Direct measurements of dinitrogen efflux from continental shelf sediments ~ndicated that denltnflcation (mean 3.2 mm01 N m-2 d-l) was very important ~n nitrogen cycling. Most dinitrogen came from sediment-nitrate. All ammonium produced in these sediments was probably nitrified and then denitrified. In a closed incubation, the linear production of dinitrogen, a s oxygen decreased, was unexpected as was the low rat10 of oxygen consumption to denitrification (3.6:l). Sin~ulation modelling suggests

more »

... the following explanation: Most carbon IS oxidised anoxically, but mtrogen (ammonium) diffuses to the oxygen zone relatively deep in the sediment, where hlgh rates of coupled nltrif~cat~on-denitrification result. As oxygen decreases, the zones of nltrification and denitrification move upward. The nitrate initially present In the enclosed overlying water decreases, but due to the decreasing diffusional path to the zone of denitrification, ~t s rate of denitrification remains constant. The concentration of n~trate from sediment nitrification increases in the overlylng water, but due to the decreasing rates of sediment nitrification, its rate of denitrlficatlon is also constant. KEY WORDS: Nitrification Denitrification To construct a model which simulates a particular sediment, a variety of factors must be known: quantity of degradable organic matter, its C/N ratio, its distribution in the sediment and the extent to which the reduced products of anoxic respiration are free to diffuse (Blackburn & Blackburn 1993a , b. c, Blackburn et al. 1994). In addition, the rate constants for the oxidation of carbon by 02, NOs-and SO4'-, and for the oxidation of NH4+ and SH-must be known. The values of the variables, outlined in the legend to Fig. 1 , gave quite a close representation of the data for a continental shelf sediment (Devol 1991). Dinitrogen was produced at a Linear rate (Fig. 1A) . The total denitrification rate of -3.7 mm01 N m-' d-' was similar to that reported (mean 3.2 mm01 N n1r2 d-l). In the simulation (Fig. lB) , -'E-mall: henry@pop.bio aau.dk there was a slight increase in No3-concentration, compared to the actual slight decrease (mean 1.3 mm01 m-2 d -l ) . Fine tuning of the nitrification and denitrification rates would correct this difference. During the time of N2 linear production (2 d), there was a fall in oxygen concentration from 200 to 20 FM. The initial rate of oxygen uptake was 11.9 mm01 m-' d-', very close to the reported value (11.4 mm01 m-2 d-l). As O2 decreased, accumulation of NH,' increased, due to the sediment becoming less oxidised (Fig. 1C) . The model predicted a time-dependent decrease in the rate of nitrification from -4.5 mmol N m-' d-', but denitrification (Ds) of sediment-NO3-(N03s) was relatively constant at -3 mmol N m-' d-' (Fig. ID) . Denitrification (Dw) of NO3-originally in the water (N03w), was also relatively constant at -1 mm01 N mY2 d-l, even though N03w decreased from 30 to < l 0 FM. The movement of the zone of denitrification closer to the sediment surface, as the depth of 0, penetration decreased with time ( Fig. 2) , can explain the apparent contradictions. The linear efflux of N2s (N, from N03s) was due to the N03s which accumulated in the water. The diffusional path for this Ds, back through the nitrification zone to the zone of denitrification, decreased with incubation time. This counteracted the decreased rate of nitrification and resulted in an almost constant, but slightly increasing, rate of Ds. Similarly, the decreased diffusional path for N03w to the zone of denitrification counteracted the decreased concentration of N03w and resulted in an almost constant rate for Dw. It was suggested that most of the NH,' produced in the sediment (-2.5 mm01 N m-2 d-l) (Christensen et al. 1987 ) would be nitrified and then denitrified (Devol 1991) . This was unlikely, as nitrification initially exceeded Ds (Fig. ID) , and N03s escaped to the overlying water. It is only in the situation where this N03s is trapped in water overlying the sediment surface that Ds 0 Inter-Research 1996