New Observational Material about Seismic and Non-Seismic Tsunamis in Greece and Surrounding Areas from 1900 to 2023

Round 1

Reviewer 1 Report

The manuscript illustrate significant new information on last ca. 124 years tsunami effects in the Egean region. The presentation is fine and well written. I have two comments. One is about the Mw resolution. For historical earthquakes, before say year 1960, Moment Magnitude values should be given with one digit after the coma. Second comment, please consider the effects of coseismic coastal uplift in your measurements of tsunami flooding during strong ca M7 events. This is described for the 1953 earthquake, but i see very similar evidence for the 1958 earthquake based on your Figure 4. Please add a discussion on this point. 

Detailed specific comments are included in the annotated manuscript.

Comments for author File: Comments.pdf

Author Response

The revised manuscript of our paper has been prepared by following the instructions by the Editorial Office and most of the recommendations given by the Reviewers. We appreciate a lot the comments submitted by the three Reviewers since have been much helpful for the improvement of the initial manuscript. Our response is analyzed in details in the next lines.

Reviewer 1.

Comment. One comment is about the Mw resolution. For historical earthquakes, before say year 1960, Moment Magnitude values should be given with one digit after the coma.

Reply. We understand the concern expressed by the reviewer. However, as noted in section 2.2. (now 3.2. Tsunami observations of the initial ms  “The earthquake origin times and magnitudes are according to the determinations in the catalogue by ISC-GEM [2] unless otherwise reported”. The ISC-GEM [2] magnitude determinations are given with two decimal digits and, therefore, we keep it.

To make it more clear, and to satisfy the reasonable comment by the reviewer, in the revised version we added the next clarification: “Earthquake magnitudes listed in the ISC-GEM [2] catalogue have been re-computed uniformly at large extent and with procedures detailed by Giacomo et al. (2015). Magnitudes reported by ISC-GEM [2] have been reproduced in the present study without changes, e.g. with two decimal digits. As a rule the uncertainty involved in the magnitudes calculated by ISC-GEM [2] as a rule is +/- 0.2”. 

Comment. Second comment, please consider the effects of coseismic coastal uplift in your measurements of tsunami flooding during strong ca M7 events. This is described for the 1953 earthquake, but Ι see very similar evidence for the 1948 earthquake based on your Figure 4. Please add a discussion on this point. 

Reply. In response to this comment we inserted relevant short explanations in lines 311-313 of the revised submission and in the legend of figure 5.

Comment. Detailed specific comments are included in the annotated manuscript.

Reply. Practically speaking, these detailed comments are focusing on issues related to the two main comments. However, we have posted detailed replies and explanations in each one the comments expressed by the reviewer in his/her annotated manuscript.

Reviewer 2 Report

see attachment

Comments for author File: Comments.pdf

Minor editing of the English form is required. 

Author Response

The revised manuscript of our paper has been prepared by following the instructions by the Editorial Office and most of the recommendations given by the Reviewers. We appreciate a lot the comments submitted by the three Reviewers since have been much helpful for the improvement of the initial manuscript. Our response is analyzed in details in the next lines.

Reviewer 2.

Comment. A section on Geological Setting is lacking and should be added.

Reply. We respect the opinion of the Reviewer. A Geological Setting section would really be useful. In view of that we present tsunami observations for the entire Greek region we considered that it would be quite difficult to describe the geological framework of such a highly complicated region from geological point of view.

We decided to comply with the reviewer’s suggestion by introducing a new section named “2. Tsunamigenic sources in the Greek region” as an alternative approach. In this section we describe briefly the main tsunamigenic zones and how they are related with the main geotectonic processes in the region. In addition, we include a map illustrating the main tsunamigenic zones as published by Papadopoulos et al. (2014), a citation which has been requested by the reviewer to include in the revised ms anyway. As a consequence, the present-day section 2 has taken number 3 and so forth.

Comment. The subsection “2.2. Tsunami Observations” must be shifted in the Results and the present-day location of the paragraph on Materials and Methods is not proper. The subsection is also very long. The authors must do an effort to reduce its length.

Reply. The subsection 2.2. Tsunami Observations contains new observational material, i.e. a set of new data. Therefore, its natural position belongs within the section “2. Materials and Methods”.

On the other hand, we have made an effort to reduce the length of “2.2. Tsunami Observations” as suggested by the reviewer.

Comment. If the compilation section (2.2 Tsunami Observations) is prevalent over the new results on the observation of tsunami, as it happens, I suggest to classify this paper as a Review Article, not as a Research article.

Reply. The strong point of this paper is that it presents for the first time plenty of new tsunami observations, which is also adopted by the reviewer in its next comment. These new observations were compiled after research based on “official reports, local archives and books, press reports, eyewitness accounts, pictorial and video material, tide-gauge records as well as from our field surveys”. Therefore, we do not share the opinion of the reviewer.

Comment. Accordingly, the title of the manuscript must be modified as “A compilation of tsunamis in Greece: new observational data from 1900 to 2023”.

Reply. The suggestion of the reviewer did not take into account two important elements of our title of the paper: (a) “Greece and Surrounding areas”, (b) seismic and non-seismic tsunamis. The first element reflects the broad geographical distribution of the audience that would be interested about the paper. The second element reflects the interest that this paper may have to both the communities dealing with seismic and non-seismic tsunamis.

Comment. The authors must clarify better, what in this paper is compilation and what are new results, clarifying which is their new contribution in the work of compilation of tsunamis in Greece.

Reply. In the section “3.1 Method Outline” we pointed that “Most of the identified tsunamis remained unknown or little-known so far to the tsunami community. In about the last 15 years several tsunamis have been reported in the region under study. Some of them have already been studied in specific papers (Table 1, Figure 1) and, therefore, our compilation did not include those tsunamis”. This paragraph makes crystal clear that the new tsunamis analyzed in our paper have been revealed by our own research and that already known and studied tsunami events have not been considered in our paper. In addition, in the same section we have clarified that “Based on the collected material we characterized each single event with a reliability index. Additionally, we assigned a respective tsunami intensity when sufficient observational data are available”, which implies that the reliability index and the intensity assigned to each one of the tsunami events examined has been made by us.

Following the reasonable suggestion by the reviewer, in the revised version the section “Introduction” has been enriched by an additional clarification which reads as follows: “Three out of 26 individual events examined are referring to already known seismic tsunamis: (a) 8 November 1905 in Mt Athos, northwestern Aegean Sea, (b) 7 February 1948 in Karpathos Isl., southeastern Aegean Sea, (c) 24 February 1981 in eastern Corinth Gulf. These three tsunamis have been included in the current investigation for three different reasons. The 1905 event has been included with the aim to attract the interest of the reader about the cascade effect which characterized the event. Namely, the earthquake caused landslides and rock falls, which triggered the tsunami by entering the sea water. The well-known 1948 tsunami remains controversial as regards the generation mechanism, i.e. seismic versus landslide. We present new observations collected during our field survey as well as new eyewitness accounts, which may support future studies to better understand the generation mechanism. The small tsunami of 1981 has been documented from eyewitness accounts as well as from a tide-gauge records. However, this record remains very little-unknown since it was published in a local journal. Here we reproduce that mareogram for the sake of the international community”.

Comment. Accordingly, I would expand the introduction, taking into account a broader range of references, including Papadopolous et al. (2014) and references therein. These authors have analyzed the rift of Corinth Gulf, Central Greece, which is highly prone to tsunamis. Regardless the tsunami size, this rift structure is characterized by the highest rate of tsunami production in the European–Mediterranean region. This is explainable by the concurrence of several favoring factors such as high seismicity, susceptibility to coastal/submarine landsliding and steep bathymetry. The tsunami rate is higher in the west side of the Gulf which is consistent with that the seismicity rate is also higher than on the east side.

Reply. We agree with the reviewer and the reference Papadopolous et al. (2014) was added in the new section “2. Tsunamigenic sources in the Greek region”, which has been created in response to a previous suggestion by the reviewer. The suggestion for adding other references is discussed in a later similar comment by the reviewer.  As regards the special note focusing on the tsunami potential of the Corinth Gulf it has been done in the new section 2.

Comment. The geological implications of tsunamis of Greece on the surrounding marine environment are poorly or not explained, referring in particular to submarine landslides. In the discussion, please add a section on these geological implications. Run up? Landslides? Please discuss.

Reply. This issue has been included in the section Results.

Comment. In their compilation work on tsunamis, the authors must consider the previous catalog existing on tsunamis by Tinti et al. and Maramai et al. (University of Bologna and INGV, Italy). Has it been considered and included in the compilation paragraph? Please answer and revise.  

Reply. Since our paper is focusing on tsunamis that occurred in Greece after 1900, it would be out of the scope of the paper to expand citations either to other parts of the Mediterranean or to historical catalogs, i.e. before 1900. The catalog by Maramai et al. (2014), which is referring to tsunamis in the entire Mediterranean, including Greece, is already cited. The contribution by S. Tinti to the study of Mediterranean tsunamis is internationally recognized. To our knowledge this author did not publish catalog(s) referring to Greek tsunamis as he did for Italy. Nevertheless, following the suggestion by the reviewer we found appropriate to quote the paper by Zaniboni et al. (2014), co-authored by S. Tinti, which is focusing on a Greek landslide tsunami.

Comment. Some figures must be merged, as the maps showing the epicenters. Possibly, these maps can be collected in a unique map, more regional in scale, with inset shows the epicentral maps. This would reduce the number of figures, which is quite high.

Reply. This is a good idea. However, merging the two figures in one is technically inappropriate for the reason that so many elements should be plotted in the new figure that would result in a complete confusion.

Reviewer 3 Report

This study collected a variety of information sources for tsunamis occurring in Greece and its surrounding areas from 1900 to 2003. Based on the collected material, they characterized each single event with a reliability index.

For the study of historical tsunami events, the most reliable method is geological surveys, which involve assessing tsunami deposits to determine the extent of the tsunami’s impact, supplemented by numerical simulations for comparative analysis. Eyewitness testimony is a valuable avenue, but relying solely on it has limitations. Please note that when I refer to eyewitness testimony, I am using it in a broad sense, which includes reports, books, eyewitness accounts, as well as pictorial and video evidence.

Frankly speaking, I am not denying the validity of the author’s approach, as their work remains meaningful. However, I suggest that the author, at the very least in the cases presented, incorporate some comparisons with geological surveys to enhance the credibility of the method. Hence, my recommendation is major revision.

Line 78: These three reliability indicators are affected by time, with descriptions of older events becoming less clear and fewer information sources reporting the events. How should this issue be considered during analysis?

Line 106: I find it hard to consider that am Mw 5.7 magnitude earthquake would trigger a storm-like wave. The author should explain why they categorized it as R=3. Could this be due to triggering landslides, or simply because there happened to be a storm on the same day?

Line 151: Please translate this sentence into English.

Line 182: Similarly, why would an Mw 6.43 magnitude earthquake cause such a high tsunami and such extensive inundation? Is it possible that the magnitude record is incorrect?

Line 229: In the field trip, did you only interact with witnesses and testimonies? Why not conduct some research on tsunami deposits? For an ancient event, relying solely on human testimony may lead to significant errors.

Line 388: How was this rock uplifted? Was there originally a portion submerged in water that was lifted? Please have the author provide a detailed description of the extent of the uplift based on eyewitness.

Line 526: What is the wavelength of this tsunami, and what evidence is there to prove it was caused by a submarine landslide?

Minor editing of English language required.

Author Response

The revised manuscript of our paper has been prepared by following the instructions by the Editorial Office and most of the recommendations given by the Reviewers. We appreciate a lot the comments submitted by the three Reviewers since have been much helpful for the improvement of the initial manuscript. Our response is analyzed in details in the next lines.

Reviewer 3

Suggestion. For the study of historical tsunami events, the most reliable method is geological surveys, which involve assessing tsunami deposits to determine the extent of the tsunami’s impact, supplemented by numerical simulations for comparative analysis… I suggest that the author, at the very least in the cases presented, incorporate some comparisons with geological surveys to enhance the credibility of the method.

Reply. Geological surveys concluding to the discovery of tsunami sediment or other geological evidence is in fact a productive method for the study of large-size historical events. However, we study tsunamis occurring in the modern period, which implies that the preservation of geological evidence, if any, is very limited due to modern human interventions (e.g. tourist activities in coastal zones) usually do not allow preservation of tsunami geological evidence. In addition, most of the tsunamis studied have not been of so large size as to leave behind clear geological evidence.

In the examined period of 1900-2023 two large and destructive tsunamis occurred in the Greek region. The first is the one of 9 July 1956 in the south Aegean which has not been included in the list of the tsunamis we studied since it is already well-known. The second large tsunami occurred in 7 February 1948 in Karpathos Isl., SE Aegean Sea. This event is also well-known but we included it in our study since we found new accounts. In addition, during our field survey in Karpathos we found and analyzed new tsunami observation points. However, we were unable to find geological evidence for this tsunami.

Following the reviewer’s suggestion, we added a relevant short note in the sub-section 3.2 Tsunami Observations.

Comment. Line 78 (now 130): These three reliability indicators are affected by time, with descriptions of older events becoming less clear and fewer information sources reporting the events. How should this issue be considered during analysis?

Reply. The comment is reasonable for old events. In the text we clarify that the main criteria for assigning a reliability index to a particular event are the credibility of the information source, the description of the phenomenon itself, and the number of individual information sources reporting the event. These three criteria apply to older and recent events.  Our new observations cover the most recent time interval of 1900-2023 and therefore the reliability indicators are more credible with respect to pre-1900 historical events.

Comment. Line 106 (now 173): I find it hard to consider that a Mw 5.7 magnitude earthquake would trigger a storm-like wave. The author should explain why they categorized it as R=3. Could this be due to triggering landslides, or simply because there happened to be a storm on the same day?

Reply. The tsunami community is well aware that the sea disturbances caused by earthquakes often are described by non-experts as “stormy sea” although in most cases they are tsunamis. The particular observation comes from an official source. However, the observation is not verified by another source and therefore we did not characterize the event with reliability score 3 instead of the maximum 4.

Comment. Line 151 (now 219): Please translate this sentence into English.

Reply. Done. 

Comment. Line 182 (now 256): Similarly, why would an Mw 6.43 magnitude earthquake cause such a high tsunami and such extensive inundation? Is it possible that the magnitude record is incorrect?

Reply. As explained in the text, the information about this supposed tsunami has not been authenticated. For this reason, we characterized it as of reliability only 1, meaning “improbable tsunami”.

Comment. Line 229 (now 304): In the field trip, did you only interact with witnesses and testimonies? Why not conduct some research on tsunami deposits? For an ancient event, relying solely on human testimony may lead to significant errors.

Reply. The reviewer refers to the 1948 strong tsunami. During our field survey we checked for tsunami deposits. However, we were unable to find geological evidence as explained in 3.2.

Comment. Line 388 (now 468 ): How was this rock uplifted? Was there originally a portion submerged in water that was lifted? Please have the author provide a detailed description of the extent of the uplift based on eyewitness.

Reply. We already provided the eyewitness account that after the earthquake and after the sea oscillation terminated, they observed that “the water level remained at least 40 cm lower since then until today”. There is no doubt the area uplifted permanently due to the earthquake. This has been verified and supported in a relevant scientific paper published in 1955 [47] in which a photograph of the uplifted rock was also published. During our recent field survey, we observed the area and verified once more the co-seismic uplift (Fig. 7).

Following the comment, we added in Fig. 7 an arrow showing the sea level before the uplift of 1953 and a second arrow showing the sea water level before an earlier co-seismic uplift in the same area.   

Comment. Line 526 (now 607): What is the wavelength of this tsunami, and what evidence is there to prove it was caused by a submarine landslide?

Reply. No evidence is available about the wavelength of this tsunami.  In the text we have included information which is relevant to the reviewer’s question: “Three waves were reported there arriving a few minutes apart each other”.

We cannot prove that it was caused by a submarine landslide. However, we suggested that it is the only remaining possible mechanism for the tsunami generation after excluding the other possible ones. Namely, in the text we say: “No earthquake was registered before or after the tsunami, while port authorities verified that no sizable boat sailed in the area at that time. On the other hand, the weather was reported calm. We suggest that the wave was likely generated by an aseismic submarine landslide. Τhe area is prone to aseismic coastal and submarine landslides causing local but powerful tsunamis, such as the ones that occurred in 1963 and 1996”.

Round 2

Reviewer 2 Report

the paper has improved and can be published

Reviewer 3 Report

I thank the authors for their careful response to my comments. It is a pity that they were unable to find geological evidence for some tsunami events, yet I accept the reasons. I also appreciated that they modified the Figure 7 to show the sea level change due to the co-seismic effects, making the evidence more convincing. I believe it is now suitable for publication in GeoHazards.