Tsunami Alert Efficiency in the Eastern Mediterranean Sea: The 2 May 2020 Earthquake (M_w6.6) and Near-Field Tsunami South of Crete (Greece)

Round 1

Reviewer 1 Report

Introduction:

- I think it is better to start with a bit explanations of the 2^nd May event, what happened, why this event raised up awareness in tsunami warning system in this region again, etc.  

- I remember that there are some small events in the Mediterranean (i.e. Bodrum-Kos 2017/your reference no. 12) too. No such evaluation on these events?

- You may tell the readers more if there were any tsunami advisories or warnings announced in this region before the 2^nd May event.

Fig. 1: What is a meaning of yellow stars in a), not the epicenter? They are the same as the epicenter of the 2^nd May event, so may be use different symbols? Please also explain about the triangle (tide stations?) and black dots.

Method:

- Please also give full names of CENALT, INGV, NOA.

- 2.2. NEAMTWS operational architecture: Having a flow chart showing how these organizations are related will be helpful for better understanding.

- If there is already more detail about NEAMTWS in UNESCO-IOC website or books, please also add these references.

- Line 126: “Four are the possible types of tsunami alert messages:” Do you mean “There are four types of …”?

- Table 1: Just my own curious that runup also use in the criteria where most system in the world only use amplitude (because performing detail simulation for the database is difficult). How the system predict runup?

- Fig. 3: This surprised me that there are peoples kept watching the water retreat. So education is also needed together with improvement of the technical warning issues.

- Fig. 5 and 6: Font size of X and Y axis is quite small

Result:

- 3.1 Tsunami size: Can I confirm if Mt=6.87 is from which station or average from these stations? And why you selected Mw6.6 from GFZ not from others?

-3.2 Upstream alert: Can you please explain a bit more on definition of I and O message? For this, I do not understand Fig. 7 and the rest of the analysis. Please also add caption of X and Y axis of Fig. 7 and 8.

- Is it possible to add a table summarizing the timeline of messages from different organizations?

- Please also define the meaning of upstream and downstream alert

Discussion:

- I think there are several types of indicator for assessing the warning system such as warning time, predicted amplitude, predicted arrival time, timing of cancelation, etc. You only discussed and evaluated based on the warning time. Please also discuss other issues if possible.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

The manuscript “Tsunami alert efficiency in the eastern Mediterranean Sea: the May 2, 2020 earthquake (Mw6.6) and near-field tsunami south of Crete Isl. (Greece)” assess the performance of the current Mediterranean tsunami warning system using the 2020 Mw 6.6 earthquake in the Hellenic Subduction Zone. The authors evaluate the upstream and upstream components of the different local agencies in terms of time and accuracy. The authors conclude that the main issue with the warning system is the near-field character of most relevant tsunami events which challenge the relatively long forecast time and insufficient accuracy of the warning system.

Here, I present some ideas which I think are very important to be included and addressed by the reviewers:

• The authors should elaborate more about the monitoring tide-gauge system. If we see Figure 5, the tsunami wave periods are very close to the resolvable temporal resolution of the tide gauge (Nyquist frequency). Thus, I think the tide gauges configuration may be inadequate for the observed tsunami waves (e.g. the 2018 Palu tsunami was apparently not accurately measured by a local tide gauge).
• The existing tsunami warning system seems to focus in tsunamis directly generated by the co-seismic deformation of earthquakes. However, I do not see any concern or comment about tsunamis generated by earthquake-triggered landslides or volcanic eruption. The authors should convey ideas of these additional tsunami sources in the manuscript. It seems to me the current warning process would fail. A good example are the 2018 Palu and Krakatau tsunamis.

Some specific comments:

• Please add more information about the HSZ (plates, convergence). Then, you can describe the earthquakes and tsunamis. Please, also present all past tsunami events in the introduction (i.e., lines 431-435 into introduction). We need more information about past earthquakes and tsunamis to properly weigh the significance of this paper in the first pages.
• Recent tsunamis have shown the important role of tsunami wave propagation and final inundation response (e.g. edge waves in Mexico 2020). Especial characteristics of tsunamigenic earthquakes for a given magnitude may also prompt the generation of exceptionally large tsunamis (e.g. tsunami earthquakes). Hence, I will better elaborate the idea of using the tsunami intensity measure M_t in the paper. I am concerned that readership believe this is the parameter the agencies should use for tsunami warning (which in this particular event worked well).
• I don’t think the tsunami alert efficiency, F, is a pertinent and useful parameter for warning. As a user, I am more concern how many minutes I have for evacuation (e.g. F~1 when this is associated with f-t=3 minutes is still not good). Please, provide a better justification to F or provided limitations as compared to f-t.
• Throughout the text. Please provide country or pertinent TSP of the stations which are described in the manuscript. Not everybody is familiar with the Mediterranean Sea.
• In line 383-384 it is mentioned that “local authorities decided to keep people away from the beach”. This is indeed as bad as Arvi experience (i.e. no action). A late action of authorities at the beach may expose them and civilian to the tsunami inundation. I suggest to emphasize these relevant problems which seem to be related with awareness.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

In the manuscript by example of the May 2, 2020 event the efficiency of tsunami warning system is analyzed. Along with some interesting details regarding this event, authors introduce and examine tsunami alert efficiency parameter. The manuscript can be published in GeoHazards after minor revisions.

List of suggested corrections.

Lines 77-78. The phrase «The 1 July 2009 tsunami was reported from Arvi, Myrtos and Chrysi islet.» seems to be inserted in the Figure caption by a mistake.

Lines 104, 116 etc. Please specify which magnitude is used (Mw, Ms?).

Lines 160-161. Sentence is misleading. Please change it to «From eyewitness accounts it results that the tsunami of 2 May 2020 was similar to the one reported after the earthquake 1 July 2009.»

Line 163. «No tide-gauges were operating in the area at the time.» At what time: 2009, 2020 or both?

Lines 223-224. Please specify that you mean records provided nearshore tide-gauge stations, but not deep-water stations.

Figure 8. The axes must be labeled. It would be much better to plot F versus ln (f). Regression curves should be shown for all 4 cases. Legends for regression curves should be depicted. It would be interesting to extrapolate the regression curves into domain of negative values of F.

Figure 9. The axes must be labeled. It would be much better to plot F versus ln (f). Regression curves should be shown for all 3 cases. Legends for regression curves should be depicted.

Author Response

Please see the attachment. 

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

I confirmed that the authors have properly replied and responsed to my questions and comments. Please make sure about the definition of definition of tsunami (runup, amplitude, etc) so that readers will not misundesrstanding.