Review Reports

Round 1

Reviewer 1 Report

A very interesting study. My comments in are in the attached manuscript. 

Comments for author File: Comments.pdf

Author Response

Thank you for this positive feedback to our manuscript. We followed the comments and notes provided in the provided annotated PDF and revised our manuscript accordingly. We added standard error bars as requested to figure 2 (see also a similar comment of reviewer 2). We added more remote sensing focused language and numbers in the text body as requested. Regarding the comment that this paper refers to policy too often, we believe this paper can give policy makers a better understanding of how remote sensing can assist in environmental policy monitoring. We thus would like to keep the policy framing of this science to make it appeal to remote sensing scientists as a novel method they can utilize, for instance to inform EU policy makers on the impacts of PA on European forests. We expect that our study will inspire the audience of Remote Sensing how remote sensing can contribute to policy-related challenges and decision support, that is important worldwide.

Reviewer 2 Report

have read the manuscript remotesensing-400615 that I was invited to review on 20 November, accepted to review 22 November and returned 10 December. The paper looks at the NPP anomalies between forested Natura2000 sites (category I-II) against matched non-protected sites with the underlaying hypothesis that smaller anomalies suggest great stability/resilience. They find little (if any) difference in NPP anomalies at the continental scale, but some at regional scale (where PA exhibited smaller anomalies in the south). Interestingly they also find a general effect of stand-age with older forest showing higher “resilience”. I found the paper interesting and well written in a clear concise English. Especially figure 5 is an important contribution to our understanding of the importance of PA age. I think the paper would benefit from in the discussion slightly better address how important a measure such as NPP anomalies are for biodiversity values and whether the study period is sufficient to make strong conclusions on the climate-change questions raised in the paper.

 

I only have few comments to the manuscript:

Line 82-85 outlines some important mechanisms, but even when selecting only IUCN category I-II PAs, how much dead wood is actually in the reserved? From my knowledge, basically only a handful of PAs (like Bialowieza) actually have amount of deadwood that resembles natural conditions. This is likely to affect the results – isn’t it? Following from this I think I would be worth discussing a bit more the link between NPP anomalies and biodiversity values and how close they align.

 

Lines 108-11: I am seriously unsure how Newbold et al. 2015 supports this statement. And I think it runs counter to most matching analysis – including global ones. It might be that Europe constitute a particular landscape where 1) there are people most places, 2) infrastructure is largely good everywhere, and 3) accessibility plays a small role. But I am not sure I believe this and that would still be scale-independent. I would like to see if the results change if including a human-pressure variable or a stronger defence for not doing it.

 

It is unclear to me, if protected areas from category III-VI was excluded from the possible control sample. This would be essential.

 

Lines 129-132: I think you need to expand on this. the description of how forest age was calculated is almost missing entirely and this is the basis for some of the most interesting results (figure 5) so it should be clear how this was done.

 

The results in general seems a bit inconsistent in its use of terminology on the power of the results. Line 137 writes: “pronounced difference”, line 162: “Slightly greater”, line 166: “Statistically significant”. It would probably be good to make clear how these things relate to statistical test of their credibility, even when they are not significant. There might be more cases where this is relevant.

 

Overall, I am also a little surprised about how many of the results are presented without accounting for uncertainty (e.g. S.D. / S.E. / deviation) – like lines 150, 137, 162, 180 – and there are likely more.

 

Minor comments:

Check abbreviations. Maybe this paper has been submitted before to one of those journals that takes the scientific process so lightly that they have the methods in the end – because several terms seems to be written abbreviated in the methods to then be explained in results. Line 17: NPP is not explained. Line 89: CDDA is not explained. Line 141 not the correct name for IUCN (and IUCN has already been explained). Line 142: now CDDA is explained. There might be more.

 

Line 88: should it be “in” instead of “on” in “on or before 1995”

Line 136: delete “did” (end of line)

Line 143: have à had

Line 145: not sure I understand

Line 213: delete “Global”

Author Response

have read the manuscript remotesensing-400615 that I was invited to review on 20 November, accepted to review 22 November and returned 10 December. The paper looks at the NPP anomalies between forested Natura2000 sites (category I-II) against matched non-protected sites with the underlaying hypothesis that smaller anomalies suggest great stability/resilience. They find little (if any) difference in NPP anomalies at the continental scale, but some at regional scale (where PA exhibited smaller anomalies in the south). Interestingly they also find a general effect of stand-age with older forest showing higher “resilience”. I found the paper interesting and well written in a clear concise English. Especially figure 5 is an important contribution to our understanding of the importance of PA age. I think the paper would benefit from in the discussion slightly better address how important a measure such as NPP anomalies are for biodiversity values and whether the study period is sufficient to make strong conclusions on the climate-change questions raised in the paper.

I only have few comments to the manuscript:

Response: Thank you for this thorough and comprehensive review. We respond to the single comments below point-by-point.

Line 82-85 outlines some important mechanisms, but even when selecting only IUCN category I-II PAs, how much dead wood is actually in the reserved? From my knowledge, basically only a handful of PAs (like Bialowieza) actually have amount of deadwood that resembles natural conditions. This is likely to affect the results – isn’t it? Following from this I think I would be worth discussing a bit more the link between NPP anomalies and biodiversity values and how close they align.

Response: Thanks for this important observation. A discussion paragraph was now added to the conclusions about how this method may be used as a proxy measure for functional biodiversity as it relates to ecological niches.

Lines 108-11: I am seriously unsure how Newbold et al. 2015 supports this statement. And I think it runs counter to most matching analysis – including global ones. It might be that Europe constitute a particular landscape where 1) there are people most places, 2) infrastructure is largely good everywhere, and 3) accessibility plays a small role. But I am not sure I believe this and that would still be scale-independent. I would like to see if the results change if including a human-pressure variable or a stronger defence for not doing it.

Response: This citation is used here to show that socioeconomic models, unlike purely ecological models, are typically used on local to national scales. Our study focus on the ecological drivers of vegetation response to conservation and thus in this study we did not include covariates like proximity to roads or settlements, that indeed may impact forest function. The reliability and local accuracy of large-scale infrastructure datasets like Open-Street-Map might be a challenge in a continental analysis like our current study. We, however, acknowledge the value of such an exercise and added new language in the text to clarify our motivation.

It is unclear to me, if protected areas from category III-VI was excluded from the possible control sample. This would be essential.

Response: Thanks for raising this important issue. Categories III-IV were not included as protected areas, because those categories allow for human intervention and thus fall into our definition of actively managed.

Lines 129-132: I think you need to expand on this. the description of how forest age was calculated is almost missing entirely and this is the basis for some of the most interesting results (figure 5) so it should be clear how this was done.

Response: Thanks, we edited the text accordingly.

The results in general seems a bit inconsistent in its use of terminology on the power of the results. Line 137 writes: “pronounced difference”, line 162: “Slightly greater”, line 166: “Statistically significant”. It would probably be good to make clear how these things relate to statistical test of their credibility, even when they are not significant. There might be more cases where this is relevant.

Response: These sections have been removed, revised or replaced accordingly.

Overall, I am also a little surprised about how many of the results are presented without accounting for uncertainty (e.g. S.D. / S.E. / deviation) – like lines 150, 137, 162, 180 – and there are likely more.

Response: Standard error has now been added to figure 2 to better represent error. All graphs and charts have some measure of uncertainty associated with them. The numbers in the text refer to these figures and we believe that repeating the uncertainty information in the text would have made the text repetitive and number heavy.

Minor comments:

Check abbreviations. Maybe this paper has been submitted before to one of those journals that takes the scientific process so lightly that they have the methods in the end – because several terms seems to be written abbreviated in the methods to then be explained in results. Line 17: NPP is not explained. Line 89: CDDA is not explained. Line 141 not the correct name for IUCN (and IUCN has already been explained). Line 142: now CDDA is explained. There might be more.

Response: Thank you. We followed these suggestions.

Line 88: should it be “in” instead of “on” in “on or before 1995”

Response: Corrected.

Line 136: delete “did” (end of line)

Response: Done.

Line 143: have à had

Response: Done

Line 145: not sure I understand

Response: New language was added to clarify this sentence.

Line 213: delete “Global”

Response: Done