the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 29 Apr 2022
Performance of the flood warning system in Germany in July 2021 – insights from affected residents
Abstract. In July 2021 intense rainfall caused devastating floods in Western Europe and 184 fatalities in the German federal states of North Rhine-Westphalia (NW) and Rhineland-Palatinate (RP) questioning their flood forecasting, warning and response system (FFWRS). Data from an online survey (n = 1315) reveal that 35 % of the respondents from NW and 29 % from RP did not receive any warning. Of those who were warned 85 % did not expect a very severe flooding and 46 % did not know what to do. Regression analysis reveals that this knowledge is influenced by gender and flood experience, but also by the contents and the source of the warning message. The results are complemented by analyses of media reports and official warnings that show shortcomings in providing adequate recommendations to people at risk. Dissemination of warnings, communication of the expected flood magnitude and adequate responses are seen as entry points for improving the FFWRS in Germany.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(1035 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(1035 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2022-244', Martijn Kuller, 11 May 2022
This paper presents the results of a survey held among the residents of a recent severe flood event in Germany. It is a valuable piece of work both for the academic community interested in risk communication and for authorities to draw lessons from and improve practices in the future. While the overall quality of the work sufficient to warrant publication, there are a few issues that need to be addressed. Most importantly, the structure of the paper could benefit from clearer research questions, hypotheses and a methods and results section following this structure.
Overall the language is good and the paper reads well, but some phrases are rather “spoken” language than written language style (e.g., line 39: “save people’s lives”. I would rewrite to: “prevent fatalities” or “prevent loss of life”. Second example in line 79: “…how the targeted population…”. I would change to: “…warning perception by the target population…”). The article would benefit from review by a native English writer. Below I provide further details on this, as well as other comments by section, and then detailed comments in a table.
Introduction
This section has a clear structure and explains the background, the need for the research and the research aim clearly. However, detailed research questions and hypotheses are missing, which are needed to justify the methods used (types of data (indicators) gathered, methods of data collection etc.).
Data and Methods
It is stated that the survey was distributed via facebook, and a press release was sent through the newspaper. However, it seems like the survey could only be accessed through facebook. Why did the authors choose this method, and how did they make sure this didn’t lead to a biased sample, as facebook is only used by a specific demographic, leaving out other specific demographics that might be important. Beyond age, general tech-savvyness and online presence play a role in the accessibility of the survey to people using facebook. If the authors somehow tried to correct for this bias, please provide some information about how this was done. If not, please discuss how the method might have influenced the outcome and what this means for the conclusions.
Also, it is explained what the questions in the survey were regarding, however I am missing some explanation as to why these questions were posed (what were the exact research questions the authors want to answer with the set of questions they selected, including the demographical questions), and what were they expecting to find? Part of this information could already be provided in the introduction. Furthermore, is the survey accessible to the reader somehow? It would be good to add them to supplementary materials for reference and future use by the readers.
The Likert scales used in the survey are analysed statistically, using the answers as numerical values. Although this is common practice, it is strictly speaking not valid, as these numbers are only category indicators, and not actual numerical values of the answer, and can thus not be used for statistical analysis (drawing mean, sd etc).
Results and discussion
If the factors investigated and presented in Table 2 explain so little in the receipt of an official warning, what then is (or could be) the most important predictor for this?
All the results are self-reported data from the survey. I wonder how this self-reported data compares to behaviour and if this could be discussed in this section (e.g., how do the numbers of “knowing what to do” compare to the observed behaviour of people on the day?). Would there be any such comparison possible here in the discussion?
The results presented in Figure 5 show very little variation. This is a common problem with asking respondents what they want, choosing from a list of options: they are going to want it all. Such results are not very helpful when resources are scarce and trade-offs have to be made (in this case, we want to keep a warning clear and concise). A better way to measure such preferences could therefore be asking respondents to rank or to distribute a limited number of points among the options.
A section on further research is missing. I already made some recommendations in this review, e.g., comparing self-reported information to actual behaviour (if this is not going to be implemented by authors).
Conclusions
Because the research questions weren’t completely clear (see other comments), the conclusion doesn’t naturally follow from the rest of the paper. When this issue is resolved, the conclusion might be written in a more concise manner, going back to these research questions.
There are many references to the main text as well as other papers in the conclusion. I recommend rewriting the conclusion according to my above instructions, and avoid any references, as a conclusion should stand on its own, not repeat the results, but rather state the significance and meaning of results. As it stands, the conclusion mostly summarises the main text.
Furthermore, I think the paper could benefit greatly from a separate section (within conclusion or at the end of the discussion) that summarises the most important recommendations from this research. This research is very practical and applied, and has great potential to aid authorities around the world.
Please find detailed comments in the table below:
Line
Comment
39-40
One or two sentences discussing why the performance might have been so good for these events could benefit this section here.
153-158
This belongs more in the methods section, see previous comments.
159
Results are presented for which no methods have been described (logistic regression is mentioned). This should be added to the methods section (also see previous comments about methods).
160
Please explain how to interpret odds ratios.
Table 2 and 3
From the table it isn’t clear how gender impacts the prediction on knowing what to do. Is it male or female that increases the knowing what to do (explain in the table).
184
Is there data available to make a more precise comparison between the warnings issued by the different authorities and the warnings received by the population as reported through the survey? This could shed light on the effectiveness of the dissemination by the authorities. (a hint to this information is given in the text in lines 183-184).
189
The authors state that they find the trust in the credibility high. I would argue that 9% of people not believing a warning is quite a high number in an emergency situation, and less than half of the people stating they find the warning highly credible is to my perception quite low. From a government-issued warning, I would strive for credibility numbers close to 100%.
198
This sentence is confusing, please re-write.
210
This is more a concluding remark.
Fig. 5 and 6
Meaning of the numerical scale points need to be explained in the caption. Furthermore, as indicated before, it is strictly speaking not correct analysing and presenting this data in this way.
298
It is not common to refer to tables and figures in the conclusion if not strictly necessary.
Citation: https://doi.org/10.5194/egusphere-2022-244-RC1 - AC1: 'Reply on RC1', Annegret Thieken, 10 Sep 2022
-
RC2: 'Comment on egusphere-2022-244', Anonymous Referee #2, 15 Jul 2022
This paper provides very important insights into the early warning, preparedness, and response of the German flood event in 2020. The major aim of the paper is to analyze the operation of the warning system in July 2021 and to compare survey results to survey results of historic floodings in Germany. The results presented in this paper will be of high value to improve the disaster management in Germany.
I agree with the major recommendations provided in the comment RC1; thus, below, I will provide only a few suggestions in addition to the already stated recommendations.
Introduction/Methods
The introduction provides a clear and in-depth overview on the event, its impacts, and on the early warning structure in Germany. Considering that one major aim of the paper is to compare the event with previous flooding events in Germany, it could be of advantage to briefly introduce these flooding events. This would improve the reader’s understanding on the context of the previous events that may support the understanding of similarities and differences identified between the historic and the recent event. It could be included in form of a table or short paragraph in the introduction or methods chapter.
Results/Discussion
The section presents the extensive results of the survey by partly comparing them to previous flooding. It could be of advantage to have on section dedicated (e.g., between 3.2 and 3.3) on the major similarities/differences identified between the events and discuss these. Considering also the following question: are there aspects that actually improved during the past decades?
Conclusion
The conclusion could be more specific in terms of recommendations for the future (e.g., ‘communication […] have to be considerably enhanced’ –> you could specify/list how it should be enhanced).
Also, it could be interesting to see who these recommendations would be addressed to - who would be in charge of addressing the criticalities identified in this research.
I also agree that there shouldn’t be any references (incl. to figures) in the conclusion.
Lastly, the last sentence (L324-325) is of high importance, but this topic was not discussed in the paper.
Citation: https://doi.org/10.5194/egusphere-2022-244-RC2 - AC2: 'Reply on RC2', Annegret Thieken, 10 Sep 2022
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2022-244', Martijn Kuller, 11 May 2022
This paper presents the results of a survey held among the residents of a recent severe flood event in Germany. It is a valuable piece of work both for the academic community interested in risk communication and for authorities to draw lessons from and improve practices in the future. While the overall quality of the work sufficient to warrant publication, there are a few issues that need to be addressed. Most importantly, the structure of the paper could benefit from clearer research questions, hypotheses and a methods and results section following this structure.
Overall the language is good and the paper reads well, but some phrases are rather “spoken” language than written language style (e.g., line 39: “save people’s lives”. I would rewrite to: “prevent fatalities” or “prevent loss of life”. Second example in line 79: “…how the targeted population…”. I would change to: “…warning perception by the target population…”). The article would benefit from review by a native English writer. Below I provide further details on this, as well as other comments by section, and then detailed comments in a table.
Introduction
This section has a clear structure and explains the background, the need for the research and the research aim clearly. However, detailed research questions and hypotheses are missing, which are needed to justify the methods used (types of data (indicators) gathered, methods of data collection etc.).
Data and Methods
It is stated that the survey was distributed via facebook, and a press release was sent through the newspaper. However, it seems like the survey could only be accessed through facebook. Why did the authors choose this method, and how did they make sure this didn’t lead to a biased sample, as facebook is only used by a specific demographic, leaving out other specific demographics that might be important. Beyond age, general tech-savvyness and online presence play a role in the accessibility of the survey to people using facebook. If the authors somehow tried to correct for this bias, please provide some information about how this was done. If not, please discuss how the method might have influenced the outcome and what this means for the conclusions.
Also, it is explained what the questions in the survey were regarding, however I am missing some explanation as to why these questions were posed (what were the exact research questions the authors want to answer with the set of questions they selected, including the demographical questions), and what were they expecting to find? Part of this information could already be provided in the introduction. Furthermore, is the survey accessible to the reader somehow? It would be good to add them to supplementary materials for reference and future use by the readers.
The Likert scales used in the survey are analysed statistically, using the answers as numerical values. Although this is common practice, it is strictly speaking not valid, as these numbers are only category indicators, and not actual numerical values of the answer, and can thus not be used for statistical analysis (drawing mean, sd etc).
Results and discussion
If the factors investigated and presented in Table 2 explain so little in the receipt of an official warning, what then is (or could be) the most important predictor for this?
All the results are self-reported data from the survey. I wonder how this self-reported data compares to behaviour and if this could be discussed in this section (e.g., how do the numbers of “knowing what to do” compare to the observed behaviour of people on the day?). Would there be any such comparison possible here in the discussion?
The results presented in Figure 5 show very little variation. This is a common problem with asking respondents what they want, choosing from a list of options: they are going to want it all. Such results are not very helpful when resources are scarce and trade-offs have to be made (in this case, we want to keep a warning clear and concise). A better way to measure such preferences could therefore be asking respondents to rank or to distribute a limited number of points among the options.
A section on further research is missing. I already made some recommendations in this review, e.g., comparing self-reported information to actual behaviour (if this is not going to be implemented by authors).
Conclusions
Because the research questions weren’t completely clear (see other comments), the conclusion doesn’t naturally follow from the rest of the paper. When this issue is resolved, the conclusion might be written in a more concise manner, going back to these research questions.
There are many references to the main text as well as other papers in the conclusion. I recommend rewriting the conclusion according to my above instructions, and avoid any references, as a conclusion should stand on its own, not repeat the results, but rather state the significance and meaning of results. As it stands, the conclusion mostly summarises the main text.
Furthermore, I think the paper could benefit greatly from a separate section (within conclusion or at the end of the discussion) that summarises the most important recommendations from this research. This research is very practical and applied, and has great potential to aid authorities around the world.
Please find detailed comments in the table below:
Line
Comment
39-40
One or two sentences discussing why the performance might have been so good for these events could benefit this section here.
153-158
This belongs more in the methods section, see previous comments.
159
Results are presented for which no methods have been described (logistic regression is mentioned). This should be added to the methods section (also see previous comments about methods).
160
Please explain how to interpret odds ratios.
Table 2 and 3
From the table it isn’t clear how gender impacts the prediction on knowing what to do. Is it male or female that increases the knowing what to do (explain in the table).
184
Is there data available to make a more precise comparison between the warnings issued by the different authorities and the warnings received by the population as reported through the survey? This could shed light on the effectiveness of the dissemination by the authorities. (a hint to this information is given in the text in lines 183-184).
189
The authors state that they find the trust in the credibility high. I would argue that 9% of people not believing a warning is quite a high number in an emergency situation, and less than half of the people stating they find the warning highly credible is to my perception quite low. From a government-issued warning, I would strive for credibility numbers close to 100%.
198
This sentence is confusing, please re-write.
210
This is more a concluding remark.
Fig. 5 and 6
Meaning of the numerical scale points need to be explained in the caption. Furthermore, as indicated before, it is strictly speaking not correct analysing and presenting this data in this way.
298
It is not common to refer to tables and figures in the conclusion if not strictly necessary.
Citation: https://doi.org/10.5194/egusphere-2022-244-RC1 - AC1: 'Reply on RC1', Annegret Thieken, 10 Sep 2022
-
RC2: 'Comment on egusphere-2022-244', Anonymous Referee #2, 15 Jul 2022
This paper provides very important insights into the early warning, preparedness, and response of the German flood event in 2020. The major aim of the paper is to analyze the operation of the warning system in July 2021 and to compare survey results to survey results of historic floodings in Germany. The results presented in this paper will be of high value to improve the disaster management in Germany.
I agree with the major recommendations provided in the comment RC1; thus, below, I will provide only a few suggestions in addition to the already stated recommendations.
Introduction/Methods
The introduction provides a clear and in-depth overview on the event, its impacts, and on the early warning structure in Germany. Considering that one major aim of the paper is to compare the event with previous flooding events in Germany, it could be of advantage to briefly introduce these flooding events. This would improve the reader’s understanding on the context of the previous events that may support the understanding of similarities and differences identified between the historic and the recent event. It could be included in form of a table or short paragraph in the introduction or methods chapter.
Results/Discussion
The section presents the extensive results of the survey by partly comparing them to previous flooding. It could be of advantage to have on section dedicated (e.g., between 3.2 and 3.3) on the major similarities/differences identified between the events and discuss these. Considering also the following question: are there aspects that actually improved during the past decades?
Conclusion
The conclusion could be more specific in terms of recommendations for the future (e.g., ‘communication […] have to be considerably enhanced’ –> you could specify/list how it should be enhanced).
Also, it could be interesting to see who these recommendations would be addressed to - who would be in charge of addressing the criticalities identified in this research.
I also agree that there shouldn’t be any references (incl. to figures) in the conclusion.
Lastly, the last sentence (L324-325) is of high importance, but this topic was not discussed in the paper.
Citation: https://doi.org/10.5194/egusphere-2022-244-RC2 - AC2: 'Reply on RC2', Annegret Thieken, 10 Sep 2022
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 1,030 | 521 | 22 | 1,573 | 24 | 19 |
- HTML: 1,030
- PDF: 521
- XML: 22
- Total: 1,573
- BibTeX: 24
- EndNote: 19
Viewed (geographical distribution)
| Country | # | Views | % |
|---|---|---|---|
| Germany | 1 | 683 | 46 |
| United States of America | 2 | 167 | 11 |
| France | 3 | 69 | 4 |
| Netherlands | 4 | 53 | 3 |
| United Kingdom | 5 | 51 | 3 |
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
- 683
Cited
9 citations as recorded by crossref.
- Resilience beyond insurance: coordination in crisis governance E. Platzer & M. Knodt 10.1007/s10669-023-09938-7
- More than heavy rain turning into fast-flowing water – a landscape perspective on the 2021 Eifel floods M. Dietze et al. 10.5194/nhess-22-1845-2022
- A multi-disciplinary analysis of the exceptional flood event of July 2021 in central Europe – Part 1: Event description and analysis S. Mohr et al. 10.5194/nhess-23-525-2023
- Kommunikationsdefizite in der Frühwarnung und Risikokommunikation G. Rusch 10.1007/s35152-024-1848-9
- Preparing for the unprecedented B. Golding et al. 10.5194/asr-20-85-2023
- The risk may not be limited to flooding: polluted flood sediments pose a human health threat to the unaware public A. Weber et al. 10.1186/s12302-023-00765-w
- Post-Flood Resilience Assessment of July 2021 Flood in Western Germany and Henan, China B. Manandhar et al. 10.3390/land12030625
- Critical Failure Factors of Flood Early Warning and Response Systems (FEWRS): A Structured Literature Review and Interpretive Structural Modelling (ISM) Analysis S. Samansiri et al. 10.3390/geosciences13050137
- Brief communication: On the extremeness of the July 2021 precipitation event in western Germany K. Lengfeld et al. 10.5194/nhess-23-1227-2023
Annegret H. Thieken
Philip Bubeck
Anna Heidenreich
Jennifer von Keyserlingk
Lisa Dillenardt
Antje Otto
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(1035 KB) - Metadata XML