This revised manuscript is much clearer and the revisions have addressed many of the reviewer comments. After reading the paper I do think there are some additional issues that remain to be addressed, listed below. The most substantive comment is the last one, concerning separation of the observation of weak association of icebergs with NPP from hypotheses about functional relationships.

1) On page 6, lines 10-15, the F test for linear trend in iceberg changes is mentioned. Although the F test is fairly standard a reference here would be helpful.
2) In the tables, where means are presented, and p values given, a measure of uncertainty in the means (for example 95% confidence interval) should also be given - really, for all reported statistics in this paper the confidence interval should be given. Where p values are given the test these are relevant to should be mentioned in the figure caption as well.
3) On page 6, lines 29-31, NPP differences between cells with and without icebergs are discussed for seasonal data. Although there are differences they are very small, are these really important?
4) Page 10, lines 20-26. This section starts by saying that the general enhancement of NPP near icebergs is consistent but then says that the small iceberg effect in the CSZ is not significant. These seem contradictory statements to me. The large iceberg effect in the PFZ also seems to be ignored in this section, and then the section ends by saying the data may be too coarse. So, I am the first sentence of this section seems at odds with its remainder - and my suggestion is to reword the first part unless you really feel you can demonstrate a consistent enhancement.
5) Page 10, lines 28-32. Significance is repeated several times here, this could be reworded.
6) Page 11, line 5, type “ad”
7) Page 11, line 11-12. This sentence seems to be saying that because icebergs did not change much then it is not surprising that NPP did not change? But the alternative hypothesis is that icebergs don’t matter to NPP, or that other things were affecting NPP, given that the relationships presented elsewhere in the paper are not that strong. I think this could be reworded, perhaps to say that if icebergs are important to NPP then it is not surprising that a trend in NPP was not found.
8) Page 12, line 26-29. It is first stated that in general grids with more icebergs have high NPP but then an exception (the CSZ) is noted. “In general” implies to me a consistent pattern, rather than one with clear exceptions, so I suggest rewording.
9) In the conclusions section I am concerned about concluding there is a functional relationship between icebergs and NPP based on the correlations in the data. Reviewers also brought this up. For example, the statements that “Therefore, icebergs, large and small, have an observable positive impact on ocean NPP at the SO scale” and that that icebergs might provide a negative feedback to warming imply that the data presented here prove there is a functional relationship, but as I understand what they show is there is a statistical association, mostly fairly weak. Readers who do not look carefully at the data in the paper may draw the wrong conclusion from these statements so I suggest some revision is needed to carefully separate the observations from hypotheses about what they mean. There are similar statements in other places in the manuscript.

Thank you for considering all of my comments and for your clear responses. I have some small mostly technical comments on the revised manuscript but would ask that you consider point #2 below particularly.

1) Abstract, lines 16-19 – the wording still, to me, strongly implies causation (“icebergs increase NPP”) but as discussed in other comments what is shown is an association in the data.
2) Abstract, lines 22-23. About the negative feedback on global warming. Because this would be an important point for carbon cycle science, I would like to ask if you can quantify the potential impact based on the data in the paper. For example, what is the total difference in NPP integrated across the different zones for regions with and without icebergs? Is it at all significant relative to the total global NPP? If not, can the statement about negative feedback be justified? Warming may produce more icebergs, but eventually it would produce less, as ice margins retreat on to land. I have the same comment about the text in the conclusions.
3) Page 3, line 2. “Has” should be “have.”
4) Page 3, line 18. Here and elsewhere, data is a plural word, please check the text.
5) Page 7, line 25. Are all differences significant? CSZ differences are not, correct? Can you give differences and p values in Table 2?