The manuscript presents a social network / graph analysis of actors mentioned in the press during the 2019 seismo-volcanic crisis in Mayotte, France. It discusses the amount of attention received by local authorities, national authorities, and scientists, and highlights the effect of the actors’ degree of connection to the society on the results.

The manuscript is well written, the results are discussed in a fair way and to a good length. The only downside of this manuscript is its editorial quality. I have found several instances of incorrect English throughout the manuscript; perhaps there are more. Please the manuscript thoroughly checked. Also, the Figures 1, 5, 6, 7, 8, 9, 10, and 11 have a low resolution, text hard to read, and too small font sizes. There is overlapping text in some of the figures.

Line 114, “eruptive activity at sea”: please consider revising to “submarine eruptive activity”

Line 114, “a newly born”: Please use a reference to confirm that it is indeed new and that it is not a preexisting that was simply discovered in 2019.

Line 115, “uncertainties were really high”: uncertainties with respect to what? The location of the volcano? The origin of the felt earthquakes? The duration of the crisis? Please specify if there is a size or a magnitude (and units of measurement or magnitude units) used to describe the uncertainty. 

Line 117, “poor instrumental network”: Poor in what sense? Not dense enough? Bad resolution in time, bad bit rate, something else?

Line 117, “instrumental network”: revise to “network of sensors”

Line 119, “to appraise the situation”: Please consider revising to “to have situational awareness”.

 

Line 126, “seisms”: replace with “earthquakes”.

Line 127, “km from the coast”: replace with “off the coast”.

Lines 128-129, “on Petite Terre island”: replace with “on the Peite Terre island, Mayotte's second-largest island, east of the largest island and closer to the Fani Maore volcano”.

Line 192, Figure 1: The image resolution is too low, which makes the text hard to read, which is already hard to read because of the small font size. Please consider improving the figure while taking into account any editorial requirements.

Line 207, “a double-reading method”: Please give a short description of this method, perhaps a reference too.

Line 214, “20190507_JDM_001”: It seems that this needs to be deleted.

Line 216, “its exact denomination(s)”: Incorrect English; replace with “their exact denominations”.

Line 225, Table 1; Line 438, Figure 5:  “sismo-volcanic”: typos; replace with “seismo-volcanic”

Line 258, “Louvain clustering method”; Line 260, “network diagrams plotting citation links”

: cite any code or software used to apply this method and to plot the diagrams or declare that you used an in-house code (specify the programming language you used in this case, and cite it).

Lines 253; 309, : replace “vs” with “versus” throughout the manuscript, e.g.: “speech vs simple” revise to “speech versus simple”; “mention vs indirect mention” to “mention versus indirect mention”, etc..

Line 367, “achieving a PhD in geography at Paul Valéry Montpellier 3 University”: Incorrect English, revise to “holding a PhD in geography from the Paul Valéry Montpellier 3 University”.

Line 374, “BRGM (French geological survey BRGM)”: revise to “BRGM (French geological survey)”.

Line 378, “representatives, ect)”: replace with “representatives, etc.)”.

Lines 395-396, “oceanographic campaigns (MayObs 1 and 2)”: please add a reference.

Line 398, Figure 3: Please consider using colorblind friendly colors.

Lines 402-403, “presented in green”, “presented in orange”: revise to “annotated with green/orange labels”.

Line 417, “Groupe d’Intervention Macrosismique”: please give a short description of what this is, and a reference as well.

Line 436, Figure 5: Not very helpful, too dense network, hard to see the connections.

Line 454, “mainland France and”: add a serial comma “mainland France, and”.

Line 456, “Macroseismic Intervention Group”: Here it is mentioned in English; elsewhere the manuscript used its French name. Please use the same names consistently through the text.

Line 489, “VLP earthquake”: Please write it out in full: “very-long-period earthquake”.

Line 491, “Twitter netwo rk”: Delete the space: “Twitter network”.

Line 586, “several limitations in our study”: Please consider briefly discussing all the limitations of this study that you are aware of.

Line 628, “relatively minimal material and human damage (only three lightly injured and cracked buildings in Mayotte)”: Please revise to “the light building damage (cracks), and the small number of people affected (three lightly injured).”

Line 636, “However, here despite”: please revise to “However, here, despite”

Article review: Le Vagueress et al “Quantitative analysis of actors' mention in press coverage of a seismo-volcanic activity in the French overseas”

Review by Lara Mani

This paper presents a network analysis of actor mentions within press coverage of the 2018 Mayotte volcanic-seismic crisis. It examines how various actors (scientists, civil protection, disaster risk managers, and civil society) interact in press coverage to identify the key actors at play during the crisis and how their influence evolves as the crisis unfolds.

Overall, the piece is well structured and well argued, and the methodology is robust and well justified. I believe it will make a unique and significant contribution to the literature. However, it is littered with editorial errors and language inaccuracies and could do with a thorough proofread (I include an annotated version of the manuscript to aid the authors in correcting mistakes). There are also some challenges with what I presume are translation inaccuracies (see various comments below), and these should be carefully revised. Further general comments on the manuscript and then more specific comments are included below.

General comments:

• The title of the article makes little sense and should be revised. E.g. “Quantitative analysis press coverage during the 2018 Mayotte volcano-seismic crisis”
• In section 2 where you describe the case study from Mayotte, it would be helpful to have a visual timeline of the eruption and key moments in the disaster response to aid the reader.
  • This would also help to map the periods used throughout the piece e.g. L440 “for the second period”.
• Section 3.1 – not sure the use of ‘corpus’ is useful. Although technically not incorrect, this is not commonly used in English and we’d likely just use ‘collection’ so I'd suggest using similar.
• Stylistic preference, but the authors shouldn’t be present in the text. The use of we and our should be removed.
• Before you begin to show the results and discuss the implications of the study, it would be useful in the early stages of the manuscript to understand who the actors are and their role and capacity during the crisis to add context to the data or at least how you’ve split them amongst the categories in Table 4.
• Ensure there is consistency in how you refer to your figures and tables (e.g. you use both Figure and Fig).
• Be careful not to discuss the results in the results section and hold that for the discussion e.g. L374-379.
• Check the consistency of how you represent dates throughout the piece e.g. L412 “05/10/18” and L441 “October 5^th 2018”.
• In your citation network maps (Figs 5-11), you have the influencer as Said Hachi,m and in the text his name is Said Said Hachim – check for consistency.

Specific comments

Please also refer to the attached marked-up PDF of the paper for additional suggestions for improving the text.

Introduction

• The style of the literature review reads more like a thesis chapter than a paper. It’s unnecessary to state why each paper is relevant in the brackets, and it’s quite annoying to read e.g. L44, L49, L57-60, L65-68, L86.
• Many of the citations are very old references and I would suggest that the author seek some more recent literature, particularly in the opening paragraphs.
• P60 – “The way journalists tend to cross speeches…” Not sure what you’re trying to say with this sentence, consider rephrasing.
• L88 – “as this approach enables; i) gaining insights … ii) identifying actors” – the English doesn’t quite work here. Maybe change “enables” to “allows for” or “supports”.
• L94 – add the year of the seismo-volcanic crises.

Case study description

• L107 – I’d suggest a more descriptive title for this section “The 2018 Mayotte eruption”.
• L113 – “turned out to be linked to a volcanic eruptive activity at sea and a newly born volcanic edifice” change to “This seismic crisis was linked to the formation of a new submarine eruption site…”
• L116 – “uncertainties were high” – what uncertainties? Also, remove ‘really’ from this line. I think this sentence just needs reforming.
• L117 – “poor instrumental network” change to “poor monitoring network”.
• L117-119 – sentence needs restructuring.
• L126 – replace “seisms” with earthquakes.
• L127 – change to “was the detection”.
• L129 “since prior to 2018” remove “since”.
• L138 – removed prevention from “risk prevention communication”.
• L139 – change to “important social inequalities”
• L142 – change “implanted in the population” to “within the population”
• Section 3.1 L165-168 – not sure these papers are well summarised above to build upon them. Ensure they are included in your literature review and well fleshed out if you intend to build upon their works.

Method

• L182 – add line break to next paragraph.
• L185 – “thus involving possible changes in perception” – this feels speculative… can you found this in the literature?
• Figure 1. – the image is too small to read well.
• Section 3.2 – L202 remove indentation at the beginning of the paragraph.
• L204 – change “Twitter” to “X (formerly known as Twitter)”.
• L206 – change "are then also selected” to “and are also included”
• L207 – remove the references to the research team “a double-reading method by human operators was used to select and…”
  • What does this mean “a double-reading method”?
• L234 – add line break.
• Table 2 – can you add the rule for each part of the matrix and then show the example?
• L257 – “two global network analysis indicators” – what are they and what does this mean?
• L258 – define the ‘Louvain clustering method’.
• L258-262 – this is a very long sentence; consider cutting it down or restructuring.

Results:

• Figure 2 L331 – remove “in Part 2”.
• L343 – add line break
• L350-351 – be careful not to stray into discussions in this section.
• Table 4. Who do you include under other populations? Is this the diaspora?
• Figure 3 should be moved up before the paragraph starting “Figure 4” L386.
• Figure 3 – can you find a way to plot this so we can see the nuance of the data in the bottom left more clearly? E.g. break the X axis
• L412-416 – double check these have translated across properly to what you intend to say here, they aren’t all clear to me who you are referring to e.g. prefect (Head of Department?), rectorate (Head of Education?).
• Figure 5 – this is unreadable! Can you strip back some of the data to show the most important relationships rather than your entire data set (or which some surely is not discussed here).
• Figure 5. L 438 – the figure caption should be expanded to explain the graph (see them as a standalone).
• Figure 6. This one is easier to read by needs to be made much larger to be able to see the finer details.
• Figures 7-11 As previous figures – these are difficult to read.

Discussion:

• It would be useful to understand the role of Associated Press here e.g. Reuters who provide articles for international media based on local reporting.
• How does your analysis fit with the planned crisis communication strategy for Mayotte? Can your analysis be used to prove the efficacy of the pre-defined communication pathways for volcanic crises?
• L 674-5 “facilitating two-way communication between institutions…” – why is this important/beneficial?
• Your results are meticulous in showing how actor citations change throughout the evolving crisis but then you neglect this temporal perspective in the discussion. How do actor relationships change as the crisis evolves (and indeed the political crisis unfolds)? Otherwise, figures 5-11 are quite useless, and could be summed up using just one graph!
• One thing I think is missing from the analysis is around how the mentions are made – are they positive, neutral or negative? Do actors place the blame on others and thus support the erosion of trust in other organisations. How does their representation in media affect their perceptions?
• Include some of the limitations of this study and how they could be overcome for future studies.

Conclusions and perspectives

• A little bit of French in the title! Conclusion and perspective.
• L 761-762 – although a great point about diversification of channels, this is a new point and shouldn’t be brought into the conclusion if not mentioned above.
• The conclusions are a little weak and could draw out more of the excellent points made in the discussion. It’s useful to hear how the method could be used more broadly but this is new content and shouldn’t be in the conclusions (maybe a new short section on future work?).