the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 17 Apr 2025
Is Europe’s transport infrastructure ready to face climate change?
Abstract. Climate extremes are intensifying in frequency and severity, posing escalating risks to Europe’s transport infrastructure. Over the past two decades, these events have caused physical damages, economic losses, user delays, and rising health emergencies. Here, we first review observed impacts of selected climate extremes on Europe’s transport network. We then present a multi-hazard analysis, designed to be intuitive and accessible for policymakers, to quantify the increasing exposure of four transport modes to river floods, heatwaves, droughts, tropical cyclones, and wildfires. We find that climate extremes are already affecting all transport modes across large parts of Europe, with billions euro worth of economic damage even for single events. For example, the 2018 Rhine River drought, which resulted in EUR 2.4 billion in economic losses in Germany. Under a medium-high emissions scenario (RCP6.0), exposure of critical European transport infrastructure to climate extremes is projected to rise, with heatwave exposure increasing up to 70 times by mid-century (2024–2075) compared to historical conditions (1954–2005). Droughts could affect 40 % of inland waterways for the first time by mid-century in this scenario. By the end of the century (2049–2100), railways and roads are projected to face up to 42 times as many wildfires compared to the historical period and exposure to river floods is projected to increase locally up to 13 times. Mitigation efforts aligned with a low-emissions scenario (RCP2.6) demonstrate clear benefits, reducing EU-mean transport mode exposure to heatwaves by 26–32 %, wildfires by 17–38 %, floods by ~20 %, and droughts by 10–20 %, in terms of relative differences, by the end of the century. Urgent action is needed to strengthen the resilience of Europe’s transport network and address emerging climate challenges. A combination of ambitious mitigation and adaptation strategies is essential to ensuring their long-term functionality in a changing climate.
- Preprint
(83035 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-1697', Anonymous Referee #1, 04 Jun 2025
General comment
Firstly, I would like to thank the authors for writing an informative paper with the objective of (I) providing a review of literature about impacts of various climate extremes on Europe’s transport network, and (II) analyzing potential future exposure under three emissions scenarios. While the work steps are generally well explained and based on freely accessible data and code, major improvements would make it easier for readers to understand the study’s findings and methodology. The results of the literature review should be more concise, highlighting key findings from relevant literature and linking them to the analysis of future exposure. Regarding the exposure analysis, the paper would benefit from including further high-impact natural hazards, namely windstorms and coastal flooding, and excluding tropical cyclones (TCs). In addition, using comparable time windows and providing more information about the sensitivity analysis would facilitate the interpretation of the findings. The paper as a whole needs to be more concise, not just the results of the literature review. The following paragraphs describe my section-specific comments and overarching issues, followed by technical corrections.
Major comments
Abstract
In my opinion, the authors should mention the median values of potential future exposure besides the maximum values. It would help to give readers a better understanding of the results, as median values show no increase in exposure to river floods, wildfires, and TCs by mid- and end-of-century.
Additionally, the abstract would be accessible to a wider audience by mentioning that RCP6.0 best resembles the extrapolation of current climate policies. It would clarify that the findings relate to a 'most likely' emissions scenario (L 8ff). Having said that, a later comment of the review relates to the use of RCP6.0. Another minor issue, the baseline period should be mentioned when reporting the relative differences in exposure under RCP2.6 (L 13ff).
Introduction
The introduction could be improved by highlighting the key aspects that demonstrate the added value of the literature review (namely the widespread impacts of past extreme events) and the analysis of future exposure, for example, by referencing relevant studies (e.g. Forzieri et al. 2018).
Forzieri, G., Bianchi, A., Silva, F. B. E., Marin Herrera, M. A., Leblois, A., Lavalle, C., Aerts, J. C. J. H., & Feyen, L. (2018). Escalating impacts of climate extremes on critical infrastructures in Europe. Global Environmental Change, 48, 97–107. https://doi.org/10.1016/j.gloenvcha.2017.11.007
Methods
In case the TEN-T network dataset provided by the European Commission is open accessible, a weblink or citation of the used shapefile would further support the open access approach of the study.
In addition, the authors should make it explicit that they assumed the same extent of the TEN-T network already for the reference period (1965-2005) to conduct the analysis (see, e.g., Figure 8 – years: 1965, 1995).
Section 2.2.2 mentions that the GDIS dataset comprises reports about disasters from 1960-2018 (L 85); the same applies for the results section (L 200). However, the study only uses events from the GDIS dataset between 2010 and 2018. It would be clearer for the reader if mentions of the GDIS dataset from before this period were removed.
The lines 99-103 mention that the case studies would be a balanced representation of extreme events in Europe. However, the paper focuses on case studies from Western and Southern Europe. The inclusion of at least one case study of Central and Northern Europe would support their pan-European approach.
On page 6, the authors state that they conducted a sensitivity analysis to determine at which threshold localized extreme events (e.g. river flooding) affect nearby transport infrastructure. Based on the ISIMIP resolution of approximately 55.5 x 55.5 km at the equator level, this would mean that approximately 9.6 x 9.6 km of this area would need to be impacted by the event to affect the local transport infrastructure. Further explanations of how the sensitivity analysis was performed (without the need to look into the respective python script) would better justify the 3% threshold.
Also, on page 6 and the subsequent page the authors mention the first time the ensemble median EMF. It would be beneficial to highlight that the median EMF is based on the ensemble of global climate and impact models that were described earlier. In addition, the terminology should not change throughout the text, for instance, it is referred to as “ensemble median EMF” (page 7), as “multi-model ensemble median EMF” (L 433, Table 6 and A6, and Figure 7-11), and as “impact-climate model ensemble median EMF” (L 513f).
Abbreviations in the figures, tables and in their captions (e.g. GCMs, EMF) should be explained close to the figure/table, so readers could interpret the content without looking up the abbreviations in the main text. Especially, this applies to Table 6 since it summarizes main findings of this study - it would be beneficial to describe briefly the meaning of EMF >100%, <1% and >1%.
The paper does not state clearly which countries from the GDIS dataset and the TEN-T network were considered - Figure 2 is misleading in this context as well as similar figures later on. The current TEN-T Network plans include Ukraine, Moldova and Turkey, but exclude Russia and Belarus. However, Figure 2 shows GDIS polygons (but without TEN-T road network) referring to regions in Russia, Belarus and Turkey.
The authors declare their focus on RCP6.0 that “it most closely resembles the extrapolation of current climate policies into the future” (L 148), but literature is missing which supports this statement.
Results
The results of the literature review (Section 3.1 and its subsections) should be more concise focusing on the important findings of the review, link them better to the second part of the results section (i.e., the future exposure) and relocate some information, such as Tables 3 to 5 to the Appendix. For instance, the paragraph about the study from Verschuur et al. 2023 is located in the results (L 280 –293) but deals with exposure of port infrastructure worldwide. However, the study area of this paper is Europe, so this paragraph (in its current form) does not fit to this section.
Related to the previous comment, a summary table with all hazard types and transport modes considered should be created, summarizing the damage mechanism and relevance. Such a table would give a clear and concise overview of the key findings from the literature review and better link them to the analysis of future exposure.
I would also suggest having a look into literature to verify some links between hazards and transport modes. For instance, the cited webpage (Systems, 2024) is about drought impacts on airports in the U.S., but is there also evidence for Europe? The impact factors, like soil type and drought duration, might differ between both continents. If there is no evidence, I would also advise removing the corresponding parts from Section 3.2. (L 474) and Table 6.
Two minor comments regarding Section 3.1:
- Remove Figure 1 as it only contains information already included in the corresponding sentence (L 190-192). In addition, this sentence should be rather part of the introduction as it is not a finding of this study.
- Adapt sentence in line 193ff in the way that it underpins the key message of Figure 3.
In Section 3.2, the authors analyze potential future impacts on different transport modes, including coastal ports. However, they do not include coastal floods as an additional extreme event category in their exposure analysis (see, Table 1), despite they highlight in line 581f the increased risk it poses in a warmer climate; the same applies to storms. In addition, the impacts of coastal floods/storms under current climate are already mentioned (e.g., Sections 3.1.1 and 3.1.2). I would recommend either including coastal floods/storms to the exposure analysis by using an additional dataset or discussing the projected changes and impacts of coastal floods/storms for Europe based on relevant literature (besides Arias et al., 2021 for windstorms).
Related to the previous comment, the inclusion of TC impacts seems to be rather confusing, primarily due that no historical impacts of TCs on Europe’s infrastructure were presented before. The literature review only mentions the impacts of TCs worldwide, such as in the paragraph about the findings from Verschuur et al. 2023 regarding damages to port infrastructure worldwide (L 280ff). Therefore, I would recommend excluding TCs from the study, or at most refer to them in one or two sentences.
The results for RCP2.6 and RCP8.5 are reported in the Appendix but only briefly mentioned at the very end of the results. Thus, the sentence “Results for RCP2.6 and RCP8.5 are reported in the Appendix A and discussed in the main text” (L 149) is misleading.
Time ranges (e.g. L 197ff, 387ff, 402ff): For the reader it is impossible to set the different time ranges in context to each other. Furthermore, it leads to additional text and distracts the reader's attention. A 52y overlapping window is not recommended; this time span is uncommon and the overlap is less useful when comparing changes in exposure over time. I propose using only a few comparable time ranges, ideally non-overlapping ones. Only when unavoidable, the authors should provide explanation why they deviate. Below, are some examples where the text should be improved:
- The time range of 2010-2018 used for the GDIS dataset should be aligned with the other time ranges of the study (e.g. a division of the 30y windows); the same applies to the DRMKC data, ranging from 2010-2022 (L 387).
- The findings about the exposure to future heatwaves/wildfires include once the changes based on 30y non-overlapping and on 52y overlapping windows (page 20 and 21). Please use only one (non-overlapping) time window.
- It seems that the time range for mid- and end-of-century spans 52 years, so line 169 is misleading.
- L 157 is misleading, too. It refers to the baseline period as a 41y timespan (1965-2005).
The statement in line 234f is only partially supported by Table 3 - the table does not provide a quantification over a longer time range as stated in the corresponding sentence. The authors should find literature to support the statement of line 234f. The same applies to Table 4 (L 294-296) and Table 5 (L 357f).
In my opinion the text part about landslides triggered by flash floods and preceding droughts should be set more into context to roads and railways (L 367-371). Furthermore, the last sentence of the text part is just an explanation of processes described in previous sentences, thus, it can be taken out.
Overall, the authors should try to make the manuscript more concise by removing redundant information and ensuring that the explanations support the messages of the text. Some of the previous comments are already related to this issue. Below, are some further examples where in my opinion the text contains redundancies or is too detailed, mainly focusing on Section 3.2.:
- It is sufficient to mention examples of indirect and non-monetary losses just once, as done when cited Bednar-Friedl et al. (2022) or Eneir et al. (2010). In the subsequent text, repetition of these examples should be avoided. If necessary, they could be referred to as non-monetary losses and indirect impacts
- L 401 - 404: I think both sentences can be condensed into one sentence.
- L 438 - 442: Also, here make it more concise, e.g. (L438f) the increase in the ensemble mean is similar across all transport modes, thus it is sufficient to report only a range with 30.8% (waterways) - 37.9% (airports).
- L 469 - 472: Try to include the second finding about road exposure to the previous sentence since all three network types face similar strong exposure
- L 479ff: It seems to be redundant to L475f
Table 6 would benefit from some adaptations (same as for Table A5 and A6): The authors should consider using Q10 and Q90 (instead of Q20 and Q80), since they report only the former in Section 3.2 and Figure 6. Also, the total number of ensemble members used to calculate the median values and percentiles should be provided in the table.
In addition, underline the finding that the Mediterranean region will face the strongest increase in number of heatwaves and wildfires across Europe by mentioning Figure 7 and 9 (L 458f). This would support both the paragraph and the corresponding sentence in the discussion that reflects this result (L 562f).
Discussion
While the study solely focuses on infrastructure exposure, it would benefit from discussing briefly the aspect of vulnerability at least in this section. For instance, how the TEN-T network is specifically designed to withstand more frequent and intense climate extremes.
Subsection 4.1: The authors should try to adapt parts of the second last paragraph to fit better to the topic of this subsection, namely past events. In addition, the text about the future impacts should be moved to the subsequent section.
Subsection 4.2: The paragraph about drought impacts is not linked to the analysis of future exposure, furthermore, it contains explanations about drought impacts which should not be part of discussion.
Subsection 4.2.1 need to be better linked to examples of compound events from the literature review (e.g. co-occurrence of wildfires and heatwaves in Sicily).
Conclusions
I would recommend revising this section completely. The additional work can improve how readers perceive the study since most of them will read after the abstract the conclusions to get an impression of the study. Thus, it should be concise and well-devised. For instance, in its current form the section lacks a summary and assessment of the study’s key findings.
Minor comments
- Table 1: The caption should be moved to the top of the table, as well as the text with the references to global climate and impact models.
- Consistent writing of “en route”
- Figures in the appendix (A1 – A4): their information should try to be condensed and presented more clearly. Readers might find it difficult to detect the differences between these figures. For instance, it might improve the quality of the figures by representing adjacent events of the same hazard types by one symbol with a size according to the number of events it represents
- Table A3: the same sentence appears twice: “The initial package of as much as 2.25 billion could be complemented by other sources like the EU’s solidarity fund, von der Leyen said.” and the table caption should refer to Greece
- L 9: missing whitespace
- L 69: abbreviated RCP without entire name at first mention, same as for GDIS dataset L 77
- L 21: is number from EEA 2023
- L 41f missing word after “,”
- Make figures colorblind friendly
- Figure 8: similar red tones should be avoided, and the black error bars are barely visible
- Figure 5: Merge both plots as they refer to the same information
- Years for in-text citations are occasionally missing, please have a look at the NHESS guidelines for citations, e.g. L 111
- Lange et al. 2020b is the same as Lange et al. 2020a
Citation: https://doi.org/10.5194/egusphere-2025-1697-RC1 -
AC3: 'Reply on RC1', Cristina Deidda, 22 Aug 2025
Dear reviewer,
We would like to thank you for the detailed review and useful suggestions to improve the manuscript.
Please find below our point-by-point responses (in bold) to each of your comments.
General comment
Comment:
Firstly, I would like to thank the authors for writing an informative paper with the objective of (I) providing a review of literature about impacts of various climate extremes on Europe’s transport network, and (II) analyzing potential future exposure under three emissions scenarios. While the work steps are generally well explained and based on freely accessible data and code, major improvements would make it easier for readers to understand the study’s findings and methodology. The results of the literature review should be more concise, highlighting key findings from relevant literature and linking them to the analysis of future exposure. Regarding the exposure analysis, the paper would benefit from including further high-impact natural hazards, namely windstorms and coastal flooding, and excluding tropical cyclones (TCs). In addition, using comparable time windows and providing more information about the sensitivity analysis would facilitate the interpretation of the findings. The paper as a whole needs to be more concise, not just the results of the literature review. The following paragraphs describe my section-specific comments and overarching issues, followed by technical corrections.
Answer:
We would like to thank you for your constructive review, which has been very helpful in improving the manuscript. We agree that the literature review should be reshaped to better highlight key findings and their link to our analysis of future exposure. We will exclude tropical cyclones from the analysis and incorporate more information about additional hazards, such as coastal flooding. We will also provide a clearer explanation of the rationale for using different time windows. As explained in the details in the comments section, one of the most difficult part of this literature review is the absence of a database uniformed that give information for the same time windows, for all hazards, of economical losses and impacts on transport modes. This limitation force to do the analysis with the available data for each different dataset, creating different time window.
Major comments
Abstract
Comment:
In my opinion, the authors should mention the median values of potential future exposure besides the maximum values. It would help to give readers a better understanding of the results, as median values show no increase in exposure to river floods, wildfires, and TCs by mid- and end-of-century.
Answer:
Thank you for the helpful comment. We initially reported maximum values to highlight critical local increases in exposure across the TEN-T network, which are particularly relevant for risk management. However, we agree that including median values will offer a more comprehensive understanding of the overall trends. We will revise the abstract and main text to include both metrics, clarifying that while median exposure may remain unchanged in some cases, local increases can still be substantial.
Comment:
Additionally, the abstract would be accessible to a wider audience by mentioning that RCP6.0 best resembles the extrapolation of current climate policies. It would clarify that the findings relate to a 'most likely' emissions scenario (L 8ff). Having said that, a later comment of the review relates to the use of RCP6.0. Another minor issue, the baseline period should be mentioned when reporting the relative differences in exposure under RCP2.6 (L 13ff).
Answer:
Thank you for this helpful remark. We will revise the abstract to indicate that RCP6.0 closely reflects the ‘most likely’ emissions scenario. We will also specify the baseline period used when reporting relative differences under RCP2.6 to improve clarity for a wider audience.
Introduction
Comment:
The introduction could be improved by highlighting the key aspects that demonstrate the added value of the literature review (namely the widespread impacts of past extreme events) and the analysis of future exposure, for example, by referencing relevant studies (e.g. Forzieri et al. 2018).
Forzieri, G., Bianchi, A., Silva, F. B. E., Marin Herrera, M. A., Leblois, A., Lavalle, C., Aerts, J. C. J. H., & Feyen, L. (2018). Escalating impacts of climate extremes on critical infrastructures in Europe. Global Environmental Change, 48, 97–107. https://doi.org/10.1016/j.gloenvcha.2017.11.007
Answer:
We will include this reference in the text.
Methods
Comment:
In case the TEN-T network dataset provided by the European Commission is open accessible, a weblink or citation of the used shapefile would further support the open access approach of the study.
In addition, the authors should make it explicit that they assumed the same extent of the TEN-T network already for the reference period (1965-2005) to conduct the analysis (see, e.g., Figure 8 – years: 1965, 1995).
Answer:
We agree with that, we will specify that we use the same extent of the TEN-T network, adding the weblink.
Comment:
Section 2.2.2 mentions that the GDIS dataset comprises reports about disasters from 1960-2018 (L 85); the same applies for the results section (L 200). However, the study only uses events from the GDIS dataset between 2010 and 2018. It would be clearer for the reader if mentions of the GDIS dataset from before this period were removed.
Answer:
Thank you for this helpful comment. You are correct that our literature review analysis focuses specifically on recent extreme events. We selected the period starting from 2010 to capture the increase in extremes over the last 15 years. The GDIS database provides data only until 2018, so this limitation is due to data availability. We agree that this is not clearly explained in the current version of the manuscript. We will revise the text accordingly to specify the period used in the analysis and to avoid any confusion regarding the broader time span covered by the GDIS dataset.
Comment:
The lines 99-103 mention that the case studies would be a balanced representation of extreme events in Europe. However, the paper focuses on case studies from Western and Southern Europe. The inclusion of at least one case study of Central and Northern Europe would support their pan-European approach.
Answer:
Thank you for this thoughtful observation. The selected case studies are the results of a large research effort and were chosen based on specific criteria: (a) recent extreme events, (b) very high economic damages, (c) impacts across multiple transport modes, and (d) availability of reports and detailed documentation on impacts to transport infrastructure. These criteria considerably reduce the number of events that can be selected, as transport-specific damage and economic loss information is often difficult to obtain. Although the selected events do not fully cover the entire European territory, we believe they still provide a representative overview of major types of extreme events and their impacts on different transport modes. The selected cases highlight some of the most significant recent events and mechanisms across Europe, while also reflecting the regions most affected by extremes in recent years. We will revise the manuscript to clarify these selection criteria and better explain our rationale.
Comment:
On page 6, the authors state that they conducted a sensitivity analysis to determine at which threshold localized extreme events (e.g. river flooding) affect nearby transport infrastructure. Based on the ISIMIP resolution of approximately 55.5 x 55.5 km at the equator level, this would mean that approximately 9.6 x 9.6 km of this area would need to be impacted by the event to affect the local transport infrastructure. Further explanations of how the sensitivity analysis was performed (without the need to look into the respective python script) would better justify the 3% threshold.
Answer:
Thank you for the comment. You are right that additional explanation of the sensitivity analysis would help clarify our choice of the 3% threshold. Some of the extreme event datasets express hazard extent as the percentage of the grid cell affected. Setting this threshold too high risks underestimating exposure, as it may overlook localized but impactful events, particularly river floods. Conversely, a threshold that is too low may overestimate exposure, including cases that are unlikely to result in significant disruption. To address this, we tested several thresholds ranging from 1% to 5% of the cell area affected. At the 5% level, we observed that most of flood events were not captured in the spatial statistics. At the 1% threshold, however, exposure to hazards such as wildfires was overestimated, with large portions of the network appearing to be affected in nearly every case. The 3% threshold emerged as a compromise: it allowed for the detection of localized events like river floods while avoiding overestimation from more diffuse hazards like wildfires. We will revise the manuscript to better explain this sensitivity analysis and the reasoning behind our threshold selection.
Comment:
Also, on page 6 and the subsequent page the authors mention the first time the ensemble median EMF. It would be beneficial to highlight that the median EMF is based on the ensemble of global climate and impact models that were described earlier. In addition, the terminology should not change throughout the text, for instance, it is referred to as “ensemble median EMF” (page 7), as “multi-model ensemble median EMF” (L 433, Table 6 and A6, and Figure 7-11), and as “impact-climate model ensemble median EMF” (L 513f).
Abbreviations in the figures, tables and in their captions (e.g. GCMs, EMF) should be explained close to the figure/table, so readers could interpret the content without looking up the abbreviations in the main text. Especially, this applies to Table 6 since it summarizes main findings of this study - it would be beneficial to describe briefly the meaning of EMF >100%, <1% and >1%.
Answer:
Thank you for pointing this out. We will revise the manuscript to ensure consistent terminology throughout, for example, by standardizing the use of "multi-model ensemble median EMF" across the text, figures, and tables. We will also make it clearer that the median EMF is based on the ensemble of global climate and impact models described earlier in the manuscript. Regarding abbreviations, we agree that figure and table captions should be self-contained. We will review all captions—especially for Table 6—to ensure that abbreviations like GCMs and EMF are clearly explained. We will also expand the caption of Table 6 to provide a clearer interpretation of EMF values >100%, <1%, and between 1–100%.
Comment:
The paper does not state clearly which countries from the GDIS dataset and the TEN-T network were considered - Figure 2 is misleading in this context as well as similar figures later on. The current TEN-T Network plans include Ukraine, Moldova and Turkey, but exclude Russia and Belarus. However, Figure 2 shows GDIS polygons (but without TEN-T road network) referring to regions in Russia, Belarus and Turkey.
The authors declare their focus on RCP6.0 that “it most closely resembles the extrapolation of current climate policies into the future” (L 148), but literature is missing which supports this statement.
Answer:
Thanks for the comment. We will revise and check the figure. We extract with QGIS the events polygons that cover the countries over the EU. Some polygons linked to the same event covered also multiple regions, also not included in the TEN-T. We will revise the figure and add more literature to support the RCP statement.
Results
Comment:
The results of the literature review (Section 3.1 and its subsections) should be more concise focusing on the important findings of the review, link them better to the second part of the results section (i.e., the future exposure) and relocate some information, such as Tables 3 to 5 to the Appendix. For instance, the paragraph about the study from Verschuur et al. 2023 is located in the results (L 280 –293) but deals with exposure of port infrastructure worldwide. However, the study area of this paper is Europe, so this paragraph (in its current form) does not fit to this section.
Answer:
We will revise the results section to place greater emphasis on the key findings of the literature review and strengthen their connection to the analysis of future exposure. The paragraph on the study by Verschuur et al. (2023) will be rewritten to focus solely on the findings relevant to Europe, ensuring that the content aligns with the scope of this paper. Thanks for this suggestion.
Comment:
Related to the previous comment, a summary table with all hazard types and transport modes considered should be created, summarizing the damage mechanism and relevance. Such a table would give a clear and concise overview of the key findings from the literature review and better link them to the analysis of future exposure. I would also suggest having a look into literature to verify some links between hazards and transport modes. For instance, the cited webpage (Systems, 2024) is about drought impacts on airports in the U.S., but is there also evidence for Europe? The impact factors, like soil type and drought duration, might differ between both continents. If there is no evidence, I would also advise removing the corresponding parts from Section 3.2. (L 474) and Table 6.
Two minor comments regarding Section 3.1:
- Remove Figure 1 as it only contains information already included in the corresponding sentence (L 190-192). In addition, this sentence should be rather part of the introduction as it is not a finding of this study.
- Adapt sentence in line 193ff in the way that it underpins the key message of Figure 3.
Answer:
We will consider adding a summary table covering all hazard types and transport modes, summarizing the damage mechanisms and their relevance, to better connect the literature review with the future exposure analysis. We will also revise the information regarding drought impacts on airports; however, we will leave the projection results in Table 6 and Section 3.2 for completeness and coherence. Regarding Figure 1, although it may appear somewhat repetitive, we believe the figure is informative, especially for policymakers and non-academic audiences, as it provides a schematic overview of all the types of impacts. We also agree to adapt the sentence in line 193 accordingly.
Comment:
In Section 3.2, the authors analyze potential future impacts on different transport modes, including coastal ports. However, they do not include coastal floods as an additional extreme event category in their exposure analysis (see, Table 1), despite they highlight in line 581f the increased risk it poses in a warmer climate; the same applies to storms. In addition, the impacts of coastal floods/storms under current climate are already mentioned (e.g., Sections 3.1.1 and 3.1.2). I would recommend either including coastal floods/storms to the exposure analysis by using an additional dataset or discussing the projected changes and impacts of coastal floods/storms for Europe based on relevant literature (besides Arias et al., 2021 for windstorms).
Answer:
Thank you for pointing this out. We are aware of this limitation and initially considered adding another dataset to include coastal floods. However, no dataset is currently available with the same RCP projections and structure as the one used for the other hazards, which would make it difficult to provide a consistent and uniform quantitative analysis. For these reasons, we chose a cleaner approach, maintaining consistency across hazards. Among the two options, we agree that discussing the projected impacts of coastal floods and storms based on relevant literature is the most suitable approach, and we will provide additional information accordingly.
Comment:
Related to the previous comment, the inclusion of TC impacts seems to be rather confusing, primarily due that no historical impacts of TCs on Europe’s infrastructure were presented before. The literature review only mentions the impacts of TCs worldwide, such as in the paragraph about the findings from Verschuur et al. 2023 regarding damages to port infrastructure worldwide (L 280ff). Therefore, I would recommend excluding TCs from the study, or at most refer to them in one or two sentences.
Answer:
Thank you for the constructive comment. We agree and will exclude TC from the analysis.
Comment:
The results for RCP2.6 and RCP8.5 are reported in the Appendix but only briefly mentioned at the very end of the results. Thus, the sentence “Results for RCP2.6 and RCP8.5 are reported in the Appendix A and discussed in the main text” (L 149) is misleading.
Answer:
We will revise the sentence accordingly.
Comment:
Time ranges (e.g. L 197ff, 387ff, 402ff): For the reader it is impossible to set the different time ranges in context to each other. Furthermore, it leads to additional text and distracts the reader's attention. A 52y overlapping window is not recommended; this time span is uncommon and the overlap is less useful when comparing changes in exposure over time. I propose using only a few comparable time ranges, ideally non-overlapping ones. Only when unavoidable, the authors should provide explanation why they deviate. Below, are some examples where the text should be improved:
- The time range of 2010-2018 used for the GDIS dataset should be aligned with the other time ranges of the study (e.g. a division of the 30y windows); the same applies to the DRMKC data, ranging from 2010-2022 (L 387).
- The findings about the exposure to future heatwaves/wildfires include once the changes based on 30y non-overlapping and on 52y overlapping windows (page 20 and 21). Please use only one (non-overlapping) time window.
- It seems that the time range for mid- and end-of-century spans 52 years, so line 169 is misleading.
- L 157 is misleading, too. It refers to the baseline period as a 41y timespan (1965-2005).
Answer:
Thank you for your helpful comment. We acknowledge that the use of different time windows throughout the manuscript may create confusion for the reader. The inconsistency in time periods come from methodological and data-related constraints, which were not sufficiently explained in the original manuscript. We will revise the text to clarify the rationale behind these choices.
- Past event databases (GDIS and DRMKC): The time windows used for the analysis of past extreme events are driven by data availability. Our objective was to review transport-related events from the fifteen years, making use of all available information. The GDIS dataset covers events until 2018, while the DRMKC database includes events up to 2022. While this results in slightly misaligned time windows (2010–2018 vs. 2010–2022), we believe including all relevant data from the most recent years improves the completeness of the analysis. We will update the manuscript to better explain that this section specifically focuses on recent historical events and that we deliberately included the most up-to-date information available, even if this leads to different end years.
- Future projections: The choice of a 52-year window for future projections is guided by the need to assess rare extreme events, such as those defined by 100-year return periods. A 30-year window is often insufficient to robustly estimate such events, as the sample size would be too small. For this reason, we opted for longer time spans to improve the statistical robustness of our extreme event analysis. The overlap between mid-century and end-of-century periods is a consequence of balancing the need for long windows with the limited time horizon of available climate projection data (which ends at the end of the century). While this does create overlap, we will clarify in the text why it was necessary and unavoidable in this context.
- Exposure trends (bar plots in Fig. 10, etc.): In contrast, the bar plots use non-overlapping 30-year windows to illustrate long-term changes in exposure across distinct periods. This choice allows for a clearer visual comparison of trends over time. We will revise the relevant text to emphasize the difference in purpose between the two approaches: the 52-year windows are used to estimate the occurrence of rare extremes, while the 30-year windows are used for analysing broader exposure trends.
- Clarifications to misleading lines: We agree that lines 157 and 169 are potentially misleading and will revise them accordingly.
We understand that different time windows can create confusion. We will adapt the text to provide more information about the underlying reasons, while ensuring that readability is not compromised.
Comment:
The statement in line 234f is only partially supported by Table 3 - the table does not provide a quantification over a longer time range as stated in the corresponding sentence. The authors should find literature to support the statement of line 234f. The same applies to Table 4 (L 294-296) and Table 5 (L 357f).
In my opinion the text part about landslides triggered by flash floods and preceding droughts should be set more into context to roads and railways (L 367-371). Furthermore, the last sentence of the text part is just an explanation of processes described in previous sentences, thus, it can be taken out.
Answer:
We acknowledge that the statements introducing the Table 3, 4 and 5, are not fully supported by the corresponding tables. will add more references to relevant literature to strengthen the argument. We also will adapt the text in lines 367-371 accordingly.
Comment:
Overall, the authors should try to make the manuscript more concise by removing redundant information and ensuring that the explanations support the messages of the text. Some of the previous comments are already related to this issue. Below, are some further examples where in my opinion the text contains redundancies or is too detailed, mainly focusing on Section 3.2.:
- It is sufficient to mention examples of indirect and non-monetary losses just once, as done when cited Bednar-Friedl et al. (2022) or Eneir et al. (2010). In the subsequent text, repetition of these examples should be avoided. If necessary, they could be referred to as non-monetary losses and indirect impacts
- L 401 - 404: I think both sentences can be condensed into one sentence.
- L 438 - 442: Also, here make it more concise, e.g. (L438f) the increase in the ensemble mean is similar across all transport modes, thus it is sufficient to report only a range with 30.8% (waterways) - 37.9% (airports).
- L 469 - 472: Try to include the second finding about road exposure to the previous sentence since all three network types face similar strong exposure
- L 479ff: It seems to be redundant to L475f
Table 6 would benefit from some adaptations (same as for Table A5 and A6): The authors should consider using Q10 and Q90 (instead of Q20 and Q80), since they report only the former in Section 3.2 and Figure 6. Also, the total number of ensemble members used to calculate the median values and percentiles should be provided in the table.
In addition, underline the finding that the Mediterranean region will face the strongest increase in number of heatwaves and wildfires across Europe by mentioning Figure 7 and 9 (L 458f). This would support both the paragraph and the corresponding sentence in the discussion that reflects this result (L 562f).
Answer:
Thank you for your valuable and detailed feedback. We agree that some parts of the manuscript, especially Section 3.2, can be made more concise and focused. We will revise the text to reduce redundancies, remove unnecessary details, and ensure that explanations directly support the main messages. For example, we will avoid repeating examples of indirect and non-monetary losses and simplify sentences where possible. Regarding Table 6 (and Tables A5 and A6), we chose Q20 and Q80, we believe that this indicators can give more information, since for some hazards Q90 approaches infinity. However, we will consider to adapt the tables adding the number of ensemble members used to calculate median and percentile values. Finally, we appreciate your suggestion to underline key findings—such as the strong increase in heatwaves and wildfires in the Mediterranean region—with references to the relevant figures. We will implement these improvements to enhance clarity and strengthen the conclusions.
Discussion
Comment:
While the study solely focuses on infrastructure exposure, it would benefit from discussing briefly the aspect of vulnerability at least in this section. For instance, how the TEN-T network is specifically designed to withstand more frequent and intense climate extremes.
Answer:
We acknowledge this point, we will modify the Introduction to give more context and relevant references on risks and vulnerabilities. We will add also more information in this section to vulnerability of TEN-T network.
Comment:
Subsection 4.1: The authors should try to adapt parts of the second last paragraph to fit better to the topic of this subsection, namely past events. In addition, the text about the future impacts should be moved to the subsequent section.
Answer:
This is a good point, we agree with that and we will modify the manuscript accordingly.
Comment:
Subsection 4.2: The paragraph about drought impacts is not linked to the analysis of future exposure, furthermore, it contains explanations about drought impacts which should not be part of discussion.
Answer:
The discussion of the drought index is included in this paragraph because it describes the index used in the analysis of future exposure. This part is important, as it explains the link between the specific drought index, based on soil moisture drought, and its impact on transport infrastructure. However, we understand that this may be too extensive for the discussion section. Therefore, we will move it, for example, to the Methods section, where we describe the variables and indices used in the analysis.
Comment:
Subsection 4.2.1 need to be better linked to examples of compound events from the literature review (e.g. co-occurrence of wildfires and heatwaves in Sicily).
Answer:
That is a good point, thank you. We will add more information to better link Subsection 4.2.1 with examples of compound events from the literature review, including details on the selected case studies that are compound.
Conclusions
Comment:
I would recommend revising this section completely. The additional work can improve how readers perceive the study since most of them will read after the abstract the conclusions to get an impression of the study. Thus, it should be concise and well-devised. For instance, in its current form the section lacks a summary and assessment of the study’s key findings.
Answer:
We will revise the section to be more concise, with more summary and assessment of study findings.
Minor comments
Comment:
Table 1: The caption should be moved to the top of the table, as well as the text with the references to global climate and impact models.
Answer:
We will review it, thanks for noticing it.
Comment:
Consistent writing of “en route”
Answer:
We will review it.
Comment:
Figures in the appendix (A1 – A4): their information should try to be condensed and presented more clearly. Readers might find it difficult to detect the differences between these figures. For instance, it might improve the quality of the figures by representing adjacent events of the same hazard types by one symbol with a size according to the number of events it represents
Answer:
The reported information, the extreme events from GDIS dataset, is the same in all figures; the differences lie only in the positions of the respective transport modes (airports, ports, inland waterways). Presenting all transport modes in a single figure was confusing, so we split them into multiple figures, one for each mode. This approach also helps stakeholders focusing on a specific transport mode to easily identify the relevant information. We acknowledge that this was not clearly explained and will clarify it better in the text.
Comment:
Table A3: the same sentence appears twice: “The initial package of as much as 2.25 billion could be complemented by other sources like the EU’s solidarity fund, von der Leyen said.” and the table caption should refer to Greece
Answer:
Thank you so much for noticing it, we will fix it.
Comment:
L 9: missing whitespace
Answer:
We will review it.
Comment:
L 69: abbreviated RCP without entire name at first mention, same as for GDIS dataset L 77
Answer:
We will review it.
Comment:
L 21: is number from EEA 2023
Answer:
Yes, we will add the citation.
Comment:
L 41f missing word after “,”
Answer:
We will review it.
Comment:
Make figures colorblind friendly
Answer:
We will change the palette for some plots, following also the feedback from the others reviewers.
Comment:
Figure 8: similar red tones should be avoided, and the black error bars are barely visible
Answer:
We will take into account of this.
Comment:
Figure 5: Merge both plots as they refer to the same information
Answer:
Thank you for the comment. Although the plots contain the same information, we believe it could be useful for stakeholders to have both. We will, however, consider merging them.
Comment:
Years for in-text citations are occasionally missing, please have a look at the NHESS guidelines for citations, e.g. L 111
Answer:
We will review it.
Comment:
Lange et al. 2020b is the same as Lange et al. 2020a
Answer:
We will review it.
Citation: https://doi.org/10.5194/egusphere-2025-1697-AC3
-
RC2: 'Comment on egusphere-2025-1697', Anonymous Referee #2, 23 Jun 2025
The authors may consider reformulating the title to make it more engaging and assertive. At present, the use of a question reduces the impact of the message in the paper. A more direct and declarative title could better reflect the findings and relevance of the study, while also capturing the attention of the reader more effectively.
While the manuscript provides a valuable compilation of bibliographic references related to individual extreme climate events—particularly in the case study sections—there appears to be a gap in terms of broader scientific framing. Specifically, the references are largely concentrated on the documentation of specific events (e.g., floods, heatwaves) and their impacts, but there is limited engagement with the wider scientific literature on climate risk assessment for transport infrastructure. This includes theoretical or conceptual discussions on vulnerability and resilience, as well as methodological references supporting the use of multi-hazard models and exposure metrics like the Exposure Multiplication Factor (EMF). Furthermore, the transition from observed events to model-based future scenarios would benefit from clearer referencing to studies that have addressed similar downscaling and ensemble techniques. Overall, the manuscript would be strengthened by a more explicit literature review that connects recent empirical findings with the methodological and conceptual framework guiding the projections.
Acronyms and technical terms (e.g., RCP6.0, EMF) are not always introduced or defined clearly upon first use. Ensuring consistent definition and explanation would make the manuscript more accessible to non-specialist readers.
The manuscript oscillates between formal and informal tones in isolated comments and segments. Standardizing the scientific tone and avoiding colloquial or overly informal phrasing.
There are some suggestions:
Line 25: Please define the acronym used in this emissions scenario to ensure clarity for all readers;
Line 116: Are there existing studies or references supporting these events? Consider adding them to strengthen the discussion;
Line 117: Please rephrase the last paragraph into a bullet list to improve readability;
Line 118: Consider clarifying the resolution used for impact models by specifying it in latitude × longitude format;
Line 135: The introduction to Table 1 should appear before the table to improve logical flow and readability;
Line 137: Please elaborate or rephrase for precision;
Line 138: Consider explaining the overlapping time periods “2049 to 2075” as they may confuse the readers;
Figure 3: Please avoid using grey tones in the graphical elements, as they visually overlap with the background text and reduce readability;
Figure 4: Please consider increasing the primary units on the y-axis to improve readability. The current scale makes it difficult to clearly interpret the number of reported events per country;
Figure 5: Please consider using a more contrasting color palette to improve readability.Citation: https://doi.org/10.5194/egusphere-2025-1697-RC2 -
AC1: 'Reply on RC2', Cristina Deidda, 22 Aug 2025
Dear Reviewer,
We would like to thank you for your constructive feedback and careful review. Your comments were very helpful in guiding us to improve the manuscript, particularly by refining the title to make it more impactful, adding more context on risk assessment, and enhancing the figures.
Please find below our point-by-point responses (in bold) to each of your comments.
Comment:
The authors may consider reformulating the title to make it more engaging and assertive. At present, the use of a question reduces the impact of the message in the paper. A more direct and declarative title could better reflect the findings and relevance of the study, while also capturing the attention of the reader more effectively.
Answer:
We appreciate your suggestion regarding the title. Although we initially preferred the use of a question, we agree that a more assertive formulation would more effectively convey the main findings. We will therefore revise the title, with possible alternatives such as “Climate Change Is Intensifying Risks to Europe’s Transport Infrastructure” or “Climate Change Is Threatening Europe’s Transport Infrastructure.”
Comment:
While the manuscript provides a valuable compilation of bibliographic references related to individual extreme climate events—particularly in the case study sections—there appears to be a gap in terms of broader scientific framing. Specifically, the references are largely concentrated on the documentation of specific events (e.g., floods, heatwaves) and their impacts, but there is limited engagement with the wider scientific literature on climate risk assessment for transport infrastructure. This includes theoretical or conceptual discussions on vulnerability and resilience, as well as methodological references supporting the use of multi-hazard models and exposure metrics like the Exposure Multiplication Factor (EMF). Furthermore, the transition from observed events to model-based future scenarios would benefit from clearer referencing to studies that have addressed similar downscaling and ensemble techniques. Overall, the manuscript would be strengthened by a more explicit literature review that connects recent empirical findings with the methodological and conceptual framework guiding the projections.
Answer:
Thank you for this insightful comment. We agree that the manuscript would benefit from a stronger connection to the broader scientific literature on climate risk assessment for transport infrastructure. While we do discuss the importance of multi-hazard analysis and introduce the Exposure Multiplication Factor (EMF) metric, with proper references, in the Methods section (lines 155–165), we acknowledge that the conceptual framing, as well as references to prior methodological work, can be better emphasized earlier in the manuscript.
We will revise the Introduction to include relevant theoretical and conceptual literature on vulnerability, resilience, and climate risk frameworks in transport systems. Additionally, we will strengthen the discussion around multi-hazard modelling approaches and exposure metrics by citing recent studies that have employed similar methodologies. In the Results and Discussion sections, we will also clarify the transition from observed events to future model-based projections , with more explicit references to studies that underpin the methodological and conceptual framework.
Comment:Acronyms and technical terms (e.g., RCP6.0, EMF) are not always introduced or defined clearly upon first use. Ensuring consistent definition and explanation would make the manuscript more accessible to non-specialist readers.
The manuscript oscillates between formal and informal tones in isolated comments and segments. Standardizing the scientific tone and avoiding colloquial or overly informal phrasing.
There are some suggestions:
Line 25: Please define the acronym used in this emissions scenario to ensure clarity for all readers;Answer:
Thank you for your valuable comments. We will revise the manuscript to properly introduce and define all acronyms and technical terms upon first use. We will also review the text to ensure a consistent scientific tone throughout, addressing the specific suggestions you provided.
Line 116: Are there existing studies or references supporting these events? Consider adding them to strengthen the discussion;
Yes, we have listed all references to similar studies using the same event definitions in the Methods section. We will also add these references here to strengthen the discussion. Thank you for the suggestion.
Line 117: Please rephrase the last paragraph into a bullet list to improve readability;Thanks for the comment, we will review it and changing the last paragraph in bullet list.
Line 118: Consider clarifying the resolution used for impact models by specifying it in latitude × longitude format;We will specify it as: “ 0.5° × 0.5° (approximately 55 km × 55 km at the equator).”
Line 135: The introduction to Table 1 should appear before the table to improve logical flow and readability;Thank you for pointing this out, we will update it.
Line 137: Please elaborate or rephrase for precision;We will provide additional details on the raw data types and the sensitivity analysis to improve transparency for the reader. In the initial dataset, heatwaves and droughts are recorded as affecting entire grid cells, whereas wildfires, tropical cyclones, and floods are expressed as the fraction of the grid cell exposed to the extreme event. For these latter hazards, a threshold is required to convert the fractional exposure into binary data and ensure consistency across the dataset. We classify a grid cell as affected by wildfires, tropical cyclones, or floods if more than 0.3% of its area is exposed. This relatively low threshold was selected because these hazards are often highly localized; using a higher area percentage would have excluded a substantial number of events.
Line 138: Consider explaining the overlapping time periods “2049 to 2075” as they may confuse the readers;
We will add further explanation of the reasoning behind the overlapping time periods. The choice of a 52-year window was made to ensure statistical robustness when assessing rare extreme events, such as those with return periods of 100 years. Shorter windows (e.g., 30 years) would not provide a sufficient sample size for capturing such events. The overlap between the mid-century and end-of-century periods results from balancing the need for long analysis windows with the limited time horizon of the available climate projection data (ending at the end of the century). While this does create overlap, it was necessary and unavoidable in this context, and we will clarify this in the revised manuscript.
Figure 3: Please avoid using grey tones in the graphical elements, as they visually overlap with the background text and reduce readability;Thank you for the comment. We will update Figure 3 to replace grey tones with more distinct colors to improve readability.
Figure 4: Please consider increasing the primary units on the y-axis to improve readability. The current scale makes it difficult to clearly interpret the number of reported events per country;Thank you, we will adjust the primary units to make it more readable.
Figure 5: Please consider using a more contrasting color palette to improve readability.Thank you for the suggestion. We will test more contrasting colour palettes to improve the readability of Figure 5.
Citation: https://doi.org/10.5194/egusphere-2025-1697-AC1
-
AC1: 'Reply on RC2', Cristina Deidda, 22 Aug 2025
-
RC3: 'Comment on egusphere-2025-1697', Anonymous Referee #3, 25 Jun 2025
The paper addresses a very relevant and urgent issue – the impact of climate change on transport infrastructure in Europe – which makes it of great interest to both policy makers and the scientific community. The topic is of great relevance, given the increasing climate-related disruptions to transport systems across Europe.
The manuscript would benefit from the following improvements:
- Since the paper addresses climate change impacts, it would be important to clarify in the Introduction or Methods section why exposure definitions from the IPCC frameworks are not adopted or cited in the paper.
- It is recommended to cite some papers that address the topic of climate risk analysis for the infrastructures such as airports (e.g. https://doi.org/10.1016/j.cliser.2024.100536; https://doi.org/10.1007/s11069-021-05066-0)
- It is recommended to streamline the Discussion section, which currently appears fragmented. A more coherent narrative structure would improve the clarity and interpretation of the results.
- It is suggested to integrate the "Challenges and future work" section in the Conclusions.Citation: https://doi.org/10.5194/egusphere-2025-1697-RC3 -
AC2: 'Reply on RC3', Cristina Deidda, 22 Aug 2025
Dear Reviewer,
I would like to thank you for your valuable feedback and constructive review. Your comments have been very helpful in improving the context of the paper, particularly by encouraging us to provide more information on risk assessment for transport infrastructure and to cite key references such as the IPCC risk definition. We will also refine the manuscript, especially in the Discussion and Conclusion sections, by improving the flow of the text and ensuring a more coherent and consistent narrative structure.Please find below our point-by-point responses (in bold) to each of your comments.
Comment:
The paper addresses a very relevant and urgent issue – the impact of climate change on transport infrastructure in Europe – which makes it of great interest to both policy makers and the scientific community. The topic is of great relevance, given the increasing climate-related disruptions to transport systems across Europe.
Answer:
Thank you for your positive comment on the relevance and urgency of the topic.
Comment:The manuscript would benefit from the following improvements:
- Since the paper addresses climate change impacts, it would be important to clarify in the Introduction or Methods section why exposure definitions from the IPCC frameworks are not adopted or cited in the paper.Answer:
Thank you for the comment. We agree that the manuscript would benefit from a stronger connection to the broader scientific literature on climate risk assessment for transport infrastructure. We will revise the Introduction to incorporate relevant theoretical and conceptual literature on vulnerability, resilience, and climate risk frameworks in transport systems. In this context, we will refer to the IPCC definition of risk and clarify how the current analysis relates to it, focusing on the hazard and exposure components.
Comment:
- It is recommended to cite some papers that address the topic of climate risk analysis for the infrastructures such as airports (e.g. https://doi.org/10.1016/j.cliser.2024.100536; https://doi.org/10.1007/s11069-021-05066-0)Answer:
Thank you for the interesting references. We will add them for the section on airports and complement them with additional references for the other modes of transport.
Comment:
- It is recommended to streamline the Discussion section, which currently appears fragmented. A more coherent narrative structure would improve the clarity and interpretation of the results.Answer:
Following this suggestion, and those of the other reviewers, we will revise the structure of the text to improve the readability and flow of the narrative in the Discussion and Conclusion sections.
Comment:
- It is suggested to integrate the "Challenges and future work" section in the Conclusions.Answer:
Thank you for the suggestion. We will revise the structure of all sections, particularly the Conclusions, to improve readability and the flow of the text.
Citation: https://doi.org/10.5194/egusphere-2025-1697-AC2
-
AC2: 'Reply on RC3', Cristina Deidda, 22 Aug 2025
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 487 | 63 | 24 | 574 | 21 | 39 |
- HTML: 487
- PDF: 63
- XML: 24
- Total: 574
- BibTeX: 21
- EndNote: 39
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1