Causal analysis of phytoplankton development in a small humic lake using structural equation modelling

The factors affecting phytoplankton development in a small boreal, dystrophic lake during summer stratification were explored using structural equation models (SEM). Phosphorus had the highest positive impact on phytoplankton, and higher temperatures also enhanced the biomass. Water colour, and to a lesser extent intense zooplankton grazing, restricted phytoplankton biomass. Grazers generally seemed to be ineffective at controlling phytoplankton, however, which could be partly due to the high abundance of Gonyostomum semen (Raphidophyceae), a large motile algae not readily grazed by zooplankton. The importance of water colour, a significant factor in dystrophic lakes, emerged clearer in SEM than from regression models. SEM proved to be an effective and informative technique for exploring the factors affecting phytoplankton development, the role of each variable, and their interactions. Incorporating Bayesian analysis into the traditional SEM enabled a more detailed examination of variation in the variable estimates and possible sources of uncertainty and provided more reliable error estimates. We used total chlorophyll a as a proxy for total phytoplankton biomass, but the results clearly indicated that some of the emerging questions could have been better addressed by separating different phytoplankton groups. Nevertheless, SEM provided new insights from standard data, and we encourage its further applications in aquatic science.