Summary
Flows of resources and movement of species across ecosystems can affect the functioning of recipient ecosystems, and these available resources can also affect how they respond to change. Meta-ecosystem research has tended to focus on the effect of subsidy quantity, but ecosystem function is affected by the quality of subsidies as well. When the movement of resources between ecosystems has stoichiometric asymmetry, such as at a terrestrial-aquatic ecotone, it will likely affect the productivity of those systems as nutrient limitation can have a role in the behaviour and growth of different species. Here, the authors present a meta-ecosystem model with explicit stoichiometry, coupling carbon and nitrogen dynamics between two ecosystems. They provide evidence that resource flow between ecosystems can lead to greater productivity at the meta-ecosystem scale through a positive spatial feedback loop. They also provide evidence spatial flows can have a negative impact on production when it perpetuates a mismatch in stoichiometry of available resources and the requirements of basal species. The authors encourage the integration of landscape heterogeneity into stoichiometrically explicit meta-ecosystem models and continued investigation into the effects of both qualitative and quantitative aspects of resource flow.

General Comments:
1) I enjoyed reading this paper. The authors tackle an important problem in spatial ecology. The paper is an impressive amount of work with transparent attempts to look at sensitivity of model predictions to a variety of scenarios. The explanation for each decision in the methods was clearly stated and referenced the very thorough supplementary information.

2) I like the blending of an empirical synthesis with a model. But the combination of data and model is not clear. Most importantly, I am struggling to understand the ecology behind the flows of bacteria (B) and consumers (C) in the aquatic system to the terrestrial ecosystem. What is the mechanism at play here or what is a real scenario where this can happen? The empirical synthesis suggests that common aquatic to terrestrial flows are amphibians, insects (like emergent aquatic insects), and carcasses – but none of these are depicted in this model. It is rather unique to not have parallel food chains (ie nutrients, plants, herbivores in each) in the two patches of a metaecosystem so this needs to be better justified based on the ecology of these coupled aquatic-terrestrial systems. In the end, I was left wondering if the authors are modelling something that does not occur in nature.

3) Related to the above, the rationale and methods behind the inclusion of empirical data is not clear (for example, some of the nitrogen to carbon ratios presented in table S1.1 in the appendix are for organisms not present in the model). I suggest clarifying the connections between the empirical data and meta-ecosystem model. Additionally, the full reference for all third-party data should be summarized in a table either within the text or in the supplementary information and should be referenced whenever these data are mentioned. The authors also chose to base some model parameters using empirical data from multiple types of terrestrial and aquatic ecosystems (streams, lakes, forest, and grassland), which have different N:C ratios from each other, to build a general terrestrial-aquatic model. The authors may consider basing the model parameters on data from specific terrestrial and aquatic ecosystems or provide further rationale for how the current approach is a reasonable one.

4) I like the diverse metrics used by the authors. The log ratio of production (eq 3) and the spatial feedback metric (eq 4) are simple and easily interpretable, the sensitivity analysis metrics and figures in appendix are clever. I can envision these methods and metrics being adopted by other researchers.

5) I like the model but I also have lots of questions about the model:
i) subscripts T and A are only used in the detritus and nutrient compartments – why not carry them throughout every compartment? This would make it easier to read and reduce cross-referencing needs. I found myself going back to Fig. 2 all the time to recall what G, P, B, and C were. I realize it would mean a double subscript but this is already done with variables D_i
ii) the authors say they use parameters that lead to coexistence of the variables. How exactly was this measured? After a certain number of simulation years? Was there a threshold used to make sure a variable was not < 0 (I.e., 0 or 0.001, or 0.1?). Did you observe many cases of variables < 0 and you only retained the feasible (ie > 0) scenarios? I would like to see more details on this as with such a complex model, coexistence of all variables is likely very challenging to achieve. I would expect exploitative competition via subsidies to lead to extinction of some trophic levels.
iii) I struggled with parts of the equation. In particular the flow from aquatic to B_cf_c to N_T or B_Gf_G to N_A – there are no arrows for these in Figure 2. What did I miss? I would find it very useful if you could add the process reflected in each part of the equation in plain text above each part. This can be done in appendix. I want to understand all parts of the equations but I am not quite there yet given the complexity.
iv) It would be useful if you could add a verbal definition of each parameter (and variable) to table S2.1.
v) why did the authors use r_i for N:C ratios – most other work uses alpha, beta, etc and r usually depict per capita growth rates in many ecological models. I found this choice of notation unnecessarily confusing (i.e., not what an informed reader would expect, thereby making it more difficult to interpret).

6) Quality vs quantity trade-off. How was this really done? I think the effects are just a result of changes in the quantity of C and or N. Resource quality can be other things – like nutritional ecology literature – see review in Raubenheimer et al. 2009 Funct Ecol and new theory in Osakapolor et al. 2023 Ecology (Omega fatty acids). The authors might consider some text on these other ways of measuring quality in order to reach a broader audience. This would be well received by those trying to integrate nutritional ecology and ecological stoichiometry. The paragraph on p. 16 lines 29-51 would be a good place for this.

7) P. 6 lines 27-28. This is an important assumption – all subsidies are detritus, ie there is no active feeding on live subsidies which is a common ecological phenomenon (e.g., spiders feeding on emergent insects). This may be a point worth mentioning in the discussion – how might the dynamics differ if the subsidy could arrive as not only detritus (see review in Allen & Wesner 2016 Ecology).

8) P. 9 lines 17-19. This is concise but it may be more useful to cite some of these figures in the main text? Alternatively, I recommend the authors provide some brief interpretation of these sensitivity analyses in the appendix. While they may be mostly qualitatively similar, there are important quantitative differences – especially in figures S5.X (ie addition of predator).

9) Fig 3. These effect sizes or variations in productions are very small – even for log response ratios. Are these biologically significant?

10) Overall, I find the results very interesting but I was left wanting to know more about the mechanistic pathways that give rise to these results. This is one drawback of log response ratios – it abstracts away specific pathways. Can the authors provide some visual aids to help the reader better understand how patterns arise as a result of the indirect paths in their model?

11) I recommend the authors provide some citations to key results figures in places in the discussion to facilitate cross-referencing.

Specific Comments:
12) P. 1 line 29. The Leibold citation is not a metaecosystem model paper.

13) P. 2 line 38. I think you mean Leroux & Loreau 2012 and not 2008 as 2012 was a meta-ecosystem model.

14) P. 6 line 15. Can you provide an explanation for why terrestrial decomposer activity is implicitly built into the model whereas the aquatic decomposers are explicit?

15) P. 9 line 48. There’s a typo; change “we observe that aquatic ecosystem is a net source” to either “we observe that the aquatic ecosystem is a net source” or “we observe that aquatic ecosystems are a net source.”

16) Figure 1 caption (lines 3-35): consider referencing the source of the empirical data in the caption (see comment 1)

17) Figure 4 (pages 25 and 26): the caption references a black line, but no black line is visible in the figure.

18) Figure 5 (page 27): line 21/22 in the caption says that “some inserts [in the figures] are displayed for some curves”. The meaning of this is not clear, what do the inserts represent?

19) Page S3. In the first paragraph you list “species < genus < family < order < phylum” but it is not clear whether this is simply showing the order of taxonomic levels or if this is a hierarchy of the N:C ratios that you were using.

20) Figure S7.1 caption (page S30): can you clarify what the terrestrial production is being subtracted from?

21) Figure S7.7 (pages S36 and S37): consider making it clear how the blue rectangle (fraction of subsidies exported to the aquatic ecosystem) is scaled. Currently not shown quantitatively.

22) Figure S4.3: the axis labels on the subplot insert did not render properly (labels overlapping each other)

The authors have done a very good job of addressing the diverse and comprehensive set of comments from the three reviewers. I found the additional details, analyses, and figures to be useful. I have a few small final suggestions.

For equations 1 and 2 in the main text, I recommend the authors state explicitly why the P_NT, G_NT, and C_NA equations are not listed. They are listed in appendix. Of course, they are simply some N:C ratio of the corresponding carbon equation for each trophic level but the lack of explicit text on this may render the work difficult to reproduce. Having struggled with deriving my first stoichiometrically explicit models, I can say that this missing information would have been a major barrier for me. Daufresne is a good example for how to communicate this concisely. You may even consider stating “As demonstrated in Daufresne…dP_NT = alpha_i*dP_CT and therefore, the set of equations reduces to…”.

Line 378. There is a typo here “Yet, yet…”

Figure 2. I recommend the authors place the change in notations included in equations 1 and 2 also in the figures, especially the model diagram.

Equations on pages S21 and S22 – why have the authors dropped the subscript T and A in the G_N, P_N, and C_N equations?

Appendix section S8. This is a good idea. But in the main text, delta does not have a subscript, so I assume all delta are equal. Here the authors look at variation in delta and they add a subscript to delta. I think this will be confusing. I recommend the authors either i) define delta with a subscript in the main text and state delta_P = delta_H = delta_B=delta_C but we look at variations in this in Appendix S8 or ii) define the subscript for delta in this appendix.