Disastrous Flash Floods Triggered by Moderate to Minor Rainfall Events. Recent Cases in Coastal Benguela (Angola)

Round 1

Reviewer 1 Report

Dear authors,

I carefully went through your manuscript and I found the topic very interesting since the data about flash flood records from developing countries are really scarce, but valuable. However, I feel that the current version of your manuscript is like a “technical note” rather than the research article. You stressed the importance of relatively low rainfall that possibly triggered the flash floods and the socio-economic conditions that are probably responsible for such damages which could otherwise not be such great in more developed regions. These factors are of course very interesting to be studied, but the overall presentation in your paper lacks clear structure of the research paper. Therefore, I think that the paper should undergo major revision, particularly focusing on clearer description of methods and your results and mention some uncertainties related to precipitation records from radar. I also recommend to check the English as some parts are hard to understand. Detail comments and suggestions can be found in the attached pdf.  

Comments for author File: Comments.pdf

Author Response

REPLY TO REVIEWERS COMMENTS

The authors truly appreciate the reviews provided by two colleagues, which helped to improve the manuscript. Below, in blue, we present explanations, point by point, on how we dealt with each comment/suggestion.

Reviewer #1

I carefully went through your manuscript and I found the topic very interesting since the data about flash flood records from developing countries are really scarce, but valuable. However, I feel that the current version of your manuscript is like a “technical note” rather than the research article. You stressed the importance of relatively low rainfall that possibly triggered the flash floods and the socio-economic conditions that are probably responsible for such damages which could otherwise not be such great in more developed regions. These factors are of course very interesting to be studied, but the overall presentation in your paper lacks clear structure of the research paper. Therefore, I think that the paper should undergo major revision, particularly focusing on clearer description of methods and your results and mention some uncertainties related to precipitation records from radar. I also recommend to check the English as some parts are hard to understand. Detail comments and suggestions can be found in the attached pdf.  

Reply: Thank you for the overall assessment. The structure of the paper was reorganised. The methods were described in more detail, as asked by the 3 Reviewers, and we believe they are also clearer. A reference to the rainfall uncertainties are also mentioned now. English was double-checked following the comments found in the annotated manuscript and by an independent native English.

Detailed comments from the annotated manuscript

1 (title). I propose to change to: Disastrous flash floods triggered by moderate to minor rainfall (coastal Benguela, Angola)

Reply: We can agree. Title was changed

2 (abstract 1). What does it mean: “frequency of relatively steep hills”? Do you mean “abundance”? and what does it mean “fine-grained geological units”? Do you mean “fine-grained sediments”? It would be good to explain in more detail through the text further

Reply: The expressions were modified. As the abstract is near the limit defined by the Journal, these geomorphological features are only further explained in other parts of the manuscript.

3 (abstract 2): However, socio-economic conditions are most likely the main reasons of flood damages.

Reply: Changed as suggested

4. (introduction): I miss more detailed info including existing knowledge about the flood records in developing countries (like Angola) and methods how they are recognized. I imagine one or two paragraphs that describe state-of-art of this topic.

Reply: We felt that a thorough review of the state of the art will be excessive for the present article. But a new paragraph, with a few examples of works performed in southern Africa and mentioning Angolan reality, was added in the introduction.

5 (introduction). This paragrpah is related to particular event and study site so I prefer moving to study site section.

Reply: We prefer to leave it in the introduction. It shows why the work was performed. However, it was modified and merged with the last paragraph of the introduction.

6 (introduction). I miss some deeper hypotheses: what did you expect? Did you know that flash floods were caused by limited rainfall? If yes, then I would expect that the state-of-art in the introduction will tell us more about the effects of moderate rainfall to flash flood hazard worldwide. If the analysis of meteorological conditions that preceded the flash floods and that trigger flash floods are the main results of your research, you have to more focus on climate factors in your introduction.

Reply: Thank you for this constructive comment. The final sentences of the introduction were modified to better present hypothesis and research aims

7 (Regional setting). "Geomorphic conditions" is better (geomorphology is the branch of science)

Reply: Changed in a way that accommodates the recommendations of reviewers1 and 2

8 (Regional setting). I am not sure, if this paragraph is so important in the context of the research.

Reply: Catumbela river and delta play a fundamental role for coastal geomorphology, but we agree that they may not deserve such a long description. The paragraph was shortened and merged with the previous one.

9. I miss more detailed description of the "channels" which seems to be like "intermittent gullies" (with occasional flows). What is their geometry, sediment grain-size, are there any differences in studied sections? Are the gullies normally dry, or periodically flown? Please, give the readers more info for better imagination.

Reply: A description of the valley net was added.

10 (Regional setting). Climate conditions

Reply: Changed

11 (methods). Please, provide more info about TRMM, especially the resolution of grid data which is quite important when analysing flash floods. Are there any uncertainties regarding the (un)recording of localized convective precipitation?

Reply: The methods section was enlarged, including the resolution of the dataset. The possibility of unrecorded convective rain was added in the discussion.

12. Chapter 4 should be named "Results"

Reply: We confess we prefer the previous title, but it was changed as recommended

13 (results). I propose: "Flash flood in March 2015 (Lobito)

Reply: Title was changed

14 (results). It would be also nice to express the percentile of the triggering daily precipitation per year (i.e., if the 12 mm rainfall is the "normal" or "extraordinary" for the study site. I would pretend to have a look on the data either from weather station or to TRMM and to calculate this percentile (if such precipitation event belongs to the extreme ones or not).

Reply: A paragraph discussing the more or less extraordinary precipitation associated with the flash floods was added in the discussion. The TRMM record is available for download and we consider unnecessary to display it in the Journal, but it is provided with this revision.

15 (Fig. 3). Did you also use the data from weather station? It should be mentioned in the methods and also add some detailed info about the station and measurements.

Reply: We did our best to obtain updated rain data with relevance for the studied events in all possible national and regional Angolan institutions. The results available are either from sites too far from the affected region (>200 km) or not reliable. For example, in Catumbela airport no data was collected in the day of the floods of 2015. Remote sensing results are thus most useful in these less developed regions.

16 (results): If you record water level marks, it should be also noted in the methods. When (how many days/months after floods?) were these marks recorded? Which type of flood marks were considered?

Reply: Measurements of water levels are now more clearly indicated in the methods. As explained, measurements were performed during the month after each event and considered references signed by local population.

17 (results). Flash flood in March 2019 (Catumbela)

Reply: Title was changed

18 (discussion). I would reformulate it - it sounds strange

Reply: Title was changed to “Natural features promoting flash floods”

19 (discussion): Please, be more specific: which particular "indurated layers" - what kind of rocks? and what are "more friable units" - sediments?

Reply: Lithologies with distinct mechanical strength were identified

20 (discussion). In this section, I miss the mention about the intensity of rainfall (that can be the main factor of triggering the flash floods) and the uncertainties regarding the data obtaining from TRMM - can you really record even spatially limited short-term rainfall events?

Reply: A paragraph focused on these issues were added in the discussion.

21 (discussion). What about "Socio-economic conditions as the important preparatory factors of flash flood damages"

Reply: Title was changed.

22 (discussion): Please, can you illustrate the river flow? With some arrows or transparent lines? It would be more telling.

Reply: Arrows were added to make flow direction more visible

23 (conclusions): This sentence is not meaningful. Actually, such rain was sufficient to trigger flash floods. The word "significant" is not proper.

Reply: Phrase was changed

Other minor typing errors identified by the reviewer were corrected

Reviewer 2 Report

Thank you for the opportunity to review an interesting paper. In my opinion, the manuscript requires major corrections. I kindly apologize for the quality of the English in my text, I hope everything I wrote is understandable.

The article deals with an important problem of flash floods in small catchments. The analyzed case shows that even small rainfall under specific conditions can cause large economic and social losses and fatalities.

In my opinion, the article is interesting but requires corrections and additions.

1. Introduction

The article deals with natural and anthropogenic parameters  influencing the course of the flood. The introduction lacks information on natural factors (geological, geomorphological or hydrological). There is a wealth of literature on this subject and should be cited.

2.1. Geomorphology

Chapter 2.1 presents information on geology (lithology, sedimentology) and geomorphology in a chaotic manner. I propose to organize the information and name the chapter Geology and geomorphology.

3. Results

The entire text of this chapter contains methods (unfortunately not all). The results are only presented in the table that should be discussed.

The methods presented in the article are very interesting and it is worth presenting them in the Metods chapter. Remote sensing methods should also be described and information on the imagery (aerial orthophotomap? VHR satellite imagery?) used should be provided.

Detailed comments:

Line 139: „DEM; generated with an accuracy of 10 m” - not with accuracy but with spatial resolution

Figs. 2A, 4A-C, 5: What images are shown in the drawings? Aerial imagery, satellite imagery, orthophotomaps?

Author Response

REPLY TO REVIEWERS COMMENTS

The authors truly appreciate the reviews provided by two colleagues, which helped to improve the manuscript. Below, in blue, we present explanations, point by point, on how we dealt with each comment/suggestion.

Reviewer #2

Thank you for the opportunity to review an interesting paper. In my opinion, the manuscript requires major corrections. I kindly apologize for the quality of the English in my text, I hope everything I wrote is understandable.

The article deals with an important problem of flash floods in small catchments. The analyzed case shows that even small rainfall under specific conditions can cause large economic and social losses and fatalities.

In my opinion, the article is interesting but requires corrections and additions.

1. Introduction

The article deals with natural and anthropogenic parameters  influencing the course of the flood. The introduction lacks information on natural factors (geological, geomorphological or hydrological). There is a wealth of literature on this subject and should be cited.

Reply: Introduction was reorganized and sentences focusing the natural factors that influence flood dynamics were added.

2.1. Geomorphology

Chapter 2.1 presents information on geology (lithology, sedimentology) and geomorphology in a chaotic manner. I propose to organize the information and name the chapter Geology and geomorphology.

Reply: Following the comment, we tried a better organization of the geological/geomorphological setting. It comprises now 3 paragraphs dedicated to (1) the main geological units observed in the coastal region, (2) overall morphology and its relation with outcropping units and (3) valley configuration.

3. Results

The entire text of this chapter contains methods (unfortunately not all). The results are only presented in the table that should be discussed.

The methods presented in the article are very interesting and it is worth presenting them in the Metods chapter. Remote sensing methods should also be described and information on the imagery (aerial orthophotomap? VHR satellite imagery?) used should be provided.

Reply: Thank you for alerting for this flaw that skipped us. The title was changed and the text substantially extended.

Detailed comments:

Line 139: „DEM; generated with an accuracy of 10 m” - not with accuracy but with spatial resolution

Reply: Changed as requested

Figs. 2A, 4A-C, 5: What images are shown in the drawings? Aerial imagery, satellite imagery, orthophotomaps?

Reply: Source information was added in the methods and in the captions of figures that showed aerial imagery.

Reviewer 3 Report

Dear Authors,

this study reports data about two coastal floods that are important climate events we must forecast. Therefore, as scientists, we should always give importance to this study both for their value themselves and for their societal significance. Besides its value, this paper highlights some shortcomings and scientific issues that I report here:

• INTRODUCTION: there is a lack of information about these flash floods, literature references and the main challenges (scientific, climatic, social aspects) related to these events are completely lost in this introduction. Therefore, I can estimate if this study only reports two events description or if authors have improved the literature about this topic. Please, also include the main research questions and the study aim.
• Keywords: I recommend using keywords not included in the manuscript title
• 2. Climate: add more information about the temperature data
• METHODS: this section is completely lost… information reported in the first part of Results are not sufficient to present in clear way materials and methods used in a complicated study like this. This section and Introduction are the most important in a scientific paper and they must be written allowing all information to readers (that probably deal with the same events and could replicate your methods)

Author Response

Reviewer #3

this study reports data about two coastal floods that are important climate events we must forecast. Therefore, as scientists, we should always give importance to this study both for their value themselves and for their societal significance. Besides its value, this paper highlights some shortcomings and scientific issues that I report here:

INTRODUCTION: there is a lack of information about these flash floods, literature references and the main challenges (scientific, climatic, social aspects) related to these events are completely lost in this introduction. Therefore, I can estimate if this study only reports two events description or if authors have improved the literature about this topic. Please, also include the main research questions and the study aim.

Reply: The introduction was entirely reformulated. A few more references concerning the main challenges were included. Taking into consideration that we do not intend to present an exhaustive review of the topic, in our understanding, the present number of references in the Introduction is quite reasonable (30). The final sentences were modified to show in a clearer way the main goals of the research.

Keywords: I recommend using keywords not included in the manuscript title

Reply: Key words were adapted to avoid replicating terms from the title

2. Climate: add more information about the temperature data

Reply: Although temperature is not very important for the studied topic, we agree that some information should be presented. It was added.

METHODS: this section is completely lost… information reported in the first part of Results are not sufficient to present in clear way materials and methods used in a complicated study like this. This section and Introduction are the most important in a scientific paper and they must be written allowing all information to readers (that probably deal with the same events and could replicate your methods)

Reply: We regret that in the first submission the Methods section was wrongly called “Results”. The section was improved, as requested by the 3 reviewers, and properly titled.

Round 2

Reviewer 1 Report

Dear authors,

I appreaciate improving the original version of your ms. Several parts are now clearer, but some parts still need moderate revision. See the notes below. Moreover, even though I am not a native speaker, I feel that the English language should be checked again (e.g., L68-70, L242 - on the other end?, L250 - outstanding? you probably mean exceptional?, L288 Previous works have shown etc.). 

L85-88 - please, be more concise and give us more quantitative description of the valley morphology (width and depth of the streams/gullies). From the description, it appears that you speak about ephemeral gullies (ephemeral streams) - please, consider this term in the ms where applicable and also to keywords.  

L157 - indicate also the resoulution in km units of grid cells (I suppose ca. 25-30 km?)

Fig. 3 - my original question/note was related to the "weathering station" - if I understand well, you mean weather/meteorological station? In the ms you tell us that the weather station data are not available, but in this figure you mention daily averages from "weathering station" in Lobito - please clarify what does it mean? 

L251-252 - so, according to this percentile, it seems that (1) it was an extraordinary rainfall for this region (even though some higher values were recorded as well) and (2) other factors like socio-economic unpreparedness, but also antecedent moisture conditions related to preceding rain and probably the intensity of rain could cause the dramatic flood damages - it should be also discussed and extended  

L252-253 - what does it mean "reached once in 6 out of 16 hydric years"? please clarify by improved English? ...also, hydric is adjective which rather belongs  to regime of soils - you probably meant "hydrological year"? 

L254-255 - This is not sufficient information. If you mention uncertainty in the data, you should provide the readers some references where these uncertainties are described. It is generally known that radar estimates can overestimate the light rain, but underestimate short-term high-intensity rainfall, which is your case. In this respect, it would be  fine to extend it, at least by some references that explain such uncertainty.  

Author Response

I appreaciate improving the original version of your ms. Several parts are now clearer, but some parts still need moderate revision. See the notes below. Moreover, even though I am not a native speaker, I feel that the English language should be checked again (e.g., L68-70, L242 - on the other end?, L250 - outstanding? you probably mean exceptional?, L288 Previous works have shown etc.).

Reply: Thank you for the alerts. Written English was checked once again

L85-88 - please, be more concise and give us more quantitative description of the valley morphology (width and depth of the streams/gullies). From the description, it appears that you speak about ephemeral gullies (ephemeral streams) - please, consider this term in the ms where applicable and also to keywords.

Reply: We tried to give more quantitative information on valley morphology, keeping it in a concise way-

L157 - indicate also the resoulution in km units of grid cells (I suppose ca. 25-30 km?)

Reply: At the investigated latitude it is ~27 km. This figure was added.

Fig. 3 - my original question/note was related to the "weathering station" - if I understand well, you mean weather/meteorological station? In the ms you tell us that the weather station data are not available, but in this figure you mention daily averages from "weathering station" in Lobito - please clarify what does it mean?

Reply: It referred to historical results reported for Lobito. To make sure that similar climatic parameters are being represented, we decided to replace it by the average TRMM data for the available time series.

L251-252 - so, according to this percentile, it seems that (1) it was an extraordinary rainfall for this region (even though some higher values were recorded as well) and (2) other factors like socio-economic unpreparedness, but also antecedent moisture conditions related to preceding rain and probably the intensity of rain could cause the dramatic flood damages - it should be also discussed and extended

Reply: During most of the year there is no precipitation in coastal Benguela and the rain events that occur during the wet season are relatively short. So, almost all 3-hour records indicate no or very minor precipitation. But events with comparable or higher precipitation than those associated with the flash floods of 2015 and 2019 were detected in 37.5 % of the years (6 years for a time series of 16 years). We changed Fig. 3 to show the entire dataset, instead of only the periods of the flash floods. This figure shows better that the rainfall associated with the events of 2015 and 2019 is not unusual. Note that the influence of antecedent is not relevant, as higher rainfall after the event of March 11 had no major consequences.

L252-253 - what does it mean "reached once in 6 out of 16 hydric years"? please clarify by improved English? ...also, hydric is adjective which rather belongs to regime of soils - you probably meant "hydrological year"?

Reply: Yes, is hydrological year, thanks! The phrase was simplified to “reached in 6 out of 16 hydric years”.

L254-255 - This is not sufficient information. If you mention uncertainty in the data, you should provide the readers some references where these uncertainties are described. It is generally known that radar estimates can overestimate the light rain, but underestimate short-term high-intensity rainfall, which is your case. In this respect, it would be fine to extend it, at least by some references that explain such uncertainty.

Reply: We can agree. This topic was extended with 2 more phrases.

Author Response File: Author Response.docx

Reviewer 2 Report

The article has been significantly improved and in my opinion can be published.

Author Response

The article has been significantly improved and in my opinion can be published.

Reply: Thank you for the help to improve the manuscript.

Author Response File: Author Response.docx

Reviewer 3 Report

Introduction: 

I appreciate the effort in recutting this part. Besides the 30 references (useful only if very close to the issues and if they report literature gaps in the field), I always fail to read something about the significance of this study in the scientific context, the differences (improvements) in comparison with other studies (from a scientific point of view: materials, analysis, interdisciplinary aspects, etc.)

Methods: 

This section must further improve. I can not read anything about your activities during the 4 field works to identify and measure flooding effects (when?how?why?how), the questionnaire used (questions? aim? related answer analysis) Google Earth images were used to verify damages due to flood events, but how? did you map and classify these damages? which classification did you use? 

Moreover, results report a mixture of data (meteorological, geomorphological, risk assessment..) without a scientific explanation of the impact assessed. For instance, Fig 4 a-b-c reports flow direction, but what is the information? why flows did follow that routes? what about runoff? there are a lot of studies in this field useful to understand how to conduct a flooding study and which parameters are necessary. Information reported in this paper seems more appropriate for a technical report (there is no much analysis besides GIS mapping and some meteorological data). 

Just for example, I suggest seeing this paper for the method description: 

https://www.researchgate.net/profile/Ying-Wang-356/publication/334628262_Assessment_of_future_flash_flood_inundations_in_coastal_regions_under_climate_change_scenarios-A_case_study_of_Hadahe_River_basin_in_northeastern_China/links/5fccf046299bf188d4f9e6ed/Assessment-of-future-flash-flood-inundations-in-coastal-regions-under-climate-change-scenarios-A-case-study-of-Hadahe-River-basin-in-northeastern-China.pdf

Author Response

Introduction: I appreciate the effort in recutting this part. Besides the 30 references (useful only if very close to the issues and if they report literature gaps in the field), I always fail to read something about the significance of this study in the scientific context, the differences (improvements) in comparison with other studies (from a scientific point of view: materials, analysis, interdisciplinary aspects, etc.)

Reply: Between the submission of the revised version, on April 18, and the arrival the decision, on April 21, there was another flood event in coastal Angola with causalities. This time in the capital, Luanda, with a death toll of 14 and more than 8000 displaced from their homes. Although it wasn’t a flash flood comparable to those presented in the submitted article, it reveals how fragile are the settlements around coastal cities of Angola. Investigating the causes of this vulnerability cannot be performed with high quality in-situ data because they are not available. The present research shows how non-sophisticated data collected in the field (e.g., water levels) coupled with remote sensing data (TRMM rainfall, DEM) provide excellent information on the causes of flash-flood damage. We are expecting that the presentation of this low-cost approach will be considered relevant for Hydrology.

Methods: This section must further improve. I can not read anything about your activities during the 4 field works to identify and measure flooding effects (when?how?why?how), the questionnaire used (questions? aim? related answer analysis) Google Earth images were used to verify damages due to flood events, but how? did you map and classify these damages? which classification did you use?

Moreover, results report a mixture of data (meteorological, geomorphological, risk assessment..) without a scientific explanation of the impact assessed. For instance, Fig 4 a-b-c reports flow direction, but what is the information? why flows did follow that routes? what about runoff? there are a lot of studies in this field useful to understand how to conduct a flooding study and which parameters are necessary. Information reported in this paper seems more appropriate for a technical report (there is no much analysis besides GIS mapping and some meteorological data). Just for example, I suggest seeing this paper for the method description:…

Reply: The methods applied are simple and the authors cannot see how they could be much more detailed. We know well the work recommended by Reviewer 3 as an example (by the way, it inspired us to extend the discussion on the links between climate change and flash flood effects, which was useful to support our perception that local factors are the leading cause of damage in coastal Benguela). But that article is for a different and more complex approach. To ensure that the methods are presented in a more structured way, it was created a new figure with the adopted workflow.

In what regards the enquires to the population, the when (4 weeks after flooding events), why (to understand flooding dynamics) and how (based on water levels during different stages of the events) are more clearly presented now. Unfortunately, only maximum water levels were considered reliable. Google Earth images served to confirm the location of the most affected areas and it was not necessary to apply any classification.

We confess we did not understand the issue on runoff and water routes. Arrows indicate water flow at valley bottoms. Lateral runoff cannot be high with a rainfall of <4 mm/hour and valley margins of just a few hundred meters.

Author Response File: Author Response.docx