Dear Dr. Pandit,
thank you for your author response to the two reviews.
Regarding your response to reviewer #1, please
1. include your response to the comment on line L156-157 in the manuscript.
2. make sure you not only add biomass to your figures, but also add the interpretation of those results (GPP change vs. biomass change after) in the manuscript text (reviewer comment to "L169-170: why". You need to thoroughly explain why GPP is reduced over several years after fire and explain it with the model assumptions and model algorithm/equations. This is unusual outcome compared to observations that see a quick GPP recovery after fire.
Response to Reviewer 2
1. Please add sentences on the conclusion and discussion points of this study in the abstract and check if you have answered (conclusions) and discussed the objectives of your manuscript in the respective sections.
2. Make sure that you address all reviewer comments regarding model description, methods and statistics used in the analysis of model results. The effect of fire on shrub GPP needs to be plausibly described. If you have benchmarked your model already in a previous publication then please cite the outcome in the GPP evaluation.
3. your response to the comment referring to L153: I think your response misses the point raised by the reviewer. Your response explain the general fire ecology in the shrub ecosystem under study, but you need to make sure you explain the curves displaying fire and GPP in your model results. Please check again this point in your model and model analysis and make sure the simulated pattern are sufficiently explained in the manscript.
4. your response to comment regarding L158: when adding this aspect to your manuscript, please consider the time scales of your simulation study and make sure the stability definition is not flawed by discussing seasonal pattern.

Best wishes,
K.Thonicke.

Dear Dr. Pandit,
thank you very much for your revisions and detailed responses. Your manuscript has advanced but still requires further improvements or explanations.
One major issue that I still have is that you write in line 285 that the simulated pattern might be influenced by the model not being in equilibrium. I think this is a crucial point. If the simulated fire and GPP pattern that you analyse in your manuscript are additionally influenced by the model getting into equilibrium then it is unclear if the pattern you describe are solely driven by climate and you can draw the conclusion as you did. I apologize for not spotting this point earlier, but I regard this as a crucial point which I ask you to cross-check in your modelling protocol, if your model results are in equilibrium with climate and you then only simulate the influence of transient climate on fire and GPP. Because this point might require more time to revise and adjust the results shown in the manuscript, I will reconsider your manuscript after major revisions (it is more work than minor revisions). In addition to this point I have the following points which I ask you to implement in the revised version:
Abstract/ Introduction
Line 4: correct to Dynamic Global Vegetation Model to come up with the abbreviation DGVM.
Line 11: Why EDv2.2 is described as “dynamic vegetation model” not DGVM? If only “dynamic vegetation model” is meant, then correct the justification for a DGVM in the sentences before and justify dynamic vegetation models instead. This also applies to your respective paragraphs in the introduction (lines 48-69).
Methods
Line 89: delete “about the model”
Line 111-112: add meaning of µ and µ0 for λ, correct to “relative area”
Line 115: decription and explanation of terms in equation incomplete and unclear, please correct. Explain in a separate sentence what the basic idea of equ. 2 is. Why does soil depth play a role in fire ignition? This is totally unclear. Even if it has been described in a previous paper, this is an important feature of the model that the reader of this manuscript needs to fully understand.
Study area
Use “RCEW watershed” throughout the manuscript when you refer to the regional simulations, and “soda fire” when you refer to the Soda fire regional simulations. This way you make the use of these terms consistent with what is shown in Fig. 1.
Caption Fig. 1: Add the information that the grey area in the large map marks the extent of the Great Basin.
Line 125: define here, clearly in the text, that you have another definition of your study area to run your regional simulations.
Line 205: replace “making comparisons with mean monthly GPP of July from the model.” with “comparing them against simulated mean monthly GPP values of July.”
Methods and results:
1) replace “long-term” with decadal or “mult-year”. “Long-term” is unspecific and people have different understanding what “long-term” stands for. Use “seasonal” instead of “short-term”, again this is a unspecific description of time scale.
2) Use subscripts of C3 and CO2 throughout the manuscript consistently.
Results
Line 241: it is “underestimated” not “underpredicted”
Fig. 2: the fire model simulates fire fractions every year. Is this realistic (how does it compare to observed data) and how can it be explained by the design of the fire model (main assumptions, refer to equ. 1 and 2).
Line 245 ff and throughout the section: Please rewrite “loss of GPP” think of recovering GPP instead, Specify which row in Fig. 4 you refer to, because Fig. 4 actually shows a comparison between the fire and no-fire experiments. GPP is low, and still less than the initial conditions in 2016. In 2019 this spatial pattern is still different from 2016. This description is missing in your results paragraph. It is not “loss”, it is rather “reduced GPP”.
Line 255: replace “later” with “latter”
Line 270: Check if corrected sentence is complete.
Line 285: if your transient simulations have not yet reached an equilibrium, then please revise your modelling protocol accordingly, and re-run the simulations so that the simulation results are a result of climate and fire and not modelling artefacts. This is crucial for the manuscript to get published!
Line 299: delete remaining part of an old sentence: “For EDv2.2 GPP”
Caption of Figure 6: explain colors
Discussion
262: Forkel et al. 2019 is a global study. It is not comparable to the spatial and temporal scale of your study, please delete.
Line 374: if you see potential errors occurring from WRF, specify which those could be and how do they influence your results. Same applies for MODIS GPP data (line 379).
Entire text: please revisit the revised text again where formulations can be sharpened more to the point. See example for line 205.
Conclusion:
The fact that GPP is still underestimated by 20% should be reflected in the confidence description of your results. The model is missing the elevation gradient in GPP which is also reflected in the flux-tower comparison. This should be clearly stated in the conclusion.

Thanks for your understanding and looking forward to see your revised manuscript.
Best wishes,
Kirsten.

Dear Karun Pandit,
thank you for implementing the requested changes. It again helped a lot to improve the manuscript which I am happy to accept now.
Best wishes,
Kirsten.