Oceanic nitrogen fixation and biogeochemical interactions between the nitrogen, phosphorus and iron cycles have important implications for the control of primary pro- duction and carbon storage in the ocean. The biological process of nitrogen fixation is thought to be particularly important where the ocean is nitrogen limited and oligotrophic. This thesis examines some of the mechanisms responsible for the distribution, rates and temporal variability of nitrogen fixation and its geochemical signature in the modern ocean. I employ simple analytical theories and numerical models of ecosystems and biogeochemical cycles, and closely refer to direct observations of the phytoplanktonic community and geochemical tracers of the marine nitrogen cycle. Time-series observations of geochemical tracers and abundances of nitrogen fixers (or diazotrophs) in the northern subtropical gyres suggest variability in nitrogen fixation on interannual and longer timescales. I use a highly idealized, two-layer model of the nitrogen and phosphorus biogeochemistry and ecology of a subtropical gyre to explore the previously proposed hypothesis that such variability is regulated by an internal biogeochemical oscillator. I find, in certain parameter regimes, self- sustained oscillations in nitrogen fixation, community structure and biogeochemical cycles even with perfectly steady physical forcing. The period of the oscillations is strongly regulated by the exchange rate between the thermocline and mixed-layer waters, suggesting a period of several years to several decades for the North Pacific subtropical gyre regime, but would likely be shorter (only a year or so) for the North Atlantic Ocean.Geochemical tracers such as DINxs (=NO3−16PO3) measure the oceanic departure from the Redfield ratio. DINxs is often used to estimate the rate of nitrogen fixation in the ocean, by quantifying the tracer accumulation along isopycnals. How- ever this tracer reflects an interwoven set of processes including nitrogen fixation, but also denitrification, atmospheric and riverine sources, differential remineralization and complex transport pathways. I examine analytical solutions of the prognostic equa- tion of DINxs and an idealized three-dimensional model of the basin-scale circulation, biogeochemical cycles and ecology of the North Atlantic Ocean. The two approaches demonstrate that the observations of a subsurface maximum in the North Atlantic Ocean and the temporal variability at the station BATS of DINxs can be explained simply by preferential remineralization of organic phosphorus relative to nitrogen. A further analysis reveals that the current geochemical estimates based on inorganic forms of phosphorus and nitrogen underestimate integrated nitrogen fixation rates by a factor of two to six, by neglecting the preferential remineralization effect. Most current understanding of oceanic nitrogen fixation is based on the Tricho-desmium, though unicellular cyanobacteria, diatom-diazotroph associations (DDA) and heterotrophic bacteria might be as important in adding nitrogen into the ocean. I employ a self-assembling global ocean ecosystem model to simulate diverse phyto- planktonic diazotrophs in the global ocean and examine how temperature, oligotrophy, iron and phosphate limitations influence the global marine diazotroph distribution. Analogs of Trichodesmium, unicellular diazotrophs and DDA are successful in the model, showing very similar distributions with observations. The total dia- zotrophic population is distributed over most of the oligotrophic warm (sub)tropical waters in the model. The model demonstrates that temperature is not the primary control, but suggests instead that diazotroph biogeography is restricted to the low fixed nitrogen oceanic regions which have sufficient dissolved iron and phosphate. The theory of resource competition is used to map out regions of iron and phosphate reg- ulation of diazotroph distribution. The theory suggests that diazotrophs are largely regulated by iron availability, in particular in the Pacific and Indian Oceans. The iron cycle is currently not well enough constrained to confidently predict the diazotroph distribution in global ocean models.
|Number of pages||185|
|Publication status||Published - 2009|