An integrated biological pump model from the euphotic zone to the sediment: a 1-D application in the Northeast tropical Atlantic

Abstract

Abstract A coupled one-dimensional biogeochemical/physical model is developed to follow the organic matter fluxes from the upper ocean to the sea floor. The biogeochemical model is a nitrogen-based seven-compartment model including nutrients, phytoplankton, zooplankton, two pools of dissolved organic matter, and two size classes of detrital material. Particle dynamics are considered through the water column as well as organic matter deposition and mineralization in the superficial sediments. The model is applied at the EUMELI oligotrophic site (21°N, 31°W) where different seasons were sampled in 1991–1992 and sediment trap data collected continuously over the same period. The model, forced with the reanalyzed ECMWF fluxes for these years, reproduces satisfactorily the weak seasonal variability of phytoplankton concentration as well as the exported nitrogen fluxes. Annual primary production (65 g C/m 2 /yr) is sustained mainly by remineralization of DON and zooplankton excretion. Export production at 150 m is ensured by large particles, the DON export contributing only 31% of the total export. The POC export represents 1.3% of the primary production. Including nutrient horizontal advection in the model to mimic any lateral Ekman transfer from the enriched neighboring subtropical gyre (5.5 mmol N/m 2 /yr over the first 150 m estimated from optimization) induces an annual primary production of 73 g C/m 2 /yr, closer to Morel et al.'s (Deep-Sea Res. I 43(8) (1996) 1273–1304) estimate (110 g C/m 2 /yr). Estimated mean carbon fluxes at 1000 and 4400 m depth compare well with sediment trap data, 2 mg C/m 2 /d and 1 mg C/m 2 /d, respectively. Remineralization and disaggregation are the dominant processes below 150 m, aggregation playing a minor role. Observed continuous particulate organic matter fluxes over both years show a more variable evolution than the modeled one. This could be due to mesoscale circulation in the area, or subduction of water masses from the Mauritania upwelling. The modeled seasonal variability of dissolved matter fluxes at the water–sediment interface is very weak, as expected.