I appreciate the modifications that the authors have made to the manuscript in response to the comments by both reviewers.

The clarification of the aim of the manuscript as focusing on the Beaufort region and comparing with the long-term traps described in Forest et al 2013 helps. Both reviewers remark that a floating (or ice moored) trap design is not commonly used in ice-covered waters. Rev #1 further says that few or none are available for the Arctic Ocean ( but see many publications by the Wassmann group with short-term floating traps). Similarly to Rev #2, I also miss a more thorough reference to other work in the Arctic Ocean. However, the authors do briefly mention a few studies done in the North Atlantic and other parts of the Arctic Ocean, i.e., "However, similar or even lower fluxes have been reported from elsewhere in the Arctic Ocean (Cai et al., 2010; Chen et al., 2003). " [p1265, L5-8]

Hence, I'm surprised that results from the Barents Sea by the Wassmann group are not used to place these Beaufort Gyre results in a broader Arctic Ocean context [even if briefly], especially given the focus this manuscript places on fecal pellets and seasonal change (eg, Wassmann et al 2006 PiO); it may be that Barents Sea export fluxes are much higher than the Beaufort Sea ones in absolute terms- that's not my point. Rather the apportioning of the fecal pellets to mesozooplankton groups and the seasonal differences are relevant despite location- eg, low fluxes early in the season in the Barents may be ecologically more similar to conditions in the Arctic basin or other Arctic regions. Same for winter/spring transition when heterotrophy is dominant - similarly to late summer/fall when the study herein occurred.

There is also the unforgiving interannual variability in these Arctic fluxes - something for which one can also draw from Barents Sea studies (eg, Reigstad et al 2008 DSR). This is not mentioned in the manuscript, though it is now discussed by the authors in their reply to Rev#2.

A few sentences in the discussion, where pertinent, would clearly place these results in a broader context; the manuscript should stand on its own, not only as a complement to the Forest et al 2013 paper.