Dear authors,
many thanks for your extensive replies to the very constructive reviewer comments. Please submit a revised version along the lines of your replies, making sure that you benefit the most from the suggestions made.
Best,
Sönke

Dear authors,

many thanks for the revised version of the manuscript. I have unfortunately failed to receive one review in time, and therefore assessed the manuscript myself. Both, reviewer and I think that your revision covers the main points of critique raised during the previous set of revisions, but have minor comments to increase clearity of the manuscript. Please respond to these suggesting by providing a point-by-point response to the comments along with an annotated manuscript of the changes made.

Many thanks,
Sönke Zaehle

Minor editorial comments:

p 4 L30: add “of the joint distribution” to avoid confusion (each variable is sampled uniformly, so for each variable, the tails and the centre are equally likely, making the last part of the sentence illogical).

p9 L 21: How is uncertainty defined here, and based on which model output. Obviously, parameters will have different effect sizes on different variables of interest, so the choice of this criterion does matter for the validiy of the approach.

P9 L24. Having problems understanding the iteratively here, maybe “for each parameter set” is clearer?

Table 2: explain the use of bold fond.

P15 L 24: remove reference to Cameron et al. in prep, unless this is now in press or open-discussion.

P18 L 17: to my knowledge, M van Ojien is not an editor for Biogeosciences

Dear authors,
many thanks for these revisions. My apologies, I should have spotted this earlier, but there are units missing for Table 2, Table A4 (alternatively add a new table corresponding to Table A2), Figure A2 and Figure A6, which you will please add. Otherwise this manuscript is fine for publications.
Best wishes,
Sönke