No abstract

Four years of temperature, salinity, and velocity data enable a direct computation of volume transport and a temporal description of water properties exchanged through the Bering Strait. The mean volume transport over the 4‐year period (September 1990 through September 1994) is 0.83 Sv northward with a weekly standard deviation of 0.66 Sv. The maximum error in this mean estimate is 30%. Interannual variability in transport is typically 0.1 Sv but can, at times, reach nearly 50% of the mean. The transport of 1.14 Sv during the first 9 months of 1994 is the largest in the last 50 years. The rate of winter salinity increase is very similar from year to year, suggesting regional average ice formation of about 5 cm d−1. The amplitude of the annual salinity cycle is about 2 psu, with salinity reaching a maximum in early April. There can be large interannual variations in the salinity (about 1), particularly in winter. Background autumn salinities average 32.0 in the eastern and 32.6 in the western channel.

Two closely instrumented arrays were deployed within Barrow Canyon during 1986–1987 in an attempt to measure the outflow of dense, hypersaline plumes created during sea ice formation along the Alaskan coast. However, no hypersaline plumes were observed. Rather, we found cold, relatively fresh waters advected downcanyon by the mean flow alternating with upcanyon flow of warm and saline water upwelled onto the shelf. Upwelling was most frequent in the fall, and upcanyon speeds reached 60 cm s−1. At times the resulting onshore heat and salt fluxes were large enough to be of possible local significance, for example, to the surface heat budget. Contrary to earlier findings, the flow was only weakly correlated with the wind and the atmospheric pressure gradient. Instead, we found both upwelling and flow reversals to be coherent along the coast at sites 400 km apart, with phase differences corresponding to a typical speed of 2.3 m s−1. We suggest that the majority of these events are manifestations of shelf waves propagating eastward along the Arctic Ocean margin.

Analysis of time series from 1981 to 1982 near Bering Strait shows the meridional wind component, the current, and the oceanic pressure gradient to be well correlated. We argue that because of the constraining coastal geometry, there are wind‐driven flow divergences which perturb the oceanic pressure field and account for most of the current variability. For example, northerly winds cause a pronounced set‐down of sea level southeast of Bering Strait, and if the winds are strong enough, they will reverse the mean slope downward to the north through the strait. Winds during 1976–1977 were also well correlated with the Bering Strait transport estimates of Coachman and Aagaard (1981), and we therefore use the calculated winds from 1946 to 1982 to examine the low‐frequency transport variability. There is a marked seasonal cycle, with summer transport about 50% greater than during winter. The long‐term mean transport is calculated as less than 0.6 Sv, which is much lower than earlier estimates. Abnormally large transports apparently occurred during a number of years from 1948 to 1967, but during the past 15 years, transports have tended to be considerably less, corresponding to the stronger northerly winds during these years. The interannual wind variability near Bering Strait, together with the corresponding variability in the transport through the strait, is in fact part of the large‐scale variability of the atmospheric circulation over the North Pacific and its oceanic effects.