Comparative Analysis of Climate Change Impacts on Meteorological, Hydrological, and Agricultural Droughts in the Lake Titicaca Basin

Round 1

Reviewer 1 Report

The manuscript addresses an interesting point of view when examining the climate change impacts on the area of interest here – Lake Titicaca basin. By assumed indications for meteorological, hydrological, and agricultural draughts with the previously established indices (SPI, SRI, and SSMI), the Authors have obtained potential future results using different climate models and a specific hydrological one.

The research setup is well organized, most steps are thoroughly explained, and the results presented.

The Authors themselves showed respectable scrutiny on GCM results by implementing bias correction procedures and evaluating their performance using Taylor diagrams.

Hydrological models should be expanded, and hydrological processes should be analysed in a more extensive matter, but as the Authors already recognized it as a limitation of this study, it would be quite interesting to see further research since that level of analysis is out of the scope of the presented work.

Both motivation and research goals are well described, and an extensive reference list is given.

Some of the minor comments for the Authors are given below:

-         Abbreviations should be explained the first time when they are used in the text (e.g., RCP8.5 in abstract or ENSO at line 438).

-         In general, some mild English editing might be applied (mostly the style of sentences). Furthermore, the Authors should be concise with usage of terms and names: e.g., RMSE at line 187 is the same as RMSD line 291, but there is called deviation instead of error, and also it is called difference in Fig 3 caption; this generates unnecessary confusion. Same goes with the name of sub-basin Puente Ramis – if it's not indicated that Ramis is a shorter for Puente – it should be used with the same name in every instance of the manuscript.

-         Study area paragraph should be one of the sections, e.g., 2. Study area and it should contain Fig1, which oddly is never mentioned within this paragraph.

-         Website links, such as those used at lines 122 or 137, etc., should be put in the references list and just referenced by number at these positions in the text.

-         There are few irregularities within the reference list – there should be consistency in the way that doi(s) are presented; reference [40] should at least have the name of the project for which the report is given, reference [36] has a faulty web address, etc.

-         Maybe the Authors should comment on why RCP 8.5 has been chosen? If the idea is to investigate the worst-case scenario – it should be annotated as such.

-         When writing about GR2M model, an original reference should be made to Makhlouf 1994 (Makhlouf, Z., Michel, C., 1994. A two-parameter monthly water balance model for French watersheds. Journal of Hydrology 162(1994), 299 – 318).

-         In fig1 there are 11 gauging stations, for apparently 11 sub-basins, however, the Authors present only 9 of them in Table 2, and later only 6 of them are used in some of the results (fig3 and fig4), 7 in fig 5 and only 3 of them in table 3.  Then, why are specifically Puente Ramis and Calacoto Maure stations given in Figure 6? There should be an explanation/description of the reasoning behind these selections.

-         Why is gamma distribution used for SPI? There are references for SSMI and SRI in the manuscript; a similar one for SPI or reasoning is preferable.

-          In figure 3 and 4, there are multiple dots presenting results for different models, and the Authors have indicated they have chosen the 5 of them (Table 1) based on these results, but it would be prudent to at least say how many models were tested in this way i.e., how many dots there are?

-         Figures 5, 9, 10, and 11 should be somewhat improved, increase resolution, or change the line types to make them more visible. More so, since 9-11 deliver key results. Furthermore, the positions of gauging stations are redundant in those figures since the reader can just be pointed to the Fig1.

-         An indication of threshold values of indices (SPI etc.) used for defining the severity of draughts is preferable (pg19-20).

-         Why are only data for the wet season given and analysed, Oct to April? What is happening in the rest of the year?

-         Sentences at lines 518-521 are not quite clear: if draught is lasting between 1-4 months, what does it mean that its duration is increased? Is it maybe the frequency of such draught higher? Or if the duration is increased, then it is not the draught of the 1 or 4 months, but the draught of 2 or 5 months? Some clarification is needed here.

-         If figure 12 is improved (which is preferable), table 5 might be redundant.

-         How is this 69% at line 609 obtained?

Assuming the Authors give these few explanations and minor changes to the manuscript, I would recommend publishing.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

The authors study changes in drought in the Altiplano region. Using three standardized drought indices changes in meteorological, agricultural and hydrological drought are evaluated. The authors study changes in drought frequency, intensity and duration within 2034-2064 with respect to 1984-2014 by applying five selected General Circulation Models under RCP8.5. The GR2M hydrological model is used to calculate the agricultural and hydrological drought indices. The study find regionally differentiated increases in the frequency, duration and intensity of meteorological, agricultural and hydrological drought that are strongly coupled to the projected changes in temperature and precipitation.

Major comments:

• The methodology behind the drought analysis needs to be explained in more detail in order to allow the reader to understand the presented results. Specifically the timescales of the indices need to be defined, as do the variables drought intensity and duration.
• The use of GCMs for this kind of regional study and the applied rescaling (downscaling) of the GCM values needs to be explained and justified.
• A separate discussion section would be valuable. A discussion of the study results with respect to previous studies mentioned in the introduction is missing yet. The discussion of the sources of uncertainties of the presented results could be extended.

More comments and suggestions are given in the attached manuscript-pdf.

Comments for author File: Comments.pdf

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

Topic of the research is very attractive and actual. Research is done with regards to the theoretical knowledge, modeling and simulation on the real case study, which is an added value of the paper.

For the goal of the improvement of the manuscript, I am suggesting a major revision.

1) Line 103 - this is title or subtitle?

2) Map of the analyzed locations must be included within the text.

3) Line 116 - please put references in the appropriate brackets [].

4) Line 125 - authors should avoid writing in the first face. Please avoid writing in style ''we have, I was''. Writing style should be neutral.

5) Line 249 - authors are declaring that monthly precipitation (P), soil moisture (S) and runoff (Q) datasets were used in research.This is in collision with the equation (1), which include insolation. Please, explain this in some short sentence(s).

6) Taylors' diagrams should be additionally described. I.e, what about red and blue dots and their position regarding ideal position?

7) Analysis is based on comparison of the observed and modelled values. One point is missing, which is time series analysis of the observed data. I am suggesting the Rescaled Adjusted Partial Sums (RAPS) method. Within this, new (obtained) subseries could be used for the insight into the possible causes of their appearance. Climate characteristics of the analyzed area are in most of the cases one of the causes of dividing the original time series. I am proposing RAPS analysis for the presented precipitation and stream flow time series.

8) Conclusion is written in the form of a discussion. Authors should remove parts where they are doing discussion and comparison of the values and put this in the discussion chapter.

9) For further research, I strongly suggest analysis of the insolation values, because this is one of the indicators of the climate changes.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

The authors have addressed most of my comments sufficiently. There are still some important aspects with regard to the methodology applied for drought analysis that require revision before the manuscript can be published.

1. Line 277/78 “The procedure can be described by the following for steps: i) fit a probability density function to the frequency distribution of the P, S and Q variables for specified monthly time scales, (…)”
   The monthly time scale that was used for the drought analysis needs to be specified for each index. Further, the authors should explain shortly, why the respective time scale was chosen. Are there any drought impact studies in the region that have tried to link overserved drought impacts to specific monthly time scales of the three drought indices?
2. Line 287-89: “Drought are defined according to their properties such as intensity (value of the index), frequency (the number of dry events) and duration (number of consecutive dry months) in various ways.”
   The definitions of frequency and duration are not helpful. The frequency of drought could be calculated based the number of drought events (according to the definition you are using to determine the drought duration) or you could define the months with an index value below a specific threshold as drought frequency. You may specify an annual frequency or the frequency over the 30-year study period or … You need to define specifically your approach in order to allow the reader to understand your subsequent analysis.
   With respect to drought duration, you need to define the conditions/threshold for starting and ending the drought events. Does the index value need to fall below zero or below -1 to start a drought? Does the drought events end if the index value rises above -1 or zero or …
3. Figure 5: Does this figure show the results of the bias-corrected GCMs? How would the curves look for the non-corrected GCMs? Please, be more specific in the text (lines 347-354), when you are referring to coorected GCMs and when to un-corrected.
   In figure 5 the abbreviation EVT needs to be changed to ET.
4. The paragraph in lines 387-392 (before Figure 7) is repeated in lines 399-404 (after Figure 7)
5. Figure 8 is not understandable without specifying the monthly timescale of the indices SPI, SSMI and SRI. How do you exactly determine the drought intensity for the two 30 year periods? Furthermore, calculating relative trends for the drought intensity might be questionable – the index values more or less range between -2 und +2 and for very small initial values you can get very high percentage changes that are not really meaningful.
6. Figure 9 and 10: The results for the frequency of specific drought durations do strongly depend on the chosen monthly time scale of the index. As the plotted drought durations are generally very short you probably used a very short timescale – 1 month? For evaluating agricultural and hydrological drought longer timescales like 3, 6, 9 or 12 months should be chosen and analysed, as drought needs some time to build up in the systems. Most preferably the chosen timescale is justified by the analysis of impact data. You need to specify your applied methods for the drought analysis.
7. Figure 12 is not understandable without specifying your calculation of drought frequency. You present very high numbers of drought events. Thus, you probably give the frequency for the entire 30 year period and you count every month under your defined drought conditions and not the number of individual drought events.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

Everything is fine now. 

Author Response

Thanks