Evaluation of Different Objective Functions Used in the SUFI-2 Calibration Process of SWAT-CUP on Water Balance Analysis: A Case Study of the Pursat River Basin, Cambodia

Round 1

Reviewer 1 Report

Main issue

There is no difference between the two objective functions, NSE and RSR as they are not just correlated, but equivalent, as NSE = 1 – RSR².  Maximising NSE will also minimise RSR. This can be demonstrated in Figures 8, 9, 10 and 11, where the graphs, R², RMSD and relationships are identical.  The paper needs to be rewritten to state that using NSE and RSR as objective functions are equivalent, and thus one of them will not be further considered.  Many changes to text and figure need to be made.

More Minor comments

Abstract – needs to be modified to address the above concern.

Line 75 – would read better as “Thus in this study the objectives …”

Line 86, could delete “which”

Line 146 – should define HRUs – presumably Hydrological Response Units?

Paragraph 3.3.2 – it should be stated in this paragraph that NSE and RSR are equivalent objective functions.

Table 3 – the modified Nash – Sutcliffe efficiency should have absolute values, ie. Not (Q_m – Q_s)^p_i , but ǀQ_m - Q_sǀ^p_i etc.

Line 233 – the word ‘ones” could be deleted

Paragraph commencing line 251 and subsequent – this paragraph should not only contain a description of the results, but discussion on why there are differences between them.

Figure 4 – It is interesting that the performance of the validation as measured by the statistical indices is better than the calibration in many cases.  This would not normally be expected.  Why this result?

Figure 8, 9, 10 and 11 – as above, delete one of either NSE or RSR results, as they are the same.  Text changes will be required as well.

Line 476 – the word “deceiving”, should this  be “deceptive”?

Paragraph commencing line 498 – it seems to this reviewer that the observed discharge pattern follows closely the monthly rainfall pattern, without the need to allow for runoff processes.  Have the authors plotted monthly discharge directly against the monthly rainfall and looked at the correlation? This may affect the discussion.

Reference 4 – should be “..Theoretical Documentation Version 2009..”, not “..Theoretical Cocumentation Version 2009..”?

Author Response

Dear Sir/Madam,

Thank you so much for your comments and suggestions for the improvement of our paper.

We have responded to each comment you made.

Please kindly see the attachment.

Sincerely,
Davy Sao

Author Response File: Author Response.docx

Reviewer 2 Report

Here some specific comments:

Line 44: Please add some references to support this, and specifically on the climate and landuse changes capabilities you are referring to.

Line 60: “has been widely used for calibration and uncertainty analysis.”: Please add some references to support this

Line 91: Give more details about basin characterization: at least report landuse shares, how much is forest?  I know the authors refers to other papers but these data should also be reported in this paper, even if in brief.

Line 105: Please increase resolution of legend in Figure1a and make more evident Discharge Station used for calibration analysis in Figure 1b.

Line 128-142: More details on data used for the analysis are required: 1) for daily precipitation clarify if only 1 station was finally used for the whole basin, in this case how was the precipitation spatially distributed across the basin? Is the Quote used to take into account the influence of elevation on precipitation? Which is the period covered for discharge data ? These information may be also added to Table 1.

Line 169-170: Clarify the difference of “Method” in Table 2: clarify that when relative is reported this means that a percentage relative change is applied to “original” parameters, and only when Replace the min and max range values allowed for calibration are reported..

Line 173: why 8 of 11 functions were selected ? is there a justification for this choice?

Line 291: General comment:

In my view most of the results of the analysis are presented in the Discussion section? Why? I would focus the analysis of results such as discharge estimations, and best parameters sets in the results section for clarity and focus the last section of the Paper on the issue related with the interpretation of such results.

Line 303: is “Revap” explained somewhere in the paper?

Line 356-357: reformulate phrase

Line 358: SOL_AWC was low (+16%): this means that the original value was low and now it is higher? clarify.

Line 402: “All of the objective functions generally overestimated the base flows”: this model behavior should be elaborated and justified in the discussion section

Line 481: “Characteristics of the River Basin”: this description can be moved in the previous section, focusing the discussion on how the basin characteristic affected the study and results.

Author Response

Dear Sir/Madam,

Thank you so much for your comments and suggestions for the improvement of our paper.

We have responded to each comment you made.

Please kindly see the attachment.

Sincerely,
Davy Sao

Author Response File: Author Response.docx

Reviewer 3 Report

This paper presents a very well-designed and well-performed work on testing a SWAT model based on various statistical indices, the metrics included in the SUFI-2 algorithm. I find the paper very well written and structured, very comprehensive and, certainly, interesting for the model community. Although the subject is covered in the literature, such an application in a new case study is useful to demonstrate how different metrics may influence the modelling overall. I have no major comments or concerns as I am satisfied with the paper as it is and the extensive evaluation presented. I think that it can be published as it is.  

Author Response

Dear Sir/Madam,

Thank you so much for your kind review and comments on our paper.

We really appreciate it.

Sincerely,
Davy Sao

Round 2

Reviewer 1 Report

The authors have provided a response to two comments, about the calibration and validation (point 10 in the response) and the annual pattern of rainfall and discharge (point 13) that are reasonable, but I would like to see this discussion included in the paper before it is published.

Author Response

Dear Sir/Madam,

Thank you very much for your comment and precious time reviewing our paper.

We have included the response to point 10 in the paper in Line 307-312 and the response to point 13 in Line 526-527 and Line 532-533.

Sincerely,

Davy Sao