Forensic Hydrology: A Complete Reconstruction of an Extreme Flood Event in Data-Scarce Area

Round 1

Reviewer 1 Report

Dear Editor.

I have finished my review on the proposed paper “Forensic hydrology: a complete reconstruction of an extreme flood event in data-scarce area” hydrology-1709996-peer-review-v1.

 

Summary of the manuscript:

In the proposed paper, the authors’ goal is to reconstruct the extreme flood events that took place in Central Greece on 18 September 2021, which had devastated effect in the area. They used satellite and ground rainfall products to construct the flood hydrograph and social network data to validate the flood extent. According to the results the social network data could be very useful to validate the flood extent of a flood event.

 

General review:

1. Generally, the manuscript presents an interesting topic and the specific research seems to include some significant points for the research community of this field.
2. The proposed paper is very well written with very good use of English language. Except some very minor grammatical mistakes and word errors, this paper is written with a very good scientific style. However, the authors used too much: “We make….”, “We did… “, “We used…”. In my opinion, passive voice should be used in scientific papers. The authors should check again the paper to correct these minor mistakes.
3. The proposed paper is very well structured. It begins with the Introduction with some limited references that helps the reader to get into the subject. In Introduction there is an effort to provide previous studies with similar scientific content, which took place in the research area and in other countries. Authors describe and set very well the scientific problem. At the end of Introduction, authors clearly state the goals of the research. However, the authors didn’t manage to provide adequate number of relative studies that recently published in Greece, which deal with the same subject (simulation of extreme flood events) and used almost the same methodology. Below, I gave some examples.
4. The methodology is generally very interesting, and well explained, so other researchers could easily repeat it. However, some parts need to be clearer. See below specific comments.
5. The results scientifically explained and are OK.
6. The quality of the work in Discussion is not adequate. See below specific comments.
7. Conclusions are appropriate for this paper.

 

Additional points for revision:

In my opinion, the proposed paper could be characterized as a good research work, complies with aims of Hydrology. 

First and most significant is that IANOS catastrophe was took place in 18-19 September of 2020.  Not at 2021!!! In abstract and in the text (line 102) you have wrong dates.

INTRODUCTION: This part of the study is weak. More literature should be added, especially from Greece (the last 4-5 years), from previous studies that deal with extreme flood events, using satellite products, field observations, hydrological and hydraulic simulations with HEC-HMS and HEC-RAS. There are at least 4-5 extreme flood events in Greece (Crete, Mandra, Olympiada-Chalkidiki, Thasos) that were studied with almost the same methodology and you didn’t refer in your paper. You used only 36 references. For the specific subject (hydrological-hydraulic modeling) less than 60 references is not adequate. You should find these studies, add them in Introduction and define what is new/novel in your research.

Lines 34-36: This statement needs support with literature of previous studies. Also, in Greece and Mediterranean region, small ephemeral streams that passing through settlements, causing the majority of flash flood events. Not only the river overflow.

Lines 44-46: I agree with this statement (Koutsoyiannis et al. 2012). However, in Greece and Mediterranean region the majority of flood events are caused by ephemeral streams, which cross-section dimensions are significantly reduced by anthropogenic activities (land use change, urban sprawl etc.) Please, rephrase this part.  

Lines 61-62: From Greece, you found only 3 studies from the same author? Please, search the literature again for the last 4-5 years.

Lines 81-84: I agree, this is probably the first hydrological–hydrodynamic analysis implemented at the greater area of Karditsa city. However, almost the same methodology was used in previous studies in Greece. The novel part of this study is mainly the use of social media.

Lines 91-92: Here, you give the area of the watersheds with the party beyond (downstream) Karditsa city. However, I see that you applied the hydrological analysis in smaller watersheds upstream Karditsa city. The area downstream the Karditsa didn’t contribute to flood hydrograph. So, I think that you should correct the area of the watershed in the text, using the area that you used as an input in the hydrological model.

FIGURE1: I think the frame with the text “Plastiras reservoir” is in wrong place. The arrow point outside Greece and the frame (a) is not visible.

Lines 113-142: REMOTE SENSING PRECIPITATION DATA: There are previous studies that used satellite products for event-based hydrological modelling of extreme floods, that you didn’t reference in your study (Sapountzis et al. 2021, Gilewski and Nawalany 2018, Varlas et al. 2017).

Line 153-154: This is a serious drawback of your study. As you geolocated the points of the social media footage, you should organized field research to find these locations and take measurements of the flood depth.

Lines 163-165: I agree with you. Globally well-known software. And it is widely used in Greece. Please, add references from Greece of similar watershed conditions. Enrich you study.

Line 177-178: Why you used the Snyder unit hydrograph? Why not SCS-CN? Support your choice with literature.

Lines 179-181: You calculated the lag time through the time of concentration (Giandotti). Did you validate the lag time? The time of the centroid of excess rainfall and the time of peak discharge were known from the satellite and social media data.

Lines 182-183: The average soil-moisture condition is not defined only by the season, but also from the 5 five previous days (if there is rainfall). Further, in the abstract and in the Conclusions, you say something about the CN (Curve Number?). But in the text, you didn’t say anything about the CN values. How you calculated CN? Which are the values? Also, throughout the seasons the AMC (Antecedent Moisture Conditions) is changing. The runoff generation is significantly different during the wet period when the soil moisture is higher, in comparison to the drought season when the soil moisture in lower. It is known, for example, that forests present finite capabilities to retain large amounts of precipitation, especially during extreme rainfall events, even if the forest cover percentage is significantly high. For that reason, when CN is estimated the AMC is taken into account according the following table (Chow et al. 1988):

Classification of antecedent moisture condition classes (AMC) for the SCS method of rainfall abstractions (source: Chow et al. 1988; table 5.5.1, p. 149).

AMC group	Total 5-day antecedent rainfall (mm)
	Dormant season	Growing season
I	Less than 13	Less than 35
II	13 to 28	35 to 53
III	Over 28	Over 53

 

So, how you calculated the CN and which are the calculated values?

Line 185: You say that you used Manning’s n coefficient equal to 0.040. This roughness coefficient was the same for all the flooded area? Is the roughness coefficient the same for urban areas, crop lands, stream bed and stream banks? Also, it is widely known that the roughness values are changing during the flood event (Ji-Sung, Chan-Joo, Won, & Yong-Jeon, 2010). For 1D modeling this is not a problem. But, for 2D modeling (as in your study) this is a serious problem. If you used one value of roughness coefficient, your results are not trustworthy.

Line 191: “It is a well-known software…”. Add literature applying the HEC-RAS. Mainly from Greece.

Lines 199-200: “similar projects”…. Add literature.

Lines 208-211: See the previous comment about the roughness coefficient (line 185).

Lines 212-223: The DEM that you have used has another serious problem, when it is used for hydraulic simulations. The first is the lateral and vertical structures (bridges, culverts, walls etc.) In the text you say how you manage this problem. However, the second problem with this DEM is dense riparian vegetation. The DEM incorporates the height of the dense riparian vegetation and significantly influence the true heights of stream bed and banks. Did you take into account this distortion of the DEM?

Lines 282-286: Again, I propose to see, discuss and add in the text the following studies that used satellite products (Sapountzis et al. 2021, Gilewski and Nawalany 2018, Varlas et al. 2017).

Lines 299-300: Here, you should calculate the specific discharge (m³/s*km²) of the flood event, using the upstream watershed area, no the whole study area that you show in figure 1. Then, compare your results with previous studies in Greece and Mediterranean (for example Gaume et al. 2009, and other studies in Greece).

Lines 340-370: It is serious disadvantage of the study, the absence of post field measurements of flood depth and flood extent.

Lines 377-382: I have already provided you 3 studies using satellite data for extreme flood analysis (Sapountzis et al. 2021, Gilewski and Nawalany 2018, Varlas et al. 2017).

Lines 407-411: If you had field data for each of social network images it would be perfect. You should underline this.

 

References

Chow V.T., Maidment D.R. and Mays L.W. (1988), Applied Hydrology, McGraw‐Hill: New York, NY, USA, 1988, p. 572, ISBN 0 07‐010810‐2.

Ji-Sung, K., Chan-Joo, L., Won, K., & Yong-Jeon, K. (2010). Roughness coefficient and its uncertainty in gravel-bed river. Water Science and Engineering, 3(2), 217–232. https://doi.org/10.3882/j.issn.1674-2370.2010.02.010.

Gaume, E., Bain, V., Bernardara, P., Newinger, O., Barbuc, M., Bateman, A.,… Viglione, A. (2009). A compilation of data on European flash floods.Journal of Hydrology, 367(1), 70–78. https://doi.org/10.1016/j.jhydrol.2008.12.028

Gilewski P. and Nawalany M. (2018). Inter-comparison of rain-gauge, radar, and satellite (IMERG GPM) precipitation estimates performance for rainfall-runoff modeling in a mountainous catchment in Poland. Water, 10, 1665. https://doi.org/10.3390/w10111665.

Sapountzis M., Kastridis A., Kazamias A-P., Karagiannidis A., Nikopoulos P., and Lagouvardos K. (2021), Utilization and uncertainties of satellite precipitation data in flash flood hydrological analysis in ungauged watersheds, Global NEST Journal, 23(3), 388-399. doi.org/10.30955/gnj.003905

Varlas G., Anagnostou M.N., Spyrou C., Papadopoulos A., Kalogiros J., Mentzafou A., Michaelides S., Baltas E., Karymbalis E. and Katsafados P. (2019). A Multi-Platform Hydrometeorological Analysis of the Flash Flood Event of 15 November 2017 in Attica, Greece. Remote Sensing, 11, 45. https://doi.org/10.3390/rs11010045.

Author Response

Dear reviewer, we thank for your comments. Please find attached our detailed responses.

Author Response File: Author Response.docx

Reviewer 2 Report

The manuscript is sufficiently well written and presents in a clear manner the phases of hydrological modelling for a flood event in the Karditsa prefecture (Greece). In general, it lacks of novel aspects, even if authors underline the importance and novelty of data collection useful for flooding map validation by social media. Apart from the contribution from social media, authors present a “habitual” hydrological modelling procedure, from watershed components analysis, to the development of models by HEC-HMS and HEC-RAS. Nevertheless, it represents a valuable contribution to the study of flooding due to very intense events, and the used methodology can be transposed to other study areas. For this reason, I recommend minor revision following the comments below.

Figure 5: It surprises me that satellite products “Early”,”Late“ and “Final” differ so much each other and highly underestimate the spatial distribution of rainfall, despite the post processing based on ground observation as mentioned in line 251. It would be interesting to have more information about how these maps were built: for example, which are the climate change adjustment factors responsible to the great difference between the Early – Late and Final versions? I think this could bring implications also in how return periods of peak discharges and discharge volumes are calculated. And, in general, a discussion would be interesting addressing why the satellite maps differ so much each other.

 

Line 195: the urban and peri-urban area of Karditsa city is quite complex,

What do authors mean for “ quite complex”? Please explain

 

Line 353: As expected, the simulated flood extent is greater than all the three observed extents

Can authors explain better why they expected the overestimation of flooded areas extent?

 

Lines 289-290: In the light of the above analysis the gauge rainfall records demonstrate a higher rainfall accuracy

The gauge rainfall records provide the observed field data. Thus, I wouldn’t say it demonstrate higher rainfall accuracy, but rather it is the basic information which can be used to validate other sources of data (e.g., satellite images).

 

 

The English form need some revision and improvement. Some examples to the sentences which should be rephrased are below:

Line 237: The statistical analysis shown only for indicative purposed since the 12-years’ time record  is insufficient for achieving high reliability.

Line 331: The most severe overflows impacted the city has been observed

Lines 396-397: we cannot excluded

Line 371: please remove the text in greek fonts.

Author Response

Dear reviewer, we thank for your comments. Please find attached our detailed responses.

Author Response File: Author Response.docx

Reviewer 3 Report

1. The authors have mentioned the importance and advantage of using remote sensing data for flood modeling in the conclusion part. However, in the results, it has been said that all the remote sensing products failed to provide accurate rainfall records compared to the gauged rainfall records. That implies the remote sensing products chosen for the study were unsuitable. Why haven’t the authors tried using some other products after the selected products failed to give satisfactory results?
2. Authors have mentioned that the hydrological model selection for each process should be based on the catchment characteristics. What characteristics have been considered other than the terrain type, i.e., low-lying and mountainous areas mentioned? 
3. The authors have assumed that the pluvial flood component is not crucial compared to the fluvial component based on the fact that the town’s sewer system was submerged during and after the flood event. Is this reason enough not to consider the pluvial flood component?
4. The authors have considered only three scenarios with a considerable gap to quantify the return period of the IANOS event, i.e., 50 years, 100 years, and 1000 years. Since the Medicane IANOS was a rare and unexpected Mediterranean tropical cyclone which could have been influenced by world climate change. So, how is the definition of an extreme flood, as the paper title mentioned, 50 years? Or 1000 years in this study?
5. Authors have mentioned that the study's novelty is using images from social media. Still, no detailed description has been given of how these images have been helpful. And how about the areas in which the photos were not available? In that case, how can data be manipulated? Also, wouldn't satellite data of the flood period give the same information?
6. Lastly, the figures seem haphazard; it would be better to align and organize them for a better aesthetic.
7.  In Fig. 8, the right graph did not mention the difference between the square gauge station and the round one.

8. Can’t not find the reference to the Giandiotti equation mentioned in this article.

1. In line 307, the “I.D.F” curves should be given the full name as Intensity-Duration-Frequency for it first mention in the article.
2. In line 315, what is the meaning of “10.000”years?
3. The article did not mention how the discharge volume was calculated in the figure 10(c) and (d).
4. In figure 11, the “deapth” should be “depth”.
5. What is the meaning in Line 317?
6. There is no value comparison between simulated and real data with the actual number, only a graph to show it is satisfactory. I don’t think it is a good way to give a conclusion. The approach of the article has its novelty, but it still can't not correctly quantify the validation part.
7. What is the meaning of hm3 in Fig. 9? And, there is no need for point J7 in the Fig. 9, because it is not analyzed and discussed.

Comments for author File: Comments.pdf

Author Response

Dear reviewer, we thank for your comments. Please find attached our detailed responses.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Dear authors.

Thank you very much for your time and to provide detailed responses to mu comments.

1. The number of references in a paper reflect the quality of the "state-of-the-art" and the Discussion. The yearly number of studies in hydrological and hydraulic simulation is tremendous. And some of them come from Greece. If you believe that you found adequate number of researches, that give you the confidence that your paper fills a gap of scientific knowledge, I can not insist to add more references. However, I believe that papers with this subject should provide an in depth literature review.
2.  I am fine with the explanation that you gave about the validation of the time of concentration.
3. Regarding the roughness coefficient I am little confused. I do not have any problem with the values that you chose. I wonder if you took into account the constant changing of roughness coefficient values during the flood event, because you applied 2D modeling. During the flood event the roughness coefficients are changing every moment. During the flood event, within the stream (for example) the riparian vegetation (trees, bushes) could be uprooted, the hydraulic characteristics are changing due to sedimentation and bank/stream bed erosion. This is what I asked. Did you take into account the changes of the roughness coefficient values during the flood event?
4. "From our point of view there is no need to provide specific discharge (m3/s*km2) of the flood event...". The specific discharge is just a simple division of max discharge and watershed area. I do not think that was a big deal to be calculated. It provides the opportunity to make comparisons with other similar watersheds in Mediterranean. 

 

Author Response

Dear Reviewer, please find enclosed our responses for your consideration.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments:

1. For the first comment and answer, what I am failing to understand here is by saying remote data didn’t give satisfactory results, are the authors trying to say that using the images taken during the event is a better way to gather data? If so, then that is a very situational remark.
2. For the second comment and answer, I did understand that the catchment delineation has been considered; however, can you further explain the second half of the first sentence (“aiming to inform the hydraulic model with the sound inflows.”.
3. For the third comment and answer, I would still like to know, in your opinion, would considering the pluvial flood component have affected the study results or not?
4. I understood that from figure 3; however, if you look at your revised manuscript, you will see some figures are missing the number, and some have an incomplete caption.
5. This paper lacks the use of proper punctuation.
6. Check line 49: It should be urban sprawl, not spraw.
7. Check lines 51-52 (grammatical mistake): If you are using the plural of the source, i.e., sources, use them; otherwise it.
8. Check line 63 (grammatical mistake): “and debris flow.”
9. Crowdsourced is a one-word either use it with a hyphen or without space.
10. There is an article usage problem in your paper when you mention rivers etc.; please use proper articles.
11. Please work on the grammar of this paper, especially punctuations and articles.
12. Check line 150: Zekkos et al. (2022, see Figure 2c)
13. Check line 181: Medias word does not exist in the dictionary. Media is a collective noun (used both for singular and plural)
14. Check 220: It's the U.S., not US Army Corps of Engineers.

Author Response

Dear Reviewer, please find enclosed our responses for your consideration.

Author Response File: Author Response.docx