Seasonal predictability in coastal planktonic communities

Predictive understanding of variability in planktonic communities could improve ecosystem response forecasts. We studied predictability in observations from a nearshore coastal site (Martha’s Vineyard Coastal Observatory) where long-term (>15 y) high-frequency observations with automated flow cytometry and imaging allow many genera and species of phytoplankton and microzooplankton to be separately quantified. Wavelet analysis emphasized that the annual time scale dominates variability for most taxa (>70% display annual periodicity in all years observed). To further examine seasonal predictability, we quantified interannual variability from several angles for each taxon: consistency in the shape of the annual cycle; shifts in seasonal cycle timing; changes in seasonal cycle amplitude; and contributions of constancy versus contingency on time of year. We used a combination of a novel cyclicity index, Colwell’s predictability indices, and a lag-adjusted seasonal model that incorporates seasonally resolved timing and amplitude fluctuations. Our findings reveal significant variability in the degree of predictability and identify the most predictable and least predictable members of the community. We find that shifts in amplitude are particularly important in driving overall variability and that the dynamics of many species or genera are more predictable than their aggregated functional or taxonomic group (e.g., particular diatoms compared to all diatoms combined). Our results underscore the importance of accounting for taxon-specific dynamics when modeling planktonic ecosystems.

Presentation Preference: Oral

Primary Presenter: Miraflor Santos, Woods Hole Oceanographic Institution (msantos@whoi.edu)

Authors:

Miraflor Santos, WHOI  (msantos@whoi.edu)

Andrew Solow, WHOI  (asolow@whoil.edu)

Emily Peacock, WHOI  (epeacock@whoi.edu)

E. Taylor Crockford, WHOI  (ecrockford@whoi.edu)

Heidi Sosik, WHOI  (hsosik@whoi.edu)