the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 28 Feb 2024
Local floods in Madeira Island between 2009 and 2021. Rainfall analysis and risk assessment in mountain streams
Abstract. The torrential floods that occurred in December 2020 and January 2021 in villages located mainly on the northern side of Madeira Island, are analysed in comparison with other local scale events that occurred in 2009, 2012 and 2013. The term torrential flood is adopted in this work to designate the hydrogeomorphologic events, characterized by the occurrence of interactions between slope movements (landslides, debris flows) and fluvial dynamics (flash floods), typical in mountainous regions, formed by deep and narrow-bottomed valleys, as is the case in Madeira Island.
The heavy rainfall episodes of 25 December 2020 and 7 January 2021 were particularly surprising for the localised incidence of numerous occurrences of landslides, debris flows, and flash floods triggered by them.
Intense precipitation episodes, with hourly maximums of more than 40 mm and accumulated amounts in 24 hours exceeding, in some rain-gauges, the critical threshold of 300 mm, triggered peak discharges in several catchments, the largest with an area of up to 50 km2 and the smallest with less than 20 km2. The flow hydrographs of some streams show several flood peaks, associated with secondary maximums of heavy rainfall.
Daily and sub-daily rainfall peaks over catchments are crucial conditions for triggering floods. It was found that there is a very strong correlation between maximum precipitation in 24 hours and 12 hours, which can be mathematically described by a linear regression law. On the other hand, antecedent daily rainfall may also influence the occurrence of torrential floods in Madeira Island: the analysis of the combinations of critical pairs, consisting of the maximum rainfall in 24 hours and the antecedent rainfall calibrated for different durations (in days), allowing the identification of polynomial regression rules, which can be used to detect, in time and space, critical scenarios of torrential floods. The constitution and maintenance of an inventory of critical precipitation values, represents a relevant source of information to ensure the quality of the results issued by any flood early warning system.
Finally, hydrogeomorphological susceptibility mapping, which was produced through the collection of field information in the aftermath of catastrophic floods, embodies a complementary approach that provides a distinct spatial perspective of the areas of sediment production, flooding and deposition.
- Preprint
(2513 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-513', Anonymous Referee #1, 03 Apr 2024
Dear Author
I reviewed your paper “Local floods in Madeira Island between 2009 and 2021. Rainfall analysis and risk assessment in mountain streams”. Despite I recognize that you and your co-authors worked a lot on it, and your English is perfectly understandable, I think that your paper is still not ready for an international journal as NHESS. In the following, I will describe the points that not convinced me and I give you some suggestions to strength them.
- First of all, the paper is shaped as a case study, a report of events occurred in a specific site, without any possible relation with other case studies/sites. I suggest to ask yourself what a reader (that is not interested to your study area) could learn from your paper. The methodology? In this case you should strength the methodology description, introducing a flow chart and keeping in mind people that are interested in reproduce the same approach. You should give less emphasis to the description of the effects of the events, which are local, and repeat themselves everywhere, even with stronger severity.
- It is difficult to appreciate the actual significance of this research because methodology and results are merged and it is difficult to understand what are the premises and what are the results. More in general, you should work to highlight: what is the aim of the paper? What you are doing in this paper? And why? What is the need that this paper is going to meet? In what way this will happen? Answering to these questions could help you to rearrange the paper in a way more similar to the classical structure of scientific papers.
- Particularly, Introduction must state the objectives of the work and provide an adequate background, placing the study in a broad context and highlight why it is important. It should define the purpose of the work and its significance and the current state of the research field reviewed using key (and recent) publications, also highlighting controversial and diverging hypotheses. Oppositely, your Introduction is an introduction to the study area and no recent papers with a similar approach are quoted and compared in terms of objectives and methodology to your paper
- Methodology must provide sufficient details to allow the work to be reproduced by an independent researcher
- Conclusions must report the conclusions (not the synthesis) obtained from your study and possible applications in other frameworks, highlighting whether some of them have generic application, rather than being case-specific
- About figures: you should do a selection of most significant figures for readers customized to see geomorphic effects of floods and landslides. Captions of figures and tables are very long: I suggest to review them.
- Tables should be formatted in a better way, avoiding redundancies and rows with more than a line (you should simply enlarge the columns…). And moreover, in an international journal you should try to summarize instead of present long lists of data. Actually, you should identify and propose only data functional to highlight something (i.e. review tab 4).
- References denotes the absence of a specific research on recent similar papers. I suggest to do this step before rearrange the paper and to add to references only the articles quoted in the paper (Blöschl is only in references)
I hope that these suggestions can help you to improve your paper.
Best regards
Citation: https://doi.org/10.5194/egusphere-2024-513-RC1 -
AC1: 'Reply on RC1', Sergio Lopes, 12 Apr 2024
Dear Referee,
First of all, thank you for your work and all the care you took to read our article carefully. All your comments are pertinent. In this sense, despite the time constraints, I will do my best to try to make the appropriate changes to improve the quality of the final text.
The primary objective of the article is to apply a specific methodology to derive precise information regarding critical rainfall thresholds that may lead to torrential floods. In practice, this involves identifying the critical combination between the rainfall during the event and the accumulated rainfall from preceding days and weeks. Such information serves as a crucial input variable for a flood early warning system. In fact, it is essential to provide a more comprehensive framework for this issue both in the introduction and in the methodology. Thank you.
I appreciate the suggestion to select the most significant figures, because in fact we can better visualise the effects of floods through an example. I also agree with the revision of table 4. It makes perfect sense.
Regarding the references, there is a lack of more recent sources, as the original version of this text was created some time ago. However, I will update the bibliography, especially to enhance the article’s introduction.
I really appreciate your contribution to improving the work I've been doing on this subject.
Best regards
Sérgio Lopes
Citation: https://doi.org/10.5194/egusphere-2024-513-AC1
-
RC2: 'Comment on egusphere-2024-513', Anonymous Referee #2, 05 Apr 2024
This study is a comparative analysis of torrential floods that occurred in specific villages on the northern side of Madeira Island during December 2020 and January 2021, compared with earlier local events from 2009, 2012, and 2013. The study focuses on understanding the hydrogeomorphological events characterized by interactions between slope movements (landslides, debris flows) and fluvial dynamics (flash floods) in mountainous regions like Madeira Island, triggered by intense precipitation episodes exceeding critical thresholds, leading to peak discharges in various catchments.
Although the study has some interesting insights, it appears to be in need of a lot of improvements and restructuring to make it of a quality that would be ready to be considered for NHESS. For example, the description of susceptibility appears to oversimplify the relationships between topography, geology, and flood hazard susceptibility. English writing should be improved on topics like sentence constructions, and the flow/narrative of the story should be improved a lot for some sections like introduction and discussion.
Specific comments:
Suggestion to write title without full stop to make it more attractive.
Abstract
- To increase interpretability and readership of the manuscript please keep to the conventional way of writing an abstract (background, rationale, main message, approach/results, implications). For me now the abstract does not give a proper overview of the manuscript.
- Some sentence are very long and hard to dissect.
- Refrain from using words like “surprising”.
- “Early warning system” is mentioned in the abstract, but not in the manuscript. Why is this mentioned only here? And if so, there should be a clear indication why it is included
Introduction
- Header numbering missing
- Layout of text is different than rest of the manuscript
- What is meant with “Bibliography” line 38?
- Impressive has a positive connotation, so maybe use another word when talking about disasters.
- Instead of reading an introduction which builds on top a problem statement with background information and scientific relevancy of the identified research gap, this introduction reads as a case study area description. Suggestion to rewrite introduction.
- The overall objective should be clearly defined based on the research gap. Now it just states it is expected to bring new scientific contributions. The research objectives are not well identified. What is meant with analyse the data? What type of analysis? How is the relevancy demonstrated of the high-resolution susceptibility map? What is meant with high-resolution? Is local-scale or small-scale more suitable?
Data
- To help the reader understand the manuscript better, a flowchart depicting the methodological steps is needed.
- The technique is commonly referred to as Annual Maxima
- As correctly stated in line 83-83, for impact analysis annual maxima might not be relevant as more than one extreme value per year can be of importance. Why is not a peak over threshold method selected here?
- Please use subheadings for sections throughout the manuscript to help the reader understand the aim of the specific section.
- Line 91, what is meant by maximum precipitation variable in 24 hours? What is the unit of this variable?
- What would be the sensitivity of using k=0.8 to the findings of the study?
Rainfall
- Please also show R2 values for Figure 6
- Where there checks done for homoscedasticity of the data?
- The plots are difficult to interpret and compare when the x-axis scale changes each subplot in Figure 6
Discharge
- How does the occurrence of secondary peaks of intense rainfall correlate with the complex flood patterns observed in the studied catchments?
- How reliable and valid are the peak flood flow estimations obtained from the HEC-HMS model?
Flood risk
- Line 436: Susceptibility to?
- I find the susceptibility map to be oversimplified as it also seems to be just an indication of previous events. There is limited quantitative analysis or data-driven assessment of susceptibility factors such as slope gradient, geology, rainfall intensity, or land cover. Without quantitative analysis, the susceptibility map may not accurately represent the true spatial distribution and intensity of flood hazards.
- The text indicates that the susceptibility map was created based on field survey work and manual identification of areas susceptible to torrential floods, slope movements, and debris flows. The term "cartographic sketch" suggests a less formal and more subjective approach to mapping susceptibility.
Discussion
- The discussion appears to be a summation of what was found in the study. There is no discussion on limitations or implications of this study. It would help if this study has a clearly defined research question to answer, which was also lacking in the introduction.
Citation: https://doi.org/10.5194/egusphere-2024-513-RC2 -
AC2: 'Reply on RC2', Sergio Lopes, 16 Apr 2024
Dear Referee,
First of all, thank you for your work and all the care you took to read our article carefully. Your specific comments will certainly help to improve the quality of the article. In this sense, despite the time constraints, I will do my best to try to make the appropriate changes to improve the quality of the final text.
Regarding the suggested application of peak over threshold method, I find it inappropriate because the adopted rule is to select only known heavy rainfall events that resulted in torrential floods. However, in specific hydrological years, there have been instances where at least two floods occurred on different dates within the same hydrological year. In such cases, both events were included in the analysis.
Line 91, what is meant by maximum precipitation variable in 24 hours? What is the unit of this variable?
The maximum rainfall in 24 hours corresponds to the highest amount of rainfall that occurs within the 24-hour moving interval, calculated based on the 10-minute readings from the udographs on the day(s) of the rain event. The unit of measurement is always in mm.
What would be the sensitivity of using k=0.8 to the findings of the study?
In theory, with the empirical parameter K = 0.9, we assume that 90% of the rainwater that occurred in a certain period prior to the day of interest (even before applying the exponent) remains accumulated in the soils of the catchment-river system. This evaluation considers rainwater from previous days and weeks, which can contribute to soil saturation, instability, and soil movement. The remaining 10% of that water has already been drained into the water network, percolated, or evaporated
By applying K = 0.8, we assume that on steeper slopes (such as narrow-bottomed valleys, as found on Madeira Island), a higher percentage of rainwater (in this case, 20%) quickly runs off the slopes into watercourses or is lost through evaporation or percolation. In other words, we aim to determine lower thresholds of antecedent precipitation calibrated for the occurrence of torrential floods.
Rainfall
After submitting the article, I realized that there was an error in the application of the formula for calibrated antecedent precipitation calculations. Consequently, all the calculations had to be revised. This process has now been completed, and the arrangement of data pairs in the graphs has been reorganized. With this new arrangement, it no longer makes sense to apply the rules for the equations of the regression lines. Instead, for the different durations shown in Figure 6, new thresholds for maximum precipitation in 24 hours and calibrated antecedent precipitation were identified, using only absolute specific values of the two variables. These updated values will be presented in the corrected version of the article.
It makes sense to standardize the x-axis scale, so thank you for the reminder.
Discharge
How does the occurrence of secondary peaks of intense rainfall correlate with the complex flood patterns observed in the studied catchments?
As a general rule, the occurrence of secondary peaks of intense precipitation contributes to prolonging flood flows for longer periods (hours), and thus increase the risk of catastrophic floods. In other words, these floods are complex because they result from several intense downpours, which tend to increase the instability of the land in the upper sectors of the basins.
How reliable and valid are the peak flood flow estimations obtained from the HEC-HMS model?
We did not validate these flood peak flow values estimated through modeling. However, for the purposes of comparative analysis between catchments, we considered that the values obtained reflect an order of magnitude sufficient to perceive the spatial variations of floods. Nevertheless, comparing the values estimated by modeling with those obtained by applying the Manning-Strickler Formula reveals that the former are underestimated compared to the latter
Flood risk
I agree with you. Figure 6 is not precisely a map of susceptibility to torrential flooding. Instead, its purpose was to create a schematic figure illustrating the areas affected by the flood that occurred on December 25, 2020. The intention was to demonstrate that in mountainous regions, even in very small catchments (1 km²), floods can pose a significant risk. In fact, this topic may need to be revisited.
I really appreciate your contribution to improving the work I've been doing on this subject.
Best regards
Citation: https://doi.org/10.5194/egusphere-2024-513-AC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 191 | 38 | 19 | 248 | 10 | 9 |
- HTML: 191
- PDF: 38
- XML: 19
- Total: 248
- BibTeX: 10
- EndNote: 9
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1