Thorium isotopes as indicators of particle dynamics in the upper ocean: results from the JGOFS North Atlantic Bloom experiment

Abstract

Measurement of 234Th and 228Th in suspended and sinking particles made during the 1989 JGOFS North Atlantic Bloom Experiment permit estimation of the rates of particle cycling. Using a simple model of thorium-particle interactions applied to water column and floating sediment trap data at 150 and 300 m, the rate constant, β2, for aggregation of small suspended particles into large rapidly sinking (∼150 m day−1) particles increases from ∼0 to ∼30 y−1 over the course of the bloom. The rate constant for disaggregation of sinking particles, β−2, similarly increases from ∼100 to ∼300 y−1 over the same period. These suggest that small particle residence times (relative to packaging or aggregation) decreasesto ∼15 days and that large particle residence times (relative to disaggregation) decrease to ∼1 day as the bloom progresses. Late in the bloom, particles are cycled such that aggregation of suspended particles (∼2 μg 1−1 day−) is comparable to particle break-up (∼3 μg 1−1 day−). Errors on the rate constants, calculated by propagating estimated errors on the individual terms in the model, are large and arise principally from uncertainty in the gradient in activity and mass fluxes between the two trap depths. However, the values calculated independently from the two tracers (234Th and 238Th) generally agree to within 30%. The 234Th balance for the upper water column (Buesseleret al., Deep-Sea Research, 39, 1115–1137, 1992) suggests that a substantial portion of the thorium and mass flux is not recorded by the traps. If it is assumed that this flux is carried on more slowly sinking particles (∼50 m day−1) that are not trapped efficiently, and these particles directly interact with the suspended particles pool in the same fashion as the trapped sinking particles, calculation of aggregation and disaggregation rate constants late in the bloom shows a higher value for β2 but a comparable value for β−2 relative to the values determined for the trapped particles. This suggests that the slowly sinking material (e.g. marine snow) is more effective at aggregating small, suspended particles than are the rapidly sinking particles. Temporal increases in β2 and β−2 for the trapped particles are matched by increases in the rate constants for decomposition of particulate organic carbon and nitrogen (2–35 y−1 for C; 4–40 y−1 for N), suggesting that increases in microbial activity are directly reflected in rates of particle aggregation and disaggregation.