ALBERT

Description: Open-ocean deep convection is a littleunderstood process occurring in winter in remote areas under hostile observation conditions, for example, in the Labrador and Greenland Seas and near the Antarctic continent. Deep convection is a crucial link in the “Great Ocean Conveyor Belt” [Broecker, 1991], transforming poleward flowing warm surface waters through atmosphere-oceaninteraction into cold equatorward flowing water masses. Understanding its physics, interannual variations, and role in the global thermohaline circulation is an important objective of climate change research. In convection regions, drastic changes in water mass properties and distribution occur on scales of 10–100 km. These changes occur quickly and are difficult to observe with conventional oceanographic techniques. Apart from observing the development of the deep-mixed patch of homogeneous water itself, processes of interest are convective plumes on scales 〈1 km and vertical velocities of several cm s−1 [Schott et al., 1994] that quickly mix water masses vertically, and instability processes at the rim of the convection region that expedite horizontal exchanges of convected and background water masses [e.g., Gascard, 1978].

Description: Moored Acoustic Doppler Current Profilers (ADCPs) were used to analyse the daily vertical zooplankton migration and its seasonality. One-year records of vertical velocity and acoustic backscatter were obtained at four stations in the Greenland Sea. Both parameters exhibited a diurnal cycle typical for vertically migrating zooplankton. Upward and downward migration occured in short periods approximately 5 h long, and peak migration velocities were around 1.5 cm s−1. Similar structures were observed at all four mooring sites in the 200–300 m depth range. Farther down, between 1000 and 1400 m, no daily migration was observed. Strong seasonal variations are evident, and both the phase and intensity of the migration pattern change with daylight as the season progresses. In summer and during the polar night the migration became very weak and was only detectable in the displacement of scattering layers. When the day/ night contrast was large, intense upward or downward motion was accompanied by sloping backscatter isopleths. We observed two main scattering layers, a deep layer that varies in depth with season and an almost invariable shallow scattering layer at about 150 m depth. The deep layer was interpreted as the “resting depth” of the migrating plankter, and the latter as their “feeding horizon”. Changes in the “resting depth” from about 400 m in autumn and spring to about 200 m in winter lead to seasonal variations in the migration distance. This behaviour is discussed with respect to environmental conditions.

Description: During the winter of 1988–1989 five acoustic Doppler current profilers (ADCPs) were moored in the central Greenland Sea to measure vertical currents that might occur in conjunction with deep mixing and convection. Two ADCPs were looking up from about 300 m and combined with thermistor strings in the depth range 60–260 m, two were looking downward from 200 m, and one was looking upward from 1400 m. First maxima of vertical velocity variance occurred at two events of strong cold winds in October and November when cooling and turbulence in the shallow mixed layer generated internal waves in the thermocline. Beginning in late November the marginal ice zone expanded eastward over the central Greenland Sea, reaching its maximum extent in late December. In mid-January a bay of ice-free water opened over the central Greenland Sea, leaving a wedge of ice, the “is odden,” curled around it along the axis of the Jan Mayen Current and then northeastward and existing well into April 1989. Below the ice a mixed layer at freezing temperatures developed that increased in thickness from 60 to 120 m during the period of ice cover, corresponding to an average heat loss of about 40 W m−2. Through brine rejection, mixed-layer salinity increased steadily, reducing stability to underlying weakly stratified layers (Roach et al., 1993). During the ice cover period, vertical currents were at a minimum. After the opening of the ice-free bay, successive mixed-layer deepening to 〉350 m occurred in conjunction with cooling events around February 1 and 15, accompanied by strong small-scale vertical velocity variations. Upward mixing of more saline waters of Atlantic origin during this phase reduced the stability further, generating a pool of homogeneous water of 〉50 km horizontal extent in the central Greenland Sea, preconditioned for subsequent convection to greater depths. Individual convection events were observed during March 6–16, associated with downward velocities at the 1400-m level of about 3 cm s−l. One event was identified as a plume of about 300-m horizontal scale, in agreement with recently advanced scaling arguments and model results, and with earlier similar observations in the Gulf of Lions, western Mediterranean. The deep convection occurred in the center of the ice-free bay; hence brine rejection did not seem necessary for its generation. Plume temperatures at 1400 m were generally higher than that of the homogeneous surface pool, suggesting entrainment of surrounding warmer waters on the way down. Mean vertical velocity over a period of convection events was indistinguishable from zero, suggesting that plumes served as a mixing agent rather than causing mean downward transport of water masses. However, different from the surface pool that was governed by mixed-layer physics, the water between 400 and 1400 m was not horizontally homogenized in a large patch by the sporadic plumes. Overall, and compared to results from the Gulf of Lions, convection activity in the central Greenland Sea was weak and limited to intermediate depths in winter 1988–1989.

Description: The seasonal cycles found in moored current measurements in the equatorial Somali Current region and along the equator between 50° and 60°E are compared with the multilayer Geophysical Fluid Dynamics Laboratory model for the tropical Indian Ocean. The remote forcing of Somali Current transport variations by incident long equatorial waves from the equatorial interior subthermocline region is investigated by analyzing the model velocities of annual and semiannual period. Amplitudes and phases of linear equatorial Rossby and Kelvin waves were least-squares fitted to the model velocities between 5°S and 5°N, 55° and 86°E from 100-m to 1000-m depth. Two cases of wave fits are distinguished: the “free” Kelvin wave case, where the Kelvin waves were fitted independently, and the “reflected” Kelvin wave case, where they were coupled to the Rossby waves by the western boundary condition for a straight slanted (45° to the north) coastline. The wave field velocities explained 70% of the spatial variance in the equatorial model subregion and also compared reasonably well with observed current variations along the equator. At the western boundary, the short-wave alongshore transport due to reflected incident long waves was determined and found to be antisymmetric about the equator. The maximum transport variation for the semiannual period due to the short waves was about 5 × 106 m3 s−1 between 150- and 800-m depth at 3° north and south of the equator. Observational evidence for the western boundary transport variations and the sensitivity to changes in the incident wave field are discussed.

Description: Ocean deep velocity profiles were obtained by lowering a self-contained 153.6-kHz acoustic Doppler current profiler (ADCP) attached to a CTD-rosette sampler. The data were sampled during two Meteor cruises in the western tropical Atlantic. The ADCP depth was determined by integration of the vertical velocity measurements, and the maximum depth of the cast was in good agreement with the CTD depth. Vertical shears were calculated for individual ADCP velocity profiles of 140-300-m range to eliminate the unknown horizontal motion of the instrument package. Subsequent raw shear profiles were then averaged with respect to depth to obtain a mean shear profile and its statistics. Typically, the shear standard deviations were about 10(-3) s-1 when using up and down traces simultaneously. The shear profiles were then vertically integrated to get relative velocity profiles. Different methods were tested to transform the relative velocities into absolute velocity profiles, and the results were compared with Pegasus dropsonde measurements. The best results were obtained by integrating the raw velocities and relative velocities over the duration of the cast and correcting for the ship drift determined from the Global Positioning System. Below 1000-m depth a reduction of the measurement range was observed, which results either from a lack of scatterers or instrumental problems at higher pressures.

Description: Recent observations within deep convection regimes of the Gulf of Lions and Greenland Sea all confirm the existence of small-scale plumes of only a few 100 m horizontal scale during cooling periods, in agreement with scaling arguments and non-hydrostatic modelling results. The integral effect of the plumes is that of a mixing agent rather than carrying water downward in a mean motion. It depends on the intensity and duration of the cooling how complete the mixing within the depth range of the plumes is. In the Greenland Sea, the role of the ice through brine rejection was found to be important in the preconditioning period (November - February) rather than for the deep convection itself (March) which occurred when the water was ice-free. After the convection period water masses are exchanged with the environment through baroclinic instability, causing increased deep T,S variance on a larger scale that continues to exist well into the next summer, allowing identification of previous-winter convection activity