Variable Stoichiometry Effects on Ocean Model Biogeochemical Cycles

Accurate modeling of major ocean primary producers is important for simulating nutrient cycling and the biological carbon pump. Here we investigate the effect of replacing the canonical Redfield ratios with variable elemental stoichiometries for phytoplankton on nutrient distributions and community composition. Using an ocean biogeochemical model, a linear P:N model (increasing P:N with increasing PO4) was implemented for the ordinary phytoplankton functional group while a nonlinear Si:N model (increasing Si:N with decreasing Fe) was extended to diatoms. Surface NO3 concentrations were increased by the variable P:N model at all latitudes, providing overall better agreement with observed concentrations and a sharper, realistic gradient between 30-70°S. Conversely, surface PO4 concentrations were moderately reduced below observed values almost globally. This also shifted the Southern Ocean (SO) meridional PO4 gradient south by 3°. The subtropical gyres were the few areas that saw very little change. The variable Si:N model improved the representation of the observed SO Si(OH)4 meridional gradient, shifting it south 10° to the correct location but decreased SO concentrations slightly below observed values. In the Bering Sea, previously low Si(OH)4 concentrations were further reduced by ~50%, however, NO3 concentrations were slightly improved by the variable P:N model. The plankton community taxonomic compositions are sensitive to the stoichiometry models. The variable P:N model caused an increase in nitrogen-fixers in the oligotrophic ocean due to the increased PO4 availability there, consistent with previous research. The variable Si:N model enhanced Si(OH)4 uptake relative to NO3 at low Fe concentrations and caused more Si(OH)4 limitation for diatoms, most notably along the Southern Subtropical Front, and induced a shift from diatoms to ordinary phytoplankton. Simulations of the Last Glacial Maximum will be analyzed and the results will be presented.