Large carbon sink potential of secondary forests in the Brazilian Amazon to mitigate climate change

Tropical secondary forests sequester carbon up to 20 times faster than old-growth forests. This rate does not capture spatial regrowth patterns due to environmental and disturbance drivers. Here we quantify the influence of such drivers on the rate and spatial patterns of regrowth in the Brazilian Amazon using satellite data. Carbon sequestration rates of young secondary forests (<20 years) in the west are ~60% higher (3.0 ± 1.0 Mg C ha−1 yr−1) compared to those in the east (1.3 ± 0.3 Mg C ha−1 yr−1). Disturbances reduce regrowth rates by 8–55%. The 2017 secondary forest carbon stock, of 294 Tg C, could be 8% higher by avoiding fires and repeated deforestation. Maintaining the 2017 secondary forest area has the potential to accumulate ~19.0 Tg C yr−1 until 2030, contributing ~5.5% to Brazil’s 2030 net emissions reduction target. Implementing legal mechanisms to protect and expand secondary forests whilst supporting old-growth conservation is, therefore, key to realising their potential as a nature-based climate solution.

Responses to Reviewers of the manuscript entitled: "Large carbon sink potential of Secondary Forests in Brazilian Amazon to mitigate climate change" By Heinrich V.H.A. and co-authors First, I hope the Reviewers and their families are well and healthy. We thank both Reviewers for their comprehensive review and supportive comments. We have addressed them in full, which has helped to improve our study. Below, we repeat all Reviewers' comments and reply to the comments one by one. Each comment is numbered, with our responses in bold. The Reviewers were numbered 2 and 3, so we have kept these names to limit confusion, notethere was no response from Reviewer 1 with any changes for us to make. Kind regards, Viola Heinrich and co-authors (17/12/2020)

Reviewer #2
1 -The manuscript entitled "Large carbon sink potential of Amazonian Secondary Forests to mitigate climate change" present a relevant study that aimed to quantify the contribution of the Secondary Forest to carbon sequestration and the influence of environmental and disturbance drivers on the rate and spatial patterns of regrowth in the Brazilian Amazon. The study addresses a very important topic for spatially monitor carbon dynamics in the tropical forest, and with the method used they provided information disaggregated by regions, with applicability in different sectors of society, such as science, economics and sustainable development.
The paper is very well written and uses current literature. As I am not a native English speaker, I cannot evaluate English.
The statistical approach seems adequate and sound and the methods were described in a clear and detailed way.
Response: We thank Reviewer #2 for their positive assessment of the paper. We are glad the inter-disciplinary applicability of the paper comes across and that the methods were described and carried out in an adequate manner.
2 -The article will be of great interest to the audience of this magazine, but I think it needs to point out more explicitly what are the novel in it, since part of the findings shown in it have already been presented by Poorter et al., 2016(doi: 10.1038/ nature16512) and Patton et al. 2020(doi.org/10.1038/s41586-020-2686.
Response: We agree that the novel nature of the paper can be made even more explicit. We have changed the last paragraph in the introduction to highlight the new aspects our work addresses relative to previous work, including very recent literature such as Patton et al., 2020, the paragraph reads:

"The primary aim of this study is to provide key advances in understanding the spatial variation of secondary forest regrowth in the Brazilian Amazon, a large and geographically complex region.
Previous studies have already provided the first steps to understanding regrowth on a biome scale, influenced by some driving variables 11,14,22 . Here we introduce additional environmental and anthropogenic disturbance drivers that affect regrowth, including local-scale drivers, and for the first time, disaggregate their effects using a spatially explicit approach 11,14,22 ." Throughout the results, discussion, and conclusion we have also imbedded further sentences highlighting the novel nature of the work. For example, the beginning of the conclusion now reads: "The models developed in our study provide a new assessment of the carbon sink potential of secondary forests in the Brazilian Amazon, considering age, persistence, local and regional drivers. This type of approach using regional and global remote sensing products has not been attempted before to such a high spatial resolution. The models have the potential to benefit both the carbon modelling and carbon-policy communities to help understand the regional variations of regrowth under different drivers. The carbon modelling community will benefit from the ability to spatially monitor carbon dynamics, which can be incorporated into models and scenarios of land cover and climate change. Our models provide a new, detailed quantification of the naturally regrowing secondary forest carbon sinks." 3 -I suggest incorporating in the discussion information about how frationation can affect regrowth in FS and whether the change in species composition, already mentioned in the literature, driven by climate change will affect this regrowth.
The manuscript represents a solid piece of work that should be published after some minor corrections.
Response: We assumed the term fractionation was a typo and was referring to "fragmentation" of old-growth forest as well as secondary forest. If this was not the case, we apologize and would ask for additional clarification. We have linked this with responses 4 and 6 from reviewer #3, which we believe was related to a similar issue of proximity to old-growth forests and species composition, respectively.
We hope that the following sentences in the text addresses these comments: Responding to species composition and climate:  Figures 1-3) Responding to fragmentation impact on regrowth:

Reviewer #3
1 -I really enjoyed reading this manuscript, which takes the next step to project the impacts of secondary forest regrowth and subsequent disturbances on rates of carbon storage in the Brazilian Amazon The key advance made in this study was examining the spatial variation in trajectories and drivers of biomass accumulation across this enormous and geographically complex region. The analyses revealed strong differences in these trajectories and also strong effects of factors that have not previously been modelled, such as short-wave radiation and fire disturbances. This study provides a state-of-the-art assessment of the carbon sink potential of regrowth forests, taking into account forest age, persistence, and local drivers of forest regrowth. It will have a major impact on how we assess this important nature-based solution for climate change mitigation.
Response: We thank the Reviewer for their very positive overall assessment of the paper and their strong words of support highlighting the importance of the research. We were delighted and encouraged to receive the comments and have responded to the specific comments below.
Specific comments: 2 -Title: I suggest referring to Brazilian Amazon in the title. THe study does not cover the entire Amazon basin.
Response: Thank you for this suggestion. We have amended the title to read:

"Large carbon sink potential of Secondary Forests in the Brazilian Amazon to mitigate climate change"
It is still within the word limit of Nature Communications for titles ( Response: Thank you. We have now cited it in the introduction to strengthen our argument that our study addresses the impact of local factors on regrowth, one of the issues mentioned in the conclusion of the Cook-Patton et al. study that was still outstanding following their study:

"The primary aim of this study is to provide key advances in understanding the spatial variation of secondary forest regrowth in the Brazilian Amazon, a large and geographically complex region. Previous studies have already provided the first steps to understanding regrowth on a biome scale, influenced by some driving variables 11,14,22 . Here we introduce additional environmental and anthropogenic disturbance drivers that affect regrowth, including local-scale drivers, and for the first time, disaggregate their effects using a spatially explicit approach 11,14,22 ."
4 -Line 77: There are also key differences in species composition and wood density between western and eastern Amazonia that influence biomass and carbon storage potential. Species composition could be a "hidden" variable that is also correlated with climate variables and affects carbon storage during forest regrowth: Response: Again, thank you pointing us to this additional reference and point to consider. We have included an additional paragraph in the discussion as well as embedded sentences throughout that addresses this point (see below). However, we expect to see the impact of species competition and wood density to emerge more over Pan-Amazonia, i.e. beyond the Brazilian Amazon, as this accounts for the largerleaved, lower wood density trees in the Peruvian, Colombian, Ecuador Amazon (i.e further West towards the Andes). Expanding beyond the Brazilian Amazon is beyond the scope of the current study, but we take on board this valuable comment for future studies using this approach for the whole Amazon region, and present the amended paragraph in the discussion: Fig. 9b), which reduces the intensity and occurrence of the burning itself (Figure 3a).  The bar graph (g) shows the average importance ranking of the drivers (af), as well as Forest age, in influencing Aboveground carbon (AGC) accumulation. The average ranking was calculated following 30 iterations of a conditional random forest model. The importance has been ranked from least important (1) to most important (7) and the vertical dotted line separates environmental drivers (left) from anthropogenic disturbance drivers (right). Shading in (a-f) denotes the 95% confidence interval of the models, based on the median value of the initial data for each agedots in figures. The error bars in (g) denote the 95% confidence interval.  Table 9 for quantitative interpretations of the regions. The average importance ranking for each of the six variables, as well as Forest age, is shown for (b) the North-West region, (c) the North-East and Central-North region, (d) the South-West and Central region, and (e) the South-East and North region. The average ranking was calculated following 30 iterations of a conditional random forest model. The importance has been ranked from least important (1) to most important (7) and the vertical dotted line separates environmental drivers (left) from anthropogenic disturbance drivers (right). The error bars in (be) denote the 95% confidence interval.