Monitoring Seedling Emergence, Growth, and Survival Using Repeat High-Resolution Imagery

Round 1

Reviewer 1 Report (Previous Reviewer 4)

The authors have done a good job to improve the paper. The authors’ response is reasonable. Thanks for your revision.

Author Response

Thank you for your time and effort reviewing our paper. We appreicate your input from the first revision. We are pleased that you felt that we had addressed all of your comments sufficiently. 

Reviewer 2 Report (New Reviewer)

General comments:

This is a very useful and timely study examining the ability of remote cameras to measure the height, plant density, and rate/cause of mortality of individual planted grasses in an experimental field setting. Using a combination of frequent field observations and cameras placed in the plots the authors do a nice job of showing the success, downsides, and tradeoffs of camera traps for plant measurements and capturing herbivory events. The results will be of interest to researchers and managers alike looking to capture fine-scale plant characteristics over time.

My major concern with the manuscript is that there are places in the Results where the figures and text contradict themselves. For example the text says that height estimates from the cameras are underestimating the true heights, but then some of the figures and later text show that the cameras are in fact overestimating height. I have highlighted this more explicitly in my line comments below. This could be due to misinterpretation on my part, and if that is the case I apologize, or perhaps to mislabeling of figures on the author’s part? Or maybe it’s the result of pooling data for certain figures differently than for others? Whatever the explanation, right now it looks to me as if the figures and text are not consistent in some fairly major and fatal ways, and that needs to be double checked and addressed before the manuscript be considered at all acceptable for publication.

 Minor comments:

 

·        Multiple of your figures have titles in the figure panels themselves, I suggest you remove those as it is general publishing convention to not have panel titles on figures. That information is what figure captions are for, the additional text on your figure is just making it more cluttered.

·       In your discussion you mention the potentially prohibitive cost of the 28 cameras in the study, but what about the feasibility of needing that density of cameras when we’re talking about more large-scale monitoring?

Line comments:

Note: Because there are no line numbers I will be referencing “line comments” by the section and paragraph number, hopefully that will be sufficiently clear!

Intro, first paragraph: The “In contrast to plant communities…” line is a little confused, right now it reads to me as though the only way invasive annual grasses alter ecological processes is by causing a change in fire regime? Granted, impacting fire regimes is an important change, but invasive annual grasses can change species composition, percent bare ground, and any number of other things independent of fire regimes.

Intro, last paragraph: The beginning of this paragraph feels very repetitive. You don’t need to tell us that you were interested in height, density, and mortality so you tested the ability to capture height, density, and mortality by measuring height, density, and mortality. 

Methods: Maybe I missed it, but I don’t see where you tell us how many plots you had? I’m assuming one plot per camera, so 28 plots total? But I should not be forced to assume that! Please tell us explicitly, something like “we had 28 plots that were some meters by some meters in size and at each plot we placed a single camera.”

Results: Are all of these figures showing your results AFTER “accounting for underestimation” of height in your picture estimates? If so that needs to be stated very clearly in your Methods and probably again at the beginning of your Results, because that’s very different than if these were raw, uncorrected results. Also, in your Methods you need to explicitly say how you “accounted for underestimation.” Did you adjust for the fact that your data show a difference in the underestimation based on date, or did you simply adjust all your picture estimates by 14% as you state in the beginning of your results?  

Results, Figure 2: It would be helpful to the reader if you plotted a 1:1 line on the graph, since it would allow for easier visualization of how much under vs over-estimation is going on.

Results, second paragraph: “Measurements were underestimated at earlier dates by as much as 27% (14 June) and became more accurate over time with a difference of 8% on July 5 and no difference between field and image estimates at the last measurement (19 September, Figure 2).” I suspect you meant to cite Figure 3 in this sentence? More importantly, this sentence is directly contradicted by these statements in the same paragraph when you say “Fenced plot estimates of average seedling height were overestimated by approximately 2 cm in June and became more accurate over time (Figure 3B) while unfenced plot estimates were consistently overestimated by about 2.5 cm (Figure 3C).” You also cite Figure 3 when I think you meant to cite Figure 4, since Figure 3 does not have multiple panels. Finally, you say “Unfenced image estimates were 2.9 cm closer to field measurements than fenced plot estimates, a difference of 16%” but in Figure 4B and C it looks to me as if the fenced picture estimates (Fig 4B) are a lot closer to the field estimates than the unfenced picture estimates (Fig 4C).   

Results, Figure 3: I find the wording of your caption slightly misleading, when you caption a figure “The difference between heights” then I’m expecting the figure to be plotting the actual difference (i.e., photo height – field height) between the measures. I suggest changing the caption to something more light “Impact of date on accuracy…” or “Variation in estimate accuracy…” or actually you could just use your figure title (which needs to be removed) in your caption.

Results, Figure 4: Is panel A showing the difference in plant height between photo vs field measures, as in those plotted numbers are field measure minus photo measure? If so you need to clarify that in the figure caption, and if that is the case then isn’t panel A showing pretty much the same thing as panels B and C (so maybe you could just have panel A)?    

Results, “Seedling Density”, first paragraph: “5.3 seedlings ¼ m^-2” what is this density estimate? I highly recommend you give density in seedlings per square meter, not per square quarter meter. Please don’t make your readers do mental math to make your results comparable to the rest of the literature, there's a reason why most manuscripts stick to a default of density per square meter or per hectare!

Results, “Seedling Density”, second paragraph: “Row order from the camera did not affect accuracy of image estimates for seedling density (p=0.069, Table 2).” This is not quite a true statement, since the interaction of row order and date IS significant so row order does in fact impact accuracy in some capacity. It would be clearer to say that “Row order alone..." or "Row order independently..." did not significantly impact density.

Results, Figures 7 and 8: Your axes are labeled “plants per square... ” should be “seedlings per square... ” to be consistent with how you reference density in the rest of the paper.

Discussion, second paragraph: Why are the last sentences in italics?

Discussion, third paragraph: I wonder if, in addition to being confounded by density, the impact of date on accuracy is partially because the larger seedlings are easier to see and measure than smaller seedlings?

Discussion, Table 5: I’m a little confused why you’re introducing new results here instead of having this table in the Results section?

Discussion, last paragraph: “If research is concerned with frequent temporal changes, we recommend that the camera take frequent time lapse images. If temporal change is less frequent, adjust the camera accordingly.” These suggestions are so incredibly vague that I am not sure they are adding much to your discussion, I suggest you cut them. In fact, this entire last paragraph is quite vague and I think a rather weak ending to this section. I would suggest that you could cut the entire paragraph and end with the much stronger and more helpful previous paragraph covering your suggestions on camera placement.

Author Response

Thank you so much for going through this paper and providing meaningful and helpful comments and edits. We have gone through each of your comments and updated the manuscript accordingly. These comments have really helped us improve this paper. Please see the attached file with the specific response to each of your comments. 

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report (New Reviewer)

I really appreciate the author's positive attitude and thoroughness in addressing my comments, thank you! And I am very relieved that the discrepancies I found in the first draft were a simple mislabeling and clarity issue that has been completely addressed in the revised version. This is a very interesting paper and I'm excited to see how camera trapping can be used in the future for plant monitoring.  

I do have one last revision though... in my version of the revised manuscript it looks like there's a rogue citation on the line above the "Reference" heading. 

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

This manuscript proposed a new method of using triggered cameras to monitor plant growth status. Combined with image processing technology, the growth indicators such as plant height and density could be extracted by this method, as well as herbivorous events detecting through time-series plant images to remotely analyze the sources of plant damage. The nondestructive measurement and remote control are worthy of affirmation, which is of better practical significance. However, there also remain to be some problems noted in the writing process of this manuscript, as follows:

 

1. The extraction method of plant height and density requires to be described in detail, such as schematic diagrams.

 

2. When comparing the measured plant height and density data with the image extracted data, utilizing scatter diagrams and calculating the evaluation indicators (such as R² and RMSE) may make the data analysis more statistically significant.

 

3. When referring to the error source of extracted plant height, the manuscript talked about the impact of herbivorous events, which may also be related to fences, but this impact was not discussed in the comparison of events.

Reviewer 2 Report

Dear authors,

thanks for your paper, I would like to have number of recommendations. All paper is written a little like some technical report, not as scientific paper.

I would like to see more detail scientific approach in your paper.

1. Please specify more detail your scientific goals in introduction
2. Methodology needs to be extended and probably will be good to move Figure 8 into methodology and describe in details all approach of your work
3. For the results will be good to extend a little every results, give some introduction
4. Extend please all text in results, there is number of images and tables, but better explanation is needed
5. Conclusion need to be extended.

Reviewer 3 Report

The manuscript by Jesse R Morris et al. “Evaluation of Remotely Triggered Cameras as a Method for… ” presents an interesting study conducted on experimental fenced and unfenced plots to better understand potential and limitations of proximal sensing techniques for measuring dynamics of vegetation to be used for degraded rangelands restoration. An appropriate statistical analysis allowed to compare direct measurements concerning height, density, and mortality recorded in the field and information derived from images automatically acquired by remote cameras.

The study presented is interesting and the manuscript is well organized and clearly presented.

Minor comments:

• In figure 1 you show photo of the plot setup with camera; it would be interesting to show also one/two picture acquired by this camera as an example of remotely sensed images to be processed.
• In your experiment 28 cameras were all placed parallel to the ground: on this basis analysis is conducted and results are derived.  In  Discussion paragraph (Lines 428-430) it is recommended, for specific conditions, to place “camera above the plants, perpendicular to the ground”. Can you specify foundation/bibliographic reference in support of this recommendation ?

Reviewer 4 Report

The manuscript describes a camera system for seedling growth monitoring. The method is thoroughly described. The result is clearly presented. The paper is well-developed and easy to follow. However, I think this work does not fit the scope of the journal. The main feature of remote sensing is using the imaging/camera system, typically equipped on satellite or aircraft, to detect Earth's surface over large area. The spatial coverage of remote sensing data is normally at the regional or global level. However, the method described in the paper is an in-situ camera system, which may be applicable to smallholders but difficult to be deployed and applied in a large area. Although the cameras are "remotely triggered", it is not regarded as an application of remote sensing technology in my point of view. The authors may consider resubmitting the paper to a photogrammetry or Ag-related journal.