the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 29 Jan 2024
Model based assessment of climate change impact on inland flood risk in coastal areas caused by compounding storm tide and precipitation events
Abstract. In addition to storm surges, inland flooding due to intense rainfall becomes an increasing threat at coastal lowlands. In particular, the coincidence of both types of events poses great challenges to regional water boards since their technical drainage capacities are limited. The evaluation of historical inland flood events at the German North Sea coast at the gauge Knock near Emden shows that in the past mainly moderate storm series in combination with large-scale, heavy precipitation led to an overload of inland drainage systems, whereas storm tides and precipitation alone could be handled well. Evaluation of the drivers of inland flood events simulated for the control period of two climate models confirms that a combination of storm tides and precipitation leads to highest drainage system overloads. Moderate system overload is also caused by heavy precipitation events alone rather than by storm tides without precipitation. Scenario projections based on a set of combinations of two highly resolved climate models and two emission scenarios suggest that the intensity of compound events of rainfall and storm tides will increase consistently against the background of mean sea level rise for all investigated climate change projections, while simulated system overload is higher for RCP8.5 compared to RCP2.6 scenario. Comparable to the past, future compound events will cause more potential damage compared to single extreme events. Such behaviour can be expected to induce an increasing frequency and intensity of inland drainage system overloads along the North Sea coast if timely adaptation measures will not be taken.
- Preprint
(1146 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-29', Wiebke Jäger, 26 Feb 2024
General comments
- This is an interesting paper showing that compound flooding at the Northwestern coast in Germany, can be driven by combination of moderate or intense drivers and not only by extreme individual drivers now as well as under future climate projections. The method is clear and appropriate. The results are relevant for the scientific community working on compound flooding as well as for adaptation and mitigation planning of local and regional water boards of the studied area. Below comments indicate elements of the manuscript that are not yet (entirely) clear to me.
Specific comments
- Title: Could be more concrete mentioning the study area instead of “coastal areas” as I think the results are mainly relevant for the particular case study rather than for coastal areas in general.
- The authors use the term “storm tide”, “storm surge”, “sea level”, “maximum sea level” and “storm flood” throughout the text. Are these used interchangeably or with different meanings? A common definition is water level = tide + surge (+ sea level rise), but it does look like the authors use them to refer to the same. An introduction and definition of the terms used would provide more clarity. I also suggest to use one term throughout for clarity.
- The authors focus on events of storm tide and precipitation events. Why do they not consider river discharge as well? It seems the authors see river discharge and precipitation as strongly linked in the area of the study (e.g., by mentioning the work of Henrich et. 2023 on storm tide and river discharge), but this is not made explicit. Could this be clarified and elaborated?
- 10 I think the abstract could be clearer when describing the methods and the results separately (e.g. “We analyzed historical data and simulations of a control period as well as a future period. The results are X”. Instead of “Our analysis of historical data showed X. Our analysis of simulation data for the control period also showed X.”) as well as more concrete (e.g. “combinations of climate model A and climate model B with emission scenarios RCP8.5 and RCP2.6” “combinations of two highly resolved climate models and two emission scenarios”)
- 23-27: Looking at the references, this seems to describe the case of the North Western Coast of Germany and should be mentioned somewhere.
- 63 what is meant by “integrative scenario-based climate impact assessments”? Is this a standard term?
- 64, the authors state that such “integrative scenario-based climate impact assessments” are not available for local to regional flooding at shallow coasts. Could the authors provide references for global analyses or non-shallow coasts? Or similar efforts focusing only on historical data, but not for future climate projections? I am missing a more comprehensive literature review here on studies related to compound flooding, and in particular approaches using physics-based models.
- 120 Different types of compound events are introduced as part of the method section, but not used or reflected upon in the remainder of the study. If there is no explicit link with the results or discussion sections, this should be left out.
- 135: An explanation (summary from van den Hurk) of the reasoning behind selecting the 15 largest events would be helpful.
- 137. “by selecting the 15 largest events according to sea level, precipitation and inland water level”. I was not sure if the authors meant 15 events in total or 45 events in total until I saw Figure 5.
- 142 I was confused by the sentence “While the analysis […] for the evaluation of the climate change projections a new model set-up was necessary.” In the end, the authors still follow the approach of van den Hurk et al. 2015, but instead of historical data they use simulated data for which they had to set up a model. Is this what is meant here?
- Figure 2: This flow-chart is helpful to understand the model chain. I suggest to expand it to include the entire methodology of the paper.
- Authors describe magnitudes with terms as “largest”, “moderate” or “intense” and “heavy”, but it remains unclear what these mean.
- 210 / Figure 3: Are the circles the underlying data points and the line the fit? Could the authors provide more context and how this function is obtained? What has been received? Can this data be shown? How was the polynomial estimated?
- 220-225: What do “very high”, “high” and “moderate” mean? Could this be made more concrete in the results and discussion section? The same holds for adjectives “intense” and “heavy” that the authors use in other parts of the results and discussion section.
- 223 “… lead to the largest observed deviations from the regulated inland water level (in total 9 of the 15 events)”: Which 9 events are these? This is not clear from Figure 4. And where and why has this threshold in deviation from regulated inland water level been set? How is the size of the circle related to the difference? Linearly? Logarithmic? Would the scale impact how we interpret the results? In general, the interpretation of the Figure is not immediately intuitive to me. Perhaps the authors could guide the reader through the three groups (of 15 events each) separately and then arrive at their overall conclusion.
- Figure 4: The caption states that the Figure shows the water level anomaly [cm] however the number in the circles run from 1- 45. These are probably event IDs and should be changed to water level anomaly values?
- 247 Why is the control period not chosen to be identical with the period for which observational data is available? Wouldn’t this make the comparison more straight forward and, for example, enable plotting using the same axes limits?
- 321 “most of the storm tides do not lead to significant drainage overload”: How is “significant” defined? Why?
- Figure 10 should be plotted in the same Figure or at least with the same y-axis. Now the results look more similar than they are.
- In some Figures HadGEM is on the right panel and in others on the left. I suggest to keep this the same.
- Figure 11: There seems to be a negative correlation between sea level and precipitation. How could this be explained and how is this of relevance to the results?
- 351: The authors cite Heinrich et al who states that compound events of storm tide and high river discharge are not observed more frequently than to be expected by chance. How will this be in the future? What are the limitations to using individual projects for winter precipitation, runoff generation and sea level rise (l. 357)
- 411: It would be great if the data could be published alongside the paper.
- This seems to be a similar paper focusing on a different case study: https://www.sciencedirect.com/science/article/abs/pii/S0022169423001087
Citation: https://doi.org/10.5194/egusphere-2024-29-RC1 -
AC1: 'Reply on RC1', Helge Bormann, 13 Apr 2024
Dear reviewer, dear Wiebke Jäger,
Thanks a lot for your helpful comments which will improve the clarity our manuscript. We go through your comments point by point to clarify how we will use your comments and questions during revision.General comments
1. This is an interesting paper showing that compound flooding at the Northwestern coast in Germany, can be driven by combination of moderate or intense drivers and not only by extreme individual drivers now as well as under future climate projections. The method is clear and appropriate. The results are relevant for the scientific community working on compound flooding as well as for adaptation and mitigation planning of local and regional water boards of the studied area. Below comments indicate elements of the manuscript that are not yet (entirely) clear to me.Thanks a lot, we appreciate your positive assessment of our study.
Specific comments
1. Title: Could be more concrete mentioning the study area instead of “coastal areas” as I think the results are mainly relevant for the particular case study rather than for coastal areas in general.Thanks for the comment, we suggest: “Model based assessment of climate change impact on inland flood risk at the German North Sea coast caused by compounding storm tide and precipitation events”
2. The authors use the term “storm tide”, “storm surge”, “sea level”, “maximum sea level” and “storm flood” throughout the text. Are these used interchangeably or with different meanings? A common definition is water level = tide + surge (+ sea level rise), but it does look like the authors use them to refer to the same. An introduction and definition of the terms used would provide more clarity. I also suggest to use one term throughout for clarity.
You are right, suggesting standard terms for the description of the extreme events we analyse. Therefore, in the original manuscript, we used the terms according to Pugh and Woodworth (2014): Sea Level Science: Understanding Tides, Surges, Tsunamis and Mean Sea-Level Changes (https://doi.org/10.1017/CBO9781139235778) without citing the paper explicitly for the definitions used.
In the revised version of the text, we will explain the terms used and reduce the use of different terms as suggested.3. The authors focus on events of storm tide and precipitation events. Why do they not consider river discharge as well? It seems the authors see river discharge and precipitation as strongly linked in the area of the study (e.g., by mentioning the work of Henrich et. 2023 on storm tide and river discharge), but this is not made explicit. Could this be clarified and elaborated?
This is a good point. The Emden water board is located downstream the Ems River, at the Ems estuary. At gauge Knock, the estuary is more than 3 km wide, so the impact of river discharge on the water level in the estuary is small compared to the impact of storm surges. Furthermore, there is a lag between high river runoff and precipitation depending on catchment size while due to the small area in this study surge and precipitation events occur simultaneously. So we can not rule out that there is a contribution, but a different time lag is needed for the largest impact. Therefore, we neglected such influence. However, for the transferability of our results e.g. to upstream water boards at the Ems River, this is essential. We will add information on the estuary to section 2.1 (target area) and address these issues in the discussion section (also requested by reviewer 3).
4. 10 I think the abstract could be clearer when describing the methods and the results separately (e.g. “We analyzed historical data and simulations of a control period as well as a future period. The results are X”. Instead of “Our analysis of historical data showed X. Our analysis of simulation data for the control period also showed X.”) as well as more concrete (e.g. “combinations of climate model A and climate model B with emission scenarios RCP8.5 and RCP2.6” “combinations of two highly resolved climate models and two emission scenarios”)
We will rewrite the abstract while considering your suggestions.
5. 23-27: Looking at the references, this seems to describe the case of the North Western Coast of Germany and should be mentioned somewhere.
You are right, we will add that information in the first sentence of the introduction.
6. 63 what is meant by “integrative scenario-based climate impact assessments”? Is this a standard term?
We did not aim at using this expression as a standard term. We used it to express the integrative calculation of the potential impacts of storm surges and rainfall based on a consistent set of climate scenarios. We will make that clear by revising this paragraph.
7. 64, the authors state that such “integrative scenario-based climate impact assessments” are not available for local to regional flooding at shallow coasts. Could the authors provide references for global analyses or non-shallow coasts? Or similar efforts focusing only on historical data, but not for future climate projections? I am missing a more comprehensive literature review here on studies related to compound flooding, and in particular approaches using physics-based models.
You make a good point. As mentioned in the manuscript, we did not find any comparable modelling study in the literature. On the global scale, water management along shallow coast is not explicitly considered. However, data based studies are available (e.g., by van den Hurk et al., 2015), and thankfully you provided the citation of Xu et al. (2023) in your review, which shows similarities coupling an urban storm water model to a hydrodynamic model. We will extend this literature review in the revision as suggested (also requested by the other two reviewers).
8. 120 Different types of compound events are introduced as part of the method section, but not used or reflected upon in the remainder of the study. If there is no explicit link with the results or discussion sections, this should be left out.
Thanks for this note. We used the classification only implicitly for the discussion of the explanatory power of our model approach. So far, we assume that the multivariate mechanism can explain the pattern of the events. But this is not sufficient, preconditions such as soil moisture in the marsh are affects the pattern significantly. We will therefore revise the manuscript in order to use the classification explicitly in the discussion of the identified compound events in the sections 3.1. and 3.2.4 (and later on in the discussion section).
9. 135: An explanation (summary from van den Hurk) of the reasoning behind selecting the 15 largest events would be helpful.
Van den Hurk et al. (2015) showed the 20 largest inland water levels and compared them to the highest 10 wind-induced surges and precipitation events. Due to limited data availability (20 year time series of observations) we made compromise between the relative small number system overloads in the observations and a minimum number of events for evaluation. We ended up in 15 events, which is in the same range as applied by van den Hurk et al. (2015), and which we applied to observed water levels, wind-induced surges and precipitation events.
10. 137. “by selecting the 15 largest events according to sea level, precipitation and inland water level”. I was not sure if the authors meant 15 events in total or 45 events in total until I saw Figure 5.
We selected 15 events each, resulting in max. 45 events if the selected events are independent. We will clarify that during revision.
11. 142 I was confused by the sentence “While the analysis […] for the evaluation of the climate change projections a new model set-up was necessary.” In the end, the authors still follow the approach of van den Hurk et al. 2015, but instead of historical data they use simulated data for which they had to set up a model. Is this what is meant here?
Yes, this is one aspect. To be able to analyse future conditions, we use simulated data instead of observed data. Another aspect is that we needed to replace the observed water level [m] by the simulated system overload [m³]. This is due to the fact that we are no able to model the whole anthropogenic drainage system by a hydrodynamic model. This is because the local water management action is not documented. Therefore, we compare runoff generation against drainage capacity, resulting in a system overload described by a water volume, instead of calculating inland water levels. We will clarify that by rewriting the paragraph.
12. Figure 2: This flow-chart is helpful to understand the model chain. I suggest to expand it to include the entire methodology of the paper.
We fear one flow chart including the whole methodology will be too complex. But we think about the possibility to visualize the data-based methodology according to van den Hurk et al. (2015), as well.
13. Authors describe magnitudes with terms as “largest”, “moderate” or “intense” and “heavy”, but it remains unclear what these mean.
This is a good point, also mentioned by the other reviewers. During revision, we will define the terms used and reduce the number of terms, if necessary.
14. 210 / Figure 3: Are the circles the underlying data points and the line the fit? Could the authors provide more context and how this function is obtained? What has been received? Can this data be shown? How was the polynomial estimated?
As mentioned in the paragraph prior to Figure 3, the drainage function of the pumps at gauge Knock is composed through an assessment of the state agency for the free flow through the gate in case of low tide (=negative hydraulic head) and a calculation of the pump capacity per pup, based on the pump characteristics and of the blade position (positive hydraulic head). According to the Emden water board, for every 50 cm (= for each 50 cm interval of hydraulic head) the blade position is adjusted, and therefore a new estimation of pump capacity is necessary using the pump curve (points). Above a hydraulic head of 3.75 m, the pumps need to be switched off. The polygon is the best fit to the pump rate estimations, selected by least squares method.
If required, we are happy to provide such explanation in the paragraph; however, we don’t think that providing the pump characteristics (e.g., through an additional figure), and the blade positions for every 0.5 m (e.g., in an additional table) is beneficial for the readers.15. 220-225: What do “very high”, “high” and “moderate” mean? Could this be made more concrete in the results and discussion section? The same holds for adjectives “intense” and “heavy” that the authors use in other parts of the results and discussion section.
As mentioned above: revision, we will define the terms used and reduce the number if necessary.
16. 223 “… lead to the largest observed deviations from the regulated inland water level (in total 9 of the 15 events)”: Which 9 events are these? This is not clear from Figure 4. And where and why has this threshold in deviation from regulated inland water level been set? How is the size of the circle related to the difference? Linearly? Logarithmic? Would the scale impact how we interpret the results? In general, the interpretation of the Figure is not immediately intuitive to me. Perhaps the authors could guide the reader through the three groups (of 15 events each) separately and then arrive at their overall conclusion.
The 9 events you ask for are shown in Figure 4. According to the figure caption, those events are the orange circles with red or/and blue edging. The remaining 6 orange circles (without any red or blue edging) are those compound events which member belong neither to the 15 highest sea levels nor to the highest 15 precipitation events. As mentioned in the caption, the diameter of the circles is proportional to the values. As suggested, we will improve the description in the text to make it more intuitive.
17. Figure 4: The caption states that the Figure shows the water level anomaly [cm] however the number in the circles run from 1- 45. These are probably event IDs and should be changed to water level anomaly values?
This is not correct; the number are the water level anomalies in [cm] and not IDs. The diameter of the circles is proportional to the values. We will clarify this in the figure caption.
18. 247 Why is the control period not chosen to be identical with the period for which observational data is available? Wouldn’t this make the comparison more straight forward and, for example, enable plotting using the same axes limits?
We chose a standard control period, which is usually used for the interpretation of climate model output. The time period of the observed data (20 years) is too short to use it as control period for climate variables (>= 30 years).
19. 321 “most of the storm tides do not lead to significant drainage overload”: How is “significant” defined? Why?
The term significant is not used in a statistical sense. From the figures 11 and 12 it can be seen that that most of the highest storm tides do not induce any drainage systems overload. We will adjust the wording to make that clear.
20. Figure 10 should be plotted in the same Figure or at least with the same y-axis. Now the results look more similar than they are.
It’s correct that the results look more similar than they are. We will adjust the axes. However, the general trend (increasing number of days with system overload at the end of the century, higher number of days with system overload in the RCP8.5 scenario) is similar for both climate models.
21. In some Figures HadGEM is on the right panel and in others on the left. I suggest to keep this the same.
Thanks for this note. We will adjust the order of the subfigures.
22. Figure 11: There seems to be a negative correlation between sea level and precipitation. How could this be explained and how is this of relevance to the results?
This is a good question. However, you should keep in mind that the correlation might be an artefact by putting the three separate ensembles (15 events each for highest sea levels, highest drainage system overloads, highest precipitation events) together. Within the three ensembles, there is no such trend visible. Anyhow, one explanation could be that the highest storm tides are generated by NW winds, while the highest precipitation is generated by westerly winds (causing only moderate storm tides).
23. 351: The authors cite Heinrich et al who states that compound events of storm tide and high river discharge are not observed more frequently than to be expected by chance. How will this be in the future? What are the limitations to using individual projects for winter precipitation, runoff generation and sea level rise (l. 357)
Heinrich et al. (2023) showed that the simultaneous occurrence of storm surges and storm surge series in conjunction with long lasting precipitation predominantly occurs during cyclonic westerly weather conditions. Therefore, westward estuaries are mainly at risk.
Based on a comprehensive ensemble study it could be shown that in all climate models and scenarios an increased frequency of atmospheric cyclonic westerlies in winter is to be expected by 2100 (Heinrich et al., 2024). So, an increasing flood risk can be assumed for westerly estuaries while not for northward estuaries like the Ems River.
We will comment on this in the revision. Heinrich et al. (2023b) (10.3389/fclim.2023.1227613) further showed that in future scenarios compound events will increase mainly because of rising mean sea level which will further increase the risk and occurrence of events. We will include this in the discussion.24. 411: It would be great if the data could be published alongside the paper.
We will discuss this option. First, data will be made available on demand.
25. This seems to be a similar paper focusing on a different case study: https://www.sciencedirect.com/science/article/abs/pii/S0022169423001087
Thanks for this note! We will include this paper into the literature review and the discussion.Heinrich P, Hagemann S, Weisse R et al. (2023) Compound flood events: Analysing the joint occurrence of extreme river discharge events and storm surges in northern and central Europe. Natural Hazards and Earth System Sciences 23(5):1967–1985. doi: 10.5194/nhess-23-1967-2023
Heinrich, Philipp; Hagemann, Stefan; Weisse, Ralf; Gaslikova, Lidia (2023b): Changes in compound flood event frequency in northern and central Europe under climate change. In: Frontiers in Climate 5, S. 1227613. DOI: 10.3389/fclim.2023.1227613.
Heinrich P, Hagemann S, Weisse R (2024) Automated Classification of Atmospheric Circulation Types for Compound Flood Risk Assessment: CMIP6 Model Analysis Utilising a Deep Learning Ensemble. doi: 10.21203/rs.3.rs-4017900/v1
van den Hurk, B., van Meijgaard, E., de Valk, P., van Heeringen, K.-J., and Gooijer, J.: Analysis of a compounding surge and precipitation event in the Netherlands. Environmental Research Letters 10, 035001, doi.org/10.1088/1748-9326/10/3/035001, 2015.
Xu, K., Zhuang, Y., Bin, L., Wang, C., Fuchang, T. (2023) Impact assessment of climate change on compound flooding in a coastal city. Journal of Hydrology 617/C, 129166.Citation: https://doi.org/10.5194/egusphere-2024-29-AC1
-
RC2: 'Comment on egusphere-2024-29', Anonymous Referee #2, 28 Mar 2024
This manuscript evaluates the role of storm tides and precipitation and their compounding effects in creating drainage system overloads for a case study are in Germany both under present-day and future climate conditions. This is an interesting and timely topic and the manuscript will be a valuable addition to the existing literature. I provide some general and several detailed comments below which should be taken into consideration before the manuscript can move toward publication.
General comments:
- I would put less emphasis on the differences in changes found between the two emission scenarios (for example in the abstract). The fact that 8.5 leads to more change than 2.6 is to be expected. Personally I find the results for present-day most interesting showing which types of events cause the most problems, and the fact that models can at least capture the broad characteristics of those events. The future change part is less exciting I would say, so I suggest not putting it too much at the forefront.
- The referencing is very biased toward self-citations (some references only available in German), and while that is in general ok and often warranted, there are parts where other literature should be cited; for example, in the intro when the broader topic of compound flooding is introduced, but other places as well.
Specific comments:
11 storm surge series?
25/26 are the “storm floods” from inland or coastal flooding drivers? Terminology is not clear in many places when referring to different flood drivers.
36 typo: the end of the century
48/49 I couldn’t follow that sentence. The fact that drainage systems worked at the edge of capacity means they weren’t overloaded, but the occurrence of regional flooding indicates otherwise.
64 see above, what does “storm flood” mean here
68 The models should not just capture tidal water level dynamics but water levels in general including storm surge
97 typo: of sea water level
108 typo: are part of the CMIP5
120-130 Zscheischler et al. (2020) do not provide a definition of compound flooding but a typology; an alternative definition to the one from the IPCC was proposed in Zscheischler et al. (2018; https://doi.org/10.1038/s41558-018-0156-3)
163 typo: for potential evapotranspiration
230 Figure 4 and other figures as well, please make clear what the numbers in the circles represent exactly
300/301 Why do you think that is? It would help the reader to get some idea about where such differences could come from.
Citation: https://doi.org/10.5194/egusphere-2024-29-RC2 -
AC2: 'Reply on RC2', Helge Bormann, 13 Apr 2024
Dear reviewer,
Thanks a lot for your helpful comments which will improve the clarity our manuscript. We go through your comments point by point to clarify how we used your comments and questions during revision.General comments
1. This manuscript evaluates the role of storm tides and precipitation and their compounding effects in creating drainage system overloads for a case study are in Germany both under present-day and future climate conditions. This is an interesting and timely topic and the manuscript will be a valuable addition to the existing literature. I provide some general and several detailed comments below which should be taken into consideration before the manuscript can move toward publication.Thanks a lot, we appreciate your positive assessment of our study.
2. I would put less emphasis on the differences in changes found between the two emission scenarios (for example in the abstract). The fact that 8.5 leads to more change than 2.6 is to be expected. Personally I find the results for present-day most interesting showing which types of events cause the most problems, and the fact that models can at least capture the broad characteristics of those events. The future change part is less exciting I would say, so I suggest not putting it too much at the forefront.
Thanks for sharing your opinion on which parts of our investigation are more interesting than others. Indeed, the present-day analysis reveals which kind of events cause most water management problems, and the results are in agreement to those found for the Netherlands by van den Hurk et al. (2015). Reviewer 1 made some suggestions to make the main findings intuitively understandable. Unfortunately, data availability of German coastal water boards is scarce so that a spatial comparison for different locations is not possible. In our opinion it seems to be crucial to test whether model based analysis can depict the impact of compound events on coastal flood risk and on the potential system overload of drainage systems. Therefore, we think that the modelling study has a significant added value and innovation aspect. Applying the model chain to the control period proves that the generation mechanism of compound events is represented. Its application to future conditions confirms that compound events will probably probably also happen in the future and will be more intense than under today’s conditions. However, we agree with you, that it’s not necessary to discuss the differences in the scenarios in too much detail. We will check in the revision where to focus such discussion without affecting the main findings. In general, it seems to be relevant to consider the generation of hydro-meteorological compound events for climate change adaptation planning.
3. The referencing is very biased toward self-citations (some references only available in German), and while that is in general ok and often warranted, there are parts where other literature should be cited; for example, in the intro when the broader topic of compound flooding is introduced, but other places as well.
Thanks for reminding us to put more emphasis on considering the available literature. In particular, in the introduction we will replace the self-citations by citing other relevant publications to describe the challenges and the status quo of investigations and available solutions. However, for some statements, we did not find suitable other references so far. Since reviewer 1 also suggested another similar study to discuss, we will be sure to improve the literature review.
Specific comments
1. 11 storm surge series?Yes, thanks for this correction.
2. 25/26 are the “storm floods” from inland or coastal flooding drivers? Terminology is not clear in many places when referring to different flood drivers.
Thanks, your right. Here, storm surges are meant, as well. We will critically check the wording throughout the manuscript during revision.
3. 36 typo: the end of the century
Thanks for this correction.
4. 48/49 I couldn’t follow that sentence. The fact that drainage systems worked at the edge of capacity means they weren’t overloaded, but the occurrence of regional flooding indicates otherwise.
Thanks for this note. In general, regional water managers report that the drainage systems more often work at the edge of their capacity in our days. But during the described event in 2022 local flooding occurred, because of a system overload.
We will rewrite the sentence.5. 64 see above, what does “storm flood” mean here
Storm flood means storm surge, here. We will critically check the wording throughout the manuscript during revision.
6. 68 The models should not just capture tidal water level dynamics but water levels in general including storm surge
You are correct. We will rewrite the sentence and consider your suggestion.
7. 97 typo: of sea water level
Thanks for this correction.
8. 108 typo: are part of the CMIP5
Thanks for this correction.
9. 120-130 Zscheischler et al. (2020) do not provide a definition of compound flooding but a typology; an alternative definition to the one from the IPCC was proposed in Zscheischler et al. (2018; https://doi.org/10.1038/s41558-018-0156-3)
Thanks for this suggestion! We will consider your suggestion and specify the sentence during revision.
10. 163 typo: for potential evapotranspiration
Thanks for this correction.
11. 230 Figure 4 and other figures as well, please make clear what the numbers in the circles represent exactly
In Figure 4, the numbers in the circles are the observed deviations from the target inland level [cm], as mention in the figure caption.
In Figure 5 the numbers quantify the calculated drainage system overload [106 m³], as mention in the figure caption.
To be able to analyse future conditions, we use simulated data instead of observed data. Here we needed to replace the observed water level [m] by the simulated system overload [m³]. This is due to the fact that we are no able to model the whole anthropogenic drainage system by a hydrodynamic model. This is because the local water management action is not documented. Therefore, we compare runoff generation against drainage capacity, resulting in a system overload described by a water volume, instead of calculating inland water levels.
We will emphasize this during revision and check the text to avoid any misunderstanding.12. 300/301 Why do you think that is? It would help the reader to get some idea about where such differences could come from.
As discussed in a later section of the manuscript, the two climate models simulate extreme most of the extreme events in different seasons, MPI in late autumn, HadGEM in winter. Also the characteristics of the simulated events differ. The HadGEM model simulated long-lasting events, resulting in a higher amount of days with system overload (compared to MPI model), while the number of critical events is similar to the MPI model.
During revision, we will discuss your question and provide insight into these details.
van den Hurk, B., van Meijgaard, E., de Valk, P., van Heeringen, K.-J., and Gooijer, J.: Analysis of a compounding surge and precipitation event in the Netherlands. Environmental Research Letters 10, 035001, doi.org/10.1088/1748-9326/10/3/035001, 2015.Citation: https://doi.org/10.5194/egusphere-2024-29-AC2
-
RC3: 'Comment on egusphere-2024-29', Anonymous Referee #3, 05 Apr 2024
The manuscript “Model based assessment of climate change impact on inland flood risk in coastal areas caused by compounding storm tide and precipitation events” shows how different combinations of sea levels and precipitation at present and in the future combine to drive inland flooding events/drainage system overloads at a water board in Germany. The manuscript is well written, and the conclusions seem mostly backed up by the results. I have a few comments, summarized below, that I think would strengthen the results presented in the manuscript.
General comments:
One of the main conclusions of the research is that the combination of moderate storm tide and precipitation events drive inland flooding or drainage system overload. However, what is considered “extreme” versus “moderate” across the distribution of events, is never sufficiently defined. The 15 highest events end up being the metric that is designated as “extreme”, so my assumption is anything below those events are considered low or moderate. Looking at Figure 4, the 15 highest sea levels (defined as the average across 5 tidal cycles) for the observations are all above 0.8m and the 15 highest 3-day precipitation events are all above 35-40 mm. This would lead me to believe any combination of sea level below 0.8m or precipitation below 35-40 mm would be “moderate” or “low.” However, when moving to the drainage capacity and future climate scenarios for the winter, “extreme” sea level increases to over 1.5 m or 2.0 m, depending on time period and climate forcing, and precipitation extremes increase to over 50 mm, but even as high as 60 or 70 mm. Thus, in the future scenario, there are many combined events that fall over what was called the “extreme” range in the observations, even if they are not the most extreme in the new climate, which leads me to conclude that extreme sea level and precipitation events are causing many of the drainage capacity extremes, rather than moderate events. So, a broader discussion on the definition of extreme vs moderate events is necessary for backing up the main conclusions of the work.I understand that the inland water level can’t necessarily be simulated for the future climate, and the authors justify values/results might be different in the simulated climate control period because they’re using a different metric – drainage system overload. A question I have is, can’t the drainage system overload be calculated for the observational data in the same way and compare the types of events they see driving drainage system overload versus inland water level difference (what is currently calculated?) This could be helpful in distinguishing whether any variation in driving events is due to the different metric or is due to the model simulations.
Figure 4 shows that some of the highest sea levels, precipitation events, and inland water levels co-occur (where blue, red, and orange align). The MPI model shows this as well (using the drainage capacity metric) while the HadGEM model rarely shows this. Can this be physically explained? Does this mean the HadGEM model is actually not a great representation of the climate system?
Following on, figures 4, 5, 6, 11, and 12 require the reader to estimate how many “extreme” precip, sea level, and inland water level/drainage capacity co-occur. It would be interesting to actually quantify the number of drainage capacity/inland water level extremes that do NOT occur versus those that do occur with the driving variables, which would more easily pull out the statistics/comparisons of co-occurrence across the plots.
Discussion:
The manuscript results are very specific to the location. A broader discussion of the results and how they may relate or be extended to other locations should be added to ensure the overall findings relate to the broader readership of NHESS.I was expecting more of a discussion on the variation across the future climate model results. The discussion presents the future scenario results as very similar to each other, but I see distinct differences. For example, why does sea level matter more for the MPI events than it does for the HadGEM?
The timing (during the year) of precipitation events is highlighted in the results and discussion, but there’s no similar analysis for sea level. Do high sea level events tend to coincide in the same months as high precipitation? If they do not, is this the reason there aren’t more extreme sea level and precipitation events combining?
Line by Line/Specific comments:
- Sea levels need a datum specified in all plots
- In the introduction different types of compound events are defined, but only one is considered here. I’m not sure all need to be brought up.
Line 32: “Zealand” spelled incorrectly
Line 43: It is not clear what is meant by “individual events mentioned above” is. One at a time? The connection isn’t clear.
Line 57: I believe the Svensson and Jones 2002 article didn’t identify more compound flood events, but a higher dependence between compounding forcing variables. They didn’t model flooding.
Line 88: How extreme are we talking?
Line 115: How well does the bias correction do – especially with the extreme data as that is the focus of this study?
- The precipitation metric is 3-day precipitation and the sea level metric is an average over 5 tidal periods. While this was used in previous literature that is cited for a specific storm event and region, this manuscript needs a justification as to why those metrics are used and relevant for this location.
Line 156: the inland water level is assumed to be constant for the drainage capacity calculations. How good of an assumption is this? How different are the results if the inland water level is varied?
- Values of future sea level projections at mid or end of century would be helpful!
Line 226: It’s mentioned that sea level and precipitation don’t represent the sole drivers of inland water anomalies – what is meant by this? Nonlinearities? Other variables? This could be expanded upon in the discussion section.
Line 380: will “not be”
- How many realizations of each different climate model was used? For example is the HadGEM results representative of one end member or an ensemble? And if an ensemble, how many endmembers were used?
Figures:
Figure 1: Needs an improved legend or caption. There are lots of components of the map that aren’t well defined. Needs datum for elevation.
Figures 4,5,6,11,12: Can the numbers within the circles be explained somewhere in the captions? Sea levels need datum.
Figure 10: I find it pretty interesting that for the MPI the days with system overload are really similar for mid century 8.5 and end of century 2.6!
Citation: https://doi.org/10.5194/egusphere-2024-29-RC3 -
AC3: 'Reply on RC3', Helge Bormann, 13 Apr 2024
Dear reviewer,
Thanks a lot for your helpful comments which will improve the clarity our manuscript. We go through your comments point by point to clarify how we used your comments and questions during revision.General comments
0. The manuscript “Model based assessment of climate change impact on inland flood risk in coastal areas caused by compounding storm tide and precipitation events” shows how different combinations of sea levels and precipitation at present and in the future combine to drive inland flooding events/drainage system overloads at a water board in Germany. The manuscript is well written, and the conclusions seem mostly backed up by the results. I have a few comments, summarized below, that I think would strengthen the results presented in the manuscript.Thanks a lot, we appreciate your positive assessment of our study.
1. One of the main conclusions of the research is that the combination of moderate storm tide and precipitation events drive inland flooding or drainage system overload. However, what is considered “extreme” versus “moderate” across the distribution of events, is never sufficiently defined. The 15 highest events end up being the metric that is designated as “extreme”, so my assumption is anything below those events are considered low or moderate. Looking at Figure 4, the 15 highest sea levels (defined as the average across 5 tidal cycles) for the observations are all above 0.8m and the 15 highest 3-day precipitation events are all above 35-40 mm. This would lead me to believe any combination of sea level below 0.8m or precipitation below 35-40 mm would be “moderate” or “low.” However, when moving to the drainage capacity and future climate scenarios for the winter, “extreme” sea level increases to over 1.5 m or 2.0 m, depending on time period and climate forcing, and precipitation extremes increase to over 50 mm, but even as high as 60 or 70 mm. Thus, in the future scenario, there are many combined events that fall over what was called the “extreme” range in the observations, even if they are not the most extreme in the new climate, which leads me to conclude that extreme sea level and precipitation events are causing many of the drainage capacity extremes, rather than moderate events. So, a broader discussion on the definition of extreme vs moderate events is necessary for backing up the main conclusions of the work.
This is a good point, also mentioned by the other reviewers. During revision, we will clearly define the terms used (extreme, moderate, low).
2. I understand that the inland water level can’t necessarily be simulated for the future climate, and the authors justify values/results might be different in the simulated climate control period because they’re using a different metric – drainage system overload. A question I have is, can’t the drainage system overload be calculated for the observational data in the same way and compare the types of events they see driving drainage system overload versus inland water level difference (what is currently calculated?) This could be helpful in distinguishing whether any variation in driving events is due to the different metric or is due to the model simulations.
Thanks for this suggestion. In principal, this is a good point, if data in sufficient resolution would be available. Unfortunately, the observed local climate data (in daily resolution) are not available in required temporal resolution (hourly). Thus, a disaggregation would be required introducing additional uncertainty. However, the system overloads calculated for the control period fit well to the recent experience of the regional water boards. We suggest commenting on this in section 3.2.1. to emphasize the plausibility of the simulation results for the control period.
3. Figure 4 shows that some of the highest sea levels, precipitation events, and inland water levels co-occur (where blue, red, and orange align). The MPI model shows this as well (using the drainage capacity metric) while the HadGEM model rarely shows this. Can this be physically explained? Does this mean the HadGEM model is actually not a great representation of the climate system?
We would be carefully to rate whether HadGEM is able or not to reproduce the climate system just based on this compound characteristic. However, it is remarkable that the 15 highest system overload events in the control period simulated by MPI occur with sea levels which are more than 80 cm higher than by HadGEM. Thus, for the HadGEM model, with similar cumulative rainfall amounts, the antecedent saturation of the soil (before intense rainfall events) seems to have a greater impact for HadGEM simulations. We will discuss this different behaviour in section 3.2.1. in order to emphasise model specific behaviour.
4. Following on, figures 4, 5, 6, 11, and 12 require the reader to estimate how many “extreme” precip, sea level, and inland water level/drainage capacity co-occur. It would be interesting to actually quantify the number of drainage capacity/inland water level extremes that do NOT occur versus those that do occur with the driving variables, which would more easily pull out the statistics/comparisons of co-occurrence across the plots.
This is a good suggestion.
We will provide the statistical numbers for the different kind of events as requested to help the reader distinguishing between the different types. This will help to assess the importance of compound events.Discussion:
5. The manuscript results are very specific to the location. A broader discussion of the results and how they may relate or be extended to other locations should be added to ensure the overall findings relate to the broader readership of NHESS.
Thanks for this important suggestion. We are aware of several other studies with similar behaviour, such as from the Netherlands, Great Britain and from China. We will introduce a paragraph on the possible extension to the locations to the discussion section. Further studies in the compounding effects of storm surges and river flooding will also be helpful for discussion regional transferability.
6. I was expecting more of a discussion on the variation across the future climate model results. The discussion presents the future scenario results as very similar to each other, but I see distinct differences. For example, why does sea level matter more for the MPI events than it does for the HadGEM?
In the original manuscript, we were hesitating to discuss the projections in too much detail due to the small ensemble of combinations of climate models and scenarios. We focused on the general characteristics of the generation of compound events.
But indeed, the scenarios themselves differ, and sea level has a different impact on compound events calculated by MPI and HadGEM models.
One reason is that MPI simulates highest precipitation rates during the storm season in late autumn, while HadGEM also simulates extreme and precipitation events in winter (mentioned already in section 3.2.4.), associated with different event characteristic (e.g., in rainfall duration). We will evaluate those events in more detail for the revised manuscript.7. The timing (during the year) of precipitation events is highlighted in the results and discussion, but there’s no similar analysis for sea level. Do high sea level events tend to coincide in the same months as high precipitation? If they do not, is this the reason there aren’t more extreme sea level and precipitation events combining?
For the climate scenario RCP8.5 we find a slight increase for storm events in fall (for the MPI model) and in winter (for the HadGEM model). But compared to precipitation, such tendency is not as clear for high sea levels. So yes, that may be one reason. The largest compound events for MPI are associated with sea levels which are about 0.5 m higher than for the HadGEM model. This seems to be an important aspect which should be definitely added to the discussion.
Specific comments
1. Sea levels need a datum specified in all plots
We will add the datum to all plots (m above NHN; "standard elevation zero")
2. In the introduction different types of compound events are defined, but only one is considered here. I’m not sure all need to be brought up.
This is a good point, mentioned by the other reviewers, as well.
We used the classification only implicitly for the discussion of the explanatory power of our model approach. In a first attempt, we assume that the multivariate mechanism can explain the pattern of the events. But this is not sufficient, preconditions such as soil moisture in the marsh are affects the pattern significantly. We will therefore revise the manuscript in order to use the classification explicitly in the discussion of the identified compound events in the sections 3.1. and 3.2.4 (and later on in the discussion section).3. Line 32: “Zealand” spelled incorrectly
Will be corrected.
4. Line 43: It is not clear what is meant by “individual events mentioned above” is. One at a time? The connection isn’t clear.
We will adjust the expression. We were aiming at storm surges on the one hand and intense precipitation on the other hand.
5. Line 57: I believe the Svensson and Jones 2002 article didn’t identify more compound flood events, but a higher dependence between compounding forcing variables. They didn’t model flooding.
Thanks for this correction. We will adjust the text accordingly.
6. Line 88: How extreme are we talking?
This is explained later on in figure 3. Pumps need to be switched off from 3.75 m geodetic head upwards. This is around 2 m above average high tide.
7. Line 115: How well does the bias correction do – especially with the extreme data as that is the focus of this study?
Quality of bias correction for the region was described by Bormann and Kebschull (2023) as well as by Ley et al. (2023). We will add the citations to the text.
8. The precipitation metric is 3-day precipitation and the sea level metric is an average over 5 tidal periods. While this was used in previous literature that is cited for a specific storm event and region, this manuscript needs a justification as to why those metrics are used and relevant for this location.
Thanks for this advice. The 3 day period is chosen based on literature (van den Hurk et al., 2015) and the experience of the regional water board members. In the past, coastal drainage infrastructure was dimensioned based on two day extremes. Since already three day storm series are observed, they aim at adaptation planning for extreme events with a duration of three days. We will add this justification to the respective section.
9. Line 156: the inland water level is assumed to be constant for the drainage capacity calculations. How good of an assumption is this? How different are the results if the inland water level is varied?
As shown in figure 4, the maximum deviation from this assumption in the available 20 years of observation was 44 cm. Such deviations only occur in case of system overload. Usually, deviations are significantly smaller and do not have a string influence on pumping rates.
10. - Values of future sea level projections at mid or end of century would be helpful!
We will provide these average projections in section 3.2.2
11. Line 226: It’s mentioned that sea level and precipitation don’t represent the sole drivers of inland water anomalies – what is meant by this? Nonlinearities? Other variables? This could be expanded upon in the discussion section.
We will expand this while relating our case to the typology of compound events. One important quantity is the antecedent soil moisture (saturation of the area prior to the event). Unfortunately, soil moisture is not measured in the region, thus a quantitative evaluation is not feasible.
12. Line 380: will “not be”
Will be corrected
13. How many realizations of each different climate model was used? For example is the HadGEM results representative of one end member or an ensemble? And if an ensemble, how many endmembers were used?
There was just one realisation per climate model available and used in hourly resolution.
14. Figure 1: Needs an improved legend or caption. There are lots of components of the map that aren’t well defined. Needs datum for elevation.
We will check the figure; datum will be added (m above NHN; "standard elevation zero").
15. Figures 4,5,6,11,12: Can the numbers within the circles be explained somewhere in the captions? Sea levels need datum.
The numbers in the circles are explained in the figure captions, but we will try to improve the description. Datum will be added (m above NHN; "standard elevation zero").
16. Figure 10: I find it pretty interesting that for the MPI the days with system overload are really similar for mid century 8.5 and end of century 2.6!
We also recognised this, but we don’t have a physical explanation for this similarity so far.
Bormann, H. and Kebschull, J.: Model based estimation of climate change impacts on the drainage demand of low lying coastal areas in Northwest Germany along the North Sea. Journal of Hydrology: Regional Studies 48, 101451, https://doi.org/10.1016/j.ejrh.2023.101451, 2023.
Ley, A., Hähnel, P., Bormann, H.: Addressing the challenges of climate scenario-based impact studies in modelling groundwater recharge on small barrier islands at the German North Sea coast. Journal of Hydrology: Regional Studies 50, 101578, https://doi.org/10.1016/515 j.ejrh.2023.101578, 2023.
van den Hurk, B., van Meijgaard, E., de Valk, P., van Heeringen, K.-J., and Gooijer, J.: Analysis of a compounding surge and precipitation event in the Netherlands. Environmental Research Letters 10, 035001, doi.org/10.1088/1748-9326/10/3/035001, 2015.Citation: https://doi.org/10.5194/egusphere-2024-29-AC3
-
AC3: 'Reply on RC3', Helge Bormann, 13 Apr 2024
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 275 | 68 | 21 | 364 | 6 | 6 |
- HTML: 275
- PDF: 68
- XML: 21
- Total: 364
- BibTeX: 6
- EndNote: 6
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1