Bauer B, Sommer U, Gaedke U
divergence, hierarchical level, mesocosms, predictability, replicates, mesocosm, Kiel indoor, Kiel, Germany, 1.4 m3
1. Models aim to predict phytoplankton dynamics based on observed initial conditions and a set ofequations and parameters. However, our knowledge about initial conditions in nature is neverperfect. Thus, if phytoplankton dynamics are sensitive to small variations in initial conditions,they are difficult to predict.2. We used time-series data from indoor mesocosm experiments with natural phyto- andzooplankton communities to quantify the extent to which small initial differences in the species,functional group and community biomass in parallel treatments were amplified or buffered overtime. We compared the differences in dynamics between replicates and among all mesocosms of1 year.3. Temperature-sensitive grazing during the exponential growth phase of phytoplankton causeddivergence. In contrast, negative density dependence caused convergence.4. Mean differences in biomass between replicates were similar for all hierarchical levels. Thisindicates that differences in their initial conditions were amplified to the same extent. Even thoughlarge differences in biomass occasionally occurred between replicates for a short time, dynamicsreturned to the same path at all hierarchical levels. This suggests that internal feedbackmechanisms make the spring development of phytoplankton highly predictable