Stochastic Generator of Earthquakes for Mainland France
Round 1
Reviewer 1 Report
Dear Editor
I reviewed the manuscript " Stochastic Generator of Earthquakes for France mainland" submitted to Applied Sciences by Gouache et al.
I m afraid that the manuscript suffers from many major/fundamental weaknesses that has to be rejected.
Overall, I often try to offer some ideas so that the authors can improve their weaknesses in a way that the submitted manuscript can be saved. However, I see no such possibility for this submission. This is a rather simplistic and old probabilistic approach that ignores important geological constrains and can severely mislead readers who are not specialists. The authors try to generate synthetic earthquakes by propagating information predominantly from past historical seismicity. However, beyond all other issues the most important thing is that their statistical sample is far from complete. Firstly their statistical sample is only a small portion of the required one and secondly they disregard the physical basis of the seismic cycle on individual faults.
Even the authors recognize this in their intro “seismicity is considered as low to moderate due to its high return periods”. Anyone who has basic knowledge on statistics realise that such cases are difficult and you need to have a long observation period much longer than the return periods of the earthquakes you study. However, despite the fact that they acknowledge this, their data input and their overall methodology simply disregard it. Their faults have recurrence intervals of several thousand of years, but their seismic catalogue is only complete for 5.8 events for only 400 years and for M>6.0 for only 700years. In Europe, we are aware that earthquake catalogues are complete for strong events >5.8 not more than 500 years. In addition, the authors do not refer to the major uncertainties from historical events that are up to 0.5 of magnitude! and up to 40-50km radius of the estimated epicentre. These are also major uncertainties that are not communicated to the readers as they should. Stucchi et al., 2013 paper that combines all the historical records of Europe clearly communicates and quantify this.
Overall what the authors recognize is only a small fraction of faults that have been activated over the last few hundreds years, therefore they miss the majority of the active sources! In addition, these faults since they have recently been activated, they DO NOT pose a threat since they are early in their cycle. On the other hand, the ones that are missing from the catalogues are approaching the end of their cycle. However, these faults are absent from their catalogues. This is what Scholz 2002 has been communicating that such approach will clearly present an erroneous view of present day hazard (see also McCalpin 2009). So overall, the authors miss the majority of the faults and also miss all the important ones that are expected to rupture in the future!
Despite all the above the authors do not communicate any of the above. Even if their methodology has a lot of weaknesses they have at least to communicate them properly to their readers. This is not the case and their reference list is clearly either irrelevant or missing major papers regarding seismic hazard assessment (all these are clearly communicated by Grutzner et al. 2013, Papanikolaou et al. 2015, Faure-Walker et al. 2019, Scotti et al. 2021 among others). My view is that the authors lack important knowledge regarding recent advances in the area of seismic hazards that are concentrated on active faults (see Working Group of California Earthquake Probabilities since the early 90ies), they ignore essential scientific arguments.
It is unclear how the faults are introduced in their model and one of the key questions is whether the map in Figure 3a shows the faults or the active faults. Most of the faults in most areas are inactive and offer no value to seismic hazard, only the active ones are recognized as seismic sources. So you can not use the faults from a typical geological map, since most of them are inactive. So I m afraid that the vast majority of the faults in their map are actually inactive and thus produce a false distribution of/non-existent seismic sources!
Finally, I’m also afraid that the authors lack the expertise to assess seismic hazard, not only because they disregard or ignore all the basic points addressed above, but also because they write that: seismic activity is driven by various and complex phenomena such as erosion! Line 46-47 glacial isostatic adjustment! This is true after the end of the glaciations 18kyrs years ago, or some seismicity remnants in Northernmost Europe, but not in France. What is surprising is that they do not clearly communicate that all major earthquakes are generated by faults! so that they should concentrate on faults for extracting seismic sources.
Making hazard and risk maps is a very important responsibility not only for science, but also for the society, therefore we have to take extra care.
Author Response
Please see the attachment.
Author Response File: 
Author Response.docx
Reviewer 2 Report
Journal: Applied Sciences
Article: Stochastic Generator of Earthquakes for France mainland
Reference Number: applsci-1426001
Comments to the authors:
This article proposes a stochastic framework for the estimation of seismic hazards in France. The work is sound and can be followed. There are some major issues with the manuscript which should be addressed before it can be recommended for publication here.
- The literature review section is not complete and should be enriched by work conducted in this area; in that case, it might be better figured out whether this is a new work or only a case-dependent study with no specific novelty. The following studies can be suggested but the authors have no obligation to use them all.
- Farahani, S., et al. (2020). "Macrozonation of seismic transient and permanent ground deformation of Iran." Hazards Earth Syst. Sci. 20(11): 2889-2903.
- Sitharam, T. G. "The Quintessence of 25 Years of Our Contributions to Geotechnical Earthquake Engineering." Indian Geotechnical Journal1 (2021): 3-49.
- Jahangiri, Vahid, and Hamzeh Shakib. "Reliability-based seismic evaluation of buried pipelines subjected to earthquake-induced transient ground motions." Bulletin of earthquake engineering8 (2020): 3603-3627
- Ji, Jian, et al. "Probabilistic investigation of the seismic displacement of earth slopes under stochastic ground motion: A rotational sliding block analysis." Canadian Geotechnical Journal7 (2021): 952-968.
- The reviewer did not find a specific framework throughout the manuscript as claimed by the authors; the authors mention that they have proposed a new framework but they have employed a conventional hazard estimation methodology as widely used in other countries. The authors are advised to provide a flowchart after the literature review section so the work novelty might be better understood.
- As well, the authors are advised to discuss clearly whether their work can address a multi-hazard scenario; if this is not the case, they can mention the limitation of their work.
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Reviewer 3 Report
In this article, the authors propose a new approach to generate synthetic earthquake catalogs and illustrate its applicability to France.
There are 37 references in the text. Of these, about 30% are less than 5 years old, and about 49% are more than 10 years old. The average date is 2000, which is acceptable.
The return periods that are presented in Table 2 seems a bit low to me. For example, for a magnitude 6.5 the value that is presented is 53 years. However, I believe that the last earthquake with a magnitude higher than 6 that occur in France was probably the Provence earthquake in 1909, which contradicts those values.
In my opinion, authors should better explain those values and must compare them with the instrumental period, because the determination of magnitudes based on observed (reported) intensities in historical earthquakes is not much reliable.
There are also some minor English grammar issues. For example, in line 64/65 where is process, probably the authors meant to write processed, and so one.
Despite considering the interesting the proposed approach, I found the description of the method a bit fuzzy, namely for someone to reproduce this approach in another country.
For all those reasons, I believe that the paper is interesting for publishing, but only after a major revision.
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Round 2
Reviewer 3 Report
In my opinion, the authors greatly improved the paper. The article is much better structured and clearer. In the current version, it is much obvious for the reader which methodology the authors propose. Figure 1 is a great contribution to a quick understanding of the adopted approach, which I found now much more interesting to be publish.
Therefore, it also became clearer what the advantages and disadvantages of the proposed methodology are.
In the caption of Figure 1 where is DPM the authors probably meant to write PMD.
There is no explanation for Fy() in lines 114 to 116.
The authors should clarify how the map faults are converted into a probability, as stated in line 134. Is the probability obtained by dividing the number of cells associated to a given earthquake divided by the total number of cells of that faulting system, or something like that? And how this is combined with the rupture parameters that are presented in section 2.1.3? In my opinion, this is not clear, as it should in order to be possible to replicate the approach.
In section 3.2.2, authors must clearly indicate how many earthquakes were generated in each synthetic catalog, using the Monte Carlo method. This number is important as it controls the minimum probability that can be obtained.
It is not very clear for me which is the scientific basis for the expression ND = (dp-df )/df that is presented in line 431.
For all those reasons, I believe that the paper is very interesting for publishing, but still only after a revision.
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Round 3
Reviewer 3 Report
The authors responded to all my questions and concerns.
In my opinion, the article is now ready to be published.
Author Response
Thanks you for your mainingful reviews that help to improve our manuscript.
