges).
R: Thanks for your encouraging comments.
However, the manuscript lacks a clear objective and a clear thread, and statements are made that should be either clarified or carefully rephrased. Additionally, I feel that the main claim of the paper – that such complex, computationally expensive framework is needed to resolve storm-surges/sea-levels at a local scale – is not thoroughly proved. There is no quantitative assessment of the improvement of these variables under the additional resolution or coupling processes included, while the opportunity to demonstrate it (at least partially) is there given the availability of multiple nested models within the framework. Additionally, most conclusions are likely to be specific to the region at hand, and the design of the framework itself elsewhere in the world might not be straightforward (for example, there is little point on targeting decameter scale resolutions – as done in the last module of the framework- if high-resolution bathymetric data is poor or not available). Furthermore, the framework relies on reanalysis data that, at least for the ocean component, comes from a dedicated, regional dataset, which is not available all around the world. Because of these conditions, and the complexity of the modelling framework which hampers an easy relocation to other places in the world, I propose to reframe the study as a dedicated study for the Adriatic Sea (and reflected in the title). The objective should be clearly addressed at the end of the introduction, following a literature review that evidences the research and knowledge gap (which is currently done from a historical point of view – overly extensive, in my opinion -, and at a global/continental scale, but a transition towards the motivation to target the Adriatic Sea is currently lacking). Since this is not a model validation paper, and since it is not suited to be the presentation of a general framework as it is currently suggested, a clear objective should be stated highlighting the expected outcomes (in my view, the demonstration of the added value of such an (expensive) modelling framework compared to more classical, simple approaches – lower resolution, less coupling) and clearly stating the focus of the assessment (in this case storm-surges, which is not clearly stated from the beginning – while it is in the title, and not properly argued in the methods and results) .The manuscript should be restructured and rephrased following the objective at hand, and the target variables and metrics. In terms of target variable, I wonder about the suitability to focus on storm-surges to highlight the added value of the framework (as opposed to total sea-levels), when (extreme) storm-surges are often modelled using comparatively much cheaper barotropic modelling chains. It would be good if, in the introduction, the author could justify the decision to use such a complex system to target storm-surges (providing literature that has, to some extent, evaluated the sensitivity of extreme surges to certain of the elements in the current modelling framework – coupling (in this case between ocean-atmosphere, to justify the use of a 3D ocean model), resolution, non-linear interactions with non-barotropic components, etc. Otherwise, the results could focus on total sea-level extremes, which are more influenced by baroclinic processes.
R: The authors acknowledge their failure to provide and present a general modelling framework for meter-scale storm-surge projections. Consequently, they will reframe the aim of their study to the specific studied area: the Adriatic Sea.
First, the title will be changed to: Next generation atmosphere-ocean climate modelling for storm surge hazard projections: the case of the Adriatic basin.
Second, as suggested by the reviewer the introduction will be rewritten in order to highlight several aspects missing in the present version. The overly extensive historical literature review that evidences the research and knowledge gaps will be shorten and a transition towards the motivations to target the Adriatic basin will be added. The objective of the article will be reframed to demonstrate the added value of the modelling framework for the projection of storm-surges in the complex Adriatic environment (e.g., topography and bathymetry, basin wide seiches, etc.) compared to more classical, simple approaches using lower resolutions and less coupling. This will include a review of the available storm surge research in the Adriatic Sea. Finally, the justifications of using the presented baroclinic framework instead of simpler barotropic models (as used in the final step of the nesting chain) will be added to the article. This will also include a clear justification for the usage of the PGW methodology.
Finally, the authors propose an approach to exploit the current framework in an efficient manner, by finding a way to identify/isolate the occurrence of extreme events directly from the CMIP6 models and then downscaling those selected events with the framework. I have serious doubts that this is a simple as it is suggested, given considerable snip-up periods associated with 3D regional ocean models (and their sensitivity to initial conditions). This idea is poorly elaborated, and it is highly speculative. I propose it is somewhat left out, or simply included as a potential prospect at the end of section 4.
R: With the idea to reframe the entire study to the Adriatic basin and not to try to provide and present a general modelling framework for meter-scale storm-surge projections, the authors agree with the reviewer that the framework presented at the end of the article could be left out. However, with the presented experiment, the authors learnt that the methodology (PGW + further downscaling of extreme events) is not suitable for representing the climate uncertainty due to its numerical cost.
Other specific comments per section
Introduction
I propose to shorten the description of the history of climate projections and focus on current gaps, gearing up towards the motivation behind this study, and finishing with a clear statement on the objectives and approach. Additionally, include other references to regional modelling exercises other than the previous publications of the author.
R: The authors fully agree and will rewrite the introduction following the structure proposed by the reviewer and presented above.
Models and Method
Here I am missing a plot showing the domains and subdomains involved in the modelling chain, plus some description of the processes included (tides – are tides forced at the boundaries? plus coupling atmosphere-ocean) and background datasets (bathymetry). What about vertical resolution?
R: The modelling framework of the AdriSC modelling suite has been presented in many other articles and the authors thought that citing this extensive literature would be enough. However, as both reviewers specifically ask for more details on the modelling suite, they will be added to the new version of the article.
How are storm-surges isolated from the total sea-level series?
R: Storm surges are selected using return periods extracted from the detrended total sea-level signal. The method will be rewritten in a clearer manner in order to properly highlight how the storm-surges were extracted.
Results
I really miss some plots demonstrating the (presumably positive) impact of the added complexity of the framework to represent (extreme) storm-surges(or sea-levels, if the focus switches to sea-levels)
R: With the new aim of the study concentrating in proving the added value of sub-kilometer modelling for storm surge projections, several plots showing the impact of resolution on the results will be added to the article.
It would be nice to have, if metrics on the improvement due to resolution are provided, a map showing the gain in the resolved spatial-scales of the maximum sea-level map between the two resolutions. In order to understand if we really resolve new scales with the additional resolution (or if the extremes – and their changes – are rather smooth instead).
R: Thanks a lot for the suggestion, such a plot will be added in the new version of the article.
Here, what is most confusing to me is that you are using a criterion to classify your extreme events based on TWL extremes, but then applying the classification to the storm-surge component alone, therefore losing the correspondence between the category (=return period range) and the targeted events/values. Please explain this clearly in the methods (2.2).
R: As mentioned above, the method section will be rewritten in order to clarify how storm surges are selected from the detrended sea-level model results.
Discussion and conclusions
The discussion section has to be much more elaborated (at the expense of other sections that are too long, you need to be more concise in the description of the results, and refer to tables/figures for a detailed overview) and improved considerably. A proper benchmarking is missing (using regional/Mediterranean assessments, e.g. Toomey et al. 2022, something more comparable to the current analysis). The limitations of the current analysis are not even mentioned. For example, the current assessment uses a single GCM model as forcing, and the method employed to introduce climate change in the model - PGW - has also it's limitation and might lead to climate change impacts completely different to other methods, so the currently presented changes in storm-surges are to be considered with care and highly uncertain. Furthermore, this would be the place to discuss the quantified effects of resolution (and why not, coupling if possible, as this is potentially one of the aspects that makes this setup so heavy). Here, the justification of the added value of this framework should be highlighted, with proper literature review – e.g. why not simply use a modelling framework with barotropic models, without coupling?)
R: The authors fully agree with the reviewer and the “discussion and conclusions” section will be rewritten. In particular, the strength (i.e., added value) and weaknesses (e.g., limitation of PGW, climate uncertainty not taken into account, etc.) of the presented method will be described in details. The literature review will also be extended in order to better acknowledge the current efforts made to bridge the gap between coastal hazard studies and climate projections.
Section 4.2 is interesting but very speculative. I propose to reduce it and shouldn’t represent a section in itself at the current state of development of the idea. Instead, mention it as a prospect (but also describing the questionable aspects about it – for example the spin-up, or the fact that GCMs will represent the Mediterranean poorly so intermediate modelling activities will be needed, such as Cordex, to avoid large propagating biases along your boundaries ) and it should not be in the abstract, in my opinion. Here you should also mention potential future directions of research given the limitations of the current study (which need to be discussed!) or further exploitation of the dataset generated in this study.
R: The authors fully agree that, at this stage, the presented methodology needs a lot of work to be properly generalized and that issues like spin-up or poor GCM results in the Mediterranean Sea are definitely serious problems for such an approach. Consequently, as already stated above, this section will be left out of the article and only the fact that more efficient sub-kilometer-scale strategies than the one presented in this article will be mentioned.
In general
Avoid using the numbering or items in phrases ( (1)…(2)), most of the times these can be replaces by a simple ‘and’ (especially in cases when just 2 items are described)
R: Numbering will be removed everywhere it is possible (and particularly when only two items are listed).
Citation: https://doi.org/10.5194/egusphere-2023-913-AC2
Status: closed
-
RC1: 'Comment on egusphere-2023-913', Anonymous Referee #1, 21 Jul 2023
The topic of the manuscript and the modelling effort put on this research are of great relevance. However, I recommend against publication of the manuscript in its current state. I appreciate the effort put by the two authors, and the potential of their analysis, so I would like to challenge authors to resubmit a new version including a specific research aim, sound scientific description of the hypothesis, experimental design, and findings, and how these represent a substantial contribution to understanding future surge hazards in the Adriatic Sea or for modelling of extreme sea levels.
Please find more detailed comments in the attachment, i wish authors understand them to be constructive.
-
AC1: 'Reply on RC1', Clea Denamiel, 03 Oct 2023
The authors truly appreciate the careful review and detailed comments provided by the reviewer and would like to thank them for their time and consideration.
Hereafter the comments of the reviewer are highlighted in italic while the responses of the authors are preceded with "R:".
The topic of the manuscript and the modelling effort put in this research are of great relevance. However, I recommend against publication of the manuscript in its current state. I appreciate the effort put by the two authors, and the potential of their analysis, so I would like to challenge them to resubmit a new version including a specific research aim, sound scientific description of the hypothesis, experimental design, and findings, and how these represent a substantial contribution to understanding future surge hazards in the Adriatic Sea or for modelling of extreme sea levels. I admit that with such large and detailed analysis it can get very complicated, but the current presentation quality is rather low. Additionally, although the scientific significance and quality of this manuscript for the understanding of future surge hazards in the Adriatic sea has the potential to be high, in the current stage is uncertain due to missing details or discussion. Although I wrote several major and minor comments on the current manuscript, I would like to first share some general insights below to support authors with the resubmission of their manuscript. Please understand them to be constructive.
R: The authors acknowledge their failure to provide and present a general modelling framework for meter-scale storm-surge projections. Consequently, they will reframe the aim of their study to the specific studied area: the Adriatic basin.
- The introduction is wide and general. There is little focus on the case study. I also miss insights on what the current scientific gaps are and what are those taken in this manuscript, what is new? is that increase on resolution? The extreme events simulations? why is the PGW used over other methods?
- There is little reflection on what others have done regarding waves and surge projections in the introduction.
R: The introduction will be rewritten in order to highlight several aspects missing in the present version. The overly extensive historical literature review that evidences the research and knowledge gaps will be shorten and a transition towards the motivations to target the Adriatic basin will be added. The objective of the article will be reframed to demonstrate the added value of the modelling framework for the projection of storm-surges in the complex Adriatic environment (e.g., topography and bathymetry, basin wide seiches, etc.) compared to more classical, simple approaches using lower resolutions and less coupling. This will include a thorough review of the available storm surge research in the Adriatic Sea. Finally, the justifications of using the presented baroclinic framework instead of simpler barotropic models (as used in the final step of the nesting chain) will be added to the article. This will also include a clear justification for the usage of the PGW methodology.
- The manuscript is difficult to read, in my opinion using too much decoration, I advise to adopt a more scientific approach on statements. In fact, the positioning is not clear, is a methodological finding or a case study presented? For example, Line 62- Finally, the local application of the (sub-)kilometre-scale methodology to the Adriatic Sea is only used as a proof of concept and all the presented approaches and results can be replicated and/or adapted at any location in the world where extreme sea level hazard assessments fully depend on the accurate representation of the complex geomorphology of the coastal areas I advise in such a case to present the experiment that validates the hypothesis that upscaling of the methodology is possible to any location in the world. More precisely, I would encourage the authors to be cautious with limitations, once their methodology might struggle to simulate future meteo-tsunamis, that are of great relevance for extreme sea level hazard assessment in the Adriatic sea.
R: As stated above the reframed aim of the study will be to demonstrate the added value of the modelling framework for the projection of storm-surges in the complex Adriatic environment (e.g., topography and bathymetry, basin wide seiches, etc.) compared to more classical, simple approaches using lower resolutions and less coupling.
As for modelling the frequency and intensity of past and future meteotsunami events, the authors recently published a methodology for this specific purpose (Denamiel et al., 2023). This is indeed another important contribution to extreme sea levels in the Adriatic Sea but it would require either to use sub-kilometer-scale atmospheric model at the climate scale (unrealistic at this stage due to the numerical cost) or the use of the methodology presented in Denamiel et al. (2023) and at the end of this article. To this date, such an approach is still under development (i.e., synoptic index performances should be increased, method should be tested in different places, etc.) and, as suggested by the other reviewer, is not necessarily relevant to this particular article.
Denamiel C, Belušić D, Zemunik P and Vilibić I. Climate projections of meteotsunami hazards. Front. Mar. Sci., 10,1167863 (2023). https://doi.org/10.3389/fmars.2023.1167863
- The manuscript requires that methods are reproducible, important elements are missing in the methodology. For example, several models such as ROMS, ADCIRC, and SWAN are used, where parameterizations, spatial, and temporal discretization are really impacting results. Details on how the pseudo-global warming method is applied are neither provided.
R: The modelling framework of the AdriSC modelling suite has been presented in many other articles and the authors thought that citing this extensive literature would be enough. However, as both reviewers specifically ask for more details on the modelling suite, they will be added to the new version of the article.
- Regarding the above, if the manuscript aims at developing a new modelling approach for investigating future extreme surges, I encourage to present the experimental design elaborating on sensitivities or uncertainties arising from model choices.
- Then, how do sensitivities from modelling choice compare with the changes seen in the future scenario simulated? If authors have addressed this in other research publications it is worth to incorporate explicitly in this manuscript.
R: Sensitivities and uncertainties introduced by the choice of the models used to downscale the regional (or global) climate model results can be important. However, in this article, we have demonstrated that the kilometer-scale modelling suite is performing well during the 1987-2017 period (comparison of the extreme sea level distributions at different tide gauges). The authors thus believe that representing the climate uncertainty (i.e., having different scenarios of climate warming from different global climate models) is much more critical than representing the uncertainty linked to the choice of the models used for the downscaling. This is something that cannot be done with the presented methodology due to its numerical cost. This will be discussed in the “discussion and conclusions” section.
- The manuscript structure is ok regarding sections used; however I found significant information out of their corresponding section; a few examples are: Line 134-138 In methods but it is part of the results section. Line 138-143 In methods but it is part of the discussion section. Line 188-196 In methods but it is part of discussion In general, in the results section new data is described.
R: The authors will make sure to move the results from the method section.
- In my opinion, there are choices that are inconsistent, and that are not discussed appropriately. For example: If interaction between mean sea level rise and other sea level components are of interest in this research, why the regional model does not account for it, but local ones do? I consider that superimposing mean sea level rise in the boundary conditions of the local models could lead into numerical artifacts in the local results, to accommodate for inconsistent boundary conditions, I would say that, if it is one of the research objectives of this manuscript, at least understanding the effect of this superposition deserves investigation.
R: The authors are not responsible for the choices made by the Med-CORDEX community (regional climate simulations of the Mediterranean Sea) that did not include sea level rise in their simulations and decided to only add the tides (as sea-level and barotropic currents) in their baroclinic simulations to not create too much imbalances between 3D current forcing and sea-levels (continuity equation) and to not count twice the effects of the thermal expansion (included in the sea level rise datasets provided by the IPCC-AR5 2015 ensemble). Further, adding sea level rise to the initial/boundary conditions of models like ADCIRC is a common practice in the extreme sea level community using barotropic models which is not known to create any instability as it just raise the initial/boundary sea levels of the domain and, in contrast with baroclinic models, the balance between sea-levels and depth-averaged currents is not critical as depth averaged currents are generally not well represented by barotropic models (i.e., often not even analyzed). Finally, during the development and implementation of the AdriSC modelling suite, the choice of the locations of the boundaries of the different atmosphere-ocean models has been made with extreme care in order to be able to represent all important processes driving the Adriatic Sea circulation. For example, to this date, the AdriSC climate model is the only model capable to represent the impact of the North Ionian Gyre (or BioS) within the Adriatic Sea (Denamiel et al., 2022). The authors thus believe that, despite all the modelling errors (see evaluation of the AdriSC climate model; Denamiel et al., 2021 and Pranić et al., 2021), the AdriSC climate modelling suite has already been proven to perform well within the Adriatic basin.
Denamiel, C., Pranić, P., Ivanković, D., Tojčić, I., and Vilibić, I. Performance of the Adriatic Sea and Coast (AdriSC) climate component—a COAWST V3.3-based coupled atmosphere–ocean modelling suite: atmospheric dataset, Geosci. Model Dev., 14, 3995–4017 (2021). https://doi.org/10.5194/gmd-14-3995-2021
Denamiel, C., Tojčić, I., Pranić, P. et al. Modes of the BiOS-driven Adriatic Sea thermohaline variability. Clim Dyn 59, 1097–1113 (2022). https://doi.org/10.1007/s00382-022-06178-4
Pranić, P., Denamiel, C., and Vilibić, I. Performance of the Adriatic Sea and Coast (AdriSC) climate component—a COAWST V3.3-based one-way coupled atmosphere–ocean modelling suite: ocean results, Geosci. Model Dev., 14, 5927–5955 (2021). https://doi.org/10.5194/gmd-14-5927-2021
- I find figure 1 too general, a proper schematic of the manuscript workflow would help very much on understanding the methodological approach and how it benefits the assessment of future extreme surge.
R: Figure 1 will be redrawn to better present the needs and added value of the sub-kilometer-scale approach in storm surge projections.
- In general, consider the use of appendixes when needed.
R: As the article will be fully restructured in order to prove the added value of the presented approach, material will definitely be moved to either appendixes or supplementary material.
Citation: https://doi.org/10.5194/egusphere-2023-913-AC1
-
AC1: 'Reply on RC1', Clea Denamiel, 03 Oct 2023
-
RC2: 'Comment on egusphere-2023-913', Anonymous Referee #2, 25 Aug 2023
General comments
The study proposes a novel, high-resolution ocean-atmosphere coupled modelling framework to increase the accuracy and therefore local relevance of storm-surge projections under climate change. Their modelling framework benefits from kilometer-scale atmospheric and ocean resolutions to reproduce sea-levels and their extremes and includes a final downscaling module to simulate selected extreme events at sub-kilometer scale with further coupling possibilities (waves, SLR). The framework is developed for the Adriatic Sea and it is validated against tide-gauges. Impact of climate change on storm-surges is assessed by comparing simulations of historical and a future climate change scenario (RCP8.5).
The complexity of the modelling framework is impressive, and together with the targeted resolutions it definitely brings added value at local scales compared to the existing coarse projections in CMIP6, which lack the relevant processes to represent extreme sea-levels (tides, storm-surges). However, the manuscript lacks a clear objective and a clear thread, and statements are made that should be either clarified or carefully rephrased. Additionally, I feel that the main claim of the paper – that such complex, computationally expensive framework is needed to resolve storm-surges/sea-levels at a local scale – is not thoroughly proved. There is no quantitative assessment of the improvement of these variables under the additional resolution or coupling processes included, while the opportunity to demonstrate it (at least partially) is there given the availability of multiple nested models within the framework. Additionally, most conclusions are likely to be specific to the region at hand, and the design of the framework itself elsewhere in the world might not be straightforward (for example, there is little point on targeting decameter scale resolutions – as done in the last module of the framework- if high-resolution bathymetric data is poor or not available). Furthermore, the framework relies on reanalysis data that, at least for the ocean component, comes from a dedicated, regional dataset, which is not available all around the world. Because of these conditions, and the complexity of the modelling framework which hampers an easy relocation to other places in the world, I propose to reframe the study as a dedicated study for the Adriatic Sea (and reflected in the title). The objective should be clearly addressed at the end of the introduction, following a literature review that evidences the research and knowledge gap (which is currently done from a historical point of view – overly extensive, in my opinion -, and at a global/continental scale, but a transition towards the motivation to target the Adriatic Sea is currently lacking). Since this is not a model validation paper, and since it is not suited to be the presentation of a general framework as it is currently suggested, a clear objective should be stated highlighting the expected outcomes (in my view, the demonstration of the added value of such an (expensive) modelling framework compared to more classical, simple approaches – lower resolution, less coupling) and clearly stating the focus of the assessment (in this case storm-surges, which is not clearly stated from the beginning – while it is in the title, and not properly argued in the methods and results) .The manuscript should be restructured and rephrased following the objective at hand, and the target variables and metrics. In terms of target variable, I wonder about the suitability to focus on storm-surges to highlight the added value of the framework (as opposed to total sea-levels), when (extreme) storm-surges are often modelled using comparatively much cheaper barotropic modelling chains. It would be good if, in the introduction, the author could justify the decision to use such a complex system to target storm-surges (providing literature that has, to some extent, evaluated the sensitivity of extreme surges to certain of the elements in the current modelling framework – coupling (in this case between ocean-atmosphere, to justify the use of a 3D ocean model), resolution, non-linear interactions with non-barotropic components, etc. Otherwise, the results could focus on total sea-level extremes, which are more influenced by baroclinic processes.
Finally, the authors propose an approach to exploit the current framework in an efficient manner, by finding a way to identify/isolate the occurrence of extreme events directly from the CMIP6 models and then downscaling those selected events with the framework. I have serious doubts that this is a simple as it is suggested, given considerable snip-up periods associated with 3D regional ocean models (and their sensitivity to initial conditions). This idea is poorly elaborated, and it is highly speculative. I propose it is somewhat left out, or simply included as a potential prospect at the end of section 4.
Because of the raised concerns, I suggest a major revision.
Other specific comments per section are:
- Introduction: I propose to shorten the description of the history of climate projections and focus on current gaps, gearing up towards the motivation behind this study, and finishing with a clear statement on the objectives and approach. Additionally, include other references to regional modelling exercises other than the previous publications of the author.
- Models and Method:
- 1 Here I am missing a plot showing the domains and subdomains involved in the modelling chain, plus some description of the processes included (tides – are tides forced at the boundaries? plus coupling atmosphere-ocean) and background datasets (bathymetry). What about vertical resolution?
- 2/2.3 How are storm-surges isolated from the total sea-level series?
- Results:
- I really miss some plots demonstrating the (presumably positive) impact of the added complexity of the framework to represent (extreme) storm-surges(or sea-levels, if the focus switches to sea-levels)
- It would be nice to have, if metrics on the improvement due to resolution are provided, a map showing the gain in the resolved spatial-scales of the maximum sea-level map between the two resolutions. In order to understand if we really resolve new scales with the additional resolution (or if the extremes – and their changes – are rather smooth instead).
- Here, what is most confusing to me is that you are using a criterion to classify your extreme events based on TWL extremes, but then applying the classification to the storm-surge component alone, therefore losing the correspondence between the category (=return period range) and the targeted events/values. Please explain this clearly in the methods (2.2).
- Discussion and conclusions:
- The discussion section has to be much more elaborated (at the expense of other sections that are too long, you need to be more concise in the description of the results, and refer to tables/figures for a detailed overview) and improved considerably. A proper benchmarking is missing (using regional/Mediterranean assessments, e.g. Toomey et al. 2022, something more comparable to the current analysis). The limitations of the current analysis are not even mentioned. For example, the current assessment uses a single GCM model as forcing, and the method employed to introduce climate change in the model - PGW - has also it's limitation and might lead to climate change impacts completely different to other methods, so the currently presented changes in storm-surges are to be considered with care and highly uncertain. Furthermore, this would be the place to discuss the quantified effects of resolution (and why not, coupling if possible, as this is potentially one of the aspects that makes this setup so heavy). Here, the justification of the added value of this framework should be highlighted, with proper literature review – e.g. why not simply use a modelling framework with barotropic models, without coupling?)
- Section 4.2 is interesting but very speculative. I propose to reduce it and shouldn’t represent a section in itself at the current state of development of the idea. Instead, mention it as a prospect (but also describing the questionable aspects about it – for example the spin-up, or the fact that GCMs will represent the Mediterranean poorly so intermediate modelling activities will be needed, such as Cordex, to avoid large propagating biases along your boundaries ) and it should not be in the abstract, in my opinion. Here you should also mention potential future directions of research given the limitations of the current study (which need to be discussed!) or further exploitation of the dataset generated in this study.
- In general: Avoid using the numbering or items in phrases ( (1)…(2)), most of the times these can be replaces by a simple ‘and’ (especially in cases when just 2 items are described)
I have included detailed comments on the manuscript text in the attached pdf.
-
AC2: 'Reply on RC2', Clea Denamiel, 03 Oct 2023
The authors truly appreciate the careful review and detailed comments provided by the reviewer and would like to thank them for their time and consideration.
Hereafter the comments of the reviewer are highlighted in italic while the responses of the authors are preceded with "R:".
General comments
The study proposes a novel, high-resolution ocean-atmosphere coupled modelling framework to increase the accuracy and therefore local relevance of storm-surge projections under climate change. Their modelling framework benefits from kilometer-scale atmospheric and ocean resolutions to reproduce sea-levels and their extremes and includes a final downscaling module to simulate selected extreme events at sub-kilometer scale with further coupling possibilities (waves, SLR). The framework is developed for the Adriatic Sea and it is validated against tide-gauges. Impact of climate change on storm-surges is assessed by comparing simulations of historical and a future climate change scenario (RCP8.5). The complexity of the modelling framework is impressive, and together with the targeted resolutions it definitely brings added value at local scales compared to the existing coarse projections in CMIP6, which lack the relevant processes to represent extreme sea-levels (tides, storm-surges).
R: Thanks for your encouraging comments.
However, the manuscript lacks a clear objective and a clear thread, and statements are made that should be either clarified or carefully rephrased. Additionally, I feel that the main claim of the paper – that such complex, computationally expensive framework is needed to resolve storm-surges/sea-levels at a local scale – is not thoroughly proved. There is no quantitative assessment of the improvement of these variables under the additional resolution or coupling processes included, while the opportunity to demonstrate it (at least partially) is there given the availability of multiple nested models within the framework. Additionally, most conclusions are likely to be specific to the region at hand, and the design of the framework itself elsewhere in the world might not be straightforward (for example, there is little point on targeting decameter scale resolutions – as done in the last module of the framework- if high-resolution bathymetric data is poor or not available). Furthermore, the framework relies on reanalysis data that, at least for the ocean component, comes from a dedicated, regional dataset, which is not available all around the world. Because of these conditions, and the complexity of the modelling framework which hampers an easy relocation to other places in the world, I propose to reframe the study as a dedicated study for the Adriatic Sea (and reflected in the title). The objective should be clearly addressed at the end of the introduction, following a literature review that evidences the research and knowledge gap (which is currently done from a historical point of view – overly extensive, in my opinion -, and at a global/continental scale, but a transition towards the motivation to target the Adriatic Sea is currently lacking). Since this is not a model validation paper, and since it is not suited to be the presentation of a general framework as it is currently suggested, a clear objective should be stated highlighting the expected outcomes (in my view, the demonstration of the added value of such an (expensive) modelling framework compared to more classical, simple approaches – lower resolution, less coupling) and clearly stating the focus of the assessment (in this case storm-surges, which is not clearly stated from the beginning – while it is in the title, and not properly argued in the methods and results) .The manuscript should be restructured and rephrased following the objective at hand, and the target variables and metrics. In terms of target variable, I wonder about the suitability to focus on storm-surges to highlight the added value of the framework (as opposed to total sea-levels), when (extreme) storm-surges are often modelled using comparatively much cheaper barotropic modelling chains. It would be good if, in the introduction, the author could justify the decision to use such a complex system to target storm-surges (providing literature that has, to some extent, evaluated the sensitivity of extreme surges to certain of the elements in the current modelling framework – coupling (in this case between ocean-atmosphere, to justify the use of a 3D ocean model), resolution, non-linear interactions with non-barotropic components, etc. Otherwise, the results could focus on total sea-level extremes, which are more influenced by baroclinic processes.
R: The authors acknowledge their failure to provide and present a general modelling framework for meter-scale storm-surge projections. Consequently, they will reframe the aim of their study to the specific studied area: the Adriatic Sea.
First, the title will be changed to: Next generation atmosphere-ocean climate modelling for storm surge hazard projections: the case of the Adriatic basin.
Second, as suggested by the reviewer the introduction will be rewritten in order to highlight several aspects missing in the present version. The overly extensive historical literature review that evidences the research and knowledge gaps will be shorten and a transition towards the motivations to target the Adriatic basin will be added. The objective of the article will be reframed to demonstrate the added value of the modelling framework for the projection of storm-surges in the complex Adriatic environment (e.g., topography and bathymetry, basin wide seiches, etc.) compared to more classical, simple approaches using lower resolutions and less coupling. This will include a review of the available storm surge research in the Adriatic Sea. Finally, the justifications of using the presented baroclinic framework instead of simpler barotropic models (as used in the final step of the nesting chain) will be added to the article. This will also include a clear justification for the usage of the PGW methodology.
Finally, the authors propose an approach to exploit the current framework in an efficient manner, by finding a way to identify/isolate the occurrence of extreme events directly from the CMIP6 models and then downscaling those selected events with the framework. I have serious doubts that this is a simple as it is suggested, given considerable snip-up periods associated with 3D regional ocean models (and their sensitivity to initial conditions). This idea is poorly elaborated, and it is highly speculative. I propose it is somewhat left out, or simply included as a potential prospect at the end of section 4.
R: With the idea to reframe the entire study to the Adriatic basin and not to try to provide and present a general modelling framework for meter-scale storm-surge projections, the authors agree with the reviewer that the framework presented at the end of the article could be left out. However, with the presented experiment, the authors learnt that the methodology (PGW + further downscaling of extreme events) is not suitable for representing the climate uncertainty due to its numerical cost.
Other specific comments per section
Introduction
I propose to shorten the description of the history of climate projections and focus on current gaps, gearing up towards the motivation behind this study, and finishing with a clear statement on the objectives and approach. Additionally, include other references to regional modelling exercises other than the previous publications of the author.
R: The authors fully agree and will rewrite the introduction following the structure proposed by the reviewer and presented above.
Models and Method
Here I am missing a plot showing the domains and subdomains involved in the modelling chain, plus some description of the processes included (tides – are tides forced at the boundaries? plus coupling atmosphere-ocean) and background datasets (bathymetry). What about vertical resolution?
R: The modelling framework of the AdriSC modelling suite has been presented in many other articles and the authors thought that citing this extensive literature would be enough. However, as both reviewers specifically ask for more details on the modelling suite, they will be added to the new version of the article.
How are storm-surges isolated from the total sea-level series?
R: Storm surges are selected using return periods extracted from the detrended total sea-level signal. The method will be rewritten in a clearer manner in order to properly highlight how the storm-surges were extracted.
Results
I really miss some plots demonstrating the (presumably positive) impact of the added complexity of the framework to represent (extreme) storm-surges(or sea-levels, if the focus switches to sea-levels)
R: With the new aim of the study concentrating in proving the added value of sub-kilometer modelling for storm surge projections, several plots showing the impact of resolution on the results will be added to the article.
It would be nice to have, if metrics on the improvement due to resolution are provided, a map showing the gain in the resolved spatial-scales of the maximum sea-level map between the two resolutions. In order to understand if we really resolve new scales with the additional resolution (or if the extremes – and their changes – are rather smooth instead).
R: Thanks a lot for the suggestion, such a plot will be added in the new version of the article.
Here, what is most confusing to me is that you are using a criterion to classify your extreme events based on TWL extremes, but then applying the classification to the storm-surge component alone, therefore losing the correspondence between the category (=return period range) and the targeted events/values. Please explain this clearly in the methods (2.2).
R: As mentioned above, the method section will be rewritten in order to clarify how storm surges are selected from the detrended sea-level model results.
Discussion and conclusions
The discussion section has to be much more elaborated (at the expense of other sections that are too long, you need to be more concise in the description of the results, and refer to tables/figures for a detailed overview) and improved considerably. A proper benchmarking is missing (using regional/Mediterranean assessments, e.g. Toomey et al. 2022, something more comparable to the current analysis). The limitations of the current analysis are not even mentioned. For example, the current assessment uses a single GCM model as forcing, and the method employed to introduce climate change in the model - PGW - has also it's limitation and might lead to climate change impacts completely different to other methods, so the currently presented changes in storm-surges are to be considered with care and highly uncertain. Furthermore, this would be the place to discuss the quantified effects of resolution (and why not, coupling if possible, as this is potentially one of the aspects that makes this setup so heavy). Here, the justification of the added value of this framework should be highlighted, with proper literature review – e.g. why not simply use a modelling framework with barotropic models, without coupling?)
R: The authors fully agree with the reviewer and the “discussion and conclusions” section will be rewritten. In particular, the strength (i.e., added value) and weaknesses (e.g., limitation of PGW, climate uncertainty not taken into account, etc.) of the presented method will be described in details. The literature review will also be extended in order to better acknowledge the current efforts made to bridge the gap between coastal hazard studies and climate projections.
Section 4.2 is interesting but very speculative. I propose to reduce it and shouldn’t represent a section in itself at the current state of development of the idea. Instead, mention it as a prospect (but also describing the questionable aspects about it – for example the spin-up, or the fact that GCMs will represent the Mediterranean poorly so intermediate modelling activities will be needed, such as Cordex, to avoid large propagating biases along your boundaries ) and it should not be in the abstract, in my opinion. Here you should also mention potential future directions of research given the limitations of the current study (which need to be discussed!) or further exploitation of the dataset generated in this study.
R: The authors fully agree that, at this stage, the presented methodology needs a lot of work to be properly generalized and that issues like spin-up or poor GCM results in the Mediterranean Sea are definitely serious problems for such an approach. Consequently, as already stated above, this section will be left out of the article and only the fact that more efficient sub-kilometer-scale strategies than the one presented in this article will be mentioned.
In general
Avoid using the numbering or items in phrases ( (1)…(2)), most of the times these can be replaces by a simple ‘and’ (especially in cases when just 2 items are described)
R: Numbering will be removed everywhere it is possible (and particularly when only two items are listed).
Citation: https://doi.org/10.5194/egusphere-2023-913-AC2
Data sets
Next-generation (sub-)kilometre-scale climate modelling for extreme storm-surge hazard projections Cléa Denamiel https://osf.io/2hgfm
Model code and software
AdriSC Climate Model: evaluation run Cléa Denamiel https://osf.io/zb3cm
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 333 | 170 | 24 | 527 | 11 | 13 |
- HTML: 333
- PDF: 170
- XML: 24
- Total: 527
- BibTeX: 11
- EndNote: 13
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1