The paper focuses on a highly relevant topic: Early Warning Systems for storm-induced overwash in a coastal region. While the paper is well written and structured, I believe three key topics require further attention from the authors before publication. 

1)    The first and most important of these is the correct definition of Overwash and Flood. In fact, the authors only evaluate “smaller” berm overtopping (traces) and “larger” berm overtopping (which they call overwash and flood) and never mention dune overwash or the flood/inundation of assets. I have made several comments on this issue below, and I think the authors should consider Sallenger’s regimes for this study. They could even introduce a “berm overtopping” regime as a complement.

2)    Secondly, the authors should also clearly indicate the indicators to be used in the EWS regarding assets. While they recognise the necessity of indicators for the EWS in the Introduction, they never mention what thresholds are used or what warning levels have been defined (and why). This is never clearly stated throughout the paper. What warning levels exist, and how will they be sent to the first responders, managers, and the population? 

3)    Thirdly, the entire validation uses all data, including a large majority of low-energy conditions, whereas the EWS is designed for storms. Thus, the authors should clearly define the thresholds above which it is considered a storm for the purposes of this study, and then validate only values above these thresholds. Validation against mild or low wave conditions is not very meaningful when dealing with this type of work. 

Detailed comments are provided below, including reinforcement of the points already mentioned.

Methodology:

Figure 1: Please include the name Poetto Beach in Figure 1b); Please include the acronym CAM after videocameras in the caption of Figure 1.

Lines 149-151: Not easy to understand and split between these two concepts. Is overwash simply when the berm is surpassed by the run-up, and is flood simply when there is extensive water over the berm? Is that correct? Please be clearer. Why don’t you use Sallenger's terminology for overwash and inundation? Why is the berm crest used instead of the dune crest? This is berm overtopping, not overwash.

Line 150: What do you mean by "limited water coverage"? Is it the maximum limit inland reached by the water? The meaning of limited and extended water coverage is unclear, which is paramount to understanding the rest of the paper.

Line 158: above you have beta in italic for the foreshore slope, but not here. Please check for consistency. Please check all acronyms.

Lines 160-161: Why is berm height used and not dune height? If there are dunes inland (even low ones), flooding should only occur after the dunes (or an obstacle where there are no dunes) have been overpassed, and not the berm. It is possible to have berm overtopping with no flooding for infrastructures. Please justify this choice.

Lines 184-185: Integrating measured bathymetry with the lower-resolution data from EMODNET is not simple, especially since EMODNet data often has relatively high vertical errors. How did you merge the data into the grid, and how were these potential vertical errors considered?

Lines 223-226: Is this enough? Waves forecasted with these global models are often underpredicted during storms.  It is most likely that the EWS will also underestimate the flood hazard.  What has been done to minimise this problem and ensure that there is no underestimation?

Lines 234-235: The tide gauge should only measure the sea level. How did you compare wind and wave data with data from the tide gauge?

Line 246: Replace Ruju and Viola 2024 by Ruju and Viola (2024). Check the references throughout the text. There are some minor typos.

Table 2: Table 2 is first included in the text and only mentioned afterwards. It should be the other way around. Please revise it for the final version. The explanation of the variables should not be inside the table. Some of the variables have already been explained in the text, so there is no need to repeat them.

Results:

Line 274: The retreat is true, but only for the berm. The dune does not appear to have major changes in position (only in height). Therefore, it does not affect the entire profile.

Line 276: Is this the reason for the flat and lowered berm profiles at P2 in the last two surveys?

Line 278: Why has the dune crest lowered? Wave overtopping? Wind removal without compensation at the front? Human trampling? This has never been explained or analysed, even though the paper is about overwash (runup exceeding dune height).

Lines 285-286: I disagree that this threshold is commonly used. In the vast majority of the existing works, the dune height is used as the threshold for coastal flooding. As can easily be seen from the two profiles, the berm can be overwashed, and there will still be enough berm for water percolation and infiltration before reaching the dune elevation. Flooding of human settlements will only occur after overwashing the dune crest and not the berm crest, unless there is significant occupation on top of the berm during winter, in which case, the risk to that occupation needs to be assessed. Is that the case?

Line 287: “flooding threshold values in NEPTUNE-EWS”. This must be clearly explained. Why is he berm crest used instead of the dune crest? In Sallenger's regimes and all other derived works, it is always the dune elevation and never the berm elevation. If there is a dune between the road, the houses, and the waves, it should be the dune crest to be used, not the berm crest. If there is no dune, then you should use the maximum elevation before any human structure. If there are buildings between the dune and the berm crest, then two thresholds are needed: one for the flood of the infrastructure between the berm and the dune (which can be the berm crest) and another for the flooding of the assets behind the dune (the dune crest elevation). However, this simple approach (using berm crest) is not capable of distinguishing between different levels of berm overtopping, dune overwash or inundation, and therefore it does not work properly as an EWS.

Line 346: "the" and not them

Line 351: Would it be possible to correct the values or apply a correction that would suit and improve the data, particularly for storms?

Line 369 and data analysis for the 45 days period: The analysis was performed on the full 45-day dataset. However, for an EWS focused on storms, the analysis should cover a longer period and focus only on deviations during storms. The authors should have selected a threshold for storms (at least a simple Hs threshold) and analysed the data above that threshold only. From the records, I can see only one or two potential storms for the analysed period. The fit is very good for the highest storm, and better than what I would expect from global models at coastal locations.

Line 382: “low wave heights”. Yes, but for a storm model, the concern with small wave heights is probably minimal. This is the reason why the analysis should focus on wave conditions with the potential to cause Overwash, rather than on all conditions.

Line 397: Are these waves relevant to the analysis? Do they produce overwash? If not, why analyse them?

Lines 426-427: It seems that there is a 10 cm vertical difference between most of the shown records. However, the BIAS shows almost no difference (1 cm). So, why is there an underestimation during storms when the opposite should occur? The model is most probably not correctly predicting storm surge.

Line 438: What happens when you average the entire set of storm data? Is there a vertical difference? According to the BIAS, the difference for the entire series is almost 0. Is it the same for storms, or is the model unable to correctly reproduce the storm surge? For a model that considers storms, the analysis should focus mostly on storm data and not on all data.

Lines 439-440: I believe so, but wouldn't it be better to apply a correction?

Table 4: You have some values of mean Hs of 0.6 m to 0.8 m and periods of around 6-7 s. Is that really a storm capable of generating Overwash, or is it because you defined the berm crest as the limit for Overwash, meaning that lower waves can produce overwash events?

Figures 14, 15 and 16 (caption): I think this is the first time you have referred to a Flood State, mentioning classes No trace, Overwash and Flood. What do these terms mean? You defined these terms in a very simple (and not entirely accurate) way at the end of 2.1 study Area, which is strange. This definition is the most important in the paper and should not be at the end of the Study Area. You should clearly state what a Flood State (or Regime, since not all conditions lead to flooding) means. You should use Sallenger's regimes. Does Flood correspond to Inundation in the Sallenger regime? I don't think so.  How do you differentiate between Overwash and Flood? I think you are only analysing berm overtopping and not true dune Overwash, as it should be.

Lines 491-581: The event-by-event description (lines 491 - 581) is tedious and should be simplified, probably using a table to summarise these results.

Line 499 and figures 14 and 15: In figures 14 and 15 you distinguish between overwash and flooding. Here, however, you combine both. Why? Why are you not adopting Sallenger's terminology and concepts?

Discussion

Lines 638-640: This is true, but in general, I would say that for beaches that are exposed to wave action, the runup must be included. In sheltered areas such as estuaries and backbarriers, however, this is not necessary because small waves with short periods do not contribute to effective flooding. Therefore, in this case, it is important to include the wave-induced runup. This does not necessarily contradict other studies, but rather represents the correct definition of what is relevant for each environment.

Line 641: “flooding predictions”. This could refer to berm flooding or berm overtopping. You never mention flooding of roads or structures, even though the entire introduction points to that. I would say that the introduction of this paper misleads the reader, and the definitions are not clear enough to effectively state that you are only analysing berm overtopping and not dune overwash or infrastructure overwash/flood. This must be clear throughout the paper.

Lines 643-657: This description from line 643 until 657 partially repeat the results.

Conclusions

Line 689: This paper presents the validation of a potential warning system, rather than the warning system itself. All the data and discussion is about the validation, not the EWS and its implementation.

Line 700: You are forecasting berm overtopping, not coastal flooding. These are quite different.

Line 701: The meaning of a storm event for the region was never defined, and this should be done before any data analysis. The data analysis should focus primarily on the storm events and exclude calm conditions.

This is definitely a good quality work that merits publication. However, while I am positive, I believe that the structure and presentation can be improved substantially. The paper presents efforts towards a potential EWS and not the EWS itself. There is a lot of information on validation and the paper is rather long. For that reason, I see a lot of potential to produce two papers and that would be my first suggestion if the authors decide to take that way:

• One paper presenting the modelling chain and the validation results. I think this is the most solid part and I would be happy to review a revision, which focusses only on that aspect of the work
• Second paper discussing about how the model output is converted to early warnings and their accuracy. I believe that here there still needs to be done some additional work in order to present something more robust and closer to the current state of the art.

Obviously, the authors can choose to try to include everything in one paper, but they need to try to prepare something easier for the reader to follow. I believe that to do that, they need to try to focus only on the important information, and even move some technical parts in the appendix (like extended model validation, different run up formulas, etc).

The system is based on forcing from other models to run SWAN. The model performs rather well in the validation for Hs, but is less accurate for the period, direction and other secondary parameters. This is a known issue of such wave models and there are several previous studies reporting similar performance that could be cited to support to present effort.

The authors use empirical run up formulas to assess potential overtopping events and in this part there are a lot of aspects that merit clarification. First of all, since there is a video system on-site that could at least provide run up maxima, the authors could validate the different fomulas, or even fit a site-specific one. Second, any discrepancy between observed and estimated R2 can be the result of a combination of model/formula error, beach face slope inaccuracy and the variable capacity of the changing bathymetry to attenuate incoming wave energy. For all of the above there is a substantial amount of literature, that’s why I am in favor of splitting the manuscript in two separate papers.

Assuming that that the use of the video imagery information can solve the run up estimation accuracy and that the beach face slope doesn’t change a lot, as stated in the manuscript, the authors need to address the effect of the nearshore bathymetry. I add below some examples from relevant literature that could help (the order is chronological and the list is not exhaustive):

https://onlinelibrary.wiley.com/doi/abs/10.1002/esp.2264

https://core.ac.uk/download/pdf/36733581.pdf

https://www.sciencedirect.com/science/article/pii/S1674237019300833

https://www.sciencedirect.com/science/article/abs/pii/S0141118722003601

There is a lot of literature about how to estimate accurately the wave run up and the authors discuss some. It is not clear that using a model like XBeach could improve the performance if the bathymetry is not updated (which is practically not possible for known reasons), but that could be investigated. In case there are data, the authors could have estimates of potential seasonal profiles that could be used. Also if the R2 cannot be accurately estimated then the authors could consider issuing warnings based on some percentile threshold of the TWL and not the berm/dune height itself. Finally, if they combine all sources of incertainty (e.g. range of slope, range of R2, range of dune height) they could issue forecasts in a probabilistic framework (i.e. probability of overtopping, rather than a binary outcome).

Of course all the above are suggestions that I hope could be helpful.