Tracking Changing Evidence of Water Erosion in Ordos Plateau, China Using the Google Earth Engine
Round 1
Reviewer 1 Report
The manuscrtipt requires the following minor improvements:
- in lines 17 and 18 it would be better to show unitary values, or at least to use “soil loss” instead of soil erosion and to use Mg instead of t,
- in line 62 the dot between on and [17] should be removed,
- in line 92 date format should be January 01, 2013 just like in line 93,
- in line 97, after clay the word separates or fractions should be regarded
- symbols in equations 1-5 should be explained,
- in lines 128, 147 and 152 should be “RUSLE” instead of “RSLEI” and “RULSE”
- it would be better to use everywhere unit for area km2, and Mg for mass,
- in line 210 “in” before bare soil should be added,
- title in Table 3 should start by capital letter,
- in lines 45, 228 and 229 should be “RUSLE” instead of RULSE”,
- in Table 3 and in line 213 the used unit is not clear, it would be better to use: Mgˑkm-2ˑyr-1,
- title of vertical axis in Fig. 7 should be: “Total amount of soil loss”, and the unit Mg instead of t,
- some titles of sub-points start from lower case letter (3.2.1, 4.1, 4.2, 4.5 and 5.1).
Author Response
Please see the attachment.
Author Response File: 
Author Response.docx
Reviewer 2 Report
The article describes the application of the RUSLE erosion model in the Ordos region (China) using the Google Earth Engine (GEE) platform. Although the RUSLE concept is not new, the idea of applying GEE data and procedures is interesting and, if it works, it can facilitate the quantification of soil loss in areas with a lack of other data. As such, the article fits the scope of the journal. However, the description of the methodology is incomplete and unconvincing, so it is not possible to verify its correctness. As a result, it is not clear whether the conclusions of the study are correct, and it would also be difficult to reproduce its results.
The Introduction section is relatively short and does not provide a concise picture of the current state of soil erosion modeling. There are very few examples of the application of the RUSLE model using satellite data. An overview of other methods of measuring or modeling erosion losses is also lacking (e.g., WEPP, SWAT, field measurements). For example, the following articles provide an overview of soil erosion modeling research:
https://doi.org/10.1016/j.scitotenv.2021.146494
https://dx.doi.org/10.22161/ijaers.4.12.22
In the methodology part, the key formula of the RUSLE model and the explanation of its parameters are absent. It is also not clear from which months of the year the Landsat images were selected. If from the whole year, then the NDVI calculation can be strongly influenced by the availability of cloudless scenes in different years. Moreover, the calculation of the annual NDVI does not make much sense, as the NDVI fluctuates significantly throughout the year. Since precipitation occurs in the studied area primarily from June to August, it would be more logical to apply the analysis in this time frame.
When describing the Random Forest classification, basic information is missing, such as the number of sampling points and the classification accuracy. It would also be appropriate to explain in more detail the derivation of the C-factor from the NDVI and the P-factor from the land cover layer. In fact, it is not possible to estimate the supporting effects of practices such as contouring, strip-cropping, and terraces from Landsat images with a pixel size of 30m.
Correct statistical methods were chosen to assess erosion trends, but there is a lack of citation of formulas 1-5 or at least description of variables. The sample size is also unknown. Model validation is essential to evaluate research results. However, the sediment transport modulus, which was used as reference data, is not described or cited. I am also not convinced of the correctness of the resulting erosion rate values. In Table 3, as well as in the text, the values are given in unclear units (t/(hm^2*a)).
The Discussion section is limited to the explanation of erosion trends, and the results and statements are not confronted with the findings of other similar studies.
Minor and formal errors:
Sections 3 and 4 have the same title: Materials and Methods The same applies to Sections 3.2.3 and 4.1.
For better understanding, the Ordos study area could be referred to as a region or cadastral territory rather than a city.
Figure 4 - Different years are mentioned in the description than in the maps. The threshold values between erosion categories are not defined either in the figure or in the text.
Figures 5 – 4 of 9 maps are identical to maps in Fig.4.
Concerning the language, I recommend reducing the use of simple verbs (is, are ...) and avoiding the distinction between domestic and foreign (home and abroad, foreign scholars …).
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 3 Report
The article “Tracking changing evidence of soil erosion control in Ordos Plateau using the Google earth engine”, deals with soil erosion in the Ordos Plateau (Mongolia). In the introduction section, the authors establish the area of study, which should go to material and methods, this introduction section is somewhat poor, we advise the authors to expand with similar studies from other territories.
In the material and methods section, a subsection 3 and another 4 are made, I understand that subsection 4 is an error, which should correspond to results.
The legends of figures 3, 4 and 5 should be improved.
In Agronomy there are different articles on plant covers that connect with soil erosion, the authors can refer to them in their discussion and they would enrich it.
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Round 2
Reviewer 2 Report
I thank the authors for acknowledging and incorporating comments. The article has been improved. Nevertheless, a few comments remain.
Methods
The calculation of the P-factor is still not clear. What land use categories were classified by the RF method, and how were these categories reclassified to the P-factor?
Perhaps it would be appropriate to add maps of individual factors in a higher resolution to the Supplementary Materials.
Results
It would seem more logical to me to present first the erosion stages (section 4.3) with maps from individual years (fig. 5), and then the erosion changes (section 4.2, fig.4). However, I leave the final decision up to you.
Figure 4 - Different years are still mentioned in the description (1998, 2005, 2012, 2018) than in the maps (2013, 2016, 2019, 2021).
Fig.6 - I would choose the opposite color scale, the decrease in erosion should be shown in green and the increase in red.
Discussion
A comparison of the results with the findings of other studies is still lacking. This would need to be supplemented, especially considering that several studies focusing on the temporal-spatial changes of soil protection in China have been published recently:
https://www.mdpi.com/2071-1050/14/10/5829/htm
https://www.tandfonline.com/doi/full/10.1080/19475705.2020.1861112
https://www.sciencedirect.com/science/article/pii/S1470160X22004174
An advancement of your article compared to the above is the calculation through the GEE, so I would describe and emphasize this process in more detail in the Methodology.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
