The contribution of various types of settling particles to the flux of organic carbon in the Gulf of St. Lawrence

摘要

The contents of 31 samples from free-drifting sediment traps deployed in the Gulf of St. Lawrence (GSL) were analyzed for the individual contribution of the different types of particles encountered to the total particulate organic carbon (POC) flux. Two trap models were used in 1993-1994: small traps at 50m depth and large traps at 50 and 150m. Total POC fluxes averaged 42mgCm~(-2)d~(-1) for the more reliable large trap and 149mgCm~(-2)d~(-1) for the small trap. The POC fluxes were attributed to different classes of particles based upon microscopically determined particle dimensions and carbon/volume algorithms available in the literature. Fecal pellets, followed by phytoplankton, were the major attributable components, with important contributions by microzooplankton, particularly during the summer of 1994. The mean fluxes for pellets (6 and 60mg Cm~(-2)d~(-1), for the large and small traps, respectively) and phytoplankton (3.2 and 42.9mgCm~(-2)d~(-1)) were in the range of those encountered in other areas of moderate primary productivity. Mean zooplankton carbon fluxes (1.8 and 8.5mgCm~(-2)d~(-1), respectively), however, reflect higher than average zooplankton abundances in the GSL. The C fluxes of specific algal groups confirmed the existence of three trophic regimes previously identified from water column studies and numeric cell fluxes: (1) a period when diatoms were dominant during the spring, (2) a longer interval, which was dominated by dinoflagellates at most others times of the year, and (3) a period of transition during summer. Carbon of animal origin dominated the attributable flux, including an important fraction associated with heterotrophic dinoflagellates. The contribution of marine snow to the total flux (estimated as the difference between the total POC flux and the sum of the attributed components) frequently amounted to more than 60%. The true importance of marine snow remains uncertain, however, because the errors associated with each of the measured components accumulate to produce large uncertainties. The methodological problems involved are discussed.