Downscaling Climatic Variables at a River Basin Scale: Statistical Validation and Ensemble Projection under Climate Change Scenarios

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This study presented a statistical downscaling at selected weather stations to evaluate future climate change. I have some questions with the study approach in the hope the authors could address. My overall recommendation is resubmission/major revision. 

1. I don't understand how did authors determine cumulative changes from single stations. Perhaps the average from all stations was calculated? Since the weather stations only locate in relative low elevation regions, it is possible that cumulative changes from all stations are not representative for the basin-wide changes. 

2. I am not confident on author's approach in selecting GCM and scenarios. CMIP6 has been out for a long time. The authors only used 3 GCMs. What are reasons behind selecting those GCMs? Are those GCMs representative of future changes in T and P? What are uncertainties related to GCM selection? Those questions need to be addressed.

3. Follow up on previous point, from 2006 to 2023, there are 17 years of data for evaluating GCMs. How well does the selected GCMs and emission scenarios perform in simulating weather during this period? this could be a useful indicator on GCM selection. 

Comments on the Quality of English Language

language quality is good for publication. 

Author Response

Response to Reviewer 1

 

Comment 1

I don't understand how did authors determine cumulative changes from single stations. Perhaps the average from all stations was calculated? Since the weather stations only locate in relative low elevation regions, it is possible that cumulative changes from all stations are not representative for the basin-wide changes. 

Response 1

The cumulative change was done for every station alone and then the resulting cumulative change was averaged over all the stations. As for the bias related to elevation, this is indeed a possibility that one cannot test for. In the revised manuscript however, we underline this important potential limitation in lines 97-107.

 

Comment 2

I am not confident on author's approach in selecting GCM and scenarios. CMIP6 has been out for a long time. The authors only used 3 GCMs. What are reasons behind selecting those GCMs? Are those GCMs representative of future changes in T and P? What are uncertainties related to GCM selection? Those questions need to be addressed.

Response 2

The CMIP6 data publication began in 2019 and the majority of the data publication was completed in 2022 while our study started in 2019 and as such CMIP5 data was used. More broadly, the downscaling models will remain unchanged if one wants to apply them to CMIP6 or other future models.

As stated in line 178-180, the GCMs were selected based on the availability of data, their spatial resolution, and their individual performance in the region.

 

Comment 3

Follow up on previous point, from 2006 to 2023, there are 17 years of data for evaluating GCMs. How well does the selected GCMs and emission scenarios perform in simulating weather during this period? this could be a useful indicator on GCM selection. 

Response 3

Performance assessment of the selected GCMs during the period 2006 to 2023 was not conducted since not all the stations data was available and the work on this study started in 2019.

As stated above, the GCMs were selected based on the availability of data, their spatial resolution, and their individual performance in the region. At the beginning of the study, we selected the newest generation of the GCMs in the Coupled Model Intercomparison Project (CMIP5). We selected the same time frame from 1981 to 2005 and also the same variables from NCEP to have consistency with the atmospheric variables available from CanESM2, GFDL-ESM2M and HadGEM-CC outputs and the outcome from the GCM ensemble versus NCEP and the observed data at the weather stations and we included 2 examples as shown in Figure 6.

 

Reviewer 2 Report

Comments and Suggestions for Authors

General comments: This paper has introduced the SD method based on MLR over a typically semiarid region, where water resource management design is highly demanded. The topic of this paper is interesting, and the overall written English is fine, however, some issues related to logic and technology need to be enhanced and clarified to sharpen the novelty of this work as follows.

 

Comment 1: 1) The issues and objectives studied in this article need to be particularly clear (such as original defects - method innovation or insufficient product strength - new data enhancement). Is the validation of downscaling methods? Data products? 2) The English language seems fine. The work has the potential to be published but before it should be considered for publication, it has to pass through professional proofreading, and all the highlighted points below need to be corrected and implemented.

 

Comment 2: Temperature is easy to calibrate and predict compared to precipitation, which was not discovered through the method design research in this article. This article should strengthen the regional applicability advantages of the method to enhance this article’s novelty.

 

Comment 3: I suggest the authors revise the introduction of the study per the comments raised. The authors can also use the following points below as a guideline to help them come out with an exciting introduction that is more scientific.

Background & Significance: (What general background does the reader need to understand the manuscript and how important is it in the context of scientific research).

Problem definition: (What are the research questions to fill in the gaps of the existing knowledge body or methodology )?

Motivations & Objectives: (Why are you conducting the study and what are the goals to achieve?)

 

Comment 4. For Section 2:

For data processing in section 2.1, the station resolution seems around or less than 1 degree (Table 2), while CMIP5 datasets of great scales (Why not CMIP6?) direct interpolation on-site scale seems too rough and unconvinced because this unmatched scale can directly affect your results. Thus, can the authors clearly explain their internal relations during your result and method sections?

For section 2.2, please provide formulas for IDW and standardization schemes for wider reading.

For section 2.4, in Lines 185-186, the choice standard of this threshold needs to be briefly explained.

 

Comment 5. For Section 3:

The results section should not only describe the graphs and tables but also include a dialectical analysis of the underlying reasons behind your Findings. This needs to be strengthened in each Finding of every subsection.

 

Comment 6: For Section 4:

In Lines 359-361, this sentence is more proper to be placed in the section Data or Method.

In Line 364, what is the “comparable performance” exactly here?

In Lines 380-391, which results in this part are consistent or contradictory to those in this article’s Findings, and further analysis is needed.

 

Comment 7: For Section 5:

The specific scenarios in which this method can be applied need to be specified. The advantages of the method need to be highlighted, and its disadvantages should also be pointed out. 

Comments on the Quality of English Language

The topic of this paper is interesting, and the overall written English is fine, however, some issues related to logic and technology need to be enhanced and clarified to sharpen the novelty of this work.

Author Response

Response to Reviewer 2

 

General comments:

This paper has introduced the SD method based on MLR over a typically semiarid region, where water resource management design is highly demanded. The topic of this paper is interesting, and the overall written English is fine, however, some issues related to logic and technology need to be enhanced and clarified to sharpen the novelty of this work as follows.

General response

We thank the reviewer for the valuable comments and their positive overall assessment of the paper. We have made every effort to address all comments which we indeed believe to have improved the quality of the paper.

Comment 1:

1) The issues and objectives studied in this article need to be particularly clear (such as original defects - method innovation or insufficient product strength - new data enhancement). Is the validation of downscaling methods? Data products?

2) The English language seems fine. The work has the potential to be published but before it should be considered for publication, it has to pass through professional proofreading, and all the highlighted points below need to be corrected and implemented.

Response 1

• We have rewritten the introduction to summarize the open challenges and goals of the study into 3 questions: Can MLR downscaling significantly improve the projected climate variables at a local scale compared to the coarse GCM fields? How should an MLR model trained on historic reanalysis data be applied to downscale future GCM projections and deal with their biases? What is the vulnerability of water resources in the arid Jordan River Basin to climate change?

These issues and objectives are included in lines 118-132.

• We have revised the writing of the MS and we think it is in better shape linguistically.

 

Comment 2:

Temperature is easy to calibrate and predict compared to precipitation, which was not discovered through the method design research in this article. This article should strengthen the regional applicability advantages of the method to enhance this article’s novelty.

Response 2

As the reviewer suggested, we did in fact find that the temperature projection had lower uncertainties than precipitation projection. However, this was not an a-priori goal of the design methods as it would potentially bias the findings.

The methodology is replicable to other regional or worldwide basins. However, as with all statistical downscaling, the exact MLR model is local and applicable only to the station for which it is developed. In this context, increasing the regional applicability may be inherently limited.

Comment 3:

I suggest the authors revise the introduction of the study per the comments raised. The authors can also use the following points below as a guideline to help them come out with an exciting introduction that is more scientific.

Background & Significance: (What general background does the reader need to understand the manuscript and how important is it in the context of scientific research).

Problem definition: (What are the research questions to fill in the gaps of the existing knowledge body or methodology)?

Motivations & Objectives: (Why are you conducting the study and what are the goals to achieve?)

Response 3

Background & Significance: The background and significance are added in lines 39-46. The importance of the manuscript in the context of scientific research is explained in lines 80-96.

Problem definition: the research questions are now formulated in lines 154-169.

Motivations & Objectives: it is clarified in lines 120-132.

 

Comment 4. For Section 2:

For data processing in section 2.1, the station resolution seems around or less than 1 degree (Table 2), while CMIP5 datasets of great scales (Why not CMIP6?) direct interpolation on-site scale seems too rough and unconvinced because this unmatched scale can directly affect your results. Thus, can the authors clearly explain their internal relations during your result and method sections?

For section 2.2, please provide formulas for IDW and standardization schemes for wider reading.

For section 2.4, in Lines 185-186, the choice standard of this threshold needs to be briefly explained.

Response 4

Section 2.1: The CMIP6 data publication began in 2019 and the majority of the data publication was completed in 2022 while our study started in 2019 and as such CMIP5 data was used.

Regarding resolution, the CMIP data set as the reviewer suggests have the resolution of the relevant model. The station data however cannot be associated with a resolution; their data is local at that point. Statistical downscaling is thus aiming to reproduce the local, pointwise time series from the coarse CMIP fields through the training on the historical data to reduce the relationship between these local time series and the coarse CMIP-scale fields.

Section 2.2: Formula for IDW is included and standardization schemes are included in the revised manuscript.

Section 2.3: Given the dominant aridity in the basin, 1-mm threshold for a wet month is considered based on other similar locations [74,76]. While it is arbitrary, the choice of this threshold will not affect the results because it is a low value and for the periods with significant rainfall this threshold is easily exceeded.

Comment 5. For Section 3:

The results section should not only describe the graphs and tables but also include a dialectical analysis of the underlying reasons behind your Findings. This needs to be strengthened in each Finding of every subsection.

Response 5

We planned the analysis and discussion in separate sections (discussion and conclusion). We can add further clarifications about the findings if the reviewer finds a particular section lacking in discussion.

 

Comment 6: For Section 4:

In Lines 359-361, this sentence is more proper to be placed in the section Data or Method.

In Line 364, what is the “comparable performance” exactly here?

In Lines 380-391, which results in this part are consistent or contradictory to those in this article’s Findings, and further analysis is needed.

Response 6

In Lines 359-361, the sentence is now placed in Section 2.4.1 in the revised manuscript (lines 269-271)

In Line 364 this is now clarified to “similar performance”. We mean comparable errors. This was corrected in the revised manuscript line 436.

In Lines 380-391, the results were compared in the revised manuscript lines 454-467.

 

Comment 7: For Section 5:

The specific scenarios in which this method can be applied need to be specified. The advantages of the method need to be highlighted, and its disadvantages should also be pointed out. 

Response 7

We have updated the paper accordingly. The specific scenarios in which this method can be applied are now outlined on lines 73-86. The advantages of this method are now listed on lines 454-463 while the disadvantages are now listed on lines 463-467 in the revised manuscript.

 

 

Comments on the Quality of English Language

The topic of this paper is interesting, and the overall written English is fine, however, some issues related to logic and technology need to be enhanced and clarified to sharpen the novelty of this work.

General response

We thank the reviewer for the valuable comments and the positive overall assessment of the paper. We have made every effort to address all comments which we indeed believe to have improved the quality of the paper.