Comparison between Quantile Regression Technique and Generalised Additive Model for Regional Flood Frequency Analysis: A Case Study for Victoria, Australia

Round 1

Reviewer 1 Report

L20-23: It is somewhat confusing.

L29-30: “The accuracy of FFA largely depends on availability of good quality flood data of adequate length”. It is redundant with what was written in the previous sentence. Either an all-inclusive sentence is made, or the latter is superfluous.

L126: For readers who do not have a detailed knowledge of the geography of Australia, nor of the State of Victoria, more graphic information is needed. In addition, the print "quality" of the figure is very low. I recommend increasing the quality of the figure and adding more graphical information on the geography of the catchments.

L192-196: It would be useful if the authors could explain, in the material and methods section, what and how they have used clustering techniques to establish the four regions. Readers who wish to "replicate" what the authors have done would appreciate it.

L217: Table 2: Why do the authors write the full p-value? Does the reader have to look at a lower value for any of the parameters to decide which is better? I cannot find any reference to this in the text, so it seems unnecessary to me.

L227: “RE spreads for ARIs of 5, 20, 50, and 100 years are significantly higher…” Significantly higher? Y... where do the authors justify such an assertion?

L230-232: “This demonstrates that higher ARI flood quantiles are associated with a higher degree of uncertainty, as indicated by a greater degree of spread in the RE”.

Well... it doesn't seem so clear to me. In fact, it is still not very clear to me that it is not the result of chance or of the dispersion of information (or its quality). If anything, "it could indicate that...".

L251: What does GCV mean? In addition to explaining it in the text, it should be explained in the heading of Table 3, as tables should be "self-explanatory". This needs to be corrected.

L257: Table 3: same comment regarding the p-value that I made for table 2.

L292: Figure 5 could be cropped vertically, when the Ratio(Qpred/Qobs) reaches the value of "-2".

Author Response

L20-23: It is somewhat confusing.

Authors’ response: Thanks for your comment. These lines are revised as below: “The GAM model performance is found to be better for smaller return periods (i.e., 2, 5 and 10 years) with a median relative error ranging 16-41%. For higher return periods (i.e., 20, 50 and 100 years), log-log linear regression model (QRT) outperforms the GAM model with a median relative error ranging 31-59%.

L29-30: “The accuracy of FFA largely depends on availability of good quality flood data of adequate length”. It is redundant with what was written in the previous sentence. Either an all-inclusive sentence is made, or the latter is superfluous.

Authors’ response: Thanks for your comment. This sentence is now deleted.

L126: For readers who do not have a detailed knowledge of the geography of Australia, nor of the State of Victoria, more graphic information is needed. In addition, the print "quality" of the figure is very low. I recommend increasing the quality of the figure and adding more graphical information on the geography of the catchments.

Authors’ response: Figure 1 is updated which shows location of Victoria in Australia.

L192-196: It would be useful if the authors could explain, in the material and methods section, what and how they have used clustering techniques to establish the four regions. Readers who wish to "replicate" what the authors have done would appreciate it.

Authors’ response: Thanks for the suggestion. A new section (Section 3.3) on cluster analysis is added in Methodology.

L217: Table 2: Why do the authors write the full p-value? Does the reader have to look at a lower value for any of the parameters to decide which is better? I cannot find any reference to this in the text, so it seems unnecessary to me.

Authors’ response: This column with p-value is deleted in the revised manuscript.

L227: “RE spreads for ARIs of 5, 20, 50, and 100 years are significantly higher…” Significantly higher? Y... where do the authors justify such an assertion?

Authors’ response: Thanks for the comment. This sentence is now deleted.

L230-232: “This demonstrates that higher ARI flood quantiles are associated with a higher degree of uncertainty, as indicated by a greater degree of spread in the RE”. Well... it doesn't seem so clear to me. In fact, it is still not very clear to me that it is not the result of chance or of the dispersion of information (or its quality). If anything, "it could indicate that...".

Authors’ response: The sentence is modified as below: Higher ARI flood quantiles are associated with a higher degree of spread in the RE, which could indicate a higher standard deviation in the estimate.

L251: What does GCV mean? In addition to explaining it in the text, it should be explained in the heading of Table 3, as tables should be "self-explanatory". This needs to be corrected.

Authors’ response: GCV means generalized cross-validation (GCV) statistic. It is corrected in the text and in Table 3.

L257: Table 3: same comment regarding the p-value that I made for table 2.

Authors’ response: The p-value column is deleted from Table 3.

L292: Figure 5 could be cropped vertically, when the Ratio(Qpred/Qobs) reaches the value of "-2".

Authors’ response: We prefer to keep this figure as it is (to be consistent with Figure 4).

Reviewer 2 Report

Interesting research but there are some issues that need to be addressed.

The aim of the research is missing scientific soundness, it should be better explained in the Introduction. 

A literature review could be presented systematically in table form.

Discussion should be added to clearly represent the obtained data, conclusions and comarision with other studies.

 

Author Response

Interesting research but there are some issues that need to be addressed.

Authors’ response: Thanks for your positive note.

The aim of the research is missing scientific soundness, it should be better explained in the Introduction.

Authors’ response: Thanks for the comment. The following sentences are added in the Introduction: “To fill this knowledge gap, the aim of this study is to test the applicability of GAM to Victoria State of Australia and compare results with linear RFFA techniques. This will assist to select more accurate RFFA technique for practical application in Victoria.”

A literature review could be presented systematically in table form.

Authors’ response: Thanks for your suggestion. However, we prefer to present literature review in descriptive form, a table will duplicate the similar information.

Discussion should be added to clearly represent the obtained data, conclusions and comparison with other studies.

Authors’ response: Thanks for the comment. We have presented a discussion section and compared results of this study with similar other studies.

Round 2

Reviewer 2 Report

The authors have included all comments.

No further comments.