Spatial Distribution of Soil Moisture in Mongolia Using SMAP and MODIS Satellite Data: A Time Series Model (2010–2025)

Round 1

Reviewer 1 Report

December 30, 2020

Manuscript: ‘Spatial Distribution of Soil Moisture in Mongolia Using SMAP and MODIS Satellite Data: A Time Series Model (2010–2025)’

In this paper the Soil Moisture Active Passive (SMAP) and Moderate Resolution Imaging Spectroradiometer (MODIS) data sets were used to estimate the distribution of soil moisture and applied a comparison with the monthly satellite-derived precipitation/temperature and crop yield from 2010 to 2020 over Mongolia (study area). Unlike previous studies, an optimized ARIMA model was used for soil moisture forecasting until 2025 in the study area. In situ soil moisture observations at different depths from the Information and Research Institute of Meteorology, Hydrology and Environment (IRIMHE) of Mongolia were used as a benchmark data to validate the effectiveness of the ARIMA models. This is a relevant topic lies within the scope of the MDPI remote sensing journal. The article is well organized and neatly written with the appropriate scientific content. Based on the above, I support the publication of this manuscript, but only after a minor revision.

********************************

Title: it fits perfectly the paper content. 

Abstract: it is quite adjusted to the paper content.

Line 26: for the sake of clarity, please indicate the meaning of ‘SMC’.

Introduction: it provides sufficient background and includes relevant references on the satellite-based techniques to estimate soil moisture, highlighting the main limitations of the different approaches. Objectives and the novelty are clearly presented.

Lines 83-86: this part should be moved to ‘2.1. Study Area’

Study Area and Data Preprocessing: the study area and data sets are effectively described.

Line 178: authors should add the data source for the CRU TS temperature and precipitation as a link (similar to SMAP data at line 157 and MODIS data at line 166)  

Methodology: the methodology is clearly described.

Line 272: for the sake of clarity, the authors should explain how to interpret the Pearson correlation coefficient (r); e.g., ‘r ranges from -1 to 1, with higher values corresponding to better coherence between the inter-compared data sets’

Results: these are clearly presented.

Line 295, figure 295; Line 420, Figure 12: for consistency, use ‘ . ‘ instead of ‘ ,‘ for residual values.

Line 323, Table 4; line 347, Table 5; line 363, Table 6: for a fair comparison, the authors should add other statistical metrics such as the root-mean-square error (RMSE) and bias (B)  

Discussion: this is clearly presented and is supported by results from the previous section.  

Conclusion: these are clear, concise, and are in line with results.

Author Response

Dear Reviewer 1, 

Thank you so much for your valuable comments, and I revised according to your comments. Please see the attachment included responses of your comments in the word file. 

Best wishes,

Authors

 

Author Response File: Author Response.pdf

Reviewer 2 Report

In the manuscript, The author proposed a monthly soil moisture estimation approach that only uses NDVI and land surface temperature (LST). As a reference SM, 9km SMAP product was used. Then the author showed the relation between monthly SM and temperature/precipitation and crop yield. Then A time series forecasting model was developed.
The study is valuable to understand SM distribution in Mongolia and it's relation with agricultural yields. Also, the study showed NDVI and LST are enough parameters for monthly SM estimation. However, the paper needs some improvements.

1) SMmod comparison with SMAP and metrological measurements was performed by calculating the correlation coefficient. However, for SM estimations studies RMSE, ubRMSE, and bias error are also important performance metrics. I advise them to add these metrics to the paper.
2) Whereas the SMmod model shows a better correlation with SMAP estimation, why the performance is very low when compared with in-situ measurements? The author claimed the reason is limited in-situ data. But I think it is not the only reason. Could you add a comparison between SMAP and metrological measurements?

3) How many ARIMA models or combinations were tested before the selected ARIMA(12,1,12)? How different the other model performance from the selected one?

4) The robustness of the ARIMA model is not clear. Because they tested the model the measurements between Dec 2019 and May 2020, but this period is also used for model training or development process. I advise the author to leave that time period from the training part and just use it for testing.

5) Coefficients in Equation 6 and Table 3 are not the same.

6) Generally, there is a problem with figure axis labels. In most of the figures, there is no y left and y right axis label. For figures 5 and 7 x-axis label should be in the below of the figures.

7) why the figure 5a and figure 7a is not a straight line?

 

Author Response

Dear reviewer, 

Thank you so much for your valuable comments and suggestions. 

We revised it according to your comments. 

Best wishes, 

Authors

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

The authors should work on figures more. There is no consistency in the text font size of the figures.

Author Response

Dear reviewer, 

Thank you so much for your comment. 

The figures were revised as requested. 

All the best, 

Authors