Flood Pulse Effects on the Growth of Pseudoplatystoma fasciatum in the Amazon Basin

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

General comment

The ms. describes a study addressing the influence of flood-Pulses on the growth of Pseudoplatystoma  fasciatum in the Amazon Basin. Growth data were obtained from six sampling stations and were related with hydrological indicators describing the duration and intensity of both, floods and droughts. The authors found an inverse relationship between growth (vertebrae increment width) and the hydrological indices, with slower growth during intense floods and intense droughts. 

The authors have used an appropriate approach (but see the specific comments) and the ms. is generally clearly written. Overall, the study is interesting and should be of interest to readers of Fishes.

Below I give some specific comments that could help the authors revising the ms.

Specific comments

Lines 207-208 “We used these same estimates of bankfull to our five study sites by applying those percentages to equivalent historical hydrological data from 1998 to 2022.” – the authors need to provide more details concerning the hydrological data (daily water levels) used in the study. Were the hydrological data gathered at gauging stations? Did a gauging station exist for each of the sampling locations?

Lines 285-286 “We found that our length-at-age estimates were not roughly but not fully congruent with those of prior studies” I believe the authors are trying to said that (…) estimates were roughly but not fully congruent (…).

Lines 191-195 “To model the growth of P. fasciatum, we fitted the Von Bertalanffy growth model (VBGF) to data from all specimens in the five sites (….)” As the VBGF was fitted to all sampling sites, it is not useful to compare growth between sampling locations and therefore it is not targeting the study objective as stated in lines 127-130.

Lines 337-363. The first two paragraphs of the discussion repeat ideas and could be much cleaner.

Lines 347-348 “Even though our results showing that the growth of P. fasciatum is slower in years of intense droughts are supported by those of other studies on other species, we find this result surprising.” How can the authors state surprise in the discussion, when, concerning drought intensity, the study introduction only refers the existence of a negative relation in other studies (lines 93-94)?

Lines 364-365 “Our result showing that the growth of P. fasciatum is slower in years of intense floods was expected, although it is not supported by some other studies [4,11,12].” Again, how can the authors state in the discussion that the growth of the studied species was expected to be slower with high floods, when, concerning flood duration, the study introduction only refers the existence of a positive relation in other studies (lines 78-84).

Comments on the Quality of English Language

Please refer to the comments concerning some english suggestions.  

Author Response

Lines 207-208 “We used these same estimates of bankfull to our five study sites by applying those percentages to equivalent historical hydrological data from 1998 to 2022.” – the authors need to provide more details concerning the hydrological data (daily water levels) used in the study. Were the hydrological data gathered at gauging stations? Did a gauging station exist for each of the sampling locations?

Answer:  We agree with your suggestion. To provide more detail concerning the hydrological data, we changed the text as follows: lines 224 to 227“Using the bankfull estimates, we calculated F and D. To calculate F, we first plotted the curve of daily water level specific to each sampling location per year and the bankfull curve. The daily water levels were obtained in the database of the Brazilian National Water Agency for each sampling site. Then, we summed daily water levels of the high-water season and subtracted the values of the bankfull. Similarly, we calculated D per year by summing the daily water levels during the low water season and subtracted the bankfull level (Figure 3).”

Lines 285-286 “We found that our length-at-age estimates were not roughly but not fully congruent with those of prior studies” I believe the authors are trying to said that (…) estimates were roughly but not fully congruent (…).

Answer:  Yes, this is a typing mistake. Thank you for noticing.

Lines 191-195 “To model the growth of P. fasciatum, we fitted the Von Bertalanffy growth model (VBGF) to data from all specimens in the five sites (….)” As the VBGF was fitted to all sampling sites, it is not useful to compare growth between sampling locations and therefore it is not targeting the study objective as stated in lines 127-130.

Answer: The objective described by the reviewer states that “Here, we evaluated how interannual variability in flood pulses influences the growth of P. fasciatum. We addressed two main research questions: Do floods lead to faster, and droughts lead to slower growth of P. fasciatum? Do floods and droughts affect all age-classes in the same manner? ”. Therefore, we did not have an objective compare the growth between sampling locations.

 Our intention in plotting the VBGF was to assess whether our growth data was in general agreement with that described in the literature. The data are plotted in aggregate form for all sampling sites because sampling sizes did not allow for site-specific growth curves to be plotted.

Lines 337-363. The first two paragraphs of the discussion repeat ideas and could be much cleaner.

Answer:  We revised the text for clarity and concision.

Lines 347-348 “Even though our results showing that the growth of P. fasciatum is slower in years of intense droughts are supported by those of other studies on other species, we find this result surprising.” How can the authors state surprise in the discussion, when, concerning drought intensity, the study introduction only refers the existence of a negative relation in other studies (lines 93-94)?

Answer:  We agree with the comment. We mentioned before that this result is surprising because food availability during intense droughts should increase for a piscivorous species as the density of prey increases with the decrease in water levels. We changed the text to improve the clarity as follows: Lines 370 to 371: “Our results showing that the growth of P. fasciatum is slower in years of intense droughts are supported by those of other studies on other species.” We believe this adresses the issue at hand.

Lines 364-365 “Our result showing that the growth of P. fasciatum is slower in years of intense floods was expected, although it is not supported by some other studies [4,11,12].” Again, how can the authors state in the discussion that the growth of the studied species was expected to be slower with high floods, when, concerning flood duration, the study introduction only refers the existence of a positive relation in other studies (lines 78-84).

Answer:  We agree with the suggestion and changed the text as follows: Line 387 to 388: “Our result showing that the growth of P. fasciatum is slower in years of intense floods was not supported by some other studies [4,11,12]” We believe this addresses the issue at hand.

Reviewer 2 Report

Comments and Suggestions for Authors

I liked this paper very much. I think their research on Fish growth within Amazon Basin rivers as it relates to drought and flood conditions is timely. I also think the use of aging is a particularly powerful technique for this type of analysis. However, I do see some problems that need to be addressed before acceptance. There is quite a bit of analysis in this paper. Which is great. However, some of it is not addressed or alluded to in the discussion. What is the point of having the results reported if they are not discussed? Please see my comments below:

General and Specific comments: 

Line 193: you are missing a closed in this equation--𝐿𝑡 = 𝐿∞ ∗ [1 − 𝑒(−𝑘(𝑡 − 𝑡𝑜))]. Also, the Lt should be used subscript as should the L infinity. I assume this be corrected by a copy editor. Also, there should be called equation 3 and equation 3 should be called equation 4.

Line 219: Weisberg et al. (2010) should also be listed as reference 40.

Line 230: equation 3. Same issue with subscripts. Rename equation 3 as equation 4.

Figure 4: What do you mean by frequency? Frequency is the number of times an event occurs. I assume that this is proportion and not frequency. Also, you might want to consider having the same total range of length classes (cm) on your x-axis for each facet. It will show the differences in size ranges between the populations.

Figure 5: What do the orange and blue dots represent? What is this figure about? You need to indicate this in the Figure Legend/text.

I find it interesting that in line 281 you suggest the population from Santarem is growing more slowly than the other populations and you imply that this is due to overexploitation. Often when populations are overexploited intraspecific competition is reduced and growth rates increase. You might want to expand on this in the discussion.

One 288: not sure what you’re getting at here when this phrase—“estimates were not roughly but not fully congruent”. Please clarify.

One 288: space needs to be added after the] for reference 34 in the word and.

Figure 6: This is inappropriate as PV has a disproportionate influence on this relationship because it represents an order of magnitude greater number of fish. If you wanted to compare the populations, you should have calculated the VonB for each population and compared them. Alternatively, you could look at the residuals for the other populations compared to the PV. I would do the former and not the latter. Also, you did not talk about Von B results in the discussion. So if you want to include this result you need to talk about it in the discussion. Looking at your figure 6 it certainly seems that the Santa population has a much lower growth rate compared to PV and the others. You might want to discuss this more fully.

Figure 7: you might want to include a better description of how F and D are calculated in the methods section or at least in the figure legend.

Figure 8: what is the use of figure 8? You might want to explain this more fully in the results and discussion. If not, you can always drop it from your paper.

Line 338: “The inverse relationships that we found between the growth of P. fasciatum and floods and droughts are both supported and not supported by the literature.” –Awkward--Not a great way to start a discussion. This indicates that there is a controversy associated with this concept as it relates to growth rates during floods and droughts.

Line 380: there should be a space between [22] and showing.

Line 401: you indicated you had limited statistical power for this analysis to truly test the inverse relationship between intensity floods and growth. If you believe that to be case, it would be a good idea to run a power analysis. This would be helpful for your paper and for people doing research in this area.

Line 404: yes, there is a problem with the lack of balance in your data set from the five sites. This is clearly shown in the von Bertalanffy growth curves.

Line 421: you indicate that the variation in the width of growth increments may have affected your results and you attributed this to ecological reasons reflecting the different River environments. What are these ecological reasons? This is where you should discuss the reasons.

Author Response

Line 193: you are missing a closed in this equation--?? = ?∞ ∗ [1 − ?(−?(? − ??))]. Also, the Lt should be used subscript as should the L infinity. I assume this be corrected by a copy editor. Also, there should be called equation 3 and equation 3 should be called equation 4.

Answer:  We changed the equation and revised its labels.

Line 219: Weisberg et al. (2010) should also be listed as reference 40.

Answer:  We added the study to the references and reviewed the text for any possible citation mistakes.

Line 230: equation 3. Same issue with subscripts. Rename equation 3 as equation 4.

Answer:  Fixed.

Figure 4: What do you mean by frequency? Frequency is the number of times an event occurs. I assume that this is proportion and not frequency. Also, you might want to consider having the same total range of length classes (cm) on your x-axis for each facet. It will show the differences in size ranges between the populations.

Answer:  We modified all the graphics according to the suggestions.

Figure 5: What do the orange and blue dots represent? What is this figure about? You need to indicate this in the Figure Legend/text.

Answer:  We changed Figure 5’s label to “Calculated mean radius of growth marks (in blue) and mean increments (in orange) of P. fasciatum at five sites in the Amazon basin.”

I find it interesting that in line 281 you suggest the population from Santarem is growing more slowly than the other populations and you imply that this is due to overexploitation. Often when populations are overexploited intraspecific competition is reduced and growth rates increase. You might want to expand on this in the discussion.

Answer:  I chose not to discuss these results in the discussion session because inter-sites comparaison of growth was not the focuss of this study (and sample size per site did not allow it either). We plotted the VBGF to ensure that our data matched the literature's descriptions.

One 288: not sure what you’re getting at here when this phrase—“estimates were not roughly but not fully congruent”. Please clarify.

Answer:  It was a typing error. We changed the text to: “We found that our length-at-age estimates were roughly but not fully congruent with those of prior studies (Table 3).”

One 288: space needs to be added after the] for reference 34 in the word and.

Answer:  We revised the sentence as suggested by the reviewer.

Figure 6: This is inappropriate as PV has a disproportionate influence on this relationship because it represents an order of magnitude greater number of fish. If you wanted to compare the populations, you should have calculated the VonB for each population and compared them. Alternatively, you could look at the residuals for the other populations compared to the PV. I would do the former and not the latter. Also, you did not talk about Von B results in the discussion. So if you want to include this result you need to talk about it in the discussion. Looking at your figure 6 it certainly seems that the Santa population has a much lower growth rate compared to PV and the others. You might want to discuss this more fully.

Answer: We plotted the VBGF to ensure that our data matched the literature's descriptions. We tried to fit one model for each basin. However, our sampling size did not allow such an analysis.

Figure 7: you might want to include a better description of how F and D are calculated in the methods section or at least in the figure legend.

Answer:  We rewrote the text for clarity as follows: “Using the bankfull estimates, we calculated F and D. To calculate F, we first plotted the curve of daily water level specific to each sampling location per year and the bankfull curve. The daily water levels were obtained in the database of the Brazilian National Water Agency for each sampling site. Then, we summed daily water levels of the high-water season and subtracted the values of the bankfull. Similarly, we calculated D per year by summing the daily water levels during the low water season and subtracted the bankfull level (Figure 3).”

Figure 8: what is the use of figure 8? You might want to explain this more fully in the results and discussion. If not, you can always drop it from your paper.

Answer:  We removed this result from the ms.

Line 338: “The inverse relationships that we found between the growth of P. fasciatum and floods and droughts are both supported and not supported by the literature.” –Awkward--Not a great way to start a discussion. This indicates that there is a controversy associated with this concept as it relates to growth rates during floods and droughts.

Answer:  We changed the text to read: “The inverse relationships that we found between the growth of P. fasciatum and floods and droughts intensity are consistent with some but not all studies.”

Line 380: there should be a space between [22] and showing.

Answer:  We added the space.

Line 401: you indicated you had limited statistical power for this analysis to truly test the inverse relationship between intensity floods and growth. If you believe that to be case, it would be a good idea to run a power analysis. This would be helpful for your paper and for people doing research in this area.

Answer:  We believe that the limitation regarding spatial sample heterogeneity, size-length variability among rivers, and variations in growth increments among basins should not affect the analysis of our assessment of whether floods and droughts affect growth rates because, to address this question, we used a linear mixed modeling approach that leads to improved estimates for groups (e.g., sites) with small sample sizes through partial pooling across sites.

Line 404: yes, there is a problem with the lack of balance in your data set from the five sites. This is clearly shown in the von Bertalanffy growth curves.

Answer:  Yes, indeed.

Line 421: you indicate that the variation in the width of growth increments may have affected your results and you attributed this to ecological reasons reflecting the different River environments. What are these ecological reasons? This is where you should discuss the reasons.

Answer:  Such differences include time magnitude and intensity of flood pulses, different anthropogenic pressures, such as deforestation, pollution, fishing pressure.

We changed the text to read: “The observed variability in growth increments could be due to ecological reasons reflecting the different river environments including water quality and productivity, time magnitude and intensity of flood pulses, different anthropogenic pressures, such as deforestation, pollution, fishing pressure.” But we unfortunately don’t have precise data to dwell further into these potential differences

Reviewer 3 Report

Comments and Suggestions for Authors

Review of Flood-Pulse Effects on the Growth of Pseudoplatystoma fasciatum in the Amazon Basin by Pereira et al.

General Comments

 

This paper addresses important and interesting questions but suffers from some aspects of presentation and apparent interpretation of results as described in the specific comments that follow.

Specific Comments

·     Abstract, line 19. Spell out genus name at first use.

·     Keywords, line 36. I would add Pseudoplatystoma fasciatum and Amazon River basin to the list of keywords

·     Introduction, lines 119, 121, and 122. “Rising-water season” is typically referred to as the ascending limb of the hydrograph and “falling-water season” is referred to as the descending limb. “Low-water season” is typically referred to as the base-flow period.

·     Materials and Methods, line 147. Provide sample sizes for each age-class in each river.

·     Data Analysis, line 206. Why are there two numbers presented for this estimate? Do they represent the range of bankfull estimates? Which value was used? How similar is the Solimões River to the rivers studied here? An alternative approach might be to use return frequency of annual peak flows to determine bankfull flows, which for many rivers is 1.5 years. If you have a decent flow data set for each river (which it appears you do), you could calculate this easily.

·     Data Analysis, line 214. “High-water season” is typically referred to as the peak-flow period.

·     Data Analysis, Figure 3. The description of the graph in the caption does not seem accurate. Why are the stippled and hatched areas described as flood and drought and the areas under the dashed lines described as intense and mild? Those areas are also within the range of flood and drought. Don’t the stippled and hatched areas represent the mean flood and drought?

·     Results, line 262. Here and elsewhere “P. fasciatum” should be italicized.

·     Results, Figure 5. This figure is a bit confusing. The caption indicates the graph shows the calculated mean radius of growth marks, but it also shows mean increment values. The caption should indicate which set of points represents mean radius and which indicates mean increment.

·     Results, line 279. Insert “pooled” between “sites” and “because.”

·     Results, line 286. Should the phrase “were not roughly but not fully congruent” read “were roughly but not fully congruent?”

·     Results, line 288. Where were the referenced studies conducted? Please specify.

·     Results, Figure 7. Provide more appropriate x-axis and y-axis labels. The x-axes should be labeled “Flood Index” and “Drought Index”, and the y-axis should be labeled “log Increment Width.” Why does the drought index have negative values?

·     Results, Figure 8. Change x-axis label to “Age (years).”

·     Results, Figure 9, Revise the size of each graph to make them equal height. Several of the drought graphs are considerably taller than the others.

·     Discussion, line 332. The statement that “P. fasciatum had higher growth rates in years of “mild” droughts in all age-classes except age-class 6.” doesn’t seem supported by Figure 9. The difference is not significant, but the mean is higher. The pattern suggests that all age-classes grew faster in mild drought years. The lack of statistical significance in this age class could be due to a smaller sample size, similar to your argument in the Discussion regarding the effect of floods.

·     Results, Figure 9. Drop “ANOVA” from figure caption. In the caption, provide a description of the box and whisker plots indicating what each element (whiskers, boxes, internal line of each box) means.

·     Discussion, line 339. Change “floods and droughts” to “flood and drought intensity.” Change “both supported and not supported by the literature” to “are consistent with some but not all studies.”

·     Discussion, line 351. Change “decreased with droughts” to “decreased with drought intensity.”

·     Discussion, line 364-365. Why was this result expected if previous studies did not show this effect?

·     Discussion, line 398-399. The statement that the sample size for each age-class was no more than 1/6 of the total sample size means the sample sizes for each age-sex class were equal to 1/6 of the total sample size. Certainly, that is not the case. Sample sizes for older age classes are most likely going to be smaller than for younger age-classes as is evident from the data plotted in Figure 8.

·     Discussion, line 432-438. The concern expressed here regarding issues related to hydroelectric dams blocking access to important upstream and downstream habitat could be offset to some extent with the placement of fish passage structures.

·     Discussion, line 439-441. It seems more could be said about the effects of climate change on growth rates. Would river regulation reduce the effects of extreme conditions under climate change? Would higher water temperatures change the observed relationships to flood and drought intensity? Would higher temperatures make mild drought conditions less beneficial?

·     Discussion, line 442. This sentence doesn’t make sense to me: “Because the growth of P. fasciatum is faster when hydrological conditions vary little, the biomass available for harvesting P. fasciatum in the system should vary relatively little interannually.” This would only be the case if hydrological conditions varied little. If intense floods and droughts become more frequent as is expected under climate change, there could be great interannual variability in harvest biomass.

Comments for author File: Comments.pdf

Author Response

Abstract, line 19. Spell out genus name at first use.

Answer:  The text was revised as suggested.

 

Keywords, line 36. I would add Pseudoplatystoma fasciatum and Amazon River basin to the list of keywords

Answer:  We added the following keywords: “Interannual hydrological variation, growth, catfish, Neotropics, Pseudoplatystoma fasciatum, Amazon Basin”

Introduction, lines 119, 121, and 122. “Rising-water season” is typically referred to as the ascending limb of the hydrograph and “falling-water season” is referred to as the descending limb. “Low-water season” is typically referred to as the base-flow period.

Answer:  We changed the text in the introduction to improve clarity as follows: Lines:116 to 128: “P. fasciatum grows to more than one meter in length [32,34]. Its growth rates are higher during the rising-water season (typically referred to as the ascending limb of the hydrograph) and lower during the low-water season (typically referred to as the base-flow period), like most fishes in river floodplains [35]. Female P. fasciatum become sexually mature later and reach larger sizes than males [34]. P. fasciatum performs feeding and reproductive migrations accompanying seasonal hydrological variations, migrating upstream to spawn during the rising-water season [34]. Eggs and larvae of P. fasciatum drift downstream with the current during the high-water season when adults and juveniles migrate laterally to the floodplains to prey on other fish and grow faster. During the falling-water season (typically referred to as the descending limb), P. fasciatum migrates back to the mainstream river and stays there for the low-water season, when it predates less and grows more slowly [31]”. We believe this adresses the issue at hand.

Materials and Methods, line 147. Provide sample sizes for each age-class in each river.

Answer:  We added Table 1 at line 164, showing the sample size per age class for the five sites in the Amazon.

Data Analysis, line 206. Why are there two numbers presented for this estimate? Do they represent the range of bankfull estimates? Which value was used? How similar is the Solimões River to the rivers studied here? An alternative approach might be to use return frequency of annual peak flows to determine bankfull flows, which for many rivers is 1.5 years. If you have a decent flow data set for each river (which it appears you do), you could calculate this easily.

Answer: The ideal approach to estimating the bankfull level would be to measure water levels over at least a hydrological year and specifically measure when certain vegetation types get inundated in the floodplain for each site. Because we could not take these measurements for this study, we followed the methodology of Castello et al. (2015), who used estimates to calculate the bankfull based on fieldwork observations in Tefe, in the Solimoes River. Castello et al. 2015, estimated that 42%  and 45 % of the water levels were the baseline for calculating the bankfull for the low and high water seasons.. Thus, we used these percentages to estimate the bankfull and calculate the hydrological indices of droughts and floods.

Our sampling sites share similarities to the Solimoes River and habitat idiosyncrasies. Although this approach is not ideal, it was the best we could do to estimate the bankfull as accurately as possible. We argue that even if there is variabtion in bankful, it should not affect results much because what matters is interannual variability in the indexes, not their exact magnitude.

Data Analysis, line 214. “High-water season” is typically referred to as the peak-flow period.

Answer:  We added the definition in the introduction in line 123.

Data Analysis, Figure 3. The description of the graph in the caption does not seem accurate. Why are the stippled and hatched areas described as flood and drought and the areas under the dashed lines described as intense and mild? Those areas are also within the range of flood and drought. Don’t the stippled and hatched areas represent the mean flood and drought?

Answer: No. The stippled area represents intense and mild floods, and the hatched area represents intense and mild droughts. We revised the figure for clarity.

Results, line 262. Here and elsewhere “P. fasciatum” should be italicized.

Answer:  We revised the whole text to address that issue.

Results, Figure 5. This figure is a bit confusing. The caption indicates the graph shows the calculated mean radius of growth marks, but it also shows mean increment values. The caption should indicate which set of points represents mean radius and which indicates mean increment.

Answer:  We changed the caption for: “Figure 1. Calculated mean radius of growth marks (in blue) and mean increments (in orange) of P. fasciatum at five sites in the Amazon basin.”

Results, line 279. Insert “pooled” between “sites” and “because.”

Answer:  We applied the suggested revision, which improved the text.

Results, line 286. Should the phrase “were not roughly but not fully congruent” read “were roughly but not fully congruent?”

Answer:  Yes, it was a typing mistake. Thank you!

Results, line 288. Where were the referenced studies conducted? Please specify.

Answer:  We detailed the study areas in lines 303 to 305: “For age class one, our estimated length at age was longer than those by Loubens and Panfilli [34] who studied in the Mamore Basin in the Bolivian Amazon, and Armas et al. [32] in the Peruvian Amazon.”

Results, Figure 7. Provide more appropriate x-axis and y-axis labels. The x-axes should be labeled “Flood Index” and “Drought Index”, and the y-axis should be labeled “log Increment Width.” Why does the drought index have negative values?

Answer: Some drought indices have negative values because I scaled the variables before running the model in R following the mwthods of the cited paper, Zuur (2010). Using the scale function in R, I divided the center number of each log of the increment width by its standard variations.

Results, Figure 8. Change x-axis label to “Age (years).”

Answer:  Figure 8 was not necessary, therefore, removed.

Results, Figure 9, Revise the size of each graph to make them equal height. Several of the drought graphs are considerably taller than the others.

Answer:  We made the graphics proportionally and symmetricly distributed.

Discussion, line 332. The statement that “P. fasciatum had higher growth rates in years of “mild” droughts in all age-classes except age-class 6.” doesn’t seem supported by Figure 9. The difference is not significant, but the mean is higher. The pattern suggests that all age-classes grew faster in mild drought years. The lack of statistical significance in this age class could be due to a smaller sample size, similar to your argument in the Discussion regarding the effect of floods.

Answer:  We changed the text for “On the other hand, for droughts, P. fasciatum had higher growth rates in years of “mild” droughts in all age-classes. The lack of statistical significance for age-class 6 could be due to the small sample size in this age-class. (Table 1; Fig. 9) in line 352 to 355. We believe this adresses the issue at hand.

Results, Figure 9. Drop “ANOVA” from figure caption. In the caption, provide a description of the box and whisker plots indicating what each element (whiskers, boxes, internal line of each box) means.

Answer:  We changed the text in Figure 9 legend to read: “Box and whisker plots showing the growth (increment width) of six age-classes of P. fasciatum for "intense" and "mild" years of floods and droughts in the Amazon with their respective p-values. The box plots indicate the lower and upper quartiles, and the means (line in the boxplots).”

Discussion, line 339. Change “floods and droughts” to “flood and drought intensity.” Change “both supported and not supported by the literature” to “are consistent with some but not all studies.”

Answer:  We rewrote the sentence as: “The inverse relationships that we found between the growth of P. fasciatum and floods and drought intensity are consistent with some but not all studies” at lines 361 to 362.

Discussion, line 351. Change “decreased with droughts” to “decreased with drought intensity.”

Answer:  We changed the text as suggested to line 371 to 374: “Bayley [4] found that growth rates of Acarichthys heckli, Colossoma macropomum, Triportheus albus, T. aungulatus, T. elongatus, Brycon melanopterus and Prochilodus nigricans, decreased with droughts intensity.”

Discussion, line 364-365. Why was this result expected if previous studies did not show this effect?

Answer:  We expected that the growth of P. fasciatum would be slower in years of intense floods because P. fasciatum is a piscivorous species. We explained in lines 401 to 402 that:” … prey density, which can decrease with the dilution effect during increasing flood intensity, thereby decreasing growth rates”

The reason why other studies did not support this relationship is that all of them investigated the ecology of omnivorous and herbivorous species.Lines 388 to 395: “Previous studies show that fish grow faster in years of intense floods; this was the case for T. andersoni, T. macrochir, T. rendalii, Acarichthys heckli, Colossoma macropomum, Triportheus albus, T. aungulatus, T. elongatus, Brycon melanopterus, and Prochilodus nigricans [4]. One possible explanation for this lack of support for our result is that these fishes are herbivorous and omnivorous species, which do find greater food availability during the high-water season when new terrestrial areas are inundated, favoring primary production [4]. However, P. fasciatum is a piscivorous species whose growth appears to be affected differently by seasonal floods.”

Biotic and abiotic factors could  be the mechanisms that explain this relationship. We explained these factors in lines 400 to 402: “One abiotic factor that could explain why the growth of P. fasciatum is slower in years of intense floods is anoxia. Junk et al. [23] showed that after the peak of the flood, dissolved oxygen levels are minimal, creating anoxic conditions that could possibly lead to lower growth rates in P. fasciatum. One biotic factor that likely influences the growth of P. fasciatum is prey density, which can decrease with the dilution effect during increasing flood intensity, thereby decreasing growth rates”

Hence, we changed the text for clarity as follows: “Our result showing that the growth of P. fasciatum is slower in years of intense floods was not supported by some other studies [4,11,12]. Previous studies show that fish grow faster in years of intense floods; this was the case for T. andersoni, T. macrochir, T. rendalii, Acarichthys heckli, Colossoma macropomum, Triportheus albus, T. aungulatus, T. elongatus, Brycon melanopterus, and Prochilodus nigricans [4]. One possible explanation for this lack of support for our result is that these fishes are herbivorous and omnivorous species, which do find greater food availability during the high-water season when new terrestrial areas are inundated, favoring primary production [4]. However, P. fasciatum is a piscivorous species whose growth appears to be affected differently by seasonal floods. Both biotic and abiotic factors could explain a decrease in growth rates as a function of flood intensity. One abiotic factor that could explain why the growth of P. fasciatum is slower in years of intense floods is anoxia. Junk et al. [23] showed that after the peak of the flood, dissolved oxygen levels are minimal, creating anoxic conditions that could possibly lead to lower growth rates in P. fasciatum. One biotic factor that likely influences the growth of P. fasciatum is prey density, which can decrease with the dilution effect during increasing flood intensity, thereby decreasing growth rates. Supporting this explanation are results from the study by Junk [22]showing that the lowest fat content occurs at the peak of the flood season for three predatory species in the Amazon, likely reflecting the ecological mechanisms that, in our study, may have led to slower growth for P. fasciatum during intense floods. Junk [22] inferred that seasonality in fat content in fishes occurs because fish need to accumulate energy to perform spawning migrations. P fasciatum performs spawning migration at the beginning of the rising-water season, so they probably had consumed all fat content after spawning in the high-water season.

Discussion, line 398-399. The statement that the sample size for each age-class was no more than 1/6 of the total sample size means the sample sizes for each age-sex class were equal to 1/6 of the total sample size. Certainly, that is not the case. Sample sizes for older age classes are most likely going to be smaller than for younger age-classes as is evident from the data plotted in Figure 8.

Answer:  We changed the text for clarity at Lines 421 to 425: “The sample size of each age-class varied with a few specimens younger than 2 years and older than 4 years old, which limited the ability of the ANOVA to reject the null hypothesis of there being no differences in growth between mild and intense floods. This limited statistical power was exacerbated by the effect of floods on growth being one-fifth that of droughts.”

Discussion, line 432-438. The concern expressed here regarding issues related to hydroelectric dams blocking access to important upstream and downstream habitat could be offset to some extent with the placement of fish passage structures.

Answer:  In the Amazon Basin, fish passages do not work for most of the migratory species because those passages are not designed according to the ecology of Amazonian species but rather for temperate species (Birnie- Gauvin et al 2019; Pelicice et al 2017; Pelicice ae at 2008, Pelicice et al 2015; Pompeu et al 2012; Silva et al 2018).

Discussion, line 439-441. It seems more could be said about the effects of climate change on growth rates. Would river regulation reduce the effects of extreme conditions under climate change? Would higher water temperatures change the observed relationships to flood and drought intensity? Would higher temperatures make mild drought conditions less beneficial?

Answer: We appreciate the comment above but we believe that could be a theme for a new paper and the deep discussion of climate change its impacts is not the main objective of our study.The effect that climate change has on hydrology is the increase in extreme events, such as floods and droughts, as seen in recent years. River regulation caused mostly by dams in the Amazon has several impacts on river ecosystems including 1) alteration of flooding; 2) change in nutrient flow; 3) change in the water temperature; 4) disruption of fish migration; 5) change in fish diversity composition; and 6) change in water chemistry.Therefore we argue that the synergetic effect of dams and climate change might be detrimental to P. fasciatum ecology, including migration, growth, survival, and reproduction.

Discussion, line 442. This sentence doesn’t make sense to me: “Because the growth of P. fasciatum is faster when hydrological conditions vary little, the biomass available for harvesting P. fasciatum in the system should vary relatively little interannually.” This would only be the case if hydrological conditions varied little. If intense floods and droughts become more frequent as is expected under climate change, there could be great interannual variability in harvest biomass.

Answer:  We removed the implication to manage and conservation regarding fisheries from the discussion.

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

General comments: Nice paper. Glad you addressed the changes requested. 

Specific comments:

Line 183: I might be wrong but shouldn't Weisberg et al (2010) be also reported as [40] so the reader knows where to look in the references?

Please change Frequency on the dependent axis on Figure 4 to Relative Frequency or keep it as proportion. Frequency is a count of how many times things happen. It does not vary between 0 and 1. That is a relative proportion.

Please check your citation numbering. It seems to be off again. 

Author Response

Responses to Reviewer 2

 

General comments: Nice paper. Glad you addressed the changes requested.

RESPONSE: We are pleased with the positive comment.

 

Specific comments:

 

Line 183: I might be wrong but shouldn't Weisberg et al (2010) be also reported as [40] so the reader knows where to look in the references?

RESPONSE: We are grateful for the reviewer’s sharp eye, and have fixed the errant citation stylistic, now at line 179.

 

Please change Frequency on the dependent axis on Figure 4 to Relative Frequency or keep it as proportion. Frequency is a count of how many times things happen. It does not vary between 0 and 1. That is a relative proportion.

RESPONSE: The reviewer is correct, and the vertical axis has been changed to read “Relative Frequency”.

 

Please check your citation numbering. It seems to be off again.

RESPONSE: We checked all reference numbers.

Reviewer 3 Report

Comments and Suggestions for Authors

General Comments

            The authors made significant changes to the manuscript and addressed most of my comments. Original comments that I feel were not adequately addressed are shown below.

Comments not Adequately Addressed

·         Original comment: Data Analysis, line 206. Why are there two numbers presented for this estimate? Do they represent the range of bankfull estimates? Which value was used? How similar is the Solimões River to the rivers studied here? An alternative approach might be to use return frequency of annual peak flows to determine bankfull flows, which for many rivers is 1.5 years. If you have a decent flow data set for each river (which it appears you do), you could calculate this easily.

My response: This comment has not been addressed adequately. It is not clear why there are two percentage values used to estimate bankfull flow. Was bankfull flow estimated twice, and are these the two percentage values determined?

·         Original comment: Data Analysis, Figure 3. The description of the graph in the caption does not seem accurate. Why are the stippled and hatched areas described as flood and drought and the areas under the dashed lines described as intense and mild? Those areas are also within the range of flood and drought. Don’t the stippled and hatched areas represent the mean flood and drought?

My response: The revised graph is not any clearer than the original. Dark shading is supposed to represent drought, but is only used for mild drought. Stippling is used for both intense flooding and intense drought. Intense what? I presume the curve to the left is for floods and the curve to the right for drought. Assuming that is the case, I would get rid of the shading, label the left curve “Flood” and the right curve “Drought” along the bankfull line and the three areas within each curve as “Mild.” “Moderate,” and “Intense.” The caption should be revised accordingly.

·         Original comment: Results, line 262. Here and elsewhere “P. fasciatum” should be italicized.

My response: For the most part this comment has been addressed, but one block of new text (line 340) has it in plain text.

·         Original comment: Results, Figure 7. Provide more appropriate x-axis and y-axis labels. The x-axes should be labeled “Flood Index” and “Drought Index”, and the y-axis should be labeled “log Increment Width.” Why does the drought index have negative values?

My response: Comments on axis labels addressed. The response to my comment on the negative drought index values indicates that these values result from an error. I think those values should be recalculated.

·         Original comment: Discussion, line 339. Change “floods and droughts” to “flood and drought intensity.” Change “both supported and not supported by the literature” to “are consistent with some but not all studies.”

My response: Comment addressed, but flood and drought should be singular not plural.

·         Original comment: Discussion, line 351. Change “decreased with droughts” to “decreased with drought intensity.”

My response: Comment addressed, but drought should be singular not plural.

·         Original comment: Discussion, line 439-441. It seems more could be said about the effects of climate change on growth rates. Would river regulation reduce the effects of extreme conditions under climate change? Would higher water temperatures change the observed relationships to flood and drought intensity? Would higher temperatures make mild drought conditions less beneficial?

My response: More detailed discussion on this issue should be included. It does not have to be exhaustive, but should briefly describe the relevance of the results under a changing climate.

Comments for author File: Comments.pdf

Author Response

Comments Adequately Addressed

Original comment: Data Analysis, line 206. Why are there two numbers presented for this estimate? Do they represent the range of bankfull estimates? Which value was used? How similar is the Solimões River to the rivers studied here? An alternative approach might be to use return frequency of annual peak flows to determine bankfull flows, which for many rivers is 1.5 years. If you have a decent flow data set for each river (which it appears you do), you could calculate this easily.

Reviewer’s response: This comment has not been addressed adequately. It is not clear why there are two percentage values used to estimate bankfull flow. Was bankfull flow estimated twice, and are these the two percentage values determined?

Authors’ answer: The two percentages were estimated by Castello et al. (2015), who used two estimates for bankfull based on fieldwork observations in Tefe on the Solimoes River. That is, Castello et al. (2015) estimated that 42% of water levels’ inundation for the low-water season and 45% of the water levels for the high-water season as the baselines for calculating the bankfull for the low- and high-water seasons. Thus, we used these percentages to estimate the bankfull and calculate the hydrological indices of droughts and floods.

We changed the text now at lines 218-224 to read: “Castello et al. [29] estimated that low, swampy chavascal woodland vegetation in várzea floodplain ecosystems similar to those in this study become flooded each year at 42% and 45%, respectively, of water-level differences between minimum and maximum water levels during the years the studies were conducted. We used these same estimates of bankfull levels to our five study sites by applying those percentages to equivalent historical hydrological data from 1998 to 2022.”

 

Original comment: Data Analysis, Figure 3. The description of the graph in the caption does not seem accurate. Why are the stippled and hatched areas described as flood and drought and the areas under the dashed lines described as intense and mild? Those areas are also within the range of flood and drought. Don’t the stippled and hatched areas represent the mean flood and drought?

Reviewer’s response: The revised graph is not any clearer than the original. Dark shading is supposed to represent drought, but is only used for mild drought. Stippling is used for both intense flooding and intense drought. Intense what? I presume the curve to the left is for floods and the curve to the right for drought. Assuming that is the case, I would get rid of the shading, label the left curve “Flood” and the right curve “Drought” along the bankfull line and the three areas within each curve as “Mild.” “Moderate,” and “Intense.” The caption should be revised accordingly.

Authors’ answer:  We have changed Figure 3 according to the reviewer’s recommendations.

 

Original comment: Results, line 262. Here and elsewhere “P. fasciatum” should be italicized.

Reviewer’s response: For the most part this comment has been addressed, but one block of new text (line 340) has it in plain text.

Authors’ answer: We have italicized species names in all contexts.

 

Original comment: Results, Figure 7. Provide more appropriate x-axis and y-axis labels. The x-axes should be labeled “Flood Index” and “Drought Index”, and the y-axis should be labeled “log Increment Width.” Why does the drought index have negative values?

Reviewer’s response: Comments on axis labels addressed. The response to my comment on the negative drought index values indicates that these values result from an error. I think those values should be recalculated.

Authors’ answer: As explained before, the negative values for the drought indices are not in error. Some drought indices have negative values because we scaled the variables before running the model in R following the methods of the cited paper, Zuur (2010). Using the scale function in R, we divided the center number of each log of the increment width by its standard variation, which resulted in negative values.

 

Original comment: Discussion, line 339. Change “floods and droughts” to “flood and drought intensity.” Change “both supported and not supported by the literature” to “are consistent with some but not all studies.”

Reviewer’s response: Comment addressed, but flood and drought should be singular not plural.

Authors’ answer: We changed the text to the singular.

 

Original comment: Discussion, line 351. Change “decreased with droughts” to “decreased with drought intensity.”

Reviewer’s response: Comment addressed, but drought should be singular not plural.

Authors’ answer: We changed the text to the singular.

 

Original comment: Discussion, line 439-441. It seems more could be said about the effects of climate change on growth rates. Would river regulation reduce the effects of extreme conditions under climate change? Would higher water temperatures change the observed relationships to flood and drought intensity? Would higher temperatures make mild drought conditions less beneficial?

Reviewer’s response: More detailed discussion on this issue should be included. It does not have to be exhaustive, but should briefly describe the relevance of the results under a changing climate.

Authors’ answer: We have elaborated this paragraph of the Discussion as follows at what is now lines 474-492: “Our results have implications for the conservation and management of P. fasciatum, particularly in relation to the effects of construction of hydropower dams and climate change. Dams alter upstream and downstream river hydrology [50]. Upstream of dams, the effects of floods and droughts are quite effectively moderated, while downstream, their intensity generally decreases, becoming milder in most cases [50,51], which according to our results might be beneficial for the growth of P. fasciatum. However, the impacts of dams go far beyond the modulation of hydrological events, and include disruption of migration or access to spawning sites [30], impacts that could overcome any growth benefits of milder hydrological events. That is, even though the growth of P. fasciatum might increase downstream of the dam with milder hydrological variation, the negative effects of dams on migratory species might be more harmful to P. fasciatum because they threaten the completion of the species’ life cycle. Another implication of our study concerns the impacts of climate change. Climate change in the Amazon basin is expected to increase the frequency and intensity of extreme flood and drought events [52,53], which would be expected to induce slower growth of P. fasciatum. Our results thus provide a hypothesis for testing using growth data to be collected over time.”