Dear Drs Zha and Zhuang,

Thank you for responding to the referee reports. Work on the manuscript has addressed many of the concerns by referees, and I am content with these (e.g. the issue of equifinality). However, I am not convinced that you have managed to alleviate the main point of referee 1, which is that the uncertainty associated with both models compared here is so large that it is not possible to judge any potential improvement from incorporating dormancy. I would therefor like to ask you to revise the manuscript further to address the issue of modelling uncertainty. I include some further specific explanations below.
I agree with referee 1 that model uncertainty is key for this comparison, and this is not well illustrated also in the revised paper. Figures 2, 4 and 5 show that in some instances the new model follows temporal dynamics better, but the improvement over the non-dormancy model is not that clear, and in absence of explicit uncertainty associated with either model, the reader can not judge whether this is a significant improvement. This then leads to similar problems when making regional predictions – Are stated differences for the contrasting models within the model uncertainty?
You also don’t address fully the point that the dormancy model increases modelling uncertainty owing to the larger number of parameters. As no modelling uncertainty is quantified at present, I think that this fundamental limitation of constructing a more complex model (i.e. with more parameters) is not reflected appropriately.
Figure 11 seems to show the wide band of uncertainty (but this is not clearly explained in the figure caption). Can the apparent difference/improvement of the model prediction be justified, given the considerable underlying uncertainty?
Some of the figures are still very hard to read. Axis labels and numbers are generally too small. Fig. 3, for example, is improved as far as resolution is concerned, but font size is far too small. The same applies to other figures.

Best regards,
Jens-Arne Subke

Dear Drs Zha and Zhuang,

Thank you for responding to the referee reports. Work on the manuscript has addressed many of the concerns by referees, and I am content with these (e.g. the issue of equifinality). However, I am not convinced that you have managed to alleviate the main point of referee 1, which is that the uncertainty associated with both models compared here is so large that it is not possible to judge any potential improvement from incorporating dormancy. I would therefor like to ask you to revise the manuscript further to address the issue of modelling uncertainty. I include some further specific explanations below.
I agree with referee 1 that model uncertainty is key for this comparison, and this is not well illustrated also in the revised paper. Figures 2, 4 and 5 show that in some instances the new model follows temporal dynamics better, but the improvement over the non-dormancy model is not that clear, and in absence of explicit uncertainty associated with either model, the reader can not judge whether this is a significant improvement. This then leads to similar problems when making regional predictions – Are stated differences for the contrasting models within the model uncertainty?
You also don’t address fully the point that the dormancy model increases modelling uncertainty owing to the larger number of parameters. As no modelling uncertainty is quantified at present, I think that this fundamental limitation of constructing a more complex model (i.e. with more parameters) is not reflected appropriately.
Figure 11 seems to show the wide band of uncertainty (but this is not clearly explained in the figure caption). Can the apparent difference/improvement of the model prediction be justified, given the considerable underlying uncertainty?
Some of the figures are still very hard to read. Axis labels and numbers are generally too small. Fig. 3, for example, is improved as far as resolution is concerned, but font size is far too small. The same applies to other figures.

Best regards,
Jens-Arne Subke

Dear Drs Zha and Zhuang,

Thank you for addressing referees concerns over the manuscript. I have reviewed your changes and have reason to request further changes before I’m able to finally accept this manuscript for publication. This addresses mostly the writing and framing of your science, not further analyses, and I hope that you will be able to perform revisions relatively swiftly.

A key concern is your framing of the study as addressing “Arctic” soils. Your study does not focus on Arctic ecosystems. Latitudes north of 45° include mostly temperate and boreal regions, with the Arctic making up a small fraction only (possibly 10 - 15%, depending on definition). Your reference sites (Table 2) include only two actual Arctic sites. It’s not clear to me why you present this as a study addressing Arctic C storage; the model addresses microbial dormancy in general, why is it framed specifically around the Arctic context? Please remove or revise sections that present this study as specifically Arctic-focused (e.g. lines 48/49 and 58 in the Abstract, and lines 114-120, and 126/127). Similarly, “high latitudes” would normally be understood to be > 60°, most commonly in fact > 66° (i.e. pole-ward from Arctic Circle). Please remove reference to “high latitude” where you include areas north of 45°N.

This aside, I also think that you have to make it much clearer that this presents the potential of including microbial dormancy in models to potential outcomes, rather than providing realistic scaled numbers for northern regions. A very substantial part of land surface north of 45° is agricultural, which is entirely absent from your data sets used for parameterisation. Please make it absolutely clear that this is an exercise of using models to estimate potential regional C dynamics, but that this applies to natural ecosystems only, so it is a partial C balance for the region. I see no problem with this being a study to demonstrate the role of dormancy in principle, but want to avoid these numbers being used to claim actual source or sink functions of ecosystems in the region. Your objectives stated at the end of the introduction should make this very clear.

Please note the detailed points below when you re-draft the manuscript. I hope that this will make the manuscript more focused and comprehensive.

141: Again, why the Arctic specifically? Your modeling approach is generic to any soil microbial dormancy, or not?

166/184: Equations 1 and 3 are identical, but use different parameters. Please keep consistent (and possibly omit one of them, as they are repetitive).

247-248: Please refer to biomes covered as “northern ecosystems ranging from temperate to Arctic”, or similar. Not “high latitude”.

260: This is misleading. Four of the validation sites are the same as those used for parameterisation. This has to be absolutely clear.

337-339: This sentence goes beyond data presentation. Please move it to the discussion, where this interpretation is more appropriate.

352 – 360: Also here, avoid lengthy explanations in the results section. Abbreviate the entire section to: “Without considering dormancy, MIC-TEM estimates more active microbial biomass, hence overestimating both RH, and NPP (due to higher simulated N mineralization and uptake by plants), but resulting in lower NEP than that calculated by MIC-TEM-dormancy.”

361-373: This section is not very well phrased. Please present the key findings of the seasonal dynamics without detailed descriptions of the shape of graphs. Also no need to reference the fact that moisture and temperature are higher in summer.

397-398: Delete last sentence (again, this should be picked up in the discussion).

399-400: Delete first sentence, and add “in the 2090s” after “scenarios” in the following sentence. Note that it is 2090s, not “2990s” as stated in line 399.

410, 413 and 415: Replace “predicted” by “predicts”.

418 – 426: This should be deleted as it has been covered in the introduction. Focus on key findings of your study in the discussion.


Figure 3: The figure caption states that you show parameter ranges for all six sites, but on three are shown. Text in the Results section also suggests that data are shown for all six sites.

Dear Drs Zha and Zhuang,

Many thanks for sending the updated version for the manuscript. This now meets all requested edits from referees as well as from me, and I'm happy for this to be published.

Best wishes,
Jens-Arne Subke