the continental slope through Herald Canyon represent the more important pathway to the Arctic Ocean. The slope offshore of the canyon mouth in this region is also much less than off Barrow, likely resulting in less mixing over the slope than is the case near Barrow Canyon.
A curious aspect of the annual salinity signal in the waters entering the Chukchi is that the trend of winter increase in salinity appears identical in the two channels of Bering Strait. This implies either that the brine rejection driving the increase is primarily the result of regional ice formation distributed over the shelf rather than concentrated in coastal polynyas, since the latter occur predominantly in the western Bering Sea and would drain most directly through the western channel of the strait, or that the brines from localized sources are very efficiently redistributed over the shelf.
Another interesting issue is the relationship between these shelf circulations and nutrient fluxes. If the cross-shelf circulations forced locally by brine rejection are weak, then only a small compensatory on-shelf flow is required, and furthermore the offshore source is likely of shallow origin. Little if any nutrient enrichment of the shelf would occur. In contrast, the regional throughflows in the Barents and the Bering-Chukchi are large, and they carry significant amounts of nutrients, especially in the Bering-Chukchi. Even in the rather deep Barents Sea, however, the depth-mean summer nitrate values are in the neighborhood of 5 μmol L-1 or more, and mixed layer values are easily twice this large during winter.
I believe our understanding of the coupling of the shelf, slope, and deep basin in the Arctic, together with its consequences, are poised for major advances, and I appreciate the initiative of our Japanese colleagues in promoting this discussion through a very timely workshop.
