Evaluation of Extreme Hydroclimatic Trends in River Basins Located in the Northeast and South Regions of Brazil

Round 1

Reviewer 1 Report

1. Several chapters are overwritten, especially in Materials and Methods, CMIP5 models don’t need to be introduced separately.

2. Two maps of the study area should be merged into one, as is the introduction of the study basins.

3. The resolution of CMIP5 mode data is relatively low, generally 250km, and whether it is downscaled during use needs to be explained in the paper.

4. The bias correction method seems too simple, and the correction effect seems to be poor from the following results. Recommend to use a more appropriate correction method, or directly use the data product that has been downscaled and corrected for deviation.

5. The CMIP5 data is only a tool, and it is not necessary to compare it with the observed data in a long space. The paper should focus on the trend of future runoff changes

6. L713-L714, the conclusion " Therefore, the climate models struggle to reproduce the natural variability in the basins of the South region, overestimating the standard deviations throughout the year" indicates that several sets of data are not applicable to the study area, so the simulated results of future climate and runoff are not reliable

7. The focus of this paper should be on a detailed analysis of future projected runoff, but it seems that the entire article is devoted to discussing the applicability of CMIP5 models in your study areas.

8. The references are relatively old, and the literatures in the past three years are few, so it is recommended to supplement

Author Response

Thank you for your review. The answers can be found in the attached file and in the new version of the article. 

Author Response File: Author Response.docx

Reviewer 2 Report

This study projects future inflow at four important basins in Brazil. Four GCM projections are used to consider uncertainties resulting from the choice of GCMs. Findings in this study can be useful information as a reference on future water availability in Brazil with considering uncertainties of flow projections at the target basins. However, the draft needs some clarifications, and I recommend the following revisions.

Major comment:

The analysis does not use temperature projections although temperature is critical for evaluating water loss. Thus, it should be clearly mentioned in the discussion or conclusion that temperature projections are not used in this study. In addition, if possible, it would be beneficial to briefly discuss possible impacts of including temperature projections into your analysis.

Minor comments:

l.32-33:

Time horizon of those projections should be mentioned.

l.196-197:

What is the temporal scale of those observed rainfal data? I recommend showing the locations of the observed rainfall data used for the anlaysis.

l.200-l.201:

It seems that extreme high flows are underestimated in Figure 3. Those underestimations may be related with the identification of outlier values which are greater than 95% of precipitation distribution. Thus, I recommend discussing the possible impacts of your outlier detection method on your inflow simulations. In addition, the distributions you assumed for the precipitation should be mentioned (e.g. uniform distribution?).

Figure 4 (and Figure 3):

Although it is mentioned that National Electric System Operator provides natural flow series from 1931-2005 (l.184-l.185), analysis in Figure 4 includes the observed flow data from 2007-2014. Please clarify the data source for the observed discharge data in Figure 4 (and Figure 3).

l. 227-228:

Your explanation on RCP8.5 says: "a high probability of global surface temperature exceeding 1.5 degree by the end of 21st centry". Was it described/discussed in van Vuuren (2011)? I recommend checking the reference.

Figure 3 and Figure 4, right axis:

'comma' in the right axis should be decimal point.

Figure 3:

Is calibration period from 1994 to 2006 (12 yrs)? Is this correspond to 60% of data for calibration? Your calibration period should be clearly explained in the body of text (not only in the caption). Validation period should be also clearly mentioned in the body of the text.

l.346-347 and l.351-352:

Were the adjustments to the initial base flow and the initial soil moisture content done against September 1994? (If this is the first month of the calibraion period.) Were those same adjustments applied to the future projections? I recommend elaborating on these points.

Author Response

Thank you for your review. The answers can be found in the attached file and in the new version of the article. 

Author Response File: Author Response.docx

Reviewer 3 Report

This study evaluates the trends in hydroclimatic extremes in river basins in the northeast and south regions of Brazil by utilizing hydrological simulation and bias-corrected climate scenarios. One of the main limitations of this work is the introduction. The paper should indeed be correctly contextualised. Similar works include: “Assessment of hydrological trends and changes in hydroclimatic and land use parameters in a river basin in northeast Brazil” and “Land use changes and hydrological trend analysis in a Brazilian Cerrado basin”. What are the main contributions of this work? Is it the case study or the methodology? Does the work involve a comparison between different basins? All points to be underlined.

There seems to be an issue with the indentation. Please verify when to start a new paragraph or not.

In Table 3, there is a small red line/bar.

Have you considered plotting the empirical cumulative distribution function (ECDF) instead of the probability density function?

Please refer to the following publication: Analysis of extreme high streamflow in climate change research: How should hydrological models be calibrated? It should be noted that calibrating on observation data and then applying climate change forcing could raise some concerns. Kindly address this issue in the discussion section.

Please verify when to start a new paragraph or not. Many time there is just a sentence per paragraph.

Author Response

Thank you for your review. The answers can be found in the attached file and in the new version of the article. 

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

1. I would still advise the authors to separate the comparison of CMIP5 models from the hydroclimatic trends, so that the research focus can be more prominent.

2. The last sentence (L868-871) of the discussion section is a bit abrupt, although other reviewers have suggested relevant changes.

3. L652-653, Histograms are not enough to draw conclusions about future trends of flood and drought, I suggest put them in the discussion section.

4. the manuscript should be complemented with the significant of the trend, I suggest giving the trend values.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

The authors addressed nearly all of my concerns and responded to all my queries. Consequently, the introduction, which was a major drawback of the article's previous version, has undergone considerable revision. However, I only partially concur with the authors' views on calibration. I recommend referring again to the cited work. It refers to the calibration of a physically-based model. I am not implying that the authors are mistaken. I would like to highlight that the calibration of a hydrological model with measured data (P in this case) has some limitations. These limitations need to be mentioned in the discussion.

Author Response

Please see the attachment.

Author Response File: Author Response.docx