the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 14 Mar 2025
Quantifying cascading impacts through road network analysis in an insular volcanic setting: the 2021 Tajogaite eruption of La Palma Island (Spain)
Abstract. Post-event impact assessments (PEIA) are essential to elucidate disasters’ drivers and better anticipate future events. The 2021 Tajogaite eruption of Cumbre Vieja (La Palma, Spain) demonstrated the various orders of impact due to compound volcanic products (i.e., lava, tephra, gas) affecting a highly interconnected and low redundant infrastructure, typical of insular environments. Using a forensic approach, we discretise the causal order of cascading impacts, from physical damage (first order) to loss of functionality of the road network (second order) and subsequent systemic disruption of emergency management and socio-economic sectors (third order). Based on graph theory, we apply a comprehensive network analysis to quantify the loss of functionality and resulting effects, based on the spatiotemporal evolution of centrality indicators. The consequences on dependent systems are expressed in terms of increased driving time syn- and post-eruption between target locations for emergency (evacuation), public health (hospital), agriculture (crops-market), and education (schools). Graph indicators are objective measures of system performance during (disturbing/degraded states) and after the eruption (restorative state), when two new roads where rapidly built to reconnect the island. This study demonstrates how network analyses, informed by comprehensive PEIA, can accurately capture complex systemic disturbances, thus highlighting its potential for risk assessments.
- Preprint
(3130 KB) - Metadata XML
-
Supplement
(5378 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-986', Anonymous Referee #1, 09 May 2025
This paper presents a detailed post-event impact assessment (PEIA) that highlights the compound nature of volcanic hazards and their far-reaching consequences on interconnected infrastructure systems. By adopting a forensic approach and applying graph theory-based network analysis, the study quantifies both direct physical damages and indirect systemic disruptions—particularly within transportation, emergency services, agriculture, and education sectors.
The objective of the work is clearly defined, and the case study is of utmost importance for the scientific community. The study is grounded in the specific context of the 2021 Tajogaite eruption, and this focused approach allows for a nuanced and thorough examination of local impacts and system responses. The depth of analysis provides clear value for understanding complex interactions in disaster scenarios. Nonetheless, the strong emphasis on a single case may somewhat narrow the study’s relevance to broader contexts. A more explicit connection to similar challenges in other regions would further enhance the paper’s applicability and reach. The case-study of La Palma is undeniably meaningful but I would suggest giving a bit more emphasis to broader implications on the aspects that would make it usable in another region and context.
I think you approach could gain relevance if discussed against similar approaches developed for other natural hazards and for tacking the deriving emergencies, as well as for multi-hazard scenarios. A wider comparative perspective or a discussion linking the results to other active volcanic areas or areas prone to other hazards (e.g. landslides) would enhance the study’s impact and utility for a more diverse audience. Expanding the contextualization would help readers better understand how the presented methods and insights could be adapted to different geographic or socioeconomic settings.
With regards to the methodology, I suggest to discuss a bit more how these ex-post assessments can help future risk management. How can these be included in capacity planning, emergency management and in the definition of future exposure and impact scenarios? I really like your idea of using past lava flow locations and this might be discussed explaining how this could be complemented (e.g. with modeling) to provide future scenarios to land use planners and infrastructure managers. You already started discussing this in line line 907-915 but I think a bit more emphasis could be given to this.
Finally, I think the manuscript could benefit from an overall polishing to avoid repetitions which are sometimes confusing for the reader and perhaps distract from the main message (See specific comments provided in the supplement file).
Overall, I congratulate the authors for their rigorous and insightful work, which makes a valuable contribution to disaster impact analysis; with the suggested enhancements to broaden its contextual reach, this study has the potential to become a significant reference for the scientific and risk management communities.
-
AC2: 'Reply on RC1', Lucia Dominguez Barragan, 16 Jul 2025
This study by Dominguez et al. presents a comprehensive post-event impact assessment of the 2021 Tajogaite eruption on La Palma Island, Spain. The authors apply an interesting approach of combining a forensic impact assessment analysis with network analysis to evaluate and quantify functionality loss and systemic impacts to infrastructure from the eruption. I think this will make an important and useful contribution within the scope of NHESS and to the academic literature on post-event impact assessment. However, there are a few issues that require consideration that I outline below.
No response needed.
Ambiguity between observed and modelled impacts and effects: It can be unclear in places whether an impact of effect described in the paper was observed/reported by stakeholders or derived from the modelling approach. This is because the text shifts at times between describing what happened (or was reported to have happened) and then what is inferred through the modelling. The differences can get a bit lost in dense paragraphs. I think the manuscript would benefit greatly from greater clarification. Perhaps through the use of a table. This would aid the reader to interpret the added value of using the network analysis approach.
We thank the reviewer for this insight. We will better clarify the methods used and the outcomes coming from the participatory discussions with stakeholders and the results from network analysis.
Clarifying and critically evaluating the graph-theory approach among existing geospatial methods: The manuscript would benefit from a clearer articulation of what value graph-theory indicators add as the indicators are somewhat theoretical. The authors state this is pioneering in volcanology, but I think that is possibly too strong of a claim. Whilst this appears to be the first time researchers have applied topological indicators in this way for a volcano case study, it is not the first time that network analysis or geospatial analysis has been used to supplement observed impact information for post-event impact assessment for volcanic eruptions. As the authors highlight, EBC and ECC are both toplogical measures and so do not necessarily reflect functional aspects of road importance (e.g., traffic capacity, importance to critical facilities etc). It would be valuable for the authors to critically evaluate the implications of this, and to situate this approach in relation to other geospatial methods that have been applied in post-event impact assessments for volcanic eruptions. The authors do have a section on potential and limitations, but I think this currently falls short of the necessary critique of the method, its added value, and how it might fit with other previously used approaches.
In the introduction of the manuscript, we highlighted pioneering studies that applied graph networks to the quantification of vulnerability, such as Weir et al. (2024) and Mossoux et al. (2019). Our intention was not to present our study as the pioneering work, but rather to emphasize that, to the best of our knowledge, this is the first time topological indicators have been applied specifically within the field of volcanology.
We would be grateful if the reviewer could point us to any additional references where similar topological measures have been used, as we would be keen to include them.
We also appreciate the valuable suggestion to expand Section 5.2 on the potential and limitations of our approach. In response, we will include a more critical discussion of the implications of using graph-based indicators, particularly in comparison with other methodologies used for volcanic post-event impact or risk assessments.
Ethics: The authors highlight that they undertook discussions with stakeholders following the eruption. Can they please include a statement somewhere indicating that the necessary ethics approvals (if required) for their institution for this type of research were obtained. Often this can go in the acknowledgements or methodology section.
Thank you for your comment. Throughout the study, we were careful to follow best-practice guidance for engagement and interactions at the Science-Practice-Policy Interface (SPPI) (Cash et al. 2003 ; Wyborn et al. 2017 ). Interactions were semi-structured, involving several recurring discussion points and adapted to current operational concerns and requirements. This model ensured that the interactions were mutually beneficial and highly adaptive. No formal interviews were conducted, no individuals have been named, nor have their personal opinions been shared. The nature of the engagement did not meet the threshold required for ethics approval at the University of Geneva, where the funding was held and the majority of the authors were based.
However, an ethics approval is in place for the project that funded this study. This funding project (SNSF #188757) is principally concerned with volcanic risk in the Central and Southern Andean Volcanic Zones and hence is tailored to engagement undertaken in South America with local research partners. However, the ethics approval granted for the funding project outlines a robust ethical approach for engagement with volcanic risk management stakeholders that heavily informed this study in La Palma. We will provide mention of this ethics approval in the acknowledgements and note that it is the ethics approval for the project that funded this study.
Some smaller points
- The manuscript in places uses a lot of jargon that could be more plain spoken (or at least clear definitions provided early in the manuscript). Examples include: insular environment/context, centrality indicators, "compound lava", single annular primary network, system performance phases named as disturbance progress, degraded, and restorative states.
We thank the reviewer for this comment. We will carefully verify the whole manuscript with special attention to remove any unnecessary jargon and provide definitions of unclear terms.
- There are a number of overly verbose and structually complex sentences throughout. As one example: "Further research should focus on the selection of indicators (and parameterisation) with a physical meaning, which must be developed as we explore better this approach, by applying to future on-going eruptions (syn- and post-event impact assessments) or to future potential scenarios (pre-event impact assessments)." This sentence would benefit from being split into a couple sentences. I suggest the authors review the entire manuscript to reduce redudant words and complex sentences.
Thanks for this comment. We will carefully revise the language and complicate sentences by addressing comments of both reviewers.
The authors mention that they assume no clean-up occurs as there is no information they can use for this. This is reasonable, but as the authors point out this will have a significant impact on the outputs and inferences made. Could the authors explore end-member scenarios similar to what was done for speed restrictions?
This is a very important aspect. In fact, in the preliminary stages of this research, we wanted to explore the effect of clean-up operations, as was done, for example, in the case of impact on buildings where a record of clean-up operations was available (see Reyes-Hardy et al. 2024, https://www.frontiersin.org/journals/earth-science/articles/10.3389/feart.2023.1303330/full).
It is important to note that clean-up operations are crucial in the case of roof collapse of buildings. However, in the case of roads, our results clearly showed that the major component influencing the evolution of graph indicators was the lava field. This means that the effect of tephra fallout only narrowly dispersed the value of indicators, but the crucial trend was defined by the lava field evolution (see, for example, Fig. 9B and 9C). This is clearly stated in lines 648–650.
For this reason, we considered that introducing fictitious scenarios of clean-up operations of roads would introduce an extra source of uncertainty that is totally difficult to validate (since we do not have any record of when and where the clean-up operations were done), and in the end, the indicator value would approximate (if not equal to) the lava scenario (brown triangles in Figure 9B, 9C).
- Line 112: "under the light of network analysis" is strange phrasing and unclear.
We will change for “in light of network analysis”
- Line 278: Why was the decision unprecedented?
With “unprecedented” we mean that the decision of allowing residents to enter into the exclusion zone to collect their belongs was never done before in any emergency that Spain has faced. This was stated by emergency stakeholders, particularly civil protection practitioners. We consider that it is pertinent to mention this decision in our manuscript not only because of its relevance in the crisis management but also for its implications on the road network accessibility and connectivity. Indeed, this measure provoked long traffic jams and important disruptions of the road network not directly impacted by lava or tephra.
- A few more sentences to expand upon the rationale for the selection of origin-destination locations. It's stated that these are key selected locations, but why?
We thank the reviewer to point out this. In fact, we wanted to demonstrate with few examples the effects on other systems given the loss of connectivity of the road network. To do that, we select 4 sectors: emergency, health, agriculture, education and we carefully selected the coordinates of locations that were relevant for the communities and stakeholders to show the changes on travel time before, during and after the eruption of Tajogaite. We will add some lines clarifying this in the manuscript to facilitate the rational understanding of this selection.
- Line 402-404: Is the sentence relating to local and global newspapers relevant? Not clear to me.
Media and the community in general were quite concerned about the lava reaching the sea. This event is relevant to our focus on the road network because on that day (29 September), the western flank of the island was disconnected (evidenced by all graph indicators dropping, Figure 10), but also because media and emergency attention were largely focused on it. There was particular concern about the secondary explosions that lava could produce upon contact with water. Indeed, some confinement measures were implemented due to poor air quality.
Line 414: How much ash fell at the port? Can you specify a thickness?
Unfortunately there is any official measure of thickness at the port. Practitioners only mention that “some ash” fell down few days during the eruption. Direct impacts due to tephra were no reported.
- Line 430: reference to Fig. 2B. Should this be 3B? Fig 2 doesn't have a B.
Totally correct. Thanks for pointing this.
- Table 5: What does "highly disruptive" mean in this context under the Tephra column?
We wanted to emphasize that although tephra has not direct physical damage on roads, it could be highly disruptive. We will modify by “disruption due to reduced visibility, loss of traction, covering of marks, reduction of skid resistance and total covering, depending on deposit thickness”
- Line 509: What level of damage occurred? Was this cosmetic or functional damage?
It was physical damage including cracks, deep scratches and abrasion of markings. We will add these specific physical impacts on the manuscript
- Line 623-624: why are percentage ranges provided? Isn't clear what the percentages refer to.
It is provided in percentage as a total measure of timing affectation during the scholar year. We will add the number of days instead.
- Line 780: "IS1b till the end" shoud this be "until the end"?
We will modify it.
I would like to finish by applauding the authors on producing what I believe is an interesting piece of research and a valuable and timely contribution. It is important to document the impacts of volcanic eruptions and the authors have taken this a step further by providing an additional tool and approach to explore post-event impact assessment for volcanic eruptions.
We kindly thank the reviewer for her/his valuable insights in this review. We believe all the comments addressed here will contribute to improve our manuscript.
Citation: https://doi.org/10.5194/egusphere-2025-986-AC2
-
AC2: 'Reply on RC1', Lucia Dominguez Barragan, 16 Jul 2025
-
RC2: 'Comment on egusphere-2025-986', Anonymous Referee #2, 04 Jun 2025
This study by Dominguez et al. presents a comprehensive post-event impact assessment of the 2021 Tajogaite eruption on La Palma Island, Spain. The authors apply an interesting approach of combining a forensic impact assessment analysis with network analysis to evaluate and quantify functionality loss and systemic impacts to infrastructure from the eruption. I think this will make an important and useful contribution within the scope of NHESS and to the academic literature on post-event impact assessment. However, there are a few issues that require consideration that I outline below.
Ambiguity between observed and modelled impacts and effects: It can be unclear in places whether an impact of effect described in the paper was observed/reported by stakeholders or derived from the modelling approach. This is because the text shifts at times between describing what happened (or was reported to have happened) and then what is inferred through the modelling. The differences can get a bit lost in dense paragraphs. I think the manuscript would benefit greatly from greater clarification. Perhaps through the use of a table. This would aid the reader to interpret the added value of using the network analysis approach.
Clarifying and critically evaluating the graph-theory approach among existing geospatial methods: The manuscript would benefit from a clearer articulation of what value graph-theory indicators add as the indicators are somewhat theoretical. The authors state this is pioneering in volcanology, but I think that is possibly too strong of a claim. Whilst this appears to be the first time researchers have applied topological indicators in this way for a volcano case study, it is not the first time that network analysis or geospatial analysis has been used to supplement observed impact information for post-event impact assessment for volcanic eruptions. As the authors highlight, EBC and ECC are both toplogical measures and so do not necessarily reflect functional aspects of road importance (e.g., traffic capacity, importance to critical facilities etc). It would be valuable for the authors to critically evaluate the implications of this, and to situate this approach in relation to other geospatial methods that have been applied in post-event impact assessments for volcanic eruptions. The authors do have a section on potential and limitations, but I think this currently falls short of the necessary critique of the method, its added value, and how it might fit with other previously used approaches.
Ethics: The authors highlight that they undertook discussions with stakeholders following the eruption. Can they please include a statement somewhere indicating that the necessary ethics approvals (if required) for their institution for this type of research were obtained. Often this can go in the acknowledgements or methodology section.
Some smaller points
- The manuscript in places uses a lot of jargon that could be more plain spoken (or at least clear definitions provided early in the manuscript). Examples include: insular environment/context, centrality indicators, "compound lava", single annular primary network, system performance phases named as disturbance progress, degraded, and restorative states.
- There are a number of overly verbose and structually complex sentences throughout. As one example: "Further research should focus on the selection of indicators (and parameterisation) with a physical meaning, which must be developed as we explore better this approach, by applying to future on-going eruptions (syn- and post-event impact assessments) or to future potential scenarios (pre-event impact assessments)." This sentence would benefit from being split into a couple sentences. I suggest the authors review the entire manuscript to reduce redudant words and complex sentences.
- The authors mention that they assume no clean-up occurs as there is no information they can use for this. This is reasonable, but as the authors point out this will have a significant impact on the outputs and inferences made. Could the authors explore end-member scenarios similar to what was done for speed restrictions?
- Line 112: "under the light of network analysis" is strange phrasing and unclear.
- Line 278: Why was the decision unprecedented?
- A few more sentences to expand upon the rationale for the selection of origin-destination locations. It's stated that these are key selected locations, but why?
- Line 402-404: Is the sentence relating to local and global newspapers relevant? Not clear to me.
- Line 414: How much ash fell at the port? Can you specify a thickness?
- Line 430: reference to Fig. 2B. Should this be 3B? Fig 2 doesn't have a B.
- Table 5: What does "highly disruptive" mean in this context under the Tephra column?
- Line 509: What level of damage occurred? Was this cosmetic or functional damage?
- Line 623-624: why are percentage ranges provided? Isn't clear what the percentages refer to.
- Line 780: "IS1b till the end" shoud this be "until the end"?
I would like to finish by applauding the authors on producing what I believe is an interesting piece of research and a valuable and timely contribution. It is important to document the impacts of volcanic eruptions and the authors have taken this a step further by providing an additional tool and approach to explore post-event impact assessment for volcanic eruptions.
Citation: https://doi.org/10.5194/egusphere-2025-986-RC2 -
AC1: 'Reply on RC2', Lucia Dominguez Barragan, 15 Jul 2025
Reviewer #1
Title: Quantifying cascading impacts through road network analysis in an insular volcanic setting: the 2021 Tajogaite eruption of La Palma Island (Spain)
Decision: Major revisions
General review:
This paper presents a detailed post-event impact assessment (PEIA) that highlights the compound nature of volcanic hazards and their far-reaching consequences on interconnected infrastructure systems. By adopting a forensic approach and applying graph theory-based network analysis, the study quantifies both direct physical damages and indirect systemic disruptions—particularly within transportation, emergency services, agriculture, and education sectors.
No response needed.
The objective of the work is clearly defined, and the case study is of utmost importance for the scientific community. The study is grounded in the specific context of the 2021 Tajogaite eruption, and this focused approach allows for a nuanced and thorough examination of local impacts and system responses. The depth of analysis provides clear value for understanding complex interactions in disaster scenarios. Nonetheless, the strong emphasis on a single case may somewhat narrow the study’s relevance to broader contexts. A more explicit connection to similar challenges in other regions would further enhance the paper’s applicability and reach. The case-study of La Palma is undeniably meaningful but I would suggest giving a bit more emphasis to broader implications on the aspects that would make it usable in another region and context.
We thank the reviewer for this important observation. Discussion of the applicability of this approach to other applications is provided in “Section 5.4. Insular and monogenetic settings”. As an example of applicability, the same methodology was applied for the 8 historical eruptions of La Palma by using the actual road network and past lava field footprints (Figure 11).
However a more explicit discussion of the applicability of this approach to other insular contexts (i.e. with closed road systems) will be provided in the reviewed version. We will also discuss its applicability to open systems (continental areas), through defining boundaries, and averaging graph indicators in delimited areas. In fact, the versatility of this approach allows application to any region and hazard footprint (e.g., floods, landslides, etc.) as it has been seen for the example of historical lava fields polygons that, regardless the hazard intensity metrics, the disconnected road graphs can be analysed. All this will be further emphasized in the reviewed version.
I think you approach could gain relevance if discussed against similar approaches developed for other natural hazards and for tacking the deriving emergencies, as well as for multi-hazard scenarios. A wider comparative perspective or a discussion linking the results to other active volcanic areas or areas prone to other hazards (e.g. landslides) would enhance the study’s impact and utility for a more diverse audience. Expanding the contextualization would help readers better understand how the presented methods and insights could be adapted to different geographic or socioeconomic settings.
Same comment as given above:
We thank the reviewer for this important observation. Discussion of the applicability of this approach to other applications is provided in “Section 5.4. Insular and monogenetic settings”. As an example of applicability, the same methodology was applied for the 8 historical eruptions of La Palma by using the actual road network and past lava field footprints (Figure 11).
However a more explicit discussion of the applicability of this approach to other insular contexts (i.e. with closed road systems) will be provided in the reviewed version. We will also discuss its applicability to open systems (continental areas), through defining boundaries, and averaging graph indicators in delimited areas. In fact, the versatility of this approach allows application to any region and hazard footprint (e.g., floods, landslides, etc.) as it has been seen for the example of historical lava fields polygons that, regardless the hazard intensity metrics, the disconnected road graphs can be analysed. All this will be further emphasized in the reviewed version.
With regards to the methodology, I suggest to discuss a bit more how these ex-post assessments can help future risk management. How can these be included in capacity planning, emergency management and in the definition of future exposure and impact scenarios? I really like your idea of using past lava flow locations and this might be discussed explaining how this could be complemented (e.g. with modeling) to provide future scenarios to land use planners and infrastructure managers. You already started discussing this in line line 907-915 but I think a bit more emphasis could be given to this.
Many thanks for your insight. We would include a more detailed discussion of the implications of this methodology for future risk management. Indeed, the methodology is well structured, rapid to apply, and the python package (OSMnx) is friendly and runs with low computational resources, which could make easy to use for risk management and during crisis management processes. This is particularly important in insular and monogenetic volcanism that involves significant challenges for land-use planning and rapid emergency response due to the reduced space and sudden volcanic changes. All this will be better discussed in the reviewed version.
Finally, I think the manuscript could benefit from an overall polishing to avoid repetitions which are sometimes confusing for the reader and perhaps distract from the main message (see specific comments below).
We will polish and deeply review the whole manuscript to avoid redundancies and simplify language when needed.
Overall, I congratulate the authors for their rigorous and insightful work, which makes a valuable contribution to disaster impact analysis; with the suggested enhancements to broaden its contextual reach, this study has the potential to become a significant reference for the scientific and risk management communities.
No response needed.
Specific comments:
Introduction:
In this work it is somehow unclear why you use the term systemic or cascading impacts, and if you are tackling both. In particular, you mention in line 93-95 that the cascading effects are triggered due to the interaction between roads and other exposed assets, so, to my understanding, we are talking about the induced impacts due to road network disruptions. However, other systemic impacts are also described. I think the reader would benefit for a clarification in the introduction so it’s clear what is meant.
Thank you for your comment. We will clearly define our use of the terms ‘systemic’ and ‘cascading’ in the introduction of the revised version. For your interest, we follow the definitions defined by the UNDRR, utilised in their reporting (e.g. GAR reports) and available in their terminology definitions. This will be made clear in the revised version.
Lines 100-104: It could be worth mentioning the study from Scaini et al. 2014 on the loss of functionality in terms of physical and systemic vulnerability induced by compound volcanic hazards in Tenerife on exposed assets including roads.
We will include this citation in the revised version.
Fig.2: The primary road (black line) should be included in the legend, in addition to the caption
Thanks for this. We suspect you mean to refer to Figure 1B; the legend will be completed in the revised version.
Methodology:
Line 293-294: I think the reader would benefit from a broader description of the involved stakeholders and how they were involved, and for example if the same stakeholders were participating to all the encounters. Also, do the mentioned discussions only include structured interviews or is there also a different kind of interaction (e.g. group meetings)? Did they show you the location of the impacts using a map, or how did they identify the impacted areas and/or the location of the assets? Also, In line with my comment on the generalization of your approach, how do you suggest to adapt the interview questions to different contexts? How was this done passing from Dominguez et al., 2011 to this context?
We will edit the text substantially to clarify the engagement approach undertaken with stakeholders in La Palma. To clarify here, we did not conduct structured interviews, but we follow a prescribed engagement approach (e.g., co-production, participatory frameworks). This was in very large part due to the fact that this study was highly reactive, involving ad hoc meetings with stakeholders that were highly time-limited and operationally busy during the emergency management, response and recovery periods of the 2021 Tajogaite eruption.
We did however, follow best-practice guidance for engagement and interactions at the Science-Practice-Policy Interface (SPPI) (Cash et al. 2003; Wyborn et al. 2017). Interactions were semi-structured, involving several recurring discussion points and adapted to current operational concerns and requirements. This model ensured that the interactions were mutually beneficial and highly adaptive. Designing a participatory research approach requires substantial time and resources, both of which were highly limited for the duration of this study. Future work will follow a prescribed research engagement approach, given we will have the liberty of time and resources post-recovery (following the approach outlined in Weir et al. (2024), Natural Hazards).
To specifically address your comment: 1) We will clarify the number of individuals engaged with throughout the study, the number of repeat participants, and the nature of the discussion (group based, 1-on-1 etc.); 2) we did not conduct structured interviews at any point in the study; 3) we will clarify the tools used to facilitate the discussion (i.e., maps, GIS), and make clearer the source of presented results; 4) we did not use pre-prepared interview questions, but we will include some discussion on how the engagement approach can be adapted to suit other contexts.
Line 300: I think this sentence explains very well what is done and could be moved to the initial methodological part rather than the data acquisition section.
Thanks for this suggestion – this will be included in the revised version.
Table 1: why ‘disrupted’ is FL-I and ‘reduced’ is FL-II? How were these functionality levels defined? Were they all observed in the case-study?
Levels of functionality were defined in collaboration with stakeholders. We wanted to differentiate between the direct reduced service of road network due to tephra fallout (FL-II) and exclusion zone (FL-III), from the indirect disrupted service due to increasing traffic and disturbances on the network, even when hazards were not present (or insignificant, i.e. in the case of low thickness tephra fallout). The boundaries of these zones were highly discussed, and the majority of stakeholders agreed that traffic disturbances reached as far as Punta Gorda and Santa Cruz (see Figure 6B). We will clarify these points in the revised version.
Line 343-346: The information in the road network data should be moved to the data section
Thanks for this suggestion, we will move this paragraph to facilitate readability,
Results:
Section 4.1 mostly presents the impacts in a narrative form and it’s hard for the reader to distinguish if these information were collected from the presented methods, or if they are part of the context and the already available information. In the second case, these narratives could be provided at the beginning to explain to the reader what is the situation, while the results should present your original findings.
This section covers information collected and compiled from discussions with stakeholders, that’s why it makes part of the Results section. We believe that after clarifications on the methodology section (discussed above), this will be made clearer.
Lines 385 -389 and Supplementary material on impacts to other critical infrastructure systems: I am unsure why these impacts are summarized here, if it’s to demonstrate the impacts due to the volcanic hazards or to the road closure. This li ks back to the question on the definition of cascading and systemic impacts, which I think could be made clearer in the text.
It has been already demonstrated that impacts on critical infrastructure involves complex and long cascading chains of effect (e.g., Rinaldi et al., 2001; Rinaldi et al. 2004) and that we need a system-of-system approach to better quantify it (i.e., systemic vulnerability) (e.g., Eusgeld et al. 2011, Weir et al. 2024b). In this particular study we focus on the cascading impacts due to road network loss of functionality and for this reason we introduce here a brief description on what were the indirect consequences on water supply and then in the agriculture sector. Clear definitions of cascading impacts due to systemic vulnerability will be added as suggested by the reviewer, thus making the manuscript clearer.
Table 5: I think this table is very relevant to understand what are the results of this work.
Row 2, column 5: ‘Highly disruptive’ is a bit generic
Row 7, column 4: is the surface missing?
Many thanks for catching this. We will address these points in the revised manuscript.
Section 4.2.2: The initial part is a repetition of the method which should be already explained in the methodology section.
We will adapt the methods and results sections to be aligned and avoid repetitions.
eSection 4.2.3: Similarly to section 4.1 there is a broad description of what happened but it is somehow unclear what fraction of this information was gathered using your methods (e.g. interviews) and which information was already available in advance. Even in this second case, if you reinterpreted this based on the impact chains that you identified and on your methods, this can be part of the results, but it should be made clear to the reader.
The results presented here are the outcomes of discussions with stakeholders based on the engagement approach we have followed. This would be better clarified in the revised Methods section (discussed above).
Discussion:
Line 730-740: I would include in the introduction some background on the ex-post analyses and the approach in the field of disaster risk reduction, with literature references (e.g. AM Ferreira et al., 2023; F. Atun, 2024), in order to highlight the relevance of you work beyond the case-study. Also, this could be discussed to show what are the improvements of the approach that you propose.
We will add a sentence or two in the introduction, stating that post-event impact assessment has great utility for disaster risk reduction as it informs future pre-event risk analysis and planning. Due to space and word constraints, we cannot expand beyond 1-2 sentences, but we will include seminal references to direct readers to further reading on this topic.
Line 745-746: ‘identify the most impactful cascading effects’: Is this shown in the figures? Or was this part of the replies that they gave you to the interviews?
We believe this is a miscommunication due to phrasing. We will reword the relevant sentence to scale back the claim that we somehow ranked the ‘most impactful’ events.
Line 760-764: It would be relevant to explain what kind of data would support a deeper and more extended deductive/inductive reasoning. This could help other researchers interested in applying this tool in other areas where such data might be available.
We tried here to discuss in these lines which data were available (specified in lines 510-515), see below and the gaps we found indeed. With specific records of impacts, questions in an event tree can be precisely defined and moreover, probabilities for each scenario can be calculated. In the absence of this data, what we propose here is to quantify these functionality scenarios in light of network indicator evolution.
Lines 510-515: “There is no exhaustive record of all disruptions that occurred. Indeed, the DGT of Spain recorded and reported only the largest disruptions to road network operations. To supplement this data gap, we attempt to reconstruct here a set of functionality scenarios based on the narrative descriptions of stakeholders and deductive logic questions presented in an event tree shown in Fig. 7A (Functionality Levels, FL 0 to IV, described in Table 1).”
Line 765-768: This sentence is not super clear to me, could you rephrase? Does it mean that, since the majority of the tephra deposition happened in the first phase and that the impacted road length are almost constant during the following phases?
We will rephrase it as follows,
“Concerning the reduced service due to tephra accumulation (FL-II), we found that almost the totality of the surface estimated from the whole tephra deposit at the end of the eruption (>190 km2) was affected since the beginning of the eruption, resulting in an almost constant road length for FL-II impact, regardless of the amount of accumulated tephra (Fig. 7B). This means that the few millimeters covering the surface immediately implied a FL-II, however, a temporal evolution of tephra deposit is required.”
(This then introduces the next paragraph, where FL-II are then sub-divided into categories in relation with tephra thickness.)
Given that you comment on different CI (e.g. water networks) and given the presence of the water reservoirs that are, both here and in other Canary Islands, so relevant for local activities, I would suggest adding some references in the introduction about the exposure and vulnerability of these (e.g. Stewart et al., 2006, 2009).
Thanks for this suggestion, we will include these references in the revised version.
Fig. 6C: Is it systemic or cascading?
We believe definitions in the introduction will help to clarify this point. We will be including clear definitions in the revised version, as discussed above.
Line 814-815: how much of this analysis can be done a priori? e.g. analyzing the existing network and its potential pitfalls using a combination of scenario modeling and network analysis?
We will add a brief discussion of the utility of this approach to pre-event risk analysis.
Apart from the road type (and associated speed), did you account for the road path (e.g. if it is mostly straight or curved) and the pavement type? Also, how different could this scenario in case of different weather conditions (e.g. under heavy rainfalls), especially in areas served by unpaved secondary or tertiary roads?
With the available information our study accounts for 3 different weights (length, travel speed and travel time) and we explore the different outcomes for these variables. Analysis was then under the umbrella of types of road to define different speeds. However, future analysis including traffic, slope or pavement type could add more real insights on the impedance of roads. We will add a sentence on this potential for future work to the revised version.
Do you foresee to tackle the lack of traffic data in some other way (e.g. using phone data, such as in the work of Yabe et al., 2022; Wilson et al., 2016; Giardini et al., 2023)? Or, if not, shall this be addressed in the future?
This is a very interesting strategy that could be used for future eruptions. One of the major limitations we have during crisis is the lack of time and guidelines to collect impact data, and we are not likely to be able to acquire this data so far after the event. It is not likely to form part of our future work package relating to La Palma, though in other contexts this would have great utility.
Conclusions:
Following my comment on the results, I think the conclusions are a bit generic and could be made stronger by giving more emphasis to the specific results achieved with the analysis (e.g. from table 5, fig. 10-11). In particular, from Fig. 10 it seems that the travel time increases a lot for R1, and this might set a priority for emergency managers.
We will work on the conclusions to add specific results of Tajogaite case study in addition with general conclusions of the methodology proposed.
Minor comments:
If possible, I would suggest to translate the tables in S2 into English
Absolutely agree. We wanted to show the original alert levels of PEVOLCA, but of course we will add a complementary English language version.
I’m not a big fan of acronyms and sometimes I feel their use could be reduced in favor of readability.
We agree that generally the use of acronyms should be restricted. Here we really try to use the least acronyms possible to convey the necessary details, and to avoid confusion we added a list of abbreviations at the beginning of the manuscript.
Supplement:
In the section regarding air traffic disruptions, I think the Sara et al. Reference Barsotti et al.:
Sara, B., Simona, S., Giovanni, M., Alicia, F., Aline, P., Georgios, V., de Zeeuw van Dalfsen, E., Lars, O., Adriano, P., Jean-Christophe, K., Susan, L., Rita, C., Mauro, C., Jordane, C., Charlotte, V. B., Mauro, D. V., de Chabalier, J. B., Teresa, F., Fontaine Fabrice, R., Arnaud, L., Rui, M., Joana, M., Roberto, M., Anne, P. M., Jean-Marie, S., Ivan, V., Kristín, V., Samantha, E., and Giuseppe, S.: The European Volcano Observatories and their use of the aviation colour code system, Springer Berlin Heidelberg, https://doi.org/10.1007/s00445-024-01712-0, 2024
We will add this reference to the revised version.
Also, on the tephra-induced impacts there are more references that could be provided.
References will be added through addressing previous comments.
With regards to the selection and classification of exposure assets (S3), I think it should be discussed and compared with other existing classification and taxonomies (e.g. the GED4ALL) and their adoption for exposure assessment in other study areas. Exposure is inherently multi-hazard, so the approach would benefit from the adoption of a more general taxonomy. Also, I suggest discussing how local-scale methods such as the one proposed here can be integrated in the overall multi-hazard risk assessment.
Changing the exposure classification scheme is not feasible at this point in the study unfortunately, however we will add brief mention of existing global taxonomies (e.g. GED4ALL) and why they were not appropriate here. And discussion of the integration of local-scale methods, like those employed here, with large scale multi-hazard risk assessment will be expanded on in the discussion
Citation: https://doi.org/10.5194/egusphere-2025-986-AC1
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 179 | 128 | 12 | 319 | 18 | 11 | 26 |
- HTML: 179
- PDF: 128
- XML: 12
- Total: 319
- Supplement: 18
- BibTeX: 11
- EndNote: 26
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1