Potential of Time-Series Sentinel 2 Data for Monitoring Avocado Crop Phenology

Round 1

Reviewer 1 Report

The objective of this manuscript, entitled “Avocado crop phenology monitoring using a time series of Sentinel-2 data” is to retrieve avocado crop phenology using Sentinel 2 derived enhanced vegetation indices. The topic falls within the scope of the Journal.

 I recommend a following minor revision before the manuscript is ready for publication.

 1, It is suggested to add a quantitative method on how to determine the date of t2 and t1 used to calculate the EVI_slope. In the other words, how to identify four phenological stages using the SG smoothed EVI time series data.

 2, It is suggested to add more explanations about the differences in the different years and two different geographical locations shown in Figure 5.

 3, Pay attention to abbreviations. Try to give abbreviations when they first appear. Please review the abbreviation of the full text.

Author Response

Dear Reviewer,

Thank you for your effort to review our manuscript “Avocado crop phenology monitoring using a time series of Sentinel 2 data”. Please find the responses below:

 

Point 1: It is suggested to add a quantitative method on how to determine the date of t2 and t1 used to calculate the EVI_slope. In the other words, how to identify four phenological stages using the SG smoothed EVI time series data.

 

Response 1: This is a fair point and certainly one that we acknowledge. In line 256, the explanation has been added “The beginning and end dates of each phenological stages were taken from the field observed data given in Table 1”. (in orange). Moreover to clarify the field observed phenological data, line 192 has been extended as “.., which was the observation recorded from past fifteen to twenty years of data”. (in orange)

 

Point 2: It is suggested to add more explanations about the differences in the different years and two different geographical locations shown in Figure 5.

 

Response 2: Thanks for the suggestion. More explantaions have been added after line 179. “The mean annual rainfall over the study period (2017–2021) varied from 307.2 mm to 1214.4 mm in Bundaberg region with highest in 2017 and lowest in 2019 respectively and from 104.8 mm to 216.8 mm in Renmark region with highest in 2017 and lowest in 2019 respectively. The mean maximum temperature varied from 27.6 ◦C to 28.5 ◦C for Bundaberg region and 24.3 ◦C to 26 ◦C for Renmark region over the study period (2017–2021)” (in orange)

 

Point 3: Pay attention to abbreviations. Try to give abbreviations when they first appear. Please review the abbreviation of the full text.

 

Response 3: The abbreviations in the whole manuscript have been amended as suggested (in orange)

Author Response File: Author Response.docx

Reviewer 2 Report

This manuscript did the statistical analysis of EVI time series at different avocado phenological stages. There are some issues for this manuscript.

1. The sample sizes are too small for the two study sites.

2. The objectives and novelty of this manuscript are not clear. For the second objective, are there any technical difficulties?

3. The title does not reflect the feature of this paper. The author only did statistical analysis of the EVI time series. But how was the phenology of avocado monitored is not clear. Can the phenology be detected using these metrices? I suggest the authors add more experiments.

 

Detailed comments

L144-L145: The weather condition is not analyzed in the results and discussion.

L179: are the phenological information data from only one year?

Table 2: why std errors are the same for different stages?

Figure 5: are they the mean EVI of each farm? 

Author Response

Dear Reviewer,

Thank you for your effort to review our manuscript “Avocado crop phenology monitoring using a time series of Sentinel 2 data”. Please find the responses below:

 

Point 1: The sample sizes are too small for the two study sites.

 

Response 1: Thanks for your valuable comments. The sample size is an important feature for any empricial study in which the goal is to make inferences about a population from a sample. In practice, the sample size used in a study is usually determined based on the cost, time, or convenience of collecting the data, and the need for it to offer sufficient statisticl power. In this study, we used Sentinel 2 data, which is available in every 5 days, from 7 blocks in two different regions over 5 years period. According to, “Chapter 13, page 215, in: Kenny, David A. (1987). Statistics for the social and behavioral sciences. Boston: Little, Brown”, the number of samples for 0.8 power is only 26. Therefore, we believe that the samples used for the study is enough for the statistical tests.

 

Point 2: The objectives and novelty of this manuscript are not clear. For the second objective, are there any technical difficulties.

 

Response 2: Thanks for your comments. The objectives and novely of this study has been clearly stated in the last paragraph of Introduction section. The second objective of this study is to associate Sentinel 2 derived EVI with 4 key pehnological stages of avocado crops. The remote sensing techniques have been used for understanding the phenological stages of avocado crops, which is an important tool for farm management practices (i.e. irrigation and application of fertilizers and pesticides), yield prediction, harvest logistics and marketing strategies, and for identifying seasonal variations associated with pests, diseases and climate change. To the best of our knowledge, the use of satellite remote sensing for describing avocado crop phenology is novel and not yet been researched in any previous study. We would like to assure that no technical difficulties arised for the sencond objective.

 

Point 3: The title does not reflect the feature of this paper. The author only did statistical analysis of the EVI time series. But how was the phenology of avocado monitored is not clear. Can the phenology be detected using these metrices? I suggest the authors add more experiments.

 

Response 3: Thanks for your suggestion. The title of the manuscript has been amended to reflect the overal study described in the manuscript. Now the tile is “Potential of time series Sentinel 2 data to monitor avocado crop phenology”

 

Point 4: L144-L145: The weather condition is not analyzed in the results and discussion.

 

Response 4: Thanks for your suggestion. In our study we showed mainly the potential of Sentinel 2 image to monitor or identify different phenological stages of avocado crops. No weather parameters were used in the study. Therefore, there was no analysis on weather parameters done in Results and Discussion sections.

 

Point 5: L179: are the phenological information data from only one year?

 

Response 5: Thanks for your question. The same issue was raised by another reviewer. In line 193, the explanation has been extended as “.., which was the observation recorded from past fifteen to twenty years of data”.

 

Point 6: Table 2: why std errors are the same for different stages?

 

Response 6: Thanks for your question. Tukey's test uses a t-test type of statistic that is the ratio of the difference between two mean values divided by the standard error of the difference of those two means. Assuming independence and the same individual standard errors, that would be √2 times the standard error of an individual mean. To find more explanation about this, please searth the textbook “Design of Experiments: Statistical Principles of Research Design and Analysis. 2nd Edition” by Robert O. Kuehl.

 

Point 7: Figure 5: are they the mean EVI of each farm?

 

Response 7: Yes, they are the mean EVI of each farm. To clarify, the figure 5 and the caption have been amended to show Mean EVI in y-axis and in legend.

Author Response File: Author Response.docx

Reviewer 3 Report

The manuscript is well written and addressed one of active research areas i.e., Avocado crop phenology monitoring using proper scientific approach. Thus, the manuscript may be accepted for publication in the current version.

Author Response

Dear Reviewer,

Thank you for your effort to review our manuscript “Avocado crop phenology monitoring using a time series of Sentinel 2 data”. Please find the responses below:

 

Point 1: The manuscript is well written and addressed one of active research areas i.e., Avocado crop phenology monitoring using proper scientific approach. Thus, the manuscript may be accepted for publication in the current version.

 

Response 1: Thanks for your enourmous compliment about our manuscript. The manuscript has been amended according to the suggestions of other reviewers.

Author Response File: Author Response.docx

Reviewer 4 Report

A very interesting paper, highly topical issue, very well presented and analysed. taking advantage not only of the high temporality of Sentinel2 but also the benefit of the incorporation of the four key phenological stages of the avocado crop (i.e. flowering (F), vegetative growth (V), fruit maturity (M), and harvest (H)); and their association to their respective seasonal growth profiles. An encouraging characterisation of all the phenological stages by the different EVI metrics

One very important issue is that EVI index must be calculated from Bottom atmosphere reflectance, while in this study EVI has been directly calculated from Sentinel 2 L1C that means that no atmospheric correction has been performed

Could you justify why the atmospheric correction has not been applied to the Level 1C data, or why level 2A data was not directly otherwise for the EVI calculation?

Nevertheless I believe that GEE does not provided level 2A for the whole of your study period (https://developers.google.com/earth-engine/datasets/catalog/COPERNICUS_S2_SR. Dataset Availability 2017-03-28)

 

I also suggest some minor changes.

-   For Figure 1, It would be interesting to see the orthophto content for the Bundaberg and Renmark plots, just outline the parcels and leave them without any filling

-   Include the farm name (Bundaberg and Renmark) in all the figures (Figures 2-3-5-6-7) that will help to better understand them, in addition to being included in the description of the figures

Author Response

Dear Reviewer,

Thank you for your effort to review our manuscript “Avocado crop phenology monitoring using a time series of Sentinel 2 data”. Please find the responses below:

 

Point 1: Could you justify why the atmospheric correction has not been applied to the Level 1C data, or why level 2A data was not directly otherwise for the EVI calculation?

Nevertheless I believe that GEE does not provided level 2A for the whole of your study period (https://developers.google.com/earth-engine/datasets/catalog/COPERNICUS_S2_SR. Dataset Availability 2017-03-28)

 

Response 1: This is a fair point and certainly one that we acknowledge. Yes, the L1C data were transformed to L2A using Sen2Cor algorithm. Section 2.3 amended and the text added “However, the TOA data have significant limitations given their sensitivity to changes in the composition of the atmosphere through time. Therefore, all L1C data were transformed to L2A (bottom of atmosphere, BOA) using Sen2Cor algorithm [53]”. (in orange). The flowchart in Figure 4 has also been amended.

 

Point 2: For Figure 1, It would be interesting to see the orthophto content for the Bundaberg and Renmark plots, just outline the parcels and leave them without any filling.

 

Response 2: Thanks for the suggestion. Figure 1 has been amended as suggested.

 

Point 3: Include the farm name (Bundaberg and Renmark) in all the figures (Figures 2-3-5-6-7) that will help to better understand them, in addition to being included in the description of the figures.

 

Response 2: Thanks for the suggestion. However, due to confidentiallity agreement with the growers, the farm names were not disclosed.

Round 2

Reviewer 2 Report

Thanks for the authors' response. 

Author Response

Dear Reviewer,

Thank you for your effort to review our manuscript “Avocado crop phenology monitoring using a time series of Sentinel 2 data”. 

Reviewer 4 Report

The suggestions mainly the major one has been solved

Author Response

Dear Reviewer,

Thank you for your effort to review our manuscript “Avocado crop phenology monitoring using a time series of Sentinel 2 data”.