Modelling the effect of cell-size-dependent nutrient uptake and exudation on phytoplankton size spectra

The effect of phytoplankton cell size on the variation of nutrient uptake and exudation rates is examined: we first present an overview of the relationship between the variation of the growth and loss parameters and cell size. We then investigate the effect of cell-size-dependent parameters on the development of an entire phytoplankton community by means of a numerical, vertically resolved nutrient–phytoplankton model. The model represents phytoplankton size distributions in three different ways, namely one configuration with explicit representation of 14 size classes, one configuration with constant-slope power-law spectral representation, and one configuration with variable-slope power-law spectral representation. The size-dependent configurations are further compared to a size-independent configuration. Consistent with theory, the explicit and variable-slope spectral model simulations predict increased importance of larger cells, or “flat” size distribution under conditions of low light and high nutrients, while smaller cells (“steep” size distributions) may dominate in oligotrophic, well-lit regimes. In some situations the variable-slope spectral model seems to be sufficient to reflect the phytoplankton size distribution; however, especially in the deep phytoplankton maximum a unimodal rather than power-law spectral description might be more appropriate to reproduce results of the explicit 14-size-class model. The assumption of a fixed spectral slope, according to which larger size classes gain importance especially during bloom periods, is not consistent with the underlying theory, and does not agree with the results of the size-discrete model. The comparison of model predictions with variations of phytoplankton size distribution observed in the field is hampered by the sparsity of data, especially for the winter season. A half-saturation constant that represents the nutrient uptake of the entire phytoplankton community ( K * ) compares well to published values.