Ocean Deoxygenation Conference | Kiel 2018

Biotic processes in biogeochemical models are usually greatly oversimplified in the sense that they are represented by closure terms unrelated to the behaviour of the organisms involved. This applies to both Redfield models of primary production, which ignore the highly variable elemental stoichiometry of phytoplankton and Holling-type models of grazing, which are at odds with observed zooplankton feeding behaviour. Two major difficulties preventing progress have been the complexity of mechanistic biotic process representations and the scarcity of data required to calibrate them.

Optimality-based models of phytoplankton and zooplankton have been developed in the SFB 754, which address the complexity problem by allowing the reproduction of observed organism behaviour with relatively simple formulations. These models have been implemented in a 0–1D modelling framework for mesocosm simulations and also in the (3D) UVic Earth System Model. The data provided by the mesocosm experiments of the SFB 754 greatly facilitate calibration of the optimality-based plankton models and identifying model deficiencies. The resulting model formulation is analysed in the UVic model with the help of novel parameter estimation methods.