Development of Flood Early Warning Frameworks for Small Streams in Korea

Round 1

Reviewer 1 Report

I enjoyed reviewing your paper about the development of FEWFs for small streams in Korea. You all did a very good job in explaining why this work is important, how you went about characterizing the test streams, how you developed the FEWF, and how you evaluated it. The Conclusions followed the study as laid out in this manuscript very well. I applaud your efforts in developing not only interesting science, but also an important tool that could have a very practical application. 

I only had a few non-substantive editorial suggestions and you can find those in the attached.

Again, this is a well executed study and a very well-prepared manuscript.

 

Comments for author File: Comments.pdf

Author Response

Response to Reviewer 1 Comments

Thank you very much for the invaluable suggestions.

Point: I enjoyed reviewing your paper about the development of FEWFs for small streams in Korea. You all did a very good job in explaining why this work is important, how you went about characterizing the test streams, how you developed the FEWF, and how you evaluated it. The Conclusions followed the study as laid out in this manuscript very well. I applaud your efforts in developing not only interesting science, but also an important tool that could have a very practical application.

I only had a few non-substantive editorial suggestions and you can find those in the attached.

Again, this is a well executed study and a very well-prepared manuscript.

 Response 1: I revised the article to reflect the parts sent by the reviewer. Thank you very much again.

Author Response File: Author Response.pdf

Reviewer 2 Report

- The author discussed the necessity of the research but I suggest adding the literature review for some related research in this field to clarify the novelty of the research
- I am not sure that ADV can work properly in flood time
How does CADMT work to measure velocity and discharge? Are there any previous experiences in this field? More clarification is required
- More clarification is required to determine the severe level on page 11. Does it constant as a caution level or can take different values?
Why do the authors emphasize the development of FEWF on small streams? - What are the limitations of the FEWF for other streams?
- What are the advantages of minimizing a sum of less rapidly increasing functions of the residuals, rather than minimizing a sum of squared errors?
- In Fig. 8, I could not accept the monograph for the whole data. What is the reason for suggesting this monograph?
- If the authors remove small rainfall and small discharges, it seems they can obtain better results for forecasting
- In Fig. 9, many of the measured data are for the low depth and discharge while the simulated data is very limited for high discharge. - I suggest removing some of the low depth, because the FEWF is applied for high discharges and rating curves must be verified for these ranges. Another important issue is the suitability of the manning equation for high discharges because in high discharge, the bed load transport will happen and the determination of the roughness coefficient is very complicated. So I do not agree that the rating curve can be effectively used to forecast depth with discharge in both measured and unmeasured small streams.
- Figure 11 is not clear to review. The authors must discuss some differences in this figure
- What is the high accuracy between measure depth and discharge with the predicted in Figure 12 while the many of input data have high uncertainty?
-The authors have only described some of the results of the FEWF without discussing the results, the limitations of the methods, and especially the generality of the method for other streams

It seems the quality of writing require minor corrections.

Author Response

Response to Reviewer 2 Comments

Thank you very much for the invaluable suggestions.

Point 1: The author discussed the necessity of the research but I suggest adding the literature review for some related research in this field to clarify the novelty of the research

Response 1: It is one of the results of the NDMI research, Korea and has been added to L82-L83.

 

Point 2: I am not sure that ADV can work properly in flood time.

How does CADMT work to measure velocity and discharge? Are there any previous experiences in this field? More clarification is required

Response 2: ADV is sometimes not suitable for use in the flood season because it requires manpower, but it is used as the verification method because it is the traditional verified measurement method. In this research, the measurements were made at human acceptable discharge.

CADMT can measure the flow velocity by capturing the particles in the images stored at each moment from the CCTV and calculating the distance, direction and traveling time. Velocity is measured at each point. The surface velocity can be averaged from the velocity field (L73).

 

Point 3: More clarification is required to determine the severe level on page 11. Does it constant as a caution level or can take different values?

Why do the authors emphasize the development of FEWF on small streams? - What are the limitations of the FEWF for other streams?

Response 3:

Caution water levels are typically set at adult knee height, but severe water levels are typically set at 70% of the design flood discharge. The summary and introduction also have been revised. Small streams have very short flood travel times, making it very difficult to construct the FEWF. However, the development of ICT-based measurement technology is helping to understand the physical characteristics of small stream floods and to forecast the discharge. We developed the FEWF for small streams to find a way to prevent human accidents.

 

Point 4: What are the advantages of minimizing a sum of less rapidly increasing functions of the residuals, rather than minimizing a sum of squared errors?

Response 4: To apply regression analysis, it is necessary to consider the effect on the residuals. Minimizing the sum of less has the advantage of reducing outlier effects compared to minimizing the sum of squared errors (L293).

 

Point 5: In Fig. 8, I could not accept the monograph for the whole data. What is the reason for suggesting this monograph?

Response 5: The nomograph is a very simple method of forecasting runoff from rainfall, and when real-time rainfall is reflected, the runoff in the small streams can be forecasted more quickly. All data were reflected so that not only the high-water level but also the low-water level effect could be considered for real-time flood and water level forecasting.

 

Point 6: If the authors remove small rainfall and small discharges, it seems they can obtain better results for forecasting.

Response 6: When creating the nomograph, all flood discharges below the caution level were removed as boundary conditions, but the results gave poor results compared to when all flood discharges were included. This research will be presented in future work.

 

Point 7: In Fig. 9, many of the measured data are for the low depth and discharge while the simulated data is very limited for high discharge.

Response 7: Only the flood event above the caution level was selected and compared to relatively large values. The number of flood events that occurred in 2021 was not more than expected, so we used the largest flood event.

 

Point 8: I suggest removing some of the low depth, because the FEWF is applied for high discharges and rating curves must be verified for these ranges. Another important issue is the suitability of the manning equation for high discharges because in high discharge, the bed load transport will happen and the determination of the roughness coefficient is very complicated. So I do not agree that the rating curve can be effectively used to forecast depth with discharge in both measured and unmeasured small streams.

Response 8: We have the cross sections of the small streams, so we use the Manning formula to calculate the water level and discharge. Therefore, there was no need to remove small water level values and it is reasonable to use all the data to build the FEWF using real time or forecasted rainfall.

 

Point 9: Figure 11 is not clear to review. The authors must discuss some differences in this figure.

Response 9: It was rewritten by dividing it into Figure 11 and Figure 12, and the flood event in 2021 for each stream was randomly selected and shown.

 

Point 10: What is the high accuracy between measure depth and discharge with the predicted in Figure 12 while the many of input data have high uncertainty?

Response 10: A Q-Q plot shows the relationship between calculations and observations. If the diagonal line intersects each marker with the center line of the diagonal, the two values are in good agreement.

 

Point 11: The authors have only described some of the results of the FEWF without discussing the results, the limitations of the methods, and especially the generality of the method for other streams

Response 11: A conclusion was added, and both the abstract and introduction were revised.

Author Response File: Author Response.pdf

Reviewer 3 Report

The manuscript provided by Koo and co-workers is focused on development of flood early warning frameworks for the small streams in Korea. In my opinion, although this article contains new aspects, the manuscript can be accepted with major revisions at Water.

 

- English writing needs further polish.

- Abstract should be contained background including aims, methods, results and conclusion.

- In the introduction section 'over the past 93 years, from 1927 to 2019 … is wrong'. 1931 to 2023 is correct.

- The quality of the discussion section must be improved. In so doing, the authors must be organized the discussion from the general to the specific, linking your findings to the literature, then to theory, then practice and avoid repetition from the introduction.

- The "literature review" section of the manuscript must be improved. It is necessary to compare the results of the present study with previous similar studies.

- The 'conclusions part' must be added. Also, conclusions of your paper are especially important for this. Therefore, please try to sharpen this further. The optimal conclusion should include:

* A summary of your key findings.

* A highlight of your hypothesis, new concepts, and innovations.

* A summary of key improvements compared to findings in the literature (provide a couple of references to indicate key improvements).

* Your vision for future work.

* Limitations of the study.

- Numbering of figures is incorrect.

- For numbers in text and tables < 1.00, use three digits beyond the decimal point; for numbers between 1.00 and 9.99 use two digits beyond the decimal point; for numbers between 10.0 and 99.9, use one digit beyond the decimal point; and for concentrations ≥ 100, use the nearest whole number.

 

- The paper has some typographical and grammatical errors which must be corrected. Therefore, the English language of the text should be strongly revised by a native English speaker with expertise in the scientific field and skills in scientific paper writing.

Author Response

Response to Reviewer 3 Comments

Thank you very much for the invaluable suggestions.

 

Point 1: English writing needs further polish.

 Response 1:  We have reviewed the terms and grammar used in the entire text in order to make the text more meaningful.

 

 

Point 2: Abstract should be contained background including aims, methods, results and conclusion.

Response 2: The abstract was rewritten with background, purpose, methods, results and conclusions:

Currently, Korea is undergoing significant local extreme precipitation, which contributes to more than 80% of flood disasters. Additionally, there is an increasing occurrence of such extreme precipitation in small stream basins, accounting for over 60% of flood disasters. Consequently, it becomes imperative to forecast runoff and water levels in advance to effectively mitigate flood disasters in small streams. The Flood Early Warning Framework (FEWF) presents one solution to reduce flood disasters by enabling the forecast of discharge and water levels during flood events. However, the application of FEWF in existing research is challenging due to the short flood travel time characteristic of small streams. This research proposes a methodology for constructing FEWF tailored to small streams using the nomograph and rating curve method. To evaluate the effectiveness of FEWF, 6-years dataset from the Closed-circuit television-based Automatic Discharge Measurement Technique (CADMT) was utilized. The results indicate that FEWF successfully forecasts discharge and depth during flood events. By leveraging CADMT technology and real-time data, the development of precise and dependable FEWFs becomes possible. This advancement holds the potential to mitigate the consequences of extreme rainfall events and minimize flood-related casualties in small stream basins.

 

Point 3: In the introduction section 'over the past 93 years, from 1927 to 2019 … is wrong'. 1931 to 2023 is correct.

 Response 3: I have corrected this as below:

According to the climatological statistics of Korea national climate data center in 2020, a frequency of heavy rainfall exceeding 100 mm per hour in South Korea has steadily increased as the average temperature has risen from 1927 to 2019, as shown by the blue line in Figure 1 a. The area of occurrence is also expanding throughout Korea. Extreme rainfall that occurred once in the 1920s increased to 5 times in the 1980s, then to eight times in the 2010s, and continued to increase gradually as shown in Figure 1a.

 

 

Point 4: The quality of the discussion section must be improved. In so doing, the authors must be organized the discussion from the general to the specific, linking your findings to the literature, then to theory, then practice and avoid repetition from the introduction.

Response 4: A discussion sub-chapter has been added and the content has been further revised.

 

 

Point 5: The "literature review" section of the manuscript must be improved. It is necessary to compare the results of the present study with previous similar studies.

Response 5: The introduction has been revised.

 

Point 6: The 'conclusions part' must be added. Also, conclusions of your paper are especially important for this. Therefore, please try to sharpen this further. The optimal conclusion should include:

* A summary of your key findings.

* A highlight of your hypothesis, new concepts, and innovations.

* A summary of key improvements compared to findings in the literature (provide a couple of references to indicate key improvements).

* Your vision for future work.

* Limitations of the study.

Response 6:  The text has been revised to reflect all of the points listed above.

 

Point 7: Numbering of figures is incorrect.

Response 7: Corrected the number of figures.

 

 

Point 8: For numbers in text and tables < 1.00, use three digits beyond the decimal point; for numbers between 1.00 and 9.99 use two digits beyond the decimal point; for numbers between 10.0 and 99.9, use one digit beyond the decimal point; and for concentrations ≥ 100, use the nearest whole number.

 

Response 8: Tables and text have been revised to reflect the above.

 

 

Point 8: The paper has some typographical and grammatical errors which must be corrected. Therefore, the English language of the text should be strongly revised by a native English speaker with expertise in the scientific field and skills in scientific paper writing.

Response 8: The text has been further edited.

Author Response File: Author Response.pdf

Reviewer 4 Report

Cheong et al. developed flood early warning frameworks for small streams in Korea, using the nomograph and rating curve methods. In general, the paper is well written, but there still exist some parts that are not clear or rigorous enough. The comments are as the followings.

1.     L31-33: It seems that the knowledge in this sentence is not your finding in this study, so you must add citations here. It is a critical issue for academic articles. Please go through the whole paper carefully, and add citations to each corresponding sentence.

2.     L33-34: How do you know the area of occurrence is expanding? If you get it from literature, please add citations.

3.     Figure 1b: Do the blue circles mean all the locations that occurred extreme rainfall from 1927-2019? I think it would be better if you can add one figure to show the occurrence region in 1927 (or the first several decades in the early years) and compare it with the current extreme rainfall occurrence region map in 2019 or the current decade. It would better support your statement that ‘the area of occurrence is expanding’.

4.     L34-35: How can you get these results? From Figure 1a? If yes, cite Figure 1a here.

5.     Figure 1: Add legend

6.     L54-55: Add a brief explanation of why numerical model-based systems cannot ensure emergency response time. Are they more time-consuming? and why? Briefly explain it in the manuscript.

7.     Section 1.2: Add citations

8.     L146: It is still not clear enough why you choose these five streams. Is the data not available in other streams? Or the data quality is not good in other streams? What is your standard to choose stream sites in terms of basin characteristics and geographic location? Please add more details to the manuscript.

9.     Table 1: 1) What is the unit of Wc? Km? 2)What does C mean? Is that the same index as the ‘Cb’ in the L152?

10.   Figure 2: The font size and the legend on the five sub-figures are too small that cannot be seen clearly.

11.   Figure 3: Legend: ’before’ and ‘after’ to ‘before flood season’, and ‘after flood season’.

12.   Figure 8, 9,11,13: The font size is too small.

13.   L474-476: Are these results significant? What is the confidence interval?

14.   L503-506: How did you choose the training dataset (data to develop the system) and the test dataset (data for evaluation)? Did you choose them randomly? What is the percentage of all available data you chose for the test dataset? Will the results change if you choose a different dataset for developing the system and evaluation?

 

 

Author Response

Response to Reviewer 4 Comments

Thank you very much for the invaluable suggestions.

Point 1: L31-33: It seems that the knowledge in this sentence is not your finding in this study, so you must add citations here. It is a critical issue for academic articles. Please go through the whole paper carefully, and add citations to each corresponding sentence.

 Response 1: The content of the text has been changed in the following way to reflect the opinion of the reviewer:

According to the climatological statistics of Korea national climate data center, a frequency of heavy rainfall exceeding 100 mm per hour in South Korea has steadily increased as the average temperature has risen from 1927 to 2019, as shown by the blue line in Figure 1 a.

 

Point 2: L33-34: How do you know the area of occurrence is expanding? If you get it from literature, please add citations.

 Response 2: The content of the text has been changed in the following way to reflect the opinion of the reviewer:

The area of occurrence is also expanding throughout Korea. Extreme rainfall that occurred once in the 1920s increased to 5 times in the 1980s, then to eight times in the 2010s, and continued to increase gradually as shown in Figure 1a.

 

Point 3: Figure 1b: Do the blue circles mean all the locations that occurred extreme rainfall from 1927-2019? I think it would be better if you can add one figure to show the occurrence region in 1927 (or the first several decades in the early years) and compare it with the current extreme rainfall occurrence region map in 2019 or the current decade. It would better support your statement that ‘the area of occurrence is expanding’.

 Response 3: The figure has been changed to reflect the opinion of the reviewer.

 

 

Point 4: L34-35: How can you get these results? From Figure 1a? If yes, cite Figure 1a here.

Response 4: Figure 1a has been cited.

 

 

Point 5: Figure 1: Add legend

Response 5: Added legend

 

 

Point 6: L54-55: Add a brief explanation of why numerical model-based systems cannot ensure emergency response time. Are they more time-consuming? and why? Briefly explain it in the manuscript

Response 6: The content was incorrect and has been corrected as follows:

The most common hydraulic model that reflects the physical characteristics of river to-pography and flow is the dynamic wave model with the Saint-Vernant equation as the governing equation [10]. It is not always an appropriate method to use for forecasting purposes. In addition, to ensure the accuracy and reliability of the flood early warning system, the data analysis method is required that considers the unique characteristics of small streams, such as steep slopes and short flood travel times.

 

 

Point 7: Section 1.2: Add citations

Response 7: The introduction has been revised.

 

 

Point 8: L146: It is still not clear enough why you choose these five streams. Is the data not available in other streams? Or the data quality is not good in other streams? What is your standard to choose stream sites in terms of basin characteristics and geographic location? Please add more details to the manuscript.

Response 8: The content of the text has been changed in the following way to reflect the opinion of the reviewer:

In order to develop and evaluate the FEWF, the research selected 5 small streams with CADMT installed: Jungsunpill, Sunjang, Unchon, Neungmac, and Insu streams, as test beds. For these small streams, NDMI conducted the field survey on 89 small streams in Korea to apply FEWF as a pilot plant, and selected those streams that are easy to install CADMT [18].

 

 

Point 9: Table 1: 1) What is the unit of Wc? Km? 2)What does C mean? Is that the same index as the ‘Cb’ in the L152?

Response 9: The table has been revised. The unit of Wc is m.

 

Point 10: Figure 2: The font size and the legend on the five sub-figures are too small that cannot be seen clearly.

Response 10: The figure has been changed to reflect the opinion of the reviewer.

 

 

Point 11: Figure 3: Legend: ’before’ and ‘after’ to ‘before flood season’, and ‘after flood season’.

Response 11: The figure has been changed to reflect the opinion of the reviewer.

 

 

Point 12: Figure 8, 9,11,13: The font size is too small.

Response 12:  The figure has been changed to reflect the opinion of the reviewer.

 

 

Point 13: L474-476: Are these results significant? What is the confidence interval?

Response 13: The coefficient of determination is one of the criteria used to evaluate the result. The coefficient of determination ranges from 0 to 1. If all observations fall on the regression line, the coefficient of determination is 1.

 

 

Point 14: L503-506: How did you choose the training dataset (data to develop the system) and the test dataset (data for evaluation)? Did you choose them randomly? What is the percentage of all available data you chose for the test dataset? Will the results change if you choose a different dataset for developing the system and evaluation?

Response 14: We selected the pilot plant from 2016 to measure the discharge and water level, and selected the measured flood events from 2016 to 2020 to create the nomograph. The flood events for verification in 2021 are the result of applying the values that exist for each small steam. It is expected that the resulting value will be affected by the flow and physical characteristics of the small streams, but the influence is not expected to be large because the nomograph is also created by reflecting such influences.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

The discussion should be more vigorous. I suggested in my first review round that a comparison with previous researchers in this area, the limitations of the method, and especially the generality of the method for other streams should be included in the discussion section. The authors have only presented some parts of the results in the discussion.

Reviewer 3 Report

The revised manuscript has addressed my concerns and I agree with its publication in present form.

 

Quality of English language is appropriate.

Reviewer 4 Report

Most of the comments are responded suitably. The current version improved compared with the first version. I suggest accepting the current manuscript.