Remote Sensing of Poplar Phenophase and Leaf Miner Attack in Urban Forests

Round 1

Reviewer 1 Report

The paper is about the comparison of satellite multispectral data and ground ones from hyperspectral camera with the aim of detecting leaf miner attack effects. The paper would fit the aims of the journal. The topic is averagely interesting and potentially acceptable. Nevertheless the work, at its present state, cannot be accepted. Scientific soundness and technical deepening are insufficient and must be rigorously improved. Some suggestions of mine are reported within the attached pdf file. It shoud be highly revised, refined and completed. I would pay more attention to focus on the ground hyperspectral camera, data processing (calibration and georeferencing) and would provide infos about data integration whose quality is at the basis of the comparision. English language needs to be carefully and significantly revised.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

This study obtained the phenophases and identified pests attack using Hyperspectral camera. The topic is interesting. The study compared the VIs calculated based on hyperspectral camera and satellite, respectively. However, it did not compare the phenophases obtained from camera and remotely sensed phenology products. My detail comments are as follow.

1.    What is “NDRE” short for?

2.    The same questions for other abbreviations, like NDVI and EVI. Give their full names the first time they appear both in the abstract and text.

3.    Figures: Fig.1  Missing north arrow and legend.   Fig.4 these colors are hard to be distinguished.

4.    Did the phenophases illustrated in the Fig. 5, Fig. 6, Fig. 7 and Fig. 8 obtain based on the images from multispectral camera?

5.    The phenophases appeared repeatedly in several figures. Combine the information illustrated in the Fig. 5, Fig. 6, Fig. 7 and Fig. 8.

6.    If possible, please compare the phenophases obtained from camera and remotely sensed phenology products, which will improve our understanding of how the phenological transitions can be reflected from the pixel-level.

7.    The satellite views the canopy from the top, while the camera views the canopy on the side. How can you use multispectral camera to calibrate satellite images?

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

The study investigates the leaf miner attack in urban forests using two types of remote sensing data. A hyperspectral camera and Sentinel-2 images were used for the analysis. The time series of three years were acquired for the indices derived from the sensors and compared. The methodology needs to be detailed, and the statements need to be clarified. Because there are confusing interpretations. The study has interesting and promising results, but it has to be revised considering the suggestions and questions given below.

1.       Figure 1: Please add the range of NDVI

2.       There are several misspellings in the paper, please check it properly. Come of them are given below:

·         Line 122: “the” should be added before “red” and “spectrum”.

·         Line 149: Check the misspelling “tthe”.

·         Line 228: Check the misspelling “significantlly”.

·         Line 240, 243: replace “time” with “temporal”.

·         Line 253: Check the misspelling “remotly”.

·         In figures 5, 6, and 8: check the misspelling “satellit”.

·         Lines 358, 361, 363: remote the dots after the years.

·         Line 660: Check the misspelling “improvment”.

3.       Line 181: the resolution of Sentinel is mentioned as “high”. Do the authors mean low compared to the hyperspectral camera? It is confusing.

4.       Line 185: Please mention the level of Sentinel. Not all levels are corrected.

5.       Line 215-217: Give the long version of the abbreviations in the first place

6.       Lines 275-281: I think the statements belong to the previous reviewer's comment. I also agree that the mentioned tests should be described and cited.

7.       Figure 4: Which part of the image given in figure 1 do these results belong to? How many pixels are included in the spectral indices? Provide more details on the sampling strategy.

8.       Line 312-313: The values of satellite NDVI do not seem stable during the leafing period. It increases from 81 to 131 and shows different behavior between 251 and 331.

9.       The results of 2019 shown in Figures 6b and 6c are not given. For example, there are jumps in the results of EVI. What could be the reasons? NDVI looks more stable during the whole year. It is also confusing.

10.   Why NDRE is not created for Sentinel-2? The spatial resolution could be resampled to 10 m as done in many previous studies.

11.   Lines 358-359: the comment is confusing because the leafing starts on 81 DOY and during the leaf miner indices increase.

12.   In general, the index values of the years are given in the text, but it is very difficult to see them in the figures (5, 6, 7, 8), so it is recommended to mark them on the figure.

13.   In figure 8a it appears that from 91 to 141 the indices in both camera and satellite increase which means they are healthier not they?

14.   Lines 402-407, lines 570-573: How many samples were used for the correlation analysis? As hyperspectral camera and satellite images are acquired from different viewpoints, how do they overlap the same location?

15.   Lines 583-584: It is mentioned as 16 different wavelengths were used for the indices. However, in the processing part, this information was not provided. Which bands are used for the calculation of indices, please provide the equations in the processing part.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Authors have addressed the most of my comments and significantly improved their paper. Nevertheless I retain that some refinements are still due. 
The most of them are listed within the revised version of the paper attached. The most critical issue concerns the comparability of S2 and proximal hyperspectral sensing, being the first operating with wide spectral bands ad the second with single (almost) wavelengths. 

I suggest a further round of English editing for the technical parts only. The paper still shows a little awareness and familiarity with the remote sensing and sensor issues. Authors probably come from a more biologically oriented formation that strongly comes out in the interpretation of phenological processes. I therefore suggest that those parts of paper describing technical features related to sensors and signals are revised by one collegue more familiar with this topic.

After these refinements I feel that the paper will be proper for publication in the journal

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

The authors considered the suggestions and improved the text, including additional information. The questions are answered properly, except the one given below:

Lines 425-431: The mentioned tests (The Spearman’s, The Shapiro-Wilk and Wilcoxon signed-rank should be described, how they work, and they should be cited.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf