Trait-based investigation of phytoplankton communities reveals predictable responses to seasonal environmental variation

Background/Question/Methods

The seasonal pattern of species turnover in temperate phytoplankton communities has inspired much investigation. A mechanistic understanding of these processes is important for applications as varied as lake management and understanding the fundamentals of community assembly. Previous work has identified patterns in species dominance, for example diatoms bloom early and cyanobacteria bloom late in the season, but a mechanistic and trait-based understanding of the environmental drivers of succession has been lacking. We measured 4 physiological traits affecting bottom up as well as top-down processes (maximum growth rate, phosphorus affinity, light affinity, and predation resistance) on 27 phytoplankton species during the course of succession in a northern temperate lake, and examined how the community trait distribution responded to changes in the physical and biotic environment over 6 summers. We asked the following questions: Are the traits of the community predicted by environmental changes? Do communities exhibit tradeoffs in their trait distributions? Are there inter-specific trait tradeoffs, especially in multivariate trait-space? How do the fitness consequences of a given trait combination vary with changing environmental conditions (i.e. how is the selective environment changing)? We investigate these questions using a time series with weekly to bi-weekly species abundance counts and environmental data.

Results/Conclusions

In some, but not all, years, we find strong correlations between abundance-weighted trait distributions and environmental predictors such as phosphorus limitation and grazer abundance. We also observe patterns in community mean traits that mirror species succession: high growth rate communities come first, followed by good alpha competitors, etc., indicating that these traits may capture the mechanism of species succession. The mean traits of communities also exhibit strong negative correlations: for example, a community with high average growth rate has a low phosphorus affinity. These pairwise tradeoffs are not seen in the traits of the species themselves until multiple trait axes are considered. Incorporating all measured traits, we see negative correlations between light affinity, phosphorus affinity, and maximum growth rate emerge. In addition, we observe the growth responses of species using change in abundance between sampling dates as a measure of growth rate and find that for many of the most abundant species the expected relationships between traits hold. This implies that species performance as well as mean community trait can be predicted by environmental conditions. In sum, this study is a step toward developing a predictive and mechanistic model of the responses of phytoplankton community to changes in the environment.