Direct and Remote Sensing Monitoring of Plant Salinity Stress in a Coastal Back-Barrier Environment: Mediterranean Pine Forest Stress and Mortality as a Case Study

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Very interesting paper. I suggest this manuscript needs to incorporate accuracy assessment of the remote sensing analysis as a table for all the classifications presented in the study. 

Author Response

Very interesting paper. I suggest this manuscript needs to incorporate accuracy assessment of the remote sensing analysis as a table for all the classifications presented in the study.

We would like to thank Reviewer 1 for the positive feedback. We have now included an accuracy assessment for NDVI and NDMI. This information has been summarized in Tables 1s and 2s in supplementary information, while a new section in Materials and Methods called “Data Analysis” has been added (lines 234-247).

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript deals with the process of progressive degradation (up to total desiccation and death) of a pine forest in southern Italy. The process was monitored by integrating both remote sensing techniques (NDVI together with NDMI were the applied indices) and direct assessment procedures on the soil (salinity and temperature), on the groundwater (depth and salinity) as well as meteorological measurements (air temperature and potential evapotranspiration) and ultimately salts from storm-borne aerosols (“sea spray”) deposited on the soil. The good amount of data obtained makes it possible to interpret the process taking place and to characterize its main and secondary causes. The methodological approach taken is sound and robust, the experimental deployment is broad and effective, and the data set obtained is very substantial and useful for the purposes of the paper.

Remote sensing (RS) techniques are only one component of the study, which is largely multidisciplinary. Although this may be considered a limitation of the work, it is important to verify the application, effectiveness, and constraints of RS.

There are two main weaknesses in the manuscript:

1. a questionable "freestyle" written English

2. a rather boring and uninspiring description of the results.

Both aspects should be greatly improved.

I have personally tried to rewrite the Abstract to give you an idea of the profound improvement the text deserves. Take this exercise of mine only as an example and advice. Some other suggestions are given as comments on the revised pdf version of your paper. Please, pay special attention to the logical order of sentences. Make sure they are consistent and logically sequential.

The Abstract (Lines 11-32) was completely rephrased, according to the following suggestions:

 

Abstract

<< The increase in atmospheric and soil temperatures in the last decades have led to unfavorable conditions for plants in many Mediterranean coastal environments. A typical example can be found along the coast of the Campania region in Italy, within the “Volturno Licola Falciano Natural Reserve”, where a pinewood suffered a dramatic loss of trees in 2021. New pines were planted in 2023 to replace the dead ones, with a larger tree layout and interspersed with Mediterranean shrubs. The newly planted pines were watered daily during that summer to prevent water stress.

A direct (in situ) monitoring program was planned to analyze the determinants of pine salinity stress, coupled with remote sensing monitoring from Sentinel-2 L2A; in particular, multispectral indices NDVI and NDMI were provided by the EU Copernicus service.

Both vadose and shallow groundwater were monitored with continuous logging probes.

Vadose zone monitoring indicated that salinity peaked at 30 cm soil depth with values up to 1.9 g/L. These harsh conditions, combined with air temperatures that peaked at over 40°C, created severe difficulties for pine growth. Shallow groundwater showed that salinity near the coast was low (0.35-0.4 g/L) because the dune allowed rapid rainwater infiltration and drainage, preventing seawater intrusion. In contrast, the salinity increased towards the inland, with a peak value of 2.8 g/L at the end of the summer. In November 2023, salts from storm-borne aerosols (“sea spray”) deposited on the soil caused the sea-facing portion of the newly planted pines to dry out. Differently, the pioneer vegetation of the Mediterranean dunes, directly facing the sea, was not affected by the massive deposition of sea spray.

The NDMI and NDVI data were useful in distinguishing the old pine trees suffering from increasing stress and final death but were not accurate in correctly detecting the stress conditions of newly planted, still rather short pine trees, because their spectral reflectance largely interfered with the adjacent shrub growth. The proposed coupling of direct and remote sensing monitoring was successful and could be applied to detect the main drivers of plant stress in many other Mediterranean coastal environments>>.

 

Please note the following: all comments and suggestions, as well as requests for clarification and insight, were incorporated directly into the text of the manuscript (a new version) using Adobe's specific annotation tools for adding comments to PDFs.

Concerning the Title (Lines 2-3). I suggest a different title, in my opinion, easier to grasp and also more representative of the research:

Direct and remote sensing monitoring of plant salinity stress in a coastal back-barrier environment: Mediterranean pine stress and mortality as a case study.

Below you will find some technical and scientific questions related to your research:

1. The spatial resolution of the satellite images is one of the essential aspects of remote sensing. In your work, this information is lacking. What is the size of each pixel? How many pixels were used in the survey? What is the total size of the monitored area and the extent of the pine forest? (please, note that I prefer the term “pine forest” rather than “pinewood”). What is the actual time frequency of Sentinel-2 image acquisition?

2. Line 3, Equation 3). Please check the equation carefully. I suspect something is wrong because the units on the right do not match the units on the left of the equation.

3. Did you use the Diver in a monitoring well or piezometer tube? What was the depth of the tube? Please give a better description of the instrument you used. How far apart are the piezometers placed?

4. Monitoring is supposed to be continuous with the use of data loggers. What is the time-frequency of data collection, and how much data is actually stored in the data logger? Is the stored data an average of direct measurements? Was an hourly average taken or what is the time window of the average? This kind of information seems to be missing…

5. Have you tried using "time series" analysis statistical techniques? You can de-seasonalize the signal and highlight both its multi-year trend and any significant anomalies. With so much data at your disposal, try to make better use of it.

6. Sometimes it is not so easy to interpret the time trend graphs you show in the paper. It would be necessary to make them more readable or to highlight some specific part of them in the features that you consider most relevant. Fewer but clearer graphs would suffice. Especially in the "Discussion" section, more results accompanied by other graphs should be avoided. This is an inexperienced mistake.

Considering all these comments, I suggest that the Authors reconsider their work after a major revision.

Comments for author File: Comments.pdf

Comments on the Quality of English Language

Please read the comment in the box above. I strongly recommend a comprehensive editing of the English language.                    

Author Response

The manuscript deals with the process of progressive degradation (up to total desiccation and death) of a pine forest in southern Italy. The process was monitored by integrating both remote sensing techniques (NDVI together with NDMI were the applied indices) and direct assessment procedures on the soil (salinity and temperature), on the groundwater (depth and salinity) as well as meteorological measurements (air temperature and potential evapotranspiration) and ultimately salts from storm-borne aerosols (“sea spray”) deposited on the soil. The good amount of data obtained makes it possible to interpret the process taking place and to characterize its main and secondary causes. The methodological approach taken is sound and robust, the experimental deployment is broad and effective, and the data set obtained is very substantial and useful for the purposes of the paper.

Remote sensing (RS) techniques are only one component of the study, which is largely multidisciplinary. Although this may be considered a limitation of the work, it is important to verify the application, effectiveness, and constraints of RS.

There are two main weaknesses in the manuscript:

1. a questionable "freestyle" written English
2. a rather boring and uninspiring description of the results.

Both aspects should be greatly improved.

I have personally tried to rewrite the Abstract to give you an idea of the profound improvement the text deserves. Take this exercise of mine only as an example and advice. Some other suggestions are given as comments on the revised pdf version of your paper. Please, pay special attention to the logical order of sentences. Make sure they are consistent and logically sequential.

The Abstract (Lines 11-32) was completely rephrased, according to the following suggestions:

Abstract

<< The increase in atmospheric and soil temperatures in the last decades have led to unfavorable conditions for plants in many Mediterranean coastal environments. A typical example can be found along the coast of the Campania region in Italy, within the “Volturno Licola Falciano Natural Reserve”, where a pinewood suffered a dramatic loss of trees in 2021. New pines were planted in 2023 to replace the dead ones, with a larger tree layout and interspersed with Mediterranean shrubs. The newly planted pines were watered daily during that summer to prevent water stress. A direct (in situ) monitoring program was planned to analyze the determinants of pine salinity stress, coupled with remote sensing monitoring from Sentinel-2 L2A; in particular, multispectral indices NDVI and NDMI were provided by the EU Copernicus service. Both vadose and shallow groundwater were monitored with continuous logging probes. Vadose zone monitoring indicated that salinity peaked at 30 cm soil depth with values up to 1.9 g/L. These harsh conditions, combined with air temperatures that peaked at over 40°C, created severe difficulties for pine growth. Shallow groundwater showed that salinity near the coast was low (0.35-0.4 g/L) because the dune allowed rapid rainwater infiltration and drainage, preventing seawater intrusion. In contrast, the salinity increased towards the inland, with a peak value of 2.8 g/L at the end of the summer. In November 2023, salts from storm-borne aerosols (“sea spray”) deposited on the soil caused the sea-facing portion of the newly planted pines to dry out. Differently, the pioneer vegetation of the Mediterranean dunes, directly facing the sea, was not affected by the massive deposition of sea spray. The NDMI and NDVI data were useful in distinguishing the old pine trees suffering from increasing stress and final death but were not accurate in correctly detecting the stress conditions of newly planted, still rather short pine trees, because their spectral reflectance largely interfered with the adjacent shrub growth. The proposed coupling of direct and remote sensing monitoring was successful and could be applied to detect the main drivers of plant stress in many other Mediterranean coastal environments>>.

Please note the following: all comments and suggestions, as well as requests for clarification and insight, were incorporated directly into the text of the manuscript (a new version) using Adobe's specific annotation tools for adding comments to PDFs.

We would like to thank Reviewer 2 for the useful comments and detailed feedback on our manuscript. We appreciate the recognition of the robustness of the methodological approach and of the completeness of the data set presented in this study. We have carefully addressed the points raised regarding the English language and the presentation of the results. The manuscript has been edited by a native English speaker with expertise in academic writing. We have also revised the description of the results to make it more captivating. Based on your feedback, we have rephrased the Abstract and we have enhanced the entire manuscript.

 

Concerning the Title (Lines 2-3). I suggest a different title, in my opinion, easier to grasp and also more representative of the research:

Direct and remote sensing monitoring of plant salinity stress in a coastal back-barrier environment: Mediterranean pine stress and mortality as a case study.

The title has been modified following the reviewer’s suggestion.

 

Below you will find some technical and scientific questions related to your research:

1. The spatial resolution of the satellite images is one of the essential aspects of remote sensing. In your work, this information is lacking. What is the size of each pixel? How many pixels were used in the survey? What is the total size of the monitored area and the extent of the pine forest? (please, note that I prefer the term “pine forest” rather than “pinewood”). What is the actual time frequency of Sentinel-2 image acquisition?

Sentinel-2 L2A NDVI and NDMI data with a pixel size of 10 m and 20 m, respectively, were collected every 5 days; while annual images (each October) were extracted from the year 2016 to the year 2023. Now it is stated at lines 139-142.  The total size of the experimental field is 0.6 km², now stated at line 100; while the total area of the Pinewood is 1540 ha as already stated at line 111.  We also substitute the term “pinewood” in “pine forest” where necessary.

2. Line 3, Equation 3). Please check the equation carefully. I suspect something is wrong because the units on the right do not match the units on the left of the equation.

The correction was made at lines 186-187.

3. Did you use the Diver in a monitoring well or piezometer tube? What was the depth of the tube? Please give a better description of the instrument you used. How far apart are the piezometers placed?

Each piezometer consisted of a 2 m long polyethylene tube with internal diameter of 3 cm and slotted with a screen of 10 cm at its bottom, covered with a nitex 50 µm mesh to prevent clogging. Piezometers were installed to a depth of 1.5 m with a motor auger and then equipped with a water level data logger to monitor groundwater level (H), temperature (T), and TDS (Soil & Water Diver® from Eijkelkamp, Giesbeek, the Netherlands). Now stated at lines 193-198. The distance between P1 and P2 was 165 m, while the distance between P2 and P3 was 101 m, now stated at lines 192-193.

4. Monitoring is supposed to be continuous with the use of data loggers. What is the time-frequency of data collection, and how much data is actually stored in the data logger? Is the stored data an average of direct measurements? Was an hourly average taken or what is the time window of the average? This kind of information seems to be missing…

Concerning groundwater, for each parameter of each diver, 12710 data were collected through continuous direct measurements reaching a total of 114390 data. Data recording was set at 30-minutes intervals and the monitoring period ran from March 2023  to November 2023. Now stated at lines 200-202. Concerning the vadose zone, data recording was set at 30-minutes intervals and the monitoring period ran from March 2023 to September 2023. For each parameter and for each depth the data logger stored 7364 data reaching a total of 95732 data collected through continuous direct measurements. Now stated at lines 206-209.

5. Have you tried using "time series" analysis statistical techniques? You can de-seasonalize the signal and highlight both its multi-year trend and any significant anomalies. With so much data at your disposal, try to make better use of it.

We have improved the data analysis with the calculation of Locally Estimated Scatterplot Smoothing (LOESS). It was calculated on NDVI and NDMI data using Statgraphics Centurion 19 to better explore the non-linear trend. LOESS is defined by Span, Family, and Degree. The Span represents the proportion of data used to fit the local polynomial at each point. The Family specifies the fitting algorithm. The Degree specifies the order in which the local polynomials are fitted to each data subset. The Span was set to 0.5. Gaussian was chosen as the Family using least squares fit. The Degree was set to 2 to fit a quadratic function to each data subset. Now stated at lines 235-241. Moreover, figure 2 was modified by separating Tmax and NDVI in two different graphs, and NDMI LOESS curve and NDVI LOESS curve were added to the graphs C and D, respectively.

6. Sometimes it is not so easy to interpret the time trend graphs you show in the paper. It would be necessary to make them more readable or to highlight some specific part of them in the features that you consider most relevant. Fewer but clearer graphs would suffice. Especially in the "Discussion" section, more results accompanied by other graphs should be avoided. This is an inexperienced mistake.

We have added in figure 2 the vertical lines to underline the starting of the monitoring periods for the new pines, as suggested in the PDF comments. Moreover, we have moved figure 7 and its description in the results section. Now figure 7 is at line 396 and the text related to it is at lines 383-395.

 

Reviewer 3 Report

Comments and Suggestions for Authors

his paper combines vegetation indices and other parameters extracted from satellite imagery with surface detection to achieve the detection of plant salinity stress in a coastal back barrier environment. Quite interesting with clear significance. Additionally, the artworks can also effectively reflect the main results. But before it can be published, I still believe further improvements are needed in the following areas.

1.The measurement process of soil physical and chemical properties, it is recommended to provide a more detailed introduction.

2.What standards are used to determine the range of soil texture , USDA?

3.In M&M section, it is recommended to clarify the combination of satellite images and ground detection more clearly.

4.Figure 2 is not clear.

5. It is suggested to provide more tables to reflect the results of data

measurement and processing.

6.Please explain whether the spatial resolution of Sentinel2 meets the research objectives。

7.The introduction and summary of the research background and existing methods in the introduction section are relatively weak. It is suggested to further add relevant explanations and present hypotheses more fully.

8.The abstract does not provide specific information on satellite imagery and its specific functions and usage methods。

9. It is suggests emphasizing specific research subjects in title, such as Pinewood.

Author Response

This paper combines vegetation indices and other parameters extracted from satellite imagery with surface detection to achieve the detection of plant salinity stress in a coastal back barrier environment. Quite interesting with clear significance. Additionally, the artworks can also effectively reflect the main results. But before it can be published, I still believe further improvements are needed in the following areas.

1.The measurement process of soil physical and chemical properties, it is recommended to provide a more detailed introduction.

The measurement process of soil physical and chemical properties is not a crucial information for meeting the research goals in this study, despite they are key factor for the description of the study area and for soil classification. For this reason, we have not provided this information in the main text not to negatively impact the fluency of the text. By the way, we have added the information about all the standards methods used for the analysis of the soil parameters in Table 1.

2.What standards are used to determine the range of soil texture , USDA?

Yes, now stated in Table 1.

3.In M&M section, it is recommended to clarify the combination of satellite images and ground detection more clearly.

We have added a statement to clarify this aspect at lines 168-170. Moreover, we have provided an explanation about how field data were used to assess the accuracy of satellite data at lines 242-247.

4.Figure 2 is not clear.

Figure 2 was modified by separating Tmax and NDVI in two different graphs, and NDMI LOESS curve and NDVI LOESS curve were added to the graphs C and D, respectively. We have also added the vertical lines to underline the starting of the monitoring period of the new pines.

5. It is suggested to provide more tables to reflect the results of data measurement and processing.

We have added Table 1s and Table 2s in the Supplementary Information to represent the level of NDVI and NDMI accuracy. Moreover, we also report the amount of data collected at lines 200-201 and 208-209.

6.Please explain whether the spatial resolution of Sentinel2 meets the research objectives.

We have added a statement about this aspect at lines 465-470.

7.The introduction and summary of the research background and existing methods in the introduction section are relatively weak. It is suggested to further add relevant explanations and present hypotheses more fully.

We have enhanced the introduction section adding sentences at lines 52-56 and 68-70. We have also carefully edited the English to improve the strength of this section.

8.The abstract does not provide specific information on satellite imagery and its specific functions and usage methods.

We have added this information at lines 18-20 and 30-33.

9. It is suggests emphasizing specific research subjects in title, such as Pinewood.

Thank you for your advice, the title has been changed accordingly.

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

I acknowledge that a decisive and profound improvement of the manuscript has been achieved by giving due consideration to the suggestions and advice made during the revision process. I believe that the work is well done and has far exceeded the criteria for being considered worthy of publication. Congratulations to the Authors.

To leave no stone unturned in improving the quality of the work, one more small detail should be mentioned. Formula 3 is still questionable:

If I understand it well, the pore water total dissolved solids (TDS_pw) should be equal to the total dissolved solids in the liquid phase (TDS), taking into account the liquid/solid (L/S) ratio (5:1) and the soil porosity. Both dilution ratio and porosity are dimensionless (-) variables.

Porosity, I suppose, can be estimated by the following ratio: 𝝆_𝒃 /𝝆_p, i.e. the ratio of dry soil bulk density (𝝆_𝒃) and dry soil particle density (𝝆_p), the latter being the density of solid soil particles only, i.e., the measurement does not include pore space. That’s it! (in Italian: “vi siete persi in un bicchier d’acqua”)

Sincerely yours

Prof. Massimo Monteleone

 

 

Reviewer 3 Report

Comments and Suggestions for Authors

The paper has been well revised. It's all right for me.