Flood exposure in Rotterdam&rsquo;s unembanked areas from 1970 to 2150: sensitivities to urban development, sea level rise and adaptation

Oerlemans, Cees; van den Boomen, Martine; Rijcken, Ties; Kok, Matthijs

doi:https://doi.org/10.5194/egusphere-2024-2910

Preprints

https://doi.org/10.5194/egusphere-2024-2910

Preprints

15 Jan 2025

| 15 Jan 2025

Flood exposure in Rotterdam’s unembanked areas from 1970 to 2150: sensitivities to urban development, sea level rise and adaptation

Cees Oerlemans, Martine van den Boomen, Ties Rijcken, and Matthijs Kok

Abstract. Uncertainties in the rate of sea level rise, coupled with ongoing urban expansion, create challenges for city planners in designing flood risk adaptation strategies. This study analyzes flood exposure rates in Rotterdam's unembanked areas from 1970 to 2150. We modeled flood hazards for 10–1000 year return periods under both low (RCP2.6) and high (RCP8.5) emission scenarios, while assessing exposure using historic and planned urban development data. Without adaptation measures, flood exposure in Rotterdam’s unembanked areas is projected to increase. For 10-year flood events under RCP8.5, exposure rates are expected to increase 7-fold by 2150 compared to 2020. For RCP2.6, a 3-fold increase is projected, reflecting uncertainties in long-term sea level rise. A retrospective analysis reveals an improvement in flood exposure: exposure levels observed in 2020 were approximately half those observed in 1996, due to construction of the Maeslant storm surge barrier. Temporal variations in exposure rates are attributed to three factors: urban development, sea level rise, and the construction of the Maeslant barrier. Exposure rates are primarily influenced by the Maeslant barrier, followed by sea level rise and urban development. Understanding the interplay of these three factors is crucial for urban planning and flood risk management in delta cities.

Received: 18 Sep 2024 – Discussion started: 15 Jan 2025

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.

Download & links

Cees Oerlemans, Martine van den Boomen, Ties Rijcken, and Matthijs Kok

Status: final response (author comments only)

RC1:
'Comment on egusphere-2024-2910', Anonymous Referee #1, 11 Mar 2025
The study conducts an exposure assessment for unembanked areas in Rotterdam for the years 1970 to 2150 accounting for different extreme water levels. Building exposure is estimated for different scenario combinations and neighborhoods to identify the contribution of sea level rise, urban development and adaptation (i.e. construction of the Maeslant barrier and elevating new buildings) to the exposure. The study presents a comprehensive framework for local exposure assessments, integrating future development of building infrastructure and adaptation. The manuscript is written clearly and the text is supported by insightful figures. Though the idea of the exposure assessment does not present an entirely new approach and the findings are sometimes relatively obvious, the study combines many different contributing factors to future exposure and the framework is interesting for further local studies. I have some general and also more specific comments that I believe could increase the scientific and presentation quality of the manuscript. I hope the authors consider the comments useful.
General comments:
The discussion and conclusion could benefit from some restructuring. Right now, the discussion mostly presents the limitations of the study, whereas the conclusion discusses the implications of the results. This should also be part of the discussion and the conclusion should then only highlight on the main findings.

The Methods describing the calculation of the flood hazards are not described sufficiently for readers to be able to reproduce them. How does the Hydra-NL model incorporate the results from the hydrodynamic model to calculate water level frequency lines at different locations? Which hydrodynamic model combinations have been used from the database? It would be helpful to add a short explanation to the manuscript.

In the results, exposure between neighborhoods is expressed and compared based on the relative share of exposed housing units in each neighborhood. You state for example that Kop van Zuid-Entrepot is the most vulnerable neighborhood with 48% of the housings being exposed. However, in absolute numbers these “only” refer to 100 exposed houses. In this context for example Feyenoord would be more exposed (200 houses). From a city planning perspective, I think it might be more relevant to know about the absolute exposure in order to identify hotspots and prioritize actions in neighborhoods where many houses and thus people are affected. The absolute numbers might be especially relevant for the 1000-year event, where 100 % of the houses are exposed for almost all neighborhoods. This could also be picked upon in the discussion to reflect, that evaluating exposure is also a matter of perspective and scale.

Throughout the manuscript the use of the terms exposure, risk and vulnerability is inconsistent and often erroneous. Since this is an exposure analysis, I would suggest to stick to this term and be careful when using the terms risk and especially vulnerability.

Specific comments:
Page 3, line 74-86: I find it unusual to refer to the findings already in the introduction. I would suggest to only describe the four-step process here without any interpretation of the results yet.
Page 4, line 91-101: The first paragraph of the Methods and data section reads more like part of the introduction. Perhaps, try to reduce this paragraph and/or include it in the introduction.
Page 4, line 109: Maybe mention that altering the design flood elevation is a scenario accounting for adaptation/accommodation.
Page 7, line 178: Why did the authors choose a scenario which involves elevating the design flood level by just 20 cm? Are there concrete actions/plans considering this as a new design height? To me it doesn’t sound like too much of a difference to the reference scenario and I am wondering why the authors didn’t choose a higher design level? Also, the authors could consider to elaborate other adaptation options in their discussion, like e.g. household-level adaptation, which might eliminate exposure in cases of lower water levels.
Page 8, line 189-191: To quantify exposure, the authors calculate the difference between water levels and the elevation of the housing unit. It appears that they are not accounting for hydrological connectivity here. This means housings behind structures higher than the water level are eventually not flooded. It would be good to mention this as a limitation. A similar issue is mentioned in the discussion about the missing seawall information.
Page 15, line 341-343: “Hence the comparative subplots show that stopping urban development after 2020 could ultimately prevent exposure to 1000-year events of all 22,600 housing units planned for construction.” I find this finding a bit too obvious. Perhaps there is no need to write this.
Page 20, line 421-427: Is the height information of the flood-proofing measures not reflected in the DTM? Referring to my earlier comment, if elevation information from the DTM is accurate for protections measures it might be valuable to assess exposed areas based on hydrological connectivity rather than comparing the water level to each housing elevation.
Page 21, line 441-450: The authors compare their estimates to the results of other exposure studies. Though from what they write it sounds like these studies assess exposure as monetary value. How is the comparison of these studies with the estimates of relative exposure of housing units from this study justified?
Page 23, line 525: Is there a reference for port relocation in Hamburg? In the introduction this is only mentioned for Houston, Copenhagen and Rotterdam.
Temporal line plots: I think it could be beneficial to explain the time line differences of the x-axis scales in each line plot caption in case readers are just looking into the figures.
Technical comments:
Page 1, line 15-18: “Temporal variations in exposure rates are attributed […] and urban development.” The two sentences mean the same and one of them could be removed here.
Page 5, Figure 1: Maybe write in the “exposure”-box that the data for urban development plans only reaches until 2040 to make sure it does not cover the full period of investigation.
Page 5, line 120: It should be Table 1 here, not Table 2.
Page 7, line 161: I suggest to mention all tested return periods here instead of writing “from 10 to 1000 years”.
Page 7, line 169: “These footprints were combined with elevation data from the Digital Terrain Model of the Netherlands (AHN3, 0.5-meter resolution raster) [to assign elevation values to each housing unit]. By assigning specific elevation values to each housing unit, we can accurately assess their potential exposure given various climate scenarios.” I would suggest the authors to remove the words in [...] to avoid repetition and add a full stop at the end of the second sentence.
Page 7, line 183: “The Design Flood Elevation is one the important components …” Look into the grammar here.
Page 7, line 183 – page 8, line 188: Paragraph is formatted like the subsequent heading.
Chapter 3: Chapter 3 appears to be a bit detached from the rest of the text. I would propose to put chapter “3 Case Study: Rotterdam’s unembanked areas” at the beginning of the Methods section as a description of the study site. Chapter 3.1. and 3.2 can be included in the Results, as they describe the findings of your flood hazard calculation and data processing.
Page 9, line 203: Should it be “have been constructed on building lots situated below the NAP+3.6 m elevation”, as this is the design flood elevation?
Page 11, line 242: “3.2 Urban development between 1970-2050” Should it rather be 2040 as this is the time span of the urban development plans?
Page 17, Figure 5: The reference scenario should be mentioned in the legend.
Citation: https://doi.org/10.5194/egusphere-2024-2910-RC1
- AC1:
  'Reply on RC1', Cees Oerlemans, 10 Apr 2025
  
  Dear Referee #1,
  Thank you for your constructive comments which have helped improve our manuscript. We highly appreciate your time, effort and support. The point by point reply attached includes our responses to your comments and how we adjusted the manuscript accordingly.
  
  Kind regards,
  
  Cees Oerlemans, on behalf of the co-authors
  
  Citation: https://doi.org/10.5194/egusphere-2024-2910-AC1
  - RC3: 'Reply on AC1', Anonymous Referee #1, 15 Apr 2025
    
    Dear authors,
    I appreciate that you considered my suggestions and thank you for your thourough responses and the justification of your choices.
    
    From your response, I believe that the manuscript has greatly benefited from the changes and that most points are clearer now.
    I have two remaining questions related to your response:
    1) You say that "For situations related to the increase in exposure, we now use the term susceptibility rather than vulnerability.". Why would you interchange these terms. To my understanding susceptbility is different to exposure. If you want to refer to an increase in exposure I think you should also phrase it as such.
    2) Regarding the comment Page 15, line 341-343: What I thought was quite obvious about this sentence is that of course the planned units will not be exposed in the future if urban development was stopped completly in 2020. And I guess that this applies irrelevant of the return period. I am not sure if my remark on this came across properly before.
    Kind regards!
    
    Citation: https://doi.org/10.5194/egusphere-2024-2910-RC3
    
    AC3: 'Reply on RC3', Cees Oerlemans, 23 Apr 2025
    
    Dear referee,
    
    Thank you for your follow-up questions:
    Indeed, if we refer to exposure or an increase in exposure we phrase it a such. We only use "susceptibility" when discussing neighborhoods that may experience increased exposure, always making this relationship with exposure explicit to avoid ambiguity.
    
    The exposure of planned housing depends on multiple factors: elevation of planned units, flood return period, sea level rise scenario (high/low) and reference year. While unbuilt houses cannot be exposed, the difference in exposure between scenarios with/without urban development varies with return period. For larger return periods, this difference equals the total number of planned units. For lower return periods, fewer housing units may be affected. We will clarify this distinction to prevent confusion.
    
    We hope to have addressed your follow-up questions. Please let us know if we misinterpreted any question or if any points require further explanation.
    Kind regards,
    
    Cees Oerlemans, on behalf of the co-authors.
    
    Citation: https://doi.org/10.5194/egusphere-2024-2910-AC3
RC2:
'Comment on egusphere-2024-2910', Anonymous Referee #2, 07 Apr 2025

This study performs an exposure assessment for the unembanked areas of Rotterdam from 1970 to 2150, considering different extreme water levels (historical sea level rise, NAP + 3.6m and future sea level rise resulting from RCP2.6 and RCP8.5). The exposure of buildings is assessed under different combinations of these flood scenarios and planned urban development, and under different adaptation options (construction of the Maeslant barrier and elevation of new buildings). Although large parts of the manuscript are clearly written, there are some concerns about the overall structure of the manuscript. In addition, the reference to other studies seems a bit tricky, as the studies the authors refer to have been carried out at a regional scale, which makes it difficult to compare the results. Nevertheless, the study is within the scope of the journal as it effectively brings together several factors influencing flood exposure. The framework presented could be valuable for similar local studies. I have a number of general comments, listed below, which may help the authors to improve the scientific rigour and overall structure of the manuscript.

Introduction
The authors deduce the research gap by referring to previous studies on exposure and changing flood risk resulting from climate change. These studies, however, have been carried out on a global to European scale, and limit the implications for local studies. Specifically, European studies based on assessments using the NUTS3 level cannot easily compared to the local, building plot level. As such, while the studies of Steinhausen et al. (2022), Praprotny et al. (2024) and others are valuable, a direct conclusion for the identified research gaps needs further elaboration (comparison is a matter of scale as the resolution of the data and therefore the preciseness is spatially variable, see for example Karagiorgos et al. (2024)). Possible other local studies in this field of research include those of Schlögl et al. (2021) focusing on similar topics, and results may be compared with the results of the present Rotterdam study.
Furthermore, the authors are kindly advised to carefully check their use of terms – as this study explicitly addresses dynamics of flood hazard and exposure taking into account adaptation measures reducing exposure, the reference to the vulnerability term should be omitted (also in other sections of the manuscript) – also increasing exposure will not necessarily result in increasing vulnerability (cited from page 2, line 27), see the underlying UNDRR definitions (https://www.undrr.org/drr-glossary/terminology).
The authors may also consider to make their focus on residential buildings much clearer already here, as building types different from residential ones are excluded here – specifically also with respect to the material provided on page 3, lines 60ff. (“…risk estimates would benefit from a more detailed consideration of objects and land-use categories”), but also with respect to the concluding sentences provided in lines 71f.
The authors may wish to further adapt the four points presented as workflow for the study, in some of the aspects results are already summarised which typically should not be presented in the intro section. Also, “various urban development strategies” (point 4) explored by only “varying the design flood elevation for new development” seems a bit overwhelming as in fact, only elevation of ground was considered (and not other local flood protection such as e.g. local structural measures) – hence, the list could benefit from a thorough re-wording and formulation.

Methods and data
The authors are kindly asked to check the wording here, and to delete the references to “vulnerability” (see my comments above, this should be made clear in the introduction).
Flood hazard: the data used to compute the flood hazard are not entirely clear to me, the authors are kindly asked to provide more details to allow the potential readers an evaluation and reproduction (see also the comment of reviewer 1). Authors may consider at least an Appendix here so that details can be given with sufficient details.
Exposure: please check the formatting of the last paragraph on page 7. Moreover, I am a bit confused about the exposure scenarios as I understood that an elevation of 3.6 and 3.8 m was considered as adaptation measure. How does such a high elevation influence the flood hazard? Does this simply imply that all the new constructions are flood safe then, and if yes, this is a bit obvious as this is a universally valid assumption? The authors could clarify their methods in this direction a bit, also with respect to other adaptation probably prescribed to homeowners in the Netherlands.

Case study section
Here the authors already present results from the historical analysis which should be moved to the results section (specifically sections 3.1 and 3.2). Moreover, the statement on page 9, lines 195f. are not clear to me as the study does not build on “societal factors” influencing flood risk on the neighbourhood level. Authors are kindly asked to re-write. Lines 211ff: This paragraph weakens the method used in the study, and such constraints or limitations should be moved to the discussion section.

Results section
I am bit confused about the fact that the authors rely on a relative comparison of exposure only, and do not provide information on the absolute values as they strongly differ between the different neighbourhoods.

Figures: resolution should be higher to enhance accessibility and readability.

Summary
The manuscript is valuable for the target journal, but will benefit from a thorough re-structuring and re-writing to allow the readers to better follow the arguments. Materials should be carefully re-ordered to the respective sections (introduction, methods, results and discussion, conclusion). Wording should be carefully checked (risk, exposure, vulnerability). Comparison to other local studies on the effects of climate change on hazard, and temporal exposure development is encouraged.

References mentioned
Karagiorgos, K. et al.: Global population datasets overestimate flood exposure in Sweden, Scientific Reports, 14, 20410, https://doi.org/10.1038/s41598-024-71330-5, 2024.
Schlögl, M. et al.: Trends in torrential flooding in the Austrian Alps: A combination of climate change, exposure dynamics, and mitigation measures, Climate Risk Management, 32, 100294, https://doi.org/10.1016/j.crm.2021.100294, 2021.

Citation: https://doi.org/10.5194/egusphere-2024-2910-RC2
- AC2: 'Reply on RC2', Cees Oerlemans, 23 Apr 2025
  
  Dear Referee #2.
  We highly appreciate your constructive comments and your dedicated time and effort to help us improving our manuscript. Please find attached our point by point reply and how we adjusted the manuscript accordingly.
  Kind regards,
  
  Cees Oerlemans (on behalf of the co-authors)
  
  Citation: https://doi.org/10.5194/egusphere-2024-2910-AC2

Cees Oerlemans, Martine van den Boomen, Ties Rijcken, and Matthijs Kok

Data sets

Scripts and data of: "Flood exposure in Rotterdam’s unembanked areas from 1970 to 2150: sensitivities to urban development, sea level rise and adaptation Cees Oerlemans https://doi.org/10.4121/d1291401-708a-4d48-9d95-8259cfd987d2

Interactive computing environment

Cees Oerlemans, Martine van den Boomen, Ties Rijcken, and Matthijs Kok

Viewed

Total article views: 394 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
263	118	13	394	9	10

HTML: 263
PDF: 118
XML: 13
Total: 394
BibTeX: 9
EndNote: 10

Views and downloads (calculated since 15 Jan 2025)

Month	HTML	PDF	XML	Total
Jan 2025	86	23	5	114
Feb 2025	36	12	0	48
Mar 2025	48	30	1	79
Apr 2025	65	38	7	110
May 2025	28	15	0	43

Cumulative views and downloads (calculated since 15 Jan 2025)

Month	HTML	PDF	XML	Total
Jan 2025	86	23	5	114
Feb 2025	36	12	0	48
Mar 2025	48	30	1	79
Apr 2025	65	38	7	110
May 2025	28	15	0	43

Viewed (geographical distribution)

Total article views: 378 (including HTML, PDF, and XML) Thereof 378 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 21 May 2025

Short summary

This study analyzes flood exposure in Rotterdam's unembanked areas from 1970 to 2150, exploring the interplay between rising sea levels, urban development, and flood protection measures. Without measures, flood exposure will increase, especially after 2100. The Maeslant storm surge barrier had the most impact on flood exposure, followed by urban development and sea level rise. Varied exposure levels across neighborhoods suggest the need for localized adaptation strategies.


Total:	0
HTML:	0
PDF:	0
XML:	0