Review Reports - Assessing Glyphosate Injury and Forage Bermudagrass Regrowth Using Canopeo

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The glyphosate dose in lines 22–23 is inconsistent with the dosage notation used elsewhere in the manuscript;
“8, 16, and 24 after glyphosate application (DAG)” should be corrected to “8, 16, and 24 ‌days‌ after glyphosate application (DAG)”;
The Introduction should include an appropriate analysis of the current main applications, advantages, and limitations of the Fractional Green Canopy Cover (FGCC) method;
The correlation between Green Canopy Cover Reduction (GCCR) and herbicide dose for non-photosynthesis-inhibiting herbicides is debatable, and this method may not be applicable to all types of herbicides. The statement in the text — “This assumes that canopy discoloration (i.e., deviation from the green spectral band, ~500–750 nm) was caused by glyphosate application” — requires support from cited literature;
The content in lines 129–131 is redundant, as it repeats the description in lines 119–121;
Phrases such as “More details on the glyphosate treatments and experimental design can be found at ” are inappropriate; the experimental design and related information should be clearly presented within the manuscript so that readers can understand them at a glance;
The statistical method used in Table 1 — conducting significance tests on GCCR values across different varieties and dates — is inappropriate, as GCCR values from different varieties and sampling dates are not directly comparable;
The active ingredient amounts mentioned in lines 204 and 206 do not exist in the described experimental design;
The references are outdated; at least 50% of the cited literature should have been published within the past five years.

Author Response

Reviewer 1 Comments:

The glyphosate dose in lines 22–23 is inconsistent with the dosage notation used elsewhere in the manuscript;

Dosage notation was properly changed.

“8, 16, and 24 after glyphosate application (DAG)” should be corrected to “8, 16, and 24 ‌days‌ after glyphosate application (DAG)”;

“Day” was added between “24” and “after” in line 24.

The Introduction should include an appropriate analysis of the current main applications, advantages, and limitations of the Fractional Green Canopy Cover (FGCC) method;

Text was added in lines 59-72 addressing the current main applications, advantages, and limitations of the Fractional Green Canopy Cover (FGCC) method based on publications within the last five years.

The correlation between Green Canopy Cover Reduction (GCCR) and herbicide dose for non-photosynthesis-inhibiting herbicides is debatable, and this method may not be applicable to all types of herbicides. The statement in the text — “This assumes that canopy discoloration (i.e., deviation from the green spectral band, ~500–750 nm) was caused by glyphosate application” — requires support from cited literature;

Text was added in lines 151-155 to support the mentioned statement above.

The content in lines 129–131 is redundant, as it repeats the description in lines 119–121;

Text in lines 129-131 was deleted to avoid redundancy.

Phrases such as “More details on the glyphosate treatments and experimental design can be found at ” are inappropriate; the experimental design and related information should be clearly presented within the manuscript so that readers can understand them at a glance;

The inappropriate sentence was deleted. Also, I checked the entire Materials and Methods section and realized that all necessary information is presented. Thus, the deleted sentence was unnecessary.

The statistical method used in Table 1 — conducting significance tests on GCCR values across different varieties and dates — is inappropriate, as GCCR values from different varieties and sampling dates are not directly comparable;

We updated the Statistical Analysis (L174-178).

The active ingredient amounts mentioned in lines 204 and 206 do not exist in the described experimental design;

The active ingredient amounts were updated.

The references are outdated; at least 50% of the cited literature should have been published within the past five years.

All new added references were published in the last five years.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The article presents results from the use of Canopeo—a software tool for automated image analysis based on color thresholding—to determine the degree of phytotoxicity in ‘Goodwell’ and ‘Greenfield’ bermudagrass cultivars following treatment with the total systemic herbicide glyphosate at different doses. The study is innovative in nature and demonstrates the possibilities of using digitalization and modern software products to refine scientific research and to provide a more precise interpretation of certain results. Based on a detailed literature review, the main reasons for conducting the experiment are clearly outlined—namely, the subjective nature of visual assessment of herbicide-induced phytotoxicity, the presence of resistance in some weed species, and the possibility of overcoming phytotoxic effects in certain species at a later stage.

In the Materials and Methods section, the experimental design is described in great detail. The results are presented clearly and are illustrated with figures and one table.

Although the authors report some deviations in the assessment of phytotoxicity—such as overestimation at low doses of glyphosate or indication of weaker phytotoxicity at higher doses—the reported correlation between these results and the data from visual observations defines the approach as a relatively successful and innovative method for evaluating the effects of herbicide treatments.

I have the following questions for the authors:

Is it possible to use Canopeo under field conditions to determine phytotoxicity from herbicide treatments in cultivated plants and to assess the risk of unintended damage?

Does the sensitivity of the Canopeo method allow for the determination of lower herbicide doses that achieve the same effect as the recommended ones, with a view to reducing the use of certain active substances?

Author Response

Reviewer 2 Comments

In the Materials and Methods section, the experimental design is described in great detail. The results are presented clearly and are illustrated with figures and one table.

I have the following questions for the authors:

Is it possible to use Canopeo under field conditions to determine phytotoxicity from herbicide treatments in cultivated plants and to assess the risk of unintended damage?

At this point, we just evaluated it in controlled conditions – greenhouse settings. Further studies must be performed to determine its efficacy in field conditions.

At this point, we are unsure about it. Perhaps, evaluating the potential regrowth of the target plant and correlating it with Green Canopy Cover Reduction would be the next study.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This manuscript addresses a relevant methodological question and has practical value. The comparison between visual injury ratings and Canopeo-derived GCCR is potentially useful, and the dataset appears suitable for this purpose. However, the manuscript requires major revision before it can be considered further. My main concerns relate to methodological clarity, internal consistency, interpretation of method agreement, and English usage.

Major Comments

The manuscript relies too heavily on previous publications for essential experimental details. In lines 69–70 and 93–97, readers are referred elsewhere for core information on experimental conditions and design. Please include the necessary details directly in this manuscript so it can stand on its own. Also, please verify the citation in line 96, because the cited reference appears inconsistent with the context.
Lines 93–96 describe the study as a factorial in a completely randomized design, but the same sentence refers to “main plots” and “subplots,” which suggests a split-plot structure. Please clarify the true design, define the experimental unit, and ensure that the statistical model matches the design described.
Lines 88–89 state that two runs were conducted, but the statistical analysis section does not clearly explain how run was incorporated into the model. Please specify whether the run was treated as a random effect, fixed effect, or pooled after testing.
The manuscript’s central claim about agreement between visual ratings and GCCR is overstated. In lines 158–163, the text states that the intercept and slope were not significantly different from zero and one, respectively, but the reported p-values do not clearly support this interpretation. Please revise the statistical explanation and clarify exactly what hypothesis test was performed for the 1:1 relationship.
Figure 1 and the Bland–Altman analysis show systematic bias across different injury ranges, even though the two methods yield similar overall biological interpretations. Therefore, the language in the Results, Discussion, Abstract, and Conclusions should be moderated. The data support partial agreement for this dataset, not strict interchangeability across all injury levels.
The glyphosate rates listed in Methods are 0.39, 0.53, 1.06, 1.54, and 3.08 kg a.i. ha⁻¹, but lines 203–205 mention 1.37, 0.95, 0.47, and 0.35 kg a.i. ha⁻¹. These values do not match the treatment structure and should be corrected.

Minor Comments

The caption for Figure 1 should be corrected. In lines 166–169, “Goodwell €” should be corrected to “Goodwell (E).”
A contradictory sentence in the Results should be rewritten. In lines 177–181, the direction of bias for Greenfield is confusing and reads inconsistently. Please rewrite this section for clarity.
Table 1 needs formatting and explanatory improvement. Decimal precision is inconsistent in the table, and the note “Different letters represent significant differences among all columns” is too vague. Please standardize formatting and clarify exactly how mean separation was performed.
There are several typographical and wording errors. Examples include “24a DAG” (line 126), “fairly equivalents” (line 173), “CGGR” instead of GCCR (line 225), “prostate growth” instead of prostrate growth (line 247), and “The versa was true” instead of “The reverse was true” (line 268).
The Abstract should be revised for precision. The statement that the methods are “interchangeable for this dataset” is too strong, given the systematic bias shown later. Also, the phrase “Visual green canopy cover and GCCR ratings were collected” is imprecise because the visual variable is injury, not canopy cover.
Language editing is needed throughout. The manuscript would benefit from a thorough English revision to improve precision, readability, and scientific tone. Examples include informal wording, such as “In a nutshell,” in the Methods section, and several awkward phrases throughout the Results and Discussion sections.

Comments on the Quality of English Language

The English could be improved to more clearly express the research.

Author Response

Reviewer 3 Comments:

Major Comments

The manuscript relies too heavily on previous publications for essential experimental details. In lines 69–70 and 93–97, readers are referred elsewhere for core information on experimental conditions and design. Please include the necessary details directly in this manuscript so it can stand on its own. Also, please verify the citation in line 96, because the cited reference appears inconsistent with the context.

I checked the entire Materials and Methods section and realized that all necessary information is presented. Thus, I simply refer to the previous article, which used the same materials and methods. I deleted “in a nutshell” and “for more details…” because they were unnecessary. Any person can look at both manuscripts and see that the same information is described in both.

Lines 93–96 describe the study as a factorial in a completely randomized design, but the same sentence refers to “main plots” and “subplots,” which suggests a split-plot structure. Please clarify the true design, define the experimental unit, and ensure that the statistical model matches the design described.

The experimental design has been revised and is now accurate.

Lines 88–89 state that two runs were conducted, but the statistical analysis section does not clearly explain how run was incorporated into the model. Please specify whether the run was treated as a random effect, fixed effect, or pooled after testing.

There were two runs of five subjects each, for a total of 10 subjects per cultivar/rate combination. The 10 subjects were included as a random effect in the analysis. These were the subjects on which the three repeated measures (8, 16, 24 DAG) were measured. Cultivar, rate, and DAG were treated as fixed effects in the linear mixed-effects model for repeated measures. All this information was added to L174-178.

The manuscript’s central claim about agreement between visual ratings and GCCR is overstated. In lines 158–163, the text states that the intercept and slope were not significantly different from zero and one, respectively, but the reported p-values do not clearly support this interpretation. Please revise the statistical explanation and clarify exactly what hypothesis test was performed for the 1:1 relationship.

The passage in L184-189 was properly revised.

Figure 1 and the Bland–Altman analysis show systematic bias across different injury ranges, even though the two methods yield similar overall biological interpretations. Therefore, the language in the Results, Discussion, Abstract, and Conclusions should be moderated. The data support partial agreement for this dataset, not strict interchangeability across all injury levels.

Moderate language was used in the Results, Discussion, Abstract, and Conclusions

The glyphosate rates listed in Methods are 0.39, 0.53, 1.06, 1.54, and 3.08 kg a.i. ha⁻¹, but lines 203–205 mention 1.37, 0.95, 0.47, and 0.35 kg a.i. ha⁻¹. These values do not match the treatment structure and should be corrected.

The glyphosate rates were corrected in the entire manuscript.

Minor Comments

The caption for Figure 1 should be corrected. In lines 166–169, “Goodwell €” should be corrected to “Goodwell (E).”

“Goodwell €” was replaced with “Goodwell (E).”

A contradictory sentence in the Results should be rewritten. In lines 177–181, the direction of bias for Greenfield is confusing and reads inconsistently. Please rewrite this section for clarity.

The sentence was rewritten for clarity.

Table 1 needs formatting and explanatory improvement. Decimal precision is inconsistent in the table, and the note “Different letters represent significant differences among all columns” is too vague. Please standardize formatting and clarify exactly how mean separation was performed.

The decimal precision was standardized across the table. And the table note was updated.

There are several typographical and wording errors. Examples include “24a DAG” (line 126), “fairly equivalents” (line 173), “CGGR” instead of GCCR (line 225), “prostate growth” instead of prostrate growth (line 247), and “The versa was true” instead of “The reverse was true” (line 268).

All listed issues were resolved.

The Abstract should be revised for precision. The statement that the methods are “interchangeable for this dataset” is too strong, given the systematic bias shown later. Also, the phrase “Visual green canopy cover and GCCR ratings were collected” is imprecise because the visual variable is injury, not canopy cover.

The passage “…and are considered interchangeable for this dataset.” was deleted.

Language editing is needed throughout. The manuscript would benefit from a thorough English revision to improve precision, readability, and scientific tone. Examples include informal wording, such as “In a nutshell,” in the Methods section, and several awkward phrases throughout the Results and Discussion sections.

The article was proofread and modified to improve clarity.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

The study evaluated Canopeo-based green canopy cover reduction (GCCR) method as an alternative to traditional visual ratings for assessing glyphosate injury in two bermudagrass cultivars. The results indicated strong correlation between the two methods and suggest that both the methods can be used interchangeably. The methodology is straight forward and simple, however, authors refer to their previous publication very often for ‘more information’. If the information is relevant in this study, then the authors should provide all details in this paper as well. The statistical analysis done for correlation is appropriate. The use of Bland-Altman Analysis is a great way to show the systemic bias and under/overestimation of glyphosate effect on bermudagrass cultivars. The results and discussion sections are short and to the point, however the authors could cite more studies that have compared visual ratings and GCCR methods.

Limitation: The authors did not address the growth reduction and physiological changes in the grasses after glyphosate application. GCCR is primarily based on color differentiation while traditional Visual injury accounts for discoloration, growth and any physiological changes. Since glyphosate shows these effects around 14 days after application, the methodology should include quantitative measurements along with color differentiation. Additionally, the authors should clarify whether they included growth reduction and physiological changes in Visual injury evaluation and add to the methodology as needed.

Author Response

Reviewer 4:

I checked the entire Materials and Methods section and realized that all necessary information is presented. Thus, I simply refer to the previous article, which used the same materials and methods. I deleted “in a nutshell” and “for more details…” because they were unnecessary. A person can look at both manuscripts and see that the same information is described in both.
Unfortunately, this is the first paper using GCCR. We developed and are exposing the GCCR methodology in this paper.

Limitation: The authors did not address the growth reduction and physiological changes in the grasses after glyphosate application. GCCR is primarily based on color differentiation while traditional Visual injury accounts for discoloration, growth and any physiological changes. Since glyphosate shows these effects around 14 days after application, the methodology should include quantitative measurements along with color differentiation. Additionally, the authors should clarify whether they included growth reduction and physiological changes in Visual injury evaluation and add to the methodology as needed.

We did not include physiological changes and growth reduction in our visual assessment. We simply assessed plant necrosis as stated. We will add these factors for future studies.

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

Thank you for revising the manuscript. The revised version addresses several of the earlier mechanical issues. For example, the incorrect glyphosate rates in the Results have been corrected, the Figure 1 caption now uses “Goodwell (E),” and the “24 DAG” typo appears fixed.

However, I do not think the revision has fully resolved the main methodological and interpretive concerns.

First, the manuscript is still not fully self-contained. The Methods continue to refer readers to a previous paper (“The experimental conditions are described by [21]”), which was the core of my original concern. Although more information is now available, the study still partly relies on another publication rather than standing on its own.

Second, the revised experimental design section remains problematic. The sentence beginning “The experimental design was a factorial arrangement in a completely randomized design...” is followed by an incomplete fragment: “For the 2 × 5 factorial (excluding both controls, one for each cultivar) with three DAG repeated measures.” This is not a complete design description. It also remains unclear exactly how the controls were handled in the inferential model.

Third, the handling of the two runs remains insufficiently clear. The revision now states that there were two runs and 10 random subjects per cultivar/rate combination, but it still does not explicitly state whether the run itself was modeled as a fixed effect, a random effect, or a blocking factor, or whether the two runs were pooled after testing. This should be stated directly.

Fourth, the statistical interpretation of the regression/1:1 comparison still needs clarification. The revised manuscript now states that the intercepts and slopes were “significantly different from zero and one, respectively,” but no explanation is provided for how the difference from one was tested. If the reported p-values came from a standard regression output, they typically test the slope against zero, not one. This remains a key point because it affects the main claim regarding agreement between methods.

Fifth, the language regarding the agreement between methods remains too strong. In the Results, the authors state that the methods were “fairly equivalent”; in the Discussion, they still describe the methods as “interchangeable”; and the Conclusions repeat that both methods were considered interchangeable for this dataset. Given the Bland–Altman trends and only moderate correlations, the manuscript supports similar general biological interpretation, but not strict equivalence or interchangeability. This concern remains unresolved.

Finally, the manuscript still needs further English editing. Several awkward or inaccurate phrases remain, including “Visual green canopy cover and GCCR ratings were collected,” “fairly correlate,” “lesser than 40%,” and “did not statistically follow the 1:1 line.” These are not only stylistic issues; some of them affect the precision of meaning.

Overall, although the revision improves the manuscript in several respects, I do not consider the key concerns fully resolved. A further round of substantial revision is still needed.

Comments on the Quality of English Language

The English could be improved to more clearly express the research.

Author Response

However, I do not think the revision has fully resolved the main methodological and interpretive concerns.

Although we kept the sentence: “The experimental conditions are described by [21]” after revision 1, all necessary information was presented (except for the statistical analysis deficiency mentioned below). Please recheck the Materials and Methods and let us know which specific information is missing that would prevent understanding and replication of the study. We deleted the sentence mentioned above if that was an issue.

The incomplete sentence was deleted and replaced by a full description of the experimental design. Please note that the Experimental Design and Statistical Analysis was moved to item 2.5. Also, we rerun the PROC REG. The SAS output is attached.

The issue stated above was resolved L158 – 168 (file with track changes).

The issue above was resolved L172-173 (file with track changes).

We removed “fairly equivalent” and “interchangeable”.

We reread the article and made appropriate changes where we judged necessary. If you have any further concerns, give us the specific lines and suggestions for improvement.

Overall, although the revision improves the manuscript in several respects, I do not consider the key concerns fully resolved. A further round of substantial revision is still needed.

Author Response File: Author Response.pdf