Search

International audience; An improved estimation of the surface currents in the Levantine Basin of the Mediterranean sea is crucial for a wide range of applications, including pollutants transport and nutrients distribution. This estimation remains challenging due to the scarcity or shortcomings of various data types used for this purpose. In this paper, we present an objective validation of a variational assimilation algorithm that blends geostrophic velocities derived from altimetry, wind-induced velocities, and drifter positions, to continuously obtain velocity corrections. The assessment of the validation impact was based on available independent in-situ data (current meters, gliders, and independent drifters) and satellite ocean color images. In all cases, the improvement was shown either qualitatively (position of the eddies) or quantitatively.

International audience

Hydrological data collected in the Strait of Gibraltar have been used to examine the distribution and spatial–temporal evolution of the water masses in the area. The spatial variability has been addressed by means of a clustering method that determines the affinity of a collection of temperature–salinity samples to one of the water masses involved in the exchange. The method, which has been applied to a nearly-synoptic data set, highlights the clear evolution of the Mediterranean Waters as they flow westward through the Strait. While up to four different Mediterranean Waters are spatially distinguishable east of the main sill of Camarinal in the Strait, most of their differentiating characteristics are eroded after flowing over this restrictive topography due to mixing. West of the sill, therefore, speaking of a unique Mediterranean Water seems more appropriate. The same applies to the North Atlantic Central Water flowing in the opposite direction, which is noticeably modified along its path to the Mediterranean Sea, most of its transformation taking place in the Camarinal sill surroundings. A series of repeated transects carried out in the eastern and western sides of the Strait, provided a temporal analysis of the water masses evolution: the temporal variability manifests seasonality in the surface waters, while interannual signal is mainly detected in the deeper water masses. It is worth remarking the statistically significant positive trend of Western Mediterranean Deep Water (0.009 °C/year) and Winter Intermediate Water (0.03 °C/year), with the latter showing also intermittent occurrence in the Strait.

The circulation of the Levantine Intermediate Water (LIW) in the Algerian subbasin (western basin of the Mediterranean sea) has been much debated for more than fifteen years now. Together with the old circulation diagrams, several numerical models claim that a branch of LIW is permanently flowing westwards across the Algerian subbasin, i.e. directly from the Channel of Sardinia towards the Strait of Gibraltar. Only a few models support the fact that the unique continuous flow of LIW is structured as an alongslope counterclockwise vein, which is thus directed northwards off Sardinia in the Algerian subbasin, and hence support the diagram published by Millot in 1987 [Millot, C. (1987a) Circulation in the Western Mediterranean. Oceanologica Acta 10(2), 143–149]. According to this diagram, any little mixed LIW found in the central subbasin corresponds to fragments which have been pulled away from the vein and entrained there by mesoscale eddies originated from the Algerian Current. The ELISA experiment (1997–1998), as a follow-up of other ones conducted since about 15 years, was designed partly to validate the diagram. In addition to about 40 current meters set in place for one year, four main campaigns were conducted with a sampling strategy guided in real time by infrared satellite information. The data set we present clearly provides additional evidence that the little mixed LIW found in the central Algerian subbasin has been entrained there by the mesoscale eddies and not by a permanent westward flow.

The relationship between mesoscale hydrodynamics and the distribution of large particulate matter (LPM, particles larger than 200 Am) in the first 1000 m of the Western Mediterranean basin was studied with a microprocessor-driven CTD-video package, the Underwater Video Profiler (UVP). Observations made during the last decade showed that, in late spring and summer, LPM concentration was high in the coastal part of the Western Mediterranean basin at the shelf break and near the continental slope (computed maximum: 149 A gCl 1 between 0 and 100 m near the Spanish coast of the Gibraltar Strait). LPM concentration decreased further offshore into the central Mediterranean Sea where, below 100 m, it remained uniformly low, ranging from 2 to 4 A gCl 1 . However, a strong variability was observed in the different mesoscale structures such as the Almeria–Oran jet in the Alboran Sea or the Algerian eddies. LPM concentration was up to one order of magnitude higher in fronts and eddies than in the adjacent oligotrophic Mediterranean waters (i.e. 35 vs. 8 A gCl 1 in the Alboran Sea or 16 vs. 3 A gCl 1 in a small shear cyclonic eddy). Our observations suggest that LPM spatial heterogeneity generated by the upper layer mesoscale hydrodynamics extends into deeper layers. Consequently, the superficial mesoscale dynamics may significantly contribute to the biogeochemical cycling between the upper and meso-pelagic layers. D 2002 Elsevier Science B.V. All rights reserved.

The Algerian Current (AC) is unstable and generates mesoscale meanders and eddies. Only anticyclonic eddies can develop and reach diameters over 200 km with vertical extents down to the bottom (∼3000 m). Algerian Eddies (AEs) first propagate eastward along the Algerian slope at few kilometers per day. In the vicinity of the Channel of Sardinia, a few AEs detach from the Algerian slope and propagate along the Sardinian one. It was hypothesized that AEs then followed a counter-clockwise circuit in the eastern part of the basin. Maximum recorded lifetimes were known to exceed 9 months. Within the framework of the 1-year Eddies and Leddies Interdisciplinary Study off Algeria (ELISA) experiment (1997–1998), we exhaustively tracked two AEs, using mainly an ∼3-year time series of NOAA/AVHRR satellite images. We show that AEs lifetimes can near 3 years, exceeding 33 months at least. We also confirm the long-lived AEs preferential circuit in the eastern part of the Algerian Basin, and specify that it may include several loops (at least three).

The biomass and production of phytoplankton and bacterioplankton was investigated in relation to the mesoscale structures found in the Algerian Current during the ALGERS'96 cruise (October 1996). Biological determinations were carried out in three transects between 0°and 2°E aimed at crossing a so-called event, formed by a coastal anticyclonic eddy associated with an offshore cyclonic eddy to the west. The concentration of chlorophyll a (Chl) was maximum (> 1.2 mg m-3) within the Cyclonic eddy and at the frontal zones between the Modified Atlantic Water (MAW) of the Algerian Current and the Mediterranean waters further north. Chl (total and > 2 μm) was significantly correlated with proxies of nutrient flux into the upper layers. Autotrophic picoplankton and heterotrophic bacterial abundance and production presented clear differences between MAW and Mediterranean water, with higher values at those stations under the influence of the Algerian Current. In general, greater differences were observed in production than in biomass variables. The photosynthetic parameters (derived from P-E relationships) and integrated primary production (range 189-645 mg m-2 d-1) responded greatly to the different hydrological conditions. The mesoscale phenomena inducing fertilization caused a 2 to 3-fold increase in primary production rates. The relatively high Values found within the cyclonic eddy suggest that, although short-lived in comparison with anticyclonic eddies, these eddies may produce episodic increases of biological production not accounted for in previous surveys in the region, We thank the people on board the R/V Hespérides for their help during the cruise. We are also grateful to Dolors Vaqué, Jordi Font, Belén Martı́n-Mı́guez and two anonymous reviewers for helpful criticism and comments on a previous version of the manuscript. X.A.G.M. acknowledges the receipt of a FPI pre-doctoral fellowship from the Spanish Ministry of Education and Culture. This work was supported by grants AMB95-0901-C02 (Spanish CYCIT) and MAS3-CT96-0051 (European Union MAST Programme, MATER Project)

The Algerian current is compared to a laboratory current created in a rotating tank. The experimental current flows along a regular wall and over a motionless denser water, with a constant flow rate and a uniform velocity at the source. Data collected in the laboratory are compared with satellite and in situ data sets. The various characteristic phenomena are shown to be similar: both currents are unstable and generate meanders and cyclonic–anticyclonic eddy-pairs. Eventually, the anticyclone remains and is associated with a marked lowering of the interface. The phase speed and growth of the meanders and the shape and size of the eddies in the laboratory are in quite good agreement with those deduced from in situ data and satellite images. The observations and the velocity and interface height measurements in the laboratory allow a better understanding of the Algerian current dynamics and its instability process.

During the Thetis-2/MAST-2 tomography experiment, T7-XBT calibrated (accuracy ∼0.05°C) probes were launched ∼28 km apart between France and Algeria, twice a month from Feb. to Sep. 1994. Combined with infrared images, altimetric data and ship drifts, they provide definite information on the structure, drift and role of the eddy-like mesoscale phenomena generated by the Algerian Current instability. When embedded in this alongslope current, these phenomena generally propagate downstream at a few km/day and are markedly asymmetrical. Because of the topography in the eastern part of the Algerian Basin, they separate from the current, become more symmetrical and follow an anticlockwise circuit in the open basin. These phenomena are deeper than ∼750 m and entrain seaward pieces of the Levantine Intermediate Water (LIW) vein flowing along the Sardinian slope, thus being responsible of the large spatial and temporal variability of the LIW distribution in the open basin. The non-existence of a LIW vein flowing westward across the Algerian Basin is definitely demonstrated. In the Gulf of Lions, new insights are provided into the formation and spreading of the Winter Intermediate Water (WIW), which is the WesternMediterranean counterpart of LIW. Considering the large amount of WIW formed during this mild winter, it is clear that this water has not received enough attention yet, and is certainly a major component of the Mediterranean outflow at Gibraltar. Finally, the XBT data account for the eastward flow of the WesternMediterranean Deep Water (WMDW) off Algeria.

Within the framework of the Tunisian – EC SALTO/AVICENNE project, four hydrological sections were performed between Tunisia, Sardinia and Sicily with a 10–20 km sampling interval in April–June 1995. To our knowledge, it is the first time that sections were repeated there at such a high frequency. These data significantly increase the number of observations available on the Tunisian side of the Channels of Sardinia and Sicily, and allow reliable specification of the hydrological characteristics of the area. For the first time, and as confirmed by infrared satellite images, these data sets show (i) the large mesoscale variability of the surface flow entering the study area, as expected from the characteristics of the Algerian Current upstream, (ii) the rapid changes it undergoes around northeastern Tunisia, as it shifts roughly from southward to eastward within less than a few weeks. These data also show a large heterogeneity at intermediate levels, clearly associated with waters of very different origins that follow converging routes and mix in both the study area and the southern Tyrrhenian Sea. They finally document the fundamental importance of bathymetric features of relatively small scale such as the Skerki passage.

Eight moorings were deployed off Algeria (1–5°E) during the Mediprod-5 experiment (June 1986–March 1987). The 24 current meter time series recorded at nominal depths of 100, 300, 1000 and 2000 m are analysed together with hydrological data (May–June 1986) and satellite infrared images. As expected, the circulation features are markedly different inside and outside of a ∼ 50 km-wide coastal zone. At ∼ 25 km from the coast, five out of six moorings are well within the Algerian Current and the current profile is strongly sheared, with low correlations at depth. All water masses flow eastward along the Algerian slope, thus completing consitently our circulation diagrams. At ∼ 75 km, the currents are more correlated between 300 and 2000 m and more dependent on the occurrence of mesoscale (100–200 km) anticyclonic eddies called “open sea eddies”. An event was recorded, propagating eastward at ∼ 3 km/day across the mooring array. It is thought to consist of a meander (width 50–150 km) of the Algerian Current, extending to ∼ 100 km from the coast and associated with teo superimposed anticyclonic eddies. One eddy, enclosed within the meander, involved the surface layer and had the infra-red signature of what we previously called a “coastal eddy” (30–120 km apparent diameter). The other eddy seemingly involved the whole deep layer, rapidly became barotropic and large in diameter (up to ∼ 150 km) which made it coastal too. Sooner or later, both coastal eddies are expected to merge together. These measurements have slightly modified our former hypotheses as, instead of assuming that “open sea eddies” are old stages of coastal surface eddies becoming larger and deeper, we now expect them to be old stages of the merged coastal eddies. This new understanding of such a coastal event is more similar to an open sea eddy and seems consistent with both theoretical models and laboratory experiments. Whatever this structure, recent data support our former hypotheses that mesoscale eddies generated by the Algerian Current can have a deep extent, propagate along the Algerian and then Sardinian slopes (where they entrap Levantine Intermediate Water) and strongly influence the circulation of all water masses.

This paper synthetizes various previous analyses of the Algerian eddies. These mesoscale anticyclonic eddies are of primary importance to the circulation of all the water masses in the Western Mediterranean. They are generated by instability processes disturbing the flow of Modified Atlantic Water (MAW) shortly after it reaches the African coast near 0°E, namely the Algerian Current. Young eddies, which have diameters of 50–100 km, and the upwelling generally attached to their southwestern edges, drift eastward along the coast at a few km per day, sometimes during several months; as they grow older and larger, the most vigorous ones leave the coastal zone and drift seaward. Some of them may reach the Sardinian continental slope where a well-defined vein of Levantine Intermediate Water (LIW) flows; we suspect them to be capable to pull fragments of LIW seaward, and in fact, the least modified LIW we encountered in the middle of the Algerian Basin had the form of rings or filaments trapped by one eddy. Eddies several months old may have diameters as large as ≈ 200km; they also entrain the Mediterranean Deep Water (MDW), and they probably extend down to the bottom. In 1984, two such eddies occupied most of the Algerian Basin; the MAW was deflected seaward by the westernmost one, from the Algerian coast at≈ 3°E as far as the Balearic Islands, and was then dispatched by other eddies through the whole basin as though by a set of paddle-wheels. Such interactions are perhaps not as exceptional as was previously expected. These mesoscale phenomena have strong biological implications. The Western Mediterranean Sea thus appears to be a very suitable place for the observation and modelling of such mesoscale coherent structures.