Investigation of Pathogenic Bacterial Transport by Waterbirds: A Case Study of Flooded and Non-Flooded Rice Systems in Mississippi

Round 1

Reviewer 1 Report

The authors present results from a well designed study investigating the influence of winter flooding on bird-mediated pathogen transport in two different rice production system types. Overall, I liked this study but feel the presentation of the results could be improved. There are a few larger issues that if resolved will make this paper suitable for publication. I list them below, followed by detailed comments.

1. Both in methods and discussion the lead author's thesis is cited. I am not sure this is appropriate. The data on bird densities and estimated fecal loads seems highly relevant to this study considering it is the backbone of what is driving the response variables. In the methods, details on how fecal loads were estimated from camera trap data need to be included rather than just referencing the thesis.
2. It is not clear why you chose to use PERMANOVA to assess seemingly univariate data. PERMANOVA was designed for distance-based analysis of primarily ecological community data. You are not assessing the whole bacterial community, just two species. I don't think PERMANOVA is providing any additional info that the ANCOVA are not. If you are concerned about assumptions there is a myriad of variance structures and data distributions that could be applied to count data to address issues. See "Mixed Effects Models and Extensions in Ecology with R (Zuur et al. 2009)" or similar references. 
3. The results are extremely brief. Please expand your description of your results some, including moving the no detect info for some pathogens from the stats methods to results, and include bird and fecal load estimates across treatments. You mention them and even present a figure in the discussion. This is results.
4. The discussion is lacking. There is little effort to link the results of this study to other literature. I have made a couple specific suggestions for expansion but there are many opportunities beyond that to broaden the discussion and include literature.

Specific comments.

L2: The title suggests you found potential for pathogenic transport. I would frame the title more like a headline that conveys the actual message of your findings.

L19: but only significantly greater than conventional non-flooded. Consider writing this in terms of the observed gradients with the endpoints that are significantly different.

L32: change “have long been” to “are”

Line 40-42: This sentence is awkward. I think the "long been recognized" language here and elsewhere could be dropped. Try and write more directly.

Line 49-50: Is it collected, translocated, or recycled? N and P bound up in organic matter is processed by waterfowl, and what is not used for growth and maintenance is excreted. How much consumption occurs outside of field and is translocated in, or vice versa, or recycled from within field? I don't expect you to answer that, just recognize it is more complicated. Reference you should consider: https://doi.org/10.1016/j.scitotenv.2014.12.068

Line 52-55: I would split this into two sentences, 1 on potential benefits, 1 on potential detriments.

Better yet start this paragraph out with a contrast statement about potential agronomic benefits vs. potential environmental impacts then back up the statement with this info.

Line 55: Is the term "free living" necessary? Is this opposed to caged?

Line 67: Is the term indicators neccessary? Are the indicators a cause of impairment or just a way to measure things?

Line 74: Another way to say “of pathogenic impairment to” might be "increasing the likelihood of pathogen transport to"

Line 75: The second half of the sentence requires you change “detect and characterize” to "characterize and compare". You are making comparisons among treatments not just random observations.

Line 77: Change The primary… to “Our primarly…”

Line 78: Consider changing “resulting from” to “associated with”

Line 78-81: The way your hypothesis is currently structured makes it sound like you actually controlled bird densities. A better way to do this would be to structure a hypotheses about bird densities associated with different flood/management treatments coupled with a hypothesized linkage with pathogen levels/transport. Put another way are you actually keeping flooded fields devoid of birds?

Line 112: Why just March? It would be good to know the timing of potential pathogen transport.

Line 117-123: I am assuming that this whole study came out of the lead author's thesis so I am not sure why you are citing it as a previously published entity here. Why not just say "We estimated fecal matter inputs....And also give us the details on how you went from camera trap info to fecal matter mass.

Line 161: What are the units for the data? Are all species on the same scale?

Line 167-169: Should these statements be in the results somehow? I wouldn’t mention it here unless it changes how you would have analyzed the data.

Line 169: Can you better justify this. Do you have a reference you can cite to justify higher alpha?

Line 171: Why are you using PERMANOVA? This analysis was designed for community data. You are only looking at 2 bacteria not the entire community? What does this approach gain you that the individual pathogen load ANCOVAs does not? If you are using PERMANOVA you should cite Anderson's paper/papers on PERMANOVA but I am not sure she would agree with you using it here.

Line 174: Previously you said Enterococci was excluded due to detection in only 1 sample.

Line 179: Need a description of what your group variable was with Mann-Whitney-Wilcoxon Rank sum test. I don’t think it was the same as previous analyses.

Line 181: I am all for brevity but surely you can describe your results with a little bit more detail. See later comments on bird density/fecal load data too. Why is this not in the results?

Table 2: The table is a bit hard to read. Is there a way for the authors to delineate columns associated with treatments better? Or better yet, scrap table 2 and 3 and provide bargraphs which would be easier for the reader to quickly interpret your results. Given color is free with this journal why not jazz it up a bit.

Simply bolding the two treatments that were signficantly different from each other is not intuitive nor does it tell the whole story. Please use traditional approach of letters (A, AB, AB, B) to show how treatments significantly differed or did not differ.

Discussion in general: There is not much effort to link results to literature in this discussion. There are very few new references incorporated into the paper through the discussion which suggests the discussion needs to be expanded.

Line 205-207: I am having a hard time understanding why some of the methods and results are referencing a thesis. Your results section is 3 sentences long. Why not incorporate some of the other related results on differences in birds/fecal inputs, and bacterial measures into this manuscript to beef up the info. Is it planned for a second manuscript?

Line 213: Reference the reader back to table 2 or future bargraph.

Line 218-220: Maybe describe this as a gradient with LF having the highest and CN having lowest and significant differences.

Line 222-224: Need to discuss possible mechanisms for higher birds and bacteria levels in LN vs CN? What is different about these treatments? Are there potential temporal (flooding history) or spatial (Proximity to LF) at play here? This is one example where the reader is wanting more thoughts about what your results mean.

Line 232 and Figure 1. Whoa. You are presenting new results in the discussion. Please move this to results.

Should have been presented first in results before referencing in discussion.

Line 242: Are there other ways microorganisms are removed from the water column. Microbial food webs, etc. Another opportunity to discuss your data more.

Line 247: Do you have water samples through time? These riser board systems are pretty leaky and still release pulses if storm events push more water through than the system can handle. What if pathogen levels were higher earlier and water was flushed out during storm events?

Line 254-257: Given your results and the current literature, would it be possible to develop a model that predicts pathogen levels with bird density and estimate flock sizes that would be potentially hazardous within your system? Something worth considering or at least mentioning as a future research effort. I think the goal is not to do regular monitoring but develop tools that tell managers when to be concerned and start monitoring.

Author Response

Please see the attachment. 

Author Response File: Author Response.pdf

Reviewer 2 Report

The manuscript Potential for pathogenic transport by waterbirds: a case study of flooded and non-flooded rice systems in Mississippi describes work in which the authors sampled sediments and water from two rice fields under different management regimes, and tested the samples for bacterial indicators and known pathogens.

The manuscript is well-written. However, I am concerned about the lack of replication of the sampling. Sediment/soil samples were collected at two time periods, before and after flooding – but were not replicated over time. Water samples were collected only once and, once again, were not replicated over time. Thus, how confident are we that the results are temporally reproducible? The authors themselves state that “pathogen levels can fluctuate substantially in bird flocks between seasons” and thus, a single season of monitoring represents a good start, but one that must be replicated for scientific relevance.

 

 

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

The study conducted describes a comparison between different rice systems at the level of microbial contamination. The work shares some important results on the contamination of the different rice systems soils and the potential impact of flooding and waterbirds. The paper is generally well written but needs further work, especially in the results section that it's mainly tables presenting very few explanations. Results section should be explored a bit further. Moreover, it is difficult to understand the two description of the microorganisms chosen. Sometimes the authors use fecal indicator bacteria, other times pathogens and sometimes use both. The authors should state which microorganisms they are considering as pathogens (possibly Salmonella app. and Campylobacter spp.) and which ones are the fecal indicators (the remaining probably). It is also not clear why the authors chose not to analyse Salmonella spp. and Campylobacter spp. in the water but only in the soil.

The authors should describe the LEISA rice system in the introduction or where they feel it's most correct to those not working with this type of system.

Minor comments:

• always use spp. following Salmonella, Campylobacter and Enterococci if not looking for a certain species;
• In the abstract section, for instance, the authors write in line 19 abbreviations of Clostridium perfringens and Escherichia coli without presenting them before and the same occurs in the introduction section. Please correct this writing first in full and with the abbreviation in brackets and use always the same abbreviation for the organisms. For instance in table 2 Clostridium perfringens that was introduced before as C. perfringens is in the table introduced as C. perf. The same designations should be used throughout the text to make it easier for the reader.
• discussion section line 204 remove the "s" from increases.

Author Response

Please see the attachment. 

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Thank you for doing an excellent job addressing my previous comments. The manuscript is much improved. I only have a few minor revisions and one comment/suggestion for future work.

Your new figures really help the reader follow the story better.

Fig 1A-B. The bold letters and the thick lines for error bars look weird.

Line 192-204: In this section you discuss results related to Fig 1A then jump to Fig 2A and 2B before returning to Fig 1B. My question is why is Fig 1 A and B together. I think you could leave the text the same but make figure 1A a stand alone Fig1, leave Fig 2A-B, and make Fig 1B actually Fig 3. This would flow better.

Current Fig. 3. This figure is pale and small and a bit fuzzy. Was it also made in R or was it made in excel. I would make it bigger, bolder, similar font size, etc. to other figures so it matches in quality.

 

General comment on stats: While you didn't do anything inherently wrong, there are better, more up-to-date methods for analyzing your data. I am not asking you to revise your stats at this point, but in the future, when analyzing count or density data, please consider exploring linear mixed model approaches with different distributions, variance structures, etc. rather than transforming data and using non-parametric anovas. Mixed effects models and extensions in ecology with R is a great resource for understanding the problems and approaches with analyzing this kind of data.