the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 07 Aug 2023
Hydrometeorological controls and social response for the 22 October 2019 catastrophic flash flood in Catalonia, north-eastern Spain
Abstract. On 22 October 2019, the Francolí river basin in Catalonia, north-eastern Spain, experienced a heavy precipitation event that resulted in a catastrophic flash flood, causing six fatalities. Firstly, this study investigates the hydrometeorological factors that concurred in the unfolding of this event by using large-scale meteorological grid analyses, high-resolution mesoscale model simulations, radar-derived precipitation estimates, post-flood field observations and hydrological modelling. Mesoscale diagnosis reveals that a persistent south-easterly airflow brought low level moisture and established convective instability over the region, while local orography was instrumental to trigger moist deep convection. A convective train promoted intense, copious and prolonged precipitation over the north-western catchment headwaters. Basin response was significantly modulated by the very dry initial soil moisture conditions. After the long-lasting rainfall, an acute burst of precipitation resulted in extreme flash flooding. Fast and abrupt increases in river flow lead to short times for protection procedures to be implemented effectively. Secondly, this study collects the social response times during the event and compares them with catchment dynamics. The idea is to examine the adequacy of monitoring and warning issuance in meeting the needs of the population at risk. Based on this information, the study provides recommendations aimed at minimizing losses and improving preparedness for similar natural hazards in the future.
- Preprint
(8195 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on nhess-2023-130', Anonymous Referee #1, 17 Dec 2023
General comments
The paper presents a detailed analysis of the meteorological, hydrological, and social dynamics of a flash flood that occurred in Catalonia (Spain) on October 22 2019, resulting in six fatalities. This very well documented interdisciplinary study combines observation and modelling to understand the meteorological and hydrological context that led to this exceptional event. The study also provides insight into the dynamics of the warning process, the social perception of this warning and of the social response to this event.
The study relies on a comprehensive data set and the modeling study (both atmospheric and hydrological modeling) are well documented. However, some results may be analyzed in greater depth (see specific comments below) and the analysis of the social response is rather qualitative.
Furthermore, the paper lacks a “discussion” section that could allow comparison with other case studies and provide information about the elements of the study that are of interest to a wider audience, beyond the single case study presented in the article. The conclusions and abstract should be revised accordingly.
Therefore I recommend major revision of the paper to allow the authors to address all the issues pointed out in this review.
Specific comments
0/ section 3.1. The atmospheric modeling is based on the new version of a 3D non-hydrostatic model, developed by some of the co-authors, which has been tested and shown to perform as well as more widely used model. This model is use to assess the predictability of the event and the main driving factors of the event. As the authors study the warning process in their paper, it would be interesting to also compare the numerical simulation performed in the paper, with the results of the operational model used in 2019 during the warning process, in particular in terms of positioning of the most intense rainfall. It would allow assessing the data that are crucial to properly anticipate this kind of event. If the TRAM model had been used for the forecasting of this event, would it have improved the forecast?
1/ Line 185. The authors mention that they optimized their simulation to get a higher performance in terms of Quantitative Precipitation Estimate (QPE). Which performance criteria were used for that?
2/ line 188: ERA5 reanalysis is used for the boundary conditions of the model. What is the reliability of the reanalysis for the conditions of October 22 2019? Did the authors perform a sensitivity study to uncertainties in the ERA5 reanalysis?
3/ Section 3.2. The authors present how they derived QPE from radar and ground station data. The method relies on radar rainfall field corrected from beam occlusion and attenuation, using a standard rainfall- reflectivity relationship. If I understood correctly, the radar fields are then corrected for possible biases using the automatic ground stations. It seems that QPE used in this study are different from those presented in Martin-Vide et al. (2023) who used co-kriging with external drift of radar and ground stations data. Have the authors compared the two estimates? Have they performed a sensitivity analysis of the hydrological response with respect to uncertainties in QPE estimates?
4/ Figure 3: this figure shows that high precipitation rainfall amounts were also recorded in the catchments located in the west of the Francoli catchment. Why did the authors focus on the Francoli catchment only?
5/ section 3.3. Hydrological modeling: the hydrological model used in the study is quite simple in terms of hydraulic transfer within the river. Martin-Vide et al. (2023) show that the event led to large geomorphological changes in the riverbed. Is this process taken into account in the hydrological modeling? Does the model make hypotheses about the river bed section (width, depth)?
6/ Line 238. Could you define the term helicity?
7/ Line 272. The simulation with an erased topography is particularly interesting, as it perfectly highlights the role of topography in enhancing precipitation amounts.
8/ Line 297. “Cumulative precipitation of 29.8 mm to 39.5 mm ...”. Does this amount refer to the third part of the rainfall event?
9/ Line 310 “the relationship between the temporal and spatial scales” and caption of Figure 8d. You should be more explicit on the variable you are plotting in Figure 8d.
10/Line 333. The low runoff ratio over the entire basin is not so surprising as half of the catchment was almost not affected by the rainfall.
11/ Section 4.3.2 The hydrological model is only evaluated by comparing with the outlet discharge, which is not well representative of the area with high specific discharge. The model evaluation would be more convincing if the model results were also compared with the maximum peak discharge and time to peak estimated by Martin-Vide et al. (2023). I would have expected this comparison in the paper. A discussion about the impact of the strong geomorphological changes and large wood transport mentioned by Martin-Vide et al. (2023) and induced by the flood on the hydrological model would also be welcome.
12/ Line 444. Was the rainfall hyetograph computed for each sub-catchment where peak discharge was estimated?
13/ Figure 11: why only 3 estimates appear in this figure whereas Table 1 provides 7 estimates of peak discharge from the post-event survey? Could the location of the estimates be plotted in Figure 1?
14/ Lines 465-470: the explanation about points of larger areas lying on the Marchi et al. (2010) curve is not very clear.
15/line 517: which part of the 44 millions euros was related to the Francoli catchment?
16: Lines 518-526. It seems that the fatalities in the Francoli catchment are more related to the obstruction of the bridges that led to particular conditions in terms of water height than to the hydrological conditions themselves. Does the model simulation point out to the same locations for critical hydrological conditions?
17/ Conclusions: the paper lacks a discussion section, before the conclusions section. The latter is quite long and should be shortened.
18/ Line 566 “belatedly”??
19/ Data availability. And what about the availability of post-event data, of the modeling results?
20/ Figure 4: The caption should provide indication on which variable is shown in colors and which variable is shown with lines
21/Figure 12. The caption of this figure must be expanded to be understandable. For instance, what do the colored lines at the top of the figure refer to? Explain what are the rainfall centroids.
References
Martín-Vide, J. P., Bateman, A., Berenguer, M. Ferrer-Boix, C., Amengual, A., Campillo, M., Corral, C., Llasat, M. C., Llasat-Botija, M., Gómez, S., Marín-Esteve, B., Prats-Puntí, A., Ruiz-Carulla, R., Sosa-Pérez, R: A flash flood with large woody debris clogged bridges. The 2019 event of Francolí River (NE Iberian Peninsula), J. Hydrol.: Regional Studies, 47, 101348, 2023.
Citation: https://doi.org/10.5194/nhess-2023-130-RC1 -
AC1: 'Reply on RC1', Arnau Amengual, 25 Jan 2024
The comment was uploaded in the form of a supplement: https://nhess.copernicus.org/preprints/nhess-2023-130/nhess-2023-130-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Arnau Amengual, 25 Jan 2024
-
RC2: 'Comment on nhess-2023-130', Anonymous Referee #2, 28 Dec 2023
The manuscript is interesting and contributes to our knowledge about the hydrometeorological factors of flash flood events, but the presentation is poor. The manuscript structure needs a lot of work; there is no section explicitly for results and discussion. These make the manuscript hard to follow. The study contribution and research gap that aims to be addressed are also unclear. I have recommended a few edits and comments in the PDF. Here are additional comments:
- Overall, the writing is OK but some improvements should be addressed.
- The illustrations need major improvements.
- Abstract should be revised to provide information about methods and results. Also, please clarify the unique aspects of this study.
- Introduction: Please explicitly discuss the unique aspects and novelty of this paper.
- Currently introduction contain some information about flash flood in the case study, but the definition of flash flood is missing. In addition, some examples have been mentioned for small watersheds, but the case study is not as small size as these examples. How is the flash flood dynamic in your case study similar to these cases? Are there other types of floods in these areas?
- In some paragraphs of introduction section, several references are presented at the end of a paragraph, but these need to be specifically cited throughout the paragraph.
- The term “social response” is too broad and should be more specific. Do you mean management actions?
- Case study section needs to discuss the watershed characteristics such as climate, annual precipitation, land cover distribution, topography and other factors related to flash floods.
- Please add a schematic view of your methodology as a figure at the beginning of Section 3.
- Sections 4 and 5 should be renamed as results and discussion.
- The control numerical simulation in Section 4.1 should be discussed in detail.
- The models (TRAM, QPEs KLEM etc.) have inconsistent spatial resolutions. How did you handle this inconsistency?
- My understanding is that the automatic gauges record data at sub-daily timescale but the number of these stations are limited, particularly for streamflow. How did you use daily data for a rapid catastrophic flash flood event? What limitations and uncertainties exist here?
- Section 4.3.3: Add a table and show the sensitivity scenarios.
- The initial soil moisture is determined based on the antecedent precipitation, as a standard proxy. Why not using global data like ERA5 and CCI that directly present the soil moisture?
- I suggest using CN as a commonly used abbreviation for curve number.
- Hydrological model calibration needs details and clarifications. Why CNs were kept invariant? Why did you use an initial abstraction ratio of 0.35 (lambda)? The sensitivity analyses should be extended by evaluating other variables like lambda.
- What fit metrics (e.g., NSE and PBIAS) were used and how the model performance was judged based on them?
- Any validation effort on the hydrological model?
- Some of the error values on Table 4 are high (<65%). How would you interpret these and the efficiency of your hydrological model?
- Can your results be generalized to other flash flood events in the study area or flash events beyond the study area? Please discuss.
- Sources of uncertainty and how they can affect your results should be discussed.
- Study limitations and potential areas for future research should be discussed.
- Table 1: What does “hydrometric section” mean? Please clarify the duration of total rainfall.
- Please remove “Color code” column from Table 5.
- Figures 1-3 should be improved by considering the size, alignment etc.
- Section 4.1: Mesoscale processes and role of orography is ambiguous. Please clarify how the simulation works and how Figures 4-5 were produced?
- Figures 4-5 can be merged.
- Figure 8d: What is the main massage of temporal relationship between drainage area and precipitation? Why is the expectation that these two should have a relationship?
- Figure 11 is odd. Why do you have “estimated” uncertainties only on a few data points? This should be for all simulated values. Why do we have only one “observed” value? How can lag time be even observed?
- Please summarize the key findings of your study (e.g., as bullets) in the Conclusions section.
- Please italicize all parameters and coefficients throughout the text.
- Please spell out all abbreviations in the figures, tables and headings; these need to stand alone.
I hope the authors find these comments useful in their research. If the authors decide to submit a revision, both sets of my comments, including the above and in the PDF, have to be addressed.
-
AC2: 'Reply on RC2', Arnau Amengual, 25 Jan 2024
The comment was uploaded in the form of a supplement: https://nhess.copernicus.org/preprints/nhess-2023-130/nhess-2023-130-AC2-supplement.pdf
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 424 | 113 | 29 | 566 | 22 | 21 |
- HTML: 424
- PDF: 113
- XML: 29
- Total: 566
- BibTeX: 22
- EndNote: 21
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1