the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 06 May 2025
Assessing effects of nature-based and other municipal adaptation measures on insured heavy rain damages
Abstract. Intense short duration rainfall events are expected to increase in severity and frequency due to climate change. Densely populated urban areas are vulnerable to these events, resulting in high losses. Implementing nature-based (e.g. green streets, rain gardens and green roofs) and other municipal adaptation measures (e.g. water storage facilities) can be a way to mitigate these damages. Little is known about the effectiveness of these measures combined in a municipality. This study assesses municipal climate adaptation measures being taken by the municipality of Amsterdam. Unique claims data of almost all Dutch insurers is used to understand the impact of these climate adaptation interventions. We study one neighborhood in Amsterdam which has been renovated using climate adaptation measures, including nature-based solutions. We implement a quasi-experimental difference-in-Differences (DiD) analysis that compares insured rainfall damages in the area to a similar neighboring area that was not renovated with climate adaptation measures. We find a negative significant relation between climate adaptation measures and insured damage when comparing the treated group to the control group, i.e. damage is reduced by climate adaptation measures by €3700 euro per rain day. Furthermore, the control variables significantly associated with insured damage are precipitation per day (positively), household size (positively), address density (negatively) and value of property (positively). We suggest that nature-based and other adaptation measures can be installed by local governments and stimulated by insurers and banks to increase climate resilience in urban areas.
- Preprint
(598 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-1882', Anonymous Referee #1, 27 Jun 2025
See PDF for comments.
For full disclosure, I would like to point out to the editor that I'm affiliated with the same research consortium (Red&Blue) as two of the authors. Hence, the decision to not also review the revised manuscript. While affiliated, I don't work with any of the authors myself, nor is there a close relationship. So, concerns about conflicts of interest should be limited.
-
AC1: 'Reply on RC1', Vylon Ooms, 15 Aug 2025
We sincerely thank the reviewer for the positive evaluation, the comments and the suggestions. We address the feedback in detail in the attached PDF-file.
Based on the suggestions of the reviewer, the main changes we plan to make to the paper are as follows:
- Analysis: The reviewer commented that the analysis would be cleaner if we would only include observations before and after the intervention in the analysis, thereby leaving out observations during the intervention period. The new analysis is performed without the intervention period. The analysis with the addition of the intervention period we plan to add as a robustness test in chapter 3. By omitting the rollout period we avoid the potential of estimating only a partial treatment effect.
- Variables: We plan to remove two control variables: ‘average number of people per household per address’, and ‘percentage of real estate built before 1945’ due to multi-collinearity problems. Because of multicollinearity we also changed ‘address density’ to a different variable called ‘population density’: the amount of people per area divided by the size of the area.
- Information campaigns: We plan to add a detailed description of the content of the information campaigns.
- Explanation results: We plan to put the results into perspective by adding numbers of damage mitigation per year instead of only numbers of damage mitigation per rain day.
We believe that the feedback will substantially improve the manuscript.
-
AC1: 'Reply on RC1', Vylon Ooms, 15 Aug 2025
-
RC2: 'Comment on egusphere-2025-1882', Anonymous Referee #2, 29 Jun 2025
The manuscript evaluated the aggregated impact of a bundle of nature-based and other municipal adaptation measures on insured rain damages by comparing the insured data from two adjacent areas within the Rivierenbuurt neighbourhood, one with flood damage mitigation (FDM) measures and one without intervention. Using the statistical difference-in-difference method, the authors identified significant relations in some variables, highlighting the causal effect of FDM measures. The discussion and conclusions were valid and supported by data evidence.
The manuscript proposes a research initiative on a topic within the scope of Natural Hazards and Earth System Sciences (NHESS). I would recommend this manuscript for publication with the attached comments, particularly regarding data processing, model interpretation, and widening the discussion.
-
AC2: 'Reply on RC2', Vylon Ooms, 15 Aug 2025
We thank the reviewer for these kind words. We revised our paper following the suggestions provided by the reviewer. We address the comments in the attached PDF-file. The main changes we plan to make, based on the suggestions of the reviewer, are the following:
- Damage data: The addition of two tables to appendix 4, showing more detailed information about the damage data.
- Extra analysis: The addition of an analysis of only the significant variables to appendix 5. This would be added to the manuscript as a robustness test.
- Variables: We plan to remove two variables that cause high multi-collinearity. For the same reason, we plan to change the variable ‘address density’ to ‘population density’.
- Difference-in-difference method: We plan to add a clarification with more detailed information on the DiD-method and the placebo tests.
We believe that the comments of the reviewer have improved the manuscript considerably.
-
AC2: 'Reply on RC2', Vylon Ooms, 15 Aug 2025
-
RC3: 'Comment on egusphere-2025-1882', Anonymous Referee #3, 30 Jun 2025
Many quantitative ex-ante studies have indicated the value of NBS interventions in urban areas, because of various Ecosystem Services, including mitigatin of urban heat island impacts, reduced impacts on the water cycle and mitigating flood risk. Especially for mitigation of flood risk many studies have indicated detailed approaches for quantification of the expected impacts.
On the other hand very few studies have been perfomed ex-post to verify the validity of the assumptions of the investments made to install NBS interventions. This study is therefore highly welcomed because the scope is exactly what the flood risk modelling community has repeatedly requested: to perform ex-post analyses to enable structured learning and inform future projects of the learnings of past projects.
However, this is also the most important limitation of the study. I would therefore encourage the authors to expand and improve their study prior to finalizing the paper. Most importantly to include information from the ex-ante study leading to the NBS interventions: What were the expected reduction in Expected Annual Damage (EAD), how was this linked to properties of rainfall, and did the reductions in damages occur at the expected locations? Validating and/or falsifying assumptions of the ex-ante study would be a truly interesting study and probably the information is readily at hand from the documents leading to the strategy of NBS implementation. The transferability would also be improved if changed to an indicator related to EAD.
Another limitation of the study is the use of insurance data for buildings. There are always losses that are not covered by insurance and hence this should be mentioned in the discussion of the findings.
Figure 1 is difficult to interpret. I assume it is the median of the daily damages for days with more than X mm rainfall for each year? Make it more explicit and use more space to explore this data, e.g. by plotting the distributions for a suitable year and also aggregating to EAD for each catchment by year. Based on the figure alone I would assume that there would be no impacts of the NBS. This points to a more general issue with providing the readers with enough information about the data to validate the findings. More plots of both input and output data and model residuals would be an asset.
Minor points
The interventions occur in 2018 in the text but in the figure it looks like the intervention is in 2017? Is the x-asis correct?
You could consider removing 2010 from the study all together, it would make for a more robust analysis
Citation: https://doi.org/10.5194/egusphere-2025-1882-RC3 -
AC3: 'Reply on RC3', Vylon Ooms, 15 Aug 2025
We thank the reviewer for the positive evaluation of the paper and the detailed review. It is pleasing to read that the study is welcomed by the community. We address the comments in the attached PDF-file. The main adjustments we plan to make are the following:
- Data description: An addition of more detailed descriptive information about damage data, with various tables including information about percentiles, range, mean, standard deviation and the median. Furthermore, we plan to add a descriptive table on rain data.
- Extra analysis: The addition of an analysis without the damage outlier that occurred in August 2010. We plan to add it as a robustness test of the results.
- Damage information: Additional information on what the data entails and what it does not cover (uninsured losses).
- Limitation description: An addition of a limitation that the data of the study does not contain uninsured damages (e.g. public infrastructure).
We believe that the feedback of the reviewer has improved the manuscript substantially.
-
AC3: 'Reply on RC3', Vylon Ooms, 15 Aug 2025
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 324 | 93 | 19 | 436 | 17 | 31 |
- HTML: 324
- PDF: 93
- XML: 19
- Total: 436
- BibTeX: 17
- EndNote: 31
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1