RSPB-2022-2203
This manuscript evaluates the importance of landscape-level forest cover in driving secondary forest recovery patterns. The study evaluates variables at various scales of effect and it is noteworthy that results indicate the importance of assessing landscape variables related to recovery at the 4 km radius range. The paper is very well written and clear and represents a nice contribution to the literature. As such I would like to commend authors for taking the time to submit a polished paper for review as that is not always the case!

I do have a few minor comments and edits that are fairly quick to remedy and one larger comment that speaks to the low representation of sites in the lower forest cover classes (i.e., <60%). I include intermediate category here because by my estimation there were only 4 sites in that category. This is unfortunate as it limits the interpretation of results (as authors duly note) and I do think that adding a few more sites to this study at the low and intermediate range would be beneficial. But I recognize that that is a big ask so this is only a suggestion/observation …and one that authors have already noted in the Dicussion. More details below.

ABSTRACT
• line 112 …the variation of key ecological factors…
• line 114 drop 2nd ‘the’

INTRODUCTION
• line 129 Perhaps ‘The importance of their role in biodiversity conservation …
• line 131 ‘…interest in determining the factors affecting forest… ( or ‘understanding the factors’)
• line 138 I would ditch this sentence. Not entirely true that we know very little … but there isn’t a need to put in this justification. More landscape level assessment is needed. So I would delete this sentence and end paragraph with previous line.
• line 157 drop 2nd ‘the’
• line 159 drop 1st ‘the’
• line 161 ….increase the number of species present at the local and landscape…
• line 164 ….uncertainty about other landscape-level effects limits … [what other effects… give a couple of examples in parentheses]
• line 168 you have ‘needed’ twice in same sentence
• line 174 …. In particular, the model predicts …
• line 187 ….biomass, resulting in successional trajectories that are less predictable than those in landscapes characterized….
• line 222 perhaps ‘factors’ instead of ‘changes’?

METHODS
• line 345 drop ‘representative’ instead of ‘well-representative’
• line 353 …as a categorical variable… presumably it should be ≥10 cm for the third category?

RESULTS
• line 369 …with more than 70% of the deviance explained by the models….
• Figure 3. It is unfortunate that part of the poor predictability for the lower forest cover area is driven by the lack of plots that fit this criteria. There are also few points in the intermediate category. I am pretty sure there are passive restoration projects with low surrounding forest cover that could be added to this database to improve that representation (I can think of a few) but I recognize that that would entail a considerable amount of additional work ….so it is only a suggestion/point to consider.

DISCUSSION
• Paragraph line 418. I would also add that part of the uncertainty is due to few samples in this category range as well (I only see 3-4)
• line 437. This is an interesting result!
…stayed open throughout …
• line 470 drop 2nd ‘the’
• line 484 I disagree with this sentence. There is quite an extensive literature base now on recovery in heavily degraded habitats. See Brancalion publications for Brazil; Cole, Holl, and Zahawi publications in Costa Rica; and Martinez-Garza, Caughlin, Howe, and de la Peña-Domene for Mexico, among many others. Please rephrase. I also think that some of these long-term studies could fill the gap on your <60% cover category at the landscape level if you want to consider expanding study and contacting additional authors.