General Comments:

The manuscript applied the existing individual-based forest gap model (FORMIND), which was developed using data from an old-growth temperate forest in the northeastern China. Authors present a novel methodology to distinguish between mature and immature trees to understand forest productivity. This approach offers a fresh perspective compared to traditional methods focusing mainly on gross primary production (GPP). The manuscript is well-organized, systematically presenting its research approach, results, and conclusions.

Manuscript is interesting and useful to international audience of the journal. However, there is room for improvement in the manuscript. The approach and conclusions are somewhat limited by the methodological framework and the absence of a comprehensive analysis of the ecological implications. Authors are suggested to address the following issues while making the revision.

• The introduction provides adequate background but lacks a critical review of previous research methodologies (other process-based models) and their limitations. Also, there is a need to introduce the limitation of the current FORMIND model. Suggest to enhance the introduction.
• Different species might exhibit significant variations in growth and carbon dynamics, even within the same maturity classification. The selection and classification of trees into mature and immature categories are not sufficiently justified. More rigorous criteria and a discussion of potential biases in these classifications are needed.
• The choice of the FORMIND model may not fully capture the complexity of forest dynamics, especially in terms of species-specific interactions and responses to environmental variables. For instance, you did not apply any space competition in the model. Suggest to compare results with those obtained from alternative models, particularly those incorporating more detailed species-specific parameters or interactions with abiotic factors.
• The acknowledgment of the model's limitations is a positive aspect, but the discussion lacks a critical assessment of how these limitations might have influenced the study's conclusions. Suggestions for alternative modeling approaches or supplementary methods to address these limitations would provide a more balanced view.
• A lot of supplement information is provided with the manuscript. I’m not sure if it can refer to other fundamental literature previously published. Are the allometric relationships part of the FORMIND model? It would be better to keep concise and easier for the readers to understand.

Detailed comments:

1. Introduction: Consider providing a brief introduction on any challenges or limitations encountered while adapting the FORMIND model to this specific old-growth temperate forest.
2. Introduction: P2 second paragraph: "Nonetheless, it has proven difficult to identify clear relationships between forest structure and NPP (Chisholm et al., 2013) as several factors interact...": Suggest elaborating on the specific factors that complicate the relationship between forest structure and NPP.
3. Method: Page 4 2.1 Field data: I have concerns about the allometry information and biomass equations provided in Supplementary A. (1) A lot of species lack allometry data and biomass equations. How did you address these species in your study? Did you use likelihood-based analysis similar to the PFTs classification? Please clarify this in the methods part. (2) The biomass equations, adopted from Chojnacky et al. (2014), are generalized primarily for North American species. Since most of the equations are empirical models, I doubt their accuracy when used directly.
4. Methods: Page 6 Model fitting, second paragraph, “We fitted these 18 parameters...”: cannot get 18 parameters based on your description, please clarify.
5. Results, first paragraph, “...This contrasts with the basal area of immature trees.” please add supporting figures or statistical results.
6. Results: “...These results are depicted in Figures 4 and 5.” Figures should be accompanied by the corresponding results in brackets. This will make it clearer for readers to correlate the interpretation with the figures.
7. Results: “...These results are shown in Fig. 6.” Similar suggestions as above.
8. Results, Figure 7, does each dot correspond to a forest patch of 0.04ha? Please clarify.
9. Discussion, Page15, second paragraph, “The proportionality can be explained by the strong connection between the individual-level basal area and GPP in conjunction with the negligible NPP of mature trees.” Consider simplifying or re-framing it for better readability.
10. Discussion, Page 15, second paragraph, “On the stand level, however, neither the GPP nor the respiration were correlated with the proportion of immature trees (Fig. 7) ...”, in the discussion section, only this sentence refers to the figures. Please maintain consistency.

Fischer et al. present a novel method for relating the net primary production (NPP) of forests to forest structure. They explore this method using the process based forest model FORMIND parameterised with data from a forest inventory plot located in an old growth temperate forest in Changbaishan, China. The method rests on the assumption that as trees grow their respiratory demands increase, resulting in a decline in NPP and carbon use efficiency with size. As a result, the proportion of immature trees in the forest predicts stand level NPP as only immature trees are putting on biomass. The authors test a number of structural metrics and find that a measure of DBH entropy is the best proxy for the proportion of immature trees and could thus be used to predict the NPP of a forest using only inventory data. 

This topic will be of interest to the forest ecology community and those interested in predicting carbon sequestration using metrics of forest structure. The manuscript has a good structure and is for the most part easy to follow. The figures show the results clearly. I think keeping most of the details of FORMIND in the supplement and only describing necessary details in the main manuscript is a good approach, although I think some more high level summaries in the main text would be helpful so that readers do not have to keep moving to the supplement which is very long.

My main suggestion is to include more discussion of how some of the assumptions in FORMIND influence results. In this manuscript, FORMIND is set up so that trees have a maximum DBH, above which they allocate all GPP to respiration meaning that NPP goes to zero. It is therefore not surprising that the proportion of immature trees predicts NPP.  The authors acknowledge in the discussion that there is still debate in the literature about whether NPP declines with size. It would be good to expand on this and discuss how uncertainty in the changes of individual NPP with size influence the ability of stand structure to predict stand level NPP.  They could also test this assumption using the field data by looking at a time series of growth rates for individual trees and identifying declines in growth with tree size. 

Specific comments: 

It would have been helpful if the manuscript had line numbers. 

On page two the authors say that CUE is expected to decline with size because large trees have higher demand for respiration and non-structural carbon. This doesn’t necessarily follow unless GPP increases with size are less than respiratory and NSC increases. Or GPP asymptotes or decreases. Do we know this to be the case in this forest? 

Bottom of page 4  -  missing word - “The key idea”.

Page 5 - how were the light requirements of the species known? Do those classifications align with growth and survival rates from the inventory data? It looks like in Fig S1 the two species with the fastest DBH increment rates are on almost opposite sites of the shade tolerance spectrum which is a little surprising. Was there any comparison of growth rates predicted by FORMIND and the mean growth rate per PFT from the inventory data? 

Please provide the Genus name of Q. mongolica in the first mention of this species. 

I didn’t understand how growth was modeled without  reading the SI. Consider some high level descriptions in the main text.  

Middle of page  6 - why did the authors use  biomass and stem count as an indicator of the size distribution rather than the actual size-distribution?

There are not many details of wood decomposition or soil respiration parameterisation in the methods section. The authors could consider adding a few sentences explaining this aspect of FORMIND. It wasn’t really clear to me what the hypotheses were for how NEE would change with forest structure. 

Page 9 section 2.5. To  identify trees that have reached growth limits in field data would it not be simpler to look at growth rates and find large trees that have declining growth rates (if there are enough census intervals, or if not, those that are growing less than some quantile of the population)? Or is the  idea that DBH entropy can be used in field studies with a single census? 

Fig 4. Why are there few forest patches with a Shannon diversity of just over 1?

Results in Fig. 4 are not that surprising since the model was set up so that NPP would not be affected by large trees because large trees allocate all GPP to respiration. But there is no test of that in the field in this study. Is that a reasonable assumption? Some evidence suggests that large trees continue to actively accumulate biomass e.g. Stephenson et al. 2014 https://www.nature.com/articles/nature12914. Ah, I see the authors bring this up later in the discussion. If possible the authors should try to address that assumption using the inventory data from Changbaishan. 

Fig 4. NEE is very low in mature patches - is this because of less wood turnover? Is all CWD in FORMIND the result of mortality or is there also a representation of branch turnover? 

Bottom of page 11. Does respiration here include soil respiration? Same comment for Fig. 7c.