Selection of Bias Correction Methods to Assess the Impact of Climate Change on Flood Frequency Curves
Round 1
Reviewer 1 Report
The work presented contains interesting topics, its abstract is accurate and informative including the study's purpose, main findings, and main conclusions; the research objectives are significant and concisely stated, the results are scientifically correct, soundly interpreted; the interpretations and conclusions are justified by the results; the subject matter is within the scope of the journal.
General comments
In my opinion the uncertainty associated to the EURO-CORDEX climate models used in the proposed study may be investigated, considering the bias errors between climate models and observations; moreover a brief description of the adopted HBV hydrological model should be introduced in the paper, focusing the attention on the evaluation of annual maximum series of streamflow using daily temporal scale for the precipitation and monthly temporal scale for the temperature; finally an interesting evaluation may be conducted (also in future works) on annual maximum series of precipitation using different temporal scales.
Specific comment
In figure 3a should be “Return Period”.
Author Response
First of all, the authors would like to thank the Reviewer for the comments that allow us to improve the paper.
The uncertainty associated to the climate models has been included in the results, including the 33rd and 67th percentiles in Figure 12 that shows the expected flood frequency curves in the future periods.
A more detailed description of the HBV model has been included in Lines 137-146.
Regarding the use of different time scales, the authors think that it would be very interesting to consider varying time scales. However, the authors consider that it is out the scope of this paper and it will be consider for future works.
Specific comment
Thank you very much for the comment. The typo has been corrected. In Figure 3a, ‘Periodo de Retorno’ has been changed to ‘Return period’.
Reviewer 2 Report
In this manuscript, the authors applied some bias corrections methods to improve the prediction of the flow frequency curves related to four Spanish catchments. The paper is clear, well written and provides useful information for regional applications such as the dam design. However, there are some critical issues the authors should address before the publication.
General comments:
GC01: First of all, it is not clear which is the novelty of the research. Are there some novel methodologies or approaches? Are the results consistent with similar literature studies or do they differ somehow? Are there similar studies in the study areas, or this is the first study that provides these information regarding the correction flow frequency curves?
GC02: The authors should illustrate the state of the art of the bias correction methods in flood frequency curves, and specify if their methodology has novel aspects.
GC03: The authors should provide more details about the hydrological model (not just citing the HBV model) and the adopted input data. Which land use and soil data did they use? Which infiltration, basin response and flow routing methods did they adopt?
GC04: The authors did not mention about the errors related to the hydrological model that could affect the following step of the bias correction. Since the authors performed a calibration, is the hydrological model affected by some bias (e.g. depending on the rainfall magnitude)?
GC05: The “Conclusions” section is mainly a summary of the results. The authors should provide more comments on the results, for example comparing these latter with similar literature studies.
Other specific comments are reported below:
SC01: Page 2, Lines 84-85 redundant as respect as lines 79-80.
SC02: Page 3, Figure 1: labels are not readable, please improve the resolution (and maybe the font). Please add a legend in Figure 1 (b) specifying the drainage network, the selected watershed and the administrative boundaries.
SC03: Page 4, lines 99-100: “Gaps in time series were filled by using observations at nearby gauging stations”. How the filling is performed? Did the authors replaced the “no data” values with the closest rain gauge station or did they use some geostatistical method? How far are the closest stations? Are the closest stations at similar altitudes?
SC04: Page 4, Eq. (2): Please introduce the equation and specify the meaning of F(x). (For example “The cumulative distribution function F(x) of the GEV distribution is expressed with the following…”
SC05: Page 4, Line 121: Specify the GAP acronym and a reference
SC06: Page 4, Lines 130-133: Specify the names and the meaning of FC, PERC, K0 and K1 parameters
SC07: Pages 6-7, Figures 3,4,5: Correct the Spanish legend and axis titles in English. (Figure 3, a). Specify the other lines in the legend (more than 10 lines are represented, only two are in the legend). There is no need of the titles if the meaning of the insets is illustrated in the figure’s caption (redundant).
SC08: Page 6, Lines 187-189. Redundant with lines 161-164.
SC09: Page11, Line 289: correct “especifically” with “specifically”
SC10: Page 12: Correct the number of the section (3.4. instead of 3.3)
Author Response
General comments:
GC01: First of all, it is not clear which is the novelty of the research. Are there some novel methodologies or approaches? Are the results consistent with similar literature studies or do they differ somehow? Are there similar studies in the study areas, or this is the first study that provides these information regarding the correction flow frequency curves?.
Most of bias correction techniques are focused on correcting precipitation and temperature projections to improve their fitting to observations. However, no attention is payed at their influence on the resulting flood frequency curve. Consequently, the novelty of this approach is the selection of the best bias correction techniques in terms of the fitting of the resulting flood frequency curves. A comment has been added to the paper (lines 59-64).
As far as the authors know, this is the first study that considers the fitting to the flood frequency curve, as other studies focus on the improvement to the complete time series of precipitation and temperature. A comment has been included in the paper (Lines 61-64).
GC02: The authors should illustrate the state of the art of the bias correction methods in flood frequency curves, and specify if their methodology has novel aspects.
A review of bias corrections methods has been included in the Introduction Section (Lines 53-61). In addition, the novelty of the approach has been included in lines 82-88.
GC03: The authors should provide more details about the hydrological model (not just citing the HBV model) and the adopted input data. Which land use and soil data did they use? Which infiltration, basin response and flow routing methods did they adopt?
The HBV model has been described in more detail in lines 137-146, as well as the methods used. In addition, the source of land uses to estimate soil parameter values has been also included.
GC04: The authors did not mention about the errors related to the hydrological model that could affect the following step of the bias correction. Since the authors performed a calibration, is the hydrological model affected by some bias (e.g. depending on the rainfall magnitude)?
A figure has been added to show the results of the HBV calibration in terms of the frequency curves (Figure 2). It can be seen that there are two catchments with no significant errors, as flood frequency curves obtained by simulation are similar to observations. In the other two catchments, some errors arise depending on the return period. However, in the paper flood changes are obtained from the comparison between the results in the control and future periods, instead of with the flood frequency curve obtained from observations. Consequently, the observed frequency curve has been replaced by the simulated frequency curve in the control period in Figures 8-11.
GC05: The “Conclusions” section is mainly a summary of the results. The authors should provide more comments on the results, for example comparing these latter with similar literature studies.
Some comments about the comparisons with previous studies have been included in the Conclusions section.
Other specific comments are reported below:
SC01: Page 2, Lines 84-85 redundant as respect as lines 79-80.
Redundant information has been removed from the paper.
SC02: Page 3, Figure 1: labels are not readable, please improve the resolution (and maybe the font). Please add a legend in Figure 1 (b) specifying the drainage network, the selected watershed and the administrative boundaries.
The resolution of the figure has been improved.
SC03: Page 4, lines 99-100: “Gaps in time series were filled by using observations at nearby gauging stations”. How the filling is performed? Did the authors replaced the “no data” values with the closest rain gauge station or did they use some geostatistical method? How far are the closest stations? Are the closest stations at similar altitudes?
Gaps in precipitation data were filled by using the inverse distance weighting method with squared distances. Consequently, the gaps were filled with the closest rain gauge stations not with only the closest one. We did not found significant differences among rain-gauging stations in terms of height. Consequently, the normal-ration method or other more complex geostatistical techniques that account for precipitation variations in height were not considered.
A comment has been added in Line 113
SC04: Page 4, Eq. (2): Please introduce the equation and specify the meaning of F(x). (For example “The cumulative distribution function F(x) of the GEV distribution is expressed with the following…”.
The suggested comment has been included in the paper in the lines 129-130.
SC05: Page 4, Line 121: Specify the GAP acronym and a reference
The GAP acronym has been included in the list of abbreviations.
SC06: Page 4, Lines 130-133: Specify the names and the meaning of FC, PERC, K0 and K1 parameters
A brief description of the parameters has been included in the paper.
SC07: Pages 6-7, Figures 3,4,5: Correct the Spanish legend and axis titles in English. (Figure 3, a). Specify the other lines in the legend (more than 10 lines are represented, only two are in the legend). There is no need of the titles if the meaning of the insets is illustrated in the figure’s caption (redundant).
The figures have been improved following the comments.
SC08: Page 6, Lines 187-189. Redundant with lines 161-164.
The redundant information has been removed from the paper.
SC09: Page11, Line 289: correct “especifically” with “specifically”
Thank you very much for the comment. The typo has been corrected. “Especifically” has been changed to ‘specifically’
SC10: Page 12: Correct the number of the section (3.4. instead of 3.3)
The number of the section has been changed.
Round 2
Reviewer 2 Report
The authors did all the minor and major corrections I requested. The manuscript has been improved and now it is easier to follow and the novel aspect are much clearer. To my view, the paper can be accepted. For future reviews, I recommend the authors to always report the line numbers in the replies to the comments. For example, lines numbers are missing in the replies to comments GC05, SC01, SC06.
