Melika Baklouti

How far details are important in ecosystem modelling: the case of multi-limiting nutrients in phytoplankton–zooplankton interactions

We try to answer the question of to what extent details in nutrient uptake and phytoplankton physiology matter for population and community dynamics. To this end, we study how two nutrients interact in limiting phytoplankton growth. A popular formulation uses a product-rule for nutrient uptake, which we compare with that on the basis of synthesizing units. We first fit different nutrient uptake models to a dataset and conclude that the quantitative differences between the models are small. Then we study the sensitivity of phytoplankton growth and zooplankton–phytoplankton interactions (ZPi) models to uptake formulations. Two population models are compared; they are based on different assumptions on the relation between nutrient uptake and phytoplankton growth. We find that the population and community models are sensitive to uptake formulations. According to the uptake formulation used in the ZPi models, qualitative differences can be observed. Indeed, although two models based on functions with similar shapes have close equilibria, these can differ in stability properties. Since stability involves the derivatives of formulas, even if two formulas provide close values, large numerical differences in the stability criterion may occur after derivation. We conclude that mechanistic details can be of importance for community modelling.

Towards a simplification of models using regression trees.

Over-parametrization in modelling is a well-known issue that makes it hard to identify which part of a model is responsible for a given behaviour. In line with that ascertainment, this work presents the outline of an empirical method to simplify models by decreasing the number of parameters. By using regression trees to classify outputs according to related input parameters, the method provides the modeller with an objective tool to reduce the range of the used parameters and, under certain conditions, to establish relations between them. Thereby, the complexity of the model is reduced on the basis of mathematical arguments. As an example, a dynamic energy budget-based model of a mesopelagic bacterial ecosystem is simplified using the presented method. The main benefits of such a method are thus highlighted: (i) more robust parameter estimations; (ii) less complex formulations; and (iii) fewer modelling assumptions. To conclude, the difficulties encountered are discussed, and several solutions are proposed to deal with them.

Biogeochemical fluxes and fate of diazotroph-derived nitrogen in the food web after a phosphate enrichment: modeling of the VAHINE mesocosms experiment

Abstract. The VAHINE mesocosm experiment in the oligotrophic waters of the Noumea lagoon (New Caledonia), where high N2 fixation rates and abundant diazotroph organisms were observed, aimed to assess the role of the nitrogen input through N2 fixation in carbon production and export and to study the fate of diazotroph-derived nitrogen (DDN) throughout the planktonic food web. A 1-D vertical biogeochemical mechanistic model was used in addition to the in situ experiment to enrich our understanding of the dynamics of the planktonic ecosystem and the main biogeochemical carbon (C), nitrogen (N) and phosphate (P) fluxes. The mesocosms were intentionally enriched with  ∼  0.8 µmol L−1 of inorganic P to trigger the development of diazotrophs and amplify biogeochemical fluxes. Two simulations were run, one with and the other without the phosphate enrichment. In the P-enriched simulation, N2 fixation, primary production (PP) and C export increased by 201, 208 and 87 %, respectively, consistent with the trends observed in the mesocosms (+124, +141 and +261 % for N2 fixation, PP and C export, respectively). In total, 5–10 days were necessary to obtain an increase in primary and export productions after the dissolved inorganic phosphate (DIP) enrichment, thereby suggesting that classical methods (short-term microcosms experiments) used to quantify nutrient limitations of primary production may not be relevant. The model enabled us to monitor the fate of fixed N2 by providing the proportion of DDN in each compartment (inorganic and organic) of the model over time. At the end of the simulation (25 days), 43 % of the DDN was found in the non-diazotroph organisms, 33 % in diazotrophs, 16 % in the dissolved organic nitrogen pool, 3 % in the particulate detrital organic pool and 5 % in traps, indicating that N2 fixation was of benefit to non-diazotrophic organisms and contributed to C export.

Modeling of PCB trophic transfer in the Gulf of Lions; 3D coupled model application

3D coupled modeling approach is used for the PCB dispersion assessment in the Gulf of Lion and its transfer to zooplankton via biogeochemical processes. PCB budgets and fluxes between the different species of PCB: dissolved, particulate, biosorbed on plankton, assimilated by zooplankton, which are governed by different processes: adsorption/desorption, bacteria and plankton mortality, zooplankton excretion, grazing, mineralization, volatilization have been estimated. Model outputs were compared with the available in situ data. It was found that the Rhone River outflows play an important role in the organism contamination in the coastal zone, whereas the atmospheric depositions are rather more important in the offshore zones. The transfer of the available contaminant to bacteria and phytoplankton species is mainly related to the biomass present in the water column. Absorption fluxes (grazing) to zooplankton are rather higher than the passive sorption fluxes, which are themselves also linked to the sorption coefficient.