Reliable NPP-age relationships are critical for carbon flux simulations and forest management. In this study, 10 NPP-age curves for different regions and forest types in China were derived based on field and satellite data. The authors also compared the performance of five different math models in deriving these relationships. The results were clearly described. The authors also compared their results to existing models in the same region and differences were interpreted. In my opinion, this study is unique in making use of the large amount of field data and satellite LAI time-series, and credible and up-to-date results provided.  I found no major problems in this study. A list of minor comments is provided below for the authors’ reference.

It is my interest to see a discussion on how the CO2 fertilization could have affected in the collected datasets (i. e., biomass inventory and LAI), and correspondingly, how the trend changes in these datasets could be propagated into these derived curves. This is important because of the potentially uneven fertilization effects in different periods of time-series, compared to a scenario in the pre-industrial era. However, this might already be out of the scope of this study since a focus of this study is to compare different math models.

Specific comments:

L27: is not it GPP the largest flux (component)?

L29: a gradual decline is not always seen, especially for some mixed forests.

Table 1: turnover rate for EBF (evergreen) is “one”, is it true?

L99: LAI data in 1981-2022 were used – did you use the LAI in a specific year to calculate the corresponding Ll (age) in the NPP-age curve? LAI(age) may not be available for the earlier stage of old forest (42 yr+), did you use spatial surrogate? This could be clear in the revision.

L191: “consistent” - not sure for mixed forests

L208: “the rectangle in each line…” – no rectangles are seen.

Figure 4, panel 8 (DNF …) – for Delta_Bc (biomass increment): any interpretation for the “increase” pattern in 120-180 yrs? Slightly increasing pattern is also observed for panel 1 (EBF).

Figure 4 again: increase, stable, and decrease patterns of LAI (therefore for leave and fine-root biomass) are seen. It would be interesting to see the interpretation of these patterns (trends).

Figure 5: this figure can be replaced by a Table, with max/min numbers bolded, but this is up to the authors. Unit for RMSE needs to be added.

Table 2: either coefficient “a” has a unit, or the unit of derived total NPP needs to be indicated.

Table 3: for the “Source – This study”, are these peak-ages derived from NPP-age curves, or from measured data (Fig. 3)? It will be useful to show/interpret any differences (e.g. 39 yrs vs 50 ys for panel 10 – MF in Fig. 3?).

This ms. tested different methodologies to fit age-NPP curves over Chinese forests. I think the topic is important and is worth to investigate. Overall, this ms. provides a clear route from reasoning, model testing and analysis, and corresponding comparison with previous studies and explanations of their results. However, I feel substantial improvements are needed before it can be accepted. First of all, the underlying mechanism of age-NPP relationship should be clearly explained and the results should be further explored and explained. It is important to present to the readers the theoretical architechture of the age-NPP relationships and the potential factors that affect these relationships at the very beginning of the ms.. Then the interpretation of the results can be very well linked to the theories or possible mechanisms.  Secondly, I’m concerning about the curve building as the data is highly accumulated at the young age end so it is questionable that whether the curves are still effective for those older forests, especially when the authors used the extension of the models to old forests to select the best model. This question has existed for long in the discussion of age effect on forest C sequestration, and with the existing dataset, it is hard to say which pattern is good or not. I would suggest the authors to cut the curves to an age range that can be reasonable covered by the data. Alternatively, one needs to provide additional evidence to indicate the decline or flat pattern of NPP along the age gradient is true for old forest. Finally, there are places of presentations and organizations that require essential refinements (see specific comments).

Specific comments:

Line 20-21: I’m not convinced about the statement on the point of as little samples from the old forests were used in this study and it is highly unknown that to what point NPP stops declining, or whether they can maintein constant over an age of hundreds of years.

Line 22: The final statement is vague to me. What is “forest carbon modelling” and “future carbon projections”.

Line 25-27: The statements here are vague again – how you define the “biomass carbon gain from atmosphere” and “the largest component of the terrestrial carbon cycle”. It seems applicable for some other terms also but is not totally right as well.

Line 30: Again what is “forest carbon modelling”? I would suggest to go through the ms. and revise all those vague terms used.

Line 31: Unique -> varied.

Line 31-46: I would introduce the mechanisms driving the relationship, rather than the data gaps here. Simply pointing out the data gaps in China to me is not a strong motivation of this study. I mean, understanding the patterns is much more important and interesting than fitting models and say which one is the best, in the context of science.

Line 64-67: reads like methodology rather than an introduction of the aims. Also I feel the aim of this study is to understand the underlying patterns, rather than comparing things..  

Line 70: rephrase the sentence to something like: Chinese forests consist of five main functional types...

Line 72: okay but it’s weird to mention two classification system together with little logical linkage..

Line 70 – 84: I feel the entire paragraph should be reorganized in a better form. As I said above, there should be necessary logical linkages between sentences.

Line 184-191: Not a result literally – seems a combination of methodology and figure head..

Line 192 – 195: Method again.

Line 207-209: Seems like something in figure head again.

Line 215: The reasoning here is very tricky to me – why you think “decreased more than 50 % in old ages” is a useful gauge to identify whether it is reasonable or not. I would expect that at least there are some numbers of corresponding percentages that can be extracted from the previous studies cited here (so you can say okay NPP dropping too fast with age might be unrealistic). But again, as I said earlier, the extension of the fittings to an age of hundreds itself is not that effective becuz little evidence is there.

For the entire results section, I suggest the authors to reorgnize it in a way of “telling story”, rather than “presenting a stack of stats”.

Fig 3 and Fig 4: I guess there are some bin processings but they are not explicitly introduced in the methods.

Line 261: I would strongly encourage the authors to expand their discussion from an asepct of the underlying mechanisms and how such mechanisms may cause different/similar patterns of age-NPP curves between PFTs, or how the mechanisms may determine the success or failure of the models. As I pointed out earlier, this study should become way more interesting than the current form, if it tells me, for example, why some PFTs reach NPP peaks ealier than the others and why NPP of some PFTs declines faster than the others. Considering the well fitness of the selected models against the massive data, it is really worth to do so. This interpretation would then lead to a more meaningful comparison between the curves from this study and the ones previously built (and potentially facilitate the illustration of the advantage and uniqueness of this study).