Phytoplankton adaptation in ecosystem models.

We compare two different approaches to model adaptation of phytoplankton through trait value changes. Both consider mutation and selection (MuSe) but differ with respect to the underlying conceptual framework. The first one (MuSe-IBM) explicitly considers a population of individuals that are subject to random mutation during cell division. The second is a deterministic multi-compartment model (MuSe-MCM) that considers numerous genotypes of the population and where mutations are treated as a transfer of biomass between neighboring genotypes (i.e., a diffusion of characteristics in trait space). Focusing on the adaptation of optimal temperature, we show model results for different scenarios: a sudden change in environmental temperature, a seasonal variation and high frequency fluctuations. In addition, we investigate the effect of different shapes of thermal reaction norms as well as the role of alternating growth and resting phases on the adaptation process. For all cases, the differences between MuSe-IBM and MuSe-MCM are found to be negligible. Both models produce a number of well-known and plausible features. While the IBM has the advantage of including more mechanistic (i.e., probabilistic) processes, the MCM is much less computationally demanding and therefore suitable for implementation in three-dimensional ecosystem models.