Identifying resilience mechanisms to recurrent ecosystem perturbations.

The complex nature of ecological systems limits the unambiguous determination of mechanisms that drive resilience to natural disturbance or anthropogenic stress. Using eight-year time series data from boreal lakes with and without bloom formation of an invasive alga (Gonyostomum semen, Raphidophyceae), we studied resilience of phytoplankton communities in relation to recurring bloom impacts. We first characterized phytoplankton community dynamics in both lake types using univariate metrics of community structure (evenness, species richness, biovolume and Simpson diversity). All metrics, except species richness, were substantially altered and showed an inherent stronger variability in bloom lakes relative to reference lakes. We assessed resilience mechanisms using a multivariate time series modelling technique. The models captured clear successional dynamics of the phytoplankton communities in all lakes, whereby different groups of species were substituted sequentially over the ice-free period. The models also identified that G. semen impacts in bloom lakes were only manifested within a single species group, not across species groups, highlighting the rapid renewal of the phytoplankton communities upon bloom collapse. These results provide empirical support of the cross-scale resilience model. Cross-scale resilience could provide an explanation for the paradox that similar species richnesses are seen in bloom-forming lakes and reference lakes despite the clear difference between the community features of the two different sets of lakes investigated.