Data-Driven Approach to Assess Spatial-Temporal Interactions of Groundwater and Precipitation in Choushui River Groundwater Basin, Taiwan

Round 1

Reviewer 1 Report

This is a very interesting paper and comprehensive analysis using some common methods for analyzing groundwater levels. I think the paper could be published at some point, but I have some concerns and would like these to be addressed first: 

• General Statement: This is a long paper that could use some shortening to help with readability and overall impact. This includes reducing the number of figures and also the number of plots within the figures. For instance, figures 7 and 10 has a lot of subfigures inside of it. I would encourage the authors to think about what are the most important ones to present. Some of the methods may not needed to be completely described (i.e., SCS Curve numbers). A reference is fine.
• Page 4, Section 2.3: What is the resolution of the MODIS data used for land use classification? Are CNs calculated for each of these rasters?
• Page 7, Section 2.3.5: How applicable is SCS method and CNs for international studies given they were developed for U.S. soils.
• Page 8, Section 3.1: There are very large differences in MoP shown in Figure 2. What is the explanation for these differences.
• Page 11, Section 3.4: Cross-validation results are shown for the two kriging methods. Why is this same analysis not done for regression in the previous section?

 

 

 

Author Response

We have attached the MS Word file to list the response to the comments raised by Reviewer #1. Thanks for the valuable comments. 

Author Response File: Author Response.docx

Reviewer 2 Report

In my opinión, the manuscript “Data-Driven Approach to Assess Spatial-Temporal Interactions of Groundwater and Precipitation in Choushui River Groundwater Basin, Taiwan” is very interesting. The subject discussed in the manuscript, related to the availability of water resources, is a current issue and requires research at various scales and from different approaches, so it seems a wise choice.

In my opinion, the introduction does not fully reflect the complexity of the research work carried out, especially with regard to the handling of different time scales. I would recommend reviewing it in depth to refer to all the aspects covered in the work.

The design of the research is appropriate, the methods are adequately described and the results are clearly presented.

Although the conclusions are supported by the results, in my opinion some conclusions about the methodology used and its possible limitations are missing.

The cartography included in the manuscript as figures, although it is relevant, presents a highly improvable design. Some aspects to check on all maps are:

• the coordinates grid is unnecessary, as these maps are not to be used in the exact location of items. If included, it would be sufficient to indicate the coordinates in the corners of the map, at a much smaller size than the current one.
• In most of the maps, legend, scale and north arrow are too large, in many maps larger than the map itself

In figure 1, I would recommend including the location of Jiji weir, which is mentioned several times in the text and whose location is unknown to those of us who are not familiar with the study area. Also, I recommend to modify the layout of the text that indicates "proximal fan", “mid fan” and “distal fan”, arranging it inside the area to which it refers, since it is not easily identified. Also, I wonder if the Choushui River does not have any tributaries or they simply have not been represented on the map, as it is done with those of the Peikang River. Another small change that would improve the readability of the map would be that the labels that fall on the coloured areas (for example, the identifiers of the groundwater stations or the labels "proximal fan", etc.) were represented with a white buffer.

Some other specific aspects to review would be:

• at the bottom of figure 2, it says “b) the MoP data calculated by adding the original CHIRPS data with the interpolated residuals obtained from the differences between the rain gauge stations and the CHIRPS data.” I think it should be “…. obtained from the regression analysis between the rain gauge stations and the CHIRPS data”.
• in section 3.2.1, line 336, where it says “… that the land surface elevation can describe 97 % GWL variation in the study area “, it would say “…that the land surface elevation can describe 95 % GWL variation in the study área”. It is R-squared the parameter that indicates the percentage of explained variance by a regression model, and not the correlation coefficient (r).
• for the results mentioned in sections 3.2.2. and 3.2.3 it would be useful to present a table that includes the values ​​of r, R-squared and p-value for each of the models.
• at the bottom of figure 6 it should be stated that the graphics represent average values or r, RMSE and NMSE for each month.
• at the bottom of figure 8 it would be a good idea to mention the land Surface elevation of each of the three monitoring stations represented.
• In line 542, when it says “… by accumulating all data sets from 2006 to 2015”, ¿should it says “… by averaging all data sets from 2006 542 to 2015”? At the bottom of figure 9, it says “(a) the annual average..”.
• In line 549-550, it says “There are also negative correlation 549 areas located along the Choushui River in the center of the CRGB and along the Peikang River…”. I cannot see this negative correlation; in fact, most of the area shows positive correlation.
• In the comments of figure 10, nothing is said about the month of June, which is clearly out of the general pattern or trend.
• In lines 571 and 572, ¿should it be event-based instead of even-based? The same in line 596 (flood event, instead of flood even).
• the source of data mentioned in section 3.7 (ERA-Interim) should be included in section 2.2
• In section 4, lines 612 to 614 (“The linear regression analysis showed that the precipitation and DEM data could 612 not be included together as the additional variables for the GWL interpolation because of the obtained 613 multicollinearity condition.”) should appear before lines 610-612 (“With the DEM 610 data included in the RK, the interpolated GWL showed considerable improvement in spatial and 611 temporal domains.”), as this analysis was performed previously, precisely to decide which variables to use in RK.

Author Response

Thanks for the valuable comments raised by Reviewer #2. We have attached the MS Word file to list the point-by-point response. Please see the attached files for details.

Thank you.

Chuen-Fa Ni

 

Reviewer 3 Report

dear author

The paper is a good contribution but you need to update the rationale and the discussion

The figures are poor and need improvements

In general the paper can be published after some minor review

Sincerely

 

Comments for author File: Comments.pdf

Author Response

The comments from the third reviewer might be because of the system error. The comments shown in the .pdf file are the same as those provided by Reviewer #2. We have listed the comments and modified the manuscript based on the suggestions. The attached WORD file shows the point-by-point responses to the comments from Reviewer #3.  Thanks for the valuable comments.

 

Chuen-Fa Ni

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

• My initial comment was that the paper is too long. The revised paper is now longer. It was originally 24 pages and the revised version is 26 pages. I am concerned that the authors did not take this comment from myself (and others) seriously.
• Comment 2: I still feel strongly the suggestions on these figures is valid to shorten the paper.
• Comment 3: I did not notice a reduction in the explanation of the methods for SCS.
• Comment 6: the response to me as reviewer is fine. I think the authors need to not the applicability of the method in the manuscript
• Comment 7: the response to me as reviewer is fine. I think the authors need to not the applicability of the method in the manuscript

Author Response

We thank reviewer #1 for the valuable comments. We have modified the manuscript based on the provided suggestions and highlighted the changes in the manuscript. Specifically, the length of the manuscript was significantly reduced, and some figures were either removed or merged to make the presentation clear. The new version includes 11 figures, and the number of pages is 22, including the list of references. We have attached a list of point-by-point responses to the comments and concerns.

Thank you for your time.

 

Chuen-Fa Ni, Ph.D.,

Professor at Graduate Institute of Applied Geology,

National Central University,

Taoyuan City, 32001, Taiwan.

Email: [email protected]

Tel: 886-3-4227151 ext. 65874

Author Response File: Author Response.docx

Round 3

Reviewer 1 Report

No concerns and thanks for the revisions.