Research on the Influence of Small-Scale Terrain on Precipitation

Round 1

Reviewer 1 Report

General Comments:

      This paper uses the bidimensional empirical mode decomposition to include the small-scale effect of topography in central China on precipitation distributions which appears to be a method for downscaling. They use the multiple regression to create four different precipitation models (for different scales). The model is trained using April data from 2000-2017. The procedures are defined correctly and used correctly. The results are verified using appropriate statistics. The figures are neat and readable. The references are appropriate in number and quality. The paper should be published after minor revisions.

 

Minor Comments:

1. Line 10-12 and line 175-178: Remove template language.

 

2. Line 246 - 250 Should be in the Data section.  It is suggested that the authors explain that this study is using one month to test the effectiveness of the method. 

 

3. Suggest leaving a space between the Tables and text that follow them.

 

Specific Comments:

1. Line 38: Suggest: Observation should be observational.

 

2. Line 48: Suggest just Naoum et al.

 

3. Line 73: Need to write (BEMD) before “to process”

 

4. Line 87 – 88 Suggest removing “it is located among North, East, South, southwest and Northwest China” since this is understood.

 

5. Line 98: Suggest replacing “come” with “can be obtained”

 

6. Line 104: Define NASA, and give a reference for this information.

 

7. Line 125: Use “Nunes et al. “

Author Response

Please refer to the attachment

Author Response File: Author Response.docx

Reviewer 2 Report

The work is devoted to the study of the influence of the DEM detail on the accuracy of the precipitation distribution model, taking into account the features of relief. On the whole, the proposed approach seems to be adequate and interesting. However, there are specific questions and comments on its implementation carried out in the work.
1.    What is the physical meaning of the PPD parameter (in equation 6), and how can this parameter be calculated from the DEM (see line 251)?
2.    How the original TRMM data resolution was converted to 1 km resolution, i.e. improved by an order of magnitude? Is not, in the opinion of the authors, such a procedure a source of large additional errors? Why did this practically not manifest itself in the corresponding quality indicators (see table 2)? Does this mean that the fields of mean monthly precipitation are actually quite smooth, and their spatial features are well described at resolution scales of about 25 km? How, then, accounting for small relief features can increase the accuracy of precipitation estimates with statistical significance?
3.    In Table 3 and Fig. 8, the authors give the absolute and relative errors averaged over all stations. In order to judge the statistical significance of the detected differences, it is necessary, at least, to calculate (according to the considered sample) and present the RMSD of the corresponding values in the table and on the graph, or better, the confidence intervals of the MRE and MAE values at a certain selected probability level.
4.    On what basis do the authors choose the weights 0.2, 0.2, and 06 in equation 7? Is there any justification for this choice? It should also be noted that “inverse proportionality” to the value of x is equivalent to direct proportionality to the value of 1/x, and not to the value (1 – x).
In addition, some technical flaws in the manuscript should be noted:
1.    Lines 10 - 12 and 175 - 177 are obviously taken from the article template and do not apply to the text of the manuscript.
2.    The expression "boundary pollution" (line 179) should probably be understood as distortions at the edges of the computational grid. However, the authors, mentioning this problem, do not describe their solution in any way. The end of this sentence is not logically connected with its beginning.
3.    There are typos in the equations and comments to them (confusion with capital and lowercase letters Mean / mean, H / h). The condition in equation 3 is k = 0, not 1. And then not k > 1, but k > 0.
4.    The text in lines 333 – 343 is completely repeated below in lines 344 – 354.
5.    I am not sure about the correct use of the word "staggered" in the context of multiscale features of the DEM.

Author Response

Please refer to the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

General Comments:

The authors have revised the paper as suggested by this reviewer. The paper can go to publication.

 

Reviewer 2 Report

I think that in the present form the paper can be recommended for publishing