Particle export and flux through the Mesopelagic in the high-latitude north and South Atlantic

The biological carbon pump (BCP) is a significant part of the global carbon cycle, exporting ∼10Gt of particulate organic carbon (POC) out of the euphotic zone each year. However, most of the exported POC is remineralized biologically within the upper few hundred metres of the mesopelagic, above the permanent thermocline. Gaining more understanding of the factors controlling the BCP is hence important for understanding and predicting the global carbon cycle better. This thesis investigates the BCP in the Iceland Basin, and during an artificial ocean iron fertilisation experiment in the South Atlantic. In the Iceland Basin, export during a spring diatom bloom was tracked using Lagrangian sediment traps and thorium-234 disequilibria. A large pulse of diatom detritus was exported suddenly at the end of the bloom, probably upon impending Si-limitation. The particles were rich in transparent exopolymer particles (TEP, sticky polysaccharides secreted by phytoplankton), and a comparatively large proportion (20–40%) of the exported POC sank past 750 m. This shows that diatom blooms can produce rapid pulses of particle sedimentation that are transferred efficiently through the mesopelagic, and suggests that aggregation and sinking are mediated by TEP. In contrast, alleviating iron limitation in low silicic acid waters of the South Atlantic with very high copepod grazing pressure only caused a modest phytoplankton response and no enhancement of downward particle flux. This was probably primarily due to grazing control and detritus-feeding by copepods, since diatom growth rates were apparently not strongly Si-limited. This suggests that future Fe-fertilisation experiments must investigate the role of zooplankton thoroughly to distinguish beween bottom-up control of export by nutrient concentrations and top-down control by zooplankton. Export measurements based on thorium-234 disequilibria compared well with net community production measured by O2:Ar ratios over the 39d experiment, suggesting that these two methods can be meaningfully compared over ∼month-long cruises. Further work was conducted with a mesoscale array of four time-series sediment traps deployed for eight moths in the Iceland Basin to study particle flux at 2000m. Large, fast-sinking acantharian cysts contributed up to 48% of POC flux during a specific flux event in early spring, demonstrating that the celestite shells of these protists do not necessarily dissolve in the upper mesopelagic as generally believed. The hypothesis is advanced that deep sinking of acantharian reproductive cysts during spring in this region enables juveniles to feed off seasonally sedimenting phytodetritus in the deep-sea. Finally, the full time-series of particle flux in the four deep traps was analysed. Fluxes peaked in late spring and again in mid-summer. Over the eight months, cumulative mass flux varied by 30% between traps without corresponding variation in the cumulative flux of thorium-230, implying genuine mesoscale variability in bathypelagic particle flux. Moreover, during any one of the two-week collection intervals total mass flux of particles varied 2–16-fold between traps, although it is unclear how much of this short-term variability was due to differences in collection efficiency between traps. Overall, the traps probably under-collected thorium-230 in absolute terms by at least 50%, but this estimate is very uncertain.