Seasonal variations in pools of porewater NH,' and exchangeable NH,+ desorbed by l-step [easily desorbed NH,') and multiple KC1 extractions (tightly bound NH,') were investigated in sediments from 3 stations in Danish coastal waters. One-step KC1 extractions gave higher yields in winter and lower yields in periods of increased sediment activity and temperature. The exchangeable NH,' pool obtained by l-step extraction was always less than the total pool removed by the multiextraction technique, independent of sediment type. The multi-extraction technique proved the existence of sediment NH,' pools with different exchange capacities. There was evidence to indicate that the tightly bound pool changed seasonally and entered the normal N cycle. The size and dynamic nature of this pool have not previously been recognized: maximum and minimum integrated values for Aarhus Bay, Norsminde Fjord and Randers Fjord sedirnents were 32, 36, 25 and 4, 9 and 6 mm01 m-2, respectively. A lack of correlation between pool sizes of easily desorbed NH,' and porewater NH,' was not compatible with literature values of the NH,' adsorption coefficient (K-value). K-values varied with station, depth and season (means of all values 3.7 and 11.7 by l-step and multiple extractions, respectively), and were in general higher than the literature values.

Abstract. The release of NH: and 'S~-labelled NH: by one-step KC1 extraction was assessed in different types of coastal marine sediments. KC1 was efficient to extract NH: from sandy sediments and less efficient in silt sediments, where an extendedextraction period was required for obtaining a maximum NH: yield. Extraction at 0 or 20 OC had only a little effect on the efficiency of KCl. KC1 gave always non complete recovery of 1 5~H $ in silt sediments. However, the added label could be fully recovered by addition of 80 pmol.cm-3 exogenous NH: prior to KCl, or when NaCl or ASW replaced KCl. "NH: was added to non-biological silt sediment, which was incubated at 0 OC up to 16 hours, to see the effect of physical processes on the partition of 'IN& among porewater $ 2 9 4 9 8 ) exchangeable (9-308) and non-extractable, organic bound pools (24-42%). Total ' N recovery was approximately 100%. KC1 failed to remove 15NH4 which entered to unknown, bound pools in sediment. Only shortly after addition of "N (0.1 h), the extraction period resulted in significantly different 15N recoveries (P < 0.05) in KC1 extractable NH: , 17% versus 9% of label was recovered after 1 min or 60 rnin extraction of sediment, respectively. Two hours of incubation time were required for complete equilibrium of "NH: among porewater, exchangeable and organic bound pools. Sediments (silt) to which ' $NH: has been added in order to measure NH: turn-over and KC1 is used as extractant, should be incubated for at least 2 hours, before taking a zero-time sample.