General comments

This interesting manuscript describes a simple/conceptual model, and secondarily the intermediate-complexity JULES model, to explore in what circumstances carbon cycle instabilities can occur. The results are interesting—instability depends on ECS and CO2 fertilization—and the authors nicely document that the two models produce qualitatively similar results, although computational limits preclude a full comparison. The methods are generally clear and well written, and this is appropriate for EGUsphere and will be of interest to a wide audience.

There are some significant problems. First, there’s very little acknowledgement of previous research and similar studies—the discussion is particularly inadequate in this respect. Second, this is not a mathematical journal, and I think it would be worth defining many terms and methods more fully, i.e. clarifying the methods in many places. There are some structural problems. Finally, some aspects of the methods and study are unclear.

In summary, this is an interesting experiment that will be of wide interest, but the current ms needs major revisions to acknowledge previous literature; make it more appropriate for the journal’s readership; and improve the structure and flow of the text.

Specific comments

1. Line 4: perhaps clearly define what you mean by “unstable”
2. 11: a final sentence would be useful, laying out implications of this study
3. Introduction: have no other papers explored this since Cox 2006?
4. 69: “fitted to the outputs of more”
5. 72: “II”?
6. 132: What’s the total quantity of carbon in the simple model system, and at steady state how much is partitioned into atmosphere, land, and ocean? How does it compare to e.g. GCP2024 numbers?
7. 157: this isn’t a specialist journal, so it seems important to define what a Hopf bifurcation is; same comment re “bifurcation diagram” in line 200 and “eigenvalues of the Jacobian” in line 208. Some readers will understand, many will not
8. 222-250: belongs in the methods
9. 262-274: this just restates previous material; remove or greatly condense
10. The discussion is completely inadequate; in particular, I doubt that no one else has performed exercises like this in the last 20 years (see #3 above), and it’s important to compare and situate these results in the context of current scholarship. Currently there are almost no literature citations.