Dear Authors:

Thank you for contribution to Biogeosciences.
I examined your replies and manuscript revised on the basis of referee’s comments. Overall, I agree that the manuscript was considerably improved in comparison with the previous version. Also, I appreciate that you made several additional analyses as found in new figures.
Two referees, one disclosed and another anonym, gave different recommendations; one is affirmative and another is critical. I guess you felt difficulty in revising and replying these contrastive reviews. However, I believe these referees also represent the diversity of readers of this journal. Your study is quite model-oriented, and so modeling people are rather generous (i.e., understanding) for inevitable difficulties in model simulation and validation. In contrast, observational people are rather strict and sometimes skeptical about discrepancy in simulation data and scales. Because you used a broad-scale model (CLM-Crop) at the resolution equivalent to that of earth-system models, it is inevitable to bring a scale-gap from field observations. In this regard, I agree that aggregating SOC data for each grid (Figure 3) is a realistic way to reduce the scale-gap. However, I don’t agree to delete correlation coefficient (R) and RMSE from the figure, because they have still informative for readers to evaluate the model performance. And, I have another concern. You cited Jobbágy and Jackson (2000) to justify the gap in soil depths between field data (1m) and the model (3.8m). However, according to the paper, globally, about 36% of SOC exists below 1m-depth (842 Pg C out of 2344 Pg C); in croplands, this is still the case (91 Pg C out of 248 Pg C: Jobbágy and Jackson 2000, Table 3). So, I recommend revising the statement “the mismatch has little impact on results since most SOC is in the top soil layer” (line 222–223, revised manuscript) in a more modest manner. Perhaps, you may perform two comparisons: one present result, and another using scaled (i.e., 1-m data are multiplied by ca. 1.36 to obtain 3-m estimates) data.
As a modeling study, this is an interesting attempt for me, and I recommend minor revision for above-mentioned points.
We are looking forward to receiving a revised manuscript.

Best regards,
Akihiko Ito

Dear Authors:

Thank you for contribution to Biogeosciences.
I examined your revision and conclude that this is acceptable for publication.
I appreciate your patient revisions.

Best regards,
Akihiko Ito