THE MTP SEDIMENT TRAP EXPERIMENTS: HIGHLIGHTS ON PARTICLE FLUXES THROUGH THE EUTROPHIC AND OLIGOTROPHIC SYSTEMS OF THE MEDITERRANEAN SEA <B>THE MTP SEDIMENT TRAP EXPERIMENTS: HIGHLIGHTS ON PARTICLE FLUXES THROUGH THE EUTROPHIC AND OLIGOTROPHIC SYSTEMS OF THE MEDITERRANEAN SEA</B> <HR><P> Heussner, S. & Monaco, A. <BR> and the EUROMARGE and PELAGOS groups <BR> Laboratoire de Sédimentologie et Géochimie Marines, CNRS URA 715, Perpignan, France<P><P> Introduction</B><P><P> Among the various studies necessary to improve our knowledge of global cycling in marine ecosystems, those dedicated to understand the complex processes controlling the fluxes of materials and associated elements occupy a central position. Biological, physical and geochemical activities transform large quantities of surface-produced and land-derived matter, before distributing it to depth and ultimately to sediments. Being characterized by contrasting systems with different structure, size, water circulation (with respect to sediment transport pathways) and inputs, the Mediterranean Sea offers an attractive system to study the interactions between such activities in the cycling of biogeochemicals and other materials on the continental shelves and how they allow transport of biogeochemicals into and within the deep water basins.<P><P> The Mediterranean Sea may be considered as a "marginal basin", which means that a much larger proportion of its area, relative to ocean basins, is dominated by physical processes that characterize the margins of larger oceans. Continental margins tend to be areas with large standing stocks of particulate carbon and related elements and increased rates of primary production. This production is fuelled by the rapid recycling of nutrients as well as the proximity of shelves to both continental and basin nutrient sources. A significant fraction of primary production can be exported from continental shelves and surveys on carbon, various inorganic materials, biochemicals and radiochemicals in the shelf and slope sediments appear to support this notion of a carbon and biochemical rich depocentre on the continental slope (e.g., Buscail <I>et al</I>., this volume). In the Northwestern Mediterranean the continental shelf is often narrow except off the two main rivers Rhône and Ebro; hence, short residence times of particles are likely. In contrast the dominant shelves of the Eastern Mediteranean, the Adriatic and North Aegean Seas are broad and either baroclinic controls or the existence of coastal fronts allow for much longer transport pathways; hence there is the likelihood of greater organic matter degradation before deposition.<P><P> Another important reason that makes the Mediterranean attractive for oceanographers is the variety of oligotrophic regimes, which show a progressive eastward gradient. North-south trophic gradients also occur, especially in the Northwestern Mediterranean, Adriatic and Aegean Seas, which are largely driven by nutrients, either as direct river inputs (e.g., Rhône and Po) or indirectly by the inflow of Black Sea waters via the Dardanelles into the North Aegean Sea.<P><P> Data concerning the circulation, transport pathways and fluxes of biogeochemicals within these systems are clearly needed. The Mediterranean Targeted Project offered the first opportunity to provide such data at the scale of the entire basin. This paper intends to summarize some major flux results and to highlight the main conclusions of several simultaneous MTP trap experiments which were performed using the same experimental strategy.<I><P><P> The MTP Sediment Trap Experiments</I><BR> The experiments which are presented here were performed within three MTP subprojects: EUROMARGE-North Balearic in the Northwestern Mediterranean, between Marseilles and the Balearic Islands (5 mooring sites), EUROMARGE-Adriatic Sea in the Southern Adriatic trough (1 site), and PELAGOS in the Kythira Strait west of the Cretan Island (2 sites)(figure). A more thorough description of these trap experiments and detailed results are given in this volume or elsewhere (e.g., Balopoulos; Canals <I>et al</I>.; Miserocchi <I>et al</I>.; and the annual scientific reports of the three subprojects). In terms of experimental conditions, it is important here to know that we used identical tools and applied the same field strategy in all three experiments: long term monitoring by means of Aanderaa current meters and Technicap PPS3 sediment traps between 1993 and 1995.<P><P> The mooring lines were of identical design and were deployed at 500 and 1000 m depth in well-identified hydrodynamical features: intermediate waters of the general thermohaline circulation in the North Balearic (within the path of the Northern Current) and Adriatic sites and on either sides of the Western Cretan Strait, i.e., in the North Cretan Sea and in the Ionian basin. Each line was equipped with a trap/current meter pair at 35 meters above bottom (mab). For the deeper moorings (around 1000 m) a second instrument pair was set at around 500 mab.<I></I><P><P> To facilitate result comparison on a sample-to-sample basis, traps were programmed to collect settling particles over a 2-week interval starting at the first and 16th day of each month. Altogether some 17 lines with 24 trap/current meter pairs were deployed (6-month deployment) for 1 to 2 years. Due to logistical reasons as well as instrument availability, the experiments did not start at the same time and no common period for all sites was achieved. For most sites, nevertheless, traps were deployed simultaneously for at least 6 months between spring and autumn 1994. We therefore choose to present results from this period. Results from the Catalan margin (off Barcelona) relate to the 93 summer period, whereas those from the Balearic margin relate to summer 95. A check of the longest flux record (2 years; Gulf of Lions, off Marseille and Banyuls) showed that interannual variability was not very pronounced and we thus assume that the data sets obtained through different years are consistent, at least for the level of comparison needed here.<P><P> All laboratories involved in these experiments (LSGM, Perpignan; ICM and UB Barcelona; IGM, Bologna; NCMR, Athens) shared a common trap sample preparation and treatment (derived from Heussner <I>et al</I>., 1990) as well as common analytical procedures. All scientists concerned were trained at Perpignan for that purpose. This represents already a first achievement of the MTP by having succeeded in the transfer of trap expertise and know-how between different laboratories.<I><BR> Major Flux Results from the MTP Trap Experiments</I><BR> Total mass fluxes observed at the various sites and averaged over the 6-month summer period are given in the figure. We present here flux results from the 1000 m mooring lines. The 500 m lines, which were deployed only by EUROMARGE-NB, followed the same trends, the main difference lying in increased mass fluxes.<P><P> If we consider these results in their broadest sense, we can see the existence of two flux gradients that match the trophic degree. A west-east decreasing gradient exists at the scale of the entire Mediterranean basin, which roughly corresponds to the increase of the degree of oligotrophy. At the subbasin scale, a second flux gradient appears -- decreasing from north to south -- as illustrated by the total mass fluxes registered in the Northwestern Mediterranean (figure).<P><P> Environmental conditions can eventually be expressed by the composition of settling material, in particular the organic and inorganic carbon contents (table). Surprisingly, no major differences were found between the different experimental sites. At the most, only the inorganic carbon contents can be considered as being slightly higher in the eastern part of the Mediterranean. As a consequence, fluxes of carbon, either organic or inorganic, essentially matches the overall variability of total mass fluxes. When expressed in broad terms of carbon transfer, oligotrophy therefore expresses itself in a strong quantitative reduction of fluxes rather than in terms of modification of the quality of settling material.<I><P><P> Factors Controlling the Temporal and Spatial Variabilities of Particle Transfer</I><I><P><P> Spatial variability</I><BR> Besides this large scale variability which is related to the degree of trophy (or at least covaries with that degree), fluxes did also vary at the meso- and local scales. This is evidenced for example on the Northwestern Mediterranean margin. Mass fluxes progressively increased from the entrance of the Gulf of Lions system in the NE (off Marseilles) down to the Catalan margin (off Barcelona) in the SW, that is in the direction of the general alongslope circulation of the water masses (Northern Current). This is a clear example of hydrodynamical forcing of particle transfer.<P><P> In the same way, the low fluxes recorded in the mid-Adriatic trough can be explained by the location of the site with respect to the riverine source (Po river) and the cyclonic circulation of the Adriatic basin. A large fraction of the particles are trapped on the continental shelf, whereas the remainder escapes the Northern Adriatic and follows the western border of the basin before being ultimately exported through the Otranto strait.<P><P> The small scale variability (i.e., km range) has been studied in the North Balearic basin, where a narrow net of observations has been applied. The mass fluxes registered over a 2-year period within and outside the canyons shows that particle transfer is enhanced within the canyon axes by a factor of 2 compared to the adjacent open slopes. This feature, already observed during previous experiments, is therefore confirmed. It stresses the incidence of topography on near-bottom circulation and underlines the role of submarine valleys in channelling mass and energy transfers to the deep basins.<I><P><P> Temporal variability</I><BR> The most obvious temporal signal which has been recorded in all three experiments is a seasonal one. It is directly related to the Mediterranean climate and the functioning of the river network on the northern border. The highest fluxes are generally recorded during late autumn and winter, that is during the rain season, when the solid river discharge is the highest. This is especially true for the large rivers in the western or central part of the Northern Mediterranean (Ebro, Rhône and Po), but also for smaller coastal rivers such as in the Eastern Mediterranean. At that period -- particularly during the period of water homogenisation -- the particulate load of the water column is increased. The bulk of settling particles presents a clear continental origin, which does not necessarily imply a direct transfer. More likely, particles introduced into the marine system undergo several deposition/resuspension events before being transferred to depth on the slope and deep basins. On the contrary, the lowest fluxes are generally observed during summer. Again, this has been observed in all three sites. Reduction of downward fluxes can be related to reduced continental inputs, less resuspension events or seasonal weakening of the water mass circulation. The extreme situation was observed in the Eastern Mediterranean where the lowest mass fluxes ever recorded (around 1 mg m<SUP>-</SUP><SUP>2</SUP> d<SUP>-</SUP><SUP>1</SUP>) coincided with a strong stratification of water masses.<P><P> Finally, superimposed on these regular seasonal signals, individual events are also registered. Occasional flux peaks can be related to storms that strongly influence shelf resuspension. This is another example of climatic control, that is external forcing. The different experiments gave also clear examples of internal control generally related to hydrodynamic forcing. Some peaks in mass flux can be related to turbidity currents for example or to resuspension events induced by internal waves, as, for example, the strong mass fluxes recorded at the canyon heads in the northwestern basin. In the Adriatic, a strong summer peak corresponded to an increased current in the shelf-slope-basin direction. And, last example, the marked summer flux peak observed at 1200 m depth on the external site of the Kythira Strait could have been provoked by turbidity currents, since very strong near-bottom currents along the canyon axis (up to 40 cm s<SUP>-</SUP><SUP>1</SUP>) were recorded at that period.<B><P><P> Conclusions</B><P><P> The first coordinated flux experiment conducted during MTP 1 at the scale of the entire Northern Mediterranean provided some interesting insights into the main features of particle transfer processes. Though largely incomplete, these results underline the interest of conducting large scale 3D experiments. Hydrodynamical forcing, but also climate (rainfall, winds) and geological framework (lithology of the surrounding continents, topography, structure of the basins) determine the functioning of the Mediterranean Sea. Mass and energy transfer present east-west and north-south gradients which match the major trophic gradients of the Mediterranean. Quantity and quality, but also spatial variability of downward fluxes of settling particles will depend upon the location of particle sources with respect to the general geostrophic circuits. Eventually they are further modulated by the interaction between currents and topography. Changes in the input terms by marine production, which are related to the degree of oligrophy of the various subbasins, are only slightly affecting the overall elemental composition of settling particles exported to depth. Seasonal control therefore essentially relies upon factors that induces changes in the rate of continental inputs to the Mediterranean system.<P><P>