Dear Dr. van Sundert and colleagues,

I have now received two thoughtful reviews of your submitted manuscript. While you address some of the reviewers concerns in your answers, I do think that you are not sufficiently open to some of their suggestions. I will give you some examples:

-Reviewer 1 makes the excellent suggestion to stratify before you do your analyses. This reviewer specifically suggests to stratify according to wetness class before exploring soil fertility x productivity relations. I think that this is an important suggestion and deserves a more careful consideration than simply doing the same analysis for four different site wetness classes. I would even argue that you can only expect soil fertility to directly influence productivity if the site is well drained. Of course the IIASA approach does not consider this since agricultural sites are preferable located in better drained sites (unless it is rice), but with forests this is obviously different. Since you have a large database, you should consider to explore the suggestion to stratify further and explore whether you find stronger relations if you stratify according to soil type. At least you are then trying to find relations for sites with comparable soil forming factors.

-Reviewer 1 also suggests to include N deposition in your analysis. Your write ‘Nonetheless, if the editor so wishes, we are willing to test the influence of N deposition separately and present it in the revised manuscript`. My answer is: yes, I wish that you analyse whether N deposition affects productivity. It is something that you should have done after this suggestion was made and before writing your answer. Please do not forget to stratify here as well, since I predict that site wetness and other soil forming factors will interact with any N-deposition x productivity relation. There is a wealth of literature showing how strong the effect of N deposition is on forest productivity and just ignoring this by declaring it is out of the scope of this study is not helpful if N deposition affects productivity in you dataset.
-While I understand your arguments why you chose the IIASA approach, it would still be interesting if a purely statistical approach would lead to better results. Why don’t you present this as well (and don’t forget to stratify here as well)?

-Reviewer 2 points at the limitations of your study for developing a global metric. I agree with this reviewer that this is not possible but I do think that a metric for e.g. ‘boreal forests on well drained soils’ would already be a major improvement compared to the present situation. I think you should address this more specifically in your revised manuscript and acknowledge that a global metric is not realistic.

-Reviewer 2 also makes the excellent suggestion that data quality (including comparability of data potentially analyzed with different chemical methods) may be an issue. In your answer you are not addressing this but it is still a very valid consideration. Were uniform methods used? How many laboratories were involved in the data? Is anything reported about detection limits? How are missing data treated? These are all valid questions that you have to address in your manuscript. You cannot expect people in dig into the original dataset to find this out.

-In you answer you refer twice to the need for using datasets where climate does not vary but nutrient availability does (e.g. local gradients). I was wondering why are you not exploring this? If you have a dataset with 2500 points I am sure that there are selections that can be considered local gradients. In my opinion you have the data to test this; it is not something that should be postponed to the future.

In summary, use these two excellent reviews to better explore your dataset and improve your manuscript. I look forward to your revision.

I will sent a revised manuscript out to for review again

Best regards,

Edzo Veldkamp

Dear authors,

I apologize for the long turn-around of the reviews of your manuscript which was related to the non-responsiveness of one of the original reviewers. Luckily I found an alternative reviewer who made a careful review (anonymous reviewer #3). Of course a new reviewer brings in different views compared to the original reviews. After reading reading your revised manuscript and the two reviews, I think you have done a thorough job in revising the mansucript based on the original reviews. Nevertheless I think there is still room for improvements:
I agree with reviewer #1 that you should adjust your abstract. My recommendation follows the recommendation of this reviewer: to write a short convincing story that clarfies the reality of developing this kind of indices. Furthermore many details in your manuscript sometimes distract from the main story. I recommend to critically look if all the details are necessary.
Reviewer #3 does not like the term 'upgraded'. While I see the point made by this reviewer, I think this is mainly semantics. You many consider to use the term 'adjusted' or some other term, but if you give me a good reason why the term 'upgraded' is preferable, I am open for it.
While the list of suggestions is relatively long, I think that most of them are relatively minor. After receiving your second revision I will decide if it is necessary to send out for review again, but I expect that if you make a careful revision that only a review by myself will be necessary.

best regards,

Edzo Veldkamp

Dear authors,,

I have read yóur revised manuscript and your response to the comments of the reviewers. I think that you manuscript has improved considerable along the review process and hope that you agree with me. Thank you for submitting your interesting study to Biogeosciences.

Best regards,

Edzo Veldkamp