Biological fixation of atmospheric nitrogen in the Mediterranean Sea

Nutrient concentration in the Mediterranean Sea is controlled by water exchanges through the Strait of Gibraltar and by atmospheric and terrestrial inputs. Various peculiarities in the nitrogen and phosphorus geochemical cycles are pointed out, namely a low N : P atomic ratio (6.4) in terrestrial discharges, and a budget well balanced for phosphorus (where terrestrial discharges amount to about 80% of the outflow) but apparently very deficient in nitrogen, despite a high N: P atomic ratio (22), in Mediterranean deep waters. This suggests the possibility of a surprisingly high rate of direct atmospheric N uptake by the Mediterranean ecosystem (possibly seagrasses Posidonia oceanica and pelagic bacterioplankton species). Nitrogen and phosphorus are often compared in studies of the marine biosphere. Nitrate-N and phosphate-P occur in seawater in such low concentrations that surface waters are usually depleted by biological uptake, and a possible limitation in primary production by both nutrients in turn has been considered. The observation of a typical N : P atomic ratio of about 16, the so-called Redfield ratio (Redfield et al. 1963), which has been found to be the same in marine organisms as in seawater, has helped support the assertion of similar marine cycles. The marine biosphere temporally links the fate of these two elements whose geochemical cycles are entirely different. Most N is present in the atmosphere as gaseous N2 in equilibrium with molecular N2 dissolved in the ocean. Exchanges between molecular N2 and combined forms (organic and inorganic) are achieved mainly through the processes of N2 fixation and denitrification (see Kaplan 1983; Hattori 1983). On the other hand, atmospheric transfer processes are unimportant for P, which is added to the sea by runoff from the continents and removed by particles settling onto the sediment. The average residence time of P in the ocean is about lo5 years (Froelich et al. 1982). Mesoscale studies on semienclosed seas, such as the Mediterranean, allow conspicuous differences between the behavior of P and N to be shown and permit some conclusions, particularly concerning atmospheric N fixation by marine organisms. On a smaller scale, such studies have been made by Smith and Atkinson (1984) in a confined aquatic ecosystem off the western Australian coast and by Smith (1984) in the main lagoon of Christmas Island and in the Canton Atoll lagoon. In the Mediterranean Sea evaporation exceeds precipitation and runoff. High salinity and high-density waters are produced which supply the bulk of the Mediterranean deep water, flowing out over the sill of Gibraltar, following the bottom slope. In the opposite direction, in the surface layer of the straits, light Atlantic surface water flows into the basin to compensate for the deep outflow and for the water deficit through the sea surface. The volumes of the Atlantic inflow, 53 x 1 012 m3 yr-‘, and the Mediterranean outflow, 50.5 X 1 01* m3 yr-l, are many times greater than the water deficit, 2.5 x 1012 m3 yr-l. (These values result from a critical analysis of previous papers and from consideration of heat, water, salt and potential energy budgets: Bethoux 1979.) The Mediterranean P and N budgets have been determined (Bethoux 198 1). Outflowing deep Mediterranean waters are nutrient enriched and have a typical N : P atomic ratio of about 22 (higher than the Redfield ratio), as observed in the western basin during different oceanographic cruises in the past two decades (e.g. Coste 1969; Miller0 et al. 1978; Copin-Mont&gut and CopinMontegut 1983). They have a mean inorganic P concentration of about 0.28 pmol liter-l and represent an annual P loss of 1.41 X lOlo mol yr-*. Due to tidal streams and internal waves propagating through the straits, the Atlantic inflowing layer is variable in thickness with time and space, from about 20 to 60 m (MEDIPROD 4 cruise stations: Coste et al. 1984). In the western part of the straits, the