Dear Authors,
I have now finished the review of the paper intitled ‘Similarity of stream insect trait profiles across biogeographic regions ’. The paper analyzes the functional traits of families across regions and identify two traits ‘combinations’- which I think should be considered ‘traits’ - that characterized the functional groups. Finally, they relate the functional groups to the number of climatic groups, suggesting that this would indicate a climatic variability (but I presume this should either be clarified, or even better, detailed).
Despite the statement in page 6, lines 16-18, the hypothesis being tested (page 5, line 24) and the conclusion (page 2, line 10) are that functional niches have 'converged' across regions; Convergence is highlighted along the text, which is misleading to the readers (e.g. page 4, lines 7-9; page 5, line 24). The text should be deeply restructured, to avoid confusion.

Page 1, line 40: Here it seems to be one single database, but in the text appears to be a combination.

Page 3, lines 43-49: The construction of the text leads to an expectation of and environment-trait analysis, which is not shown. Environment ends up not being assessed but roughly estimated through the region. I suggest either re-structuring the text, so it is not misleading, or to include a few environmental variables to actually assess the environment-trait relationship.

Page 6, lines 33-34: null models need more explanation on species occurrences/abundances, species pool and randomizations. It is clear from the supplementary material that randomizations across regions were made, which limits the conclusions, and should be in the text.
Also, the species pool used (or gathered in the datasets from the literature) should also be better detailed, given its major importance in this analysis (see Götzenberger et al. 2016 J Veg Sci https://doi.org/10.1111/jvs.12452).
PCoA axes can be used as new traits to compute FD indices, which could then be tested comparing the regions (Laliberte & Legendre 2010 Ecol. https://doi.org/10.1890/08-2244.1). In any case, some measure of the environmental variability in each region should be provided. I suggest either the authors provide more references and details on the statistical procedures.

Page 7, line 17: I do not think you have 22 traits, and the Grouping feature, in Table 1, seems to be the actual traits. This is also what I understand from Crabot et al. 2021, and I quote: ‘Each trait (e.g. locomotion mode) was described by several modalities (e.g. crawler, burrower).’. I believe that this will change the analysis, which might have considered 22 traits creating some sort of bias and a zero-inflated table – since only one modality is possible for each trait. You can correct for the number of levels (modalities) in each trait, to prevent biasing the importance of traits with different number of levels.

You are using categorical traits only – this should be highlighted in the text.

Page 9, lines 16-17: Is there a reference for separating Cold and Temperate regions in Europe? Why not to include a greater difference, which you already included in the other analysis (other regions with accentuated differences?). I presume that ‘minor differences’ can be due to the comparison within Europe? I see after reading the supplementary material that detailed data was not available for the other continents, which seems odd to me. In any case, I think this information should be in the main text, to clarify why you tested only for one region.

Lines 42-44: is there a reference for this threshold?

Page 10, lines 37-58: It is not clear why to answer this question with a regression. Looking at the figure, it seems that you should compare among the groups, instead of considering the climatic zones as quantitative (it is the same to each region, so not a true replicate). Given that climate (e.g temperature and precipitation) are relatively easy to get, I wonder why not to consider them at community level?

Page 14, lines 31-32: not clear the connection with taxonomic level.

Page 28, line 22: however you had traits at family level, can you can still gather information for the site, right? Species richness (or taxa?, as said) can (should?) be used as random factors, if you applied mixed-models to answer your questions.

Lines 31-36: maybe you should provide more information – shouldn’t New Zeland be the most out of the range region in your comparison? Why did you include taxonomic information only for this region?
This section could include the reason why you analyzed the climatic patterns only for Europe.

If I’m not mistaken, they can still calculate the community mean values, despite the species-or-family level of the traits. And with that, calculate the functional redundancy of communities. Even if this is not possible, after running the PCoA, a formal test such as a generalized mixed model is possible and recommended (Laliberte & Legendre 2000) – which could include climatic variables, if provided.