Relationship between Vegetation and Soil Moisture Anomalies Based on Remote Sensing Data: A Semiarid Rangeland Case

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

 

In this study, the authors further understand the relationships between vegetation and SM content indices in semiarid rangelands as a complex agricultural dynamic system , and study the feasibility of using anomalies in SM (measured by the Z_WCI) as an advanced warning index to predict anomalies in the vegetation activity, measured by the  Z_WCI. The following aspects can be further explored:

1.It is mentioned in the manuscript that the probability of WCI anomalies and VCI anomalies exceeding different thresholds is calculated every 10 days, and three thresholds, -0.5, -0.7, and -1, were selected based on the thresholds used by the Standard Precipitation Index (SPI) and Standard Precipitation Evapotranspiration Index (SPEI), which are commonly used in drought monitoring. The manuscript lacks a detailed explanation of why the thresholds of -0.5, -0.7 and -1 were chosen.

2. The manuscript suggests that an average increase of 20-30% in the predictability of vegetation anomalies, knowing moisture soil anomalies 4 lags in advance, how is the accuracy of this prediction verified? Are there other data or methods that can be used to validate the results of the study?

3.The conclusion section of the manuscript summarizes the results of the study without clearly stating the significance of these findings for guiding drought monitoring and management on rangelands in semi-arid regions. In addition, the limitations of remote sensing technology are mentioned, but no alternative approach to address this issue in the future is described.

4.Some of the pictures in the manuscript are unclear and poorly readable; the formatting of formulas and other elements needs to be adjusted again.

 

Author Response

Authors: We are very grateful to the reviewer for the comments and suggestions to improve the study.

First of all, we regret to report a mistake in figures 3 and 4. The WCI series for Bajo Aragon was wrong in figure 3 and consequently the cross-correlation graph in figure 4. We have replaced both figures.

Comment 1: It is mentioned in the manuscript that the probability of WCI anomalies and VCI anomalies exceeding different thresholds is calculated every 10 days, and three thresholds, -0.5, -0.7, and -1, were selected based on the thresholds used by the Standard Precipitation Index (SPI) and Standard Precipitation Evapotranspiration Index (SPEI), which are commonly used in drought monitoring. The manuscript lacks a detailed explanation of why the thresholds of -0.5, -0.7 and -1 were chosen.

Response 1: Thank you for pointing this out. The choice of the thresholds is based on the following table included in the reference [56]:

[56] Ma, Q., Li, Y., Liu, F., Feng, H., Biswas, A., and Zhang, Q.: SPEI and multi-threshold run theory based drought analysis using multi-source products in China. J. Hydrol., 616, 128737, https://doi.org/10.1016/j.jhydrol.2022.128737, 2022.

SPEI is a standardized index, with average equal 0 and standard deviation equal 1, as Z_VCI and Z_WCI used in this study. Anomalies begin to be considered when they exceed the first threshold -0.5 (mild dry). When they exceed -1, they are considered moderate, severe or extreme (more than one standard deviation). We have included an intermediate value between -0.5 and -1 (-0.7) to consider intermediate situations.

Comment 2: The manuscript suggests that an average increase of 20-30% in the predictability of vegetation anomalies, knowing moisture soil anomalies 4 lags in advance, how is the accuracy of this prediction verified? Are there other data or methods that can be used to validate the results of the study?

Response 2: Thank you for pointing this out. Thank you for pointing this out. The results obtained are based on historical values ​​of the VCI and WCI anomaly indices. It is a probabilistic study on events that occurred between 2002 and 2019. Therefore, this study is extrapolating the behaviour of anomalies in future years. The obtained probabilities can be improved if new real values ​​are included.

Comment 3: The conclusion section of the manuscript summarizes the results of the study without clearly stating the significance of these findings for guiding drought monitoring and management on rangelands in semi-arid regions. In addition, the limitations of remote sensing technology are mentioned, but no alternative approach to address this issue in the future is described.

Response 3: We agree with this comment. We have added a new paragraph in Conclusion section reinforcing the importance for farmers of having an advance warning index. We have also added some additional explanations to the limitations of using remote sensing.

New paragraph added in conclusion section: " Z_VCI is one of the indices used to monitor vegetation growth anomalies. Additional monitoring of the Z_WCI index allows predicting VCI anomaly events in advance and farmers can make early decisions about their rangelands."

New lines added in conclusion section: "Temporal lengths and pixel resolution are limited by satellite data availability, a common problem in remote sensing. Exploring new satellite data with better temporal and spatial resolution could help to improve and strengthen the conclusions of this study".

Comment 4: Some of the pictures in the manuscript are unclear and poorly readable; the formatting of formulas and other elements needs to be adjusted again.

Response 4: We agree with this comment. We have noticed that some images have lost quality when converting from Word to PDF. In any case, we have changed some images to increase their quality. For example, figures 3 and 4 are new. We hope to improve the quality of the images in the editing stage.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This paper presents a case study on semiarid rangelands, exploring the relationship between the vegetation index, water condition index, and soil moisture component. The primary objective is to enhance the predictability of vegetation anomalies by identifying soil moisture anomalies with a lead time of four lags. The paper is well-organized and provides a detailed analysis of the two study sites. I recommend publication with minor revisions. Below are some of my comments.

 

1. For the two study areas, the majority is covered by crops, with about 20%-30% covered by forest. Is there any irrigation system in the cropland areas? If so, the behavior between these two land cover types could differ significantly. I suggest separating these two areas for analysis.
2. All indices are derived from optical sensors, which may be affected by cloud and atmospheric conditions. The time series data are averaged over approximately one week. How many data points are used to calculate this average, and how representative is the data for the given period? Is there an uncertainty estimate for the indices derived from the satellite data? In other words, how accurate is this data?

Author Response

Authors: We are very grateful to the reviewer for the comments and suggestions to improve the study.

First of all, we regret to report a mistake in figures 3 and 4. The WCI series for Bajo Aragon was wrong in figure 3 and consequently the cross-correlation graph in figure 4. We have replaced both figures.

Comment 1: For the two study areas, the majority is covered by crops, with about 20%-30% covered by forest. Is there any irrigation system in the cropland areas? If so, the behavior between these two land cover types could differ significantly. I suggest separating these two areas for analysis.

Response 1: Thank you for pointing this out. For the two study areas, a rangeland pixel selection was carried out. Finally, 621 and 1952 pixels defined the Los Vélez and Bajo Aragón areas, respectively. This selection was done by Tragsatec Company in collaboration with Entidad Nacional de Seguros Agrarios (ENESA).

Comment 2: All indices are derived from optical sensors, which may be affected by cloud and atmospheric conditions. The time series data are averaged over approximately one week. How many data points are used to calculate this average, and how representative is the data for the given period? Is there an uncertainty estimate for the indices derived from the satellite data? In other words, how accurate is this data?

Response 2: Thank you for pointing this out. All indices were built using frequency bands measured by optical sensors in satellites as TERRA or AGUA. The products used in this study were MOD09GQ, MYD09GQ and MOD09Q1.006. In the case of MOD09GQ, MYD09GQ products, the temporal resolution is daily. For this study the temporal resolution was reduced to 10 days using a composite method. The reason is these data were used in the Spanish satellite-based insurance in pastures. The design of this insurance needed a temporal resolution of 10 days. In the case of MOD09Q1.006, the temporal resolution is 8 days. For each pixel, a value is selected from all the acquisitions within the 8-day composite period. The criteria for the pixel choice include cloud and solar zenith.

For clarity, we have added the text: (10-days composite period) and (8-days composite period) in the Data Collection sections.

For MOD09GQ product, you can look for more information at:

https://lpdaac.usgs.gov/products/mod09gqv006/

For MYD09GQ product, you can look for more information at:

https://lpdaac.usgs.gov/products/myd09gqv006/

For MOD09Q1.006 product, you can look for more information at:

https://lpdaac.usgs.gov/products/mod09q1v006/

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Reviewer1:

The paper titled "Relationship between vegetation and soil moisture anomalies based on remote sensing data: A semiarid rangeland case" investigates the connection between vegetation and soil moisture in semiarid regions of Spain using MODIS satellite data from 2002 to 2019. The study calculates vegetation and soil moisture indices (VCI and WCI) and explores whether soil moisture anomalies can serve as an early warning for vegetation droughts. The results indicate that soil moisture anomalies can predict vegetation anomalies, particularly during cooler months (November to February), with a predictability increase of 20-30%. The study suggests that remote sensing data combined with early warning indices can effectively manage semiarid rangelands. However, the article is not innovative enough, and the content needs to be improved.

 

Comments:

1. Why was the OPTRAM model chosen for estimating soil moisture over other available models? Are there alternative models that could be considered?

 

2. How might the choice of a 10-day temporal resolution for the analysis affect the study's findings? Could different temporal resolutions lead to different results?

 

3. In addition to Z-scores, could other statistical tests or models, such as time series forecasting models, have been applied to strengthen the analysis?

 

4. How was the choice of a four-period lag in the conditional probability analysis determined? Were other lag periods considered, and if so, what were the results?

 

5. The results indicate an increase in the predictability of vegetation anomalies during certain months. What might be the underlying reasons for this seasonal variation?

 

6. How do the results from the study regions in Spain compare to other semiarid regions globally? Could a comparative analysis provide broader context to the findings?

 

7. The form of the picture in the article is extremely simple. Could some of the figures, such as the box plots, be improved with more explicit labels and descriptions to enhance readability?

 

8. Could the study discuss potential methods for validating the findings through ground-truthing or other observational data?

 

9. How does seasonal variability influence the relationship between soil moisture and vegetation in different climatic conditions, and could this be explored further?

 

10. What are the limitations of the OPTRAM model, and how might they affect the study's findings?

 

11. Consider expanding the comparative analysis to include other semiarid regions globally to provide a broader context for the findings.

 

12. Explore the possibility of adding additional statistical models or validation methods to strengthen the analysis and confirm the robustness of the findings.

 

Author Response

Authors: We are very grateful to the reviewer for the comments and suggestions to improve the study.

First of all, we regret to report a mistake in figures 3 and 4. The WCI series for Bajo Aragon was wrong in figure 3 and consequently the cross-correlation graph in figure 4. We have replaced both figures.

Comment 1: Why was the OPTRAM model chosen for estimating soil moisture over other available models? Are there alternative models that could be considered?

Response 1: Thank you for pointing this out. OPTRAM model was chosen following the results presented in the references [24,25]:

In the first reference we can read: “In conclusion, OPTRAM can estimate temporal soil moisture dynamics with reasonable accuracy for a range of climatic conditions (semi-arid to humid), soil types, and land covers, and can potentially be applied for agricultural drought monitoring.”

In the second reference we can read: “Results indicate that the prediction accuracies of OPTRAM and TOTRAM are comparable, with OPTRAM only requiring observations in the optical electromagnetic frequency domain…We also demonstrate that OPTRAM only requires a single universal parameterization for a given location, which is a significant advancement that opens a new avenue for remote sensing of soil moisture.”

Both papers compared OPTRAM to other alternative models. Of course, OPTRAM have advantages and disadvantages, but we have chosen OPTRAM for this study due to its reasonable accuracy and because only requires a single universal parameterization for a given location.

[24] Babaeian, E., Sadeghi, M., Franz, T. E., Jones, S., and Tuller, M.: Mapping soil moisture with the OPtical TRApezoid Model (OPTRAM) based on long-term MODIS observations. Remote Sens. Environ., 211, 425–440. https://doi.org/10.1016/j.rse.2018.04.029, 2018.

[25] Sadeghi, M., Babaeian, E., Tuller, M., and Jones, S. B.: The optical trapezoid model: A novel approach to remote sensing of soil moisture applied to Sentinel-2 and Landsat-8 observations. Remote Sens. Environ., 198, 52–68. https://doi.org/10.1016/j.rse.2017.05.041, 2017.

Comment 2: How might the choice of a 10-day temporal resolution for the analysis affect the study's findings? Could different temporal resolutions lead to different results?

Response 2: Thank you for pointing this out. For this study the temporal resolution of the products MOD09GQ, MYD09GQ was reduced to 10 days using a composite method. The reason is because these data were used in the Spanish satellite-based insurance in pastures. The design of this insurance needed a temporal resolution of 10 days.

One of the effects of using more temporal resolution is to increase the prediction accuracy. In this study we have found an improvement in the prediction four 10-days periods in advance. If the temporal resolution increases, we could define better the interval of prediction.

Comment 3: In addition to Z-scores, could other statistical tests or models, such as time series forecasting models, have been applied to strengthen the analysis?

Response 3: We agree with you, time series forecasting models could be another way to tackle the issue. We are thinking of preparing another study with this scope.

Comment 4: How was the choice of a four-period lag in the conditional probability analysis determined? Were other lag periods considered, and if so, what were the results?

Response 4: Thank you for pointing this out. In this conditional probability analysis we obtained results for different lags from 0 to 6. We noticed that predictability augmented when increasing lags in most of the periods. From 4 and up, predictability maintains stable in most of the periods.

If you consider it appropriate, we could add this analysis in an annex so as not to detract from the clarity of the main study.

Comment 5: The results indicate an increase in the predictability of vegetation anomalies during certain months. What might be the underlying reasons for this seasonal variation?

Response 5: Thank you for pointing this out. We have commented in the Discussion section: ”From October to January, precipitation is more or less abundant in both areas but is consistent with low temperatures. Therefore, this could explain why the predictability of vegetation anomalies using the SM content index is improved from November to the beginning of February. When precipitation continues to fall, but temperatures begin to rise, from February to the end of April, the predictability starts to decrease due to temperature and evapotranspiration playing a major role in SM content [59,60]”

Therefore, if we had to summarize the reasons for this seasonal variation we said temperature, precipitation and phenological status of the pasture.

Comment 6: How do the results from the study regions in Spain compare to other semiarid regions globally? Could a comparative analysis provide broader context to the findings?

Response 6: We agree with you, other global semiarid regions could be analysed. This methodology could be applied to any region. The results obtained in other semiarid regions could be different depending on climate conditions.

Comment 7: The form of the picture in the article is extremely simple. Could some of the figures, such as the box plots, be improved with more explicit labels and descriptions to enhance readability?

Response 7: We agree with this comment. We have noticed that some images have lost quality when converting from Word to PDF. In any case, we have changed some images to increase their quality. For example, figures 3 and 4 are new. We hope to improve the quality of the images in the editing stage.

Comment 8: Could the study discuss potential methods for validating the findings through ground-truthing or other observational data?

Response 8: Thank you for pointing this out. The question is important because we are trusting that indices reflect the truth. But the scope of this study is to use accepted indices. In this study we have chosen reliable indices as NDVI and W to consider vegetation health and soil moisture. In future studies we are thinking to use other indices that estimates better biomass.

Comment 9: How does seasonal variability influence the relationship between soil moisture and vegetation in different climatic conditions, and could this be explored further?

Response 9: We agree with you, seasonal variability influences the relationship between soil moisture and vegetation in different climatic conditions. The methodology used in this study could be applied to other regions with different results depending on climatic dynamics. In future studies this complex relationship could be explored and interesting conclusions could be obtained.

Comment 10: What are the limitations of the OPTRAM model, and how might they affect the study's findings?

Response 10:  Thank you for pointing this out. In the reference [25] we can read in the Conclusion section:

”The disadvantage of OPTRAM when compared to TOTRAM is its higher sensitivity to oversaturated and shadowed pixels. When the optical trapezoid consists of too many oversaturated pixels, solving for the wet edge needs some refinements. This, however, may not be a significant limitation because of the feasibility of a single universal model parameterization.”

“Previous studies indicated that SWIR reflectance is not only sensitive to the leaf water content, but also to the leaf internal structure. Hence, combining the SWIR signal with an NIR band (primarily sensitive to the leaf internal structure) has been suggested to minimize the uncertainty in retrieving vegetation water content. This idea may be followed to reduce the site-dependency of OPTRAM parameters.”

In any case, OPTRAM model was chosen for its promising results compared to other methods to estimate soil moisture.

[25] Sadeghi, M., Babaeian, E., Tuller, M., and Jones, S. B.: The optical trapezoid model: A novel approach to remote sensing of soil moisture applied to Sentinel-2 and Landsat-8 observations. Remote Sens. Environ., 198, 52–68. https://doi.org/10.1016/j.rse.2017.05.041, 2017.

Comment 11: Consider expanding the comparative analysis to include other semiarid regions globally to provide a broader context for the findings.

Response 11: Thank you very much for your suggestion. In next studies we are considering using this methodology to other regions, not only semiarid but other climatic regions.

Comment 12: Explore the possibility of adding additional statistical models or validation methods to strengthen the analysis and confirm the robustness of the findings.

Response 12: Thank you very much for your suggestion. In further studies we want to compare conditional probability methodology to time series forecasting. Combination of both methods will increase the quality of predictions.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

As for the proposed modification request, the author has made detailed modifications. Accept in present form

Author Response

Comment: As for the proposed modification request, the author has made detailed modifications. Accept in present form.

Response: We are very grateful to the reviewer for all the comments and suggestions to improve the study. In this version of the manuscript, we have improved the quality of figures 5 and 6.

Reviewer 3 Report

Comments and Suggestions for Authors

The author has made many changes according to the suggestions. However，I think many of the pictures in the article can be further beautified and improved. For example, the font of many pictures is small and not clear enough.

Author Response

Authors: We are very grateful to the reviewer for all the comments and suggestions to improve the study.

Comment: The author has made many changes according to the suggestions. However, I think many of the pictures in the article can be further beautified and improved. For example, the font of many pictures is small and not clear enough.

Response: Thank you for pointing this out. In this version of the manuscript, we have changed the quality format of most of the images. We hope that the visualization has improved. We have also modified the fonts of figures 5 and 6. We agree with your comment. They were very small and the visualization was difficult.