Below the canopy: global trends in forest vertebrate populations and their drivers

Global forest assessments use forest area as an indicator of biodiversity status, which may mask below-canopy pressures driving forest biodiversity loss and ‘empty forest’ syndrome. The status of forest biodiversity is important not only for species conservation but also because species loss can have consequences for forest health and carbon storage. We aimed to develop a global indicator of forest specialist vertebrate populations to improve assessments of forest biodiversity status. Using the Living Planet Index methodology, we developed a weighted composite Forest Specialist Index for the period 1970–2014. We then investigated potential correlates of forest vertebrate population change. We analysed the relationship between the average rate of change of forest vertebrate populations and satellite-derived tree cover trends, as well as other pressures. On average, forest vertebrate populations declined by 53% between 1970 and 2014. We found little evidence of a consistent global effect of tree cover change on forest vertebrate populations, but a significant negative effect of exploitation threat on forest specialists. In conclusion, we found that the forest area is a poor indicator of forest biodiversity status. For forest biodiversity to recover, conservation management needs to be informed by monitoring all threats to vertebrates, including those below the canopy.


Introduction
In the introduction the authors note that various international agreements and conventions use forest cover as a proxy for forest biodiversity. I feel this is a simplification as nowhere does it say that, for example, CBD uses forest cover as a measure of forest biodiversity. The reference in Target 5 to 'other habitats' instead indicates it is being used as a measure of forest habitats and not biodiversity. Perhaps some re wording is needed.
Citation needed to support statement in line 87 about carbon and health and line 90 on meeting the 1.5 degrees climate change target.
On line 109 it would be good to clarify why the authors hypothesise that forest biodiversity is declining. There is information in the introduction about forest loss, but this relates to forest loss not species. I think there is a missing link here to get to the hypothesis.

Methods
Line 191 -are the two measures of tree cover loss correlated? Errors in the tree cover data could explain an absence of correlation with observed species trend. It would be good to know if the tree data agree at least. At the very least add some of the accuracy assessments for the Song data from their analysis to help the reader assess how accurate such 5 km data are, but for preference add a correlation between change on the site studies here. This is of particular importance given that you then note in lines 214 -216 that there was considerable fluctuation in tree loss estimates. You need to convince the reader that these 5 km data really reliable and appropriate for this analysis.
Line 195 -the Song et al and Hansen et al data refer to tree cover, not forest cover. The consistence of use of tree cover should be checked through the manuscript and standardised. You make this point in line 206, but i you need to be consistent. Also, please be consistent in the use of negative or inverse correlations in the manuscript.
Line 224 -please give a summary of the methods of Spooner et al so readers who do not have access to it can understand what you did. In the calculation of these figures of population change did you consider weighting by a transformed SE in the subsequent analysis? This is reasonably common place in these types of analysis and ensures that the better relationships carry more weight into the analysis. The same applies to tree cover change estimates (ie those derived from the beta for the year effect). Also on the statistical analysis why did you consider variables individually rather than in a multi variate model? I can foresee potential interactions between variables that might be good to examine. For example, a decrease in tree cover by 50% might be expected to have a different impact if initial cover was 100% than if initial cover was 50%. Figure 1 is the key result here, but I feel that the presentation of geographical / taxanomic break downs in the same format as the LPI graph would be more valuable. This would enable a rapid comparison of the trends in, for example, the Nearctic with Africa. Presentation of this trend, and the errors around it, would help in the geographical interpretation of the results. Additionally, could the LPI trend for all species be added for comparison?

Results
The tables summarizing the models raised some questions for me. First is why the authors use AIC and statistical significance. This is odd as these are two philosophies for assessing variable importance. I suggest they pick one or the other and not both. This is particularly important when the results conflict -In table 1 there is apparently a significant relationship between change and body mass (P = 0.002), but the associated AIC value indicates that AIC increases, suggesting the addition of the variable did not improve the model. This needs to be explained.
In Table 2 and subsequent tables, how was delta AIC generated for the null model? Was this in fact not the null model but the model with the random effect for location added? In which case please give the original AIC too for the null model. Please give the direction of the change in AIC too to improve the clarity of tables. Finally, the parameter estimates are often small -there is little value in giving a value of 0.00. Could you rescale the variables so that you can quote informative parameter estimates.
Line 341 -I am not sure these are drivers as much as correlates. Please change wording.
On the cross correlation / literature review part of the analysis I feel that more information is needed to give the results more context. First, while I see the logic behind the assessment of positive and negative correlations with cover change (ie the use of 2*√n), I feel that there are more robust methods that could be used that make better use of the data (see above in comments on methods). I also think that the numerical comparisons presented in this section of the results could be improved upon, even with % bar graphs rather than text describing results. Importantly, how many of the species for which no relationship (positive or negative) were threatened by the pressures listed? You only present information for the 71 with relationships, but we need to know how many with no relationships were also threatened to put the results in context.

Discussion
The discussion is generally well framed, but consideration of some of the points raised above, especially the threats to the species for which no correlation with loss was detected, as noted above, might change some of the content of this discussion.
In presenting the forest LPI I was hoping there would be a comparison with the overall LPI, whether visually, or just in text. Ideally the LPI would be added to figure 1 so readers can make the comparison. Even better would be a statistical comparison of the LPI and forest LPI. This prospect is raised on line 140.
The contention that satellite data alone are insufficient to monitor biodiversity is not new, but is worth stressing as done here to ensure that conservation monitoring is complete. One question that does come to mind here though is what proportion of the world's forests / forest species are covered by the LPI compared to the potential of remote sensing for tracking / highlighting concerns across the globe? Until resources are available for more complete field based monitoring is it not better that we have some metric of the global situation? The need for better monitoring, as mentioned in lines 435 -436 is to me a key point of the paper and should be repeated in the closing paragraph.

Review form: Reviewer 2
Recommendation Accept with minor revision (please list in comments)

Scientific importance: Is the manuscript an original and important contribution to its field? Good
General interest: Is the paper of sufficient general interest? Good Quality of the paper: Is the overall quality of the paper suitable? Good Is the length of the paper justified? Yes

Do you have any concerns about statistical analyses in this paper? If so, please specify them explicitly in your report. No
It is a condition of publication that authors make their supporting data, code and materials available -either as supplementary material or hosted in an external repository. Please rate, if applicable, the supporting data on the following criteria.

Do you have any ethical concerns with this paper? No
Comments to the Author GENERAL COMMENTS: In this paper the authors follow the LPI methodology to develop a global indicator of forest specialist vertebrates to detect population trends over time . They also assess the influence of potential drivers of forest vertebrate population change, and compare trends across species groups, realms and biomes. The results show significant declines in forest vertebrate populations over the study period, but no consistent effects of most of the predictors analysed (except for 'over-exploitation'). The paper is well written and generally clear. The analyses seem robust. The resulting forest specialist index seems useful. The lack of association of population trends with potential drivers of change supports the conclusion that forest area (alone) is a poor proxy of forest biodiversity status, and that more detailed (e.g. forest quality) information is need. Below I provide some specific comments which I hope the authors will find useful to improve the quality of their paper.
SPECIFIC COMMENTS: L31 -this line is repetitive with L 25, I suggest removing it. L59 -fragmentation is a result of habitat loss and can be caused by many drivers. I suggest rewording / being more specific (e.g. fragmentation caused by roads). L78 -'the pertinence of forest area as a proxy indicator of forest biodiversity has never been tested'. Presumably the authors mean on a global scale? There have been lots of studies relating forest area to biodiversity (though not on a global scale as far as am aware). L87 -is a citation missing at the end of this sentence? L120 -what does 'traceable' mean in this context? L141 -'each taxa'… I think this should be the singular 'taxon' L147 -change 'tends' to tend' L165 -why were forest generalists included here, but not in the forest specialist index described above? Please clarify/justify. L184 -what's the temporal scale of the 'shorter-term' dataset? L195-207 -how was this mismatch in the inclusion of 'plantations' dealt with in analyses and interpreted? I.e forest area datasets include plantations, but the forest specialists index does not (stated in L134) L209 -some text / references are needed to justify the use of 5 km radii. L214-218 -this text is a bit unclear and might benefit from some rewording. L223 -'≥2 years of data covering a ≥5 year period'… this is unclear, please reword/clarify. L246-247 -please add supporting references for this or clarify if it's a hypothesis-driven sentence. L292 -please clarify the rationale behind repeating the analyses with the genus omitted. L294 -this subtitle seems to refer to text in L314-321 only; I suggest moving further down and renaming the section in L296-312. L310-311 -a citation is needed for this method and why it's appropriate in this case. L314-321 -was this a qualitative assessment only? L317 -I thought plantation forests were not included; please clarify. L338-340 -in SM 6, in addition to total number shown at the moment, it would be useful to see the percentage of species in each category (increasing, decreasing, stable) for each realm & biome. L355 -I'm not used to seeing negative AIC values. It would perhaps be useful to clarify (for other readers like myself) that a 'more negative' AIC value indicates a better model fit (if this is indeed the case?) L358 -although according to table 2 delta AIC for this model was zero (so no model improvement)… it's hard to figure out what criteria the authors are using in each instance (p value, confidence interval, AIC); this should be made clearer and ensure consistency.
L403 -It would be useful to put these results in a wider context by commenting on how trends forest vertebrate populations compare to overall LPI, e.g. are they faring better/worse/similar? L411 -I suggest rephrasing to '…whether the latest upturn in the FSI is a sign of a significant, longer-term increase in the abundance of…' L416 -and from declining to stable? L418 -In what way is the data turnover taking place? E.g. x declining species 'leaving' the dataset vs x increasing species 'entering'… L458-466 -I suggest including a few sentences about the potential mechanisms driving some of these trends.

13-Sep-2019
Dear Dr Hill: I am writing to inform you that your manuscript RSPB-2019-1785 entitled "Below the canopy: global trends in forest vertebrate populations and their drivers" has, in its current form, been rejected for publication in Proceedings B.
This action has been taken on the advice of referees and the AE, who have recommended that substantial revisions are necessary (please see their detailed comments, below). With this in mind we would be happy to consider a resubmission, provided the comments of the referees are fully addressed. However please note that this is not a provisional acceptance and we will send your revised manuscript back out for review.
The resubmission will be treated as a new manuscript. However, we will approach the same reviewers if they are available and it is deemed appropriate to do so by the Editor. Please note that resubmissions must be submitted within six months of the date of this email. In exceptional circumstances, extensions may be possible if agreed with the Editorial Office. Manuscripts submitted after this date will be automatically rejected.
Please find below the comments made by the referees, not including confidential reports to the Editor, which I hope you will find useful. If you do choose to resubmit your manuscript, please upload the following: 1) A 'response to referees' document including details of how you have responded to the comments, and the adjustments you have made. 2) A clean copy of the manuscript and one with 'tracked changes' indicating your 'response to referees' comments document. 3) Line numbers in your main document.
To upload a resubmitted manuscript, log into http://mc.manuscriptcentral.com/prsb and enter your Author Centre, where you will find your manuscript title listed under "Manuscripts with Decisions." Under "Actions," click on "Create a Resubmission." Please be sure to indicate in your cover letter that it is a resubmission, and supply the previous reference number.
Sincerely, Dr Sarah Brosnan Editor, Proceedings B mailto: proceedingsb@royalsociety.org Associate Editor Board Member: 1 Comments to Author: The study by Hill et al. uses the Living Planet Index methodology to develop an index for forest specialist vertebrates and explore its changes over time. It then examines the underlying drivers of the observed temporal trends with specific focus on forest cover. This paper has been reviewed by two referees. While reviewer 2 was more positive, reviewer 1 had a number of serious concerns about the methodology of the study, which I share. Both reviewers questioned the claim by the authors that forest cover is commonly used as a proxy for forest biodiversity while the relationship between the two has never been tested. This claim needs to be substantiated by references. It is also important to distinguish between relationship between forest cover and biodiversity and changes in forest cover and biodiversity. Both were tested in this study, but there is no reason why both should show the same trends, e.g. if there is a threshold of % forest cover above which changes in forest cover are not significantly affecting biodiversity, then different outcome might be expected (see comment by reviewer 1 about the importance of exploring the interactions between predictors instead of considering predictors individually). Another problem picked up by the reviewer 1 is that for the second part of their analyses, the authors have focused only on 71 populations for which significant relationship with forest cover was found. This resulted in exclusion of 104 populations with no correlation with forest cover from the analysis. The reviewer pointed out that it would be also very important to know the characteristics of species for which no relationship with forest cover has been detected and how many of these species are also threatened by the various pressures examined. In addition, this reviewer has questioned the methodology used for categorizing cross-correlations into positive, negative or uncorrelated with forest cover, which is very much akin to vote-counting. I share this concern. The 2*sqrt n approach needs to be better justified and ideally a more robust methodology used. Overall, while the study addresses an important topic, is timely and well written, I am afraid due to the identified methodological shortcomings it is not acceptable for publication in Proc Roy Soc B in its current form. If the authors are able to address the comments raised by the referee, the study might become acceptable, but it would require significant changes to the methodology and analyses presented, which is beyond the revision option.
Reviewer(s)' Comments to Author: Referee: 1 Comments to the Author(s) This manuscript describes the production of a Living Planet Index for forest species. It extracts trend data for forest species from the existing LPI database. The manuscript then goes on to compare these population trends with tree cover change assessed from satellite remote sensing data at the relevant location. Although no overall correlation between species population trends and tree cover change trends are found the authors do suggest that there are relationships for a large number (c 40%) of species examined. They then go on to undertake a simple description of the characteristics of the species for which relationships are detected, albeit without comparison back to the characteristics of species for which no relationships were detected.
The production of the LPI for forest species and a comparison between observed trends and satellite remote sensing tree cover data is an original and worthwhile topic for a manuscript. As such the manuscript has much to commend it. However, there are some areas that need to be considered.

Introduction
In the introduction the authors note that various international agreements and conventions use forest cover as a proxy for forest biodiversity. I feel this is a simplification as nowhere does it say that, for example, CBD uses forest cover as a measure of forest biodiversity. The reference in Target 5 to 'other habitats' instead indicates it is being used as a measure of forest habitats and not biodiversity. Perhaps some re wording is needed.
Citation needed to support statement in line 87 about carbon and health and line 90 on meeting the 1.5 degrees climate change target.
On line 109 it would be good to clarify why the authors hypothesise that forest biodiversity is declining. There is information in the introduction about forest loss, but this relates to forest loss not species. I think there is a missing link here to get to the hypothesis.

Methods
Line 191 -are the two measures of tree cover loss correlated? Errors in the tree cover data could explain an absence of correlation with observed species trend. It would be good to know if the tree data agree at least. At the very least add some of the accuracy assessments for the Song data from their analysis to help the reader assess how accurate such 5 km data are, but for preference add a correlation between change on the site studies here. This is of particular importance given that you then note in lines 214 -216 that there was considerable fluctuation in tree loss estimates. You need to convince the reader that these 5 km data really reliable and appropriate for this analysis.
Line 195 -the Song et al and Hansen et al data refer to tree cover, not forest cover. The consistence of use of tree cover should be checked through the manuscript and standardised. You make this point in line 206, but i you need to be consistent. Also, please be consistent in the use of negative or inverse correlations in the manuscript.
Line 224 -please give a summary of the methods of Spooner et al so readers who do not have access to it can understand what you did. In the calculation of these figures of population change did you consider weighting by a transformed SE in the subsequent analysis? This is reasonably common place in these types of analysis and ensures that the better relationships carry more weight into the analysis. The same applies to tree cover change estimates (ie those derived from the beta for the year effect). Also on the statistical analysis why did you consider variables individually rather than in a multi variate model? I can foresee potential interactions between variables that might be good to examine. For example, a decrease in tree cover by 50% might be expected to have a different impact if initial cover was 100% than if initial cover was 50%. Figure 1 is the key result here, but I feel that the presentation of geographical / taxanomic break downs in the same format as the LPI graph would be more valuable. This would enable a rapid comparison of the trends in, for example, the Nearctic with Africa. Presentation of this trend, and the errors around it, would help in the geographical interpretation of the results. Additionally, could the LPI trend for all species be added for comparison?

Results
The tables summarizing the models raised some questions for me. First is why the authors use AIC and statistical significance. This is odd as these are two philosophies for assessing variable importance. I suggest they pick one or the other and not both. This is particularly important when the results conflict -In table 1 there is apparently a significant relationship between change and body mass (P = 0.002), but the associated AIC value indicates that AIC increases, suggesting the addition of the variable did not improve the model. This needs to be explained.
In Table 2 and subsequent tables, how was delta AIC generated for the null model? Was this in fact not the null model but the model with the random effect for location added? In which case please give the original AIC too for the null model. Please give the direction of the change in AIC too to improve the clarity of tables. Finally, the parameter estimates are often small -there is little value in giving a value of 0.00. Could you rescale the variables so that you can quote informative parameter estimates.
Line 341 -I am not sure these are drivers as much as correlates. Please change wording.
On the cross correlation / literature review part of the analysis I feel that more information is needed to give the results more context. First, while I see the logic behind the assessment of positive and negative correlations with cover change (ie the use of 2*√n), I feel that there are more robust methods that could be used that make better use of the data (see above in comments on methods). I also think that the numerical comparisons presented in this section of the results could be improved upon, even with % bar graphs rather than text describing results. Importantly, how many of the species for which no relationship (positive or negative) were threatened by the pressures listed? You only present information for the 71 with relationships, but we need to know how many with no relationships were also threatened to put the results in context.

Discussion
The discussion is generally well framed, but consideration of some of the points raised above, especially the threats to the species for which no correlation with loss was detected, as noted above, might change some of the content of this discussion.
In presenting the forest LPI I was hoping there would be a comparison with the overall LPI, whether visually, or just in text. Ideally the LPI would be added to figure 1 so readers can make the comparison. Even better would be a statistical comparison of the LPI and forest LPI. This prospect is raised on line 140.
The contention that satellite data alone are insufficient to monitor biodiversity is not new, but is worth stressing as done here to ensure that conservation monitoring is complete. One question that does come to mind here though is what proportion of the world's forests / forest species are covered by the LPI compared to the potential of remote sensing for tracking / highlighting concerns across the globe? Until resources are available for more complete field based monitoring is it not better that we have some metric of the global situation? The need for better monitoring, as mentioned in lines 435 -436 is to me a key point of the paper and should be repeated in the closing paragraph.

Referee: 2
Comments to the Author(s) GENERAL COMMENTS: In this paper the authors follow the LPI methodology to develop a global indicator of forest specialist vertebrates to detect population trends over time . They also assess the influence of potential drivers of forest vertebrate population change, and compare trends across species groups, realms and biomes. The results show significant declines in forest vertebrate populations over the study period, but no consistent effects of most of the predictors analysed (except for 'over-exploitation'). The paper is well written and generally clear. The analyses seem robust. The resulting forest specialist index seems useful. The lack of association of population trends with potential drivers of change supports the conclusion that forest area (alone) is a poor proxy of forest biodiversity status, and that more detailed (e.g. forest quality) information is need. Below I provide some specific comments which I hope the authors will find useful to improve the quality of their paper. SPECIFIC COMMENTS: L31 -this line is repetitive with L 25, I suggest removing it. L59 -fragmentation is a result of habitat loss and can be caused by many drivers. I suggest rewording / being more specific (e.g. fragmentation caused by roads). L78 -'the pertinence of forest area as a proxy indicator of forest biodiversity has never been tested'. Presumably the authors mean on a global scale? There have been lots of studies relating forest area to biodiversity (though not on a global scale as far as am aware). L87 -is a citation missing at the end of this sentence? L120 -what does 'traceable' mean in this context? L141 -'each taxa'… I think this should be the singular 'taxon' L147 -change 'tends' to tend' L165 -why were forest generalists included here, but not in the forest specialist index described above? Please clarify/justify. L184 -what's the temporal scale of the 'shorter-term' dataset? L195-207 -how was this mismatch in the inclusion of 'plantations' dealt with in analyses and interpreted? I.e forest area datasets include plantations, but the forest specialists index does not (stated in L134) L209 -some text / references are needed to justify the use of 5 km radii. L214-218 -this text is a bit unclear and might benefit from some rewording. L223 -'≥2 years of data covering a ≥5 year period'… this is unclear, please reword/clarify. L246-247 -please add supporting references for this or clarify if it's a hypothesis-driven sentence. L292 -please clarify the rationale behind repeating the analyses with the genus omitted. L294 -this subtitle seems to refer to text in L314-321 only; I suggest moving further down and renaming the section in L296-312. L310-311 -a citation is needed for this method and why it's appropriate in this case. L314-321 -was this a qualitative assessment only? L317 -I thought plantation forests were not included; please clarify. L338-340 -in SM 6, in addition to total number shown at the moment, it would be useful to see the percentage of species in each category (increasing, decreasing, stable) for each realm & biome. L355 -I'm not used to seeing negative AIC values. It would perhaps be useful to clarify (for other readers like myself) that a 'more negative' AIC value indicates a better model fit (if this is indeed the case?) L358 -although according to table 2 delta AIC for this model was zero (so no model improvement)… it's hard to figure out what criteria the authors are using in each instance (p value, confidence interval, AIC); this should be made clearer and ensure consistency. L403 -It would be useful to put these results in a wider context by commenting on how trends forest vertebrate populations compare to overall LPI, e.g. are they faring better/worse/similar? L411 -I suggest rephrasing to '…whether the latest upturn in the FSI is a sign of a significant, longer-term increase in the abundance of…' L416 -and from declining to stable? L418 -In what way is the data turnover taking place? E.g. x declining species 'leaving' the dataset vs x increasing species 'entering'… L458-466 -I suggest including a few sentences about the potential mechanisms driving some of these trends.

Scientific importance: Is the manuscript an original and important contribution to its field? Excellent
General interest: Is the paper of sufficient general interest? Excellent Quality of the paper: Is the overall quality of the paper suitable? Excellent

Do you have any concerns about statistical analyses in this paper? If so, please specify them explicitly in your report. No
It is a condition of publication that authors make their supporting data, code and materials available -either as supplementary material or hosted in an external repository. Please rate, if applicable, the supporting data on the following criteria.

Do you have any ethical concerns with this paper? No
Comments to the Author This is an interesting and timely manuscript that attempts to highlight the need for field collected data in monitoring. As we head into the next CBD cycle, we need informative studies such as this to help identify relevant indicators that link to biodiversity. This manuscript describes an indicator that could be complimentary to the satellite image derived assessments of forest change. The manuscript does ned some areas of clarification in addition to the specific comments below: The authors need to be clear whether they are looking at forest cover from remote sensing data or forest change. The compare forest change with index change, but in the text refer to both forest cover and cover change. Consequently, it is unclear to me what they want the reader to focus on. The matter is confused more by their presenting a correlation between forest cover from Song et al and Hanse for 2010, but not forest change. I see this was requested in the previous set of reviews too (1.4) but was not implemented. It is essential that they present a correlation of forest change between a comparable time period (2000 -2010) for their study sites using both sets of satellite image data. Without this it is not possible to assess how consistent the measures of change are. Related to this, I disagree with sentence lines 103 to 105 which is a fundamental point of the paper language but not content -you look at change in indices. Forest cover will not correlate with FSI for obvious reasons (100 ha boreal forest will have very different FSI than 100 ha humid tropical forest), but instead it might correlate with change. This brings me to another point which is would we expect change in 10 ha of one to result in change of similar magnitude in biodiversity in another. Fitting interaction terms to the models of FSI and satellite data would help examine this effect.
The authors need to present a better explanation of why they chose the correlates that they did (line 262) as currently none given. Otherwise it appears an eclectic pick and mix selection. Was the selection based on previous studies or assessment of level 2 or 3 IUCN threats?
Finally, in their summation, the authors correctly and importantly state the need to have field measurements of biodiversity rather than relying on remote sensing alone. They are not the first to state this, and perhaps some recognition that this is oft stated would balance the conclusions a bit.

Specific comments.
Title -It's a personal thing, but I do not like titles with colons as one part of the title can almost always be deleted. I suggest deleting below the canopy here.
Ln 20 -you look at the correlation with forest change first -change wording to note this. Ln22 -It is not effects of satellite derived tree cover trends, it is a correlation between SPI and tree cover from sat data that you look at. There is a very important difference. Ln 25 is okay. Ln 27 Monitoring alone will not allow forest species to recover. Much more must be done. Change wording here to reflect fact that monitoring allows us to track recovery and implement adaptive management or wording along those lines Ln35 -Not really a deadline, rather review date. Ln 35 -delete legally established unless in FAO wording -it would be interesting to see an illegally established protected area Ln 86 -use °C Ln 90 -delete 'Recent' as it is a time limited term in this case Ln 102 -derived from rather than using as you use correlation to assess the relationship Ln 106 -109. Point 1 is not a hypothesis as currently worded. Ln 119 -Data are, not data is. Use if 'is' jars when reading. Also ln 179. Ln 180 -six, not 6 Ln 184 -what is a biologically useful resolution? Is it a resolution that is appropriate to the size of your study sites from which the population data came? That is probably a more important reason.
Ln 218 -222 -this section appears to be out of place and should be in the additional correlates section. LN 253 -255 -this is a massive reduction in the number of populations used. You drop specialists by 200 for Hansen based on my calculation. Is there a reason for this as to me this drop indicates a serious issue with these data? I think you need to add more of an explanation as to why you go from 268 (ln 137) to 74. It might be I have missed something in the text.
Ln 257 -260 -This is a measure of correlation in area, but you are looking at change in the analysis. Would it be possible to look at the correlation between change 2000 -2010 in both datasets as a more relevant measure of agreement?
Ln 311 -delete 'wished' and reword, but otherwise good to see the inclusion of this analysis here.
Ln 342 -Are these SE or 95% Cis around parameter estimates? The non-overlapping nature of the bars in Figure 2 is interesting and worth mentioning. It isn't mentioned anywhere in the manuscript, but is the massive decline in herptiles due to Chytrid? Your manuscript has now been peer reviewed and the reviews have been assessed by an Associate Editor. The reviewers' comments (not including confidential comments to the Editor) and the comments from the Associate Editor are included at the end of this email for your reference. As you will see, the reviewers and the Editors have raised a few concerns with your manuscript and we would like to invite you to revise your manuscript to address them.
We do not allow multiple rounds of revision so we urge you to make every effort to fully address all of the comments at this stage. If deemed necessary by the Associate Editor, your manuscript will be sent back to one or more of the original reviewers for assessment. If the original reviewers are not available we may invite new reviewers. Please note that we cannot guarantee eventual acceptance of your manuscript at this stage.
To submit your revision please log into http://mc.manuscriptcentral.com/prsb and enter your Author Centre, where you will find your manuscript title listed under "Manuscripts with Decisions." Under "Actions", click on "Create a Revision". Your manuscript number has been appended to denote a revision.
When submitting your revision please upload a file under "Response to Referees" in the "File Upload" section. This should document, point by point, how you have responded to the reviewers' and Editors' comments, and the adjustments you have made to the manuscript. We require a copy of the manuscript with revisions made since the previous version marked as 'tracked changes' to be included in the 'response to referees' document.
Your main manuscript should be submitted as a text file (doc, txt, rtf or tex), not a PDF. Your figures should be submitted as separate files and not included within the main manuscript file.
When revising your manuscript you should also ensure that it adheres to our editorial policies (https://royalsociety.org/journals/ethics-policies/). You should pay particular attention to the following: Research ethics: If your study contains research on humans please ensure that you detail in the methods section whether you obtained ethical approval from your local research ethics committee and gained informed consent to participate from each of the participants.
Use of animals and field studies: If your study uses animals please include details in the methods section of any approval and licences given to carry out the study and include full details of how animal welfare standards were ensured. Field studies should be conducted in accordance with local legislation; please include details of the appropriate permission and licences that you obtained to carry out the field work.
Data accessibility and data citation: It is a condition of publication that you make available the data and research materials supporting the results in the article. Datasets should be deposited in an appropriate publicly available repository and details of the associated accession number, link or DOI to the datasets must be included in the Data Accessibility section of the article (https://royalsociety.org/journals/ethics-policies/data-sharing-mining/). Reference(s) to datasets should also be included in the reference list of the article with DOIs (where available).
In order to ensure effective and robust dissemination and appropriate credit to authors the dataset(s) used should also be fully cited and listed in the references.
If you wish to submit your data to Dryad (http://datadryad.org/) and have not already done so you can submit your data via this link http://datadryad.org/submit?journalID=RSPB&manu=(Document not available), which will take you to your unique entry in the Dryad repository.
If you have already submitted your data to dryad you can make any necessary revisions to your dataset by following the above link.
For more information please see our open data policy http://royalsocietypublishing.org/datasharing.
Electronic supplementary material: All supplementary materials accompanying an accepted article will be treated as in their final form. They will be published alongside the paper on the journal website and posted on the online figshare repository. Files on figshare will be made available approximately one week before the accompanying article so that the supplementary material can be attributed a unique DOI. Please try to submit all supplementary material as a single file.
Online supplementary material will also carry the title and description provided during submission, so please ensure these are accurate and informative. Note that the Royal Society will not edit or typeset supplementary material and it will be hosted as provided. Please ensure that the supplementary material includes the paper details (authors, title, journal name, article DOI). Your article DOI will be 10.1098/rspb.[paper ID in form xxxx.xxxx e.g. 10.1098/rspb.2016.0049].
Please submit a copy of your revised paper within three weeks. If we do not hear from you within this time your manuscript will be rejected. If you are unable to meet this deadline please let us know as soon as possible, as we may be able to grant a short extension. The revised ms has been re-reviewed by one of the original referees. They found the paper improved, but there are still a few points to address. Most of them are minor, but the most important point is to clarify whether the authors are looking at forest cover from remote sensing data or forest change. The reviewer notes that "it is essential that the authors present a correlation of forest change between a comparable time period (2000 -2010) for their study sites using both sets of satellite image data. Without this it is not possible to assess how consistent the measures of change are". I therefore recommend the revision to address the above point and other comments by the reviewer.

Reviewer(s)' Comments to Author:
Referee: 1 Comments to the Author(s). This is an interesting and timely manuscript that attempts to highlight the need for field collected data in monitoring. As we head into the next CBD cycle, we need informative studies such as this to help identify relevant indicators that link to biodiversity. This manuscript describes an indicator that could be complimentary to the satellite image derived assessments of forest change. The manuscript does ned some areas of clarification in addition to the specific comments below: The authors need to be clear whether they are looking at forest cover from remote sensing data or forest change. The compare forest change with index change, but in the text refer to both forest cover and cover change. Consequently, it is unclear to me what they want the reader to focus on. The matter is confused more by their presenting a correlation between forest cover from Song et al and Hanse for 2010, but not forest change. I see this was requested in the previous set of reviews too (1.4) but was not implemented. It is essential that they present a correlation of forest change between a comparable time period (2000 -2010) for their study sites using both sets of satellite image data. Without this it is not possible to assess how consistent the measures of change are. Related to this, I disagree with sentence lines 103 to 105 which is a fundamental point of the paper language but not content -you look at change in indices. Forest cover will not correlate with FSI for obvious reasons (100 ha boreal forest will have very different FSI than 100 ha humid tropical forest), but instead it might correlate with change. This brings me to another point which is would we expect change in 10 ha of one to result in change of similar magnitude in biodiversity in another. Fitting interaction terms to the models of FSI and satellite data would help examine this effect.
The authors need to present a better explanation of why they chose the correlates that they did (line 262) as currently none given. Otherwise it appears an eclectic pick and mix selection. Was the selection based on previous studies or assessment of level 2 or 3 IUCN threats?
Finally, in their summation, the authors correctly and importantly state the need to have field measurements of biodiversity rather than relying on remote sensing alone. They are not the first to state this, and perhaps some recognition that this is oft stated would balance the conclusions a bit.

Specific comments.
Title -It's a personal thing, but I do not like titles with colons as one part of the title can almost always be deleted. I suggest deleting below the canopy here.
Ln 20 -you look at the correlation with forest change first -change wording to note this. Ln22 -It is not effects of satellite derived tree cover trends, it is a correlation between SPI and tree cover from sat data that you look at. There is a very important difference. Ln 25 is okay. Ln 27 Monitoring alone will not allow forest species to recover. Much more must be done. Change wording here to reflect fact that monitoring allows us to track recovery and implement adaptive management or wording along those lines Ln35 -Not really a deadline, rather review date. Ln 35 -delete legally established unless in FAO wording -it would be interesting to see an illegally established protected area Ln 86 -use °C Ln 90 -delete 'Recent' as it is a time limited term in this case Ln 102 -derived from rather than using as you use correlation to assess the relationship Ln 106 -109. Point 1 is not a hypothesis as currently worded. Ln 119 -Data are, not data is. Use if 'is' jars when reading. Also ln 179. Ln 180 -six, not 6 Ln 184 -what is a biologically useful resolution? Is it a resolution that is appropriate to the size of your study sites from which the population data came? That is probably a more important reason.
Ln 218 -222 -this section appears to be out of place and should be in the additional correlates section. LN 253 -255 -this is a massive reduction in the number of populations used. You drop specialists by 200 for Hansen based on my calculation. Is there a reason for this as to me this drop indicates a serious issue with these data? I think you need to add more of an explanation as to why you go from 268 (ln 137) to 74. It might be I have missed something in the text.
Ln 257 -260 -This is a measure of correlation in area, but you are looking at change in the analysis. Would it be possible to look at the correlation between change 2000 -2010 in both datasets as a more relevant measure of agreement?
Ln 311 -delete 'wished' and reword, but otherwise good to see the inclusion of this analysis here.
Ln 342 -Are these SE or 95% Cis around parameter estimates? The non-overlapping nature of the bars in Figure 2 is interesting and worth mentioning. It isn't mentioned anywhere in the manuscript, but is the massive decline in herptiles due to Chytrid?

Do you have any concerns about statistical analyses in this paper? If so, please specify them explicitly in your report. No
It is a condition of publication that authors make their supporting data, code and materials available -either as supplementary material or hosted in an external repository. Please rate, if applicable, the supporting data on the following criteria.

Do you have any ethical concerns with this paper? No
Comments to the Author Thank you for giving considerations to my comments and providing clear feedback on how they were addressed.

01-May-2020
Dear Dr Hill I am pleased to inform you that your manuscript entitled "Below the canopy: global trends in forest vertebrate populations and their drivers" has been accepted for publication in Proceedings B.
You can expect to receive a proof of your article from our Production office in due course, please check your spam filter if you do not receive it. PLEASE NOTE: you will be given the exact page length of your paper which may be different from the estimation from Editorial and you may be asked to reduce your paper if it goes over the 10 page limit.
If you are likely to be away from e-mail contact please let us know. Due to rapid publication and an extremely tight schedule, if comments are not received, we may publish the paper as it stands.
If you have any queries regarding the production of your final article or the publication date please contact procb_proofs@royalsociety.org Open Access You are invited to opt for Open Access, making your freely available to all as soon as it is ready for publication under a CCBY licence. Our article processing charge for Open Access is £1700. Corresponding authors from member institutions (http://royalsocietypublishing.org/site/librarians/allmembers.xhtml) receive a 25% discount to these charges. For more information please visit http://royalsocietypublishing.org/open-access.
Your article has been estimated as being 10 pages long. Our Production Office will be able to confirm the exact length at proof stage.
Paper charges An e-mail request for payment of any related charges will be sent out after proof stage (within approximately 2-6 weeks). The preferred payment method is by credit card; however, other payment options are available Electronic supplementary material: All supplementary materials accompanying an accepted article will be treated as in their final form. They will be published alongside the paper on the journal website and posted on the online figshare repository. Files on figshare will be made available approximately one week before the accompanying article so that the supplementary material can be attributed a unique DOI.
Thank you for your fine contribution. On behalf of the Editors of the Proceedings B, we look forward to your continued contributions to the Journal.

No. Reviewer Comment
Response 1.1 1 In the introduction the authors note that various international agreements and conventions use forest cover as a proxy for forest biodiversity. I feel this is a simplification as nowhere does it say that, for example, CBD uses forest cover as a measure of forest biodiversity. The reference in Target 5 to 'other habitats' instead indicates it is being used as a measure of forest habitats and not biodiversity. Perhaps some re wording is needed.
This is a valid comment about the exactitude of the language. As suggested, some rewording has been made, and we now state: 'Due to the importance of forests as habitat for many species, forest area is often employed as an indicator in global agreements and processes aimed at slowing and reversing the decline of biodiversity.' (lines 64-66). We also used "biodiversity relevant indicator" rather than "proxy indicator of forest biodiversity" in lines 74-75. The term "proxy indicator" has also been removed from the abstract Also on the statistical analysis why did you consider variables individually rather than in a multi variate model? I can foresee potential interactions between variables that might be good to examine. For example, a decrease in tree cover by 50% might be expected to have a different impact if initial cover was 100% than if initial cover was 50%.
Preliminary analyses on a reduced dataset included multivariate models with interactions (including mean tree cover and tree cover trend), but the results did not support the inclusion of these interactions. We therefore selected to run univariate models on the complete dataset in the first instance, and to build complexity if significant effects were found. After noting we carried out univariate models in the first instance, we have added: "with the aim of fitting multivariate models where evidence of a relationship was found for more than one predictor variable." (lines 296-298) 1.10 1 Figure 1 is the key result here, but I feel that the presentation of geographical / taxanomic break downs in the same format as the LPI graph would be more valuable. This would enable a rapid comparison of The bar graphs showing positive and negative years have been moved to supplementary materials (SM 10). We tried the reviewer's suggestion of examining the trends in, for example, the Nearctic with Africa. Presentation of this trend, and the errors around it, would help in the geographical interpretation of the results disaggregations of the overall index. However, the variable data availability over time for some of the subsets meant that a robust index could not be produced, making comparison difficult. Instead we took the mean and standard error of the mean of species trends for each of the subsets which we thought made a useful comparison. These have been added as a new figure (Figure 2) into the main body of the manuscript.
Text was added to the methods to describe the process for producing this figure (lines 163-164) 1.11 1 Additionally, could the LPI trend for all species be added for comparison?
Yes, this is a good suggestion as we allude to already in the methods. We have added the global terrestrial LPI in Figure 1 as we thought this was the most suitable comparison. The global LPI contains freshwater and marine species so the comparison is perhaps not as appropriate. The text was also amended in line 143 to state that the use of the same method enables the forest index to be compared to the global terrestrial LPI rather than the global LPI. Also we added text in lines 332-335 in the results to describe the terrestrial LPI. 1.12 1 The tables summarizing the models raised some questions for me. First is why the authors use AIC and statistical significance. This is odd as these are two philosophies for assessing variable importance. I suggest they pick one or the other and not both. This is particularly important when the results conflict -In table 1 there is apparently a significant relationship between change and body mass (P = 0.002), but the associated AIC value indicates that AIC increases, suggesting the addition of the variable did not improve the model. This needs to be explained.
P-values and AIC were both included because there isn't universal agreement about which is best. However, we agree this can be confusing when results conflict. Therefore we have opted to report only on AIC and all references to Pvalues in tables and text have been removed.

1
In Table 2 and subsequent tables, how was delta AIC generated for the null model? Was this in fact not the null model but the model with the random effect for location added? In which case please give the original AIC too for the null model. Please give the direction of the change in AIC too to improve the clarity of tables. Finally, the parameter estimates are often smallthere is little value in giving a value of 0.00. Could you rescale the variables so that you can quote informative parameter estimates.
Re. Tables 1 and 2: Previously delta AIC was based on the AIC of the model with the lowest AIC, not (necessarily) the null model. This has been changed so that delta AIC is now based on the difference in AIC between each model and the null model. The parameter estimates have now been presented in scientific notation to make the values more meaningful. 1.14 1 Line 341 -I am not sure these are drivers as much as correlates. Please change wording.
Amended to "correlates" (line 167) 1.15 1 On the cross correlation / literature review part of the analysis I feel that more information is needed to give the results more context. First, while I see the logic behind the assessment of positive and negative correlations with cover change (ie the use of 2*√n), I feel that there are more robust methods that could be used that make better use of the data (see above in comments on methods).
Cross correlation analyses have been removed. This is because we do not have qualitative information extracted for the forest specialist populations that showed no correlation with tree cover and therefore could not address comment 1.16. The correlations were carried out in order to identify the forest populations that were most of interest for the literature review.
As the literature review has been removed, there is no need to include the correlation analyses in the paper. 1.16 1 I also think that the numerical comparisons presented in this As we do not have qualitative information for the species with section of the results could be improved upon, even with % bar graphs rather than text describing results. Importantly, how many of the species for which no relationship (positive or negative) were threatened by the pressures listed? You only present information for the 71 with relationships, but we need to know how many with no relationships were also threatened to put the results in context.
no correlation with tree cover, all reference to the lit review and correlation analyses have been removed from the paper.

1
The discussion is generally well framed, but consideration of some of the points raised above, especially the threats to the species for which no correlation with loss was detected, as noted above, might change some of the content of this discussion.
See comments above.
1.18 1 In presenting the forest LPI I was hoping there would be a comparison with the overall LPI, whether visually, or just in text. Ideally the LPI would be added to figure 1 so readers can make the comparison. Even better would be a statistical comparison of the LPI and forest LPI. This prospect is raised on line 140.
Yes, this is a good suggestion as we allude to already in the methods. We have added the global terrestrial LPI in Figure 1 as we thought this was the most suitable comparison. The global LPI contains freshwater and marine species so the comparison is perhaps not as appropriate. The text was also amended in line 143 to state that the use of the same method enables the forest index to be compared to the global terrestrial LPI rather than the global LPI. Also we added text in lines 332-335 in the results to describe the terrestrial LPI.
We didn't feel that a statistical comparison was appropriate given that the forest specialist species are a subset of the terrestrial LPI 1.19 1 The contention that satellite data alone are insufficient to monitor biodiversity is not new, but is worth stressing as done here to ensure that conservation monitoring is complete. One question that does come to mind here though is what proportion of the world's forests / forest species are covered by the We note there is great value in remote-sensing derived datasets (e.g. "While remote sensing data has undoubtedly improved our ability to independently monitor and assess changes in forest cover..") and we state that these datasets should be LPI compared to the potential of remote sensing for tracking / highlighting concerns across the globe? Until resources are available for more complete field based monitoring is it not better that we have some metric of the global situation? The need for better monitoring, as mentioned in lines 435 -436 is to me a key point of the paper and should be repeated in the closing paragraph.
complemented by on the ground species-level monitoring. Therefore we are not saying they should not be used. Rather we argue that relying on them alone "risks masking below-canopy processes, such as defaunation, with grave consequences not only for forest biodiversity but also longterm forest health and carbon storage", which supports our point about the need for better monitoring (which we repeat in the closing sentence of the discussion: "Therefore, we must not lose sight of the crucial role that site-level species monitoring plays in understanding trends and drivers of forest biodiversity change.") Agreed a justification should be included. Added in: "Range size estimates were not known for every species, but body mass estimates were available (see Additional drivers of forest population change below). We calculated the correlation between body mass and range size (both log-10 transformed) for the species with both estimates available, and found a strong relationship (Pearson correlation coefficient = 0.87, p < 0.001). We therefore used body mass to predict range size for all populations using the 'predict.lm' function in the 'stats' package (R Core Team, 2018) and calculated the mean range size across all populations as 58.5 km 2 , equivalent to a circle with a radius of 4.32 km. We rounded this up and fitted buffers with a 5-km radius around the central coordinates of each forest population." (lines 218-225) 2.12 2 L214-218this text is a bit unclear and might benefit from some rewording.
amended for clarity to: "By plotting annual tree cover values from the Song dataset against year, we were able to visually assess the change in tree cover …" (lines 231-232) 2.13 2 L223 -'≥2 years of data covering a ≥5 year period'… this is unclear, please reword/clarify.
Amended to "In both cases we removed populations that no longer had ≥2 years of data spread over at least a 5 year period" (lines 240-241) 2.14 2 L246-247please add supporting references for this or clarify if it's a hypothesis-driven sentence.
Amended to clarify: "We then examined the influential genus to determine the cause and, if the genus was known to be responding to a driver other than those included in our analyses (e.g. disease, poisoning), we repeated our analyses with the genus omitted." (lines 318-320) 2.16 2 L294this subtitle seems to refer to text in L314-321 only; I suggest moving further down and renaming the section in L296-312.
The lit review and crosscorrelation analyses have been removed.
2.17 2 L310-311a citation is needed for this method and why it's appropriate in this case. It is also important to distinguish between relationship between forest cover and biodiversity and changes in forest cover and biodiversity. Both were tested in this study, but there is no reason why both should show the same trends, e.g. if there is a threshold of % forest cover above which changes in forest cover are not significantly affecting biodiversity, then different outcome might be expected (see comment by reviewer 1 about the importance of exploring the interactions between predictors instead of considering predictors individually).
See response to comment 1.9

Associate Editor
Another problem picked up by the reviewer 1 is that for the second part of their analyses, the authors have focused only on 71 populations for which significant relationship with forest cover was found. This resulted in exclusion of 104 populations with no correlation with forest cover from the analysis. The reviewer pointed out that it would be also very important to know the characteristics of species for which no relationship with forest cover has been detected and how many of these species are also threatened by the various pressures examined.
Cross correlation analyses//lit review have been removed.

Associate Editor
In addition, this reviewer has questioned the methodology used for categorizing cross-correlations into positive, negative or uncorrelated with forest cover, which is very much akin to vote-counting. I share this concern. The 2*sqrt n approach needs to be better justified and ideally a more robust methodology used.
Cross correlation analyses//lit review have been removed.

Response to Referees
Number Comment Response (line numbers refer to clean manuscript) 1 The authors need to be clear whether they are looking at forest cover from remote sensing data or forest change.
The compare forest change with index change, but in the text refer to both forest cover and cover change.
Consequently, it is unclear to me what they want the reader to focus on. The matter is confused more by their presenting a correlation between forest cover from Song et al and Hanse for 2010, but not forest change. I see this was requested in the previous set of reviews too (1.4) but was not implemented. It is essential that they present a correlation of forest change between a comparable time period (2000 -2010) for their study sites using both sets of satellite image data. Without this it is not possible to assess how consistent the measures of change are.
We have changed the correlation analysis to look at tree cover change 2000-2010 as requested (lines 243-249): "In order to examine the agreement between the two tree cover datasets, we calculated tree cover change per population from 2000 to 2010 using values derived from the Song dataset and from the Hansen dataset. We then assessed the correlation between the two sets of tree cover change values for the 685 populations included in the Hansen analyses. The correlation between the two datasets was highly significant but had a low correlation coefficient (Pearson correlation coefficient = 0.171; p < 0.001). This is in agreement with other studies that have found discrepancies between tree cover datasets when assessing tree cover change or area (Gross et al., 2017;Sexton et al., 2016)." We have added to the Discussion (lines 418-420): "Further, discrepancies between satellite-derived tree cover datasets in estimates of tree cover change or area indicate the uncertainties associated with tree cover assessments (Sexton et al., 2016;Gross et al., 2017)." Related to this, I disagree with sentence lines 103 to 105 which is a fundamental point of the paper language but not contentyou look at change in indices. Forest cover will not correlate with FSI for obvious reasons (100 ha boreal forest will have very different FSI than 100 ha humid tropical forest), but instead it might correlate with change. This brings me to another point which is would we expect change in 10 ha of one to result in change of similar magnitude in biodiversity in another. Fitting interaction terms to the models of FSI and satellite data would help examine this effect.
We agree with the reviewer that we need to make the language clearer here and that it's not the magnitude of the forest cover that we were explicitly testing, but how much it has changed. This allows us to compare large areas of forest that may support large populations, alongside small areas that may support small populations. Changes in these two may be similar, while reflecting large changes in actual areas (or animals). The FSI is a measure of change rather than an absolute value of abundance so we have made this distinction clearer.
We changed the sentence in Ln 106 to refer to tree cover change.
We also made this reference to change clear in the Discussion (Ln 418 and 453) On the second point, we would not necessarily expect a linear relationship between habitat change and biodiversity change. This has been raised as a caveat for using remote sensing data for predicting changes in species population trends.
Our approach of modelling abundance change in response to change in forest cover was to capture that while large areas may support larger populations, and smaller areas support smallerproportional changes should be similar. Particular areas may have differential responses (which would in part be captured by the 'location' random effect), but any consistent relationship between changes in cover and changes in abundance should be captured.
The authors need to present a better explanation of why they chose the correlates that they did (line 262) as currently none given. Otherwise it appears an eclectic pick and mix selection. Was the selection based on previous studies or assessment of level 2 or 3 IUCN threats?
We have now specified that correlates were selected a) if they were encoded in the LPD, or b) through literature review (lines 255-256).

4
In their summation, the authors correctly and importantly state the need to have field measurements of biodiversity rather than relying on remote sensing alone. They are not the first to state this, and perhaps some recognition that this is oft stated would balance the conclusions a bit.
We agree, and have added "as has been advocated elsewhere" to this statement, with three supporting references (line 449).

5
Title -It's a personal thing, but I do not like titles with colons as one part of the title can almost always be deleted. I suggest deleting below the canopy here.
We have discussed and would prefer to keep the title as it is, as the request was based on the reviewer's personal preference.
6 Ln 20you look at the correlation with forest change firstchange wording to note this.
Changed "drivers" to "correlates" (line 20) 7 Ln22 -It is not effects of satellite derived tree cover trends, it is a correlation between SPI and tree cover from sat data that you look at. There is a very important difference. Ln 25 is okay.
Reworded to "We analysed the relationships between the average rate of change of forest vertebrate populations and satellite-derived tree cover trends, as well as other pressures." (lines 21-22) Ln 27 Monitoring alone will not allow forest species to recover. Much more must be done. Change wording here to reflect fact that monitoring allows us to track recovery and implement adaptive management or wording along those lines This has been changed to "For forest biodiversity to recover, conservation management needs to be informed by monitoring all threats to vertebrates, including those below the canopy." (lines 26-28) 9 Ln35 -Not really a deadline, rather review date.
We have replaced '2020 deadline for the ABTs' with '2020 expiration of the ABTs'. Ln 102derived from rather than using as you use correlation to assess the relationship Changed "using" to "derived from" (line 101)

14
Ln 106 -109. Point 1 is not a hypothesis as currently worded Deleted "given the decline in area of natural forest over time (Keenan et al. 2015) and the link between habitat loss and biodiversity loss (Brooks et al. 2002)" from the first hypothesis.
We moved this sentence to the preceding paragraph to explain our hypothesis that forest vertebrates were declining (as an explanation behind this hypothesis was requested in earlier reviewer comments) (lines 98-100).

15
Ln 119 -Data are, not data is. Use if 'is' jars when reading. Also ln 179.
"Is" changed to "are" (line 114), and "was" changed to "were" (twice) (line 171-172) Ln 311 -delete 'wished' and reword, but otherwise good to see the inclusion of this analysis here.

22
Ln 342 -Are these SE or 95% Cis around parameter estimates? The nonoverlapping nature of the bars in Figure  2 is interesting and worth mentioning. It isn't mentioned anywhere in the manuscript, but is the massive decline in herptiles due to Chytrid?
These are SEs shown on Figure 2. We have added the missing information to the figure legend. On reviewing this figure we also noticed that part of the y axis label was incorrect and have also amended that.
The lack of overlap between error bars in both the taxonomic plot and the one for biomes has been added to the results (Ln 322 and 329) and in the discussion (Ln 365 onwards).
We were not able to quantify the effects of different threats on forest populations but we suspect that Chytrid could be one of the main drivers in the amphibian declines, given the threat information we have available. We have added a sentence to the discussion on this to highlight the likely importance of this threat (Ln 368-370) 23 Figure 3 given the notable differences in trends in figure 2 b and c, I think it is useful to use different symbols for either realm or biome in this figure (as long as it is not a mess) as I am curious to see what this looks like.
We have amended Fig 3 using different symbols to reflect biome (tropical or temperate). We chose the biome option over realm because realm was too messy. We would be interested to know if the reviewer thinks this is now messy and can provide the previous figure without the symbols if preferable.

24
Ln 387on average halved or just halved?
The index value is the average of the species trends, so we have kept the reference to 'on average' in line 358 but moved it to further along in the sentence to make it clearer: '..the global abundance of forest specialists more Our results do support some of the caveats noted when using remotesensing data for extinction risk assessments, for example the assumption of a linear relationship between habitat change and population trends (Tracewski et al 2016) and the influence of threats other than habitat loss (Santini et al 2019).