Assessing shaded-leaf effects on photochemical reflectance index (PRI) for water stress detection in winter wheat

Yang, Xin; Liu, Shishi; Liu, Yinuo; Ren, Xifeng; Su, Hang

doi:https://doi.org/10.5194/bg-16-2937-2019

Articles | Volume 16, issue 15

https://doi.org/10.5194/bg-16-2937-2019

© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/bg-16-2937-2019

© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 16, issue 15

Research article

|

01 Aug 2019

Research article |

| 01 Aug 2019

Assessing shaded-leaf effects on photochemical reflectance index (PRI) for water stress detection in winter wheat

Xin Yang, Shishi Liu, Yinuo Liu, Xifeng Ren, and Hang Su

Download

Final revised paper (published on 01 Aug 2019)
Preprint (discussion started on 29 Oct 2018)

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC1: 'Review of Yang et al.', Anonymous Referee #1, 22 Nov 2018
- AC1: 'We appreciate reviewer’s valuable comments and suggestions. Here are our responses to reviewer’s major concern.', Xin Yang, 27 Nov 2018
RC2: 'Review of Yang et al. Assessing shadow effects on Photochemical Reflectance Index (PRI) for the water stress detection in winter wheat', Anonymous Referee #2, 04 Feb 2019
- AC2: 'We appreciate reviewer’s valuable comments and suggestions. Here are our responses to reviewer’s major concerns.', Xin Yang, 07 Feb 2019

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

ED: Reconsider after major revisions (05 Mar 2019) by Christopher Still

AR by Xin Yang on behalf of the Authors (12 Apr 2019) Author's response Manuscript

ED: Referee Nomination & Report Request started (06 May 2019) by Christopher Still

RR by Anonymous Referee #1 (27 May 2019)

Suggestions for revision or reasons for rejection

It is great to see that the authors worked hard to incorporate the suggestions. The revisions have considerably improved the manuscript in terms of focus. I think that the manuscript can improve even more when the authors are more explicit in their data analysis. I will mention a couple of suggestions.

Specific Comments
1. For the analysis on the slope and intercept the authors have chosen to use a linear regression. However, when you look at figure 5 it is clear that for example an exponential fit could be a better representation. The model misses the point of 100% shaded, for 3 out of 4 subfigures). A quick pseudo data analysis led me to the following equation: PRI3_intercept=a*exp(shaded fraction * b)-c, where a=9.1e-7, b=13.37, and c=0.31, this results in an R-squared value of 0.997. Why is it that the authors decide to use a linear model?
2. IF, the authors decide to use above mentioned model the conclusion of the paper that “PRI of the pure shaded leaves may yield inaccurate estimates of plant water status” could be more refined. I do understand that when they use the linear models for estimating RWC using PRI the authors chose to use the regression from the 50/50 ratio sun-shaded leaves. However, if an estimation of the true shaded area can be made, the model geared towards that specific ratio could be used for RWC estimation.
3. It would be interesting to quantify the impact of using a particular linear model for the estimation of RWC using all different models (trained on the different rations). This sensitivity analysis might show whether the decision of the authors to used 50/50 leaves/shadow ratio for RWC estimation is acceptable. How does the RMSE look when 100% sunlit-model is used? And how far off is the estimation from the ‘best’ estimation (probably reached with the model that it’s trained on).

4. I also strongly recommend that a native English speaker thoroughly examines the manuscript for language errors. I’ve pointed out a couple, but I’m sure that I’ve overlooked others.

Page 6, section 3.1: There is no reference in the text to Figure 3 and Table 2.
Page 8, figure 3: The authors should consider using different symbols for the sunlit and shaded leaves data points. There is also no mention what the error bars mean. Is it for example a standard deviation, or a quantile? I would also like to know what the authors think causes the peak in shaded leaves at RWC 0.5-0.6.
Page 8, line 16: The terms “strong” and “weak” seem somewhat ambiguous. I don’t see how a R-squared of 0.31 is strong, and a R-squared of 0.28 is weak. Figure 4 a) does not show a strong relationship either.
Page 9, figure 4: I would like to compare this figure with the data in table 3. However, the authors use a different statistical metric. They should consider changing from Pearson’s r to R-squared.
Page 10, line 6: I assume the authors are showing Pearson’s r in table 4. But it would be appreciated if they explicitly say this. Again, reconsider changing this to R-squared so it matches figure 5.

Technical Corrections
Page 3, line 12: “water stress experiments of”  “a water stress experiment on”
Page 3, line 22: “the volumetric”  “a volumetric”
Page 3, line 22: “at the field”  “at field”
Page 3, line 26: “were growing outdoor under the natural condition”  were grown outdoors under natural conditions”
Page 3, line 28: “of the field capacity”  “off field capacity”
Page 3, line 28: “The water stress”  “Water stress”
Page 3, line 29: “which was the tiller”  “which was during tiller”
Page 3, line 29: “prevent the external”  “prevent external”
Page 5, line 1: “the radiometric”  “radiometric”
Page 5, line 10-11: “in the MATLAB software”  “in MATLAB”.
Page 10, line 3&4: make the “P” lower case.
Page 13, line 27: “proved”  “proven”
Page 13, line 27: “of the structural”  “of structural”
Page 14, line 30: “for crops have different”  “for crops that have a different”

Hide

ED: Reconsider after major revisions (03 Jun 2019) by Christopher Still

Dear Xin and Shishi

Thank you for your revised manuscript. It seems greatly improved from the previous one and you have worked to address reviewer criticisms and concerns. The most recent review has some important questions and suggestions (the most important ones are copied below). Please respond to them, as well as the minor comments, and modify your manuscript accordingly.

Thank you,
Chris Still
Associate Editor

"Specific Comments
1. For the analysis on the slope and intercept the authors have chosen to use a linear regression. However, when you look at figure 5 it is clear that for example an exponential fit could be a better representation. The model misses the point of 100% shaded, for 3 out of 4 subfigures). A quick pseudo data analysis led me to the following equation: PRI3_intercept=a*exp(shaded fraction * b)-c, where a=9.1e-7, b=13.37, and c=0.31, this results in an R-squared value of 0.997. Why is it that the authors decide to use a linear model?
2. IF, the authors decide to use above mentioned model the conclusion of the paper that “PRI of the pure shaded leaves may yield inaccurate estimates of plant water status” could be more refined. I do understand that when they use the linear models for estimating RWC using PRI the authors chose to use the regression from the 50/50 ratio sun-shaded leaves. However, if an estimation of the true shaded area can be made, the model geared towards that specific ratio could be used for RWC estimation.
3. It would be interesting to quantify the impact of using a particular linear model for the estimation of RWC using all different models (trained on the different rations). This sensitivity analysis might show whether the decision of the authors to used 50/50 leaves/shadow ratio for RWC estimation is acceptable. How does the RMSE look when 100% sunlit-model is used? And how far off is the estimation from the ‘best’ estimation (probably reached with the model that it’s trained on).
4. I also strongly recommend that a native English speaker thoroughly examines the manuscript for language errors. I’ve pointed out a couple, but I’m sure that I’ve overlooked others."

Hide

AR by Xin Yang on behalf of the Authors (16 Jun 2019)

ED: Publish as is (11 Jul 2019) by Christopher Still

AR by Xin Yang on behalf of the Authors (14 Jul 2019) Manuscript

Short summary

The photochemical reflectance index (PRI) derived from remotely sensed data has emerged to be a pre-visual indicator of water stress. This study evaluated the impact of the varying shaded-leaf fractions on estimating relative water content (RWC) across growth stages of winter wheat using PRI derived from hyperspectral imagery. Results showed that PRI of the pure shaded leaves may yield inaccurate estimates of plant water status, but the accuracy of RWC predictions was not significantly affected.