the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 11 Sep 2026
A disruption scoring framework for critical infrastructure exposed to multiple hazard sources: tephra fall on Japan's electricity network
Abstract. Assessing how hazards disrupt critical infrastructure systems is challenging, as these systems often comprise spatially distributed assets whose importance to service functionality varies across the network. This challenge is particularly acute in volcanic regions, where infrastructure may be exposed to hazards from multiple volcanoes with different eruption frequencies and impact footprints. This makes it hard to know where to focus resources for response, recovery and resilience planning ahead of an event. Here we present a probabilistic disruption scoring framework that combines hazard intensity, spatial extent, and occurrence probability with infrastructure exposure, asset criticality, and vulnerability to quantify the relative disruption potential of different hazard sources and locations. Asset criticality is represented by the number of people served, enabling impacts on individual infrastructure assets to be translated into a common disruption metric. We apply the framework to mainland Japan to assess potential electricity disruptions from tephra fall, considering impacts on power plants, transmission towers and substations exposed to simulated tephra fall events from 62 active volcanoes. A disruption score was calculated for each hazard footprint and used to rank the disruption potential of different tephra fall events and source volcanoes. We also quantified the contribution of different spatial areas to the annual expected electricity disruption across mainland Japan to identify locations with high disruption potential and disruption hotspots. Conditional on an eruption occurring, Fujisan is the volcano with the highest disruption potential because of its proximity to Tokyo and high criticality electricity infrastructure in central Japan. When eruption probabilities are incorporated, Asamayama emerges as the largest contributor to annual expected electricity disruption. Disruption hotspots are concentrated in major metropolitan centres, including Tokyo, Osaka, and Nagoya, as well as regions containing high criticality electricity generation facilities such as Tohoku. These findings provide a quantitative basis for prioritising volcanic risk reduction and infrastructure resilience measures. As the framework uses open-source globally available datasets and models, it can be readily applied to other hazards, infrastructure sectors, and geographic regions, providing a scalable screening tool for identifying priority hazard sources, locations, and infrastructure assets where more detailed risk and resilience investigations are likely to be most beneficial.
- Preprint
(2770 KB) - Metadata XML
-
Supplement
(417 KB) - BibTeX
- EndNote
Status: open (until 23 Oct 2026)
- RC1: 'Review of egusphere-2026-5468', Anonymous Referee #1, 05 Oct 2026 reply
-
RC2: 'Comment on egusphere-2026-5468', Anonymous Referee #2, 06 Oct 2026
reply
Review of the paper "A disruption scoring framework for critical infrastructure exposed to multiple hazard sources: tephra fall on Japan's electricity network"
Summary
This manuscript presents a probabilistic disruption scoring framework to quantify the impact of explosive volcanic eruptions on Japan’s electricity network. The approach integrates probabilistic hazard modeling and occurrence probability with infrastructure exposure, asset criticality, and vulnerability to quantify the disruption potential resulting from tephra fall. A total of 62 active Holocene volcanoes was considered, from which 310,000 simulated tephra fall footprints were derived for different eruption scenarios and atmospheric conditions. The criticality of each infrastructure asset (power plants, transmission towers, and substations) was quantified by considering the number of people served, and a footprint disruption score was computed for each tephra footprint. From these, an average annual disruption score was calculated by grouping footprints by VEI class, volcano, prefecture, or grid cell. In a scenario-based evaluation, Fujisan exhibits the highest disruption potential owing to its proximity to Tokyo and the high concentration of critical power infrastructure in central Japan. However, once eruption probabilities are integrated, Asamayama becomes the primary contributor to the expected annual electricity disruption.
General comments
I found this manuscript interesting and generally well written. The proposed framework provides a useful way to combine tephra hazard, infrastructure vulnerability and asset criticality at a national scale, and I think the approach could also be applied to other regions and infrastructure systems. The figures and tables are generally clear, and the Supplementary Information provides useful detail on the modelling choices. I particularly appreciated the distinction between unweighted FDS and probability-weighted AADS, which helps separate potentially severe scenarios from those contributing most to long-term expected disruption. I have a few comments that I think would help clarify some methodological assumptions and strengthen the interpretation of the results.
My main comments are the following:
1. Although clearly acknowledged in the 'Limitations and future work' section, I believe the assumption of using tephra fall simulations from Fujisan to represent all volcanoes needs further justification. Atmospheric variability plays a major role in volcanic hazard assessment (as also highlighted in the manuscript), making the spatial shifting of Fujisan's footprints across all 62 volcanoes a rather strong assumption. While I understand the computational rationale behind this choice, the authors should provide a comparison of tephra footprints for selected VEIs and seasons using actual wind profiles for 2-3 representative volcanoes (e.g., in northern and southern Japan). Such a comparison would help assess the robustness of this assumption and further support its use in a national-scale screening framework.
2. The limitations section could also briefly acknowledge the uncertainty associated with the eruption frequency estimates. Although the methodology accounts for incompleteness in the eruption record, VEI-specific probabilities are necessarily constrained by the available historical/geological record and, for some magnitude classes, by information from analogue volcanoes. Since these probabilities directly influence the probability-weighted disruption scores and volcano rankings, it would be useful to mention this source of uncertainty when interpreting the AADS results.
3. I understand the need to use relatively simple proxies for asset criticality at this scale. However, different approaches are used for power plants, transmission towers and substations. It would be useful to briefly discuss whether these differences may affect the relative contribution of the three asset types to the final disruption score.
Point by Point comments
131: Could the authors provide the name of the power plant that was removed from the dataset?
164: The statement that “networks containing more towers offer greater redundancy and alternative pathways for power transfer” could be clarified. The number of towers within a service area may also reflect factors such as terrain, transmission-line length, or engineering constraints, and may not necessarily correspond directly to network redundancy. I suggest rephrasing or further justifying this assumption underlying the tower criticality metric.
190: Tephra2 could be very briefly described here, particularly its main function and role within TephraProb, for readers who may not be familiar with the model.
224: Please add the acronym (CW) after "consequence weights" for consistency with the rest of the text and figures
Figure 2: The panels could be arranged in a more intuitive order. I also suggest increasing the font size of the tick labels for better readability. In addition, using different line styles in combination with colours for the different months would make the profiles easier to distinguish and improve accessibility.
Figure 8: This is an important figure, but the volcano names, rankings, and rank-change values are difficult to read at the current size. I suggest enlarging the figure substantially or considering whether panels (a) and (b) could be presented separately to improve readability.
Citation: https://doi.org/10.5194/egusphere-2026-5468-RC2
Model code and software
CI-Disruption-JP Rui Xue Natalie Teng et al. https://github.com/vharg/CI-Disruption-JP
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 263 | 113 | 48 | 424 | 63 | 51 | 58 |
- HTML: 263
- PDF: 113
- XML: 48
- Total: 424
- Supplement: 63
- BibTeX: 51
- EndNote: 58
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The manuscript addresses an important question in volcanic risk assessment on how to prioritise mitigation measure when an electricity network is exposed to tephra from volcano eruptions. Integrating probabilistic hazard information with infrastructure vulnerability and criticality provides a potentially useful national-scale screening framework. It aimed to have a great contribution to further support more targeted risk assessment and resilience planning, especially for country with many volcanic hazard sources can affect interdependent critical infrastructures.
However, the operational and physical assumptions underlying the disruption scores require more justification and substantial revision. Power-plant criticality is derived from annual generation without taking account on types of power generation technology, operating role, or the availability of replacement supply. Furthermore, electricity vulnerability is represented using thickness-based fragility functions without explicit moisture conditioning, despite stating the importance of wet-ash electrical behaviour. The exposure (e.g., population data) and electricity datasets also require updating or validation against current official Japanese statistics (such as national census). These assumptions directly affect the interpretation of the scores as indicators of electricity-service disruption and consequences for the population.
I recommend major revision. The revision should include targeted changes to the criticality and vulnerability representations, and further testing of the resulting rankings and hotspots. Mainly, the revision centred on:
These substantive revisions are necessary and not matters that can be resolved solely by expanding the limitations section. The objective remains worthwhile, but the conclusions should be supported by more robust evidence, given assumptions relevant to electricity-system operation and tephra-induced electrical faults.
Specific comments available in the attached pdf.