ALBERT

Description: Historical data from the region between the Greenwich meridian and the African continental shelf are used to compute the offshore geostrophic transport of the Benguela Current. At 32°S, the Benguela Current is located near the African coast, transporting about 21 Sv (1 Sv = 106 m3 s−1) of surface water toward the north relative to a potential density surface lying between the upper branch of Circumpolar Deep Water and the North Atlantic Deep Watar. Two warm core eddies of probable Agulhas Current origin an observed west of the Benguela Current at 32°S. Near 30°S, the Benguela Current turns toward the northwest and begins to separate from the eastern boundary. It carries about 18 Sv of surface water across 28°S. The current then turns mainly toward the west to flow over a relatively deep segment of the Walvis Ridge south of the Valdivia Bank. A surface current with northward surface of about 10 cm s−1 flows along the western side of the Valdivia Bank, while another northward surface current flows at about 20 cm s−1 some 300 km west of the bank. About 3 Sv of surface now do not leave the Cape Basin south of the Vaidivia Bank, but instead drift northward as a wide. sluggish flow out of the northern end of the Cape Basin. Because of the more southerly seaward extensions of most of the Benguela Current, there are no deep-reaching interactions observed between this current and the cyclonic gyre in the Angola Basin east of the Greenwich meridian. Beneath the surface layer, about 4–5 Sv of Antarctic Intermediate Water are carried northward across 32° and 28°S by the Benguela Current, essentially all of which turns westward to cross the Greenwich meridian south of 24°S.

Description: We report a study of a coastal frontal zone of the southeastern United States based on a field experiment and numerical modeling. The study was conducted in the spring of 1985 during weak to moderate wind stress and strong input of buoyancy from solar radiation and river discharge. The study confirms that the structure and slope of the frontal zone depends on a combination of wind stress and cross-shelf advection of buoyancy. A cross-shelf/depth two-dimensional (x, y), time-dependent numerical model illustrated the response of the frontal zone to the local wind stress regimes. A comparison of model results with field data showed that the model successfully predicted onsets of stratification and mixing. When alongshore wind stress was negative (southward), isopycnals in the frontal zone steepened due to a combination of horizontal advection and vertical convection. When stress was positive (northward), the offshore advection of low density water flattened the isopycnals and potential energy decreased, demonstrating that horizontal advection terms are important in the equation of conservation of buoyancy. The model predicts die offshore advection of lenses of less dense water during upwelling-favorable wind stress. These lenses are of the order of 20 km in cross-shelf scale and represent an efficient mechanism to export nearshore water. The lenses consist of a mixture of low-salinity coastal water and continental shelf water originating further offshore and advected onshore along the bottom. The mean flow inside the frontal zone opposed the mean alongshore wind stress. Part of the alongshore flow was in geostrophy with the cross-shore pressure gradient; the other part was due to an alongshore pressure gradient force (kinematic) of about 1 × 10−6 m s−2 (equivalent sea surface slope = 1 × 10−7), which was trapped along the coast with an offshore width scale of O(10 km). It is likely that the alongshore extent of this pressure gradient was governed by the scale at which freshwater is injected to the continental shelf, i.e., 20–30 km. The pressure gradient force immediately outside of the frontal zone was about −5 × 10−7 m s−2 in the direction of the mean alongshore wind stress. It is hypothesized that, as a result of wind setup and freshwater influx, the northward pressure gradient forced over outer shelf/slope by the Gulf Stream decreases in magnitude onshore, and can even change sign across a nearshore frontal zone of O(10 km). The implied flow field near the frontal zone is therefore highly three-dimensional with |∂v/∂y|≈|∂u/∂x|, where (u, v) are velocities in the cross-shore (x) and alongshore (y) directions, respectively.

Description: The sensitivity of the global climate system to interannual variability of he Eurasian snow cover has been investigated with numerical models. It was found that heavier than normal Eurasian snow cover in spring leads to a “poor” monsoon over Southeast Asia thereby verifying an idea over 100 years old. The poor monsoon was characterized by reduced rainfall over India and Burma, reduced wind stress over the Indian Ocean, lower than normal temperatures on the Asian land mass and in the overlying atmospheric column, reduced tropical jet, increased soil moisture, and other features associated with poor monsoons. Lighter than normal snow cover led to a “good” monsoon with atmospheric anomalies like those described above but of opposite sign. Remote responses from the snow field perturbation include readjustment of the Northern Hemispheric mass field in midlatitude, an equatorially symmetric response of the tropical geopotential height and temperature field and weak, but significant, perturbations in the surface wind stress and heat flux in the tropical Pacific. The physics responsible for the regional response involves all elements of both the surface heat budget and heat budget of the full atmospheric column. In essence, the snow, soil and atmospheric moisture all act to keep the land and overlying atmospheric column colder than normal during a heavy snow simulation thus reducing the land–ocean temperature contrast needed to initiate the monsoon. The remote responses are driven by heating anomalies associated with both large scale air-sea interactions and precipitation events. The model winds from the heavy snow experiment were used to drive an ocean model. The SST field in that model developed a weak El Niño in the equatorial Pacific. A coupled ocean-atmosphere model simulation perturbed only by anomalous Eurasian snow cover was also run and it developed a much stranger El Niño in the Pacific. The coupled system clearly amplified the wind stress anomaly associated with the poor monsoon. These results show the important role of an evolving (not specified) sea surface temperature in numerical experiments and the real climate system. Our general results also demonstrate the importance of land processes in global climate dynamics and their possible role as one of the factors that could trigger ENSO events.

Description: We examine the diffusive behavior of the flow field in an eddy-resolving, primitive equation circulation model. Analysis of fluid particle trajectories illustrates the transport mechanisms, which are leading to uniform tracer and potential vorticity distributions in the interior of the subtropical thermocline. In contrast to the assumption of weak mixing in recent analytical theories, the numerical model indicates the alternative of tracer and potential vorticity homogenization on isopycnal surfaces taking place in a nonideal fluid with strong, along-isopycnal eddy mixing. The eastern, ventilated portion of the gyre appears to be sufficiently homogeneous to allow the concept of an eddy diffusivity to apply. A break in a random walk behavior of particle statistics occurs after about 100 days when along-flow dispersion sharply increases, indicative of mean shear effects. During the first months of particle spreading, eddy dispersal and mean advection are of similar magnitude. Eddy kinetic energy is of O(60–80 cm2 s−2) in the model thermocline, comparable to the pool of weak eddy intensity found in the eastern parts of the subtropical oceans. Eddy diffusivity in the model thermocline (Kxx = 8 × 107, Kyy = 3 × 107 cm2 s−1) seems to be higher by a factor of about 3 than oceanic values estimated for these area. Below the thermocline, model diffusivity decreases substantially and becomes much more anisotropic, with particle dispersal preferentially in the zonal direction. The strong nonisotropic behavior, prominent also in all other areas of water eddy intensity, appears as the major discrepancy when compared with the observed behavior of SOFAR floats and surface drifters in the ocean.

Description: A coupled ocean-atmosphere general circulation model has been developed for TOGA related problems. The coupled model consists of an ocean model of the tropical Pacific and a global low-order spectral atmosphere model. The two models interact via wind stress and sea surface temperature. In order to avoid a climate drift within the coupled model, a flux correction method is applied.Experiments were performed by introducing a westerly wind stress burst over the western equatorial Pacific for one month. Thereafter, the wind burst is turned off and the response of the coupled model to the initial disturbance is investigated. The results are compared with the response of the ocean model run with the same disturbance in an uncoupled mode.It is shown that the coupling leads to a significant increase of the duration of anomalous conditions in the ocean. SST anomalies persist for about 12 months in the coupled run, while they have already disappeared after 4 months in the uncoupled case. The increase in persistence is due to the feedback of the atmosphere, which responds with an eastward shift of the ascending branch of the Walker Circulation.In a second experiment with the coupled model the initial disturbance was introduced within another season. The results show no basic differences to the results of the first experiment.An interesting result of the coupled model runs is the occurrence of spontaneous westerly wind bursts over the western Pacific, which developed by internal dynamics. Location and duration of these spontaneous wind bursts show some correspondence with the time-space structure of observed westerly wind stress episodes over the western Pacific.

Description: The existence of energetic anticyclonic mid-depth vortices of Mediterranean Water (meddies) questions the validity of a conventional advective–diffusive balance in the eastern Atlantic subtropical gyre. A mesoscale experiment in the Azores–Madeira region reveals a link of these meddies to large-scale subsurface meanders. For the first time it is shown that meddies may have strong surface vorticity, indicative of a generation process involving the Azores Current—a deep reaching near-surface jet.

Description: Quasi-homogeneous layers in vertical profiles of temperature and salinity in the eastern North Atlantic near Madeira indicate the existence of a subtropical Mode Water in the Eastern Basin. Temperature sections show a maximum horizontal extent of at least 500 km. The frequency distribution analysis of homogeneous layers in a historical XBT dataset shows a Mode Water formation region near and to the north of Madeira. This Mode Water is found at increasing depths and displaced to the west and southwest during the course of the year after its formation by wintertime convection. It disappears almost completely, due to mixing, before the next winter. Volume estimates suggest that this Madeira Mode Water in the eastern Atlantic accounts for 15–20% of the total Central Water formation in the corresponding density range as obtained from tracer studies in the North Atlantic gyre.

Description: A primitive equation model of the equatorial Pacific Ocean was forced by realistic wind stress distributions over decades. Results were presented for a set of two experiments. In the first experiment the model was forced by an objectively analyzed wind field, while for the second experiment a subjectively analyzed wind field was used. The results indicate a strong sensitivity of the model to the choice of the wind fields. Especially, model results in the eastern Pacific show big differences between the two model runs. Taking the results of the second model run the performance of the model with respect to interannual variability is investigated. Sea level, temperature and zonal currents show pronounced interannual variations within the equatorial belt from 10°N to 10°S. Special attention is given to the simulation of the 1982/83 El Niño event. The model reproduces most of the basic features, which were observed during this El Niño event. In particular the deceleration of the equatorial undercurrent, the evolution of eastward surface currents and the zonal redistribution of heat associated with an eastward propagation of warm water are simulated by the model.

Description: Data from a surface mooring located in the Sargasso Sea at 34°N, 70°W between May 1982 and May 1984 were compared with satellite data to investigate large diurnal sea surface temperature changes. Mooring and satellite measurements are in excellent agreement for those days on which no clouds covered the site at the time of the satellite pass. During the summer half-year at this site, there is a 20% charm of diurnal warming of more than 0.5°C, with values of up to 3.5°C observed in the two-year period. Diurnal warming observed at the mooring has been simulated well by a one-dimensional model driven by local beat and momentum fluxes. Under the conditions of very light wind and strong insolation that produce the Largest surface warming, the surface mixed-layer depth reduces to the convection depth, and wind-mixing becomes unimportant. The thermal response is then limited to depths between 1 and 2 m, making it likely that such events have been underreported in routine ship observations. In all cases observed, the spatial extent of warming events as determined by satellite data are well correlated with the corresponding atmospheric pressure patterns. Conditions giving rise to the largest diurnal warming events are often associated with a westward-extending ridge of the Bermuda high. In the region studied, 57°–75°W and 29°–43°N, diurnal warming of more than 1°C was found on occasion to cover areas in excess of 300 000 km2, with warming of more than 2°C coveting areas in excess of 130 000 km2.

Description: An XBT interface is described for use with Commodore and other 6502 based microprocessors. This interface takes the form of a single circuit board mounted inside the microcomputer and is completely software controlled. The application of this digital XBT system to the real-time computation of density and dynamic height, using historical or recent temperature-salinity relationships, is also described. Comparison between XBT and CTD measured temperatures from the Northeast Atlantic yield a mean temperature difference of −0.08°C and an rms temperature difference of 0.33°C for the upper 800 m. Examples of dynamic topography maps and a temperature section computed using this technique are also presented and comparison between objectively analyzed XBT and CTD dynamic topographies demonstrates the reliability of the method for mapping the baroclinic flow.

Description: The inclination of oceanographic mooring lines due to current drag causes errors in time series observations of currents and temperatures. The prediction of this effect requires knowledge of the drag coefficients for the mooring components. Drag coefficients, known for simple geometric shapes such as spheres or cylinders, are commonly used for mooring response computations. Selected mooring components (buoyancy elements and instruments) were tested in a tow tank to determine their actual drag coefficients. Over the Reynolds Number range, typical of oceanic conditions, deviations of the drag coefficient up to 50% are found when compared with the appropriate simple geometric shape coefficients. A set of model moorings and model current profiles is used to determine the resulting changes in component depth level and displacement. The changes in horizontal displacement of the upper part of the mooring are on the order of 10% in extreme cases and 1% under typical conditions. Their effects on current measurements will usually be negligible. However, the related vertical displacements are on the order 100 to 10 m. Such vertical displacements may carry instruments to depth levels where currents and particularly thermocline temperatures are sufficiently different from the intended level to cause errors in the time series observations.

Description: Simulated transient-tracer distributions (tritium, 3H3, freons) on the isopycnal horizons σ0=26.5 and 26.8 kg m−3 are presented for the East Atlantic, 10° −40°N. Tracer transport is modeled by employing a baroclinic flow field based on empirical data in a kinematic isopycnal advection-diffusion numerical model, in which winter convection is taken as the mechanism of communication with the ocean surface layer, and the isopycnal diffusivity is a free parameter. Diapucnic transport is ignored. The simulations employ time-dependent tracer boundary conditions, which are constructed on the basis of available observations. Simulations are compared to data obtained on a meridional section in 1981 (F/S Meteor, cruise 56/5). Best simulations were obtained by means of a subjective optimization procedure. On both levels, the observed distributions and the best simulated distributions agree well. The fact that the surface boundary conditions and interior distributions of the tracers are distinctly different leads us to the conclusion that our model provides a consistent description of upper main-thermocline ventilation and interior transport Surface-water densities in February are found to represent adequately the winter outcrop boundaries with an uncertainty of about ±300 km across. The required isopycnal diffusivity south of 29°N is 1700 m2 s−1, and 2900 m2 s−1 further north (+70/−40%). Interior transport is found to be predominantly advective. Advective ventilation across 30.5°N east of 33°W amounts to only 12% and 40% for the 26.5 and 26.8 horizons of the total ventilation rates reported by Sarmiento. The North Atlantic/South Atlantic Central Water boundary near 15°N is found to be predominantly determined by advection.

Description: Long-term temperature and current-meter records from moorings in the northern Canary Basin display strong current events with time scales between one and three months and large vertical scales of several thousand meters. The data are compared to hydrographic surveys in the area that show a meandering subtropical front. The strong current events are found to be related to the passage of the front through the mooring positions. An analysis of composite time series, for selected depths, indicates cases of westward and of eastward propagation of frontal meanders. The frontal pattern is also found in geopotential anomalies inferred from historical XBT data sets, suggesting that the front is a persistent feature of the density field. In two cases strong current events appear to be related to a Mediterranean Water lens.