Review of the article titled “Monitoring the record-breaking wave event in Melilla harbour (SW Mediterranean Sea)" by Lorente, P., et al. 2023

The manuscript "Monitoring the record-breaking wave event in Melilla harbour (SW Mediterranean Sea)" by Lorente, P., et al. 2023 uses different database such as reanalysis, forecasting model, radar tide-gauge and in situ coastal buoys, to describe an oceanic extreme event that occurred in the Melilla port during April 4th and 5th, 2022. It also analyzes the extreme regime in the Alboran Sea. The impacts of extreme wave events on harbors and the need to revise the level of security within them regarding the new climatic conditions are interesting points to study. However, the reviewer considers that the article needs crucial improvements throughout the manuscript before being considered for publication in the journal

State of Planet.

OVERALL COMMENTS

• The abstract should be rewritten to provide a more comprehensive explanation of all the values presented by the authors.

• One of the main shortcomings of the manuscript is the explanation of the different datasets used. To consider the article for publication, a comprehensive restructuring of the data section is necessary to address the following issues:

1. What is the source of the data?

2. What is the period during which they were used?

3. What are the temporal and spatial resolutions?

4. When and why were these data used?

All this information can be included in Table 1. I suggest including the following columns in Table 1: Variables (SWH, wave period, wave direction, etc.), temporal resolution, spatial resolution, and time span.

• The time span for the different datasets should be standardized. Sometimes the time period is from 1993 to 2022, while other times it is from 2010 to 2022, or from 2008 to 2022, or even from 2015 or 2011 to 2022. This inconsistency extends throughout the article, including the methods section

and various figures. If standardization is not possible due to the different scales analyzed, it must be specified why and reference the database being used.

• Why is the “wave forecsast model” of Puertos del Estado used? Would not it be more consistent to use the same database for atmospheric and oceanic variables (such as ERA5)?

• Another deficiency of the manuscript is the lack of consistency in calculating the 99th percentile. The authors use both the annual and monthly 99th percentile, as well as climatology (the average of each of the months, e.g., January, February, etc.), interchangeably, even though these values are

statistically different.

• The methods section should be rewritten and restructured, as the method described as "the percentile method" is essentially the peak over threshold (POT) method. Why was the 99th percentile threshold chosen as a reference instead of other values?

• The use of tables is excessive in the manuscript, making it challenging for the reader to follow the narrative. Tables 2 and 3 should be integrated into the introduction section to improve readability. Additionally, Table 4 should be removed, as the results presented there are better visualized in

Figures 1 and 3.

• The figures should be renumbered according to their order in the manuscript.

• A climatic analysis is recommended, including an examination of correlations with different climatic indices influencing the area and an analysis of temporal variability using for example, wavelet-type tools.

• The manuscript neglects the value of tides, even though the tidal range in the Mediterranean can reach up to 1 meter. However, it has been proven that the 99th percentile of the IG is 0.28 m, and of the agitation range is 0.38 m, which is within the order of magnitude of tides in the Mediterranean.

Therefore, a sensitivity study of the tidal value in the port should be conducted before neglecting this factor.

• The third major deficiency in the work is the study of extreme event trends in Melilla port. In Figure 4, it can be seen that for the area marked with a black rectangle, most of the pixels do not show a significant trend for April or July (the two months selected for a comparison between P99 and P50

in Annex 4). In my opinion, it cannot be concluded that the regression line is significant based on the time series shown in Figure 4 of the manuscript; the series exhibit too much variability.

• In this work, the analysis of wave height is detailed, while the analysis of wave period is given less attention, even though, for agitation activity, the period is more relevant than the wave height (Eq. 4). This is why in event E7, the agitation is so high compared to the time series, as the period at

that time is significantly higher than in the rest of the time series. This fact should be given more emphasis, and the atmospheric conditions that could have caused this remarkable event should be explored.

• The conclusion section could focus more on how ports need to revise their security protocols based on studies of extremes in the surrounding area, taking into account the analysis of return periods.

SPECIFIC COMMENTS

Introduction

L41. Modify the order of the tables according to when they appear in the text.

L44. Provide the link to the ECCLIPSE website.

L55. Infragravity waves have a period ranging from 25 seconds to 5 minutes, as indicated by [Munk, 1950].

L59. Table 4 could be omitted as it is redundant with figures 1 and 3.

L60. In the study area, significant wave heights (SWH) exceed 7m, the same order of magnitude than in the Gulf of Lion.

Data

L110. When does the multi-year wave product reanalysis end and the interim dataset begin?

Methodology

L129. Why if there are buoy data from 2008, do the authors choose to use them only from 2010?

L137. Which spiking method did you use? Were the gaps small enough to ensure that the time series was not totally distorted after processing?

L140. Pearson correlation coefficient.

Eq 2 and 3. Why do you use the sample variance instead of the population variance?

L155. The correct reference was Stockdon et al. (2006), not Inch et al. (2017).

L160. Specify the data that were used.

Results

L173. Specify the time span.

L180. Why do you consider data for wave directions only for the period between 2011 and 2022?

L186. How do you calculate the exceedance threshold and the time between two independent storms?

L233. Could you provide spectra to demonstrate how the infragravity waves dominate the energy during the analyzed events?

L235. It is not possible to see all these results in Table 6. Could you display them graphically?

L243. Would you mean "20 minute time-series"?

L253. Instead of "the 655 hourly", it would be clearer to mention the time span.

L268. How do you calculated the "monthly P99"? Is it the P99 of all the January data (February, March, etc.)? Or is it the mean value of all the P99 from all the January, February, etc. months?

Conclusions

L313-321. These points should be included within the introduction section.

L336. It is not the "percentile’s method", it is the peak over threshold.

Bibliography

L421. Berta, et al. (2020) should appear after Bensoussan, et al. (2019).

Annexes

Annex 3. Adjust all the colorbars, as P99 seems smaller than P50.

Annex 5. Consider removing this annex because the most of the pixels show non-significant trend values.

REFERENCES

Walter H Munk. On the wind-driven ocean circulation. Journal of meteorology, 7(2):80–93, 1950.

1. General Comments

The manuscript discusses the extreme wave event that affected the Melilla coast and harbour between the 4^th and 5^th April 2022, with unprecedented impact. The observations collected by an offshore wave buoy and by a tide gauge installed inside the Melilla port are used to characterize the extreme wave events that affected the area from 2008 to 2022, to evaluate the impact of these events on the conditions inside the port of Melilla (separating IG wave motions and agitation due to wind waves) and to show the distinctive extreme nature of the April2022 event when compared with the previous extreme events observed. Estimates of the return period for these events show the importance of retaining the April 2022 event in evaluation of return periods for the planning of coastal structures or harbour operations. A regional wave model (WAM based) for the Iberian area is used to detail the wave conditions at affecting the Melilla area on the 4^thApril2022, at the peak of the storm. ERA5 reanalysis are used to characterize the meteorological conditions associated with the extreme wave events observed in the Melilla area from 2008 to 2022. A reanalysis dataset from a regional model (WAM-based) for the Mediterranean Sea, covering the period 1993-2022, is used to provide the long term perspective of evolutions of wave conditions (in terms of significant wave height) in the Alboran Sea.

The central subject of the paper, the extreme events affecting the SW Mediterranean coast and its impacts on the harbour conditions, addresses relevant scientific and societal questions.

The manuscript title reflects the overall content of the paper, highlighting the focus on the extreme wave event that affected the Melilla harbour in April 2022.

The abstract does provide a concise and complete summary of the key questions focused, of the work developed in the paper and main results

Given the scientific and societal relevance, the new data that is explored, the important integration between observations and modelling results, it is the opinion of the referee that the paper should be considered for publication after some improvements on the structuration, presentation of matters and discussion, as indicated in the detailed analysis below.

1. Specific Comments

2.1 Text

Lines 90-95: The regional WAM-based wave forecast system described in section 2.2 is used in figure1.a to characterize the wave conditions offshore the Iberian Atlantic and Mediterranean areas and in figure 1.b to characterize the wave conditions that affected the Melilla area at 21:00 of 4 April 2022, during the peak of the storm. A more extensive use of the results from this model could have been conducted. The wave directions could be included in figure 1b, in this way illustrating eventual refraction effects that could be associated with the shelf topography in the Melilla area and that could, in some way, be relevant to interpret the offshore buoy measurements or the effects inside the harbour. Also, if results from this model are available for the global period 2008-2022, they could be used to characterize the wave field in the area during the different extreme events that were observed, perhaps providing some relevant information regarding different impacts of these events on the area of the harbour.

Line 94 – Although readers can access the product documentation indicated in Table 1 it would facilitate the reading of the paper if indication of the geographical area covered by the model and of the spatial resolution of model outputs is provided.

Line 101: The indication of the geographical position of the wave buoy would facilitate the reader.

Lines 101-102: Radar tide gauge data was, apparently, available since October 2007. Could you justify why you have not extended the analysis of these measurements to the longer period 2008-2022, that could be consistent with the analysis of the offshore wave buoy?

Lines 97-102, Section 2-3 Tide Gauges: could you better describe here how the IG and agitation bands were defined from the 2Hz data?

Line 129: “…for two different periods (a) 2010-2020 (before the record-breaking storm and (b) 2010-2022…”. Why didn’t you use the complete dataset that was available 2008-2022?

Line 131 : “…Three-Parameter Weibull…”, assure consistency with nomenclature introduced below  (here Three)

Lines 137-138: Gaps linearly interpolated no matter the gap extension or did you used a limit time gap?

Lines 137-138: Should this paragraph be moved to section “2. Data”, given that there you also provided some information regarding the use of tide gauge data to infer wave conditions inside the port?

Lines 141-143, equations (1),(2),(3) are the well know definitions of mean, standard deviation and correlation. Is it really necessary to introduce them here? Or could you just give a reference of a statistical or methods book/paper?

Line 147:

(a) “….the spectra of 2 Hz sea level oscillations measured by the tide gauge revealed…” should become “….the spectra of 2 Hz sea level oscillations measured by the tide gauge (not shown) revealed…”.

(b) Also, this phrase will fit better in section "4. Results". In the present section (“3. Methodology”) it would be more relevant to mention that spectra of the 2 Hz data was build to identify energetic sea level variability inside the port, and in the following section (“4. Results”, particularly on “4.4 Sea state inside the port”) then to refer that the spectra revealed a high energy content in the IG band. 

Line 148-149: “As the IG energy in the nearshore has been documented to be positively correlated with offshore SWH… “. The phrase seem to suggest that in the two references indicated the IG energy measured by the Melilla port tide gauge was found to be positively correlated with the SWH measured by the offshore Melilla buoy, which is not the case. The two references indicated in the text focus on the nearshore (surf zone) areas of beaches with a large variety of conditions, but are not including harbour areas.

Line 150: “….a scatter plot was computed…” . A scatter plot between what and what?

Line 154: SWH was already defined as the significant wave height, but we are using this notation both for the SWH calculated from wave measurements and for the SWH obtained from models. Perhaps it could help the reader if this is distinguish. For example, in this line we are referring to SWH calculated from wave measurements while a few lines below (e.g. line 161) we are referring to SWH obtained from models. It could help readers if a distinction is made between significant wave heights obtained from observations collected by the wave buoy (noted, for example, SWH_o ) and significant wave heights obtained from models (noted, for example, SWH_m )

Line 156:

(a) Equation (4), suggesting that IG height as runup is proportional to (SWH L)^1/2 was proposed by Stockdon et al. (2006) (their equation 18). Inch et al. (2017) instead suggested that IG wave height could be better predicted from offshore wells using the relation SWH²*T_p , instead of the previous one.

(b) The studies of Stockdon et al. (2006) and Inch et al. (2017) focus on IG motions near sloping beaches. Can you comment about the validity of using their results in an area such as the harbour of Melilla?

Line 158: In the work of Stockdon et al. (2006) the experimental conditions presented waves not exceeding more than 3.5m. In that case, the measurements from an offshore buoy located over the bathymetric of 20m could be used as characterizing the waves in deep water. In the present work, the focus are on extreme events, with waves with heights exceeding 6 or 7m. In this case, can you justify that measurements from a buoy deployed a 14m depth can indeed be used as representative of waves propagating in deep water and, particularly, that the deep water limit of the dispersion relation that is used in equation (5) can indeed be used to calculate the wavelength in equation (4)?

Line 172-179: The extreme events analysis is developed here in terms of SWH but can you also indicate the values of maximum wave height measured by the wave buoy offshore during the extreme storms?

Line 173: “… and derived from a long-term time series…” , perhaps change to “… and derived from a long-term (14 years) time series…” to become more clear when the term “long-term mean and extreme wave climate” is latter use to refer model results for the 1993 to 2022 period.

Lines 173-174: “ and derived from a long-term time series provided by a Melilla coastal buoy (Figure 1, b-c),…”. The long-term time series that is referred in the text is presented in figure 1d so this should be indicated in the text.

Line 174: “… was abruptly exceed during 42 hours.” The duration of the extreme conditions above P99 is an important aspect but the reader cannot derived it from the presented tables or pictures. This could be included as an additional column in the table that reports the main characteristics of the 7 extreme events presented in the offshore buoy time series (table inserted in figure 1). This table could be moved to the “Table” section, allowing that more information could be added.

Lines 175: “…. Coincident with a maximum value of the mean wave period (9.42s)…” these are results that are presented in figures 1d,e and so the reference to these figures should be included here.

Line 180: “… were NE … “ . To be consistent, please include the angular sector for NE as you have done for NE-E

Line 183: It would be interesting if you could include some information regarding how the direction of the extreme events compares with the prevailing incoming wave directions for the period 2011-2022.

Lines 185-188: The description of how the return periods was calculated would be better located in section 3 (Methodology), in the text block dedicated to the return period associated with the extreme episodes (starting on line 123). Here

Line 192: “…3-parameter Weibull probability…” assure consistency with nomenclature introduced above (e.g., in line 131)

Lines 188-194: This text include the results for the return periods that are associated to each one of the 7 extreme events indicated in the table that is inserted in figure1. These results are not presented in Table 5, which presents the return periods for different classes of SWH (3m, 4m …, 8m). It would be useful for the reader that the return periods calculated for each of the 7 extreme events could appear as a separated column in the table inserted in figure 1.

Line 201: Should you use SI units consistently along the text  (in this case, 2 x 10^-3 Pa/m) ? In some parts of the text the atmospheric pressure is reported in hPa, here differences in atmospheric pressure appear in Pa.

Lines 208-209: “… where easterlies blew persistently…”, could you in some way quantify what you mean persistently (how many days?) and if this persistence is similar for the different events or not?

Lines 211-212: “… a 6-week period between late February and early April (Figure 1, d).” This is only evident in the table inserted in figure 1 and not in figure 1d.

Line 214: And why is that? What are the conditions leading to the development of the low pressure system over N Africa and the high latitude position of the Azores High?. Relation with winter regimes (e.g. North Atlantic Oscillation winter )? – This could be brought to the discussion/conclusion section

Lines 215-216: This statement is not correct as can readily be seem comparing figure 2-a (event 7) with Annex1- a (events E1). Both figures show a comparable distance between the atmospheric pressure centers (perhaps event shorter for event E1) but very different SLP valus associated with these centers. It is the (atmospheric) pressure gradient, or SLP gradient, that was the relevant aspect in having a much strong wind fields (and wave generation mechanis) in E7 by regard to E1. Note that in lines 222-224, when you describe the E1 event, you are stating that the relevant factor is the SLP gradient and not the separation between the two main pressure systems.

Lines 217: “.. leading to very strong, persistent easterlies…”, how long is "persistent" here? Can you provide a quantification (how many days?)

Line 219: “… was abruptly exceeded during 42 consecutive hours (Figure 1, d).”. We cannot infer from Figure 1, d, how long did SWH was above P99. Perhaps better: “… was abruptly exceeded (Figure 1, d).during 42 consecutive hours.”.

Lines 220-221: SLP gradient units and consistency with SI units along the text, if required.

Line 232: “According to the spectra content of 2Hz data …” change to “According to the spectra content of 2Hz data (not shown) …”

Lines 232-233: As is mentioned in the caption and label of Figure 3, this figure presents the hourly time series of maximum sea level height in the IG band. So what is the sense of referring that “these oscillations are highly dominated by the IG band energy during the events..”? I guess that what you are saying is that the spectra of 2Hz data show that port variability, during those extreme events, is dominated by the IG band? Rephrase this paragraph so that this becomes clearer.

Lines 235-236: “… records of sea level oscillations (30s-1h) height…” should be changed to “… records of sea level height oscillations (30s-1h)…”

Lines 235-238: These two paragraphs are confuse. It is not from Table 6 that we see that sea level height oscillations are below or above this or that level. This is derived from the analysis of Figure 3a and this figure is not indicated in the text. Table 6 is providing the framework that relates these observations with the thresholds used to guide port management .

Line 243: “ …the analysis of 20-m timeseries of agitation…” should be corrected to “ …the analysis of 20 min averaged time series of agitation…”

Lines 246-248:

(a) The definition of agitation was given in line 156 and, based on the Stockdon et al. 2006) work, it relates to the IG wave inshore the port that are forced by the waves offshore (measured by the Melilla buoy). Here it seems that we are using the term “agitation response” to design the sea level variability inside the port, which can be linked to the penetration of waves but also to astronomical tides or storm surge. These seem to be different concepts for the term “agitation”.

(b) Why not to include also wind forcing conditions together with astronomical tide and storm surge? Storm surge was (in the text) only associated with the isostatic response of sea surface to the atmospheric pressure. Besides this effect, the wind can force upwelling or downwelling responses along the coast, which are associated with sea level changes. In the case of the extreme events reported, the easterly winds would expectedly promote strong upwelling along the Melilla coastline, leading to a sea level fall near the coast.

Line 262: “… closest to the moored buoy…” . Could you indicate at which distance from the buoy location?

Line 272: But a similar maximum of P99 is also present in February for the  buoy measurements.

Line 287: “… we select only March and July…”. Although March corresponds in fact to the maximum P99 values of SWH for the time series 1993-2022, we saw that April presented the maximum values in P99 o SWH for the period 2009-2022 and for the time series of the wave buoy measurements. So why haven’t you taken April as the representative month for the stormy season (and to show to the reader the spatial distribution of P50 and P99, in Annex 4 a,c), given that this is the month for which the discussion in this paper is centered and also that this is the month that you selected to highlight the trends in figure 4a?

Line 298: “… of 2 cm year^-1 “, again, check if  SI units should be used in the text

Lines 305-307: You state that although trend maps of P99 SWH for March and May showed relevant trends, they are not commented because they are located in areas far from the Melilla port area. But, is it not that the goal of this exercise, to examine the conditions in the Alboran Sea to try to understand what conditions or change of conditions could be leading to the development of extreme storms such as the one measured in April 2022?

General comment to 4.5 “Trends in Extreme Wave Climate”: the analysis is presented in terms of the SWH percentiles. But the conditions affecting the Melilla port are associated not only with the offshore wave height but also with the offshore wave period and direction (for the particular environment of the port, this can be seen in the correlation diagrams that you present in figure 3 b,d). Was this analysis made for the 1993-2022 reanalysis? If so, can you introduce some of these results in the discussion in section “5. Conclusions”?

Lines 322-231: This block of text is reproducing the information that you already provided in section 4.

Line 324: “…derived from the hourly time series….”, to become “…derived from the hourly time series for the period 2008-2022….”,

Line 341-342: “…this common atmospheric configuration seems to predominantly feature during the same stage of the year, a 6-weekperiod between late February and early April (Figure 1, d)”. This conclusion is not derived from figure 1-d, which only show us that the extreme wave events occur during that period, without relating them to a specific atmospheric pattern. The conclusion can be derived from the analysis of figures 2a and Annex 1 a to f that link the events to the atmospheric patterns.

General comments to “5. Conclusion” section:

Comment 1: The estimates of the return period are key to guide the design of coastal structures and the planning of port operations. Given such a huge importance, could you say something about how robust are the estimates of return period that you provided in section 4?

Comment2: "Section 5. Conclusions" could accommodate a more extensive discussion about the conditions leading to the extreme events that affect the SW Mediterranean (in particular, the Melilla area) and about our perception regarding the future occurrence of storms such as the one of 4-5 April 2022. You conducted an analysis of wave conditions (based on the SWH) fro the global Alboran Sea, characterizing the mean conditions for the 1993-2022 period and the trends for each month. It would be interesting to put these results in a more global perspective, for example by comparing them with other studies that focus the larger area of the W Mediterranean Sea, such as the study of Amarouche, Akpinar and Semedo (2022) (“Wave storm events in the Western Mediterranean Sea over four decades”, Ocean Modelling 170, 101933,2022.

2.2 References

All references indicated in the text and tables are reported in the reference list.

Correct the following references indicated in the list of references:

• “Chiggiato et al. 2023:” - line 435. Change to “Chiggiato, J., Artale, V.,de Madron, X. D., Schroeder, K.,Taupier-Letage. I., Velaoras, D., Vargas-Yáñez, M.:” (the data of the publication is correctly indicated in line 436, at the end of the reference).
• “Fanti V., Ferreira, Ó., Kümmerer, V. et al.” – line 472, not justified the use of et al. , change to “Fanti V., Ferreira, Ó., Kümmerer, V., Loureiro, C.”
• “Garcia-Valdecasas, J., Pérez Gómez, B., Molina, R. et al.” – line 474 – not justified the use of et al., in order to maintain consistence with other references present in the table
• (note: “Garrabou, J., Gómez-Gras, D., Medrano, A. et al.”- line 476 – justified the use of et al. given the extensive list of authors.)
• “Eyring….., 2021: Human….” – line 466. Remove 2021, the date of publication appears correctly at the end of reference, in line 471.

2.3 Tables

Following the comments made in 2.1 above, a table indicating the different systems/models used, providing the geographical location covered by those systems/models (e.g. mooring position or model area of coverage), the time period covered by data and the spatial/time resolution of the data  will be very helpful to the reader

Table 5 Caption: Perhaps you could indicate which parameters of Weibull distribution corresponds to what (slope, shape, threshold).

Table 6, 3^rd line: negative vale “-0.15m” should be an error and correspond to “=15 m”.

Table 6 caption, line 635. Indicate references for the thresholds presented or refer to the text.

2.4 Figures

Figure 1:

(a) The change of color scale from figure 1a to figure 1b can confuse the reader (although the color scales are clearly indicated)

(b) Table inserted in figure:

b.1 Specific question: indicate clearly what is the period (mean period, peak period,…?) and what is the direction (mean direction, peak direction,..?) that are indicated in the table.

b.2 General comment: This table would be better located in the "Tables" section. This would allow to include additional information that can be relevant (such as the return period for each event and the duration of SWH above P99) as proposed in the comments inserted in 2.1.

Figure 2.

(a) The vectors in figure 2b correspond to what? They should correspond to the wind at 10m vectors but it's confuse since the vectors lengths seem not to correspond to the wind speed color scale. Note for example that, in the figure, vectors are longer offshore the NW Spain (offshore Cape Finisterre and Galician coast), where speeds are about 15 m/s, than in the Gulf of Cadiz or in the Alboran Sea, where speeds approach 20 m/s

(b) Sea level pressure units: used the classical units used in meteorology (hPa), change if consistency with SI units along the text is required

Figure 2 Caption: “wind at 10m..” should become “… wind at 10m height”

Annex 1: Sea level pressure units: used the classical units used in meteorology (hPa), change if consistency with SI units along the text is required

Annex 2: The vectors in figure (a) to (f) correspond to what? They should represent the wind at 10m vectors but then the vector length does not correspond to the wind speed color scale. Note for example that, in figure a, the vectors along the W Moroccan coast are similar with the ones represented in the Aloran Sea although there is a strong difference in the speeds represented by the colors

Annex 3 Caption: please indicate to what corresponds the red and blue fits in figure 3a.

1. Technical Aspects

Line 122: “… the wave climate Melilla area… “ should read “… the wave climate affecting Melila area…” or equivalent.