Hydrologic Consistency of Multi-Sensor Drought Observations in Forested Environments
, Dean Meason 2,†, Priscilla Corbett-Lad 2,†, Barbara Höck 3,†and Narendra Das 4,5,†
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsReviewer response to the authors.
This manuscript described a study to assess drought events using precipitation, soil moisture, evapotranspiration, and land surface temperature data based on multiple satellite sensors. The results demonstrated that soil moisture observations were better than precipitation satellite observations. This study is very useful for scientists who want to apply it to the indicator for monitoring drought over large regions. The authors performed well-designed research to explore the multi-sensor drought observations in forest environments. The paper was generally well-written and comprehensive. However, there are a few shortcomings that need attention:
1. The abstract section should provide more information of soil moisture for readers to understand. Moreover, the soil moisture should be shown in the keywords.
2. I would suggest you improve Figure 1. such as scale bare and arrow. Additionally, the description of the figure should be provided in more details.
3. Did you use only terra MODIS or aqua MODIS? Or did you use both sensor data? If yes, please indicate it in the MODIS dataset.
4. The authors did not explain statistical methods for identifying the relationship between soil moisture and the other variables. I would recommend you add the information in the methodology section.
5. The main study of your work that explored the hydrologic consistency of multi-sensor data for monitoring drought of forest environments. What do you think about using these multi-sensor data to monitor drought events in highly complex landscapes and cloudy regions?
6. In the results, the spatial distribution of drought occurrence is significant and useful as an essential tool for sustainable forest management and developing mitigation strategies. Also, spatial heterogeneity of droughts in the study region can provide valuable information for the related agencies and scientists. The map results need to be shown and provided in this section.
7. In the discussion, you have to provide more details of the soil moisture based on the Soil Moisture Active Passive (SMAP) satellite. I would suggest you add more information and compare it to the previous study.
8. In this study, the satellite observation datasets have different spatial resolutions. Has this point influenced your study or not?
9. The implemented workflow is missing. I would suggest the authors provide it to show an overview of your study.
Comments on the Quality of English LanguageMinor editing of the English language required
Author Response
> This manuscript described a study to assess drought events using precipitation, soil moisture, evapotranspiration, and land surface temperature data based on multiple satellite sensors. The results demonstrated that soil moisture observations were better than precipitation satellite observations. This study is very useful for scientists who want to apply it to the indicator for monitoring drought over large regions. The authors performed well-designed research to explore the multi-sensor drought observations in forest environments. The paper was generally well-written and comprehensive. However, there are a few shortcomings that need attention:
We thank the reviewer for their helpful comments and suggestions. We have tried to address them as best as possible in conjunction with the comments from the other reviewers.
> 1. The abstract section should provide more information of soil moisture for readers to understand. Moreover, the soil moisture should be shown in the keywords.
We have added "soil moisture" as a keyword, and have made slight modifications to the abstract.
> 2. I would suggest you improve Figure 1. such as scale bare and arrow. Additionally, the description of the figure should be provided in more details.
We have added a scale bar and arrow to Figure 1. We have also modified the captions in most of the figures.
> 3. Did you use only terra MODIS or aqua MODIS? Or did you use both sensor data? If yes, please indicate it in the MODIS dataset.
We compared both Terra and Aqua MODIS products, when available, and the differences at the study sites were minimal. Therefore we used the Terra products in this study, and we have clarified this in the text.
> 4. The authors did not explain statistical methods for identifying the relationship between soil moisture and the other variables. I would recommend you add the information in the methodology section.
We have expanded the description of the Granger causality test to better describe the statistical method of identifying the relationship between soil moisture and other variables.
> 5. The main study of your work that explored the hydrologic consistency of multi-sensor data for monitoring drought of forest environments. What do you think about using these multi-sensor data to monitor drought events in highly complex landscapes and cloudy regions?
That is a very interesting idea and a worthwhile investigation. We have added this as a recommendation for future research.
> 6. In the results, the spatial distribution of drought occurrence is significant and useful as an essential tool for sustainable forest management and developing mitigation strategies. Also, spatial heterogeneity of droughts in the study region can provide valuable information for the related agencies and scientists. The map results need to be shown and provided in this section.
Although we agree with the reviewer that the spatial distribution of drought occurrence is significant, the objective of this study is the assessment of hydrologic consistency of the satellite observations which requires ground measurements. Unfortunately that limits us to evaluating the satellite observations at the locations of the ground measurements and makes the inclusion of satellite imagery beyond the scope of the study.
> 7. In the discussion, you have to provide more details of the soil moisture based on the Soil Moisture Active Passive (SMAP) satellite. I would suggest you add more information and compare it to the previous study.
We have included some additional discussion on the SMAP soil moisture observations and comparison with other studies that estimated drought characteristics using those observations.
> 8. In this study, the satellite observation datasets have different spatial resolutions. Has this point influenced your study or not?
We agree with the reviewer that the different spatial resolutions can affect the comparison of satellite observations with ground measurements, and have highlighted that in the Discussion section.
> 9. The implemented workflow is missing. I would suggest the authors provide it to show an overview of your study.
It is not clear what the format of the workflow would be most helpful, e.g. flow diagram? In terms of implementation we have included the code that performs the analysis as a Jupyter notebook (see Code availability statement).
Reviewer 2 Report
Comments and Suggestions for AuthorsThis paper systematically assesses the capability of remote sensing satellite data in monitoring drought and quantifies its consistency with results from ground-based observations. The study is innovative. However, some of the contents need to be clarified:
Section 2.3 of the paper describes the calculation of a variety of indicators. However, there needs to be some clarification, and it might be helpful to have a table summarizing all the assessment indicators and what they indicate.
L206-L207,” Therefore we applied a time series matching procedure…” The expression is not very clear, suggesting rewriting the sentence.
L213，“… between December and March of each year.” Why was this period chosen, and is there any basis or literature to support it?
L263-L264, “Both satellite datasets perform relatively poorly when compared to the ground measurements…” Satellite data is raster data for a particular area, and ground-based station data corresponds to a smaller area; how do you find the raster data corresponding to the station? Nearest or distance weighted? Does this have an impact on the calculation of R-values?
Comments on the Quality of English LanguageMinor editing of English language required. for example,
L206-L207,” Therefore we applied a time series matching procedure…” The expression is not very clear, suggesting rewriting the sentence.
Author Response
> This paper systematically assesses the capability of remote sensing satellite data in monitoring drought and quantifies its consistency with results from ground-based observations. The study is innovative. However, some of the contents need to be clarified:
We thank the reviewer for their helpful comments and suggestions. We have tried to address them as best as possible in conjunction with the comments from the other reviewers.
> Section 2.3 of the paper describes the calculation of a variety of indicators. However, there needs to be some clarification, and it might be helpful to have a table summarizing all the assessment indicators and what they indicate.
We have added a table with all the indicators calculated.
> L206-L207,” Therefore we applied a time series matching procedure…” The expression is not very clear, suggesting rewriting the sentence.
We have rewritten the sentence.
> L213，“… between December and March of each year.” Why was this period chosen, and is there any basis or literature to support it?
That period is the southern hemisphere's summer, making the potentially driest period and the most relevant to drought conditions. We have added a reference (<https://doi.org/10.2307/26169741>).
> L263-L264, “Both satellite datasets perform relatively poorly when compared to the ground measurements…” Satellite data is raster data for a particular area, and ground-based station data corresponds to a smaller area; how do you find the raster data corresponding to the station? Nearest or distance weighted? Does this have an impact on the calculation of R-values?
This is a general problem when performing comparisons between areal and point observations, and we have mentioned it as a limitation of study such as ours. In terms of finding the raster data, we used the raster pixel which contains the point measurements.
Reviewer 3 Report
Comments and Suggestions for AuthorsThis manuscript (remotesensing-2844934) explores the impact of drought on forests and the use of satellite remote sensing for monitoring. It evaluates various hydrological indicators in New Zealand's forests, finding soil moisture to be the most reliable drought indicator. The importance of understanding satellite data for effective forest management and climate change strategies.
The introduction is suitable. 'Material and Methods' requires a better description, to encompass data collection and software usage. The 'Results' section is adequate, as is the discussion, though it is relatively brief. The conclusions support the objectives of the work. However, this approach is good and consistent and requires no major changes. I suggest the authors standardise to the models of Remote Sensing using the appropriate template available on the website.
The images are of good quality and require only corrections to the units on the x- and y-axes. I suggest including more recent references. Check for language corrections, especially in the synthesis of some sentences for better fluidity.
The text is very verbose. Several sentences throughout the manuscript are inappropriate and need to be removed, e.g., L122-123, among others.
Add more keywords.
L65-82. Consider starting a new paragraph to write the objectives (line 76).
Describe all figure and table captions adequately. They are brief, unfocused, and not descriptive enough. It is not possible to determine precisely what they refer to. Complete descriptions are necessary.
In Figure 1, a scale bar and north indication are added. Additionally, include a legend with information on the elements mentioned.
The units were standardised from “0.2 cm3/cm3” to “0.2 cm3 cm−3”.
In Figure 2, what is the sample size/number of observations?
Comments on the Quality of English LanguageCheck grammar and synthesis.
Author Response
> This manuscript (remotesensing-2844934) explores the impact of drought on forests and the use of satellite remote sensing for monitoring. It evaluates various hydrological indicators in New Zealand's forests, finding soil moisture to be the most reliable drought indicator. The importance of understanding satellite data for effective forest management and climate change strategies.
We thank the reviewer for their helpful comments and suggestions. We have tried to address them as best as possible in conjunction with the comments from the other reviewers.
> The introduction is suitable. 'Material and Methods' requires a better description, to encompass data collection and software usage. The 'Results' section is adequate, as is the discussion, though it is relatively brief. The conclusions support the objectives of the work. However, this approach is good and consistent and requires no major changes. I suggest the authors standardise to the models of Remote Sensing using the appropriate template available on the website.
We have added explicit references to the Python libraries used, and have used the official MDPI LaTeX template.
> The images are of good quality and require only corrections to the units on the x- and y-axes. I suggest including more recent references. Check for language corrections, especially in the synthesis of some sentences for better fluidity.
We have attempted to include more recent references where appropriate, and modify/add sentences throughout the manuscript to improve continuity and correct and language/grammar errors.
> The text is very verbose. Several sentences throughout the manuscript are inappropriate and need to be removed, e.g., L122-123, among others.
We have removed the lines in question, and have attempted to make the text less verbose to the best of our abilities.
> Add more keywords.
We have included additional keywords.
> L65-82. Consider starting a new paragraph to write the objectives (line 76).
We have split l. 65-82 into two paragraphs.
> Describe all figure and table captions adequately. They are brief, unfocused, and not descriptive enough. It is not possible to determine precisely what they refer to. Complete descriptions are necessary.
We have modified the captions to the figures to improve their descriptiveness.
> In Figure 1, a scale bar and north indication are added. Additionally, include a legend with information on the elements mentioned.
We have added a scale bar and a north arrow in Figure 1.
> The units were standardised from “0.2 cm3/cm3” to “0.2 cm3 cm−3”.
We have changed the units to match "cm3 cm-3".
> In Figure 2, what is the sample size/number of observations?
The observations have been aggregated to daily and cover the period 2018/6/13 to 2019/5/7 with 329 observations.
Reviewer 4 Report
Comments and Suggestions for AuthorsThis manuscript assesses the hydrological consistency of observations of precipitation, soil moisture, evapotranspiration, and surface temperature based on in situ measurements from seven plantation forests in New Zealand where soil moisture sensors and weather stations have been installed, and combining datasets from different sensors. The overall content reveals an interesting phenomenon and a richer workload, but there are still some issues to be addressed.
1. The elements of Figure 1 are incomplete and need to be added on such elements as compass, legend and scale.
2. In Method 2.3, the application of interpolation based on Gaussian process regression to interpolate the missing data in the soil moisture time series is not very reasonable, in other words, why was this interpolation method selected? The authors could have compared several interpolation methods for comparative screening.
3.In Result 3.3, in the manuscript "we can evaluate the MODIS observations against measurements of 2-m air temperature at the study sites.", is there any other literature or data to support the choice of height here, and what is the basis?
4. It can be seen from this manuscript that there seems to be some problems in using satellite observations to characterize meteorological droughts, especially in the part of precipitation comparison. Consider adding a few more relevant indices to further validate your point.
Comments on the Quality of English LanguageThe article is clearly expressed, with a concise and clear sentence structure and few spelling errors, and a relatively clear logical structure that helps to understand the thinking behind the author's research. However, there are still some areas that need attention to detail, and consideration should be given to the use of transitional vocabulary to improve the flow between sentences.
Author Response
> This manuscript assesses the hydrological consistency of observations of precipitation, soil moisture, evapotranspiration, and surface temperature based on in situ measurements from seven plantation forests in New Zealand where soil moisture sensors and weather stations have been installed, and combining datasets from different sensors. The overall content reveals an interesting phenomenon and a richer workload, but there are still some issues to be addressed.
We thank the reviewer for their helpful comments and suggestions. We have tried to address them as best as possible in conjunction with the comments from the other reviewers.
> 1. The elements of Figure 1 are incomplete and need to be added on such elements as compass, legend and scale.
We have added an arrow and scale bar to Figure 1.
> 2. In Method 2.3, the application of interpolation based on Gaussian process regression to interpolate the missing data in the soil moisture time series is not very reasonable, in other words, why was this interpolation method selected? The authors could have compared several interpolation methods for comparative screening.
Compared to other methods, Gaussian Process (GP) models are non-parametric that also provide uncertainty estimates. They are also very flexible models and we argue that they can better handle sparse data (such as the SMAP/Sentinel observations) compared to simpler methods such as piecewise regression. As the focus of the paper is the hydrological consistency of the multi-sensor satellite observations, the comparison of different time series imputation methods is beyond the scope. Nonetheless, we do agree with the reviewer that such a study would be indeed valuable.
> 3. In Result 3.3, in the manuscript "we can evaluate the MODIS observations against measurements of 2-m air temperature at the study sites.", is there any other literature or data to support the choice of height here, and what is the basis?
The 2-m height for air temperature measurements has been standard in hydro-meteorology, and it is the measurement reported for the majority of weather stations since the late 19th century (see <https://data.giss.nasa.gov/gistemp/faq/abs_temp.html>).
> 4. It can be seen from this manuscript that there seems to be some problems in using satellite observations to characterize meteorological droughts, especially in the part of precipitation comparison. Consider adding a few more relevant indices to further validate your point.
Although we do agree with the reviewer that additional indices would further validate the point of satellite observations not capturing drought in the study sites, we did choose to primarily use SPI as it has been found to be the superior index (see <https://doi.org/10.1175/2010BAMS3103.1>). In fact, "the inter-regional workshop on Drought indices and Early Warning Systems organized by the World Metrological Organization (WMO) in 2019 December, 22 countries agreed to use SPI as the best index for meteorological drought monitoring" (<https://doi.org/10.1080/19475705.2022.2044394>).
Reviewer 5 Report
Comments and Suggestions for AuthorsDear Authors,
I have thoroughly reviewed your paper, which presents an exciting research topic on drought assessment using in-situ observations with multi-source satellite data under forested environments in New Zealand. The study not only exhibits promise but also sheds light on the limitations and challenges associated with remote sensing for drought monitoring. The paper is commendably written, displaying a well-structured narrative enriched with sufficient details. However, please allow me to offer a few minor comments aimed at further refining the manuscript:
In the abstract: Please briefly present the numeric results from the Granger causality test.
The introduction is very well structured and contains the right amount of details.
Although the study area is well-known globally, I recommend adding to Figure 1 a world map locator of New Zealand.
Line 96: Please clarify the meaning of "canopy closure."
Line 104: Please briefly define the term "capacitance."
Table 2: Spatial resolution should be in m or km (a uniform SI unit)
Line 165: Do you mean Kelvin by K; this should be in degrees Celcius.
Line 229: It should be "the former with ground-based estimates."
Line 326: Figure 8 should be under this paragraph.
Line 424-426: I recommend these articles (https://doi.org/10.1002/asl.1161;
https://doi.org/10.3390/rs14235961;
https://doi.org/10.5194/nhess-21-481-2021) on the "Rainfall Gini Index" for European drought assessment to be cited.
Line 438-440: This sentence should be revised for clarity.
Line 442: These causality relationships should be briefly summarized in the conclusion section.
Figure A1: I recommend using the same style for NDVI plots like LAI in Figure A2.
Best regards.
Comments on the Quality of English LanguageLine 229: It should be "the former with ground-based estimates."
Author Response
> I have thoroughly reviewed your paper, which presents an exciting research topic on drought assessment using in-situ observations with multi-source satellite data under forested environments in New Zealand. The study not only exhibits promise but also sheds light on the limitations and challenges associated with remote sensing for drought monitoring. The paper is commendably written, displaying a well-structured narrative enriched with sufficient details. However, please allow me to offer a few minor comments aimed at further refining the manuscript:
We thank the reviewer for their helpful comments and suggestions. We have tried to address them as best as possible in conjunction with the comments from the other reviewers.
> In the abstract: Please briefly present the numeric results from the Granger causality test.
We have added a summary of the causality tests with the range of p-values in the abstract.
> The introduction is very well structured and contains the right amount of details.
Thank you.
> Although the study area is well-known globally, I recommend adding to Figure 1 a world map locator of New Zealand.
We have modified the figure to include an inset of a world map.
> Line 96: Please clarify the meaning of "canopy closure."
We have added a definition of canopy closure.
> Line 104: Please briefly define the term "capacitance."
We have added a definition for capacitance.
> Table 2: Spatial resolution should be in m or km (a uniform SI unit)
As the resolution of the official products is in degrees, it is latitude-dependent and therefore we would have to use a range of values for m or km. Therefore we opt to report the resolutions directly from each data product's specification.
> Line 165: Do you mean Kelvin by K; this should be in degrees Celcius.
We have changed the units to degrees Celsius, as the quantity represents a difference the value stays the same.
> Line 229: It should be "the former with ground-based estimates."
We have fixed this sentence.
> Line 326: Figure 8 should be under this paragraph.
We have used the official MDPI LaTeX template which automates figure placement. However, in case the paper is accepted for publication we will work with the editorial office to move Figure 8 accordingly.
> Line 424-426: I recommend these articles (https://doi.org/10.1002/asl.1161;https://doi.org/10.3390/rs14235961;https://doi.org/10.5194/nhess-21-481-2021) on the "Rainfall Gini Index" for European drought assessment to be cited.
We have added these citations in the text.
> Line 438-440: This sentence should be revised for clarity.
We have modified the sentence to improve clarity.
> Line 442: These causality relationships should be briefly summarized in the conclusion section.
We have added a couple of sentences summarizing the causality relationships in the conclusion section.
> Figure A1: I recommend using the same style for NDVI plots like LAI in Figure A2.
We do agree with the reviewer that consistency would be beneficial. Unfortunately due to the higher frequency of the NDVI measurements, adding lines along with dots would make it difficult to discern the soil moisture observations (see image below). Therefore we opted to retain the original figure.
Round 2
Reviewer 1 Report
Comments and Suggestions for AuthorsTwo points need to be revised and improved.
i) In Figure 1, please improve the scale bar.
ii) A flow diagram is needed to add to the paper.
Author Response
We thank the reviewer for reviewing the revision of the manuscript.
> Two points need to be revised and improved.
> i) In Figure 1, please improve the scale bar.
It's not clear what specific changes the reviewer is requesting, but we have made an attempt to improve it. We have removed the background making it transparent and moved the entire bar to a more appropriate location.
> ii) A flow diagram is needed to add to the paper.
The workflow is very straightforward making a flow diagram redundant. However, we have added a diagram that shows the different datasets, indices and analyses performed on each.
Reviewer 4 Report
Comments and Suggestions for AuthorsTable 3 in the manuscript can be converted into a three-line table.
Author Response
We thank the reviewer for reviewing the revision of the manuscript.
> Table 3 in the manuscript can be converted into a three-line table.
We have chosen to remove the table in favor of a flow diagram.
