Constraining Flows of Nitrogen by Assimilating Data into a Multi-Element Marine Ecosystem Model

We have developed a new, multi-element ecosystem model to simulate a set of batch incubation experiments of natural phytoplankton assemblages. The model divides organisms into classes based on differences in size and function and simulates the flexible-composition of phytoplankton. Nutrient concentrations and plankton community differed among the incubations, allowing us to examine the functioning of the ecosystem by comparison. Data included concentrations of nutrients, organic matter (particulate and dissolved) and plankton (biomass by species). We used a Monte Carlo Markov Chain method to assimilate these data into our model and examined the distributions of simulated values from the ensemble of simulations. The model simulated well the changes in C:N ratio of bulk particulate organic matter (POM), and its difference between experiments. We examined the simulated gross flows of carbon and nitrogen (which cannot be directly measured), dividing the ecosystem between the Microbial Food Web (MFW) and Grazing Food Web (GFW) based on the size of organisms. The MFW dominated the flow of nitrogen in all incubations. The bulk POC:PON ratio varied inversely with the gross amount of nitrogen remineralized in a given incubation. The flexible composition of phytoplankton is a key link between remineralization and the dynamic stoichiometry of POM.