Synoptic patterns associated with high-frequency  sea level extremes in the Adriatic Sea

Ruić, Krešimir; Šepić, Jadranka; Vojković, Marin

doi:https://doi.org/10.5194/os-21-1183-2025

Articles | Volume 21, issue 3

https://doi.org/10.5194/os-21-1183-2025

© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.

Special issue:

Special issue on ocean extremes (55th International Liége...

https://doi.org/10.5194/os-21-1183-2025

© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 21, issue 3

Research article

|

01 Jul 2025

Research article |

| 01 Jul 2025

Synoptic patterns associated with high-frequency sea level extremes in the Adriatic Sea

Krešimir Ruić, Jadranka Šepić, and Marin Vojković

Download

Final revised paper (published on 01 Jul 2025)
Supplement to the final revised paper
Preprint (discussion started on 04 Dec 2024)
Supplement to the preprint

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2024-3711', Anonymous Referee #1, 02 Jan 2025

The study focuses on identifying synoptic conditions leading to extreme high-frequency sea level oscillations in the Adriatic Sea. Extreme episodes were classified into two types: HF extremes and Compound extremes, using sea level data from six tide gauge stations. Using the k-medoid clustering method on ERA5 reanalysis data, characteristic synoptic patterns were identified for both types of extremes. These patterns included "summer-type" and "winter-type," as well as a third "Bora-type" pattern for strong northeast winds. The data was divided into training and testing sets, with the extracted medoids of clusters used to label days in the testing period. It was found that extreme episodes occurred during synoptic situations similar to the identified medoids. The study discusses the potential of using k-medoid clustering to predict high sea level rise events and displays evidence of particular synoptic patterns related to such events.
General comments: The study furthers the work of Ruić et al (2023) where this manuscript introduces us to a third type of synoptic situation and takes our assumptions outside of the winter/summer parameterisation for these types of events. It also considers the concept of compound extremes which is frequently overlooked. The scientific methods used are valid, with the authors displaying due diligence in their choice of method used. It contains an excellent definition and summary of previous work on high frequency sea level extremes. A good use of the English language requiring only a few alterations (as listed below), fluent and precise. The figures are of a suitable amount, clear and concise where you can easily distinguish between seasonal patterns. Overall, a well-structured, concise, thought provoking and valid contribution to the growing field of understanding and forecasting high frequency extremes (meteotsunami).
Specific comments:
Line 207: Don’t need to include, just curiosity on my part but any thoughts on why the events are more abundant between 2016 and 2021at Rovinj?
The conclusion may benefit from bullet points summarising the key findings of the work.
Again, just curiosity on my part, but how did the testing period fair against the training period? Did it meet your expectations? I know it is difficult to say given the small number of events, maybe a better question should be do you expect it to perform to trend in the future with a longer-term dataset?
Technical corrections:
Line 32: Spell check ‘centimeter’
Line 200: Fig 3, move a) and b) on graphs to be more easily seen and maybe reword the latter part of the description
Line 273-4: Punctuation needed “Bora; Grisogono”
Line 298-99: Spell check “isobars”
Line 338: Should A2 be both a transitional and a winter type event?
Line 409: “or” should be “our”
Line 415 and 417: Reword as a little confusing
Line 421: “other” should be “others”
Line 423: “clusters” should be “cluster”
Line 426: Reword as a little confusing
Line 439: “even” should be “event”

Citation: https://doi.org/10.5194/egusphere-2024-3711-RC1
- AC1: 'Reply on RC1', Krešimir Ruić, 07 Feb 2025
  
  We thank the reviewer for a thorough review of our manuscript and for positive and constructive comments. We will address all comments, make the suggested corrections and add the necessary additional explanations to the manuscript.
  Specific comments:
  Comment: Line 207: Don’t need to include, just curiosity on my part but any thoughts on why the events are more abundant between 2016 and 2021 at Rovinj?
  Answer: Thank you for raising this important question and drawing our attention to this issue. Following your comment, we have contacted the provider of the sea level data, the Hydrographic Institute of the Republic of Croatia. Apparently, there was a change in the characteristics of the stilling well in 2016 and 2017. The connecting pipe first broke during a storm in 2016 and was then replaced in 2017 with a new connecting pipe with a different shape and different attenuation properties (compared to the original pipe). Both changes probably resulted in weaker damping of the high-frequency oscillations in the stilling well. This is now noted in the manuscript.
  Despite the changed properties, we decided to keep the Rovinj station in our manuscript (and in the supplementary figures) because: (1) more than half of the extreme events detected at Rovinj in the period 2016-2021 were also detected at other tide gauge stations; (2) the k-medoid analysis provided results that are consistent with the results for other stations, especially with the results for the closest station in Bakar Bay.
  
  Comment: The conclusion may benefit from bullet points summarising the key findings of the work.
  Answer: We agree, we will add bullet points.
  
  Comment: Again, just curiosity on my part, but how did the testing period fair against the training period? Did it meet your expectations? I know it is difficult to say given the small number of events, maybe a better question should be do you expect it to perform to trend in the future with a longer-term dataset?
  Answer: We are partially satisfied. Almost all "extreme days" in the test period had high similarity scores - i.e. the synoptics resembled those of the characteristic medoids. Nevertheless, the predictive potential appears to be relatively low. Following your comments, we will list points that could help to improve the methodology, including the suggested extension of the training set.
  
  Citation: https://doi.org/10.5194/egusphere-2024-3711-AC1
- AC2: 'Reply on RC1', Krešimir Ruić, 07 Feb 2025
  
  Technical corrections:
  Comment: Line 32: Spell check ‘centimeter’
  Answer: Corrected.
  
  Comment: Line 200: Fig 3, move a) and b) on graphs to be more easily seen and maybe reword the latter part of the description
  Answer: We will move a) and b) on the graphs and rewrite the second part of the caption.
  
  Comment: Line 273-4: Punctuation needed “Bora; Grisogono”
  Answer: Corrected.
  
  Comment: Line 298-99: Spell check “isobars”
  Answer: Corrected.
  
  Comment: Line 338: Should A2 be both a transitional and a winter type event?
  Answer: Indeed not. Thank you for noticing this. A2 should just be “winter-type”. We will correct it.
  
  Comment: Line 409: “or” should be “our”
  Answer: Corrected.
  
  Comment: Line 415 and 417: Reword as a little confusing
  Answer: We will rewrite the sentence.
  
  Comment: Line 421: “other” should be “others”
  Answer: Corrected.
  
  Comment: Line 423: “clusters” should be “cluster”
  Answer: Corrected.
  
  Comment: Line 426: Reword as a little confusing
  Answer: We will rewrite the sentence.
  
  Comment: Line 439: “even” should be “event”
  Answer: Corrected.
  
  Citation: https://doi.org/10.5194/egusphere-2024-3711-AC2
RC2:
'Comment on egusphere-2024-3711', Anonymous Referee #2, 11 Jan 2025

The study makes a significant contribution to understanding the occurrence of extreme sea levels in the Adriatic Sea. Drawing on data from six tide gauges in this basin and meteorological data provided by the ERA5 reanalysis, it seeks to identify the synoptic conditions that drive extreme episodes of high-frequency sea level oscillations (< 2 h), either independently or in combination (compound events) with lower-frequency extremes (> 2 h) primarily caused by storm surges. The manuscript advances previous research in the area by employing the k-medoid clustering method to clasify the meteorological patterns and obtain the characteristic synoptic conditions associated with both types of extremes. They are able to identify not only the well known summer-type pattern responsible for meteotsunamis but also a winter-type pattern linked to compound events, and even a third novel one (Bora-type pattern) characterized by extreme northeast winds.
General comments
Building on the previous work of Ruic et al., 2023, the authors take a significant step forward in identifying the meteorological conditions driving extreme sea level events recorded by coastal tide gauges in the Adriatic Sea. In line with other recent studies, comprehensively and well described in the introduction, they highlight the critical role of higher-frequency sea level variability in achieving a more accurate assessment of extreme coastal sea level hazards. This was not attainable in the past due to lack of sea level data with sufficient temporal resolution (<= 1 min). The upgrade of the observing networks to monitor tsunamis over the past 15 years will now enable similar studies also in other regions. I think this could lay the groundwork for a more effective implementation of early warning systems, including predictive capabilities, in the region, as the authors correctly describe in the conclusions section.
The study is rigourous, with the methodology clearly described and well justified. It includes a detailed description of the sea level data quality control and processing, and a well reasoned discussion on the selection of the k-medoid method and its advantages over other well-known classification methods. The manuscript is well written and structured, with a concise and correct English, and with an adequate number of high-quality relevant figures.
Specific comments
I find particularly important in this study the emphasis on the compound events. Despite it is well justified by the authors that the focus of this work is on the combination of meteotsunamis and storm surges, I would find interesting a brief mention to other potential contributors to total sea level such as the tide or the seasonal cycle, either in the introduction or in the discussion. Perhaps a brief description of the tide characteristics and amplitude in this area could be added. Information about which of the stations used could be presenting amplification of sea level oscillations due to more local effects such as harbour resonance would also be interesting. Finally, I recommend to include in the discussion the convenience of future studies which could better identify the contribution of waves (infragravity waves), the limitations of 1-min sampling for this objective, and recommendations to solve this in the implementation of sea level networks.
Do the authors have an explanation for the increased number of events in Rovinj between 2016-2021? This kind of analysis will become particularly important, as the time series grow, to analyse potential changes driven by climate change.
Technical corrections
Line 165: typo: “To determine medoids”
Line 439: typo: “ in the background of each event”

Citation: https://doi.org/10.5194/egusphere-2024-3711-RC2
- AC3: 'Reply on RC2', Krešimir Ruić, 07 Feb 2025
  
  We thank the reviewer for a thorough review of our manuscript and for positive and constructive comments. We will address all comments, make the suggested corrections and add the necessary additional explanations to the manuscript.
  
  Comment: I find particularly important in this study the emphasis on the compound events. Despite it is well justified by the authors that the focus of this work is on the combination of meteotsunamis and storm surges
  Answer: Thank you!
  
  Comment: I would find interesting a brief mention to other potential contributors to total sea level such as the tide or the seasonal cycle, either in the introduction or in the discussion. Perhaps a brief description of the tide characteristics and amplitude in this area could be added.
  Answer: We will add (in the introduction) additional information on the processes leading to the extreme sea levels of the Adriatic, especially tides, but also other longer period changes/oscillations, including seasonal signal, annual, decadal and climate scale changes.
  
  Comment: Information about which of the stations used could be presenting amplification of sea level oscillations due to more local effects such as harbour resonance would also be interesting.
  Answer: We will add this information in our Discussion section
  
  Comment: Finally, I recommend to include in the discussion the convenience of future studies which could better identify the contribution of waves (infragravity waves), the limitations of 1-min sampling for this objective, and recommendations to solve this in the implementation of sea level networks.
  Answer: we agree. We will include a few lines about the contribution of wind waves, swell and infragravity waves to the Adriatic sea level extremes into Introduction. In the discussion we will address the importance of including higher sampling rate measurements in sea level networks.
  
  Comment: Do the authors have an explanation for the increased number of events in Rovinj between 2016-2021? This kind of analysis will become particularly important, as the time series grow, to analyse potential changes driven by climate change.
  Answer: Thank you for raising this important question and drawing our attention to this issue. Following your comment, we have contacted the provider of the sea level data, the Hydrographic Institute of the Republic of Croatia. Apparently, there was a change in the characteristics of the stilling well in 2016 and 2017. The connecting pipe first broke during a storm in 2016 and was then replaced in 2017 with a new connecting pipe with a different shape and different attenuation properties (compared to the original pipe). Both changes probably resulted in weaker damping of the high-frequency oscillations in the stilling well. This is now noted in the manuscript.
  Despite the changed properties, we decided to keep the Rovinj station in our manuscript (and in the supplementary figures) because: (1) more than half of the extreme events detected at Rovinj in the period 2016-2021 were also detected at other tide gauge stations; (2) the k-medoid analysis provided results that are consistent with the results for other stations, especially with the results for the closest station in Bakar Bay.
  In view of this, we consider the series too short to discuss the effects of climate change. Nevertheless, we agree that it is important to assess the change in high-frequency extremes in a changing climate – and for this reason to ensure the continuation of high-resolution sea level measurements. We will emphasise this in the discussion.
  
  Technical corrections
  Comment: Line 165: typo: “To determine medoids”
  Answer: Thank you, corrected.
  
  Comment: Line 439: typo: “ in the background of each event”
  Answer: Corrected
  
  Citation: https://doi.org/10.5194/egusphere-2024-3711-AC3

Peer review completion

AR: Author's response | RR: Referee report | ED: Editor decision | EF: Editorial file upload

AR by Krešimir Ruić on behalf of the Authors (24 Feb 2025) Author's response Author's tracked changes Manuscript

EF by Mario Ebel (24 Feb 2025) Supplement

ED: Referee Nomination & Report Request started (03 Mar 2025) by Matjaz Licer

RR by Anonymous Referee #1 (08 Mar 2025)

RR by Anonymous Referee #2 (31 Mar 2025)

ED: Publish as is (31 Mar 2025) by Matjaz Licer

AR by Krešimir Ruić on behalf of the Authors (31 Mar 2025)

Short summary

This study investigates the synoptic weather patterns that cause extreme high-frequency sea level oscillations in the Adriatic Sea. Using synoptic data, 17 years of tide gauge data, and advanced clustering techniques, we identify distinct weather patterns linked to these events, some of which were previously unknown. These insights improve the understanding of sea level variability and have potential applications in forecasting coastal hazards.

Synoptic patterns associated with high-frequency sea level extremes in the Adriatic Sea

Download

Interactive discussion

Peer review completion

Suggestions for revision or reasons for rejection