In this paper, the authors aim at depicting the early state of tree regeneration after disturbance, quantifying the possible reorganization of stem composition relative to undisturbed reference conditions, and assessing how management and patch size affect forest reorganization post-disturbance. To do so they relied on an elegant methological approach, for both i) the ‘experimental’ desig (with triplets of plots: undisturbed, disturbed without post-disturbance management, and disturbed with post-disturbance management), and ii) the way to quantify regeneration structure and composition.
Overall this study deals with an important issue in current forestry, as global change and related disturbances strongly impact forests, especially in Europe, which raises great concerns for forest managers about the ‘renewal ability’ of their forests. Part of these concerns focus on stand regeneration, and thus tree recruitment, and possible shifts in species composition. The topic of the study is thus timely. Of course such a design calls for future developments, and I hope that at least some of the authors have already planned to revisit these plots in a few years!

I also have a few comments about the paper.
In general, I think this paper may better explain specificities of the studied forests in terms of regeneration. After a disturbance, depending on the size of the impacted plot, one may expect ‘classic’ succession, with a mix of early seral and mid-successional species colonizing the gap, being later replaced by late-successional species. Yet, in this study, the main trend is that advance regeneration plays the prominent role. I think that this finding deserves to be better presented, especially for a reader not familiar with this forest dynamics.

I was wondering whether management history (if not taken into account) could possibly bias the results. In other words, are all plots assigned to a the same target species (‘species groups’ = spruce, beech, pine, oaks) experienced the same management over the last years/decades?
For instance (see l. 51-68), Regarding the hypothesis of forest resilience ‘sensu strictu’, I think it is important to mention whether the Picea abies stands were originally natural or planted.

l. 268 : What were the mortality rates among the ‘species groups’? Were they comparable? If not (ie. with large differences in terms of mortality rate), could this impact the findings? (For instance, are plots experiencing larger mortality rates more prone to experience species replacement?)

l. 328-330 : Management has a stronger effect in spruce stands, but did this happen because the intensity of post-disturbance management was also stronger in spruce stands? (because of larger mortality rates in spruce – see comment above)

Were all plots of the same ‘species group’ at the same stage of development ? If not, could this also affect the outcome of the ranking (restructiring, replacement, resilience)?

l. 451-452: (very minor comment) this value (2500 t/ha) is a common target for adult trees, ie. with effect of thinning and mortality. The ‘reference’ stem density values should take plot age into account. For instance, for beech, the reference value should be the stem density after 2-4 years under natural regeneration. This being said, I agree with the general conclusion in l. 453-455.