Composition and Spatial Variation of Germinable Seed Bank in Burned Nothofagus pumilio Forests in Patagonia Argentina

Round 1

Reviewer 1 Report

The submitted manuscript contains a sufficient number of references, which are used for an excellently written introduction and a fairly well-written discussion. However, the design of the study is insufficient, but it could be partly saved by an appropriately chosen data analysis, which the authors certainly did not perform. 

L15: Abbreviations of localities are presented even before the full names are presented - the full names are more suitable for the abstract. The abstract must function independently of the rest of the text. Moreover, on the contrary, full names appear throughout the manuscript, despite the fact that abbreviations are introduced in the methodology (and from this point they should function as a substitute for the full name). There is also no need to specify exact coordinates in the abstract, it is advisable to move this to the methodology.

L55: Rather moreover than however?

Materials and Methods: LC, MZ, and RT differ not only in the time when the fire occurred but in a lot of other parameters (altitude, rainfall, livestock presence,...), it is, therefore, reasonable that the authors attribute the differences to specific localities and not to the time after the fire across the manuscript (with exception of abstract). However, I still see several fundamental problems in the analysis methodology:

First of all, there were only two sampling transects, so for each distance level, only two replicates are available (should be at least three).

L138–L139: “The response variable was the emergence… recorded once a week.” This sounds like data suitable for survival analysis, such as a Cox model, if emergence was tracked over time. But if the records made once a week were not used, and in the end the total abundance was monitored, then there is no point in reporting once-a-week records. Similarly, on L152.

L155–L156: „The results were expressed as the average number of seeds belonging to each class.“ Versus L179: „The germination capacity of N. pumilio seed varied between 4 % … and 16 %.“ This contradicts itself. Additionally, the results should be reported with an estimate of accuracy. In addition, N. pumilio is not in italics and there should be no space in "4 %".

L159 and further: ANOVA. No, this is not an appropriately chosen analysis. 1) Analyzing each site separately sounds reasonable, but if you additionally work with individual distances as a qualitative variable (ANOVA), you only have two replications for each site (as you state on L163), which is too few. The number of degrees of freedom could be increased so that you would work with the distance completely simply and intuitively as with a quantitative variable (instead of ANOVA rather a linear model, or more precisely a Generalized linear mixed model - see further). 2) Individual localities could be used in the model as a random term - this way generalizable trends could be searched for. 3) The response variable is the "number of emerged seedlings", which is also very often zero. It's funny that you care about the sphericity of the covariance matrix, but you ignore the most important assumptions of ANOVA, which is the normal distribution of the data (and homoscedasticity of variances among groups), which are certainly violated. This could be solved in GLMM by choosing the right data distribution, in this case apparently negative binomial. 4) The Mauchly test is important if you have a repeated measurement design, which I don't really see here. In addition, this test is heavily criticized, because it often fails to detect departures from sphericity in small samples (which two measurements per site certainly are).

L171 and further: PERMANOVA: This is quite fine, but remember that the Bray-Curtis index is based on abundance data. If you have presence/absence data, you are probably using the Sorensen/Jaccard index (or using B-C will automatically flip in their use).

L208: In the results, I lack information about the significance of the results. The only exception is the P-value given on this row, but it is not clear whether it refers to the autumn or spring measurement (or both).

Figure 5. This figure is handled strangely. I prefer if the same size subplots are displayed with the same axis range. If this is not possible for the original data due to large ranges, I suggest a logarithmic transformation.

Figure 6. Some of the subplots are missing - even if they should be all zeros, it's better to show them.

L265 and further: NMDS not NMSD

L351: “by White and Pickett (1885) White and Pickett [46]” should be just White and Pickett [46]. Besides, it is 1985.

L409: An unnecessary part of the template is left here, please delete it.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

Thank you to the editor for inviting me to review the M. Florencia Urretavizcaya et al.’s paper "Composition and spatial variation of germinable seed bank in burned Nothofagus pumilio forests in Patagonia Argentina". Fire is a very common disturbance in forest systems, which plays a very important role in community succession or replacement, especially in deciduous broad-leaved forests. I am interested in the topic of this paper and think it has a certain contribution in revealing the ecological role of fire. However, I am not satisfied with the results of the present presentation, because it is too routine and ordinary.

 

I think the fires at three different years give us a very good object to study which factors that play a role in seed banks and seed germination after fires. However, this paper only analyzed the differences in seed bank and seed germination at the distance scale for each site, rather than comparing the differences in the restored forest after different fire ages. Moreover, the level of fire is a very important factor, which is not mentioned.

 

As mentioned in the paper, the other disturbances of the three sites after fire are different (sheep farming, cattle rearing and biological livestock). It is also very important whether they have any effect on seed germination and species composition, but it is not appeared in the current version.

 

In addition to the effect of distance (to the remnant forest), I believe that seed germination and species composition between the three sites were also related to soil properties and climate, because it located in different geographic coordinates and altitude. It is not enough to use distance to explain the change, which would not lead to convincing conclusions.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

I thank the authors for a very thorough revision in which they decided to address and respond to all my comments. I have only two minor comments about the new version:

L234-L237: "The distance to the limit of the remnant forest was treated as a fixed effect. The seven spatial locations per transect were nested in the transects and treated as random effects." Does this mean that transects were not used as a random term, but some coordinates of those patches in space?

L243: Readers deserve to know what type of transformation was used - logarithmic?

Author Response

Reply to Reviewer #1

I thank the authors for a very thorough revision in which they decided to address and respond to all my comments. I have only two minor comments about the new version:

RE: Thank you very much for your appreciation.

L234-L237: "The distance to the limit of the remnant forest was treated as a fixed effect. The seven spatial locations per transect were nested in the transects and treated as random effects." Does this mean that transects were not used as a random term, but some coordinates of those patches in space?

Response: In the random term we considered the transects and, since we took three soil samples per sample plot, we nested the soil samples to the corresponded transect. To clarify this aspect, we rewrote the sentence to "The distance to the limit of the remnant forest (0, 15, 30, 45, 60, 75, 90 m) was treated as a fixed effect. The transect (with two levels) and the soil sample plots (with three levels) nested in the transect, were treated as random effects ." (lines 186-189).

L243: Readers deserve to know what type of transformation was used - logarithmic?

Response: We used logarithmic transformation for some variables and square root transformation for others. To clarify this aspect we mentioned the transformation used in the methodology (line 195). Besides, for the adjusted models the used transformation is mentioned in the equation stated in the figure.

 

Reviewer 2 Report

The manuscript has been substantially improved by the authors. Compared to the original version, the updated version is well-structured, and well-specified. All flaws in the manuscript have been resolved. As it stands, I would like to suggest a few minor changes:

 

Point 1: The abstract should be followed by 1-2 sentences that summarize the manuscript and explain the research significance.

Point 2: This current paper discusses some deficiencies, such as the impact of grazing, but lacks the prospects regarding future research for these deficiencies. 

 

Point 3: The DOI numbers in the reference section are missing. It is necessary for authors to include DOI numbers for references in their manuscripts. 

Author Response

Reply to Reviewer #2

Comments and Suggestions for Authors

The manuscript has been substantially improved by the authors. Compared to the original version, the updated version is well-structured, and well-specified. All flaws in the manuscript have been resolved. As it stands, I would like to suggest a few minor changes:

 RE: Thank you very much for your appreciation

Point 1: The abstract should be followed by 1-2 sentences that summarize the manuscript and explain the research significance.

Response: We included two sentences in the abstract to highlight both points (Lines 27-29 and 30-31)

Point 2: This current paper discusses some deficiencies, such as the impact of grazing, but lacks the prospects regarding future research for these deficiencies. 

Response: We included one sentence in the Conclusion Section taking into account this topic (Lines 459-460)

Point 3: The DOI numbers in the reference section are missing. It is necessary for authors to include DOI numbers for references in their manuscripts. 

Response: We checked and included the missed DOI in the new version.