Benthic Macroinvertebrates and Zooplankton Communities as Ecological Indicators in Urban Wetlands of Argentina

Round 1

Reviewer 1 Report

The study describes zoobenthos and zooplankton communities of 6 Argentinian ponds (sometimes called wetlands in the text), aiming at finding water quality indicators for South American fresh waters and assess the effects of water quality on these communities (which seems a better term than assemblage, used in the text).

However, the problem is that the study promises more than it gives, in my opinion being just a purely descriptive survey, one of many similar ones carried out in the past, and not addressing the goals stated in the Introduction.

Major issues:

1. The study is missing the analysis of some independent indicators of ecosystem quality, that might show whether the assessment made on the basis of the communities is correct or not. As it is now, the quality is assessed on the basis of biotic indices and then it is concluded that biotic indices performed well in determining the quality. This looks like eating your own tail… The criterion to assess the index performance seems to be just whether it gives a strong discrimination between the ponds: but what if this discrimination was false and the ponds in fact were of the same quality?
2. Expectations stated at the end of the Introduction: #2 is pretty trivial, #1 and 3 seem contradictory (habitat quality is as it is, only we want to assess it properly on the basis of organisms: so, either both assemblages show the same, or one of them is wrong…), #4 was not tested (I can’t see any common analysis of both communities except CCA, but it is not interpreted in terms of ecosystem quality)
3. Critical methodical details are missing. These include: (i) Seasonality of sampling – nothing is said on this in methods, suddenly seasons appear out of the blue in result reports (though at least months should be mentioned. (ii) Why were different ponds sampled in different seasons (Table 1)? This seems to make ponds difficult to compare with one another, particularly if there was a seasonal variability of data. (iii) On which dissimilarity index was the PERMANOVA based? Bray-Curtis? (iv) I don’t understand the use of PERMANOVA for abiotic variables. This can only tell there are some differences in some parameters, nothing more… Why not to test each variable separately with ANOVA to see which of them make a difference? (v) The ponds are not sufficiently described: what are their sizes, surrounding areas, distance to urban areas, trophic status, human impact, etc.? (vi) Add more details on sampling sites: distance from shore, justification of selection… Are the depths in table 1 typical for a given pond (like mean, max, etc) or specific for the sampling sites? If latter, why are they so different? (vii) What was the depth of zooplankton and water sampling? (ix) Is anything known on the fish communities in the ponds? If they are different for some reason, it can also affect invertebrate communities

 

Other issues:

Page 3, line 3. “are” is not necessary before “they”

Page 3, line 11 from the bottom of Introduction. The term “functional response” has a precise definition in ecology and this is not what you tested here.  

Page 3, line 5 in Methods. S1-S4 vs. RNUO: you state you are going to use S1-S4 names in the text, but they are used interchangeably, which causes confusion. Limit to one version and be consistent

Page 4, line 1 in Environmental variables. Why did you think temperature would be a good parameter differentiating ponds located in a short distance from each other? It is mainly dependent on season, day and some casual changes, like clouds covering the sun… Moreover, abiotic parameters seem pretty basic for the analysis aiming at assessment of ecosystem quality…

Page 5, line 1. I would say B-C distance shows differences in relative density (composition) rather than just abundance (absolute)

Page 5, line 2. What were the replicates in this analysis? If samples collected in different seasons, then what about interseasonal differences and the fact that not all ponds were tested in exactly the same seasons? This may cause that for instance a pond tested in summer can differ in temperature from a pond tested in winter…

Page 5, line 4. Say here what the factors in this analysis are

Page 5, line 1 in Results. Please provide numerical results for the PERMANOVA (as you did for communities)

Page 5, line 3-5. What is the basis for these statements? This is not directly shown by a multivariate analysis, such as PERMANOVA. Also, it seems pretty trivial that temperature in summer is higher than in winter…

Table 1. Please note that if sampling depth = Secchi depth, then you actually don’t know what the real Secchi is…

Table 2. Why did you report organic matter as proportions of its different types (summing up to 100%)? It would be much more important to show real amounts of OM, in mg/ml or %, but of the total sediment mass, not of the total OM mass. If an animal feeds on, say FPOM, it is important for it how much of this resource is available, rather than what is its percentage among other OM types.

Page 6, last line above Table 3 caption. Why do you consider this particular value as “very high”? What is the threshold for “high”, “very high”, etc.? I guess we can only consider these values relatively, i.e. one is higher than the other

Page 7, paragraph 1. How is it possible that such F values (which are ratios of between group vs. within group dissimilarities) produce significant results? Also, PERMANOVA doesn’t test densities.

Page 7, line 4 above Fig. 2. Transparency doesn’t seem to have a very long vector in Fig. 2, I would say sand content si more explanatory. Also, these should be considered separately for each axis.

Page 7, lines 1-3 above Fig. 2, and continued text on the next page. This looks completely messed up. I can’t see this in Fig. 2 at all. For instance, Flon and Aspl are located around RNUO sites (not PGB), Chironomus is stated to be typical for PGB (page 7) and then S1-3 (page 8), whereas in fact it is located near the center of the plot and does not discriminate anything, etc. Please have another look at this figure…

Fig. 2 caption. This is not a triplot, as environmental variables are shown in a separate plot. Also, on which basis you say that some variables are significant (any test)? What is the meaning of different colours in panel B?

Page 8, line 3. On which basis you state they these taxa were sensitive?

Page 8, line 6. Please note that sporadic species can be located randomly on the plot, selected by the algorithm as the “most explanatory” ones, which in fact results from their casual occurrence in one sample and absence in the others. Thus, they should not be attributed with too much attention in the interpretation of results

Page 8, line 6 in Functional groups. What do you mean by “lost dominance”? did they dominate previously and the situation changed?

Page 8, line 9 in Functional groups. Density cannot be expressed in %

Page 9, line 2 in Biotic indices. What do you mean by “regular”?

Page 10 and following. In general, the discussion is full of results. They should be reported in Results section, here provide only interpretation and literature background. You don’t need to repeat numerical results in discussion, and certainly cannot add new ones.

Page 10, line 3. Why “despite”? I would say the first part of the sentence is the reason for what you propose in the second one. Anyway, I don’t see any basis for drawing conclusions on ecosystem services from your study

Page 10, line 5. Please be careful with the word “effect”, it is very difficult, if not impossible at all, to show effects in a correlational study as this one. Anyway, I can’t see such a discussion in the following text.

Page 10, line 1 in Environmental variables. Truism

Page 10, line 3 in Environmental variables. Do you mean that pH is unstable or that it differs from pond to pond?

Page 10, line 5-6 in Environmental variables. What do you mean by “edge of waterbodies”? Nearshore? If so, I don’t understand discussing depth here, it would be another truism…

Page 10, line 10-12 in Environmental variables. Maybe, but this is not shonw in your resuls (e.g CCA) at all

Page 10, line 3 in Benthic… Most polluted based on which criterion?

Page 10, line 1 of the last paragraph. Considered tolerant based on what?

Page 11, line 2 of the last paragraph. “filter feeders” rather than “filters”

Page 11, line 5 of the last paragraph. Abundance needs a unit

Page 12, line 8 in Biotic indices. How can this be compared to the results of PERMANOVA? The latter just shows there are some differences between the communities

Page 12, line 2-3 in paragraph 6 of Biotic indices. What do you mean by “different… but consistent with stat analysis”?

Page 12, line 3 of the last paragraph. Did you determine any threshold?

Page 13, line 8-9 below Fig. 5 caption. I cannot see any basis for such conclusion in your results.

Conclusions. (1) Not possible to draw without an independent assessment of environmental quality. (3) Result, not a conclusion. (4) What is the basis for that statement? (6). I cant see any joint analysis of both communities…

Page 14, line 3-4 from the bottom of the main text. What is “acting” in this sentence? Cities or biodiversity?

Author Response

Please see the attachment. 

Author Response File: Author Response.pdf

Reviewer 2 Report

The use of the Shannon index as a comparative value to assess the quality of ecosystems between different sites presents a risk of confusion. To better speak of this index, it must be accompanied by the Piélou equitability index, or equidistribution index. Its formula corresponds to the ratio between H' (Shannon index) and Hmax (maximum Shannon index): this index therefore varies between 0 and 1. comparison between stations becomes easier and more logical.

Author Response

Please see the attachment. 

Author Response File: Author Response.pdf

Reviewer 3 Report

Review for the paper "Benthic Macroinvertebrates and Zooplankton Assemblages as Ecological Indicators in Urban Wetlands of Argentina" by María Julieta Arias, Pablo Vaschetto, Mercedes Marchese, Luciana Regaldo, Ana María Gagneten submitted to "Sustainability".

 

 

General comment.

 

Pelagic and benthic assemblages together represent the main base of any ecosystem. Their role in carbon fluxes and energy dynamics is obvious. In aquatic ecosystems, zooplankton and benthos have a great significance as predators and secondary producers. Bottom invertebrates and planktonic animals are considered to be good indicators of environmental forcing from climatic fluctuations and various human activities. The authors investigated the responses of benthic and zooplankton communities to different conditions in six urban wetlands located in Santa Fe (Argentina). They aimed to evaluate changes in community structure, functional feeding groups, and biotic indices of the benthos and zooplankton in aquatic ecosystems with various environmental states. They revealed that some aquatic organisms and biotic indices demonstrated a good potential to be indicators of environmental health in wetland ecosystems. The present study contributes to our knowledge on the assessment of the water quality in urban water bodies and may be interesting for environmentalists and ecologists dealing with monitoring of aquatic ecosystems. Standard methods to collect samples and to treat the data were used in the study. Main results are illustrated with relevant Figures and Tables. Statistical methods are carefully described and used. Discussion is focused on the main findings. The paper is well structured and sounds wise. I have some minor suggestions to improve the article.

 

Specific remarks.

 

Abstract. Santa Fe city should be replaced with Santa Fe City here and throughout the entire ms.

Fig. 1. The map must be updated with coordinates.

Section 2.2. Please, provide data on the total number of sampling for zooplankton and benthos separately.

Section 2.4, 3.2. Consider replacing "ind. m 2 " with "ind. m -2".

Table 1 must be updated with mean values.

Fig. 2. The resolution of the figure must be improved. Also, the font must be enlarged for better presentation.

Author Response

Please see the attachment. 

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

First of all, as I have already been exposed as having extensive lack of knowledge in the topic of the study, I can only say that I attempted to use the common sense and logic to evaluate the MS. And these are the effects:

1. I don’t buy arguments like “it is correct because the same thing was done in another paper…”
2. What actually is bioindication? I understand it like this: we want to know whether a given habitat is polluted or not. To do so directly, we would actually need to conduct a lot of chemical analyses, like heavy metals, pesticides, etc., as well as assess the level of human impact, eutrophication, and many others. As all the above are difficult, costly, problematic, etc., we attempt to find a proxy which will tell us the same in a simpler way. Here come the biotic indices and bioindicators. So, I will repeat my statement: there cannot be so that a given site is polluted according to one index and non-polluted to another one. It either is polluted or not, only some indices may detect it better or worse.
3. Moreover, if we want to check if our indices indicate pollution in a proper way, we need to have an independent assessment using abiotic analyses of pollution. Otherwise, we just have to believe that the indices work properly. But in such a situation, we cannot evaluate hypotheses stated in the introduction. We can only tell that some ponds differ from another in zooplankton/zoobenthos community. In other words: either we know which index is proper in a given context use it to determine ecosystem quality, or want to learn which index is proper, and then we have to relate biotic data with abiotic measurements of disturbance/pollution.
4. Whether the observed differences and gradients are due to disturbance or anything else, it is just a speculation. When I look at new Fig. 1, it is clear that the most similar sites are those situated close to one another, while the distant sites differ to a greater extent. This may be equally plausible explanation of difference as the disturbance gradient (even if it really exists). Moreover, now we can see that actually S1 & S2 and S3 & S4 are located in the same water bodies, so this challenges the issue of data independence and the statement that there were 6 ponds studied.
5. Despite your imputation, I realize that some of the taxa found in the study are known to be sensitive and others are tolerant to various types of pollution. But if it is all known, then what is novel in your study showing again that they are sensitive or tolerant?
6. When a study deals with the differences between communities inhabiting several local ponds, then this is a descriptive study and a statistical analysis doesn’t change that. To be more than that, it needs testable hypotheses and a general context, but, as I argue above, the hypotheses stated by the authors cannot be tested with the provided data.
7. In Fig. 5, you show a gradient of taxa from tolerant to sensitive to disturbance, established on the basis of your data. But earlier you determined which sites are more or less disturbed on the basis of indices based on exactly the same species!
8. I still insist that the CCA is not discussed at all in the Discussion, or at least I cannot find such a part. Anyway, what I see in Fig. 2 is different than the gradient postulated in the discussion (with RECU and PGB at two extremes of the gradient – in CCA, PGB in in the middle). Moreover, I still cannot see CCA results in the description which comes below the figure. I.e. the CCA is briefly described above the figure, and then a totally different picture of species typical for particular sites and dominating in them is shown, probably on the basis of the raw data inspection. But, in this case: what is the reason for the CCA at all?
9. I still don’t understand what was used for replicates in K-W analyses, particularly after your statement about using seasonal extremes… What was your number of samples in K-W? Were they independent as they should be? You say now different seasons were not compared, but they still were there and could affect the data and results.
10. To conclude: the MS is ok showing the differences in community composition among several ponds in South America and discussing them in the light of known sensitivity/tolerance of these communities to pollution. But it cannot be considered as assessing particular taxa or biotic indices as better or worse at determining the level of disturbance in these ponds, i.e. providing new information on these species.

Author Response

1. I don’t buy arguments like “it is correct because the same thing was done in another paper…”

This might be, however, it is always very important to count with scientific support from previous research works when carrying out an assessment.

 

1. What actually is bioindication? I understand it like this: we want to know whether a given habitat is polluted or not. To do so directly, we would actually need to conduct a lot of chemical analyses, like heavy metals, pesticides, etc., as well as assess the level of human impact, eutrophication, and many others. As all the above are difficult, costly, problematic, etc., we attempt to find a proxy that will tell us the same in a simpler way. Here come the biotic indices and bioindicators. So, I will repeat my statement: there cannot be so that a given site is polluted according to one index and non-polluted to another one. It either is polluted or not, only some indices may detect it better or worse.

Precisely, both biotic indices proposed for the benthic community classified all the sampling sites as polluted, but differed in the pollution degree. We must clarify again that we use the word “polluted” in reference to the categories established by the biotic indices.

 

1. Moreover, if we want to check if our indices indicate pollution in a proper way, we need to have an independent assessment using abiotic analyses of pollution. Otherwise, we just have to believe that the indices work properly. But in such a situation, we cannot evaluate the hypotheses stated in the introduction. We can only tell that some ponds differ from others in zooplankton/zoobenthos community. In other words: either we know which index is proper in a given context to use it to determine ecosystem quality, or want to learn which index is proper, and then we have to relate biotic data with abiotic measurements of disturbance/pollution.

We agree with you that to accurately talk about polluted sites we must include an extensive chemical assessment to reveal which pollutants are present and how they are affecting the different communities. However, this was not the aim of our work. We assessed the water quality of six urban ponds using bioindicators of both communities, based on different community attributes, functional feeding groups, and biotic indices to establish possible disturbances degrees.

The conceptual framework of this work considers bioindication as an early, practical, and economical warning: in the case of possible imbalances detected by indices such as those applied, government agents could justify carrying out batteries of more expensive studies-besides the ecological ones here carried out- that include analytical chemical determinations of water and sediment such as those proposed by the reviewer.

We insist that this was not the scope of our work. In this sense, environmental quality indicators can be divided into three broad categories: i) characteristic or indicator species, ii) univariate indices and iii) multimetric indices.

1. i) They provide information about a given environmental condition or stress. These organisms may be absent from impacted sites, be indicators of impacted sites, or exhibit some measurable response of a system under stress, such as behavioral and/or physiological changes.
2. ii) Commonly called biotic indices, they generally use the abundance and richness of species. The formulation of these indices can also be based on diversity (Margalef, Pielou, Shannon-Wiener indices) or on ecological groups or trophic groups.

iii) They were developed as integrative tools for the evaluation of the ecological condition of ecosystems. They include community descriptors as well as the percentage abundance of the different ecological and/or trophic groups.

Despite the complexity of such approaches due to the necessary knowledge of systematics, ecology and life history of organisms, in this work i), ii) and iii) were analyzed independently and comparatively for two communities (zooplankton and zoobenthos) that allowed to arrive in a well-founded way to the conclusions reached in this work.

 

1. Whether the observed differences and gradients are due to disturbance or anything else, it is just speculation. When I look at new Fig. 1, it is clear that the most similar sites are those situated close to one another, while the distant sites differ to a greater extent. This may be an equally plausible explanation of difference as the disturbance gradient (even if it really exists). Moreover, now we can see that actually S1 & S2 and S3 & S4 are located in the same water bodies, so this challenges the issue of data independence and the statement that there were 6 ponds studied.

It is correct that S1, S2, S3, and S4 present certain similarities and were in fact very close to each other in the CCA’s ordination. It would be possible –as the Ref. comment-to attribute this to the sampling site’s distance from one to another and because they are part of the same protected area, RNUO. However, when we assessed the benthic functional feeding groups, we found that S4 showed a much higher richness than the other sites in RNUO. This sampling site also presented a higher number of sensitive species that were not found in any of the other sampling sites. S1 reached a lower IMRP score than the other sites of RNUO (heavily polluted, along with PGB), but S4 conversely reached the highest IMRP score and was the only sampling site to reach the slightly polluted category. We also found significant differences in benthic species richness between S2 and S3.

 

1. Despite your imputation, I realize that some of the taxa found in the study are known to be sensitive and others are tolerant to various types of pollution. But if it is all known, then what is novel in your study showing again that they are sensitive or tolerant?

First of all, we did not affirm that “it is all known”, as the Ref. comment. Also, we did not intend to-and it is not part of the aim of our work- to determine the sensitivity or tolerance of certain species but to evaluate which ones are present or absent in the urban ponds we studied to establish, along with other analyses and community attributes, a disturbance degree of the sampling sites.

 

1. When a study deals with the differences between communities inhabiting several local ponds, then this is a descriptive study and a statistical analysis doesn’t change that. To be more than that, it needs testable hypotheses and a general context, but, as I argue above, the hypotheses stated by the authors cannot be tested with the provided data.

The three predictions stated at the beginning of this paper were tested and the results we obtained prove that: (1) It was in fact possible to establish differences between the urban ponds and also a disturbance gradient for benthic and zooplankton communities; (2) This was also possible to accomplish using both communities together and (3)* The simultaneous study of both benthic and zooplankton communities revealed more information on the urban ponds, because for some sampling sites benthic and zooplankton taxa were relevant for their ordination in CCA (e.g. S2 and S3), whereas for PGB those taxa were mainly benthic and for RECU, S1 and S4 they were zooplanktonic. While we performed a simple and small-scale assessment of local ponds, we believe we were able to accomplish the aim of our work and provided new and valuable information about these water bodies, which can be of use when assessing other wetlands in the area.

 

1. In Fig. 5, you show a gradient of taxa from tolerant to sensitive to disturbance, established on the basis of your data. But earlier you determined which sites are more or less disturbed on the basis of indices based on exactly the same species!

Precisely, Fig. 5 is a graphic that summarizes some of the results of our work, showing which species were present and were relevant to establish differences between the sampling sites.

 

1. I still insist that the CCA is not discussed at all in the Discussion, or at least I cannot find such a part. Anyway, what I see in Fig. 2 is different than the gradient postulated in the discussion (with RECU and PGB at two extremes of the gradient – in CCA, PGB in the middle). Moreover, I still cannot see CCA results in the description which comes below the figure. I.e. the CCA is briefly described above the figure, and then a totally different picture of species typical for particular sites and dominating in them is shown, probably on the basis of the raw data inspection. But, in this case: what is the reason for the CCA at all?

The gradient that is being described here is not the one with both benthic and zooplankton species, but the one with only zooplankton bioindicators, FFG, and biotic indices. Effectively, the gradient established through the assessment of both communities through the CCA is different from the ones proposed for each community separately. The CCA was important to accomplish the aim of our work because it revealed that benthic taxa were more relevant for the ordination of some sampling sites, for others it was zooplankton taxa and in some cases taxa from both communities together were important. We explained the same point about CCA above in *.

 

1. I still don’t understand what was used for replicates in K-W analyses, particularly after your statement about using seasonal extremes… What was your number of samples in K-W? Were they independent as they should be? You say now different seasons were not compared, but they still were there and could affect the data and results.

The replicates for K-W were 4 for each sampling site, which is why we decided to change the initial PERMANOVA as it was suggested. All replicates were independent. It is known that biotic and environmental variables can possibly show natural differences between seasons. It would be impossible to compare pond 1 in winter with pond 2 in summer. To avoid this, we sampled all RNUO sites in all four seasons and, since we were unable to repeat that with PGB and RECU, we sampled twice in summer and twice in winter to attempt to lessen the natural variation of the data and so be able to compare the urban ponds between each other.

 

1. To conclude: the MS is ok showing the differences in community composition among several ponds in South America and discussing them in the light of known sensitivity/tolerance of these communities to pollution. But it cannot be considered as assessing particular taxa or biotic indices as better or worse at determining the level of disturbance in these ponds, i.e. providing new information on these species.

The aim of our work was not to determine which biotic index was more suitable for the tested urban ponds if there is any. As it was established in the introduction, we intended to compare the urban ponds using different biotic indices and assess if they showed different results when one or another was used. BCI resulted in all sampling sites having regular water quality, which is plausible. However, other results from our research such as FFG richness and relative density, presence of sensitive/tolerant taxa, the K-W analysis, and B-C dissimilarity lead us to believe that perhaps IMRP results were more explanatory of the urban ponds disturbance degree.

 

Author Response File: Author Response.pdf