Developing Indicators of Nutrient Pollution in Streams Using 16S rRNA Gene Metabarcoding of Periphyton-Associated Bacteria

Round 1

Reviewer 1 Report

This study is interesting as a new methodology of defining indicator species or biological metrics of aquatic pollution. But, the data are not suitable to examine the effects of nutrient pollution on bacterial community. First, this study didn’t examine stream basic condition (e.g., size, slope, flow, light) and water/bed quality other than nutrients (e.g., organic matter, turbidity, quality of periphyton), which often correlate to periphyton and nutrients. Second, all samples should be independent each other basically in such analysis, but repeatedly collected samples were used in this study.

Line 18: …. to quantify effects.

I don’t think this study quantified effects of TP and TN. Need modification.

Line 26: Changes associated with TN gradually occurred from 275-855 µg/L.

Is the change gradual? From 275 to 855?

Line 28: Names of different taxonomic levels are included. I guess Proteobacteria should be provided in prior to the order name.

Line 31: I don’t think this study highlight quantification of biotic response. Need modification.

Line 71: A reason why this study focused on periphyton is needed.

Line 89: How were tree shades, turbidity, flow velocity, and recent flood history, which are important determinants of the development of periphyton, sometimes more important than nutrients.

Line 91: How did authors treat with different n between periphyton and nutrient samples?

Line91: Each site had 10 – 14 samples. So, the data are not independent. n should be 25 to be independent.

Line 94: I guess the five periphyton samples were pooled. This study didn’t provide periphyton basic characteristics such as Chl-a and inorganic %. It is unclear how nutrients can affect bacterial community.

Line 108: How about green algae and cyanobacteria in the periphyton. Is the diatom dominant?

Line 165: both first and second axis scores?

Line 165: Please show what are the environmental variables.

Line 166: Please show what are the multivariate statistics.

Line 166: It is unclear whether different criteria of species were used for different analyses and why.

Line 198: how about watershed area? Nutrients are the focus of this study, but we also doubt about spurious relationship.

Line 241: I want to know frequency distribution of TN and TP to understand whether it affected some of the results of indicator analyses.

Figure 2: It is not clear how the vectors were drawn. Spearman correlation coefficients of axis 1 and axis 2?

Table 1 and 2: It is better to show these tables after showing all relevant analyses. It is very difficult for me to follow these tables here. What is the purpose of summarizing? Does Figure 6 show the same thing?

Figure 3: Labels for the y-axis of b and d help understanding.

Table 4: What is CV correlation? How it is calculated.

Line 276: In my PC, it sometimes shows @g TP/L. Correct?

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 2 Report

This is an interesting study, the experiments were well conducted, and analysis was well performed. Overall, this is a high-quality manuscript.

Author Response

We appreciate the positive feedback.

Reviewer 3 Report

The study demonstrates clearly how the composition of the bacterial microbiome in water could indicate different nutrient status levels. The research methodology is convincing and the results are justified by the evidence. However, I recommend adding some more discussion about the role of bacterial metabolic activities, because the proposed methods are not taking this into account while there should be some interrelation between nutrient level and growth (metabolic activity) markers (not being used). 

Author Response

We appreciate the positive feedback. We appreciate the suggestion, but we feel adding further detailed discussion of metabolic activities and their relation to growth were beyond the scope of our study. 

Round 2

Reviewer 1 Report

This study is interesting and worth publishing if the analysis was done appropriately. I still don’t understand how samples from the same site could be independent each other as same as the samples from other sites. I recommend to reduce samples size to 25 (number of sites) if the same analysis is possible or show some logical assumptions to use all data as independent to avoid future criticisms after publication. In addition, because the effect of confounding factors were not surveyed, I don’t recommend to strongly mention about the effects of TN and TP. If the problem has been already solved in your past studies, this is not the new finding of this study.