Total particulate phosphorus (TPP), particulate inorganic P (PIP), and particulate organic P (POP) concentrations were measured in a year-long series of sediment trap samples collected throughout the oxic-anoxic water column (275 m, 455 m, 930 m, and 1,255 (PON). The lack of a relationship between POC and PIP fluxes and the large fraction of TPP associated with the PIP pool in both oxic and anoxic traps suggests that future analyses must separate PIP and POP when evaluating biological relationships between C, N, and P. The strong relationships between POC, PON, and POP also suggest that POP is not preferentially remineralized relative to PON and POC with increasing depth in this anoxic environment. P composition was also determined using solid state 31 P nuclear magnetic resonance (NMR), and it was found that phosphonates, chemically and thermally inert compounds, are a significant fraction of the TPP pool. Furthermore, these compounds were preferentially removed relative to more bioavailable P esters during a low flux event. Their selective removal suggests that these compounds may be an unrecognized source of bioavailable P under anoxic conditions.

Sediment traps provide vital information on the magnitude and composition of sinking particles. Unfortunately little is known about the integrity of various constituents, particularly phosphorus (P) measured in these samples. We report concentrations of total, inorganic, and organic P in sediment trap particles, supernatants, and rinse water collected from both oxic (275 m) and anoxic waters (455, 930, and 1255 m) in Cariaco Basin, Venezuela. On average 30% of the total P measured in the traps was in the supernatant, with an additional 10% of total P in the rinse water. Greater than 80% of the P in the rinse water and supernatant was in the form of inorganic P, higher than the 50% to 60% of inorganic P found in trap particles. Possible sources of inorganic P to supernatants include swimmer herniation, dissolution of inorganic phases, and solublization of particulate organic P. The good agreement between particulate organic carbon (C) and organic P suggests that losses of organic C to trap supernatants must also be considered. Although fluxes were underestimated by approximately 30% when supernatant concentrations were not included, temporal and depth trends were maintained. This was confirmed by incubation experiments that suggest that P loss to supernatants occurs rapidly (< 2 weeks), potentially due to particle agitation during transport. While traps may provide insight into the temporal and spatial variability of P within sinking particles, we recommend minimal sample handling and that supernatants be analyzed to determine overall P fluxes.