the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 23 Jan 2026
Rating telecommunication towers by social importance and physical vulnerability to wildfire
Abstract. Telecommunication infrastructure and the service it provides is vital for notifying threatened populations during wildfire events to ensure timely and safe evacuations. Much of the telecommunication infrastructure in the province of Alberta, Canada, is situated in fire-prone regions, with over 600 individual cell tower sites within the Forest Protection Area (FPA). Proactive measures such as fuel reduction treatments can reduce risk to communication infrastructure and support community resilience, but with limited resources available, there is a need to prioritize the most important and vulnerable towers for protective action. This study presents a structured framework for rating telecommunication infrastructure based on multiple factors of vulnerability. We developed an index that combines indicators of social importance (network redundancy, robustness, and intersection with developed lands and transportation infrastructure) and indicators of physical vulnerability to wildfire (fire exposure, directional vulnerability, slope, and suppression capability) to rate telecommunication towers in Alberta's FPA. In total, 34 telecommunication towers were identified as having both high social importance and high physical vulnerability to wildfire, with almost a quarter located in a single region, the Grande Prairie Forest Area. Potential applications of our Social Importance and Physical Vulnerability (SIPV) Index for informing protection and mitigation decisions are discussed.
- Preprint
(5702 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2026-276', Anonymous Referee #1, 03 Feb 2026
-
RC2: 'Comment on egusphere-2026-276', Anonymous Referee #1, 03 Feb 2026
Publisher’s note: this comment is a copy of RC1 and its content was therefore removed on 3 February 2026.
Citation: https://doi.org/10.5194/egusphere-2026-276-RC2 -
RC3: 'Comment on egusphere-2026-276', Anonymous Referee #1, 03 Feb 2026
Publisher’s note: this comment is a copy of RC1 and its content was therefore removed on 3 February 2026.
Citation: https://doi.org/10.5194/egusphere-2026-276-RC3 -
RC4: 'Comment on egusphere-2026-276', Anonymous Referee #1, 03 Feb 2026
Publisher’s note: this comment is a copy of RC1 and its content was therefore removed on 3 February 2026.
Citation: https://doi.org/10.5194/egusphere-2026-276-RC4 -
RC5: 'Comment on egusphere-2026-276', Anonymous Referee #1, 03 Feb 2026
Publisher’s note: this comment is a copy of RC1 and its content was therefore removed on 3 February 2026.
Citation: https://doi.org/10.5194/egusphere-2026-276-RC5 -
RC6: 'Comment on egusphere-2026-276', Anonymous Referee #1, 03 Feb 2026
Publisher’s note: this comment is a copy of RC1 and its content was therefore removed on 3 February 2026.
Citation: https://doi.org/10.5194/egusphere-2026-276-RC6 -
RC7: 'Comment on egusphere-2026-276', Anonymous Referee #2, 06 Jul 2026
I have read the manuscript “Rating telecommunication towers by social importance and physical vulnerability to wildfire.” The manuscript addresses an important and timely problem, i.e., how to prioritize wildfire mitigation for critical telecommunications infrastructure by considering both physical wildfire-related vulnerability and the social consequences of service disruption. The topic is relevant to natural-hazard risk assessment, emergency management, and critical-infrastructure resilience.
The manuscript is clearly written overall, and the proposed Social Importance and Physical Vulnerability Index has potential as a practical screening tool. However, I do not think the manuscript is ready for publication in its present form. The main areas of opportunity concern the transparency and reproducibility of index construction, the conceptual framing of exposure, vulnerability, importance, and consequence, and the absence of robustness or validation checks.
My comments below focus on issues that should be addressed before the manuscript should be considered for publication in my opinion. I also concur with several points raised by RC1, especially regarding the missing classification table, the need for Social Importance and SIPV equations, and clearer explanation of the scoring procedure.
Specific comments:
- Clarify the terminology relative to standard risk concepts.
The manuscript should more clearly position the SIPV framework relative to standard hazard-risk terminology, especially the distinction among hazard, exposure, vulnerability, and consequence. “Fire exposure” is included as a physical-vulnerability indicator, but exposure and vulnerability are usually treated as separate concepts. Similarly, the “social importance” index assigns higher scores to towers with less redundancy, because Coverage Overlap and Nearby Towers enter negatively. This appears to measure irreplaceability or criticality-given-loss rather than social importance in a general sense. The authors should explain where each component sits in the risk framework and either justify the current labels or consider alternative terminology, such as “criticality of loss,” “service criticality,” or “consequence potential.” - Provide a complete and reproducible index-construction workflow.
The current description does not allow the reader to reproduce the SIPV Index. Physical-vulnerability indicators are rescaled and binned, while social-importance indicators are z-score standardized and summed. The resulting sub-indices are then classified using different breakpoints before being collapsed into the final three-level bivariate classification. The manuscript would be improved by including one consolidated table showing, for every indicator: raw metric, units, directionality, transformation, bin thresholds or standardization method, weighting, aggregation rule, and final classification. Equations should be provided for the Social Importance Index and the final SIPV classification, not only for the Physical Vulnerability Index. - Justify cross-axis comparability.
Because the two SIPV axes are constructed using different normalization and classification procedures, “high” social importance and “high” physical vulnerability are not necessarily equivalent statistical or conceptual categories. It will be helpful if the authors justify why these two differently constructed composites can be placed on a common three-level bivariate axis. This is especially important because the principal applied finding depends on identifying towers in the high-high class. - Add sensitivity, robustness, or uncertainty analysis.
The manuscript uses equal weighting and classification thresholds that are acknowledged as simplifying assumptions, but it does not test whether the results are stable under alternative assumptions. For a composite-index manuscript, this is arguably a substantive limitation. The manuscript will be improved if the authors test whether the high-high tower group remains stable under reasonable alternative weighting schemes, different classification thresholds, and alternative normalization approaches. At minimum, the authors should report how many of the 34 high-high towers remain in that category under a small set of alternative specifications. - Assess potential multicollinearity and double-counting.
Several social-importance indicators are derived by intersecting the same modelled tower coverage polygons with different spatial layers. Larger coverage polygons may tend to have greater road coverage, greater human-footprint coverage, and greater coverage overlap simply because of their size. Road density and human-footprint density may also be correlated. The additive z-score approach assumes that indicators contribute independent information. The authors should report an inter-indicator correlation matrix and either adjust for redundancy, for example using PCA or another dimensionality-reduction approach, or justify retaining the equal-weight additive construction. - Clarify the implications of sample truncation.
The analysis is limited to towers already identified as having extreme wildfire exposure in Kuiper and Beverly (2025). This is a defensible screening design, but the scope condition should be stated more clearly throughout the manuscript. The SIPV Index is not ranking all telecommunications towers in Alberta’s Forest Protection Area; it is re-ranking a subset of already high-exposure towers. This also affects the distribution of the fire-exposure variable and its discriminatory power within the Physical Vulnerability Index. The authors should also discuss how this restricted sample affects normalization, classification, and interpretation. - Add validation or temper claims.
The manuscript identifies 34 towers as high social importance and high physical vulnerability, but no validation or plausibility check is provided. The authors should consider comparing the results with documented telecommunications outages, tower damage, historical wildfire perimeters, agency priority sites, provider knowledge, or expert elicitation. If validation data are not available, claims such as “most at-risk” could be softened. The index would then be better described as a screening or prioritization framework rather than a validated measure of actual risk. - Clarify the aggregation behavior of the Physical Vulnerability Index.
Averaging binned indicators can dilute extreme values. A tower with very high fire exposure or poor suppression access may receive only a moderate composite score if other indicators are lower. This matters for a screening tool intended to identify priority infrastructure. The authors should explain why an arithmetic mean is appropriate and consider whether alternative aggregation rules, such as weighted averages, maximum operators, or non-compensatory methods, would better preserve single-indicator extremes. - Clarify the role of the human footprint indicator.
The human footprint layer appears to be used as a proxy for people, development, or societal consequence of service loss, not as a hazard, ignition source, or direct physical-vulnerability measure. This should be stated explicitly. The authors should also discuss what the indicator may miss, such as transient recreation users, seasonal populations, emergency responders, Indigenous communities, or road users outside mapped developed footprints. - Discuss operational telecommunications limitations.
Coverage overlap and nearby tower counts are reasonable proxies for redundancy, but they do not necessarily represent actual service continuity. Redundancy depends on provider ownership, roaming arrangements, backhaul, power supply, antenna orientation, network capacity, battery and generator availability, and whether adjacent towers can absorb traffic after a failure. The authors should discuss these limitations more directly.
Technical corrections:
- Table 3 appears to be missing, despite being referenced in the manuscript.
- The Nearby Towers radius is inconsistent: Table 1 states 20 km, while Section 2.3.2 states 40 km.
- Directional vulnerability requires clarification. Table 1 refers to continuous high fire exposure greater than 60%, while Section 2.4.1 refers to an 80% intersection threshold.
- The Suppression Capability metric should be stated more precisely. “Nearest average distance” is ambiguous, and airtanker bases and skimmable lakes represent different operational capabilities.
- Equations should be added for the Social Importance Index and the final SIPV classification.
- Equation (1) appears corrupted in the PDF and should be reset.
- Section numbering should be corrected; “2.4” appears to be used twice.
- Section 2.3.5 should be titled “Social importance index,” not “Social vulnerability index.”
- In Table 1, consider changing the first column heading from “Variable” to “Indicator,” consistent with the caption.
- Tables 4 and 5 should define the numbers in parentheses, for example count and within-class percentage.
- FPA should be defined clearly at first use as Alberta’s Forest Protection Area and as the zone of provincial wildfire-management responsibility.
- The coverage-area estimation method from Kuiper and Beverly (2025) should be briefly summarized because it underlies several social-importance indicators.
- The exposure definition from Kuiper and Beverly (2025) should also be summarized rather than left entirely to the cited paper.
- The manuscript should avoid describing the collapsed high-high SIPV class as “critically vulnerable” unless “critical” is formally retained as a final SIPV category.
- “Slake Lake” should be corrected to “Slave Lake.”
- The DEM source in Table 1 appears to contain a formatting problem: “[NO_PRINTED_FORM].”
- The manuscript should check minor grammar issues, including “for all tower,” “each towers communication viewshed,” and “used as threshold.”
Citation: https://doi.org/10.5194/egusphere-2026-276-RC7 - Clarify the terminology relative to standard risk concepts.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 1,130 | 609 | 153 | 1,892 | 89 | 196 |
- HTML: 1,130
- PDF: 609
- XML: 153
- Total: 1,892
- BibTeX: 89
- EndNote: 196
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
I have now read the manuscript with the title: “Rating telecommunication towers by social importance and physical vulnerability to wildfire”. The paper is among the few studies that focus on the vulnerability of critical infrastructure to wildfire and that integrate physical vulnerability with societal importance. Research in this direction is both necessary and highly timely, particularly in light of the 2023 EU Directive on the Resilience of Critical Entities. The paper is well written and clearly structured. Nevertheless, it would benefit from a clearer presentation of the method. I also have some additional comments that would improve the paper significantly and that I strongly believe should be addressed prior to publication. These include:
Score
Indicator
1
2
3
4
Slope
0-10°
10°-20°
20°-30°
> 30°
Agueda et al. 2023: https://doi.org/10.1071/WF22134
Papathoma-Köhle et al. 2025: https://doi.org/10.1071/WF24114