Quantitative Evaluation of Soil Water and Wind Erosion Rates in Pakistan

Round 1

Reviewer 1 Report

The author must  mention timeframe of the data used (start to end year) in the model eg  wind speed, precipitation, temperature etc

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

1.         In order to solve the coarse resolution of RUSLE and RWEQ, this study uses GIS and machine learning to predict soil water erosion rates.

2.         The topic is original in the field.

3.         Similar macroscopic patterns of SAER and SIER with other published studies without verification of field investigation is the weak point of this study.

4.         It is difficult to get satisfied resolution of input parameters for predicting soil or wind erosion rates so that the machine learning technique might solve this weakness. The contribution of this study could not be endorsed without any field investigation.

5.         The authors are suggested to add some description about how to use data sources with coarse resolution in this study and discuss the potential errors.

6. With the overlay map of SAER and SIER, the authors showed three risk regions in Pakistan. This information is very important for decision maker during planning national development.

7. Data points scattered in Figure 7 so that the regression equations are not suggested.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

In this study, a soil erosion rate caused by water and wind in Pakistan was evaluated and related results was represented. The data of this research were rich and method was scientific. And the results were reliable. I think this research can be published with major revised.

Special comments

1. Section of 'Spatial prediction of soil water erosion' is too much detail. And most of them are common-sense. I think authors can only show the main parts and special parts and the rest part can be moved to the appendix.

2. Figure 7 the figures seems no significant linear relationship between X and Y. Does these results can pass the significant test? Please show related results and P value.

3. Please clearify the data resourece of the sampling survery points showed in figure 2. And express how to use these data.

4. Figure 13 and 14, I think the line of y=x will be more valuable compared with fitted function showed in figure.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 4 Report

General comments

The objective of the study is to predict the soil water erosion rate (SAER) using a combination of (Geographic Information Systems (GIS) and machine learning) and evaluated the soil wind erosion rate (SIER) using Revised Wind Erosion Equation (RWEQ) model. Finally, the author concluded that Potohar Plateau and its surrounding regions were mostly impacted by water erosion, have a soil erosion rate of 2500–5000 t•km^-2•a^-1, while wind erosion predominated in the Kharan and Thar desert, with a soil erosion rate exceeding 15000 t•km^-2•a^-1; the Sulaiman and Kirthar mountain ranges were susceptible to wind and water compound erosion, which was more than 8000 t•km^-2•a^-1.

The major issues that should be addressed to make the paper better are as follows:

·       The research is based on previous survey studies, but the authors did not give enough information about this survey. What is the survey about? What information the survey compiles? Why is the research based on this survey? I suggest having a separate section about the survey and provide enough brief information about it.

·       The research used machine learning of RF. But the authors have not given enough info about the modelling process. How much of the data used for training and how much for testing? There is no enough explanation how the Random Forest modelling approach was in terms of setting parameters and finding the best fit?

·       There are contradicting ideas within the paper. The soil water erosion is higher from the forest land use but the authors are suggesting as remedial measures on sloping croplands to change its land use to forest and grassland while those areas are high soil erosion producing areas? reason out why those areas are producing high erosion? I don’t think it is only sparsely forested as justified in the paper. Are there landslides from the area because of the forest and saturated soil during storm?

·       Section 4.4.1 shall be part of introduction, the introduction lacks what has been done before and what could be better as justification of this research work. The reader should get the new intended work in the introduction not at the end of the manuscript.

Minor comments

·       In the introduction section line 45 the author mentioned $400 amount of lose due to wind erosion.  Can you check the order of magnitude of the number?

·       Page 2 line 67, be consistent in units. Change 1000m to 1km

·       Page 10 line 178, are you talking about organic carbon loss?

·       Figure 5 and Figure 6, improve the caption by explaining about s1 and s2. In the other figures, they are also indicated but given explanation. Please give a better explanation of all figures in their captions.

·       Page 11 line 295; ‘some regions have soils’, what kind of soils are you talking about?

·       Figure 13 and 14, use only two digits after decimal

·       In result section Line 282 it say 5–2000 t•km-2•a-1 in forests it need numerical correction

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

The comments I pointed out has been reivised and I have any other comments to authors.