Faroe Bank Channel overflow 1995–2005

Abstract

The Faroe Bank Channel (FBC) is the deepest passage across the Greenland–Scotland Ridge and there is continuous overflow of dense water through the channel, which provides a substantial contribution to the global thermohaline circulation. Since November 1995, Acoustic Doppler Current Profilers (ADCPs) have been moored at the sill of the channel continuously, except for short annual servicing periods. In addition to this, CTD sections have regularly been occupied across the channel and instruments, recording temperature and salinity, have been moored for short periods. These observations demonstrate that the channel, due to its small width, can be monitored fairly well with only one ADCP mooring and they allow the generation of a time-series of overflow volume flux for the 1995–2005 period. The kinematic overflow flux, defined solely from the velocity field, was found to have an average value of 2.1 ± 0.2 Sv with a seasonal and with inter-annual variations of approximately 10% amplitude, but with no discernible trend for the whole period. The average flux of water with potential density exceeding 27.8 kg m −3 was found to be 1.9 ± 0.3 Sv with average properties: θ = 0.25 °C, S = 34.93, γ θ = 28.01 kg m −3 for this period. No evidence was found for a weakening overflow flux, but the salinity of the FBC-overflow, especially the warmer parts, has increased substantially during the period, which implies an increased average density on the order of 0.01 kg m −3. Previous observations of a helical cross-channel circulation are confirmed and may explain the thin pycnocline layer, but shear-instability induces intensive mixing that puts a lower limit on the layer thickness. Critical control at one point of the sill cross-section, rather than friction, seems to govern the overflow dynamics and simple layered models can explain much of the observed features as well as the seasonal variation, but potential vorticity, as defined for a single overflow layer, is not conserved. A previously suggested 1950–2000 weakening of the FBC-overflow, which was partly based on a subset of the data presented here, is not supported by the full dataset.