Initial Floristic Response to High Severity Wildfire in an Old-Growth Coast Redwood (Sequoia sempervirens (D. Don) Endl.) Forest

Round 1

Reviewer 1 Report

This paper reports on 6‐month recovery of vegetation following the CZU fire, one of an increasing number of fires in California. I applaud the authors for getting the data out so quickly and agree that it is important to document post‐fire recovery following the growing number of fires given increases in temperature and aridity due to climate change. The paper is generally clearly written and informative but I have a number of general and specific comments.
1. The paper refers almost entirely to post‐fire recovery along the central California coast. If this paper is going to be published in an international journal then the authors need to spend a little more time putting the paper in a broader context and at least briefly discussing how the patterns observed here compare to other forests in California or Mediterranean climates elsewhere. The paper is currently written more for a regional journal. See Holl, K. D. 2010. Writing for an international audience. Restoration Ecology 18:135‐137.
2. I have several questions about the statistical analyses used which I discuss further below.
3. The authors need to make sure they aren’t writing in what I refer to as statistic‐ese. Tell the reader what the result in English without quite as much statistical jargon, as the tests used should be described in the methods. There is currently quite a bit of this tone in the abstract and results that should be rewritten to make the results clearer to readers. For example, in the abstract “In regard to tree sizes, Pearson’s correlation indicated that canopy retention and canopy regeneration were positively associated with tree height and diameter” could be written to say that “Both canopy survival and regeneration were greater for larger height and diameter trees.” I have noted a few but not all cases below.
4. One conclusion that I think is understated is that there is a large amount of resprouting quite quickly in this ecosystem. Since I live in the Santa Cruz Mountains, I know there is a lot of discussion about whether to plant or see following fire and a major contribution this paper makes is to dcoument the extent of natural regeneration. That point could be better highlighted.
5. The figure formatting needs work. It appears the authors used the Excel defaults and didn’t update them. While Excel isn’t great for graphics, almost any feature can be changed. Specifically, (a) the y‐axis line is missing – both the x and y‐axis lines should be black, (b) putting the percentages on the bars is redundant, (c) the default horizontal lines should be removed, (d) the box around the figure should be moved, (e) the font sizes need to be slightly larger to be readable with reduction, (f) figures 1 and 2 could be a single two panel figure and the species names removed on the top panel to save space, (g) there are currently two figure 4s.
Lines 21‐24 – Rewrite these more clearly for a general audience, reducing statistical jargon.
Line 27 – Rather than just saying that “tree size has a role” it is more informative to say that larger trees
have higher survival. Also, I would add a short call out to point 3 above.
Lines 98‐99. How large of an area were the sample points distributed across and was there a minimum distance between points?
Generally the data analysis section needs some work.
Line 114 – Can be deleted. Just say what statistical tests were used.
Line 115 – Linear correlations between which variables.
Lines 116‐117 “Linear regression analysis was employed as an additional predictive correlational test for the selected variables…” This wording is confusing as correlation and linear regression make different assumptions about causality so they shouldn’t be mixed in the same sentence. Also, it is not clear what the dependent and independent variables in this analysis are. Later in lines 174‐178 there also seems to be some confusion between correlation and regression that needs to be clarified.
Line 118 – How was plot included in the two sample t‐tests? A comparison of resprouting or not should not just treat each individual as randomly selected but should include some consideration of plot to at least partially account for spatial autocorrelation. In general, it seems like the individual tree was used as the replicate, whereas the plot was really the replicate so that needs to be incorporated in the model structure.
Line 121. R‐commander is a GUI not a program. The version of R used and packages are typically listed. What tests were done to ensure that the data fit the presumed distribution? i.e. Were the data or residuals tested for normality and homoschedasticity?
Line 128 – Avoid passive voice here and elsewhere.
Line 133‐134. Start the sentence with “Trees that survived…” Rest of text is unnecessary.
Line 145 – “between” (two only) should be “among” (comparison of >2).
Lines 162‐165 and Table 1. The authors report too many significant digits. The rule is one digit beyond the precision of the measurement which means to the tenths digit for number of basal sprouts. Likewise, in table 1 to the tenths digit should be sufficient and the table could be reformatted to present the numbers as mean ± SE to save space rather than having separate columns. That table might be moved to supplementary materials as I presume that the full species list will only be of interest to a small percentage of readers.
Line 224‐228. The authors note that more time is needed for seedling recruitment. They should more explicitly note that water year 2020‐2021 precipitation was well below normal which is likely one factor affecting the low recruitment from seed.
Line 231 – “correlated with”.

Author Response

Dear Reviewer 1,

Thank you for your thoughtful comments on our manuscript.  They were very helpful, and have resulted in an improved paper.  We have outlined our responses in bold to your suggestions below:

This paper reports on 6‐month recovery of vegetation following the CZU fire, one of an increasing number of fires in California. I applaud the authors for getting the data out so quickly and agree that it is important to document post‐fire recovery following the growing number of fires given increases in temperature and aridity due to climate change. The paper is generally clearly written and informative but I have a number of general and specific comments.
1. The paper refers almost entirely to post‐fire recovery along the central California coast. If this paper is going to be published in an international journal then the authors need to spend a little more time putting the paper in a broader context and at least briefly discussing how the patterns observed here compare to other forests in California or Mediterranean climates elsewhere. The paper is currently written more for a regional journal. See Holl, K. D. 2010. Writing for an international audience. Restoration Ecology 18:135‐137. – Broadened and revised abstract, introduction, and discussion to include discussion of Mediterranean fire regimes as a whole.
2. I have several questions about the statistical analyses used which I discuss further below.
3. The authors need to make sure they aren’t writing in what I refer to as statistic‐ese. Tell the reader what the result in English without quite as much statistical jargon, as the tests used should be described in the methods. There is currently quite a bit of this tone in the abstract and results that should be rewritten to make the results clearer to readers. For example, in the abstract “In regard to tree sizes, Pearson’s correlation indicated that canopy retention and canopy regeneration were positively associated with tree height and diameter” could be written to say that “Both canopy survival and regeneration were greater for larger height and diameter trees.” I have noted a few but not all cases below. – Manuscript was revised accordingly to remove statistical jargon throughout.
4. One conclusion that I think is understated is that there is a large amount of resprouting quite quickly in this ecosystem. Since I live in the Santa Cruz Mountains, I know there is a lot of discussion about whether to plant or see following fire and a major contribution this paper makes is to dcoument the extent of natural regeneration. That point could be better highlighted. – Added statement to this effect in the Discussion section.
5. The figure formatting needs work. It appears the authors used the Excel defaults and didn’t update them. While Excel isn’t great for graphics, almost any feature can be changed. Specifically, (a) the y‐axis line is missing – both the x and y‐axis lines should be black, (b) putting the percentages on the bars is redundant, (c) the default horizontal lines should be removed, (d) the box around the figure should be moved, (e) the font sizes need to be slightly larger to be readable with reduction, (f) figures 1 and 2 could be a single two panel figure and the species names removed on the top panel to save space, (g) there are currently two figure 4s.  – Revised as requested, with the exception of merging figures 1&2 (we think they work better separately).
Lines 21‐24 – Rewrite these more clearly for a general audience, reducing statistical jargon. – Abstract was revised to remove statistical reference.
Line 27 – Rather than just saying that “tree size has a role” it is more informative to say that larger trees
have higher survival. Also, I would add a short call out to point 3 above. – Revised accordingly.
Lines 98‐99. How large of an area were the sample points distributed across and was there a minimum distance between points? – Added information on study area size and distance between plots.
Generally the data analysis section needs some work.
Line 114 – Can be deleted. Just say what statistical tests were used. – Line deleted
Line 115 – Linear correlations between which variables. – Added list of variables
Lines 116‐117 “Linear regression analysis was employed as an additional predictive correlational test for the selected variables…” This wording is confusing as correlation and linear regression make different assumptions about causality so they shouldn’t be mixed in the same sentence. – Clarified and revised accordingly.  Also, it is not clear what the dependent and independent variables in this analysis are. – indicated independent and dependent variables in the text.  Later in lines 174‐178 there also seems to be some confusion between correlation and regression that needs to be clarified. – Revised and clarified this section.
Line 118 – How was plot included in the two sample t‐tests? A comparison of resprouting or not should not just treat each individual as randomly selected but should include some consideration of plot to at least partially account for spatial autocorrelation. In general, it seems like the individual tree was used as the replicate, whereas the plot was really the replicate so that needs to be incorporated in the model structure. – Revised and clarified this section.

Line 121. R‐commander is a GUI not a program. The version of R used and packages are typically listed. What tests were done to ensure that the data fit the presumed distribution? i.e. Were the data or residuals tested for normality and homoschedasticity? – Removed reference to R-commander 
Line 128 – Avoid passive voice here and elsewhere. – Revised as suggested.
Line 133‐134. Start the sentence with “Trees that survived…” Rest of text is unnecessary. – Revised as suggested.
Line 145 – “between” (two only) should be “among” (comparison of >2).  – Revised as suggested.
Lines 162‐165 and Table 1. The authors report too many significant digits. The rule is one digit beyond the precision of the measurement which means to the tenths digit for number of basal sprouts. Likewise, in table 1 to the tenths digit should be sufficient and the table could be reformatted to present the numbers as mean ± SE to save space rather than having separate columns. That table might be moved to supplementary materials as I presume that the full species list will only be of interest to a small percentage of readers. – Reduced digits to the tenths place for basal sprouts, and to the hundredth place for percent cover.  Reformatted table as suggested, but would prefer to leave in text, rather than as supplementary material, as understory flora is an important part of forest ecosystems and will be of interest to many readers.
Line 224‐228. The authors note that more time is needed for seedling recruitment. They should more explicitly note that water year 2020‐2021 precipitation was well below normal which is likely one factor affecting the low recruitment from seed. – A statement regarding the impact of the drought on seedling recruitment was added.
Line 231 – “correlated with”. – Revised as sugested.

Reviewer 2 Report

This study is timely as climate change is expected to increase the frequency of high severity fires in the studied forests. Nevertheless, my overall opinion is that the documented studies have not been thoroughly reviewed. I am afraid that the results of this study are difficult to regard as new findings because most information have already been explored, I believe. In addition, this manuscript lacks detailed description of the method, and the analysis and explanation of the results. This manuscript needs thorough revision.

More specific concerns and comments are as follows.

Methods: First, the authors should describe the fire occurrence in more detail. For example, what is the rationale for ‘high intensity wildfire’? They must also provide a reference. Second, what are the size criteria of the investigated trees? The authors said they counted ‘All trees (live and dead), as well as all live understory species’ (L. 101-102). But only 487 trees appeared in 30 plots? It seems that trees over a certain size have been investigated. Third, they used the words ‘understory species’ and ‘understory plants’ in the entire manuscript. I think they meant to indicate ‘associate trees’ and ‘herbaceous plants’, respectively. Does this make sense? In Table 1, however, the understory species seemed to be herbaceous plants. It is confusing.

Results: In general, forests damaged by wildfires vary greatly by location. Therefore, I would like the authors to present the results at the plot level as well. For example, in Figures 1 & 2, survival or canopy retention will show differences between plots.

Why is 'trees with live crown (%)’ 0? (Figure 2), even though Myca's survival is 75% (Figure 1). What was the decision criterion for ‘survival’ or ‘death’ of trees? Myca and Arme have too few samples.

Was it analyzed by species in Pearson correlation? Since each species has different characteristics, it must be analyzed separately. And before the Pearson analysis, did the authors test whether the variables were normally distributed? The result should be written to methods.

In Figure 5, as the tree height is an independent variable, this variable should be placed on the x-axis, I think. And please analyze it by species. Results may be different.

In Table 1, The authors examined 30 plots, and 16 understory species presented. There are so few number of species, why?

Figure 3 & 4 Please remove ‘Chart area’ in the figures.

Author Response

Dear Reviewer 2,

Thank you for your thoughtful comments on our manuscript.  They were very helpful, and have resulted in an improved paper.  We have outlined our responses in bold to your suggestions below:

This study is timely as climate change is expected to increase the frequency of high severity fires in the studied forests. Nevertheless, my overall opinion is that the documented studies have not been thoroughly reviewed. I am afraid that the results of this study are difficult to regard as new findings because most information have already been explored, I believe. In addition, this manuscript lacks detailed description of the method, and the analysis and explanation of the results. This manuscript needs thorough revision.

More specific concerns and comments are as follows.

Methods: First, the authors should describe the fire occurrence in more detail. For example, what is the rationale for ‘high intensity wildfire’? – Replaced “intensity” with “severity.”They must also provide a reference. – an added a reference with regard to thew use of “intensity” and “severity.”  Second, what are the size criteria of the investigated trees? The authors said they counted ‘All trees (live and dead), as well as all live understory species’ (L. 101-102). But only 487 trees appeared in 30 plots? It seems that trees over a certain size have been investigated. – Indication of tree size added (i.e. > 10 cm dbh) Third, they used the words ‘understory species’ and ‘understory plants’ in the entire manuscript. I think they meant to indicate ‘associate trees’ and ‘herbaceous plants’, respectively. Does this make sense? In Table 1, however, the understory species seemed to be herbaceous plants. It is confusing. – The term “understory is used throughout the manuscript to refer to woody shrubs and herbaceous plants (table 1 includes both of these).  The term “subcanopy” is used to describe trees species that do not reach the upper canopy in this forest type.  The term “associated” was is used to describe any species, plant or otherwise, that is a native part of the coast redwood ecosystem.

Results: In general, forests damaged by wildfires vary greatly by location. Therefore, I would like the authors to present the results at the plot level as well. For example, in Figures 1 & 2, survival or canopy retention will show differences between plots. – Clarified in the text that analysis was done at the plot level, so that error bars in Figures 1&2 in indicate this variation

Why is 'trees with live crown (%)’ 0? (Figure 2), even though Myca's survival is 75% (Figure 1). What was the decision criterion for ‘survival’ or ‘death’ of trees? Myca and Arme have too few samples. – Clarified in the text that the designation of “survival” was based on two factors crown retention, or resprouting.

Was it analyzed by species in Pearson correlation? – Clarified in the text. Since each species has different characteristics, it must be analyzed separately. And before the Pearson analysis, did the authors test whether the variables were normally distributed?  The result should be written to methods.  – Added information regarding tests for normality.

In Figure 5, as the tree height is an independent variable, this variable should be placed on the x-axis, I think. – Switched axis for figure 5. Results may be different.  – Revised as suggested. Added regression analyses for each species.

In Table 1, The authors examined 30 plots, and 16 understory species presented. There are so few number of species, why? – Added discussion in this regard.  The severity of the fire appears to have affected both the cove and richness of understory species.

Figure 3 & 4 Please remove ‘Chart area’ in the figures. – Revised as suggested.

Round 2

Reviewer 1 Report

Generally, the authors have thoroughly addressed my comments. I just have just two minor items and a follow up on figures 1 and 2.

Line 151. While the authors don’t need to mention that they used R-commander they do need to add a citation to which version of R they used (i.e. that was being run by the R-commander interface).

Line 253. There is a typo in “appears”.

There is no reason that Figures 1 and 2 cannot be combined. They are straightforward bar graphs and people make multipanel figure (often with many panels) all the time. The captions are mostly the same, so that would reduce duplication and the x-axis tick labels could be removed on the top figure. I actually find it clearer to have them on the same figure so one can compare different responses of the same species. Also, the numbers on the bars should be removed (an unnecessary Excel default) as they are duplicative and clutter the figure. The height of the bars indicates the values.

Author Response

We have done our best to incorporate the reviewers suggestions, as described below:

Generally, the authors have thoroughly addressed my comments. I just have just two minor items and a follow up on figures 1 and 2.

Line 151. While the authors don’t need to mention that they used R-commander they do need to add a citation to which version of R they used (i.e. that was being run by the R-commander interface). - Added version of R to the analysis section of the methods.

Line 253. There is a typo in “appears”. - Revised as suggested.

There is no reason that Figures 1 and 2 cannot be combined. They are straightforward bar graphs and people make multipanel figure (often with many panels) all the time. The captions are mostly the same, so that would reduce duplication and the x-axis tick labels could be removed on the top figure. I actually find it clearer to have them on the same figure so one can compare different responses of the same species. Also, the numbers on the bars should be removed (an unnecessary Excel default) as they are duplicative and clutter the figure. The height of the bars indicates the values. - Revised figures as suggested.  Combined figures 1+2 into one figure, and removed numbers on columns.

Reviewer 2 Report

The revised manuscript became much clearer by supplementing the missing or vague descriptions in the previous manuscript.

Author Response

No suggestions were given by the reviewer for revision 2.